Trans Fatty Acids

BLUK122-Dijkstra September 28, 2007 19:12 Trans Fatty Acids i BLUK122-Dijkstra September 28, 2007 19:12 Trans ...

Author: Albert J. Dijkstra | Richard J. Hamilton | Wolf Hamm

380 downloads 1804 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

BLUK122-Dijkstra

September 28, 2007

19:12

Trans Fatty Acids

i

BLUK122-Dijkstra

September 28, 2007

19:12

Trans Fatty Acids Edited by

Albert J. Dijkstra Richard J. Hamilton Wolf Hamm

iii

BLUK122-Dijkstra

September 28, 2007

19:12

C 2008 by Blackwell Publishing Ltd

Blackwell Publishing editorial offices: Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK Tel: +44 (0)1865 776868 Blackwell Publishing Professional, 2121 State Avenue, Ames, Iowa 50014-8300, USA Tel: +1 515 292 0140 Blackwell Publishing Asia Pty Ltd, 550 Swanston Street, Carlton, Victoria 3053, Australia Tel: +61 (0)3 8359 1011 The right of the Author to be identified as the Author of this Work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The Publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. First published 2008 by Blackwell Publishing Ltd ISBN: 978-1-4051-5691-2 Library of Congress Cataloging-in-Publication Data Trans fatty acids / edited by Albert J. Dijkstra, Richard J. Hamilton, Wolf Hamm. p. ; cm. Includes bibliographical references and index. ISBN: 978-1-4051-5691-2 (hardback : alk. paper) 1. Trans fatty acids. I. Dijkstra, Albert J. II. Hamilton, R. J. (Richard John) III. Hamm, Wolf. [DNLM: 1. Trans Fatty Acids. QU 90 T774 2007] QP752.T63.T82 2007 612.3 97 – dc22 2007032665 A catalogue record for this title is available from the British Library Set in 10/12 pt Times by Aptara Inc., New Delhi, India Printed and bound in Singapore by COS Printers Pte Ltd The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. For further information on Blackwell Publishing, visit our website: www.blackwellpublishing.com

iv

BLUK122-Dijkstra

September 28, 2007

19:12

Contents

Contributors Preface

1

Fatty acids: structure, occurrence, nomenclature, biosynthesis and properties Richard J. Hamilton 1.1 1.2

1.3 1.4

1.5

1.6 2

Introduction Fatty acid nomenclature 1.2.1 Saturated acids 1.2.2 Monounsaturated acids 1.2.3 Diunsaturated acids 1.2.4 Triunsaturated acids Occurrence Fatty acid biosynthesis 1.4.1 Saturated fatty acids 1.4.2 Monoenoic fatty acids 1.4.3 Polyunsaturated fatty acids Properties of trans fatty acids 1.5.1 Melting points 1.5.2 Ultraviolet spectra 1.5.3 Infrared spectra 1.5.4 Nuclear magnetic resonance spectroscopy Labelling and legislation

ix xi

1 1 2 2 4 7 7 7 12 12 12 14 15 17 18 20 22 23

Trans fatty acids intake: epidemiology and health implications Geok Lin Khor and Norhaizan Mohd Esa

25

2.1 Introduction 2.2 Food sources of trans fatty acids 2.3 Trans fatty acids intake 2.4 Trans fatty acids in human milk 2.5 Trans fatty acids intake and health implications 2.5.1 Coronary heart disease 2.5.2 Diabetes 2.5.3 Cancer 2.6 Concluding remarks

25 26 30 39 40 40 43 44 45

BLUK122-Dijkstra

vi

3

September 28, 2007

19:12

Contents

Conjugated linoleic acid effects on body composition and clinical biomarkers of disease in animals and man: metabolic and cell mechanisms Klaus W.J. Wahle, Marie Goua, Simona D’Urso and Steven D. Heys 3.1 General introduction: conjugated linoleic acids and health 3.2 Structure, dietary origins and consumption of CLAs in man 3.2.1 Structure 3.2.2 Origins of CLAs in the human diet 3.2.3 Dietary consumption of CLAs in man 3.3 CLAs in cancer prevention and treatment 3.3.1 Epidemiology of dietary fats and cancer risk 3.3.2 CLAs and breast cancer 3.3.3 CLAs and prostate cancer 3.3.4 CLAs in gastrointestinal cancer 3.3.5 CLAs and other cancers (hepatic, pancreatic and dermal) 3.4 Cellular mechanisms of CLAs’ anti-cancer effects 3.4.1 Inhibition of angiogenesis 3.4.2 Attenuation of cancer metastasis 3.4.3 Reduction of cancer cachexia 3.5 Effect of CLAs on body composition and energy metabolism in animals and men 3.5.1 Body composition in animals 3.5.2 Body composition in man 3.5.3 Possible mechanisms underlying reported changes in body composition 3.5.4 Efficacy of different CLA isomers in regulating body composition 3.6 Other reported health benefits of CLAs 3.6.1 Effects on insulin resistance and diabetes 3.6.2 Modulation of immune functions 3.6.3 Effects of CLAs on biomarkers of cardiovascular disease 3.7 Reported adverse health effects of CLAs in vivo and in vitro 3.8 Conclusions

4

54 54 55 55 56 59 59 60 60 62 64 66 67 72 73 74 75 75 76 78 78 79 80 81 87 90 91

Analysis of trans mono- and polyunsaturated fatty acids Jean-Louis S´eb´edio and W.M. Nimal Ratnayake

102

4.1 4.2 4.3 4.4

102 102 106 106 106 111 113

Introduction Isomeric fatty acids in the human diet Gas chromatography and Fourier transform infrared spectroscopy Direct GC analysis 4.4.1 Analysis of monounsaturated isomers 4.4.2 Isomers of linoleic and -linolenic acids 4.4.3 Resolution of eicosenoic and -linolenic acid isomers 4.4.4 Effect of the type of carrier gas and flow rate on cis and trans isomer resolution and fatty acid quantification 4.4.5 Conjugated fatty acids 4.5 Silver nitrate thin-layer and high-performance liquid chromatography separation of cis and trans isomers

114 116 123

BLUK122-Dijkstra

September 28, 2007

19:12

Contents

4.6 4.7 5

Controlling physical and chemical properties of fat blends through their triglyceride compositions Albert J. Dijkstra 5.1 5.2 5.3 5.4

5.5 6

4.5.1 Monounsaturated fatty acid isomers 4.5.2 Conjugated fatty acids Utilisation of pre-fractionation steps prior to chromatographic analysis: the case of dairy fats Conclusion

Introduction Defining triglyceride compositions Melting points and sfc The effect of oil processing on triglyceride groups 5.4.1 Hydrogenation 5.4.2 Fractionation 5.4.3 Interesterification 5.4.4 Other oil treatments Using triglyceride groups in product development

vii

123 125 127 128

132 132 133 135 136 136 138 139 141 143

Trans isomer control in hydrogenation of edible oils Annemarie Beers, Rob Ariaansz and Douglas Okonek

147

6.1

147 147 147 147 148 148 149 149 149 149 150 150 151 151 153 157 158 160 160 162 162 163 169 169 175

6.2

6.3

6.4

6.5

6.6

Introduction 6.1.1 Hydrogenation process 6.1.2 History of hydrogenation 6.1.3 Reasons for hydrogenation Isomerisation 6.2.1 Geometric and positional isomerisation 6.2.2 Controlling isomerisation Reaction mechanism 6.3.1 ‘Half-hydrogenated’ intermediate 6.3.2 Saturation, positional and geometric isomerisation Industrial hydrogenation 6.4.1 Batch process 6.4.2 Reactor types and features 6.4.3 Reaction parameters 6.4.4 Influence of feedstock on trans 6.4.5 Influence of reaction conditions on trans 6.4.6 Influence of catalyst on trans 6.4.7 Influence of reactor design on trans 6.4.8 Trans isomer control New developments in low trans hydrogenation 6.5.1 Alternative reaction conditions 6.5.2 Alternative hydrogenation processes 6.5.3 Hydrogenation additives 6.5.4 Alternate catalysts Summary

BLUK122-Dijkstra

viii

7

9

19:12

Contents

Fractionation and interesterification Wim De Greyt and Albert J. Dijkstra

181

7.1 7.2

181 182 182 183 185 185 187 191 191 192 196 198

7.3

8

September 28, 2007

Introduction Fractionation 7.2.1 Historical 7.2.2 Fat crystallisation theory 7.2.3 Fat crystallisation practice 7.2.4 Separation processes 7.2.5 Fractionation products Interesterificaton 7.3.1 Historical 7.3.2 Interesterification mechanism 7.3.3 Interesterification practice 7.3.4 Interesterification products

Food applications of trans fatty acids John Podmore

203

8.1 Introduction 8.2 Margarine 8.2.1 Table margarine 8.2.2 Cake margarine 8.2.3 Pastry margarine 8.3 Biscuit fats 8.3.1 Dough fats – short dough biscuits 8.3.2 Dough fat – laminated biscuits 8.3.3 Cream filling fat 8.4 Fats for chocolate confectionery 8.5 Fats for sugar confectionery 8.6 Vanaspati 8.7 Synthetic creams 8.7.1 Whipped toppings 8.7.2 Coffee whiteners 8.8 Concluding remarks

203 205 205 208 209 210 210 211 211 211 214 215 216 216 216 217

Food products without trans fatty acids Pernille Gerstenberg Kirkeby

219

9.1 Introduction 9.2 Fat phase 9.3 Margarine and related products 9.4 Manufacturing process 9.5 Optimal processing conditions 9.6 Final remarks

219 219 222 225 230 233

Index The colour plate section follows page 228

235

BLUK122-Dijkstra

September 28, 2007

19:12

Contributors

Rob Ariaansz BASF Nederland B.V. De Meern, The Netherlands Dr Annemarie Beers BASF Nederland B.V. De Meern, The Netherlands Dr Wim De Greyt De Smet Technologies & Services Zaventem, Belgium Dr Albert J. Dijkstra Consultant to the Oils and Fats Industry St Eutrope-de-Born, France Simona D’Urso Department of Zootechnological Sciences and Nutrition Frederico II University of Napoli Naples, Italy Dr Norhaizan Mohd Esa Department of Nutrition and Dietetics Faculty of Medicine and Health Sciences Universiti Putra Malaysia Serdang, Malaysia Pernille Gerstenberg Kirkeby Gerstenberg Schroeder A/S Brondby, Denmark Dr Marie Goua The Robert Gordon University School of Life Sciences Aberdeen, UK Professor Richard J. Hamilton Consultant in Oils and Fats Chemistry Merseyside, UK

Wolf Hamm Harpenden, UK Professor Steven D. Heys Department of Surgical and Nutritional Oncology Medical School, Aberdeen University Aberdeen, UK Professor Geok Lin Khor Department of Nutrition and Dietetics Faculty of Medicine and Health Sciences Universiti Putra Malaysia Serdang, Malaysia Douglas Okonek BASF Catalysts LLC Iselin, NJ, USA John Podmore Consultant to the Oils and Fats Industry Liverpool, UK Dr W.M. Nimal Ratnayake Nutrition Research Division Food Directorate Health Products and Food Branch Health Canada Ottawa, Ontario, Canada Professor Jean-Louis S´eb´edio INRA, Unit´e de Nutrition Humaine Mass Spectrometry Platform Saint Gen`es Champanelle, France Professor Klaus W.J. Wahle The Robert Gordon University School of Life Sciences Aberdeen, UK

BLUK122-Dijkstra

September 28, 2007

19:12

Preface

Over the last several decades, a great deal of work has been carried out on trans fatty acids within a number of interrelated fields, such as nutrition, health, food science and industrial processing. In chapters written by leading experts, this volume offers a clear perspective of the current position of trans fatty acids in commerce and academic research. The book is designed as an aid to researchers and professionals in nutrition and health and those providing analytical services to the food industry. Readers seeking ways to formulate oil blends without trans fatty acids and those wishing to alter the composition of oils and fats by means of interesterification, fractionation and hydrogenation will find a large amount of research described and many applications outlined. They will also find methods of adjusting the formulation of their products and their processing. The book is written in a readerfriendly style, which will permit newcomers to the area to grasp the ways in which the field is progressing. Each chapter contains many of the latest references and significant areas of research. Chapter 1 introduces trans fatty acids and puts them in their proper context in relation to the many other saturated and unsaturated acids found in nature. It shows how trans acids are produced both in industry and by natural biohydrogenation in animals and plants. It outlines some of the properties of trans fatty acids and contrasts them with those of their cis isomers. Chapters 2 and 3 deal with the health implications and the epidemiology of trans fatty acids. Intakes of trans acids in the USA, Central America, Nigeria, Iran, India, China, Hong Kong, New Zealand, Australia, Hungary, the Czech Republic, Poland, Bulgaria and Spain are considered. A number of studies, e.g. the Zutphen Elderly, the Scottish Heart and Health, the Seven Countries, the Nurses Health and the EURAMIC case control, have been referenced. The results are considered without bias and the authors are not afraid to point out where further research is needed to confirm the original conclusions. In the consideration of conjugated linoleic acid (Chapter 3), the authors discuss the synthetic products that can be used for human and animal supplementation. With the realization that not all trans fatty acids have the same biological effect came the realization that it is important to know more than just the total trans fatty acid content in a food. Chapter 4 explains how direct GC, GC-MS, AgNO3 -TLC and HPLC are used to determine the trans fatty acid composition of various food products. Some recommendations for the best analysis of cis/trans monounsaturated fatty acids are given, whilst the need for pre-fractionation in some instances is highlighted. Chapter 5 introduces the concept of triglyceride groups and demonstrates that, for most components used in constituting fat blends, the triglyceride group composition can be calculated. It illustrates how hydrogenation, fractionation and interesterification cause the triglyceride group composition to change and it also highlights how product development can be facilitated by specifying fat blends on the basis of their triglyceride group composition. Chapter 6 provides a wide survey of the industrial hydrogenation process in the field of edible oil processing, its mechanism and how process parameters affect the trans fatty acid

BLUK122-Dijkstra

xii

September 28, 2007

19:12

Preface

content of the hydrogenation product. In addition, the chapter provides insight into the latest developments comprising the use of catalysts other than the usual nickel, the use of additives and unconventional process conditions and their effect on trans fatty acid formation. Chapter 7 deals with major modification processes that do not alter the trans fatty acid content of the oil or fat being processed: fractionation and interesterification. After outlining the basics, examples are provided that illustrate what kind of products suitable for low-trans or trans-free oil blends can be arrived at by these modification processes. The last two chapters are food product oriented and cover both compositional and processing aspects. Chapter 8 reviews products such as various margarines, dough fats and shortenings that still contain trans fatty acids. In Chapter 9, their low-trans or trans-free equivalents are introduced and their processing requirements highlighted. Solutions are provided for dealing with slow crystallisation of low-trans fat blends, which cover both the use of adjuvants and the adaptation of the cooling equipment and its process conditions. We express our thanks to the authors for their excellent contributions that provide fresh insight into this interesting and exciting field of study. We are indebted to our friends and colleagues for their helpful comments and criticisms. Finally, we are grateful for all the help we have received from Blackwell Publishing. Albert J. Dijkstra Richard J. Hamilton Wolf Hamm

BLUK122-Dijkstra

September 25, 2007

1

19:41

Fatty acids: structure, occurrence, nomenclature, biosynthesis and properties

Richard J. Hamilton

1.1

INTRODUCTION

Trans fatty acids have been present in the Western diet for as long as milk and butter have been staple commodities. However, in the last century with the discovery of catalytic hydrogenation by Sabatier and Senderens (Hastert, 1996; Hoffmann, 1989), food technologists came to recognise the improved physical characteristics which trans fatty acids could bestow on food products. The protection of the foodstuffs from the off flavours, which developed when highly unsaturated oils were incorporated into foods, was an added advantage which hydrogenation gave. However, in the last 50 years, studies have been conducted into the effects of increased quantities of trans fatty acids on human health and nutrition. The result has been the requirement for food processors to be able to claim that they have low or no trans fatty acids in their products (Korver and Katan, 2006). To appreciate the reason for this changed consideration, we first need to look at the constituents of oils and fats. As far as the world production is concerned, the major vegetable oils and fats are soya, palm, rape (canola), sunflower, cotton, groundnut, coconut, palm kernel and corn. The major animal fats, by comparison, are butter, tallow, lard and fish. During the year 2005, the production split between the animal and vegetable groups of oils and fats was 78.5% vegetable oils and 21.5% animal fats. In Chapters 8 and 9 on applications, we will see how the two sources of oils and fats are utilised. Oils and fats are made up of:

r r r r

lipids, viz. triacylglycerols (also called triglycerides), diacylglycerols (diglycerides), waxes, phosphoglycerols, sphingolipids, free fatty acids and hydrocarbons; certain vitamins; pigments and antioxidants.

These lipids cover a wide range of different chemical structures but there are two common features. Most lipids are water insoluble and they can all be biosynthetically related to fatty acids. The triacylglycerols account for 90–95% by weight of oils and fats and in many senses are the most important part of these items of commerce. A generalised formula for a triacylglycerol is shown in Fig. 1.1.

BLUK122-Dijkstra

2

September 25, 2007

19:41

Trans Fatty Acids O

H2C

O

R2

C

O

CH

H2C Fig. 1.1

O

C

R1

O

O

C

R3

General formula for a triacylglycerol.

If the fatty acids in this triacylglycerol, R1 COOH, R2 COOH and R3 COOH, are all identical, i.e. R1 = R2 = R3 , the triacylglycerol would be referred to as a monoacid triacylglycerol or a single-acid triacylglycerol. More usually, each triacylglycerol will have two or three different fatty acids. Gunstone (1967) claimed that over 300 fatty acids were known in nature. By the time of a more recent book in 1996, he estimated that there were over 1000 fatty acids (Gunstone, 1996). Thus the diversity of these oils and fats (Gunstone, 2004) is considerable as will be manifested in Chapter 4 on analysis. One simplifying feature is that the major fatty acids, in nature, have an even number of carbon atoms. In addition, there are usually only five to seven major fatty acids in most commercially important oils and fats.

1.2

FATTY ACID NOMENCLATURE

Fatty acid nomenclature is complicated by the fact that many acids were well known before any system of naming them had been determined. Thus the names of oleic, stearic and palmitic acids were well established before any rules were developed.

1.2.1

Saturated acids

Fatty acids are named according to the number of carbon atoms in the chain. In turn, the name of the fatty acids refers back to the name of the saturated hydrocarbon with the same number of carbon atoms. So stearic acid has 18 carbon atoms and is related to the alkane with 18 carbon atoms, i.e. octadecane. To obtain the name of the acid, the ‘e’ is removed from octadecane giving ‘octadecan’ and the ending ‘oic’ is added to indicate the carboxylic acid. Thus, octadecan(e) → octadecan(oic) acid → octadecanoic acid, which is the full and correct name for stearic acid. Whilst it is convenient to use the trivial names, such as oleic and linoleic acid, many of the acids encountered later in our discussions have no simple trivial names. Even the use of formulae, as given in Tables 1.1 and 1.2, is not very quick and easy. An alternative shorthand method has been devised. This system reduces the acid to the minimum statement that is needed to define it.

Chain length 4

6

8

10

12

14

16

18 20 22

4:00

6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00

Docosanoic

Eicosanoic

Octadecanoic

Hexadecanoic

Tetradecanoic

Dodecanoic

Decanoic

Behenic

Arachidic

Stearic

Palmitic

Myristic

Lauric

Capric

Caprylic

Caproic

Butyric

Common name

C H2

C H2

C H2

C H2

C H2

C H2

C H2

C H2

H2 C

H2 C

H2 C

H2 C

H2 C

H2 C

H2 C

H2 C

CH3 (CH2)20 COOH

CH3 (CH2)18 COOH

H3C

H3C

H3C

H3C

H3C

H3C

H3C

H3C

Structure

C H2

C H2

C H2

C H2

C H2

C H2

C H2

H2 C

H2 C

H2 C

H2 C

H2 C

H2 C

H2 C

COOH

C H2

C H2

C H2

C H2

C H2

C H2

H2 C

H2 C

H2 C

H2 C

H2 C

H2 C

COOH

C H2

C H2

C H2

C H2

C H2

H2 C

H2 C

H2 C

H2 C

H2 C

COOH

C H2

C H2

C H2

C H2

H2 C

H2 C

H2 C

H2 C

COOH

C H2

C H2

C H2

H2 C

H2 C

H2 C

COOH

C H2

C H2 H2 C

H2 C

COOH

C H2

H2 C

COOH

COOH

September 25, 2007

Octanoic

Hexanoic

Butanoic

Proper name

Structures of saturated fatty acids.

Shorthand notation

Table 1.1

BLUK122-Dijkstra 19:41

Structure, occurrence, nomenclature, biosynthesis and properties 3

BLUK122-Dijkstra

4

September 25, 2007

19:41

Trans Fatty Acids

In the case of stearic acid, first the total number of carbon atoms in the chain is stated, i.e. 18, and then the number of double bonds is given which, in the case of stearic acid, is 0. The shorthand system then inserts a colon between the number of carbon atoms and the number of double bonds and so, for stearic acid, the shorthand is 18:0. Stearic acid is shown in Table 1.1, where the long straight chain is given by the zigzag representation. Some of the main straight-chain saturated fatty acids are also given in Table 1.1.

1.2.2

Monounsaturated acids

Oleic acid is an unsaturated fatty acid that can be represented by the formula shown in Fig. 1.2. Thus oleic acid has 18 carbon atoms, and it has one double bond at position 9 from the carboxyl end. Since oleic acid has 18 carbon atoms and one ethylenic double bond, the name is based on octadecene. In this instance the ‘e’ is removed and the ending for the carboxylic acid group octadecen(e) → octadecen(oic) acid. Thus is added 9-octadecenoic acid. In the case of oleic acid, the double bond is in the cis configuration (also called the Z configuration from the German zusammen, meaning together). Thus to specify oleic acid precisely, the full name would be 9c-octadecenoic acid or 9Z -octadecenoic acid. An isomer of oleic acid is elaidic acid, which has a trans double bond at the 9-position. The shorthand for this acid would therefore be 9t-octadecenoic acid. If the EZ system is to be used, the letter referring to the trans configuration is E, which stands for the German word entgegen, meaning opposite. These two acids are shown in Fig. 1.2. From a chemist’s point of view, the most important part of a fatty acid is the carboxylic acid group. The position of the double bond is therefore quoted with reference to the carboxylic acid group, i.e. 9 in the case of oleic acid. Using the shorthand method oleic acid is 18:1. Since the double bond is at the ninth carbon atom and the configuration of the double bond is cis, the name becomes 9c-18:1. It is also possible to denote the position of the double bond by using the symbol . Oleic acid is described as a 9 acid, whilst petroselinic acid is a 6 acid. Some of the main monounsaturated fatty acids are given in Table 1.2.

O C H Oleic acid

9c-octadecenoic acid

H or

OH

9Z-octadecenoic acid

O C Elaidic acid Fig. 1.2

H H 9t-octadecenoic acid or

Structures of oleic and elaidic acids.

9E-octadecenoic acid

OH

14 14 16 18 18 18 18 22

14:1 5c

14:1 9c

16:1 9c

18:1 6c

18:1 9c

18:1 9t

18:1 11c

22:1 13c

Erucic

Oleic

Petroselenic

Palmitoleic

13c-Docosenoic

Erucic

11c-Octadecenoic Vaccenic acid

9t-Octadecenoic

9c-Octadecenoic

6c-Octadecenoic

9c-Hexadecenoic

Myristoleic

H3C

HC

CH

CH3 (CH2)7 CH=CH (CH2)11 COOH

CH3 (CH2)5 CH=CH (CH2)9 COOH

CH3 (CH2)7 CH=CH (CH2)7 COOH

CH3 (CH2)7 CH=CH (CH2)7 COOH

CH3 (CH2)10 CH=CH (CH2)4 COOH

CH3 (CH2)5 CH=CH (CH2)7 COOH

CH3 (CH2)3 CH=CH (CH2)7 COOH

CH3 (CH2)7 CH=CH (CH2)3 COOH

Structure

C OH

O

September 25, 2007

9c-Tetradecenoic

5c-Tetradecenoic

Common name

Structures of monoenoic acids.

Shorthand Chain Proper notation length name

Table 1.2

BLUK122-Dijkstra 19:41

Structure, occurrence, nomenclature, biosynthesis and properties 5

COOH

-Linolenic acid

-Linolenic acid

Rumenic acid

6c,9c,12c-18:3

9c,12c,15c-18:3

9c,11t-18:2

COOH

September 25, 2007

COOH

COOH

Linoleic acid

9c,12c-18:2

Structure

Name

Shorthand notation

Structures of polyunsaturated acids.

6

Table 1.3

BLUK122-Dijkstra 19:41

Trans Fatty Acids

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties

1.2.3

7

Diunsaturated acids

Linoleic acid is a diunsaturated acid with two double bonds and 18 carbon atoms and is named from the diunsaturated hydrocarbon octadecadiene (Table 1.3). Octadecadien(e) → octadecadien(oic) acid → 9, 12-octadecadienoic acid Again, the stereochemistry of the double bonds is known to be cis and so the correct name for linoleic acid is 9c,12c-octadecadienoic acid, with its shorthand name 9c,12c-18:2. There is another system of numbering the position of the double bond, which came into operation because of the way in which the fatty acid is built up during biosynthesis. In Section 1.4, it will be seen that the starting point for biosynthesis is a two-carbon unit that becomes the methyl end of the final fatty acid. Each time that another two-carbon unit is added to the chain, the name of the new fatty acid would alter and the position of any double bond would also alter with respect to the chemist’s fixed point, i.e. numbering from the carboxyl group. It was recognised that it might be advisable to use a system of nomenclature, which started at the methyl end of the acid chain. This is called the n-x system or the system. Thus linoleic acid is 9c,12c-18:2, where the carboxyl group is the starting point for the numbering. The alternative name for linoleic acid starts the numbering at the methyl end. In this case the double bond is now of six carbon atoms from the methyl group, and the position of the double bond is represented as n-6 or 6. The tells us that we start counting from the methyl end. Linoleic acid would be described as 6,9-18:2 in this alternative system. The notation for monounsaturated acids is given in Table 1.2. Rumenic acid is a conjugated diene fatty acid, 9c,11t-18:2, which is dealt with in Chapter 3.

1.2.4

Triunsaturated acids

The structures of two of the major triunsaturated acids -linolenic acid and -linolenic acid are given in Table 1.3. Their full names are 6c,9c,12c-octadecatrienoic acid and 9c,12c,15coctadecatrienoic acid respectively. The name derived as above from octadeca with the trienoic added shows that there are three ethylenic double bonds. Octadeca(ne) → octadecatrienoic acid → 9c,12c,15c-octadecatrienoic acid

1.3

OCCURRENCE

Of the saturated fatty acids, palmitic acid is the most widely occurring in both animal fats and vegetable oils, whilst stearic acid is found in lesser quantities in vegetable oils. Stearic acid is present in large quantities only in animal tallows and in vegetable fats, such as cacao butter and Borneo tallow. Butyric acid is found in butterfat (also referred to as anhydrous milk fat) produced from cow’s milk. Caprylic, capric and myristic acids are present in coconut and palm kernel oil. Oleic acid is the most widely distributed monounsaturated fatty acid. In some oils it is found in high proportions, ranging from 50 to 80%, e.g. olive, cashew and pistachio.

BLUK122-Dijkstra

8

September 25, 2007

19:41

Trans Fatty Acids 40 35

Percentage

30 25 C16

20

C18

15 10 5 0 5

6

7

8

9

10

11

12

13

14

15

16

Position of double bond Fig. 1.3

Trans isomeric monoene C16 and C18 fatty acids in butter.

Whereas most of the unsaturated fatty acids in nature have a cis double bond, there are some acids that have the trans configuration. We can concern ourselves mainly with trans fatty acids from now on. There are three main sources of trans fatty acids in the human diet; viz., they can be derived from animals or from the plant kingdom, or produced in the processing of oils and fats. In animals, trans fatty acids are derived from dietary lipids. It is believed that biohydrogenation by bacteria in the rumen of the dietary lipids results in a mixture of trans fatty acids. Such fatty acids are found in all ruminant milk fats. Rumenic acid (9c,11t-18:2) is the major conjugated fatty acid in ruminant fats. Rossell (2001) reported the trans content of subcutaneous adipose tissue in beef, sheep and pig to be 1.3–6.6%, 11.0–14.6% and 1.1–1.4% by weight respectively. In the case of farm animals, where the feed may contain trans fatty acids, the animal will metabolise some of the trans fatty acids and place some trans fatty acids in the adipose tissue. Hay and Morrison (1970) showed that amongst the trans isomers in butterfat, the monenoic C16 and monoenoic C18 are the major components (Fig. 1.3). The major isomer for C16 is palmitelaidic acid 9 (32%) and for C18 trans vaccenic acid 11 (36.1%). Trans fatty acids in most vegetable oils are present, if at all, in very minor proportions and in some oils, at the trace level. In the vegetable kingdom, trans fatty acids do occur naturally and sometimes in significant quantities; i.e. there is 6–12% of eleostearic acid 9c,11t,13t-18:3 in cherry oils, which have now been accepted as safe for food oils (Comes et al., 1992). Petroselaidic acid, 6t-18:1, is found along with petroselinic acid in Heracleum nipponicum, Conium maculatum, Phelopterus litoralis, Ligusticum acutifolium, Bupleurum falcatum, Osmorhiza aristata, Conioselinum univittatum, Hedera japonica, Panax schinseng and Aralia elata (Placek, 1963). In the plant kingdom, conjugated triene fatty acids often have one or more trans double bonds, e.g. jacaric acid 8c,10t,12c-18:3, calendic acid 8t,10t,12c-18:3, catalpic acid 9t,11t,13c-18:3, punicic acid 9c,11t,13c-18:3 and -eleostearic acid 9t,11t,13t-18:3. There are also conjugated tetraenoic acids - and -parinaric acids 9c,11t,13t,15c-18:4 and 9t,11t,13t,15t-18:4 respectively. In addition the biosynthetic pathways given in Section 1.3

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties

9

involve trans double bonds even when cis double bonds are being generated. So it can be seen that trans fatty acids occur naturally in both animals and plants. In 1983, Sommerfeld stated that ‘hardened oils do NOT contain trans fatty acids isomers other than those produced by the microflora of ruminants. Therefore claims that trans fatty acids isomers are “synthetic” “non-physiologic” or “unnatural” are unjustified if these words are used to imply “not produced by the living organism” ’. The presence in nature of conjugated linoleic acid double bonds contains trans which is further confirmation of Sommerfeld’s statement. Conjugated linoleic acid is covered in full in Chapter 3. The third source of trans fatty acids in foods is where they are produced in processing. Why was hydrogenation introduced into the oils and fats industry? Initially, it was to remedy a shortage of solid fats. At its simplest, hydrogenation is the addition of two hydrogen atoms across the ethylenic double bond of the fatty acid. It was recognised that the more unsaturated the fatty acid, the more likely it was for the fatty acid to be oxidised, which leads to oxidative rancidity. By removing two double bonds from linolenic acid, a monoenoic acid would be formed, which would resist oxidation better. Triene → diene → monoene → saturated acid

(1.1)

If the hydrogenation could proceed by the route suggested by Equation 1.1, the triene linolenic acid would yield the saturated acid, i.e. stearic acid. However, under industrial conditions, hydrogenation with a nickel catalyst is partial, giving rise to a mixture of products. From the above, it is still not obvious why there should be any trans fatty acid formed. Dijkstra (2002) suggested an amendment to the Horiuti–Polanyi mechanism in which the monoene M forms a semihydrogenated intermediate MH (Eq. 1.2). M + H → MH,

where M represents monoene

(1.2)

Dijkstra explains that the hydrogen concentration is too low for these intermediates to go on to form stearic acid. In turn this allows dissociation to occur as in Eq. 1.3. MH → M + H

(1.3)

When an individual acid, e.g. oleic acid, is considered in these reactions, the changes can be represented as shown in Fig. 1.4. When a fatty acid with a single cis double bond is partially hydrogenated, adsorption to (Step 1) and desorption from (Steps 3–5) the catalyst surface occurs, which produces a mixture of fatty acids. Some of the acids have a trans double bond. It is believed that the adsorption mechanism (Step 1) involves the formation of carbon nickel bonds between the metal catalyst and the carbon atoms of the 9 double bond C9 and C10 to give a structure (a). One hydrogen atom is then transferred (Step 2), probably from a Ni–H atom on the surface of the catalyst near the adsorbed fatty acid, to the carbon atom C9 to give structure (b). If the addition goes further, another hydrogen atom is added, the C Ni bond is broken and the hydrogen adds to C10 , with the formation of stearic acid as a desorption (Step 3).

BLUK122-Dijkstra

10

September 25, 2007

19:41

Trans Fatty Acids H2 C

R

10

H2 C

9

R1

1

H2 10 C

R

H

H

H

Ni

cis 9

H2 C

9

Ni

R1

H

(a)

2 H2 C 10

R

H2 C

9

H2 C

R1 R

H2 C

R1

3 H

Ni

H

H

H

(b)

H H Stearic acid

H

4 R

H

H2C 11

10

R1 R

H

(c)

H

Fig. 1.4

11

H

10 H2C (f) trans 10

R

R1

C H2 +

trans 9

+

R

H

9

H

(e) cis 10

H2 10 C

H2 C

10

9

R1

H

H R1

H2 C

(d) cis 9

Partial hydrogenation of oleic acid, where R1 = (CH2 )7 COOH and R2 = CH3 (CH2 )7 .

The interactions between these species are reversible (Fig. 1.4), with the result that in structure (b) the hydrogen from C9 and the C Ni bond can be eliminated to reform a double bond between C9 and C10 (Step 4). This results in a mixture of cis and trans isomers (c) and (d). It is also possible that when the C Ni bond breaks, the hydrogen which is involved in the elimination comes from the C11 . This will produce a mixture of cis and trans 10 monoenes (e) and (f). The production of the trans isomers can be seen more easily in Fig. 1.5. The cis monoene, in an addition reaction (Step 1), gives the intermediate with two Ni atoms (a). Structure (a) can now react with a hydrogen on a neighbouring nickel atom, as explained above, to form the semihydrogenated intermediate (b) (Step 2); this is the structure MH in Equation 1.2. The elimination of the Ni–H atoms from (b) with the breaking of a C H bond and a Ni H bond results in the reformation of the cis double bond between C9 and C10 (Step 3). This step is a desorption from the catalyst surface.

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties 1R

1

R

H

C R

1

R

1

H 1

H cis

C

H Ni

C

(b)

H 3

(b) R

1

H

C R

R

R

H

2

H

C

H Ni

C

11

R

H

C

H Ni

C

R

H

Ni (a)

cis

2

1R

C R

1R

H

C

4 H

H Ni

C

C

H

H

R Ni

H (b1)

(b)

5

1R

H

H R trans

Fig. 1.5 Partial hydrogenation of oleic acid, where R1 = (CH2 )COOH and R2 = CH3 (CH2 ), showing free rotation at Step 4.

However the semihydrogenated intermediate (b) can undergo free rotation about the C9 to C10 bond (Step 4) when the fatty acid is attached to the metal atom at just one point to give the conformation (b1). When the Ni–H atoms from (b1) are eliminated in a desorption (Step 5), it is a trans isomer which is formed. The oils and fats industry and catalyst manufacturers are working to permit hydrogenation without the production of trans fatty acids (see Chapter 6).

BLUK122-Dijkstra

12

September 25, 2007

19:41

Trans Fatty Acids

As will be explained later in Chapters 5, 7 and 9, industrial processors have reduced the level of trans fatty acid in foodstuffs from this source. In processing, thermally induced geometrical isomerisation can occur as described in Chapter 5.

1.4 1.4.1

FATTY ACID BIOSYNTHESIS Saturated fatty acids

When the pathways for fatty acid synthesis were being elucidated, it was realised that fatty acids were built up from acetic acid units. This finding made it easy to understand why so many naturally occurring acids had an even number of carbon atoms. It was subsequently found that only two carbon atoms of palmitic acid came directly from acetic acid, the carbons at positions 15 and 16 from the carboxyl end. The remainder of the carbon atoms came from malonyl coenzyme (CoA), as in Equation 1.4 (Gurr and James, 1980). CH3 COOH + 7CH2 (COOH)2 → CH3 (CH2 )14 COOH

(1.4)

The synthesis of palmitic acid is carried out by fatty acid synthetase (FAS). In the case of FAS in Escherichia coli, the enzymes involved in individual steps are shown in Fig. 1.6. Malonyl CoA:ACP transacylase (a) activates the malonyl unit, whilst acetyl CoA:ACP transacylase (b) activates the acetic acid unit. The joining of these two activated forms to form a C4 unit is catalysed by 3-ketoacyl-ACP synthetase (c). The ketone group in this C4 unit is then reduced to a hydroxyl group in the presence of 3-ketoacyl-ACP reductase (d). It is in the next step that we see the formation of a trans fatty acid derivative as the hydroxyl group, and a hydrogen is eliminated in the presence of 3-hydroxyacyl-ACP dehydrase (e). The trans fatty acid is not released as such but the double bond is reduced to give a new C4 fatty acid still in the activated form of ACP under the influence of enoyl-ACP reductase (f). These same reactions are performed to convert the C4 up to the normal C16 palmitic acid, with the addition of further six malonyl units. This is the natural end point of this series of reactions, and longer chain length fatty acids depend on elongation reactions using FAS III. From the viewpoint of trans fatty acids, it is important to recognise that the system accepts and utilises trans acids; i.e. trans fatty acids are not unnatural. There are in fact three fatty acid synthetases FAS I, II and III. Type I consists of large molecular mass multifunctional proteins containing covalently bonded acyl carrier proteins (ACP) and is found in animals. Type II consists of individual enzymes that normally act as one complex and are found in bacteria and plants (Gunstone et al., 1994). Type III FAS can elongate already formed fatty acids. The differences in these three synthetases relate to the subunits and the sequence of domains.

1.4.2

Monoenoic fatty acids

In the case of monoenoic acids in E. coli formed by type II FAS, there is a branch point when the chain length reaches ten carbon atoms, in an anaerobic pathway. In Fig. 1.7, it can be seen that when the dehydrase enzyme has worked on -hydroxydecenoyl ACP, the resulting trans2-decenoyl ACP can be elongated as normal and finish as palmitic acid. Alternatively, the dehydration step can lead to cis-3-decanoyl ACP which is then elongated to 9-palmitoleoyl

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties ACP-SH

CoA-SH

O

O H2 C

COOH

13

S

C

H2 C

COOH

CoA

S

C

ACP

(a) ACP-SH

CoA-SH O

O H3C

C

S

H3C

CoA

C

S

ACP

(b) O H3C

C

S

ACP

+ ACP-SH

CO2

O

O COOH

H2 C

S

C

C

H 3C

ACP

O H2 C

C

S

ACP

S

ACP

(c) NADPH + H +

O

O H3C

C

NADP +

H2 C

OH

C

S

ACP

H3C

C H

(d) OH H3C

O H2 C

C H

C

H3C S

ACP

C

H C

(e)

O H2 C

+ H2O

C C

H

S

ACP

O NADPH + H

H3C

O

C

H3C C

S

ACP

O Fig. 1.6

NADP

+

H C

H

+

Partial biosynthetic reactions of FAS.

(f )

H2 C

H2 C

C

S

ACP

BLUK122-Dijkstra

14

September 25, 2007

19:41

Trans Fatty Acids ACP

OH

3-Hydroxydecanoyl ACP

O

O H

H

H H CP A

O

CP A

trans-2-Decenoyl ACP

cis-3-Decenoyl ACP

I or II

I

16:0

9-16:1 II

18:0

11-18:1

Fig. 1.7 Anaerobic pathway of fatty acid biosynthesis in bacteria showing the branching point in the formation of cis vaccenic acid, where I is 3-ketoacyl-ACP synthetase I, and II is 3-ketoacyl-ACP synthetase II.

ACP and then to 11-cis-vaccenyl ACP (the Lipid Library, www.lipidlibrary.co.uk). This results in the n-7 double bond being retained. There are aerobic desaturases that remove two hydrogens from a saturated acyl chain stereospecifically. This is the system which is common in all organisms, where oxygen and a reducing cofactor are needed. The first double bond is usually produced at the 9 carbon atom catalysed by stearoyl–CoA 9 desaturase in plants and algae with the production of oleic acid. Palmitoleic acid is derived in a similar way from palmitic acid. Bacteria, uniquely, are able to produce 10 monoenoic acids. In addition some bacteria can remove a second pair of hydrogen atoms, giving rise to a diunsaturated double bond system, which is not a methylene interrupted one as is found in plants.

1.4.3

Polyunsaturated fatty acids

Mammals can introduce a second double bond to a monounsaturated fatty acid chain, but usually this new double bond cannot be inserted towards the methyl end of the chain.

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties

15

Oleic acid COOH

−2H

COOH

+C2

COOH −2H

COOH Oleic acid family Fig. 1.8

Biosynthetic relationships in the oleic acid family of fatty acids.

A second double bond can be introduced into oleic acid 9c-18:1, with the formation of 6c,9c-18:2 (Fig. 1.8). Addition of a further two carbon atoms under the influence of an elongase converts the C18 chain length to 8c,11c-20:2. If this acid is desaturated, it produces 5c,8c,11c-20:3. Further elongation and desaturation steps will convert this acid to other acids in the oleic acid family which all have the first double bond from the methyl end at the 9-position, i.e n-9 or 9. By contrast, when linoleic acid is subjected to a similar set of desaturation and elongation, the acids that are produced are shown in Fig. 1.9. Linoleic acid is converted to -linolenic acid, i.e. 6c,9c,12c-18:3, which is elongated to 8c,11c,14c-20:3. Further desaturation gives arachidonic acid, i.e. 5c,8c,11c,14c-20:4. Again, it is easy to see in Fig. 1.9 that all these acids in the linoleic acid family have the first double bond from the methyl end at the 6-position, i.e. n-6 or 6. Starting from -linolenic acid, desaturation yields 6c,9c,12c,15c-18:4. Elongation of this acid gives 8c,11c,14c,17c-20:4, and a further desaturation leads to 5c,8c,11c,14c,17c-20:5. This is called eicosapentaenoic acid (67A) which gives the next member of the series, docosapentaenoic acid. This acid in turn loses 2 hydrogen atoms to give docosahexaenoic acid (DHA). Both EPA and DHA are found in fish oils. All of these acids in the linolenic acid family (Fig. 1.10) have the first double bond from the methyl end at the 3-position, i.e. n-3 or 3. These pathways represent the formation of three different families of polyunsaturated fatty acids.

1.5

PROPERTIES OF TRANS FATTY ACIDS

The physical properties of trans fatty acids are different from the corresponding cis isomers.

BLUK122-Dijkstra

16

September 25, 2007

19:41

Trans Fatty Acids Linoleic acid COOH

−2H

CO2H +CH2

CO2H

−2H

CO2H Arachidonic acid Linoleic acid family n-6

Fig. 1.9

Biosynthetic relationships in the linoleic acid (n-6) family of fatty acids. Linolenic acid CO2H

−2H

CO2H +C2

CO2H −2H

COOH Linolenic acid family

Fig. 1.10

n-3

Biosynthetic relationships in the linolenic acid (n-3) family of fatty acids.

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties

17

Table 1.4 Melting points of monoacid triacylglycerols with C18 monounsaturated acids.

1.5.1

Position of the double bond

Cis isomers

Trans isomers

4 5 6 7 8 9 10 11 12 13 14 15

34 14 28 7 24 5 27 10 32 26 44 43

53 41 52 39 49 41 49 43 51 44 58 56

Melting points

In Table 1.4 and Fig. 1.11 we can see that the melting points of the two series are very different with the trans isomers having the higher melting points (Hagemann et al., 1975). It was this higher melting behaviour which made the trans fatty acids so valuable in commerce. These melting characteristics made it possible to produce the desirable properties for a plastic shortening by hydrogenation of cottonseed oil (and subsequently of other oils, e.g. soya bean oil). The polymorphism of trans fatty acid triacylglycerols is indicated in Table 1.5 (Hagemann et al., 1975).

70 60

Melting point

50 40 30 20 10 0

cis trans 4

5

6

7

8

9

10

11

12

13

14

15

DB position

Fig. 1.11 Melting points of C18 triacylglycerols with differing positions of the double bond. DB, double bond; MP, melting point. (Adapted from Hagemann et al., 1975.)

BLUK122-Dijkstra

18

September 25, 2007

19:41

Trans Fatty Acids Table 1.5 Melting points (◦ C) of polymorphs of monoacid triacylglycerols with C18 monounsaturated acids. Position of double bond

Polymorph

Polymorph

5 27 −36 16 −34 15 −20

34 53 5 39 5 41 27 49 32 51 26 44

cis 4 trans 4 cis 7 trans 7 cis 9 trans 9 cis 10 trans 10 cis 12 trans 12 cis 13 trans 13

21 23

Adapted from Hagemann et al., 1975.

These differences in melting point are attributable to the different shapes of the trans fatty acids in comparison with the cis isomers. In Fig. 1.12a, we can see the straight chain of the trans isomer elaidic acid that hardly alters the overall shape compared with a saturated acid, stearic acid, as in Fig. 1.12b. By contrast the cis double bond in Fig. 1.12c inserts a bend in the chain with the result that the molecules do not pack together as well. The melting points of a selection of trans compounds are given in Tables 1.6 and 1.7 (Hagemann et al., 1972; Jackson and Callen, 1951; Markley, 1947).

1.5.2

Ultraviolet spectra

Ultraviolet (UV) spectra are not used very often for the determination of the major fatty acids, because the UV -maximum for the cis unsaturated group is at 176 nm and for the trans double bond at 187 nm. The UV spectrum is much more informative when conjugated double bonds are present in the fatty acid (Hamilton and Cast, 1999). In cyclohexane, -eleostearic acid, 9c,11t,13t-octadecatrienoic acid, has max 262, 272 and 283 nm, as shown in Fig. 1.13, whilst -eleostearic acid, 9t,11t,13t-octadecatrienoic acid, has very similar absorption maxima at max 259, 270 and 281 nm (O’Connor et al., 1947).

Table 1.6 Melting points (◦ C) of the polymorph of selected triacylglycerols. Trielaidin Triolein Tripetroselenin Tripetroselaidin Trierucin Tri-trans-13-docosenoin Adapted from Hagemann et al., 1972.

41 5 28 52 32 58

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties Table 1.7

Melting points of selected cis and trans lipids. Melting point (◦ C)

Lipids Elaidic acid Oleic acid Petroselaidic acid Petroselenic acid Vaccenic acid Methyl elaidate Methyl oleate Methyl petroselaidate Triolein Trielaidin

45.0–45.5 16.3 53.2 33 43.5–44.1 11.5 −5 19.9 5.1 42.2

Adapted from Jackson and Callen, 1951; Markley, 1947.

(a)

(b)

(c) Fig. 1.12

Three-dimensional representations of (a) elaidic acid, (b) stearic acid and (c) oleic acid.

19

BLUK122-Dijkstra

20

September 25, 2007

19:41

Trans Fatty Acids

200

Absorbance

150

A 100

B

50

0 220

250 Wavelength (nm)

300

Fig. 1.13 UV absorption spectra of (A) -eleostearic acid and (B) -eleostearic acid. (Adapted from O’Connor et al., 1947.)

1.5.3

Infrared spectra

The maxima for the cis double bonds are at 1660–1630 and 730–650/cm, and for the trans double bonds at 1680–1670 and 980–865/cm. When the double bonds are in conjugation, we get cis,trans conjugated bonds at 990–980 and 968–950/cm, whilst trans,trans double bonds are at 990–984/cm. For triunsaturated acids, we have a maximum at 989/cm for cis,cis,trans conjugated, at 991/cm for cis,trans,trans conjugated and at 994/cm for trans,trans,trans conjugated (Chapman, 1965). A typical spectrum for trielaidin is shown in Fig. 1.14, with the characteristic band at 980/cm, in contrast with the spectrum for triolein (Feuge et al., 1951).

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties

21

100 Triolein 80 60 40 20 100 Trielaidin Transmission (%)

80 60 40 20 100 Tristearin 80 60 40 20 0

1

3

5

7

9

11

Wavelength (µm)

Fig. 1.14 Infrared absorption spectra of triolein, trielaidin and tristearin in chloroform. (Adapted from Feuge et al., 1951.) Table 1.8 Chemical shifts for the ethylenic carbon atoms in octadecenoic acid isomers by nuclear magnetic resonance. Position of double bond 5 6 7 8 9 10 11 12 13 14 15

Trans 131.98 131.13 130.81 130.67 130.54 130.47 130.43

Cis 128.72 129.49 129.89 130.13 130.23 130.21 130.34

131.44 130.62 130.31 130.17 130.09 130 129.96

130.14 130.39 130.63 131.93

Adapted from Gunstone, 1993.

128.21 129.01 129.45 129.65 129.78 129.83 129.89 129.94 129.9

130.13 129.45

130.16 131.54

129.66 129.39

BLUK122-Dijkstra

22

September 25, 2007

19:41

Trans Fatty Acids

Fourier transform infrared methods have been developed for the measurement of the trans content of oils, as noted by Ismail et al. in 1999.

1.5.4

Nuclear magnetic resonance spectroscopy

Gunstone (1993) has shown that the 13 C spectra of cis and trans fats can be used as a way of analysing partially hydrogenated fats. The values given in Table 1.8 show that there are differences between the cis and trans isomers in cases where the double bond is at C5 through C15 of the octadecenoic acids.

Nutrition facts Serving size 1 cup (228 g) Serving per container 2 Amount per serving Calories 260 Calories from fat 120 % Daily value* Total fat

13 g

20%

5g

25%

Saturated fat Trans fat

30 mg

10%

660 mg

28%

31 g

10%

Dietary fiber

0g

0%

Sugars

5g

Cholesterol Sodium Total carbohydrate

Protein

5g

Vitamin A Calcium

4% 15%

Vitamin C

2%

Iron

4%

*Percent daily values are based on a 2000 caloric diet. Your daily values may be higher or lower depending on your calorie needs:

Calories

2000

3000

Total fat

less than

65 g

Sat fat

less than

20 g

25 g

Cholesterol

less than

200 mg

200 mg

Sodium

less than

2400 mg

2400 mg

Total carbohydrate Dietary fiber

80g

300 g

375 g

26 mg

30 mg

Calories per gram Fat 9

Fig. 1.15

Carbohydrate 4 Protein 4

A United States of America nutrition facts label. (Adapted from Moss, 2006.)

BLUK122-Dijkstra

September 25, 2007

19:41

Structure, occurrence, nomenclature, biosynthesis and properties

1.6

23

LABELLING AND LEGISLATION

The Danish government has issued an order (Order no. 160), which came into operation on 31 March 2003. The order applies to oils and fats, including emulsions with fat as the continuous phase, which, either alone or as part of processed foodstuffs, are intended, or are likely, to be consumed by humans. The order does not apply to the naturally occurring content of trans fatty acids in animal fats or products governed under any other legislation. The order only applies to products sold to the final consumer. It states that it is prohibited to sell oils and fats covered by the order to consumers if they contain a higher level of trans fatty acids defined in the Annex than that quoted in Section 3. Section 3 states that, as from 1 June 2003, the content of trans fatty acids in the oils and fats covered by this order must not exceed 2 g/100 g of oil or fat. In products that are claimed to be ‘free from fatty acids’, the content of trans fatty acids in the finished product shall be less than 1 g/100 g of the individual oil or fat. Such has been the success of the Danish manufacturers/authorities that the level of intake of trans fatty acids from margarines and shortenings has fallen away completely from 4.5 g trans fatty acids per day in 1975, 2.2 g trans fatty acids per day in 1993 through 1.5 g trans fatty acids per day in 1995 to almost zero by 2005 (Leth et al., 2006). From 1 January 2006, the US government amended its regulations on nutrition labelling. This regulation is available on http://vm.cfsan.fda.gov/-Ird/fr991117.html and requires that all foodstuffs or products containing trans fatty acids, e.g. dietary supplements, should have the amount of trans fatty acids stated on the label. A typical example is shown in Fig. 1.15 (Moss, 2006), which is a listing of the grams of trans fat in a serving defined as the sum of all the unsaturated fatty acids that contain one or more isolated (non-conjugated) double bonds in the trans configuration. If the serving contains less than 0.5 g, it is possible to state the content as zero trans (Yurawecz, 2004). In October 2006, the Food navigator.com http://www.foodnavigator.com reported that the Australian government plans to work with industry to reduce trans fatty acids in Australian food.

REFERENCES Chapman, D. (1965) Infrared spectroscopy of lipids. J Am Oil Chem Soc 42 (5), 353–371. Comes, F., Farines, M., Aumelas, A. & Soulie, J. (1992) Fatty acids and triacylglycerols of cherry seed oil. J Am Oil Chem Soc 69, 1224–1227. Dijkstra, A.J. (2002) Hydrogenation and fractionation. In: Fats in Food Technology (ed. K.K. Rajah). Sheffield Food Technology, Sheffield, pp. 123–158. Feuge, R.O., Pepper, M.B., O’Connor, R.T. & Field, E.T. (1951) Modification of vegetable oils. XI The formation of trans isomers during hydrogenation of methyl oleate and triolein. J Am Oil Chem Soc 28, 420–426. Gunstone, F.D. (1967) An Introduction to the Chemistry and Biochemistry of Fatty Acids and Their Glycerides. Chapman and Hall, London. Gunstone, F.D. (1993) Composition of hydrogenated fats by high resolution 13 C nuclear magnetic resonance spectroscopy. J Am Oil Chem Soc 70 (10), 965–970. Gunstone, F.D. (1996) Fatty Acid and Lipid Chemistry. Blackie Academic and Professional, Glasgow. Gunstone, F.D. (2004) The Chemistry of Oils and Fats, Sources, Composition, Properties and Uses. Blackwell Publishing, Oxford. Gunstone, F.D., Harwood, J.L. & Padley, F.B. (1994) The Lipid Handbook. Chapman and Hall, London. Gurr, M.I. & James, A.T. (1980) Lipid Biochemistry, an introduction. Chapman and Hall, London. Hagemann, J.H., Tallent, W.H., Barve, J.A., Ismail, I.A. & Gunstone, F.D. (1975) Polymorphism in singleacid triglycerides of positional and geometric isomers of octadecenoic acid. J Am Oil Chem Soc 52, 204– 207.

BLUK122-Dijkstra

24

September 25, 2007

19:41

Trans Fatty Acids

Hagemann, J.W., Tallent, W.H. & Kolb, K.E. (1972) Differential scanning calorimetry of single acid triglycerides: effect of chain length and unsaturation. J Am Oil Chem Soc 49, 118–123. Hamilton, R.J. & Cast, J. (1999) Spectral Properties of Lipids. Sheffield Academic Press, Sheffield. Hastert, R.C. (1996) Hydrogenation. In: Bailey’s Industrial Oil and Fat Products, Vol. 4 (ed. Y.H. Hui). J. Wiley and Sons, New York, pp. 213–300. Hay, J.D. & Morrison, W.R. (1970) Isomeric monoenoic fatty acids in bovine milk fat. Biochim Biophys Acta 202, 237–243. Hoffmann, G. (1989) Chemistry and Technology of Edible Oils and Fats and Their High Fat Products. Academic Press, London. Ismail, A.A., Nicodemo, A., Sedman, J., van de Voort, F.R. & Holzbaur, I.E. (1999) Infrared spectroscopy of lipids: principles and applications. In: Spectral Properties of Lipids (eds R.J. Hamilton & J. Cast). Sheffield Academic Press, Sheffield, pp. 235–269. Jackson, F.E. & Callan, J.E. (1951) Evaluation of the Twitchell isooleic method: comparison with the infrared trans isooleic method. J Am Oil Chem Soc 28, 61–65. Korver, O. & Katan, M.B. (2006) The elimination of trans fats from spreads: how science helped to turn an industry around. Nutr Rev 64 (6), 275–279. Leth, T., Jensen, H.G., Mikkelsen, A.A. & Bysted, A. (2006) The effect of the regulation on trans fatty acid content in Danish food. Atheroscler Suppl 7, 53–56. Markley, K.S. (1947) Fatty Acids: Their Chemistry and Physical Properties. Interscience Publishers, New York. Moss, J. (2006) Labeling of trans fatty acid content in food, regulations and limits – the FDA view. Atheroscler Suppl 7, 57–59. O’Connor, R.T., Heinzelman, D.C., McKinney, R.S. & Pack, F.C. (1947) The spectrophotometric determination of alpha and beta isomers of eleostearic acid in tung oil. J Am Oil Chem Soc 24, 212–216. Placek, L.L. (1963) A review of petroselenic acid and its derivatives. J Am Oil Chem Soc 40, 319–329. Rossell, J.B. (2001) Oils and Fats, Volume 2 Animal Carcass Fats. Leatherhead Publishing, Leatherhead. Sommerfeld, M. (1983) Trans unsaturated fatty acids in natural products and processed foods. Prog Lipid Res 22 (3), 221–233. Yurawecz, M.P. (2004) FDA requires mandatory labeling of trans fat. Inform 15 (3), 184–185.

BLUK122-Dijkstra

2

September 25, 2007

19:43

Trans fatty acids intake: epidemiology and health implications

Geok Lin Khor and Norhaizan Mohd Esa

2.1

INTRODUCTION

The benefits of the functional roles of dietary fat in providing calorie density to the diet as a source of essential fatty acids and as a vehicle for fat-soluble vitamins, nutrients and antioxidants for the human body are long established. However, concerns for the ill-health effects of high intake levels of total fats and saturated fats are also well recognised. Epidemiological and experimental studies have demonstrated the links between high intake levels of saturated fats, elevated serum cholesterol levels, risk of atherosclerosis and coronary heart disease (CHD) mortality. In the light of the evidence, controlling the intake of dietary cholesterol, saturated fat and total fat has become a key dietary guideline in many countries. Heightened concerns for the deleterious effects of saturated fat intake have spurred interest in vegetable oils that are low in saturated fat. Total consumption of vegetable oils worldwide increased over twofolds from about 60 million tonnes in 1985 to over 130 million tonnes in 2005, and the level is projected to exceed 150 million tonnes by 2010 (Fig. 2.1). Averagely, global consumption of vegetable oils grew at an annual rate of 4.2% over the past decade. In terms of consumption per capita, countries of the European Community and United States are the largest consumers of vegetable oils accounting for approximately one-third of world consumption, with intake levels of approximately 50 kg/capita (Jank, 2006). Brazil and Japan are also large consumers at 20–25 kg/capita followed by China and India at 10–20 kg/capita (Fig. 2.2). The remarkable increase in vegetable oils consumption in recent decades may be attributed to rapid growth in population in developing countries and expanding economies in both developing and developed countries. Soya bean oil was the major edible vegetable oil in use globally, but after 2000, palm oil has taken the lead, owing to its relatively lower price per volume basis. The predicted global growth in palm oil usage is likely to continue through to 2016 (Drummond, 2005) (Fig. 2.3). In the food industry, vegetable oils are hydrogenated in the presence of metal catalysts and hydrogen to produce solid or semi-solid fats. As a result of this process termed partial hydrogenation, the double bonds in unsaturated fatty acids are reduced and some double bonds are converted from their normal cis configuration to the trans isomer. Vegetable oil is also partially hydrogenated to remove some fatty acids, such as linolenic and linoleic acid, which easily oxidise causing fat to go rancid with time. Partial hydrogenation of vegetable oils (PHVO) brings about desirable physical and chemical characteristics to foods cooked in PHVO, giving them distinctive flavour, crispness, creaminess, plasticity and oxidation stability.

BLUK122-Dijkstra

26

September 25, 2007

19:43

Trans Fatty Acids 200

Million tonnes

4% volume growth per year 150 100 50 0 1985 Fig. 2.1

1990

1995

2000

2005

2010

Worldwide vegetable oils consumption. (From Jank, 2006.)

EU-15 USA Brazil Japan World China India 0 Fig. 2.2

10

20

30

40

50

60

World per capita consumption of vegetable oils (in kg/capita). (From Jank, 2006.)

% of total fat

30 25 20 15 10 5 0

1976/1980

1986/1990

1999/2000

2000/2010

2016/2020

Period Soybean Fig. 2.3

2.2

Palm/palm kernel

Rape seed/canola

Sunflower seed

Trends in global edible oil production. (From Drummond, 2005; Gunstone, 2002.)

FOOD SOURCES OF TRANS FATTY ACIDS

Partially hydrogenated vegetable oil containing trans fatty acids (TFA) is used widely in the food industry, for reasons including its low cost compared with other fats, and ability to prolong the shelf-life of products and impart desirable characteristics to the food. For example, PHVO are extensively used to cook French fries and fast food, as well as in the preparation

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications Animal products (21%)

27

Cakes, cookies, pies and bread (40%)

Candy (1%) Breakfast cereal (1%) Salad dressing (3%) Shortening (4%) Potato chips (5%) Fried potatoes (8%)

Margarine (17%)

Fig. 2.4 Trans fats in US food supply. (From Perryman and Stone, 2006. Data based on FDA’s economic analysis for the final TFA labelling rule, ‘Trans Fatty Acids in Nutrition Labeling, Nutrient Content Claims, and Health Claims’ (11 July 2003).)

of baked products. In this way, foods containing industrially produced hydrogenated fats and oils contribute the major share of TFA to the diet. Dairy products and ruminant meats also contribute a small proportion of the dietary TFA (generally making up 2–6% of the fat content in dairy products and ruminant meats). These ‘natural’ TFA are formed as a result of the normal process of bacterial transformation of unsaturated fatty acids in the rumen (biohydrogenation). A small quantity of TAF is also found in poultry and pork fat through the feed. The predominant trans isomer in ruminant fat is vaccenic acid (11t-18:1), while in PHVO, elaidic acids (9t-C18:1) and its isomer 10t-18:1) are the major TFA. Data on food sources of TFA are mainly for North American and European countries. Figure 2.4 shows the proportions of TFA found in various food categories in the United States. The major sources of TFA are baked products comprising cakes, biscuits, pies and bread (40%) followed by animal products (21%) and margarine (17%) (Perryman and Stone, 2006). Snack foods, principally fried potatoes and potato chips, contribute 13%, while another 7% comes from salad dressing and shortening. Overall, approximately 80% of the food items containing partially hydrogenated oils constitute the major sources of TFA intake. A similar pattern prevails in Canada where food items containing PHVO are found predominantly in baked and fried foods, such as crackers, cookies, doughnuts, muffins, croissants, French fries and breaded foods (Health Canada, 2006). These food items were also found to be the main sources of TFA of a group of pregnant women (Elias and Innis, 2002). Among these women, bakery foods contributed 33% of the TFA consumed, fast foods 13%, bread 10%, snacks 10% and margarines/shortenings 8%. In the European multicentre TRANSFAIR Study undertaken in 1995–1996, samples of foods contributing to 95% total fat intake in 14 countries were analysed centrally for TFA content (Hulshof et al., 1999; Institute of Food Science and Technology (IFST), 2004). Among the hydrogenated products, oils and fats are the main contributors (35%) followed by biscuits and cakes (16.5%) (Table 2.1). As for the biohydrogenated products, dairy products provide 18.8%

BLUK122-Dijkstra

28

September 25, 2007

19:43

Trans Fatty Acids Table 2.1 Contributions (%) of various foods to trans fatty acids intake in the TRANSFAIR Study. Milk and cheese Butter Eggs Meat and meat products Oils and fats Biscuits and cakes Savoury pies Chips and French fries Others Total

18.8 5.9 0.9 10.3 35.5 16.5 3.5 4.5 4.1

Natural Natural Natural Natural Mainly resulting from hydrogenation Mainly resulting from hydrogenation Mainly resulting from hydrogenation Mainly resulting from hydrogenation Mainly resulting from hydrogenation

100.0

From IFST, 2004.

of the total TFA, while meat and meat products contribute 10.3% and butter another 5.9%. Overall, approximately, two-thirds of TFA in European diets are from foods with PHVO, while the remaining one-third is from natural food sources. Thus, in general, Europeans consume a lower proportion of TFA from PHVO compared with North Americans (∼80%). The TFA content in commonly consumed food items is available in databases such as those from the US (US Department of Agriculture (USDA) Nutrient Data Laboratory, 2007), Canada (Health Canada, 2006) and Denmark (Stender and Dyerberg, 2003). Among the natural products in the US market, the TFA content of ground beef, butter and whole milk are 1 g/100 g, 2–7 g/100 g and 0.07 g/100 g respectively (expressed as grams per 100 grams food). In contrast, the TFA content in shortenings, margarines/spreads and bread/cake products are relatively higher at 10–33 g/100 g, 3–26 g/100 g and 0.1–10.0 g/100 g respectively. Stick margarine typically contains a higher amount of TFA compared with the softer (tub) margarines (Table 2.2). Commercially baked products, e.g. pies and cookies, provide significant amounts of TFA. Table 2.2

Trans fatty acids in selected foods.

Food (ingredients) Stick margarines Tub margarines Shortening (blend of sunflower, soya bean and cottonseed oil) Apple pies Brownies (hydrogenated soya bean and cottonseed oil, and butter oil) Coffee cake (soya bean oil, margarine and shortening) Noodles and oriental prepared food (hydrogenated soya bean oil, cottonseed oil and chicken fat) Noodles and oriental prepared food (palm oil) Potato chips (sunflower oil) Potato chips (vegetable oil) Peanut butter, creamy (hydrogenated vegetable oil and soya bean oil) Adapted from Health Canada, 2006.

Total fat (g/100 g of food)

Trans fatty acid (g/100 g of fat)

70–72 61–73 100

40 ± 4 20 ± 5 0.6

11–12 19–27

29–33 24–25

17

0.9

21

39

22 30 32–40 48–56

0 0.6 0.5–2.7 0–0.2

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications Table 2.3

29

Wide variability in the trans fatty acids content of foods. Mean (range)

Food description (number of samples) White bread (n = 8) Whole-wheat bread (n = 8) Crackers (n = 14) Breakfast cereals (n = 11) Cake mixes (n = 3) Cookies (n = 19) Potato chips (n = 6) Doughnuts (n = 13) Sauces and gravy (n = 16) French fries (n = 16) Hard margarine (n = 14) Soft margarine (n = 14)

Trans fatty acid (% total fat in food)

Trans fatty acid (g/100 g food)

18.5 (1.3–34.9) 15.6 (1.0–36.3) 40.3 (23.5–51.3) 4.2 (0.2–24.3) 29.6 (28.7–30.1) 23.0 (1.4–45.7) 5.9 (0.4–25.3) 29.6 (3.9–42.7) 33.2 (1.7–60.3) 37.7 (4.9–56.9) 39.8 (31.1–44.6) 16.8 (1.1–44.4)

0.4 (0.0–1.0) 0.5 (0.0–1.3) 6.4 (0.7–12.9) 0.1 (0.0–1.1) 2.3 (1.4–2.8) 3.5 (0.3–8.1) 1.4 (0.1–5.7) 3.9 (0.5–7.8) 3.6 (0.0–23.1) 2.1 (0.2–3.7) 39.8 (31.1–44.6) 16.8 (1.1–44.4)

Adapted from Innis et al., 1999.

Depending on the type of vegetable oil(s) used, the TFA content in the food may vary considerably, as shown in Table 2.3 for oriental noodles and potato chips. The wide variability of TFA in the same food category was clearly delineated by Innis et al. (1999). For example, 16 bread samples in the US were found to have levels of TFA, ranging from 1% to over 30% of total fat (Table 2.4). Fourteen samples of crackers of different brands contained 23.5–51.3% Table 2.4

Variability of trans fatty acids content in foods depends on the fat added.

Fat added White bread, commercially prepared Partially hydrogenated soya bean oil, mono- and diglycerides and butter Partially hydrogenated soya bean oil and mono- and diglycerides Partially hydrogenated soya bean oil and mono- and diglycerides from hydrogenated vegetable oil and butter Partially hydrogenated vegetable shortening (may contain soya bean oil and/or cottonseed oil, butter and mono- and diglycerides) Margarine, tub Liquid and partially hydrogenated corn oil, canola oil and mono- and diglycerides Partially hydrogenated soya bean oil, mono- and diglycerides and lecithin Liquid sunflower oil and partially hydrogenated soya bean oil Partially hydrogenated corn oil, mono- and diglycerides and soy lecithin From USDA Nutrient Data Laboratory, 2007.

TFA Total fat Total TFA (% total (g/100 g food) (g/100 g food) fatty acid) 5.2

0.71

15.61

1.4

0.11

9.24

6.3

1.39

25.46

4.5

0.82

21.02

67.5

11.30

17.5

70.1

11.29

16.9

56.6

8.06

14.9

40.3

3.05

7.9

BLUK122-Dijkstra

30

September 25, 2007

19:43

Trans Fatty Acids

of total fat as TFA. Likewise, potato chips of different brands were found to have a wide range of TFA (0.4–25.3%). The wide variability of TFA content in commercially prepared white bread depends upon whether the bread is made from PHVO (e.g. soya bean oil or cottonseed oil) mixed with other oils and fats. Accordingly, the TFA content of white bread may vary from 0.11 to 1.39 g/100 g bread (Table 2.4). Similarly for the manufacture of tub margarine, the amount of TFA differs from 3.05 to 11.30 g/100 g, depending on the vegetable oil used (partially hydrogenated or liquid form of corn oil, canola oil, soya bean oil and sunflower oil). While some foods, like cake mixes and hard margarines, have a narrow range of TFA content, the wide variation within a food category indicates that the use of average values for TFA content would be of limited value (Innis et al., 1999).

2.3

TRANS FATTY ACIDS INTAKE

Various methods are used to estimate TFA at the population level. These include market share data, laboratory analysis of typical composite diets, food consumption data and the use of biomarkers, such as human milk, red blood cell membranes and adipose tissue (CraigSchmidt, 2006). Assessing intakes of TFA by conventional methods of dietary assessment is fraught with challenges because few national food composition databases include TFA content of foods, especially indigenous foods. An indirect approach to estimate population level of TFA intake is to use food availability data in the country’s fats and oils supply. ‘The availability of food for human use represents disappearance of food into the marketing system, and food disappearance measures food supplies for consumption through all outlets – at home and away from home’ (US Department of Agriculture and Center for Nutrition Policy and Promotion). While generally food disappearance data tend to overestimate consumption, the disappearance trends do provide researchers with relative changes in fats and oils consumption. For example, the US food supply data showed that vegetable oils availability in the market had risen from 35 to 68 g/capita/day between 1965 and 2003, an increase of nearly 100% (Fig. 2.5). In 2003, 81.3% of total edible fats and oils consumed comprised vegetable oils, having increased from 56.7% in 1965. This increase has been attributed to the rapid expansion of the food industry, particularly fast food outlets and restaurants and their extensive use of fried foods. Demand for convenient ready-to-eat food, including processed foods and food prepared away from home, is evident worldwide in tandem with increases in household income, resulting in lifestyle shifts in favour of higher intake of away-from-home food. In the United States, such dietary behaviour shifts have been extensively documented, partly out of concern for its growing obesity problem (Kennedy et al., 1999; Nielsen et al., 2002; Popkin et al., 2001). Data on non-pregnant adults aged 18 and older from nationally representative surveys, including the Nationwide Food Consumption Survey (NFCS) and the Continuing Survey of Food Intake by Individuals (CSFII), showed a reduction in the proportion of fat consumed from foods identified as major sources of saturated fat – red meat, butter, whole milk and eggs (Fig. 2.6). At the same time, there has been an increased trend in away-from-home food, leading to marked increases in fat intake from foods such as pizza, French fries, Mexican dishes, cheeseburgers and Chinese dishes. This dietary shift has led to a marked reduction in the ratio of visible fat (e.g. meat fat and butter) to invisible fat (French fries and pizza). The latter involves the use of vegetable oils extensively, as they are an important ingredient in processed food and food prepared in food service outlets.

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

31

% 50

g/person/day 180

45

160

40

140

35

120

30

100

25

80

20

60

15 10

40

5

20 0

0 1965

1970

1975

1980

1985

1990

1995

% veg oils of total fat

g fat from plant and animal products

g fat from animal products

g vegetable oils

2000

2003

g fat from plant products

Fig. 2.5 Trends in the availability of fat from plant and animal products in the United States between 1965 and 2003. (From FAO Food Balance Sheets, 1965–2003.)

40 35

Percentage

30 25 1965

20

1995

15 10 5 0

Me

at

ir Da

ye

gg

s A

e dd

df

ats

De

sse

rts

a Bre

dc

a ere

lp

ast

Ve

a

ge

ta

s ble

fru

its M

d ixe

dis

he

s

Ot

he

rs

Fig. 2.6 Changes (%) in the proportion of total fat contributed by major food groups between 1965 and 1995 in the United States. Data on non-pregnant adults aged 18 and older from two nationally representative surveys, namely the 1965 Nationwide Food Consumption Survey (NFCS 65) and the 1994– 1996 Continuing Survey of Food Intake by Individuals (CSFII 96). Meat: beef, pork and poultry; added fats: butter, margarine and salad dressing; mixed dishes: grain-based with beef, e.g. beef burrito; mixed dishes: grain-based without beef, e.g. pasta with tomato sauce, cheese pizza, macaroni and cheese; others: beverages and seafood. (Reproduced from Popkin et al. (2001), Copyright Elsevier 2001.)

BLUK122-Dijkstra

32

September 25, 2007

19:43

Trans Fatty Acids

In the 1980s, the average intake of TFA of the US population was estimated at approximately 8 g/day/capita, with 85% coming from foods containing partially hydrogenated oils and the rest from meat and dairy products. This figure was based on analytical data by Enig et al. (1983), Slover et al. (1985) and the USDA production and sales figures (Tarrago-Trani et al., 2006). In fact, based on food supply or food disappearance data, the mean intake levels were reportedly as high as 13.3 ± 1.1 g/day/capita (Enig et al., 1990). Also, based on food supply data, Hunter and Applewhite (1991) reported 7.6 and 8.1 g/day/capita for 1984 and 1989 respectively. By the 1990s, however, the mean consumption level of TFA in the US was reported at relatively lower levels. Based on food intake data from the 1989 to 1991 USDA CSFII and the 1995 USDA database of TFA content of selected foods, Allison et al. (1999) estimated mean consumption level of about 5.3 g/day/capita (Table 2.5). The Minnesota Heart Study, as an ongoing observational epidemiological study commencing in 1980–1982, has provided evidence in support of a declining trend in the consumption of TFA in the US. For men, the intake level fell from 8.4 g/day in 1980–1982 to 6.4 g/day in 1995–1997, while the corresponding figures for the women were 5.4 and 4.7 g/day respectively (Harnack et al., 2003). The large prospective Nurses Health Study also reported a decrease trend in TFA consumption, dropping from 2.2% to 1.6% of total fats between 1980 and 1998 (Oh et al., 2005). The decline is a consequence of a reduction in the degree of hydrogenation to retain more of the original polyunsaturated fats (Ascherio and Willet, 1997). Further reduction in TFA consumption on per capita basis may be expected in the US, following the Food and Drug Administration (FDA) imposing the regulation that requires TFA to be declared in the nutrition label of conventional food and dietary supplements. Enforced on 1 January 2006, foods and dietary supplements that contain more than 0.5 g of TFA per serving are required to make the declaration, while servings containing less than 0.5 g of TFA shall be expressed as zero (FDA, 2003). In Canada, the level of TFA intake was described as ‘one of the highest in the world’ in the mid-1990s (Health Canada, 2006). Estimations placed the level at 8.4 g/day/person (Ratnayake et al., 1998) and at 5.9–7.2 g/day/person based on the 1991 nutrition surveys in Nova Scotia and Quebec (Ratnayake, 2002). The study on Canadian pregnant women by Elias and Innis (2002) recorded average intake of TFA of 3.8 ± 0.3 g/day and 3.4 ± 0.3 g/day in the second and third trimester, respectively. These intake levels are comparable to that of non-pregnant women in the US (Allison et al., 1999; Ascherio et al., 1994), indicating that the TFA intake level in Canada has declined from the high levels in the mid-1990s. The decline has been attributed to mandatory nutrition labelling and greater consumer awareness, encouraging food manufacturers to reduce or eliminate TFA from many processed foods sold in grocery stores (Health Canada, 2006). For example, almost all bread products and salad dressings are now free of TFA. Nonetheless, while ‘significant progress has also been achieved in certain food categories such as French fries and snack foods, many other foods – including some varieties of baked goods, oriental noodles, snack puddings, liquid coffee whiteners, microwave popcorn, toaster pastries, hard margarines and shortenings – still contain high amounts of trans fats’ (Health Canada, 2006). Consequently, the Trans Fat Task Force has recommended to the Canadian government that for all vegetable oils and soft, spreadable (tub-type) margarines for sale to consumers or for use as an ingredient in the preparation of foods on-site, the total TFA content be limited by regulation to 2% of total fat content. For all other foods, the Trans Fat Task Force recommended that the total TFA content be limited by regulation to 5% of total fat content. This limit does not apply to food products for which the fat originates exclusively

Oh et al. (2005)

USA

Hunter and Applewhite (1991) Enig et al. (1990) Van den Reek et al. (1986)

Ascherio et al. (1996) Ascherio et al. (1994) Willet et al. (1993) Troisi et al. (1992)

Garland et al. (1998)

Allison et al. (1999)

Availability data/food supply Analysis of weighed diets for 7 days; 8 girls aged 12–15 yr

FFQ semi-quantitative; the Normative Aging Study; 748 men aged 43–78 yr Market sizes and shares data

24-h recall for 3 days; 1989–1991 CSFII; 11 258 aged 3 yr and above FFQ semi-quantitative+diet records; 140 women 40–65 yr from Nurses Health Study FFQ; Health Professionals Follow-up Study; 43 757 men 40–75 yr FFQ semi-quantitative; 282 men and women

Semi-quantitative FFQ; Nurses Health Study 78 778 women aged 30–55 yr FFQ semi-quantitative; Nurses Health Study (NHS 498) women aged 44–70 yr; NHSII 473 women aged 32–50 yr 24-h recall; Minnesota Heart Survey 3766 men and 4183 women aged 25–74 yr

Method of estimate; subjects

13.3 ± 1.1 3.14 ± 0.26

1984: 7.6 1989: 8.1

3.4 ± 1.2

Men: 4.4 ± 2.3 Women: 3.6 ± 2.2

First to fifth quintile: 1.5–4.3

2.8 ± 1.3

Men 1980–1982: 8.4 1995–1997: 6.4 Women 1980–1982: 5.4 1995–1997: 4.7 5.3 ± 0.08

gram

First to fifth quintile: 0.8–1.6 Men: 1.5 Women: 1.7 2.2

2.6 ± 0.02

% total energy

6.53 ± 0.42

Men: 4.3 Women: 4.8

4.7 ± 1.4

7.4 ± 0.06

(Continued )

1980– 1998: 2.2% reduced to 1.6% 4.7 range: 1.5–9.2

% total fat

Mean intake (per day per capita)

September 25, 2007

Harnack et al. (2003)

Mozaffarian et al. (2004a)

Reference

Estimated per capita intake of trans fatty acids in various countries.

Country

Table 2.5

BLUK122-Dijkstra 19:43

Epidemiology and health implications 33

Iceland Holland Belgium Norway UK Denmark Sweden France Germany Finland Spain Italy Portugal Greece Denmark Stender and Dyerberg (2003) Holland Oomen et al. (2001) Finland Pietinen et al. (1997) Zutphen Elderly Study; FFQ; 667 men 64–84 yr ATBC Cancer Prevention Study; diet history; 21 930 men 50–69 yr follow-up from 1985–1988 till 1993

TRANSFAIR Study: food records+food analysis of 1300 foods from 14 countries in 1995–1996; men and women 19–64 yr 19–64 yr 18–63 yr 19–64 yr 0–75+ yr 19–64 yr 19–64 yr 19–64 yr 19–64 yr 25–84 yr 0–70+ yr 1–80 yr 38 yr 23–64 yr

FFQ semi-quantitative; 60 pregnant women Laboratory analysis food composites; 1991 Nova Scotia and Quebec nutrition surveys

Method of estimate; subjects

5.4 4.3 4.1 4.0 2.8 2.6 2.6 2.3 2.2 2.1 2.1 1.6 1.6 1.4 1994: 2.5 2000: 1–2 1985: 10.9 ± 6.3 1995: 4.4 ± 1.7 2.0

1985: 4.3% 1995: 1.9% 0.95

2.0 1.6 1.4 1.5 1.3 1.0 1.1 1.2 0.8 0.9 0.7 0.5 0.6 0.6

Second and third trimester: 1.3 ± 0.1 3.0

Second trimester: 3.8 ± 0.3 Third trimester: 3.4 ± 0.3 5.9–7.2

8.4

% total energy

gram

% total fat

Mean intake (per day per capita)

September 25, 2007

Europe

Elias and Innis (2002) Ratnayake (2002)

Canada

Ratnayake et al. (1998) Stender and Dyerberg (2003)

Reference

(Continued)

34

Country

Table 2.5

BLUK122-Dijkstra 19:43

Trans Fatty Acids

Men: 7.1 ± 3.1 Women: 6.4 ± 2.9 1968: 4.9 1975: 3.0 Men: 6.0 Women: 5.0 174 men and women (control group); FFQ 3.0 Calculation based on assumptions for Men: 6.4 trans fats content in foods Women: 4.4 5.0 Life in New Zealand survey Men: 5.4 Women: 3.4 Food content in foods and nutrition 1.56 consumption profiles

Semi-quantitative FFQ; mean age: men (99) 57 yr and women (101) 62 yr 3-day food records; 275 male and female Urban male: 4.96 ± 0.82 adolescents 12–18 yr and female: 4.75 ± 0.68 Rural male: 4.04 ± 0.78 and female: 4.35 ± 0.67

Baylin et al. (2005)

0.7

1.17 ± 0.4 Men: 2.5 Women: 2.1 — 1.9 both sexes

Men: 2.7 ± 2.9 Women: 3.3 ± 3.0 1968: 14.1 1975: 13.6

2.2 ± 0.3

1968: 5.0 1975: 5.1

FFQ stands for Food Frequency Questionaire used to assess frequency of intake of food items over a specified period e.g. in past month, how often did you eat bread?

Costa Rica

Monge-Rojas et al. (2005)

328 Japanese-Brazilians aged 40–79 yr

Rural men: 4.1 ± 0.8 and women: 3.2 ± 0.6 Urban men: 12.2 ± 2.5 and women: 9.4 ± 1.8 1993: men 4.7 and women 5.1 2000: men 3.3 and women 3.4 —

Bertolino et al. (2006)

Food records; 1769 rural 1806 urban men and women 25–64 yr

Brazil

India

Semma (2002); Okamoto et al. (1999) Singh et al. (1996)

FFQ semi-quantitative; 10 359 men and women 40–59 yr in 1984–1986 Analysis of diets

September 25, 2007

Japan

Bolton-Smith et al. (1995) Sweden Akesson et al. (1981) Germany Heckers et al. (1979) Australia Clifton et al. (2004) Noakes and Nestel (1994) New Zealand Eyres (2000) Wilson et al. (1990)

Scotland

BLUK122-Dijkstra 19:43

Epidemiology and health implications 35

BLUK122-Dijkstra

36

September 25, 2007

19:43

Trans Fatty Acids

from ruminant meat or dairy products. These regulations are recommended to be finalised by June 2008 (Health Canada, 2006). In Europe, the TRANSFAIR Study showed considerable variations in the levels of TFA intake among adults in the 14 countries studied (Stender and Dyerberg, 2003). On average, the values ranged from the lowest at 1.4 g/day/capita in Greece to the highest at 5.4 g/day/capita in Iceland (Table 2.5). In general, the Mediterranean countries appeared to have the lowest average intake levels (below 3 g/day/capita). In contrast, countries such as Iceland, Holland and Belgium showed mean consumption levels exceeding 4 g/day/capita, which was comparable to intake levels in North America. Trans fatty acids consumption in Europe used to be higher than the levels reported in the TRANFAIR Study. In Sweden, Akesson et al. (1981) reported intake of 4.9 g/day/capita in 1968 and 3.0 g/day/capita in 1975, compared with 2.6 g/day/capita reported in the TRANSFAIR Study. In Germany in the 1970s, men and women were estimated to consume 6 and 5 g TFA daily respectively (Heckers et al., 1979), in contrast to the TRANSFAIR figure of 2.2 g/day/capita. Scotland in 1984–1986 had mean total TFA intakes of 7.1 ± 3.1 and 6.4 ± 2.9 g/day for men and women respectively (Bolton-Smith et al., 1995), compared with the TRANSFAIR finding of 2.8 g/day/capita for United Kingdom. Finland seems to have maintained its intake level at about 2 g/day/capita since the 1980s. The Alpha-Tocopherol, BetaCarotene (ATBC) Cancer Prevention Study reported TFA consumption of 2.0 g/day/capita or 0.95% energy on average, compared with the TRANSFAIR intake level of 2.1 g/day/capita. Besides margarine, there are no other important sources of TFA in the Finnish diet (Pietinen et al., 1997). Since the TRANSFAIR Study, the TFA levels in industrially produced foods in some European countries have decreased markedly. In the Netherlands, a major reduction in TFA content of retail foods has been achieved through efforts of industry with minimal government intervention. Societal pressure also helps to reduce the TFA content of fast foods. The French fries sold in a popular franchise in the Netherlands now have less than 4% trans and 24% saturates, as opposed to 21% trans and 21% saturates in the USA. This illustrates the feasibility of reducing TFA in fast foods without increasing saturates (Katan, 2006). In the context of declining intake of TFA, Denmark is a noteworthy case. The average daily intake of industrially produced TFA in Denmark in 1994 was approximately 2.5 g/person and has since fallen primarily as a result of a reduction in the TFA content of table margarines produced in Denmark (Stender and Dyerberg, 2003). Nonetheless, owing to the import of products with a high PHVO content and since ‘population groups with eating habits including frequent consumption of fast food, French fries, microwave popcorn, chocolate bars and the like have daily intakes of industrially produced TFA well above average levels in the general population, the Danish Nutrition Council has recommended that industrially produced TFA should not be used in food, and the use of industrially produced TFA in food should be phased out as soon as possible’ (Stender and Dyerberg, 2003). Since 1 January 2004, Denmark has set an upper limit on the percentage of industrially produced TFA in foods, limiting TFA from sources other than meats and dairy products, to a maximum of 2% of total fat in each food item. The content of TFA in Danish food has been monitored for the last 30 years. In margarines and shortenings, the content of TFA has substantially declined, e.g. from levels in 1995, in which only 42% of the samples were free of industrially produced TFA (less than 1%) whereas 88% were free in 1999 (Stender and Dyerberg, 2003). In 2005, an investigation of a broad range of food showed that the TFA content had been reduced or removed from the products with high TFA content originally, like French fries, microwave popcorn and various bakery

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

37

products. Hence, industrially produced TFA is no longer a significant component of the diet in Denmark (Leth et al., 2006). An evaluation was also undertaken to assess the potential exposure of consumers by analysing popular foods in Denmark and in 25 other countries (Stender et al., 2006). ‘Fiftyfive servings of French fries and chicken nuggets, 87 packages of microwave popcorn, and 393 samples of biscuits/cakes/wafers with “partially hydrogenated vegetable fat” listed high on the food label were bought between November 2004 and February 2006’. A ‘high trans menu’ was defined as a large-size serving of French fries and nuggets, 100 g of microwave popcorn and 100 g of biscuits/wafers/cakes. The amounts of TFA in a ‘high trans menu’ was 30 g in 2001 in Denmark but was reduced to less than 1 g in 2005. By contrast, a ‘high trans menu’ provided more than 20 g in 17 out of 18 countries, with Hungary, Czech Republic, Poland, Bulgaria and USA, ranking highest with 42, 40, 38, 37 and 36 g respectively. Information on intake of TFA in Australia and New Zealand is rather limited. Through the Food Standards Australia and New Zealand’s dietary modeling work, it is estimated that Australians obtain only 0.6% of their daily energy from trans fatty acids and New Zealanders only 0.7% (Food Standards Australia and New Zealand, 2007). Previous intakes were estimated at higher levels. According to a review by Booker and Mann (2005), mean TFA intakes were estimated as 6.4 g/day or 2.5% total energy for men and 4.4 g/day or 2.1% total energy for women based on simulated Australian diets collected in 1987 and estimated amounts of trans fats in the food consumed (Noakes and Nestel, 1994). Among Australians, beef and dairy fat are the major sources of trans vaccenic acid (11t-18:1), whereas margarine is the main source of elaidic acid (9t-18:1). However, since mid-1996 when trans-enriched margarines were no longer available in the supermarket, animal sources of TFA are now the major contributor to the dietary intake of TFA (Clifton et al., 2004). In New Zealand, a survey in 1989/1990 estimated intakes of TFA as 5.4 g/day for men and 3.4 g/day for women. Data accrued from studies conducted in Adelaide between 1995 and 1997 recorded dietary intake of TFA as 3.0 g/day/capita or 1.17 ± 0.4% of energy (Clifton et al., 2004). Skeaff and Gowans (2006) compared TFA intake between butter users (mean 25.8 ± 12.9% of dietary fat) and margarine users (20.4 ± 10.5%) in New Zealand, using plasma phospholipids as an indirect biomarker of TFA intake. Butter users were estimated as consuming 4 g/day of TFA. Presently, the Australia New Zealand Food Standards Code does not require food manufacturers to label the TFA content unless they make a nutrition claim about cholesterol, saturated or unsaturated fats or TFA. However, voluntary labelling is permitted, and many edible oil spread manufacturers in Australia and New Zealand have chosen to voluntarily label their products (Food Standards Australia New Zealand, 2005). Margarines with no TFA are available in New Zealand, and major brands of margarines in Australia also are trans free. Asian countries generally have lower dietary fat intake than Western countries, whose total fat intakes are approximately 35–40% of total energy. However, the situation is rapidly changing in tandem with expanding economy in recent decades. China is a noteworthy example. Consumption of vegetable oils on a per capita basis increased by 440% between 1970 and 1999. China was self-sufficient in vegetable oil production until about 1985 and since then has to rely on importing vegetable oils, including palm oil from Malaysia, which soared from 114, 000 tonnes in 1980 to 1.4 million tonnes in 1993–1994. This rapid increase in utilisation of vegetables oils is reflected in China’s dietary fat intake, which rose from 15.9% of total energy in 1982 to 29.5% by 2003 (Food and Agricultural Organization (FAO) Food Balance Sheets, 1965–2003).

BLUK122-Dijkstra

38

September 25, 2007

19:43

Trans Fatty Acids

Table 2.6

Mean content of trans fatty acids in human milk (g/100 g fatty acids).

Country

Reference

n

Mean ± SD (range)

Canada

Friesen and Innis (2006)

87

Nov 2004–Mar 2005: 6.2 ± 0.48 Apr 2005–Aug 2005: 5.3 ±0.49 Sept 2005–Jan 2006: 4.6 ± 0.32 7.1 ± 0.32 (2.2–18.7) 7.2 ± 3.0 (0.1–17.2)

Innis and King (1999) Chen et al. (1995)

103 198

USA

Mosley et al. (2005) Aitchison et al. (1977)

81 11

Poland

Mojska et al. (2003)

Czechoslovakia

Wiererov´ a et al. (2002)

35

4.22 ± 1.87

France

Chardigny et al. (1995)

10

∼ 2% (1.2–3.0)

Spain

Boatella et al. (1993)

38

∼ 1%

Germany

Koletzko et al. (1988)

15

4.4 (2.2–6.0)

Iran

Bahrami and Rahimi (2005)

52

11.3 ± 3.4

Hong Kong

Chen et al. (1997)

51

0.88 ± 0.61

China

Chen et al. (1997)

33

0.22 ± 0.06

100

7.0 ± 2.3 (2.5–13.8) (2.68–5.43) 2.36–2.77 depends: season, week of lactation

Consumption data on TFA in Asian countries is sparse. In Japan, based on the determination of the TFA content in commercially available foods and food consumption surveys, the daily intake of TFA was estimated to be 1.56 g/capita or 0.7% of total energy (Okamoto et al., 1999; Semma, 2002). In India, the major source of TFA is ‘vanaspati’, which is ghee made from PHVO. Vanaspati has a high TFA content and is used in confectionery, bakery and ready-to-eat foods. Ghee may also be prepared from milk or butter (clarified butter), and as such, it contains a high proportion of short- to medium-chain saturated fatty acids. Ghee prepared from milk or butter is traditionally more affordable to the urban affluent and middle-income groups, being more expensive than vanaspati made from vegetable oils (Ghafoorunissa, 1996). Vanaspati use, however, has risen rapidly in urban areas, as various socio-economic groups are drawn to the cities, for example, to seek opportunities in employment and education. Based on the amount of vanaspati consumed, Indian adults in urban areas were found to consume three times higher TFA (9–12 g/day/capita) than those in rural areas (3–4 g/day/capita) (Singh et al., 1996). High consumption of ghee was one of the risk factors, besides tobacco use, raised fasting glucose, high cholesterol and paternal history of cardiovascular disease, for risk of non-fatal myocardial infarction (MI) in Pakistan (Ismail et al., 2004). Compared with studies on TFA intakes in the adult population, the literature on consumption of TFA among the younger age groups is more meagre. The study by Monge-Rojas et al. (2005) among Costa Ricans aged 12–18 years showed significant urban/rural differences. Urban males and females had intake levels at nearly 5 g/day compared with just over 4 g/day among the rural subjects (Table 2.5). The authors attributed the higher TFA intake in urban areas to greater consumption of partially hydrogenated soya bean oil, whereas the rural subjects consume primarily palm shortening, which is more affordable to the lower income rural population. Overall, the TFA intake of the Costa Rican adolescents can be deemed as rather high. Adolescents in the United States are consuming even higher levels of TFA. Allison

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

39

et al. (1999) found that boys and girls aged 12–19 years consume 7.1 ± 0.4 g/day and 5.1 ± 0.2 g/day respectively, or almost 2.8% of total energy. These findings, albeit limited, present a matter of health concern, as they not only revealed that the diet of the younger age group is high in TFA, but also that a significant proportion of the source of the trans fats is red meat and dairy products, which are also major sources of saturated fat. High consumption of trans and saturated fats is known to induce atherogenic changes to plasma lipids and run the risk of CHD.

2.4

TRANS FATTY ACIDS IN HUMAN MILK

Women’s dietary intake influences the fatty acid composition of breast milk (Chappell et al., 1985). When the diet is calorie deficient, the milk’s fatty acid pattern reflects that of the adipose tissue, but when calories are adequate for energy needs, the diet becomes the main influence. Also, a high carbohydrate intake, which is typical of the diet of women from low-income households, leads to preferential synthesis of lauric acid (12:0) and myristic acid (14:0), with a concomitant decrease in oleic acid (18:1) and stearic acid (18:0). As for TFA in human milk, most of it is directly influenced by the mother’s recent dietary intake. Accordingly, the TFA level in human milk is generally higher in North America than in Europe (Table 2.6). In Canada and United States, the mean concentrations of TFA in human milk are found to exceed 7% of total fatty acids content, with values ranging from 0.1 to 17%, compared with mean levels of 1–4% among European countries. However, it is reported that milk TFA concentration in Canada has been decreasing from 7.1% in 1998 to 4.6% in 2006 (Friesen and Innis, 2006). The decline was attributed to a decrease in consumption of TFA among women, following the introduction of labelling regulations on trans fats in food. The decline may also be due to a decrease in the use of partially hydrogenated fats and oils in foods in Canada. In Asian countries (Hong Kong and China), levels of milk TFA are low, at below 1% (Table 2.6). Similarly, the mean TFA content in breast milk of African women (Nigeria) was found to be below 1%, as reported by Koletzko et al. (1991) and Mosley et al. (2005). A high level of TFA (11.3%) in human milk was reported in Iran, and this was attributed to high consumption of partially hydrogenated vegetable oils that contain high trans fat content (up to 38%) (Bahrami and Rahimi, 2005). The significance of the presence of TFA in breast milk is that it has been implicated in the ‘displacement’ of essential fatty acids in the milk. Studies showed that the percentage of milk TFA is inversely related to the percentages of milk linoleic acid (LA; 18:2, n-6) and -linolenic acid (ALA; 18:3, n-3), and these proportions are reflected in the endogenous lipids of the breast-fed infants (Chen et al., 1995; Innis and King, 1999; Ohrigge et al., 1982). Studies also revealed a significant reverse association between TFA content and arachidonic acid (AA; 20:4, n-6) and docosahexaenoic acid (DHA; 22:6, n-3) in infants’ cord blood lipids (Desci et al., 2001). It is thus suggested that TFA are involved in the inhibition of the desaturation of LA and ALA to AA and DHA. Both AA and DHA are crucial for fetal and infant growth and central system development (Carlson et al., 1997). DHA is also required in visual and neural function and neurotransmitter metabolism. Thus, ‘TFA may have adverse effects on growth and development through interfering with essential fatty acid metabolism, direct effects on membrane structures or metabolism, or secondary to reducing the intakes of the cis essential fatty acids in either mother or child’ (Innis, 2006). Maternal dietary TFA may also compromise foetal physical development since Koletzko (1992) showed that TFA can cross the placenta, and significant correlation was also shown

BLUK122-Dijkstra

40

September 25, 2007

19:43

Trans Fatty Acids

between the concentration of trans 18:1 in maternal plasma and foetal tissue (Hornstra, 2000). Nonetheless, Carlson et al. (1997) called for caution in interpreting possible association found between TFA exposure and fetal growth because of influence of complex confounding factors. Another potential adverse health effect arising from maternal intake of high amounts of TFA is the resultant reduction in milk fat, thus rendering breast milk to be less energy dense for the exclusively breast-fed infants. The potent inhibitor of milk fat synthesis found in cows was trans-10,cis-12-conjugated linoleic acid (CLA), which is formed as a result of biodehydrogenation of LA (Baumgard et al., 2000). 10t-18:1 has also been implicated in reducing milk fat synthesis (Bauman et al., 2004). In humans, results are inconsistent and reduced fat in milk was found in leaner, but not obese, women (Anderson et al., 2005). The authors suggested that women with smaller fat stores may have less substrate to mobilise and utilise as milk fat.

2.5 2.5.1

TRANS FATTY ACIDS INTAKE AND HEALTH IMPLICATIONS Coronary heart disease

The effects of TFA on serum lipids are well recognised since the experimental evidence demonstrated by Mensink and Katan (1990) and others (Nestel, 1994; Sundram et al., 1997). Epidemiological investigations that showed support for the metabolic evidence of dietary TFA increasing the risk of CHD include the Seven Countries Study and the Nurses Health Study. The Seven Countries Study, one of the earliest prospective cohort studies, found a strong positive association between TFA and CHD mortality risk (Kromhout et al., 1995). Based on 25 years of mortality data (between baseline in 1958–1964 and 1987–1988) on ∼13 000 middle-aged men constituting 16 cohorts in Europe, United States and Japan, the results of these cross-cultural analyses suggested that TFA, as well as dietary saturated fat and cholesterol, were important determinants of differences in population rates of CHD death. In the Nurses Health Study on ∼80 000 US women aged 30–55 years, those in the highest quintile of TFA intake from PHVO at baseline (median intake of 3.2% of total energy) had a 35% higher risk of CHD than women in the lowest quintile (median 1.3% of total energy) (Ascherio and Willet, 1997; Oh et al., 2005). In addition, the Nurses Health Study revealed that TFA derived from vegetable oils increase the risk of CHD, whereas the naturally occurring TFA of animal origin do not and may decrease CHD risk (Willet et al., 1993). The highest intake of vegetable TFA was associated with a 78% increase in the risk of CHD (RR 1.78), while the highest intake of animal TFA had a 41% reduction in the risk (RR 0.59). In a case-control study on 239 patients from the Boston area (Ascherio et al. 1994), the highest quintile of TFA intake was associated with a doubling of the risk of first MI. In the Netherlands, the Zutphen Elderly Study on 667 Dutch men aged 64–84 years showed that TFA intake at baseline was positively associated with the 10-year risk of CHD (Oomen et al., 2001). After adjustment for confounding factors (age, energy intake, body mass index, smoking, use of vitamin supplements and intake of alcohol, specific types of fat, dietary cholesterol and fibre), the relative risk amounted to 1.28 (95% CI 1.01–1.61). Based on a meta-analysis of four large prospective cohort studies in the US and Europe (Nurses Health Study, Health Professionals Follow-up Study, ATBC Cancer Prevention Study and Zutphen Elderly Study), Oomen et al. (2001) estimated that a 2% increase in energy intake from TFA

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

41

was associated with a 25% increase in the incidence of CHD. In the ATBC Cancer Prevention Study on ∼22 000 male smokers aged 50–69 years in Finland, after about 6 years from 1985– 1988, men in the top quintile of TFA intake (median 6.2 g/day) were found to have a relative risk for CHD at 1.39, as compared with men in the lowest quintile intake (median 1.3 g/day) (Pietinen et al., 1997). There are however some other large prospective studies that failed to find unequivocally significant associations between TFA intake and CHD risk. The Scottish Heart Health Study on ∼11 000 men and women aged 40–59 years found that the odds ratios for CHD in the quintile with the highest intake of total and commercially derived TFA were elevated, but not significantly so when compared with the quintile having the lowest intake (1.26 in women and 1.08 in men). The authors concluded that their results did not ‘support a major effect of dietary TFA from commercial hydrogenation on CHD risk in Scottish men’. In the Health Professionals Follow-up Study on ∼44 000 US men aged 40–75 years (Ascherio et al., 1996), the subjects completed food-frequency questionnaires every 2 years between 1986 and 1992. The authors found that the incidence of MI and fatal CHD was not significantly associated with TFA intake after adjustment for dietary fibre. In another follow-up study on a cohort of ∼52 000 male health professionals, no significant association between TFA intake and risk of stroke was found (He et al., 2003). The EURAMIC case-control study on 671 men with acute MI in eight European countries also did not find a significant overall association between the 18:1 TFA content in adipose tissue and the risk of first MI (Aro et al., 1995). No significant association between TFA intake and risk of MI was found. (Multivariate OR comparing the top versus the bottom quartile was 0.97.) Mozaffarian et al. (2006) suggested that by excluding the two centres in Spain in the EURAMIC Study, where CHD rates are very low and extremely low trans levels, the odds ratio in the third and fourth quartile increased to 1.53 (95% CI 1.02, 2.28) and 1.44 (0.94, 2.20) respectively. It was also suggested that in Spain, unlike in the other countries, most dietary TFA were from animal sources and this factor might have interaction with other dietary factors, or confounding factors, by unmeasured or poorly measured covariates. Using adipose tissue TFA as a biomarker of dietary intake, a study on British men with sudden death showed an odds ratio of 0.4 for the highest quintile of adipose tissue TFA (Roberts et al., 1995). However, it is suggested that the result may not be conclusive owing to its small sample size (n = 66), and it was observed that its 95% confidence interval figures for the highest TFA intake quintile (0.85, 2.84) were comparable to those in the Nurses Health Study (Ascherio and Willet, 1997). In Norway, Pedersen et al. (2000) also found adipose tissue content of TFA to be significantly higher in patients (n = 100) with first MI than in controls (n = 98) aged 45–75 years. The odds ratio for risk of MI between the fifth and first quintile of adipose tissue TFA was 2.81. The adipose tissue has been found to be a suitable biomarker of dietary intake of TFA, total polyunsaturated fatty acids, as well as n-3 and n-6 cis polyunsaturated fatty acids (Baylin et al., 2002; Garland et al., 1998; London et al., 1991). As the half-lives of fatty acids in adipose tissues are estimated to exceed 2 years, the adipose tissue is considered a stable and long-term store of fat. In this way, adipose tissue fatty acids provide an alternative approach that precludes recall bias to questionnaire-derived indices of long-term intake for trans and polyunsaturated fatty acids. The latter also suffers from ‘the problem of determining brand name usage of margarines, savoury snacks, and baked goods over time and would be limited by the fact that food composition has dramatically changed in recent history in this regard, making it almost impossible to estimate the aggregate of intakes over a number of years (Kohlmeier et al., 1997).

BLUK122-Dijkstra

42

September 25, 2007

19:43

Trans Fatty Acids

In Costa Rica, Colon-Ramos et al. (2006) reported a substantial change in adipose tissue TFA before 2000 and after. Before 2000, Baylin et al. (2003) reported a significant association between total adipose tissue TFA with risk of MI among Costa Rican adults. After adjusting for established risk factors and other confounders, the odds ratio for the highest quintile of total adipose tissue TFA was 2.94, compared with the lowest quintile. This association was attributed mainly to 18:2 TFA that were abundant in both adipose tissue and partially hydrogenated soya bean oil, margarines and baked products popular among Costa Ricans. However, following industrial modification leading to a reduction in the TFA content in the food supply, the association of TFA intake and MI was no longer significant during the study in 2000–2003 (Colon-Ramos et al., 2006). In Australia, a similar decrease in adipose tissue TFA was found following the removal of a major vegetable source of TFA in 1996 (Clifton et al., 2004). Prior to that, in the casecontrol study involving 209 cases, subjects in the highest quintile of TFA intake had an odds ratio of first MI of 2.25 (95% CI 1.16–4.32), compared with the lowest quintile intake. This positive association between levels of trans fat in adipose tissue and the risk of non-fatal MI was mitigated with the elimination of TFA from margarines. The Costa Rica and Australian studies provided evidence, based on adipose tissue TFA, that when the trans fats levels in the food supply are eliminated or substantially reduced, the association between TFA intake and development of CHD is no longer significant. Recent evidence has elucidated the biological mechanism exerted by TFA on the risk of CHD beyond changes on the lipid profile alone (Willet, 2006). Metabolic studies have revealed a positive association between dietary TFA and markers of systemic inflammation in women from the Nurses Health Study (Mozaffarian et al., 2004a) and patients with heart failure (Mozaffarian et al., 2004b). Higher intakes of TFA were associated with significant increase in the activity of the tumour necrosis factor (TNF) system among those with higher body mass index (Mozaffarian et al., 2006). Higher intake of TFA raised the concentrations of soluble TNF- receptors 1 and 2 (sTNF-R1 and sTNF-R2), interleukin-6 and C-reactive protein, all of which are known to be elevated in disease states such as CHD, diabetes and heart failure. The inflammatory effects of TFA are as damaging as those caused by saturated fats (Baer et al., 2004). In addition, higher intake of TFA was positively related to plasma concentrations of biomarkers of endothelial dysfunction, including soluble intercellular and vascular cell adhesion molecules (sICAM-1 and sVCAM-1) and E-selectin (Lopez-Garcia et al., 2005). Substitution of TFA for saturated fatty acids was also found to impair endothelial function, leading to a decrease in brachial artery flow-mediated vasodilatation by 29%, compared with intake of saturated fatty acids (de Roos et al., 2001). Thus, higher TFA intake promotes systemic inflammation and endothelial dysfunction, both of which play an integral part in the development and exacerbation of atherosclerotic lesions. As for the implication of high intake of TFA in elevating the level of Lp(a), an independent risk factor for the development of CHD, some studies have found a positive association (Almendingen et al., 1995; Lichtenstein et al., 1999; Mensink et al., 1992) but not others (Judd et al., 1998; Louheranta et al., 1999; Tholstrup and Samman, 2004). The latter investigated the effect of stearic, palmitic, oleic, trans-18:1 and linoleic acid on postprandial plasma Lp(a) on 16 young, healthy man and found pronounced increase in Lp(a) concentration after intake of all the test fats. However, trans fat did not change Lp(a) concentration during the study period. Intake of TFA contributes to a reduction in the size of low-density lipoprotein (LDL) particles, which confers an independent risk to CHD, since small and dense particles are more atherogenic than larger, less dense LDL particles (Lamarche et al., 2001). In a randomised

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

43

trial including 18 men and 18 women, consumption for 35 days of diets with different amounts of TFA (0.6, 9.4, 13.6 and 26.1 g) showed that LDL particle size decreased significantly in a dose-dependent fashion with increasing amounts of dietary TFA. Cholesterol concentrations in large and medium-sized LDL particles also increased proportionately to the amount of TFA in the diet (Mauger et al., 2003). Lemaitre et al. (2002) examined the role of TFA intake with sudden cardiac death, which is usually due to ventricular fibrillation and can occur as a first manifestation of heart disease. While the effects of free fatty acids in cardiac arrhythmias have been studied for more than 30 years (Katz, 2002), the link between TFA intake, TFA levels in erythrocyte membranes and sudden cardiac death is less well elucidated. Different trans isomers were found to exert their effects differently. High levels of trans 18:2 (cis-9, trans-12 and trans-9, cis-12 isomers of LA) in red blood cell membranes were found to be associated with markedly higher risk of sudden cardiac death than cell membrane levels of trans 18:1 (trans isomers of oleic acid), the major TFA in foods (Lemaitre et al., 2006). The author called for further studies to investigate the possible effects of trans 18:2 on arrhythmia, and meanwhile it would be prudent to limit dietary intake of trans 18:2. The majority of studies that examined the association between TFA intake and CHD risk have focused on industrially produced TFA, especially PHVO. TFA from partially hydrogenated fish oil (PHFO) has also been found to be as potent as butterfat, and significantly more potent than partially hydrogenated soya bean oil in raising plasma total and LDL-cholesterol levels in a study in Norway (Almendingen et al., 1995). In addition, Muller et al. (1998) also reported that, after 2 weeks of consuming a margarine based on PHFO, the LDL-cholesterol concentrations and the LDL:HDL ratios of subjects were 19% and 12.6% respectively, higher than those who consumed margarine based on vegetable oils. Cantwell et al. (2006) compared the postprandial effects of three solid fats (PHFO, lard and palm oil) on lipid and lipoprotein levels in eight normocholesterolaemic males. Consumption of all three high-fat test meals resulted in pronounced postprandial lipaemia, which is known to be associated with adverse metabolic events. Since TFA is known to interfere with cell membrane function, studies have been conducted to examine the effects of TFA on insulin sensitivity and consequently diabetes risk.

2.5.2

Diabetes

Compared with the evidence linking saturated fatty acids to the impairment of insulin sensitivity when substituted for unsaturated fatty acids, data relating TFA intake with diabetes risk is scanty, as reviewed by Ris´erus (2006). While there are studies that supported a positive association of TFA intake having adverse effects on insulin sensitivity, there are also studies that failed to do so. In the large prospective Nurses Health Study on ∼85 000 US female nurses who were followed for 16 years, Salmeron et al. (2001) found that while polyunsaturated fatty acid intake was associated with a substantial reduction in diabetes risk, TFA and dietary cholesterol were associated with increased risk. The authors estimated that ‘replacing 5% of energy from saturated fatty acid with energy from polyunsaturated fatty acid was associated with a 35% lower risk and that replacing 2% of energy from TFA with polyunsaturated fatty acid was associated with a 40% lower risk’. Based on the findings, it was suggested that in the US, given that the average intake of TFA from PHVO was ∼3% of energy, the incidence of type 2 diabetes could be reduced by ∼40% if these oils were consumed in their original, unhydrogenated form (Salmeron et al., 2001). Such findings have serious implications for the country food supply and questions were raised about the study’s method of analyses of

BLUK122-Dijkstra

44

September 25, 2007

19:43

Trans Fatty Acids

TFA intake. For example, the results could have been influenced by industrial changes in the TFA content of foods during the study period and wide variations that prevailed among foods in the same category. In contrast to the positive finding in the Nurses Health Study, others did not find consumption of TFA significantly associated with diabetes risk. These include the prospective studies on male health professionals (Meyer et al., 2001; Van Dam et al., 2002) and randomised crossover studies (Louheranta et al., 1999; Lovejoy et al., 2002). As pointed out by Ris´erus (2006), the literature suggests that TFA ‘has no significant effect on insulin sensitivity in lean healthy subjects’.

2.5.3

Cancer

The relationship between dietary fat and risk of cancer has been the subject of copious research studies worldwide. In epidemiological studies, cancers of the breast, colon, rectum and prostate are the common types of cancers that have been extensively investigated in relation to fat intake. Most of these studies examined the association of intakes of total fat and individual categories of fat with cancer risk, such as saturated fats, polyunsaturated fats and monounsaturated fats. There are relatively fewer studies that examined the relationship from the perspective of specific fatty acids intake, such as the essential fatty acids and omega-3 polyunsaturated fatty acids in the light of the critical roles these fatty acids play in human health. The literature on the association between TFA intake and risk of cancer is meagre, although in the past decade or so, increasingly more epidemiological studies have emerged, reflecting growing concerns for the adverse health effects arising from TFA intake. Early experimental animal studies dating back to the 1940s and international correlational observations provided support for a positive association between fat intake and development of cancer. The national recommendation by the US Committee on Diet, Nutrition and Cancer (Committee on Diet, Nutrition and Cancer, 1982) for consumption of saturated and unsaturated fats to be less than 30% of total calories in the diet constitutes an important influence on dietary guidelines of many countries. In relations to breast cancer, an extensive review by Lee and Lin (2000) of studies conducted in the 1980s and 1990s showed conflicting results for dietary fat as a risk factor. A pooled analysis of combined data from seven prospective studies in four countries found no significant association between total dietary fat intake and the risk of breast cancer (Hunter et al., 1996). There was no reduction in risk even among women whose energy intake from fat was less than 20% of total energy intake. Evidence from large cohort studies, including the Nurses Health Study and the Health Professionals Follow-up Study, did not support a positive association (Willet, 2001). In examining TFA intake and breast cancer risk, the data are equivocal. During a 14-year follow-up of the Nurses Health Study, there was no evidence of a significant association between intake of TFA and risk of breast cancer (Holmes et al., 1999). The authors reported that in multivariate models, the relative risk (95% CI) for a 1% increase in energy from TFA, the values were 0.92 (0.86–0.98). In the Nurses Health Study II, Cho et al. (2003) also did not find TFA intake to be significantly related to breast cancer risk, the multivariate relative risk for fifth and first quintile intake being 0.96, 95% CI (0.70, 1.31). In contrast, the EURAMIC Study in several European countries and Israel found that cancers of the breast and colon were associated negatively with cis monounsaturated fatty acids and positively with TFA, based on mean fatty acid composition of adipose tissue samples

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

45

(Bakker et al., 1997; Kohlmeier et al., 1997). A critique of the latter study was that it failed to take into account confounding risk factors for colorectal cancer (Nkondjock et al., 2003b). In relation to colorectal cancer, the overall evidence seems to indicate a lack of consistent positive association between intake of TFA and risk of colorectal cancer (Table 2.7). Among cohort studies, the Women’s Health Study on about 40 000 US women found that total fat intake was not related to colorectal cancer risk, neither were intakes of individual fat types nor major fatty acids including TFA (Lin et al., 2004). However, they found a positive association for fried foods away from home (e.g. French fries, fried chicken and fried fish; RR = 1.86 between fifth and first quintile intake). This result warrants further research to ascertain whether it is the presence of TFA, or acrylamide, a mutagenic chemical, in the brown surface of many heated food products, or other dietary factors that resulted in the positive association. Also, no strong association was established for TFA intake and risk of colorectal cancer among some recent case-control studies conducted in the US and Canada (Lin et al., 2004; Nkondjock et al., 2003a). As for risk of prostate cancer, there is some evidence that dietary fat is associated with increased prostate cancer risk and that specific fatty acids may have unique effects on prostate cancer risk; e.g. omega-6 fatty acids generally promote and omega-3 fatty acids generally prevent tumourigenesis (Bartsch et al., 1999). As for TFA intake, epidemiological evidence from a recent study on about 15 000 US physicians reported that plasma levels of TFA were associated with increased risk of developing prostate cancer, which were specific to organconfined and non-aggressive tumours (Chavarro et al., 2006). The -Carotene and Retinol Efficacy Trial (CARET), a randomised trial of supplemental -carotene and retinol for the prevention of lung cancer among 18 314 heavy smokers and asbestos-exposed workers, began in 1985 and ended prematurely in 1996 when it was determined that the supplements increased risks of lung cancer, cardiovascular disease and total mortality but had no effects on prostate cancer incidence or mortality (King et al., 2005). The authors found consistent trends for increasing prostate cancer risk with higher serum levels of C18 but not C16 trans fatty acids. Table 2.7 summarises the main finding from some epidemiological studies relating TFA intake and risk of cancer of the breast, colon, rectum and prostate undertaken in recent decades. Lack of consistent results from epidemiological studies may be due to several factors, including methodological issues, e.g. dietary assessment tools used (food-frequency questionnaire versus dietary records), and errors in estimating the TFA content, as few countries have the database for TFA content in foods. In the light of conflicting results, more rigorous studies are warranted. In the meantime, one could adopt the stand of the Danish Nutrition Council, which made the conclusion that ‘there was no evidence that dietary levels of trans fatty acids had any carcinogenic effect. Studies published since 1994 do not warrant revising this conclusion, but provide a basis for continued watchfulness of this possibility’ (Stender and Dyerberg, 2003).

2.6

CONCLUDING REMARKS

Epidemiological studies have shown a positive association between the intake of TFA and the risk of CHD and possibly diabetes, primarily accounted for by industrially produced TFA. TFA exerts its adverse effects on blood lipids, including increasing LDL-cholesterol concentration, decreasing the concentrations of HDL-cholesterol, lipoprotein(a) and triglycerides,

BLUK122-Dijkstra

46

September 25, 2007

19:43

Trans Fatty Acids

Table 2.7

Human studies relating trans fatty acids intake and risk of cancer.

Reference Colorectal cancer Lin et al. (2004)

Nkondjock et al. (2003a) Slattery et al. (2001) McKelvey et al. (1999)

Breast cancer Cho et al. (2003)

Voorrips et al. (2002)

Byrne et al. (2002)

Holmes et al. (1999) Kohlmeier et al. (1997)

Bakker et al. (1997) Petrek et al. (1997)

London et al. (1993)

Study setting; subjects; dietary assessment method US; Women’s Health Study began in 1993, 39 876 health professionals aged above 45 yr; semi-quantitative FFQ* French-Canadians; case (n = 402) – control (n = 668) aged 35–79 yr; FFQ US; case (n = 1993) – control (n = 2410) study; diet history 519 cases of colorectal adenomatous polyps and 551 controls aged 50–74 yr; FFQ by food groups containing PHVO

Findings Intake of TFA, total fat and major fatty acids not significantly associated with colorectal cancer risk; positive significant association found for intake of fried foods away from home TFA intake was not significantly associated with risk of colorectal cancer Weak association in women but not in men between TFA intake and colorectal cancer; OR = 1.5; 95% CI (1.0, 2.0) No significant association between adenomas and food groups that contain PHVO, namely sweetened baked goods, candy bars, French fries/chips and margarines/dressings

US Nurses Health Study II; 90 655 premenopausal women aged 26–46 yr; 1991–1999

TFA intake not significantly related to breast cancer risk; multivariate RR for fifth and first quintile 0.96; 95% CI (0.70, 1.31) The Netherlands Cohort Study on Diet Higher incidence of breast cancer with and Cancer; 62 573 postmenopausal higher intake of TFA; multivariate RR for women aged 55–69 began 1986; highest compared with lowest quintile: TRANSFAIR FFQ; 6.3 yr follow-up of 1.30; 95% CI (0.93, 1.80) cases US Nurses Health Study; 44 697 No significant increased rate of breast postmenopausal subjects without cancer with greater intake of TFA, total benign breast disease from 1980 to dietary fat and fat subtypes 1994; semi-quantitative FFQ US Nurses Health Study; 88 795 women No evidence of association between intake in 1980 follow-up for 14 yr; of total fat, TFA and other specific major semi-quantitative FFQ types of fat, and risk of breast cancer EURAMIC Study of five European cities; Positive association between adipose stores 698 postmenopausal cases primary of TFA and occurrence of breast cancer; breast cancer aged 50–74 yr; TFA covariate-adjusted RR between 75th and from gluteal fat biopsies 25th percentiles of total adipose TFA was 1.40; 95% CI (1.02, 1.93) Ecological study in 11 centres from 8 Cancers of the breast and colon associated European countries (EURAMIC Study); positively with TFA and negatively with adipose fatty acids cis monounsaturated fatty acids US 161 women with T1NO breast Odds of having positive lymph nodes cancer followed up for mean 7.3 yr; significantly lower with higher proportion adipose tissue breast and abdomen of TFA OR = 0.24, 95% CI (0.07, 0.77) aspirated Case-control postmenopausal women No major association between risk of 380 cases stage I or II breast cancer, breast cancer or proliferative benign 176 proliferative benign breast breast cancer and adipose tissue TFA cancer and 397 control; aspirates of and any categories of polyunsaturated subcutaneous fat from buttocks fatty acids (Continued)

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications Table 2.7

(Continued)

Reference

Study setting; subjects; dietary assessment method

Prostate cancer Chavarro et al. (2006)

US physicians case-control study; n =∼ 15 000; blood samples

King et al. (2005)

Hodge et al. (2004) Bakker et al. (1997)

47

US; CARET Study; 272 prostate cancer cases and 426 control; serum fatty acids methyl esters Australia; case (n = 858) – control (n = 905) aged <70 yr; FFQ

Findings Blood levels of TFA associated with increased risk of developing prostate cancer, specific to organ-confined and non-aggressive tumours Higher levels of serum C18 but not C16 TFA increased risk of prostate cancer

Margarine intake positively associated with prostate cancer OR = 1.3, 95% CI (1.0, 1.7) Ecological study in 11 centres from 8 No significant association between C18 European countries (EURAMIC Study); TFA and prostate cancer adipose fatty acids

*FFQ, food-frequency questionnaire.

reducing LDL-cholesterol particle size, impairing endothelial function, as well as promoting insulin resistance and thrombosis (Lopez-Garcia et al., 2005; Ris´erus, 2006). The average impact of TFA-induced changes in the LDL:HDL ratio correspond to tens of thousands premature deaths in the US alone (Ascherio, 2006). On a per-calorie basis, trans fat is said to increase CHD risk more than any other macronutrient (Mozaffarian et al., 2006). The adverse effect of TFA on metabolism of the essential fatty acids (all-cis) n-6 and n-3 fatty acids is a matter of concern. Findings of inverse associations between TFA and the essential n-6 and n-3 fatty acids, and TFA inhibition of AA biosynthesis in infants and young children indicate potential adverse effects of TFA on the normal child growth and development (Desci and Koletzko, 1995; Elias and Innis, 2001). Recent evidence provided support for less favourable neurologic development in neonates with high TFA in the umbilicus vein (Bouwstra et al., 2006). In the light of the vast evidence on the damaging health effects of TFA intake, legislations and modifications by the food industry have brought about substantial reduction of industrially produced TFA in the food supply in several countries. The Danish experience demonstrated that reduction in TFA exposure at the individual level can be brought about without noticeable effect on availability, price and quality of the affected foods (Stender et al., 2006). In Costa Rica, it was indirect international influence that has led to a TFA reduction in the food supply and, consequently, to a reduction in the risk of non-fatal MI (Colon-Ramos et al., 2006). As a result of these developments, dairy and meat have become the major remaining source of TFA in Europe. The question whether these ruminant TFA have the same effect on CHD risk as industrial TFA has not been settled (Katan, 2006). Jakobsen et al. (2006) called for controlled metabolic studies of the effect of intake of total and specific ruminant TFA on CHD risk. Procuring accurate estimates of TFA intake is fraught with challenges, including the fact that TFA content of foods is often not available in food composition tables. Alternative approaches have been used to assess TFA exposure, and these include measuring the TFA content in human milk, adipose tissue, plasma phospholipids and erythrocyte membranes.

BLUK122-Dijkstra

48

September 25, 2007

19:43

Trans Fatty Acids

Studies on relating TFA intakes to disease risk need to be mindful of the influence of other components in the diet, as well as lifestyle factors that may impinge on the diet–disease interaction.

REFERENCES Aitchison, J.M., Dunkley, W.L., Canolty, N.L. & Smith, L.M. (1977) Influence of diet on trans fatty acids in human milk. Am J Clin Nutr 30, 2006–2015. Akesson, B., Johansson, B.M., Svensson, M. & Ockerman, P.A. (1981) Content of trans-octadecenoic acid in vegetarian and normal diets in Sweden, analyzed by the duplicate portion technique. Am J Clin Nutr 34, 2517–2520. Allison, D.B., Egan, S.K., Barraj, L.M. et al. (1999) Estimated intakes of trans fatty and other fatty acids in the US population. J Am Diet Assoc 99, 166–174. Almendingen, K., Jordal, O., Kierulf, P., Sandstad, B. & Pederson, J.I. (1995) Effects of partially hydrogenated fish oil, partially hydrogenated soybean oil and butter on serum lipoproteins and Lp(a) in men. J Lipid Res 36, 1370–1384. Anderson, N.K., Beerman, K.A., McGuire, M.A. et al. (2005) Dietary fat type influences total milk fat content in lean women. J Nutr 135, 416–421. Aro, A., Kardinaal, A.F.M., Salminen, I. et al. (1995) Adipose tissue isomeric trans fatty acids and risk of myocardial infarction in nine countries: the EURAMIC study. Lancet 345, 273–278. Ascherio, A. (2006) Trans fatty acids and blood lipids. Atheroscler Suppl 7, 25–27. Ascherio, A., Hennekens, C.H., Buring, J.E. et al. (1994) Trans-fatty acids intake and risk of myocardial infarction. Circulation 89, 94–101. Ascherio, A., Rimm, E.B., Giovannucci, E.L. et al. (1996) Dietary fat and risk of coronary heart disease in men: cohort follow up study in the United States. BMJ 313, 84–90. Ascherio, A. & Willet, W.C. (1997) Health effects of trans fatty acids. Am J Clin Nutr 66 (suppl), 1006S– 1010S. Baer, D.J., Judd, J.T., Clevidence, B.A. & Tracy, R.P. (2004) Dietary fatty acids affect plasma markers of inflammation in health men fed controlled diets: a randomized crossover study. Am J Clin Nutr 79, 969–973. Bahrami, G. & Rahimi, Z. (2005) Fatty acid composition of human milk in Western Iran. Eur J Clin Nutr 59, 494–497. Bakker, N., van’t Veer, P. & Zock, P.L. (1997) Adipose fatty acids and cancers of the breast, prostate and colon: an ecological study. Int J Cancer 72, 587–591. Bartsch, H., Nair, J. & Owen, R.W. (1999) Dietary polyunsaturated fatty acids and cancers of the breast and colorectum: emerging evidence for their role as risk modifiers, Cardiogenesis 30, 2209–2218. Bauman, D.E., Perfield, J.W. & Lock, A.L. (2004) Effect of trans fatty acids on milk fat and their impact on human health. http://animal.cals.arizona.edu. Accessed 17 October 2006. Baumgard, L.H., Corl, B.A., Dwyer, D.A., Saebo, A. & Bauman, D.E. (2000) Identification of the conjugated linoleic acid isomer that inhibits milk fat synthesis. Am J Physiol 278, R179–R184. Baylin, A., Kabagambe, E.K., Ascherio, A., Spiegelman, D. & Campos, H. (2003) High 18:2 trans-fatty acids in adipose tissue are associated with increased risk of nonfatal acute myocardial infarction in Costa Rican adults. J Nutr 133, 1186–1191. Baylin, A., Kabagambe, E.K., Siles, X. & Campos, H. (2002) Adipose tissue biomarkers of fatty acid intake. Am J Clin Nutr 76, 750–757. Baylin, A., Mi Kyung, K., Donovan-Palmer, A. et al. (2005) Fasting whole blood as a biomarker of essential fatty acid intake in epidemiologic studies: comparison with adipose tissue and plasma. Am J Epidemiol 162, 373–381. Bertolino, C.N., Castro, T.G., Sartorelli, D.S., Ferreira, S.R. & Cardosa, M.A. (2006) Dietary trans fatty acid intake and serum lipid profile in Japanese-Brazilians in Bauru, Sao Paulo, Brazil. Cad Saude Publica 22, 357–364. Boatella, J., Rafecas, M., Codnoy, R. et al. (1993) Trans fatty acid content of human milk in Spain. J Pediatr Gastroenterol Nutr 16, 432–434. Bolton-Smith, C., Woodward, M., Fenton, S., McCluskey, M.K. & Brown, C.A. (1995) Trans fatty acids in the Scottish diet. Br J Nutr 74, 661–670.

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

49

Booker, C. & Mann, J. (2005) The relationship between saturated and trans unsaturated fatty acids and LDL-cholesterol and coronary heart disease. A review undertaken for Food Standards Australia New Zealand, Edgar National Centre for Diabetes Research, University of Otago, New Zealand. www.foodstandrds.gov.au. Accessed on 11 October 2006. Bouwstra, H., Dijck-Brouwer, J., Decsi, T. et al. (2006) Neurologic condition of healthy term infants at 18 months: positive association with venous umbilical DHA status and negative association with umbilical trans fatty acids. Pediatr Res 60, 334–339. Byrne, C., Rockett, H. & Holmes, M.D. (2002) Dietary fat, fat subtypes, and breast cancer risk: lack of an association among postmenopausal women with no history of benign breast disease. Cancer Epidemiol Biomarkers Prev 11, 261–265. Cantwell, M.M., Flynn, M.A.T. & Gibney, M.J. (2006) Acute postprandial effect of hydrogenated fish oil, palm oil and lard on plasma cholesterol, triacylglycerol and non-esterified fatty acid metabolism in normocholesterolemic males. Br J Nutr 95, 787–794. Carlson, S.E., Clandinin, M.T., Cook, H.W., Emken, E.A. & Filer, L.J. (1997) Trans fatty acids: infant and fetal development. Am J Clin Nutr 66, 715S–736S. Chappell, J.E., Clandinin, M.T. & Kearney-Volpe, C. (1985) Trans fatty acids in human milk lipids: influence of maternal diet and weight loss. Am J Clin Nutr 42, 49–56. Chardigny, J.M., Wolff, R.L., Mager, E. et al. (1995) Trans mono- and polyunsaturated fatty acids in human milk. Eur J Clin Nutr 49, 523–531. Chavarro, J., Stampfer, M.J., Campos, H. et al. (2006) A prospective study of blood trans fatty acid levels and risk of prostate cancer. In: Proceedings of the American Association for Cancer Research, Vol. 47, Abstract #4004, 943, Washington DC April 1-5, 2006. Chen, Z.Y., Kwan, K.Y., Tong, K.K. et al. (1997) Breast milk fatty acid composition: a comparative study between Hong Kong and Chongqing Chinese. Lipids 32, 1061–1067. Chen, Z.Y., Pelletier, G., Hollywood, R. & Ratnayake, W.M.N. (1995) Trans fatty acid isomers in Canadian human milk. Lipids 30, 15–21. Cho, E., Spiegelman, D., Hunter, D.J. et al. (2003) Premenopausal fat intake and risk of breast cancer. J Natl Cancer Inst 95, 1079–1085. Clifton, P.M., Keogh, J.B. & Noakes, M. (2004) Trans fatty acids in adipose tissue and the food supply are associated with myocardial infarction. J Nutr 134, 874–879. Colon-Ramos, U., Baylin, A. & Campos, H. (2006) The relation between trans fatty acids levels and increased risk of myocardial infarction does not hold at lower levels of trans fatty acids in the Costa Rican food supply. J Nutr 136, 2887–2892. Committee on Diet, Nutrition and Cancer (1982) Assembly of Life Sciences, and National Research Council. In: Diet, Nutrition, and Cancer. National Academic Press, Washington. Craig-Schmidt, M.C. (2006) World-wide consumption of trans fatty acids. Atheroscler Suppl 7, 1–4. de Roos, N.M., Bots, M.L. & Katan, M.B. (2001) Replacement of dietary saturated fatty acids by trans fatty acids lowers serum HDL cholesterol and impairs endothelial function in health men and women. Arterioscler Thromb Vasc Biol 21, 1233–1237. Desci, T., Burus, I., Moln´ar, S., Minda, H. & Veitl, V. (2001) Inverse association between trans isomeric and long-chain polyunsaturated fatty acids in cord blood lipids of full-term infants. Am J Clin Nutr 74, 364–368. Desci, T. & Koletzko, B. (1995) Trans fatty acids impair linoleic acid metabolism in children? Ann Nutr Metab 39, 36–41. Drummond, L. (2005) Deep Frying in New Zealand – A Review and Technology Update. National Heart Foundation of New Zealand’s Food and Nutrition Working Group, Auckland, New Zealand. Elias, S.A. & Innis, S.M. (2002) Bakery food are the major dietary source of trans-fatty acids among pregnant women with diets providing 30 percent energy from fat. J Am Diet Assoc 102, 46–51. Elias, S.L. & Innis, S.M. (2001) Infant plasma trans, n-6, and n-3 fatty acids and conjugated linoleic acids are related to maternal plasma fatty acids, length of gestation, and birth weight and length. Am J Clin Nutr 73, 807–814. Enig, M.G., Atal, S., Keeney, M. & Sampugna, J. (1990) Isometric trans fatty acids in the U.S. diet. J Am Coll Nutr 9, 471–486. Enig, M.G., Pallansch, L.A., Sampugna, J. & Keeney, M. (1983) Fatty acid composition of the fat in selected food items with emphasis on trans components. J Am Oil Chem Soc 60, 1788–1795. Eyres, L. (2000) Fats, fatty acids and cholesterol. NZ Food J 29, 143–146.

BLUK122-Dijkstra

50

September 25, 2007

19:43

Trans Fatty Acids

Food and Agriculture Organization (FAO) Food Balance Sheets for 1965–2003. http://www.fao.org/es/ess/ faostat/foodsecurity/Files/FoodConsumptionFoodGroups en.xls.Accessed 14 June 2007. Food and Agriculture Organization (FAO) (2006) Food Outlook Global Market Analysis, 1 June 2006 issue. http://www.fao.org/docrep/009/J7927e/J7927e00.htm. Accessed 14 June 2007. Food and Drug Administration (FDA) (2003) Federal Register 21 CFR Part 101 Part III. In: Food Labeling; Trans Fatty Acids in Nutrition Labeling; Final Rule and Proposed Rule, Vol. 68, no. 133. Department of Health and Human Services, United States. Food Standards Australia New Zealand (FSANZ) (2005) Trans Fatty Acids. http;//www.foodstandards.gov. au/newsroom/factsheets2007/transfattyacidsmay203552.cfm. Accessed 14 June 2007. Friesen, R. & Innis, S.M. (2006) Trans fatty acids in human milk in Canada declined with the introduction of trans fat food labeling. J Nutr 136, 2558–2561. Garland, M., Sacks, F.M., Colditz, G.A. et al. (1998) The relation between dietary intake and adipose tissue composition of selected fatty acids in US women. Am J Clin Nutr 67, 25–30. Ghafoorunissa R.V. (1996) Fats in Indian diets and their nutritional and health implications. Lipids 31, S287–S291. Gunstone, F.D. (2002) Production and trade of vegetable oils. In: Vegetable Oils in Food Technology (ed. F.D. Gunstone). CRC Press, Oxford, 1–17. Harnack, L., Lee, S., Schakel, S.F. et al. (2003) Trends in the trans fatty acid composition of the diet in a metropolitan area: the Minnesota Heart Survey. J Am Diet Assoc 103, 1160–1166. He, K., Merchant, A., Rimm, E.B., Rosner, B.A., Stampfer, M.J. & Willet, W.C. (2003) Dietary fat intake and risk of stroke in male US healthcare professionals: 14 year prospective cohort study. BMJ 327, 777–382. Health Canada (2006) TRANSforming the food supply. Report of the Trans Fat Task Force submitted to the Minister of Health, June 2006. www.hc-sc.gc.ca. Accessed 4 April 2007. Heckers, H., Melcher, F.W. & Dittmar, K. (1979) Zum t¨aglichen Verzehr trans-isomerer Fetts¨auren. Eine Kalkulation unter Zugrundelegung der Zusammensetzung handels¨ublicher Fette und verschiedener menschlicher Depotfette. Fette Seifen Anstrichm 81, 217–226. Hodge, A.M., English, D.R., McCredie, M.R. et al. (2004) Foods, nutrients and prostate cancer. Cancer Causes Control 15, 11–20. Holmes, M.D., Hunter, D.J., Colditz, G.A. et al. (1999) Association of dietary intake of fat and fatty acids with risk of breast cancer. J Am Med Assoc 281, 914–920. Hornstra, G. (2000) Essential fatty acids in mothers and their neonates. Am J Clin Nutr 71 (suppl), 1262S– 1269S. Hulshof, K.F.A.M., van Erp-Baart, M.A., Anttolainen, M. et al. (1999) Intake of fatty acids in Western Europe with emphasis on trans fatty acids: the TRANSFAIR study. Eur J Clin Nutr 53, 143–157. Hunter, D.J., Spiegelman, D., Adami, H.O. et al. (1996) Cohort studies of fat intake and the risk of breast cancer – a pooled analysis. N Engl J Med 334, 356–361. Hunter, J.E. & Applewhite, T.H. (1991) Reassessment of trans fatty acid availability in the US diet. Am J Clin Nutr 54, 363–369. Innis, S.M. (2006) Trans fatty acids intakes during pregnancy, infancy and early childhood. Atheroscler Suppl 7, 17–20. Innis, S.M., Green, T.J. & Halsey, T.K. (1999) Variability in the trans fatty acid content of foods within a food category: implications for estimation of dietary trans fatty acid intakes. J Am Coll Nutr 18, 255–260. Innis, S.M. & King, D.J. (1999) Trans fatty acids in human milk are inversely associated with concentrations of essential all-cis n-6 and n-3 fatty acids and determine trans, but not n-6 and n-3 fatty acids in plasma lipids of breast-fed infants. Am J Clin Nutr 70, 383–390. Institute of Food Science and Technology (IFST) (2004) Information Sheet on Trans Fatty Acids, 12 pp. http://www.Ifst.org. Accessed 12 September 2006. Ismail, J., Jafar, T.H., Jafary, F.H. et al. (2004) Risk factors for non-fatal myocardial infarction in young South Asian adults. Heart 90, 259–263. Jakobsen, M.U., Bysted, A., Niels, A.L. et al. (2006) Intake of ruminant trans fatty acids and risk of coronary heart disease. Atheroscler Suppl 7, 9–11. Jank, M.S. (2006) Soybeans. Expansion of production, consumption and trade barriers. Presented at the 2nd Round Table on Responsible Soy, Assuncion, Paraquay. Judd, J.T., Baer, D.J., Clevidence, B.A. et al. (1998) Effects of margarine compared with those of butter on blood lipid profiles related to cardiovascular risk factors in normolipemic adults fed controlled diet. Am J Clin Nutr 68, 768–777.

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

51

Katan, M.B. (2006) Regulation of trans fats: the gap, the Polder, and McDonald’s French fries. Atheroscler Suppl 7, 63–66. Katz, A.M. (2002) Trans-fatty acids and sudden cardiac death. Circulation 105, 669–671. Kennedy, E.T., Bowman, S.A. & Powell, R. (1999) Dietary-fat intake in the US population. J Am Coll Nutr 18, 207–212. King, I.B., Kristal, A.R., Schaffer, S., Thornquist, M. & Goodman, G.E. (2005) Serum trans-fatty acids are associated with risk of prostate cancer in -Carotene and Retinal Efficacy Trial. Cancer Epidemiol, Biomarkers Prev 14, 988–992. Kohlmeier, L., Simonsen, N., van’t Veer, P. et al. (1997) Adipose tissue trans fatty acids and breast cancer in the European Community Multicenter Study on antioxidants, myocardial infarction, and breast cancer. Cancer Epidemiol Biomarkers Prev 6, 705–710. Koletzko, B. (1992) Trans fatty acids may impair biosynthesis on long-chain polyunsaturates and growth in man. Acta Paediatr 81, 301–306. Koletzko, B., Mrotzek, M. & Bremer, H.J. (1988) Fatty acid composition of mature human milk in Germany. Am J Clin Nutr 47, 954–959. Koletzko, B., Thiel, I. & Abiodun, P.O. (1991) Fatty acid composition of mature human milk in Nigeria. Z Ernahrungswiss 30, 289–297. Kromhout, D., Menotti, A., Bloemberg, B. et al. (1995) Dietary saturated and trans fatty acids and cholesterol and 25-year mortality from coronary artery disease: the Seven Countries Study. Prev Med 24, 308– 315. Lamarche, B., St-Pierre, A.C., Ruel, I.L. et al. (2001) A prospective, population-based study of low density lipoprotein particles size as a risk factor for ischemic heart disease in men. Can J Cardiol 17, 859–865. Lee, M.M. & Lin, S.S. (2000) Dietary fat and breast cancer. Annu Rev Nutr 20, 221–248. Lemaitre, R.N., King, I.B., Mozaffarian, D., Sootodehnia, N. & Siscovick, D.S. (2006) Trans-fatty acids and sudden cardiac death. Atheroscler Suppl 7, 13–15. Lemaitre, R.N., King, I.B., Trivellore, E. et al. (2002) Cell membrane trans-fatty acids and the risk of primary cardiac arrest. Circulation 105, 697–701. Leth, T., Jensen, H.G., Mikkelsen, A.A. & Anette, B. (2006) The effect of the regulation on trans fatty acid content in Danish food. Atheroscler Suppl 7, 53–56. Lichtenstein, A.H., Ausman, L.M., Jalbert, S.M. & Schaefer, E.J. (1999) Effects of different forms of dietary hydrogenated fats on serum lipoprotein cholesterol levels. N Engl J Med 340, 1933–1940. Lin, J., Zhang, S.M., Cook, N.R., Lee, I.M. & Buring, J.E. (2004) Dietary fat and fatty acids and risk of colorectal cancer in women. Am J Epidemiol 160, 1011–1022. London, S.J., Sacks, F.M., Caesar, J. et al. (1991) Fatty acid composition of subcutaneous adipose tissue and diet in postmenopausal US women. Am J Clin Nutr 54, 340–345. London, S.J., Sacks, F.M., Stampfer, M.J. et al. (1993) Fatty acid composition of the subcutaneous adipose tissue and risk of proliferative benign breast disease and breast cancer. J Natl Cancer Inst 85, 785–793. Lopez-Garcia, E., Schulze, M.B., Meigs, J.B. et al. (2005) Consumption of trans fatty acids is related to plasma biomarkers of inflammation and endothelial dysfunction. J Nutr 135, 562–566. Louheranta, A.M., Turpeinen, A.K. Vidgren, H.M., Schwab, U.S. & Uusitupa, M.I. (1999) A high-trans fatty acid diet and insulin sensitivity in young health women. Metabolism 48, 870–875. Lovejoy, J.C., Smith, S.R. & Champagne, C.M. (2002) Effects of diets enriched in saturated (palmitic), monounsaturated (oleic), or trans (elaidic) fatty acids on insulin sensitivity and substrate oxidation in healthy adults. Diab Care 25, 1283–1288. Mauger, J., Lichtenstein, A.H., Ausman, L.M. et al. (2003) Effect of different forms of dietary hydrogenated fats on LDL particle size. Am J Clin Nutr 78, 370–375. McKelvey, W., Greenland, S., Chen, M.J. et al. (1999) A case-control study of colorectal adenomatous polyps and consumption of foods containing partially hydrogenated oils. Cancer Epidemiol Biomarkers Prev 8, 519–524. Mensink, R.P. & Katan, M.B. (1990) Effect of dietary trans fatty acids on high-density and low-density lipoprotein levels in healthy subjects. N Engl J Med 323, 439–445. Mensink, R.P., Zock, P.L., Katan, M.B. & Hornstra, G. (1992) Effect of dietary cis and trans fatty acids on serum lipoprotein(a) levels in humans. J Lipid Res 33, 1493–1501. Meyer, K.A., Kushi, L.H., Jacobs, D.R., Jr, & Folsom, A.R. (2001) Dietary fat and incidence of type 2 diabetes in older Iowa women. Diab Care 24, 1528–1535. Mojska, H., Socha, P., Socha, J. et al. (2003) Trans fatty acids in human milk in Poland and their association with breastfeeding mothers’ diets. Acta Paediatr 92, 1381–1387.

BLUK122-Dijkstra

52

September 25, 2007

19:43

Trans Fatty Acids

Monge-Rojas, R., Campos, H. & Rojas, X.F. (2005) Saturated and cis- and trans-unsaturated fatty acids intake in rural and urban Costa Rican adolescents. J Am Coll Nutr 24, 286–293. Mosley, E.E., Wright, A.L., McGuire, M.K. & McGuire, M.A. (2005) Trans fatty acids in milk produced by women in the United States. Am J Clin Nutr 82, 1292–1297. Mozaffarian, D., Katan, M.B., Ascherio, A., Stamper, M.J. & Willet, W.C. (2006) Trans fatty acids and cardiovascular disease. N Engl J Med 354, 1601–1613. Mozaffarian, D., Pischon, T., Hankinson, S.E. et al. (2004a) Dietary intake of trans fatty acids and systemic inflammation in women. Am J Clin Nutr 79, 606–612. Mozaffarian, D., Rimm, E.B., King, I.B. et al. (2004b) Trans fatty acids and systemic inflammation in heart failure. Am J Clin Nutr 80, 1521–1525. Muller, H., Jordal, O., Seljeflot, I. et al. (1998) Effect on plasma lipids and lipoproteins of replacing partially hydrogenated fish oil with vegetable fat in margarine. Br J Nutr 80, 243–251. Nestel, P.J. (1994) Effect of trans fatty acids on serum lipoprotein levels in man. In: Fatty Acids and Lipids: Biological Aspects (eds C. Galli, A.P. Simopoulos & E. Tremoli). World Review Nutrition Dietetics, Karger, Basel, Switzerland, pp. 187–189. Nielsen, S.J., Siega-Riz, M. & Popkin, B.M. (2002) Trends in energy intake in U.S. between 1977 and 1996: similar shifts seen across age groups. Obes Res 10, 370–378. Nkondjock, A., Shatenstein, B., Maisonneuve, P. & Ghadirin P. (2003a) Assessment of risk associated with specific fatty acids and colorectal cancer among French-Canadians in Montreal: a case-control study. Int J Epidemiol 32, 200–209. Nkondjock, A., Shatenstein, B., Maisonneuve, P. & Ghadirian, P. (2003b) Specific fatty acids and human colorectal cancer: an overview. Cancer Detect Prev 27, 55–66. Noakes, M. & Nestel, P. (1994) Trans fatty acids in the Australian diet. Food Aust 46, 124–129. Oh, K., Hu, F.B., Manson, J.E. et al. (2005) Dietary fat intake and risk of coronary heart disease in women: 20 years of follow-up of the Nurses’ Health Study. Am J Epidemiol 161, 672–679. Ohrigge, J.B., Gulley, R.M. & Emken, E.A. (1982) Occurrence of octadecenoic fatty acid isomers from hydrogenated fats in human tissue lipid classes. Lipids 17, 521–557. Okamoto, T., Matsuzaki, H., Maruyama, T., Niiya, I. & Sugano, M. (1999) Trans-fatty acid content hydrogenated oils and estimated intake. J Jpn Oil Chem Soc 48, 1411–1414. Oomen, C.M., Ocke, M.C., Freskens, E.J. et al. (2001) Association between trans fatty acid intake and 10years risk of coronary heart disease in the Zutphen Elderly Study: a prospective population-based study. Lancet 357, 746–751. Pedersen, J.I., Ringstad, J., Almendingen, K. et al. (2000) Adipose tissue fatty acids and risk of myocardial infarction: a case-control study. Eur J Clin Nutr 54, 618–625. Perryman, S. & Stone, M. (2006) Trans fats . . . new labels. www.fshn.cahs.colostate.edu. Accessed 9 October 2006. Petrek, J.A., Hudgins, L.C., Ho, M., Bajorunas, D.R. & Hirsch, J. (1997) Fatty acid composition of adipose tissue, an indication of dietary fatty acids, and breast cancer prognosis. J Clin Oncol 15, 1377–1384. Pietinen, P., Ascherio, A., Korhonen, P. et al. (1997) Intake of fatty acids and risk of coronary heart disease in a cohort of Finnish men. Am J Epidemiol 145, 876–887. Popkin, B.M., Siega-Riz, A.M., Haines, P.S. & Jahns, L. (2001) Where’s the fat? Trends in the U.S. diets 1965–1996. Prev Med 32, 245–254. Ratnayake, W.M.N. (2002) Dietary fat, coronary heart disease, serum cholesterol and quality of fat of the Canadian diet. Presented at the 17th Annual Meeting of the Canadian Section of AOCS on Advances in Nutraceuticals, Toronto. Ratnayake, W.M.N., Pelletier, G., Hollywood, R., Bacler, S. & Leyte, D. (1998) Trans fatty acids in Canadian margarines: recent trends. J Am Oil Chem Soc 75, 1587–1594. Ris´erus, U. (2006) Trans fatty acids and insulin resistance. Atheroscler Suppl 7, 37–39. Roberts, T.L., Wood, D.A., Riemersma, R.A., Gallagher, P.J. & Lampe, F.C. (1995) Trans isomers of oleic and linoleic acids in adipose tissue and sudden cardiac death. Lancet 345, 278–282. Salmeron, J., Hu, F.B., Manson, J.E. et al. (2001) Dietary fat intake and risk of type 2 diabetes in women. Am J Clin Nutr 73, 1019–1026. Semma, M. (2002) Trans fatty acids: properties, benefits and risks. J Health Sci 48, 7–13. Singh, R.B., Niaz, M.A., Ghosh, S. et al. (1996) Association of trans fatty acids (vegetable ghee) and clarified butter (Indian ghee) intake with higher risk of coronary artery disease in rural and urban population with low fat consumption. Int J Cardiol 56, 289–298.

BLUK122-Dijkstra

September 25, 2007

19:43

Epidemiology and health implications

53

Skeaff, C.M. & Gowans, S. (2006) Home use of margarine is an important determinant of plasma trans fatty acid status: a biomarker study. Br J Nutr 96, 377–383. Slattery, M.L., Benson, J., Ma, K.N., Shaffer, D. & Potter, J.D. (2001) Trans fatty acids and colon cancer. Nutr Cancer 39, 170–175. Slover, H.T., Thompson, R.H., Davis, C.S. & Merola, G.V. (1985) Lipids in margarines and margarine-like foods. J Am Oil Chem Soc 62, 775–786. Stender, S. & Dyerberg, J. (2003) The influence of trans fatty acids on health. A report of the Danish Nutrition Council. www.ernaeringsraadet.dk. Accessed 15 September 2006. Stender, S., Dyerberg, J., Bysted, A., Leth, T. & Astrup, A. (2006) A trans journey. Atheroscler Suppl 7, 47–52. Sundram, K., Ismail, A., Hayes, K.C., Jeyemalar, R. & Pathmanathan, R. (1997) Trans (elaidic) fatty acids adversely affect the lipoprotein profile relative to specific saturated fatty acids in humans. J Nutr 127 (suppl), 514S–520S. Tarrago-Trani, M.T., Phillips, K.M., Lemar, L.E. & Holden, J.M. (2006) New and existing oils and fats used in products with reduced trans-fatty acid content. J Am Diet Assoc 106, 867–880. Tholstrup, T. & Samman, S. (2004) Postprandial lipoprotein (a) is affected differently by specific individual dietary fatty acids in healthy young men. J Nutr 134, 2550–2555. Troisi, R., Willet, W.C. & Weiss, S.T. (1992) Trans-fatty acid intake in relation to serum lipid concentrations in adult men. Am J Clin Nutr 56,1019–1024. United States Department of Agriculture (USDA) Center for Nutrition Policy and Promotion (CNPP). Nutrient Content of U.S Food Supply Report 1909–2000. http;//www.cnpp.usda.gov/publications/foodsupply/ FoodSupply1909-2000.pdf. Accessed 14 June 2007. United States Department of Agriculture (USDA) Nutrient Data Laboratory. Fat and Fatty Acid Content of Selected Foods Containing Trans Fatty Acids. http://www.nal.usda.gov. Accessed 5 April 2007. Van Dam, R.M., Rimm, E.B., Willett, W.C., Stampfer, M.J. & Hu, F.B. (2002) Dietary patterns and risk for type 2 diabetes mellitus in US men. Ann Int Med 136, 201–209. Van den Reek, M.M., Craig-Schmidt, M.C., Weete, J.D. & Clark, A.F. (1986) Fat in the diets of adolescent girls with emphasis on isomeric fatty acids. Am J Clin Nutr 43, 530–537. Voorrips, L.E., Brants, H.A.M., Kardinaal, A.F.M., Hiddink, G.J. & van den Brandt, P.A. (2002) Intake of conjugated linoleic acid, fat, and other fatty acids in relation to postmenopausal breast cancer: the Netherlands Cohort Study on Diet and Cancer. Am J Clin Nutr 76, 873–882. Wiererov´a, P.D., B´ılkov´a, D., Rambouskov´a, J. & Andel, M. (2002) Trans fatty acids in subcutaneous fat of pregnant women and in human milk in the Czech Republic. Ann NY Acad Sci 967, 544–547. Willett, W.C. (2001) Diet and cancer: one view at the start of the millennium. Cancer Epidemiol Biomarker Prev 10, 3–8. Willet, W.C. (2006) Trans fatty acids and cardiovascular disease. Atheroscler Suppl, 7, 5–8. Willet, W.C., Stampfer, M.J., Manson, J.E. et al. (1993) Intake of trans fatty acids and risk of coronary heart disease among women. Lancet 341 (8845), 581–585. Wilson, N., Russell, D., Paulin, J. et al. (1990) Life in New Zealand Summary Report. University of Otago, Dunedin.

BLUK122-Dijkstra

3

September 25, 2007

20:0

Conjugated linoleic acid effects on body composition and clinical biomarkers of disease in animals and man: metabolic and cell mechanisms

Klaus W.J. Wahle, Marie Goua, Simona D’Urso and Steven D. Heys 3.1

GENERAL INTRODUCTION: CONJUGATED LINOLEIC ACIDS AND HEALTH

Conjugated linoleic acids (CLAs) are fatty acids that are found naturally in foods derived from ruminant animals and they are mainly trans fatty acids; cis,cis conjugated double bonds also occur in extremely low concentrations in nature. CLAs were first reported by Pariza and his group whilst investigating the carcinogenic components of grilled beef (Pariza and Hargreaves, 1985). Surprisingly, since trans fatty acids are generally regarded as detrimental to health, these modified dienoic trans fatty acids derived from the parent linoleic acid (18:2n-6) were found to have anti-cancer rather than pro-cancer properties. Since these early observations, a multiplicity of purported beneficial as well as detrimental effects of CLAs on health have been reported. These have focused mainly on their effects, through dietary inclusion, in animal models of human diseases or through addition to various types of animal and human cells in culture. Different CLAs are reported to be anti-cancer, anti-atherogenic, anti-adipogenic, anti-diabetogenic and/or anti-inflammatory (see Table 3.1). They also appear to elicit both beneficial and detrimental regulatory effects, or be without effect, on immune function, lipid and eicosanoid metabolism, cytokine and immunoglobulin production. They can also modulate the expression of a number of important genes, either directly by activating specific fatty acid response elements in their reporter regions or through regulation of the expression of specific transcription factors involved in the many metabolic processes they affect (see below). The majority of studies with CLAs have been carried out using chemically synthesised products that contain a mixture of cis and trans isomers, either as capsules or added to certain foods (see below). More recent studies report using individual, highly purified isomers of either cis-9, trans-11 and/or trans-10, cis-12 in animals (mainly mice). A small number of studies also report using these purified isomers individually in dietary supplementation in man. The findings contrast with the earlier findings using CLA mixtures and suggest that individual isomers of CLA have distinct effects on whole-body physiology and cell mechanisms. The trans-10, cis-12 isomer appears to elicit detrimental effects on health, mainly through increased peroxidative processes (see Bhattacharya et al., 2006; Ledoux and Laloux, 2006; Park and Pariza, 2007; Salas-Salvado et al., 2006; Tricon et al., 2005; Wahle et al., 2004; below). These observations need to be assessed critically and further substantiated before they can be regarded as fact. It is intriguing to consider why no detrimental effects are observed when the uncommon trans-10, cis-12 isomer is given as a supplement with varying proportions of the more common cis-9, trans-11 isomer? Also, why can a mixture of two isomers elicit effects that are different from those observed with

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects Table 3.1

55

Some reported general health benefits ascribed to consumption of CLAs.

r

Anti-cancer

r

Inhibit tumour growth/metastasis Inhibit cell proliferation Inhibit angiogenesis Elicit pro-apoptotic conditions

Anti-atherosclerotic

r

Reduce plaque formation Reduce adhesion molecule expression Inhibit cytokine production Inhibit angiogenesis Elicit plaque regression

Anti-obesity

r

Reduce fat deposition Reduce insulin resistance/diabetes Increase adipocyte apoptosis Inhibit lipoprotein lipase activity

Modulation of immune functions

Inhibit inflammatory cytokines Inhibit inflammatory eicosanoids Modulate immunoglobulin formation Regulate TH 1/TH 2 reactions

individual isomers? Furthermore, why do mice generally, but not always, respond differently, often opposingly, to other animal species, including humans? This chapter on CLAs attempts to unravel some of the conflicting data relating to CLAs and their putative beneficial and detrimental effects on human health, albeit through the surrogate use of animal models and cell cultures. Nutritional intervention studies on healthy volunteers and a variety of patients are still insufficient to provide an unequivocal perspective of the beneficial or detrimental effects of these intriguing fatty acids. This chapter is not all-encompassing, but somewhat subjective and eclectic in scope. Hopefully, it will present a balanced picture by highlighting some of the recent positive as well as negative data reported for these fatty acids in relation to various aspects of human health. It is hoped that it will provide a platform for further debate whilst seeking some form of current consensus relating to the safety and health aspects of various types of CLAs. Readers are referred to a number of excellent recent reviews (Bhattacharya et al., 2006; Ledoux and Laloux, 2006; Park and Pariza, 2007; Tricon et al., 2005; Wahle et al., 2004) for fuller details and to the website of the University of Wisconsin, Madison, which provides the most recent updates on all published material relating to CLAs (http://www.wisc.edu/fri/clarefs.htm).

3.2 3.2.1

STRUCTURE, DIETARY ORIGINS AND CONSUMPTION OF CLAs IN MAN Structure

CLAs, as their name implies, consist of a series of positional and geometric isomers of linoleic acid (cis-9, cis-12-18:2n-6), where one or both of the double bonds are in either the cis or the trans configuration at different positions along the acyl chain. The term conjugation in this context means that the double bonds are separated by a simple carbon–carbon linkage rather than by the usual methylene (CH2 ) group (Fig. 3.1). A number of cis,cis; cis,trans; trans,cis and trans,trans isomers with the double bonds at various locations along the acyl chain,

BLUK122-Dijkstra

56

September 25, 2007

20:0

Trans Fatty Acids 9

12 Linoleic acid

11

cis-9,trans-11 CLA

trans-10,cis-12 CLA Fig. 3.1

12

9

10

COOH

COOH

COOH

Structure of the parent omega-6 fatty acid linoleic acid and its two main conjugated derivatives.

from C6 to C15 but predominantly in positions 8–10, 9–11, 10–12 and 11–13, have been identified by various chemical reductive, chromatographic and spectroscopic techniques. These methods are described in Chapter 4, and readers are also referred to some excellent reviews on this subject (Adlof, 2003; Christie, 2003; Dobson 2003). CLAs are conjugated isomers of the dienoic, 18-carbon linoleic acid. Other polyunsaturated fatty acids (PUFAs) with longer carbon chains and two or more double bonds can also be utilised as substrates for biohydrogenation, catalytic, chemical hydrogenation or chain elongation and desaturation processes resulting in 18-carbon or possibly, 19-carbon trienoic and even 20- and 22-carbon tetraenoic, pentaenoic and hexadecanoic fatty acids, with conjugated cis,trans; trans,cis; cis,cis or trans,trans double bonds (Banni et al., 2004; Ringseis et al., 2006; Tsuzuki et al., 2006). The latter can be formed by chain elongation and further desaturation of the dienoic or trienoic precursors in mammalian tissues (Banni et al., 2004). Although not as extensively studied as CLAs, some of these less usual isomers appear to have metabolic effects similar to those of the CLAs and will not be discussed here.

3.2.2

Origins of CLAs in the human diet

PUFAs, other than CLAs, with a conjugated double bond system and with more than two double bonds, occur naturally in nature in various seed oils but their biological activity has to date not been extensively investigated when compared with data available for CLAs. Similarly, a variety of conjugated fatty acid isomers of longer chain length than CLAs (e.g. C19 and C20 ) have been identified as derivatives of long-chain PUFA metabolism in mammalian systems (Banni et al., 2004). These fatty acids, although intriguing, will not be discussed in this chapter. Suffice it to say that preliminary evidence suggests that the C19 conjugated non adecenoic acid (CNA –19:2) in some instances has similar properties to that of CLA (K.W.J. Wahle, unpublished findings). The major natural sources of CLAs are the body tissues, predominantly the fat tissues, of ruminant animals or those with ruminant-like, fermentative digestive processes, like wallabies and kangaroos. Consequently, the main food source of CLAs in the Western diet is from meat and dairy products derived from cattle, sheep, goats and deer. The rumen of these animals is a large anaerobic fermentation vat that contains a variety of microbes capable of biohydrogenating, either alone or in concert, the ingested PUFAs derived from forages and

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

57

other PUFA-containing feed sources, natural or otherwise (e.g. added grain or fish and plant oils) (mainly, 18:2n-6 and 18:3n-3 in plant oils and 20:4n-6, 20:5n-3 and 22:6n-3 in fish oils). This chapter will focus mainly on CLAs derived from plant oils. The initial process in the biohydrogenation of dietary linoleic acid results in the production of the cis-9, trans-11 isomer, as a result of the isomerisation and transposition of the delta-12 double bond to produce the conjugated, non-methylene interrupted double bond. This is the most abundant natural isomer present in ruminant tissue fats (over 90% of total CLA) and has been termed rumenic acid (RA) (McGuire et al., 1999b; Parodi, 1997, 2003). Further hydrogenation of RA produces trans-11-18:1 vaccenic acid (TVA), which is the major trans monounsaturated fatty acid present in the fats of ruminant food products (milk, yoghurt, cheese, butter and meats). Other isomers of trans monoenes are also formed but in small quantities (McGuire et al., 1999b; Parodi, 1997, 2003). This contrasts with partial hydrogenation of feed PUFA in the ramen that predominantly produces the trans-9-18:1 isomer, elaidic acid (McGuire et al., 1999b; Parodi, 1997, 2003). Biohydrogenation processes are invariably incomplete, otherwise only saturated fatty acids would flow from the rumen for absorption and tissue incorporation instead of the mix of RA, TVA, minor isomers and saturated products. The high relative proportion of RA in cows’ milk (2–52 mg/g fat depending on diet) cannot be derived entirely from the modest level of RA flowing from the rumen that is absorbed into blood for distribution to tissues, particularly adipose and mammary. It has been estimated that circa 70% of RA in milkfat is derived from TVA by the activity of delta-9 desaturase (SCD1) in endoplasmic reticulum of the mammary tissue and possibly other tissue cells (e.g. adipocytes) (Grinari and Bauman, 1999). This is interesting because both the precursor TVA and the product RA of the 9-desaturation reaction can inhibit the activity of the SCD1 enzyme in cow mammary tissue microsomes (Baumgard et al., 2002; K.W.J. Wahle and J.M. Elliot, unpublished observations). Apparently, this is not the case when TVA, trans-10, trans-12-18:2 or trans-10-18:1 is infused into cow mammary glands (Lock et al., 2007, Perfield et al., 2006). Presumably, the precursor-product concentrations are kept sufficiently low to allow the reaction to proceed in the in vivo situation or the mammary enzyme undergoes compensatory induction. This has been suggested but requires substantiation (Lock et al., 2007; Wahle et al., 2004). Human tissues are also capable of desaturating TVA to form cis-9, trans-11 CLA, indicating that TVA in cow’s milk can be converted to the bioactive CLA isomer in man (Kuhnt et al., 2006; Moseley et al., 2006). As mentioned above, a wide spectrum of minor geometrical and positional isomers of CLA is produced during rumen biohydrogenation by bacterial isomerases. These range from trans-6, trans-8-18:2 to trans-13, trans-15-18:2, with a number of cis,trans; trans,cis and cis,cis positional isomers in between these extreme positions on the acyl chain (McGuire et al., 1999b; Parodi, 1997, 2003). It is not known if these minor components elicit beneficial or detrimental cell metabolism or function. The predominant isomer in milk and other dairy products is the cis-9, trans-11-18:2, with only minor, but significant, proportions of trans-10, cis-12-18:2 (McGuire et al., 1999b; Parodi, 1997, 2003). This contrasts with synthetic, commercial preparations of CLA where proportions of the two main isomers are usually equal. The chemical method for synthesis will allow the production of a variety of CLAs with different ratios of the two isomers in the end product (McGuire et al., 1999b; Parodi 1997, 2003; Saebo, 2003). There are numerous, recent and not so recent, publications describing ways to enhance the TVA and CLA content of cow’s milk in order to obtain a food product with perceived added health benefits. Detailed appraisal of this literature is beyond the scope of this chapter. Interested readers are referred to a number of excellent reviews on the subject (Bhattacharya et al., 2006; Chilliard et al., 2001; Collomb et al., 2006; Gulati et al., 2000; McGuire et al.,

BLUK122-Dijkstra

58

September 25, 2007

20:0

Trans Fatty Acids

1999b; Parodi, 1997, 2003; Parrish, 2003; Saebo, 2003; Stanton et al., 2003; Tricon et al., 2005; Wahle et al., 2004) and, for the most recent information, to the website mentioned above. It is apparent that a cow’s diet has a major impact on the concentrations of TVA and CLA in its rumen and eventually in its milkfat. Fresh pasture rather than concentrates (grain) elicit the highest natural levels of these fatty acids in milk and alpine pastures yield greater concentrations than lowland pastures (McGuire et al., 1999b; Parodi, 1997, 2003). Reasons for this are not clear. The greatest concentrations of CLA (and TVA) in cows’ milk are obtained when supplemental oils, particularly fish oil, are added to their diets. This increases the PUFA concentration that is available for biohydrogenation. Such feeding regimes do affect the rumen microbial populations and generally result in lowering of the milkfat (Chilliard et al., 2001; Collomb et al., 2006; Gulati et al., 2000; Parodi, 1997; Parrish, 2003; Stanton et al., 2003). Genetic variations in the ability of different cows to produce high levels of CLA in their milk occur. This may be due to individual differences in the expression of mammary SCD1 activity that is responsible for the TVA conversion in the gland (Chilliard et al., 2001; Collomb et al., 2006; Gulati et al., 2000; Parrish, 2003; Stanton et al., 2003). As mentioned previously, this could reflect a greater genetic capability for induction of the enzyme in the face of substrate and product inhibition or an innate ability to regulate or compartmentalise the inhibitor concentrations in the tissues. Genetic selection programmes could produce animals with optimum capacities for CLA production. Genetic engineering is also capable of enhancing the expression of various desaturases in mammalian tissues, including the plant-derived desaturases responsible for synthesising long-chain omega-3 PUFA (Kang, 2005) whereby the requirement for dietary long-chain omega-3 PUFA for optimum health in monogastric mammals is precluded. Intriguingly, the highest natural levels of CLA observed to date in nature occur in wallaby milk (Parodi, 1997). The reason for this is unclear at present but may reflect the animal’s diet. Perhaps, the high intake of CLAs in Australian populations is due to consumption of foods derived from kangaroo and wallaby! The CLA content of other dairy products like cheese and yoghurt is largely dependent on the CLA content of the milk they are derived from since processing appears to have either no or only a little effect on the final content of CLA in these products (McGuire et al., 1999b; Parodi, 1997, 2003). Meat from ruminant animals, particularly the fat associated with meat, is also an important source of CLA, contributing in the region of 25–30% of the total intake in Western populations (McGuire et al., 1999b; Parodi, 1997, 2003). The content of CLA in fish and food products derived from fish is negligible in relation to dairy products, ranging from 0.1 to 0.9 mg/g fat in the common marine foods (McGuire et al., 1999b; Parodi, 1997, 2003). Fatty acids with conjugated double bonds do occur in various seed oils from a number of plant species (e.g. columbinic acid from columbine, a trienoic fatty acid with a trans double bond in the delta-5 position trans-5, cis-9, cis-12-18:3). CLA is not found in vegetable oils used in the food chain. Small amounts of CLA (0.1–0.7 mg/g oil) can be produced during the heating, bleaching and deodorising in the refining process for edible oils (Saebo, 2003). Partial hydrogenated edible oils, used in the production of margarines and shortenings until recently, contain a wide spectrum of cis and trans isomers, including TVA and CLA. In the 1990s, metabolic and epidemiological studies indicated that dietary trans fatty acids in hydrogenated oils could be detrimental to human health with particularly adverse effects on the risk of coronary heart disease. These adverse effects were originally thought to relate largely to the induction of high-risk lipoprotein profiles (increased total and LDL-cholesterol and decreased HDL-cholesterol) and attenuation of cell-regulatory eicosanoid synthesis (Hwang

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

59

and Kinsella, 1979; Katan et al., 1995). The edible-oil industry responded to these criticisms and developed methods that virtually eliminated partially hydrogenated isomers from its products so that modern table margarines and shortenings contain negligible quantities of trans acids, including TVA and CLA (Wahle, 1994; Wahle and James, 1993).

3.2.3

Dietary consumption of CLAs in man

The methods used to estimate the daily intake of CLA in human populations range from large dietary surveys to more detailed dietary assessments of small population subgroups, employing dietary records over periods from 24 h to 7 days to retrospective food-frequency questionnaires or various combinations of these methods (Fritsche et al., 1999; Parodi, 2003). These methods have inherent deficiencies, particularly the long-term, retrospective foodfrequency questionnaire with obvious errors due to unreliable recall and variations between individuals. Despite these shortcomings and the lack of reliable figures for CLA content of more than a small sample of food (Ledoux and Laloux, 2006; S´eb´edio et al., 1999), data have been published from different countries relating to estimated daily CLA intakes that range from negligible in non-dairy and meat-eating populations to circa 1500 mg in Australian populations (Fritsche et al., 1999; Parodi, 2003). The average CLA intake in the US has been estimated at between 52 and 137 mg/day for men and women; in the UK it is about 400–600 mg/day and in German men and women, about 430 and 350 mg/day, respectively. Interestingly, intakes in women are generally lower than those in men, which is possibly a result of their lower dairy fat consumption (McGuire et al., 1999a,b; Parodi, 1997, 2003). The average intake of CLA probably does not reflect the absolute amounts of CLA available to an individual because of endogenous conversion of TVA in dairy products to CLA via the SCD1 enzyme, as mentioned above. It has been estimated that circa 20% of TVA could be converted to CLA in situ (Bauman and Griinari, 2003; Turpeinen et al., 2002). This could increase the CLA available to people on an average UK diet to between 600 and 800 mg/day. Achieving an intake of CLA in the region of 3–6 g/day, which appears to be the level at which benefits on health can be expected, is difficult with present natural, unfortified foods and can occur only by increasing the CLA concentration in milk, and consequently in dairy products, through manipulation of the cow’s feed (see above) or through ingestion of CLAenriched oils as capsules or fortified foods. The latter approach is using natural components in pharmaceutical dosages to derive specific health benefits, like a reduction in inflammation or tumour growth (i.e. as nutraceuticals or functional foods). The food industry is particularly interested in the development of a variety of such functional foods, using different forms of bioactive fatty acids like omega-3 and omega-6 LCPUFAs and CLAs.

3.3

CLAs IN CANCER PREVENTION AND TREATMENT

The first published observations relating to the beneficial effects of CLAs were as anti-cancer agents derived from grilled minced beef by Pariza’s group (Pariza and Hargreaves, 1985). The objective of their research was the identification of pro-carcinogens in the meat, and serendipity indeed to discover lipid compounds that have the opposite effect. Since these initial observations, a wealth of data, mainly in animal models of various cancers and cancer cell cultures, have accumulated that support an anti-cancer role of these interesting fatty acids. Findings range from effects on the growth and metastasis of tumours in animals with chemically induced or implanted human tumours to the influence on the regulation of specific

BLUK122-Dijkstra

60

September 25, 2007

20:0

Trans Fatty Acids

cell signalling mechanisms and expression of oncogenes involved in cell cycling and apoptosis in various cancer cell lines. A small number of studies have attempted to establish if there was an association between CLA intake and a decreased risk or incidence of breast cancer (Bhattacharya et al., 2006; Park and Pariza, 2007; Tricon et al., 2005; Wahle et al., 2004). The following section will present a critical appraisal of the recent, relevant publications and concepts. Again, the data included are not exhaustive.

3.3.1

Epidemiology of dietary fats and cancer risk

Over the past 30 years, many epidemiological studies have examined the relationship between dietary fat intake and the risk of developing many types of cancer (Bhattacharya et al., 2006; Howe et al., 1990; Hunter et al., 1996; Wahle et al., 2004). The limitations and difficulties in undertaking such studies, particularly where food recall is involved, are well recognised. Not surprisingly, conflicting results have been published that have created some confusion regarding the role of fat in the aetiology of cancer.

3.3.2

CLAs and breast cancer

In an attempt to understand further the relationships between dietary fat intake and breast cancer, two significant meta-analyses have been reported, one examining case-control studies and the other focusing on cohort studies (Howe et al., 1990; Hunter et al., 1996). Metaanalysis of case-control studies that revealed an increase in breast cancer risk was correlated with increased dietary fat intake (Howe et al., 1990). In contrast, no relationship was found when the cohort studies were analysed individually (Hunter et al., 1996). More recently, a detailed study of over 13 000 women, which overcame some of the problems associated with the previous two studies, indicated that breast cancer risk was associated with increased fat intake, in particular, saturated fat (Bingham et al., 2003). In recent years the suggestion that all types of dietary fats are associated with increased breast cancer risk and incidence have been questioned, particularly in the light of studies on animal models of the disease where tumours were either induced chemically or human breast tumour cells were transplanted (see below). Such studies have shown that certain types of fats (fatty acids) are pro-tumourigenic, whilst others are anti-tumourigenic. In the latter category, evidence for supporting the beneficial effects of fish-oil consumption and the intake of omega-3 PUFA in reducing carcinogenesis in animal models of disease is persuasive (Karmali et al., 1984). Similar studies in animals implanted with breast cancer cells and fed CLAs suggested the possibility that this type of fatty acid can also be anti-tumourigenic (see below). In contrast, publications indicating that CLAs do not inhibit mammary tumour growth in animals are also extant. Furthermore, a recent study clearly showed that trans-10, cis-12 CLA stimulated mammary tumourigenesis and metastasis in transgenic mice overexpressing the erbB2 gene in mammary epithelium (Ip et al., 2007). This is despite previous evidence to the contrary where both isomers were shown to be anti-tumourigenic and anti-metastatic in rat models of breast cancer (Ip et al., 1995, 2000, 2002). Could this be yet another example where CLAs have different, often opposing, effects in mice compared to rats and other animal models? Despite being in the minority, these reports make it difficult to form a clear consensus about the anti-cancer effects of CLAs (Wong et al., 1997). The possibility that the earlier studies showing no effect of CLAs were carried out with impure products cannot be excluded and more recent studies with clearly defined products should be given greater credence.

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

61

Because the main sources of CLA in the human diet are milk, milk products and the meat from ruminants, it is interesting to review those studies where these dietary components have been analysed separately with regard to breast cancer risk. The literature is again confusing, with some reports showing no relationship (Chajes et al., 2002; Voorrips et al., 2002), some an increased risk of breast cancer with increasing dietary intake (Aro et al., 2000; Knekt and Jarvinnen, 1999) and yet others finding that increased intake resulted in a decreased risk of breast cancer (Landa et al., 1994; Salamini et al., 1984). Reasons for such conflicting observations are not clear but could include experimental design, source of dairy products or analytical sensitivity. Milk also contains other substances known to have anti-cancer activity, such as sphingomyelin and butyric acid, all of which can vary in concentration and thereby complicate the interpretation of the findings. Sphingomyelin is the precursor of ceramide through the action of sphingomyelinase. Both ceramide and butyric acid are potent apoptotic metabolites; the latter primarily in the intestinal tract but possibly also in breast since systemic levels of butyrate can be significant in humans. Nevertheless, on the basis of these and other studies (see below), the World Cancer Research Fund concluded that at present insufficient data existed to confirm either a positive or a negative relationship between breast cancer risk and intake of milk, milk products and meat. A case-control study in Finland compared 193 women with breast cancer with a population-based control group, matched for age and the area in which they lived (Ledoux et al., 2003). The authors found that both CLA intake and serum levels of CLA were significantly lower in postmenopausal women with cancer. It was initially suggested that diets rich in CLA-containing foods may protect against breast cancer in postmenopausal women. However, on further evaluation of the data it was impossible to determine effects attributable to CLA alone and those to other components in milk in these studies (Landa et al., 1994). Another Finnish study also reported a decreased breast cancer risk in women consuming whole milk (Salamini et al., 1984) but again the findings, whilst intriguing, cannot be taken as definitive. A cohort study in the Netherlands recently determined the relationship between dietary intake of CLA and breast cancer risk (Voorrips et al., 2002). The study included 941 patients who had developed breast cancer. Their diet was determined using a 150-item food-frequency questionnaire, with its attendant errors due to recall. Follow-up was over 6 years. The foods were those containing CLA (milk, milk products, butter, cheese and meat). The analytic data (CLA content) for these food items were obtained through linking with the European TRANSFAIR study. Findings actually demonstrated a very weak, but positive, correlation between breast cancer risk and CLA intake, suggesting a possible detrimental effect of milk intake. However, no significant association between intake of CLA-containing foods and breast cancer incidence was observed. An inverse relationship between breast cancer and intake of monounsaturated and cis unsaturated fatty acids was found, indicating a possible positive benefit of these dietary fatty acids. These observations contrast with the abovementioned epidemiological studies, indicating a link between total fat intake and breast cancer incidence. Determining food-intake data and then extrapolating to CLA intake is extremely difficult because of the lack of significant data for the CLA content of foods and also because of their inherent variability in different ruminant food products from different sources and even from the same source in different seasons because of the effect of nutrition and genetics on the CLA availability (Ledoux et al., 2003). Two studies focused on the CLA content of adipose tissue in relation to breast cancer risk and also the risk of developing metastatic disease in women. In the first (Chajes et al., 2002) study, no association between breast cancer risk and CLA levels in adipose tissue

BLUK122-Dijkstra

62

September 25, 2007

20:0

Trans Fatty Acids

was apparent, but the authors did highlight the limitations of their study. In particular, there was a narrow range of CLA levels. The findings lead the authors to question the validity of CLA content in adipose tissue as a biomarker for dietary intake of CLA. The authors then determined adipose tissue CLA levels on diagnosis of breast cancer and after subsequent development of metastatic disease in 209 patients (Chajes et al., 2003). They found no relationship between CLA content in adipose tissue and various prognostic indicators, including overall survival after 7.5 years of follow-up. A recent study in breast cancer patients did find a small association between CLA intake and tumour biology in pre- but not post-menopausal women, but no association was found between CLA intake and risk of developing the disease (McCann et al., 2004). Taken together, these epidemiological/clinical studies have not established a definitive link between normal dietary intakes of CLA from dairy products, the presence of CLA in breast adipose tissue and the risk or incidence of breast cancer. The possibility that consumption of CLAs in greater concentrations than those achievable through intakes of normal dairy and meat products (i.e. in the form of nutraceuticals, functional foods or capsules) may confer benefit cannot be precluded at present. This is relevant in the light of the anti breast cancer effects observed with the relatively high concentrations of these fatty acids, compared to normal intakes, in animal models of disease and in studies with breast cancer cell lines (see below). It is very difficult to draw any firm conclusions about CLA intake, adipose CLA and breast cancer risk in human populations from the studies that are currently available. Nevertheless, the epidemiological evidence to date has neither refuted nor confirmed a positive or negative relationship between ingestion of CLA or CLA-containing foods and breast cancer in women. Human intervention studies using well-defined CLA supplements (isomer types and relative concentrations) at different levels of intake over longer periods of time and at different physiological stages of development (pre- and post-pubertal, see below) in females are needed in order to conclusively clarify these relationships.

3.3.3

CLAs and prostate cancer

3.3.3.1 In man To our knowledge, there are no published epidemiological studies relating to the intake of CLAs and prostate cancer risk or incidence in man. It can, however, be surmised that the high and increasing incidence of the disease in industrialised societies is probably linked to environmental factors, not least of which will be nutrition and particularly fat nutrition. 3.3.3.2 In animal models of disease Studies in animal models of human prostate cancer using transplanted DU145 cells showed clear anti-tumourigenic effects of dietary CLAs similar to those observed with breast cancer models when implanted into SCID mice. The effects were opposite to those observed with linoleic acid feeding that elicited pro-tumourigenic rather than anti-tumourigenic effects. Average tumour volume in the SCID mice was reduced by CLA by circa 70 and 75% when compared with control or linoleate-fed animals respectively (Cesano et al., 1998). Similar effects of linoleic acid have been found with human cancer cell proliferation in vitro (see below). Considering the high intake of linoleic acid in industrialised populations, it is tempting to suggest that these findings could offer an explanation for the high cancer incidence in these

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

63

populations. Unfortunately, two further studies where cells were inoculated into the prostate gland or subcutaneously failed to show any inhibition of prostate cancer growth in athymic mice (Scimeca, 1999). 2-Amino-1-methyl-6-phenylimidazol [4,5-b]pyridine (PhIP) is a potent mutagen and carcinogen derived from meat cooked at high temperatures that produces prostate and mammary tumours when fed to male and female rats respectively (Yang et al., 2003). PhIP feeding for 47 days at 100 ppm induced a fivefold increase in mutation frequency in the prostate of transgenic Big Blue rats. Addition of 1% CLA mixture (w/w) to rat diets 1 week prior to exposure to PhIP decreased mutagenesis by 38% and changed the mutation spectrum, decreasing –1 frameshifts and G:C to A:T transitions (Cohen et al., 2003). These findings suggested a possible use of dietary CLA as a chemopreventative in human prostate cancer. However, observations in an established rat model of implanted hormone-refractory prostate cancer (R-3327-AT-1) indicated a significant increase in tumour volume with CLA feeding compared to controls and no inhibition of growth, or development of the tumour was observed (Cohen et al., 2003). The findings suggesting enhanced tumourigenesis with CLA clearly contrast with the majority of published findings indicating anti-tumour effects in animals and cells and may relate to the use of transgenic rather than normal rats. Such contrasting findings in only a small number of experiments, mostly in mice, when compared to the wealth of published data on mammary cancer, still make it difficult to formulate an absolutely definitive conclusion regarding the effect of CLAs on prostate cancer. Observations that CLAs can inhibit angiogenesis in mammary cancer (Ip et al., 2002) (see below) indicate that these fatty acids may inhibit tumour growth through a reduced blood supply and this could apply to all tumours, not only those of the mammary. Inhibition of angiogenesis in prostate cancer patients by CLAs has not been reported but could explain the dramatic decrease in growth of transplanted tumours in animal models (Cesano et al., 1998; Scimeca, 1999). The underlying conditions and cell mechanisms relating to the possible pro-carcinogenic effects of CLAs on prostate cancer initiation and development (see above) must be investigated further if the use of these fatty acids in cancer therapy is to be considered. 3.3.3.3

In cell studies in vitro

Studies with immortalised prostate cancer cell lines in vitro tend to support the anti-cancer effects of CLAs and show that they inhibit cell proliferation, inhibit eicosanoid formation and induce apoptosis by enhancing p53 and attenuating bcl-2 expression in human prostate cancer (PC3) cell lines (Ochoa et al., 2004). The two main isomers of CLA appear to elicit different, specific effects on eicosanoid production and oncogene regulation in order to induce apoptosis. The trans-10, cis-12 isomer attenuated mainly eicosanoid formation and the LOX5 and COX-2 enzyme expression, whilst cis-9, trans-11 affected the expression of oncogenes p53, p21WAF1/CIP1 and bcl-2 in a pro-apoptotic manner. Palombo et al. (2002) also showed that both isomers of CLA had anti-proliferative and pro-apoptotic effects on human prostate cancer cells but these authors did not emphasise the specific differences between the isomers. The inhibition of proliferation and induction of apoptosis in prostate cancer cells (LNCaP) by CLAs has also been linked with positive and negative regulatory effects on specific protein kinase C (PKC) isoforms; mainly upregulation of PKC- and PKC- and downregulation of PKC- (Song et al., 2004). These are highly specific effects of CLAs on important cell signal kinases and emphasise the multifaceted role of nutrients like CLAs in cell regulation and anti-tumourigenesis.

BLUK122-Dijkstra

64

September 25, 2007

20:0

Trans Fatty Acids

Reports to date indicate that both cis-9, trans-11 and trans-10, cis-12 isomers of CLA appear to have beneficial effects in attenuating prostate cancer proliferation and progression but possibly through entirely different cellular mechanisms (lipid metabolism or oncogene expression). Further investigations into the role of CLAs in attenuating prostate cancer and identifcation of the underlying cellular mechanisms are urgently required, as this disease is increasing in the population and few therapies are available to counter it. As with breast cancer, clinical intervention studies with CLAs alone or as adjunct therapies to the classical therapies, which now include taxane and tamoxifen, are overdue, particularly as CLAs appear to have minimal side effects and do not appear to pose a significant hazard to health other than the reported oxidative formation of isoprostanes, which in cancer prevention and apoptosis may actually be beneficial in killing cancer cells.

3.3.4

CLAs in gastrointestinal cancer

3.3.4.1 Human studies Studies relating CLA intake to possible reduced risk (or increased risk) of colorectal cancer in man are greatly lacking. To our knowledge, only one very recent epidemiological study relating intake of dairy foods containing CLA with a reduction in gastrointestinal cancer risk has been published. The Swedish Mammography Cohort Study found that for each daily increment of dairy food (two servings per day increments), the risk of colorectal cancer was reduced by circa 13% but the risk of distal colon cancer was decreased by a very significant 34%. However, as in other milk and dairy food studies attempting to correlate consumption of dairy products with decreased breast cancer incidence/risk (see above), the authors conceded that it is difficult to attribute the effects entirely to CLA content of the product (Larsson et al., 2005). Clearly, these observations and those of the Finnish studies on breast cancer indicate a possible benefit in milk and dairy consumption. 3.3.4.2 Animal studies Beneficial effects of CLAs in gastrointestinal cancers have been observed mainly in chemically induced tumours in animal models. Such studies showed that CLA inhibited the benzo(a)pyrene-induced neoplasia in mouse forestomach (Ha et al., 1990). CLA also protected against 2-amino-3-methylimidazo [4,5-f]quinoline-induced colon carcinogenesis in F344 rats (Liew et al., 1995). CLA, at 1% of the diet, also reduced colon tumour incidence in 1,2-dimethylhydrazine (DMH)-treated rats (Park et al., 2001). This was the result of a 251% increase in the apoptotic index as determined by the terminal deoxynucleotide transferasemediated dUTP nick-end labelling method. The role of eicosanoids was evident because CLA also reduced mucosal levels of prostaglandin E2 (PGE2 ), thromboxane B2 and their precursor arachidonic acid. The effects were dose dependent. As with the mammary tumours induced by dietary PhIP (see above), this chemical also induced an 8–26-fold increase in mutation frequency in the distal colon of transgenic Big Blue rats, but the proximal colon and caecum were unaffected (Yang et al., 2003). Mutation frequencies were significantly reduced (14–24%) by dietary CLA supplementation in the rats. Contradictory observations have also been reported for the effects of CLAs in these as in other tumour types. In an interesting comparison with dietary fish-oil fatty acids in ApcMin mouse model of colon cancer, CLA did not reduce tumour load, whereas omega-3 LCPUFA from fish elicited significant beneficial effects (Petrik et al., 2000). This suggested that not all effects of omega-3 PUFA compare

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

65

precisely with those of CLA in animal models of the disease, but yet again, mice appear to be different to other rodents in their responsiveness to CLAs (see below). A recent study used individual diets containing one of the two main isomers of CLA at 1% (w/w) and fed them to an ApcMin mouse model of intestinal carcinogenesis. Surprisingly, the trans-10, cis-12 isomer was reported to actually promote rather than inhibit colon carcinogenesis (Rajakangas et al., 2003). It was postulated that promotion of carcinogenesis in these studies with mice was due to activation of the oxidative stress-induced nuclear factor B (NF-B) transcription factor pathway and cyclin D1 , although the authors could not demonstrate NF-B activation by increased p65 protein binding to DNA in mouse tissues. These findings contrast with the reported inhibitory effects of CLAs on NF-B activation (IB-phosphorylation; DNA binding and transfection/reporter studies) in human umbilical vein endothelial cells (HUVEC) and prostate cancer cells in vitro (see below). This group also reported adverse effects of trans-10, cis-12 CLA because of increased lipid peroxidation determined by increased isoprostane excretion in urine (see above). Unfortunately, an enhancement of carcinogenesis with the increase in trans-10, cis-12 isomer excretion that also correlated with an increase in PGE2 formation was not reported. An increase in peroxidative tone elicited by the trans-10, cis-12 isomer could enhance prostaglandin production that might be expected to elicit pro-carcinogenic effects (see above). However, most studies to date show inhibition of eicosanoid production with CLA treatment in various tissues/cells in vivo and in vitro (see below). In contrast to the above observations, a study of the effects of CLAs on DMH-induced colon carcinogenesis in Sprague Dawley rats found decreased PGE2 , TXB2 and an increased apoptotic index which suggested that the beneficial effects of CLAs on colon carcinogenesis could be mediated through cell signalling systems in the rat colonic mucosal cells (Kim et al., 2003). These observations were supported by a recent study in rats where CLA fed at 1% of the diet resulted in a decrease in implanted human colon tumour development and peritoneal metastasis that correlated with a decrease in PGE2 levels and an increase in the pro-apoptotic:anti-apoptotic Bax:Bcl-2 ratio (Park et al., 2004). 3.3.4.3 Studies with cancer cells in vitro Similar effects of CLA on the expression of the oncogenes Bax:Bcl-2 and p53,p21WAF1/CIP1 to those reported for mouse tissues above have been reported by our group in breast and prostate cancer cells in vitro (Majumder et al., 2002; Ochoa et al., 2004; Song et al., 2006; Wahle et al., 2004). In contrast to the findings in mice in vivo, where CLAs activated the NF-B pathway, our group has also shown that CLAs can actually inhibit the activation of NF-B at the stage of IB phosphorylation in human prostate cancer cells (and in HUVEC) in vitro (Song et al., 2006). In these endothelial cells, CLAs also reduced the oxidative and inflammatory stress-induced expression of adhesion molecules, the gene reporters of which are a major target for activated NF-B (Goua and Wahle, 2007; Sneddon et al., 2006). This may again reflect differences due to tissue, species (mice appear to respond differently to CLAs than other animals) or to cell versus whole animal studies. Significant anti-proliferative effects on human HT-29 colorectal cancer cells in vitro were elicited by CLA isomers through their inhibition of DNA synthesis and induction of apoptosis (Cho et al., 2003). CLA also inhibited the heregulin--induced phosphorylation of ErbB2 and ErbB3, recruitment of p85 subunit of PI3-kinase to the ErbB3 receptor, ErbB3 associated PI3-kinase activities and phosphorylation of Akt. CLA decreased ErbB2 and ErbB3 mRNA

BLUK122-Dijkstra

66

September 25, 2007

20:0

Trans Fatty Acids

and protein expression in a dose-dependent manner. Further studies suggested that the CLAassociated benefits could be related to their ability to inhibit insulin-like growth factor-II (IGF-II) expression and attenuate the kinase 1/2 pathway and IGF-receptor signalling in these cells (Kim et al., 2003). In Caco2 cells, this group showed that the effects on IGF-II regulation were isomer specific, with the trans-10, cis-12 isomer being effective but not the cis-9, trans-11 isomer (Kim et al., 2002). Similarly, it was only the trans-10, cis-12 isomer that decreased ErbB3 expression in HT29 human colon cancer cells (Cho et al., 2005). It was suggested that these inhibitory effects of CLAs were mediated through effects on IGF-II, Akt activation and the inhibition of the cell-cycle-regulatory protein p21CIP1/WAF1, the latter by physiological concentrations of the fatty acid (Cho et al., 2003; Kim et al., 2003; Lim et al., 2005). This pathway is also involved in the inhibitory effects of CLA on breast cancer and prostate cancer cells, as shown previously by our group (Majumder et al., 2002; Ochoa et al., 2004). A recent study again showed that trans-10, cis-12 but not cis-9, trans-11 or linoleic acid elicited an inhibition of cell proliferation and induced apoptosis by inducing the expression of another pro-apoptotic gene, non-steroidal anti-inflammatory drug-activated gene-1 (NAG-1) in human colorectal cancer cells (Lee et al., 2006). Further studies, both in vivo in animal models of prostate cancer and in vitro, with clearly defined CLA isomers and mixtures and varying amounts are required before the anti-carcinogenic effects of CLAs, either one or both of the isomers, on prostate cancer can be regarded as unequivocal. Suffice it to say that current findings, particularly with cells in culture, are encouraging. The reason for the distinctive differences between isomers in relation to inhibiting or promoting tumourigenesis and metastasis in rat and mice models respectively requires further elucidation.

3.3.5

CLAs and other cancers (hepatic, pancreatic and dermal)

3.3.5.1 In man Epidemiological studies on intake of CLAs and hepatic, pancreatic or epidermal cancer risk have, to our knowledge, not been reported in man. 3.3.5.2 In animals In an animal model of liver cancer, CLAs at 0.5 and 2.0% (w/w) of the diet actually enhanced the growth of transplanted rat hepatoma dRLh-84 cells in rats compared with control animals (Yamasaki et al., 2002). Surprisingly, this was despite a reduction in PGE2 and COX-2 elicited by CLA in the tumour. These effects are usually associated with decreased tumourigenesis. Similarly, CLAs did not improve body weight (inhibition of cachexia) or the nutritional status of rats implanted with the Morris 7777 hepatoma (McCarthy-Beckett, 2002). These negative effects of CLAs in rodents in vivo are not clearly understood at present. They give cause for concern regarding the possible use of these compounds in cancer therapy and require substantiation. Despite these in vivo observations, CLAs were strongly cytotoxic to the same rat dRLh-84 hepatoma cells in vitro (see above) at concentrations as low as 1M when compared to control cells, and the main active isomer was again trans-10, cis12 and not cis-9, trans-11 (Yamasaki et al., 2002). The active isomer also increased the sub-G1 population of cells, activated caspase-3 and -9 (indicators of cell-cycle inhibition and apoptosis, respectively) and elicited a time-dependent cleavage of poly(ADP-ribose) polymerase. These cytotoxic effects of CLAs were weaker in normal hepatocytes, indicating a

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

67

possible enhanced sensitivity to CLAs in cancer cells. CLAs also inhibited growth of a human hepatoma cell line (HepG2) in vitro. Effects were not due to increased lipid peroxidation in the cells but apparently to an altered fatty acid metabolism (Igarishi and Miyazawa, 2001). The trans-10, cis-12, but not the cis-9, trans-11 isomer of CLA apparently inhibited the delta-9, -6 and -5 desaturation of linoleic and -linolenic acids in human HepG2 cells (Eder et al., 2003). Authors estimated desaturase activities from the relative concentrations of precursors and products and did not determine the enzyme activity directly. This surrogate method of assessing desaturase enzyme activity can be inaccurate because of the dynamics between the reaction components being affected by other reactions like oxidation, esterification and compartmentalisation of substrates and products. Trans-10, cis-12 CLA also increased the prostaglandin production in HepG2 cells at 100 M concentrations, whilst cis-9, trans-11 was without effect. This finding correlates with the observation in vivo in implanted rat hepatoma dRLh-84 and Morris 7777 (McCarthy-Beckett, 2002; Yamasaki et al., 2002). However, these negative effects contrast significantly with a number of reports that clearly show that CLAs inhibit eicosanoid production in both cancer and normal cells, as well as with the report from the same authors which shows that the enzymatic pathway for the formation of the eicosanoid precursor arachidonic acid (ARA) is decreased (see above). Enhanced eicosanoid formation is generally regarded as a characteristic of malignancy and its reduction is a central tenet for anti-tumour effects of CLAs (Fisher, 1995). This is not the only observation of enhanced prostaglandin formation rather than inhibition in the presence of CLAs. It has also been observed in HUVEC and blood platelets when cells are treated with high concentrations of CLAs (see below in vascular section). These differences in prostaglandin formation elicited by CLAs may reflect relative isomer concentrations or total CLA concentration. Dietary CLA at 1% (w/w of diet) also induced hepatocyte proliferation as well as apoptosis in F33 rats (Lu et al., 2000) and ornithine decarboxylase activity in mouse liver, the latter through activation of hepatic peroxisome proliferator-activated receptor (PPAR) (Belury, 2002a,b). Findings contrast with those observed in many other cell types and may be a reflection of specific CLA effects on liver. The anti-carcinogenic properties of CLAs were first identified using chemical initiation of carcinogenesis in the skin of mice and its attenuation by topical application of the fatty acid (Ha et al., 1987). Induction of skin tumours in mice by phorbol esters, as determined by hyperplasia and ornithine decarboxylase activity, was not attenuated by dietary CLA, despite the observation that c-myc mRNA and prostaglandin formation (pro-carcinogenic factors) were reduced (Kavanaugh et al., 1999). Table 3.2 summarises some of the published observations, positive, negative and neutral, relating to the effects of CLAs on aspects of carcinogenesis

3.4

CELLULAR MECHANISMS OF CLAs’ ANTI-CANCER EFFECTS

Regression of an established tumour may occur because of either a decrease in cellular proliferation, an increase in programmed cell death (apoptosis) or necrosis of the tumour due to nutrient and oxygen deprivation (inhibition of tumour angiogenesis). CLAs, both the 9,11 isomer and the mixture, are capable of eliciting an increase in apoptosis in normal rat mammary epithelium in a primary culture system in addition to reducing cellular proliferation (Ip et al., 1999). Other studies assessing the effects of CLA in a rat mammary tumour cell

BLUK122-Dijkstra

68

September 25, 2007

20:0

Trans Fatty Acids

Table 3.2 Summary of some of the reported findings, positive and negative, relating to effects of CLAs on various forms of cancer. Breast cancer Human studies

No relationship between cancer risk and dairy products intake ↑ cancer risk with dairy products intake ↓ cancer risk with dairy products intake No effects on tumour growth No relationship between cancer risk and CLA intake ↓ CLA serum levels in woman with cancer No relationship between CLA content in adipose tissue and some prognostic indicators

Prostate cancer Animal models

In vitro studies

↓ tumour volume ↓ mutagenesis No inhibition of cancer growth ↓ cell proliferation and ↑ apoptosis ↑PKC- and -, ↓PKC-

Gastrointestinal cancer Human studies ↓ cancer risk Animal studies ↓ induced neoplasia Protection against induced carcinogenesis ↓ tumour incidence No effect on tumour load ↓ PGE2 , TXB2 and ↑ apoptotic index ↑ carcinogenesis (CLA t10,c12) Other cancers Animal studies

↑ growth of hepatoma cells in animal model No inhibition of cachexia in rats with hepatoma Cytotoxic effect on hepatoma cells in vitro ↑ hepatocyte proliferation and apoptosis No effect on induced skin tumour

Chajes et al. (2002); Voorips et al. (2002) Aro et al. (2002); Knekt and Jarvinnen (1999) Landa et al. (1994); Salamini et al. (1984) Wong et al. (1997) McCann et al. (2004) Ledoux et al. (2003) Chajes et al. (2003)

Cesano et al. (1998) Cohen et al. (2003) Cesano et al. (1998) Ochoa et al. (2004); Palombo et al. (2002) Song et al. (2004) Larsson et al. (2005) Ha et al. (1990) Liew et al. (1995) Park et al. (2001) Petrik et al. (2000) Kim et al. (2003); Park et al. (2004) Rajakangas et al. (2003) Yamasaki et al. (2002) McCarthy-Beckett (2002) Igarishi and Miyazawa (2001); Yamasaki et al. (2002) Lu et al. (2000) Belury (2002a,b)

line (NMU) showed an increase in apoptosis as assessed by light microscopy and a DNA laddering technique (Ip et al., 2000). This study also determined the effects of feeding rats in vivo with CLA and observing its effects on pre-malignant lesions within the mammary tissue (Ip et al., 2000). However, CLA did not stimulate apoptosis in either the alveoli or the TEB cells that are both normal components of the mammary tree. It was not possible to determine which of the isomers had the greatest effect. The finding by these authors that feeding female rats on CLA until puberty reduced the susceptibility of mammary tissue to carcinogen-induced tumour development indicated a possible beneficial, protective effect of CLA on mammary tissue development (Ip et al., 1995). This is an interesting observation indicating the possibility that CLA might influence developmental mechanisms in mammary tissue and if shown to be equally effective in human pre-pubertal females, could be exploited as a dietary preventative therapy, particularly in individuals at risk of developing the disease. Whilst these studies have focused on breast cancer, others have examined the effects of CLA on apoptosis in experimental models of colon cancer. In colon cancer cells in culture

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

69

(SW480), CLA treatment will also result in apoptosis (Miller et al., 2002). Dietary supplementation with 1% CLA (mixture of isomers) to rats resulted in a significant decrease in colon cancers induced by DMH. However, a key point to note here was that this study also examined the effects on apoptosis in ‘normal’ colonic epithelium before the tumour was actually visible. The apoptotic index more than doubled in the colonic epithelium of CLAsupplemented animals (0.5–1.5% of the diet). This may explain, in part, the effects of CLA in preventing colon cancer development in these animals. The mechanisms of these anti-tumour effects have not been clarified. However, preliminary studies have revealed some important insights that may start to explain the molecular basis for anti-tumour activity of CLAs, in particular, with regard to their effects in reducing cellular proliferation and increasing apoptosis, as described above. Recent research has focused on the molecular mechanisms underlying the control of these pathways. As mentioned previously, our group observed that, in a prostate cancer cell line (LNCaP), CLAs, whilst reducing proliferation and enhancing apoptosis, also altered the specific PKC isoform activities in cell. This was determined as an increased expression of the different isoforms in cell membranes as opposed to cytosol. There was an increased PKC-, PKC- and PKC- but a decreased PKC- level (Song et al., 2001, 2004), and this may result in inhibition of cell growth. However, others found no effect on PKC isoforms (Masso-Welch et al., 2001). These differences may be due to differences in tumour-type, species differences or the ability of the various techniques used to detect the isoforms (Wahle and Heys, 2002). The effects of CLAs on the pro- and anti-apoptotic pathways and their controlling genes have also revealed some interesting key facts. The expression of bcl-2, a key anti-apoptotic proto-oncogene, was decreased in rat mammary tumours and tumour cells by feeding or treating with CLAs; other oncogenes involved in apoptosis such as bax or bak (Banni et al., 2003; Ip et al. 1999, 2000) were either not affected or not determined (e.g. p53, p21WAF1/CIP1, bad and bcl-X). CLA was found to elicit an oncostatic, cell-cycle inhibitory effect on human, oestrogen-receptor-positive MCF-7 breast cancer cells, whereas the parent linoleic acid had the opposing effect (Durgam and Fernandes, 1997). This is another example of the pro-carcinogenic effect of the parent omega-6 fatty acid, linoleic acid, and it is tempting to speculate that high cancer incidence could be related to high linoleic intake. In contrast to the MCF-7 cells, the ER-negative MDA-MB-231 cells were unresponsive to CLA, indicating a possible involvement of the oestrogen receptors in CLA effects. The responsiveness to CLA of the two types of breast cancer cells was mirrored by a decreased expression of c-myc, a major cell-cycle-regulatory oncogene present in MCF-7 but not in MDA-MB-231 cells, which is also involved in anti-apoptosis. Our group (Majumder et al., 2002; Ochoa et al., 2004) conducted a detailed evaluation of the effects of a mix of CLAs and individual isomers on the expression of proand anti-apoptotic oncogenes in human breast cancer and prostate cancer cells. Oestrogensensitive MCF-7 and oestrogen-insensitive MDA-MB-231 breast cancer cells and androgensensitive LNCaP prostate cancer cells were studied. Expression of major oncogenes (p53, p21 WAF1/CIP1, bcl-2, Bax and bcl-Xs) were determined at the transcriptional (mRNA) and translational (protein) level, using northern- and western blotting and specific ELISAs. Treatment with CLA inhibited proliferation and induced apoptosis, and this correlated strongly with an increased expression (mRNA and protein) of pro-apoptotic p53 and p21 WAF1/CIP1 genes by circa two- to fivefold but a reduced expression of anti-apoptotic bcl-2 by circa 20–30% in MCF-7 cells. MCF-10a cells that are regarded as normal or benign cells (not malignant) were not affected. The bcl-2 protein expression in the malignant cells correlated with the expression of this protein in rat tumour tissue (Ip et al., 1995, 1999, 2000),

BLUK122-Dijkstra

70

September 25, 2007

20:0

Trans Fatty Acids

but these authors did not investigate gene (mRNA) expression in either animal or human tissues. In oestrogen-insensitive MBA-MD-231cells that express the mutant form of p53, which prevents apoptosis, p53 mRNA was unaffected by CLAs, but p21WAF1/CIP1 and the anti-apoptotic bcl-2 mRNAs were increased. The protein expression of the oncogenes largely reflected the mRNA expression. This suggested that CLAs might not be pro-apoptotic in these cells, yet they inhibited proliferation and induced apoptosis. This correlated with the inhibition of mutant p53 and an increase in p21WAF1/CIP1. Further investigation of oncogene protein expression revealed an increase in Bax and Bcl-Xs proteins. The ratio of Bcl-2 to Bax and Bcl-Xs was decreased, and this favoured apoptosis despite the reduction in Bcl2 protein. Similar effects on p53, p21WAF1/CIP1 and bcl-2 gene expression (mRNA and protein) by CLA mixture and individual isomers to the above were observed by our group in prostate cancer cells (Ochoa et al. 2004). In these studies, the cis-9, trans-12 isomer was most effective in modulating oncogene expression. The trans-10, cis-12 isomer markedly inhibited the eicosanoid production and COX gene expression; this would be expected to prevent tumour development. These observations clearly show that the two most important isomers of CLAs, in relation to nutritional and commercial CLAs, inhibit tumour growth and development through different cell pathways in prostate cancer cells. This adds support to the possible use of a mixture of isomers for maximum effects on tumour inhibition. Such findings offer a possible mechanistic explanation for the observed inhibitory effects of CLAs on transplanted human prostate cancer cells in animals (Cesano et al., 1998). Inhibition of COX and the concomitant inhibition of eicosanoid formation by CLAs in various cells/tissues have been reviewed by Belury et al. (Belury, 2002a; Belury et al., 2002; Belury and Vanden Heuvel, 1997). Effects may be due to displacement of arachidonic acid, the precursor of the two-series prostaglandins and four-series leukotrienes or attenuation of enzyme expression (COX-2 and LOX-5, -12 and -15). As mentioned above, certain concentrations of CLAs appear to enhance eicosanoid production, making it difficult to arrive at a consensus view as to their overall effects in different tissues/cells and on different diseases. It has been suggested that CLAs may exert some of their effects through metabolism of the parent isoforms to unique eicosanoids but this requires substantiation. The role of the nuclear transcription factors NF-B and PPARs in the putative beneficial effects elicited by CLAs on cell function, particularly the attenuation of stress-induced signal cascades, has been mentioned briefly (see above). NF-B activation can be inhibited by CLAs by a reduction in the level of IB phosphorylation on its serine residues and possibly reduction of binding to the B regions in respective gene promoters (see above, Fig. 3.2). We have previously postulated that CLAs (and omega-3 PUFA from fish) can induce the intrinsic redox enzymes in cells (GPx1 and GPx4) and that this may be the mechanism whereby transcription is affected (Crosby et al., 2001; Sneddon et al., 2003). It is also conceivable that CLAs elicit their effects directly on specific response elements in the affected genes or through the action of derived eicosanoids on these specific response elements. CLAs also activate the nuclear hormone receptors PPARs, particularly PPAR- , in various tissues and cell types (Belury, 2002a; Belury et al., 2002; Belury and Vanden Heuvel 1997). The ability of CLAs to activate PPAR- was reduced when delta-6 desaturase activity was inhibited by a specific inhibitor (SC-26196) in CV-1 cells. This indicates that a metabolite or metabolites of CLA, derived by desaturase, elongase and possible eicosanoid enzyme action, may be the active effector(s) of PPAR-elicited cell functions. CLAs can induce PPAR- -responsive genes and the level of the protein in vivo. Recent studies have suggested that activators of PPAR- may be protective against cancers of the mammary gland, colon and prostate (Sporn et al., 2001). Increased activation of PPARs can result in decreased activity of stress-induced

iκB

p65

p50

p65

Proteosomal degradation

p50 P

NF-κB binding

Kinase-activated phosphorylation

Nucleus

PPAR

?

Transcription

PPAR

??

GSSG

GSH

Redox enzymes

Cytokines Adhesion molecules Heat-shock proteins

GPx

CLA

September 25, 2007

Fig. 3.2 Some intracellular mechanisms whereby CLAs and fatty acids are likely to influence the expression of stress-induced genes. Stress-stimuli elicit a signal cascade which activates the inactive NF-B/IB complex in the cytoplasm that in turn releases active NF-B which translocates to the nucleus and binds specific B response elements in the promoter regions of various genes. These include genes for adhesion molecules, cytokines, redox enzymes, heat shock proteins, cyclooxygenases etc. Fatty acids like n-3 PUFA can result in increases in conjugated derivatives (CLAs) and can upregulate redox enzyme gene expression which can regulate the activation and/or nuclear binding of transcription factors like NF-B. PPAR activation by fatty acids, particularly CLAs, may also play a role in regulating NF-B activity.

Positive regulation

Negative regulation

Oxidative/inflammatory stimulus

BLUK122-Dijkstra 20:0

Conjugated linoleic acid effects 71

BLUK122-Dijkstra

72

September 25, 2007

20:0

Trans Fatty Acids

Table 3.3 apoptosis.

Summary of some reported effects, positive and negative, of CLAs on cell mechanisms and

↑ apoptosis in normal rat mammary epithelium ↑ apoptosis in rat mammary tumour cells ↑ PKC-, PKC- and PKC-, decreased PKC- No effect on PKC isoforms ↓ bcl-2 expression ↓ cell cycle of ER-positive MCF-7 breast cancer cells No effect on ER-negative MDA-MB-231 cells ↓ proliferation, ↑ apoptosis, ↑ expression of pro-apoptotic genes in MCF-7 cells No effect on MCF-10a No effect on p53 mRNA, ↑ p21 WAF1/CIP1 and blc-2 mRNA in ER-insensitive MBA-MD-231 cells and prostate cancer cells ↑ redox enzymes in cells Activation of the nuclear hormone receptors PPARs

Ip et al. (1999) Ip et al. (2000) Song et al. (2001, 2004) Masso-Welch et al. (2001) Banni et al. (2003); Ip et al. (1999, 2000) Durgam and Fernandes (1997) Durgam and Fernandes (1997) Majumder et al. (2002); Ochoa et al. (2004) Majumder et al. (2002); Ochoa et al. (2004) Ip et al. (1999, 1995, 2000); Ochoa et al. (2004) Crosby et al. (1996); Sneddon et al. (2003) Belury (2002a); Belury et al. (2002); Belury and Vanden Heuvel (1997)

NF-B, possibly through direct interaction of the two transcription factors where the PPAR resembles IB in its action (see above). It is therefore conceivable that activation of PPAR- is an important mechanism that may explain some of the anti-cancer effects of CLAs. This hypothesis, whilst appealing, requires substantiation. Table 3.3 summarises some of the published observations, both positive and negative, relating to CLA effects on cell signalling mechanisms in carcinogenesis and tumour growth.

3.4.1

Inhibition of angiogenesis

Development of new blood vessels (neovascularisation or angiogenesis) is of pivotal importance to ensure growth and maintenance of all types of tumours through adequate supply of oxygen, growth factors and nutrients. This process is also important for growth and maintenance of complex atherosclerotic plaques. Inhibition of angiogenesis would be expected to be an excellent therapeutic objective that could reduce rapid tumour growth and plaque progression. Masso-Welch et al. (2002) are, to our knowledge, the first group to show that CLA feeding reduced aspects of angiogenesis in vivo in implanted rat breast tumours when CD2/F(1) mice were fed either 1% or 2% CLA for 6 weeks. The CLA-fed mice had lower serum and mammary tissue vascular endothelial growth factor (VEGF) concentrations, an essential factor for growth of new blood vessels. A further study from this group (MassoWelch et al., 2001, 2002, 2004) again found that both isomers of CLA, when fed at 0.5 and 1% of the diet, decreased VEGF concentration and inhibited angiogenesis in vivo in the CD2/F91 mice. Surprisingly, they found that leptin, which is also a pro-angiogenic factor as well as important in energy regulation and obesity, was only decreased by the trans-10, cis-12 isomer of CLA. The mechanisms elicited by different CLAs appear to be multifactorial even when the end point is the same. These authors also observed the inhibition of angiogenesis in an in vitro model by both cis-9, trans-11 and trans-10, cis-12 isomers and

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

73

attributed these effects to the attenuation of VEGF expression and the reduction in expression of its receptor Flk-1. Fibroblasts are an essential requisite for angiogenesis to occur, possibly through the production of basic fibroblast growth factor (bFGF). CLA has been shown to inhibit bFGF-induced angiogenesis in vivo, as well as the endothelial cell proliferation and DNA synthesis induced by this growth factor in vitro (Moon et al., 2003). It is also conceivable that the effects of CLAs on the regression of atherosclerotic plaques reported by Kritchevsky’s group (Kritchevsky et al., 2004) and Toomey et al. (2006) are the result, at least in part, of a reduction in angiogenesis, but this requires substantiation. Increased angiogenesis is also a problem in diabetic retinopathy where it can result in retinal malfunction (Oettgen, 2001). CLAs have been reported to decrease hyperglycaemia and increase insulin sensitivity in diabetes (but also to elicit the opposite effect; see below). No study of the possible amelioration of neovascularisation by CLAs in retinopathy has been conducted to our knowledge. The cellular mechanisms whereby CLAs can reduce angiogenesis warrant further investigation in the light of possible preventative clinical applications in cancer, diabetic retinopathy and atherosclerosis.

3.4.2

Attenuation of cancer metastasis

Cancer involves three major stages, initiation, tumour progression/growth and metastasis, and CLAs have been implicated in the attenuation of all three, particularly with regard to mammary cancer in rodent, non-murine, models (Belury, 2002a,b; Ip et al., 1994, 1999, 2000, 2002). Two stages have been discussed in preceding sections above. The difficult stage to deal with clinically and the one with the greatest threat to life is the latter stage of metastasis since this is responsible for the spread of malignancy throughout the body and for producing multiloci in numerous tissues that are difficult to treat with conventional therapies. CLAs, both the main isomers and a mix thereof, at concentrations of 0.5 and 1% (w/w) of the diet, had a significant and dose-dependent inhibitory effect on pulmonary tumour burden, an index of metastasis, in mice with transplantable tumours (Hubbard et al., 2000). The latency period or suppression of tumours was also enhanced significantly compared to mice not receiving CLAs. Surprisingly, the efficacy of CLAs was greater than indomethacin, a recognised inhibitor of eicosanoid synthesis and suppressor of tumour growth and metastasis, indicating the recognised role of prostaglandins in the metastatic progression of many tumour types. A recent study further substantiated the anti-metastatic effects of CLAs through inhibition of peritoneal metastasis of human gastrointestinal cancer cells in an animal model (Kuniyasu et al., 2006). Similar effects in human metastatic patients would offer a novel therapy, but such studies have not been conducted. Initiation of metastasis involves increased adhesion of circulating cancer cells to the vascular endothelium of the tissue capillaries prior to extravasation and tissue invasion. This resembles the adhesion mechanism utilised by circulating monocytes during the initiation of atherogenesis (see below). We and others have shown that CLAs can reduce the expression of adhesion molecules at the level of gene expression as well as the actual adhesion of cells on monolayers of HUVEC in culture (see below). The reduced metastasis in tumourbearing animals fed CLAs could be explained by such an attenuation of adhesion molecule expression elicited by the dietary CLAs. This is further supported by observations that CLAs can attenuate the expression of the stress cascade through decreasing the activation of the NF-B, which in turn is responsible for the upregulation of adhesion molecules on vascular endothelial cells (Goua, 2003). Omega-3 LCPUFAs from fish oil have similar anti-metastatic

BLUK122-Dijkstra

74

September 25, 2007

20:0

Trans Fatty Acids

effects in animal models implanted with malignant cells (Adams et al., 1990; Karmali et al., 1984).

3.4.3

Reduction of cancer cachexia

Prevention of cachexia in cancer patients and anorexic patients would enhance their quality of life and well-being and allow more time for conventional therapies to be effective. Omega-3 PUFAs have been shown to attenuate cachexia in animal models of cancer (Wigmore et al., 2000) and in human pancreatic cancer (Barber et al., 2001) and prostate cancer patients (S.D. Heys, S. McClinton & K.W.J. Wahle, unpublished observations). The observed effects are regarded as largely due to the attenuation of the inflammatory cytokine tumour necrosis factor (TNF-) and a specific cachectic factor. CLAs also reduce TNF- production in immune cells at the level of gene expression (see below). A recent study reported a similar decrease in cachexia, in TNF- formation by macrophages and an attenuation in inflammatory cytokine formation by endotoxin-challenged splenocytes in a mouse model of cachexia treated with CLA (Yang and Cook, 2003). These observations illustrate a similarity between omega-3 PUFA and CLA effects on certain cells and tissues that could also explain in part their similarities in conferring certain health benefits. Interleukin-6 (IL-6) was also postulated to play a role in cancer cachexia induction due to its effect in reducing lipoprotein lipase activity in mice and in 3T3-L1 adipocytes (Greenberg et al., 1992). In concert with IL-1 and TNF, it is believed that IL-6 induces weight loss mainly through effects on eicosanoid production (Cook et al., 2003). The anti-cachectic effects of CLAs in animal models of the disease (Cook et al., 2003) suggest that they also inhibit IL-6, but direct evidence for this is currently lacking. In contrast, IL-6 production in rat primary adipocytes was not affected by a mixture of CLAs consisting of 36.5% cis-9, trans-11 and 56.6% trans-10, cis-12 at concentrations from 100 nM to 1 mM (Ha et al., 2003). The treatment did, however, decrease leptin production in the adipocytes, but surprisingly, TNF- formation was increased. This contrasts with the observations that CLAs attenuate inflammatory cytokine production in vivo in animals (see below) and more recently in man (Tricon et al., 2004, 2005). Why these differences occur is not apparent at present, but it highlights once again the complexity of in vivo versus in vitro responses to CLAs mentioned above. Table 3.4 summarises some of the reported variable effects of CLAs on angiogenesis, metastasis and cachexia. Table 3.4

Summary of effects of CLAs on angiogenesis, metastasis and cachexia.

Angiogenesis ↓ angiogenesis in vivo and in vitro ↓ angiogenesis and endothelial cell proliferation and DNA synthesis

Masso-Welch et al. (2002, 2004) Moon et al. (2003)

Metastasis ↓ metastasis and ↑ latency period and suppression of tumour in pulmonary tumour ↓ metastasis in human gastrointestinal cancer cells in animal models

Kuniyasu et al. (2006)

Cachexia ↓ cachexia and TNF- production by macrophages in mouse ↓ cachexia in animal model No effect on IL-6 production and ↑ TNF- in rat adipocytes ↓ inflammatory cytokines production in vivo in animals and humans

Cook et al. (2003) Cook et al. (2003) Ha et al. (2003) Tricon et al. (2004, 2005)

Hubbard et al. (2000)

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

3.5

75

EFFECT OF CLAs ON BODY COMPOSITION AND ENERGY METABOLISM IN ANIMALS AND MEN

One of the major interests in the health effects of CLAs relate to their perceived ability to regulate energy metabolism and reduce adiposity. These effects are relatively well established in various animal species, but their role in simple dietary weight reduction strategies is less clear (Bhattacharya et al., 2006; Park and Pariza, 2007; Tricon et al., 2005; Wahle et al., 2004).

3.5.1

Body composition in animals

The effect of CLA in attenuating body fat accumulation appears to be dose responsive and independent of dietary fat content (Delany et al., 1999; West et al., 1998). Dietary intake of CLAs at 0.5–2.0 g/100 g of diet clearly reduced body fat content and in some cases, increased lean body mass in growing animals, like mice, rats and pigs (Bhattacharya et al., 2006; Keim, 2003; Salas-Salvado et al., 2006; Tricon et al., 2005; Wahle et al., 2004). The proposed mechanisms to explain these effects of CLAs include increased lipolysis, enhanced fatty acid oxidation or reduced fatty acid uptake into adipocytes and decreased adipocyte size, decreased energy intake or inhibition of lipogenic enzymes. CLA upregulation of uncoupling protein-2 in white adipose tissue offered another mechanism to explain the increased energy expenditure on CLA feeding in high-fat-fed rats (Choi et al., 2004). These effects are reported to be independent of any changes in food intake. The anabolic effects appear to precede the reduction in body fat (Bhattacharya et al., 2006; Keim, 2003; Park and Pariza, 2007; Tricon et al., 2005; Wahle et al., 2004; Wang and Jones, 2004). CLA-enriched diets reduced adipocyte cell size rather than numbers in rats (Azain et al., 2000). This contrasts with findings in mice where CLA (1% wt/wt) induced apoptosis in adipocytes and reduced white adipose tissue mass (Tsuboyama-Kasaoka et al., 2000). However, in 3T3-L1 adipocytes in vitro, adipocyte proliferation and differentiation was significantly inhibited by CLA (Brodie et al., 1999). The efficacy of CLAs in changing body composition depends on the species and even the specific strain of laboratory animals. Mice are the most responsive to CLAs with levels of 0.5% of diet eliciting between circa 40 and 80% decrease in body fat (Keim, 2003). Increased lean mass and protein accretion was also observed in some studies with mice, sometimes, but not always, without any changes in body weight (Keim, 2003). Rats respond to CLAs in a similar manner to mice but accrete less body fat during active growth phase, and the magnitude of the decreased accretion is less than that observed in mice and is dependent on strain and the specific anatomical site of the adipose tissue (Azain et al., 2000; Keim 2003). CLAs as 0.5% of diet elicited a reduction of circa 25–30% in peritoneal/parametrial fat pad weight in Sprague Dawley rats, only 14% in omental fat but 44% in the epididymal fat pad of Long-Evans rats (Azain et al., 2000; Keim, 2003; Rahman et al., 2001). The effects of 0.5% dietary CLAs in Zucker rats were not as expected. A reduction in retroperitoneal fat pad in the lean phenotype but an increase in the obese phenotype when compared with its control was opposite to the expected effect (Keim, 2003; Sisk et al., 2001). The reasons for this difference are not clear at present but suggest that the genetic predisposition to fat accumulation cannot be attenuated and may even be augmented by CLAs in obese Zucker rats. It is tempting to speculate that the obese phenotype is physiologically equivalent to mature rats. Studies in adult, non-growing animals appear not to reproduce the same dramatic attenuation of fat accretion observed in young, growing animals (Bhattacharya et al., 2006; Tricon et al., 2005;

BLUK122-Dijkstra

76

September 25, 2007

20:0

Trans Fatty Acids

Wahle et al., 2004). Adult Wistar rats fed diets containing 1% of CLA mix or individual isomers in combination with moderate exercise for 6 weeks did not exhibit any effect on body composition or weight (Mirand et al., 2004). This group found that in mature, adult rats, whether sedentary or exercised, CLA isomers induced adipose tissue lipogenesis but without affecting adipose tissue weight (Faulconnier et al., 2004). However, feeding CLA mix to middle-aged C57BL/6 female mice for 10 weeks significantly decreased fat mass, prevented age-dependent loss of lean mass and maintained higher muscle weights compared to corn-oil-fed controls (Bhattacharya et al., 2006; Tricon et al., 2005; Wahle et al., 2004). Most of the earlier research on the body compositional effects of CLAs in animals used mixed isomers. Recent studies show that purified, individual isomers of CLA, cis-9, trans-11 and trans-10, cis-12, have distinct effects on lipid metabolism in these animals. Accumulating evidence suggests that the trans-10, cis-12 isomer and not the cis-9, trans-11 isomer is the one that affects body composition through decreased adipocyte size, increased fatty acid oxidation, increased oxygen consumption and inhibition of lipogenic enzymes. A CLA mix that had the highest ratio of trans-10, cis-12 than cis-9, trans-11 elicited the greatest reduction in body fat in mice and hamsters (Bhattacharya et al., 2006; Gavino et al., 2000; Park et al., 1999; Salas-Salvado et al., 2006; Tricon et al., 2005; Wahle et al., 2004). The effects of CLAs, mainly the trans-19, cis-12 isomer, in modifying body composition through reduction of adiposity seem to be reasonably well established, at least in young, growing laboratory animals if not in older animals. The reduced effect of CLAs in older animals could explain the similar lack of significant effects observed in human volunteers.

3.5.2

Body composition in man

Despite the compelling evidence that CLAs can reduce body fat content in animals, at least in young, growing animals, surprisingly few studies have been conducted to ascertain if the same applies in human volunteers. This is particularly relevant when considering the potential impact on the increasing number of obese subjects in Western populations and the increase in the accompanying diseases of the metabolic (or dysmetabolic) syndrome with its attendant health costs. However, effects of CLAs on human body composition are sparse and contradictory, and most have not shown the significant reduction in body fat observed in animal studies. Small-scale, randomised clinical trials conducted in Norway were the first to investigate the effect of CLA supplementation on body composition in humans (Blankson et al., 2000; Thom et al., 2001). Physically active men and women (ten per group) took either 1.8 g/day of a CLA mix (Tonalin, Natural, Norway) or an olive-oil-control supplement for 12 weeks. No changes in body weight were observed, but the CLA group registered a small, circa 4% decrease in body fat percentage compared to the placebo group. It was interesting that the greatest effects on body fat reduction were found with trained athletes and the suggestion was that CLAs prevented the uptake of fat into the adipocyte. This would make CLAs ideal for the prevention of weight gain common after a weight-reducing regime. A similar study from this Norwegian group was first to examine a possible dose–response relationship between CLA and body fat mass in obese and overweight volunteers (Blankson et al., 2000). Groups were assigned to either a placebo supplement or a 1.7, 3.4, 5.1 and 6.8 g/day CLA mix again for 12 weeks. Body fat mass was significantly decreased in volunteers in the 3.4 and 6.8 g/day groups, but no change in body weight was found. The 6.8 g/day group observed a significant increase in strenuous exercise compared to the other groups. Why the 6.8 g/day group should increase its exercise rate is intriguing but difficult to explain unless CLAs at this upper level of intake influenced brain functions and induced a positive

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

77

mood. It is known that omega-3 LCPUFA can elicit specific changes in brain chemistry at the level of neurotransmitters (Hibbeln et al., 2006). The authors concluded that 3.4 g/day of CLA was sufficient to cause a reduction in body fat in overweight and obese volunteers over the 12-week period. In contrast, when 17 healthy women in a residential study, using a metabolic research centre, were supplemented with CLA (3.9 g/day), no significant effect on body composition or energy expenditure was observed when compared to a sunflower oil placebo. Fat mass changes in these women were variable and ranged from a gain of 0.9 kg to a loss of 3.2 kg (Zambell et al., 2000, 2001). Similarly, an intake of 2.1 g/day of CLA mix for 45 days in sedentary young, but mature, women did not induce any changes in body composition (Petridou et al., 2003). Vessby’s group in Uppsala, Sweden, determined the effect of 4.2 g/day of CLA on body weight and composition in men and women, using a double-blind, placebo-controlled protocol (Smedman and Vessby, 2001). CLA elicited a small reduction in body fat percentage by 1.1%. (Average loss was 0.7 kg compared with 0.2 kg in placebo group.) In a similar study with middle-aged, obese males receiving 4.2 g/day of CLA for 4 weeks there was a decrease in the sagittal abdominal diameter, suggesting a reduction in visceral fat (Riserus et al., 2001). Since visceral fat is a risk factor in the metabolic syndrome, it is conceivable that even a slight reduction due to CLA supplementation might confer benefit in this disease state (see below). A study from Greece with healthy volunteers fed 0.7 g/day for 4 weeks and then 1.4 g/day for another 4 weeks showed a decrease in fat mass in these individuals compared to controls fed a placebo (Mougios et al., 2001). Another randomised placebo-controlled trial in Wisconsin-Madison University was designed to assess the long-term effect (6 months) of 2.7 g/day of CLA intake on body fat and weight loss in obese volunteers (Atkinson, 1999). Body weight and fat mass was reduced by 2.5 and 1.0 kg respectively in both control and CLA groups. No significant effect of CLA was evident and contrasts with the above studies. Some studies with overweight men and women have provided individual CLA isomers as supplements in dairy products (milk drinks and butter) at levels of 1.5, 3.0 or 4.22 g/day for 4–18 weeks but did not find any differences in fat or lean mass. Observations from Gudmundsen’s group in Norway suggest that CLA intake in healthy or overweight, obese subjects undergoing moderate-to-hard exercise have reduced body fat and enhanced lean mass. This suggested that exercise might enhance the effects of CLA in reducing fat mass (Blankson et al., 2000; Thom et al., 2001). Another factor in animal experiments compared to human experiments is that the period of supplementation is often a far greater proportion of a rodent’s life than that of human volunteers, so that the prolonged supplementation period might be required in order to more correctly reflect the animal studies. Recent studies where long-term supplementation protocols with CLA were carried out for 12 and 24 months clearly showed reduced body fat mass in healthy, overweight volunteers and that the supplements were well tolerated (Gaullier et al., 2004, 2005). Clearly, the changes in body composition in man in response to CLA differ between published clinical trials and also between these trials and observations in animals. This is most likely due to the lower dosage (g/kg body weight) of CLAs used in human compared to animal studies and also the fact that growing animals may be more responsive than mature humans. The findings with mature, adult animals suggest that age plays a significant role in the susceptibility to the fat-reducing effects of CLAs. The majority of human CLA studies to date is for non-growing, mature volunteers. Studies with growing, non-mature human volunteers are urgently needed to answer this question. Furthermore, the marked effects of CLA on body composition in rodent models could be due to activation of brown adipose tissue in young animals. This is lost in older animals and does not occur in humans (Klingenberg and

BLUK122-Dijkstra

78

September 25, 2007

20:0

Trans Fatty Acids

Huang, 1999). Body composition in humans, unlike animals, cannot be measured directly, and various prediction equations are used to estimate lean and fat content of volunteers. These predictions have inherent errors. The fact that most of the published human trials used different predictions could be another reason for the differences observed between trials (Keim, 2003). Prediction errors typically range from 1.5 to 2.0 kg for body fat mass to circa 3% for percentage body fat. Hence, the decreases reported in the human trials of about 1.0 kg and 4% in body fat mass are difficult to assess with confidence (Keim, 2003). A greater number of longer term supplementation studies (12–24 months) with growing children/adolescents or with weight-losing people predisposed to weight regain using both a CLA isomer mix of the two main isomers and the individual isomers would clearly determine if CLAs have significant effects on body composition in man. Human studies with a greater number of volunteers than hitherto reported would also give a clearer answer to the question of whether CLAs affect body composition in man.

3.5.3

Possible mechanisms underlying reported changes in body composition

As might be expected from their whole-body effects, in animals, CLAs modulate the activity of key enzymes involved in fat metabolism, both fat storage and mobilization/oxidation. Adipose tissue in mice and rats fed CLAs have a reduced lipoprotein lipase activity and increased lipolysis with enhanced palmitoyl-carnitine acyltransferase (PCAT) activity in various tissues, suggesting a generally enhanced fat mobilisation and oxidation in the aetiology of body fat reduction (Keim, 2003). This is supported by the reports of higher energy expenditure and fat oxidation in CLA-supplemented mice (Keim, 2003; Ohnuki et al., 2001; West et al., 2000). Augmented energy loss does not appear to be related to changes in the expression of mitochondrial uncoupling proteins (Keim, 2003; West et al., 2000). The reduced lipoprotein lipase activity in CLA-supplemented rodents and reduced levels of triacylglycerol and glycerol inside the cells but increased glycerol outside the cells suggest that adipocyte replenishment is impaired (see above). Available evidence suggests that CLAs reduce body fat in animals not only by altering the key enzymes of lipid storage, mobilization and oxidation but also by reducing adipocyte proliferation and differentiation and stimulating apoptosis in pre-adipocytes (Brodie et al., 1999; Keim, 2003; Miner et al., 2001; Satory and Smith, 1999). Observations that feeding mice trans-10, cis-12 isomer markedly reduced leptin and to a lesser extent, adiponectin mRNA in adipose tissue compared with mice fed control or cis-9, trans-11 isomers suggest that the trans-10, cis-12 form may elicit its effects on body composition through changes in levels of these hormones (Warren et al., 2003).

3.5.4

Efficacy of different CLA isomers in regulating body composition

CLAs are usually given as a 50:50 mixture of the two main cis-9, trans-11 and trans-10, cis-12 isomers, although in natural food sources, the cis-9, trans-11 isomer predominates by 30–70:1 (McGuire et al., 1999b; Parodi, 1997, 2003) . However, the effects of the individual isomers on body composition can vary dramatically, and the available evidence from animal and cell culture work suggests that the cis-9, trans-11 isomer is the active anabolic agent responsible for weight gain (not fat accretion), whilst the trans-10, cis-12 isomer is the

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects Table 3.5

Summary of some reported effects of CLAs on body composition in animals and man.

In animals

↓ body fat accumulation ↓ body fat content and ↑ lean body mass ↓ adipocyte size in rats ↑ adipocyte apoptosis and ↓ adipose tissue mass in mice ↓ adipocyte proliferation and differentiation in vitro ↓ body fat in mice ↑ lean mass and protein accretion in mice ↓ adipose tissue in rats No effect on body composition in exercised rats ↓ fat mass and prevention of age-dependent loss of lean mass ↓ body fat in mice and hamster

In humans

No effect on body weight and ↓ in body fat percentage in active men and women No effect on body composition and energy expenditure in healthy women ↓ body fat percentage in men and women ↓ sagittal abdominal diameter in obese men ↓ fat mass in healthy volunteers ↓ body weight and fat mass No effects on fat and lean mass in overweight men and women ↓ body fat and ↑ lean mass in exercised healthy or overweight subjects ↓ body fat mass in healthy overweight volunteers ↑ energy expenditure and fat oxidation in mice

79

Delany et al. (1999); West et al. (1998) Bhattacharya et al. (2006); Keim (2003); Tricon et al. (2005); Wahle et al. (2004) Azain et al. (2000) Tsuboyama-Kasaoka et al. (2000) Brodie et al. (1999) Keim (2003) Azain et al. (2000); Keim (2003) Azain et al. (2000); Keim (2003); Rahman et al. (2001) Mirand et al. (2004) Bhattacharya et al. (2006); Tricon et al. (2005); Wahle et al. (2004) Bhattacharya et al. (2006); Gavino et al. (2000); Park et al. (1999); Tricon et al. (2005); Wahle et al. (2004) Blankson et al. (2000); Thom et al. (2001) Zambell et al. (2000, 2001) Smedman and Vessby (2001) Riserus et al. (2001) Mougios et al. (2001) Atkinson (1999)

Blankson et al. (2000); Thom et al. (2001) Gaullier et al. (2004, 2005) Keim (2003); Ohnuki et al. (2001); West et al. (2000)

effective catabolic agent leading to increased lipolysis and fat oxidation (McGuire et al., 1999b; Parodi, 1997, 2003). The specific effect of trans-10, cis-12 compared to cis-9, trans-11 on hormones regulating energy metabolism have been mentioned (see above; Warren et al., 2003). These isomers have different, sometimes opposing, effects on other physiological systems, and this may be related to their fat-metabolising effects (see below). Table 3.5 summarises some of the reported effects of CLAs on body composition in animals and man.

3.6

OTHER REPORTED HEALTH BENEFITS OF CLAs

CLAs have been reported to elicit both beneficial and detrimental effects on a variety of other important biomarkers of disease. Again, definitive conclusions are difficult to make because of the variety of types and concentrations of the isomers and the different systems studied.

BLUK122-Dijkstra

80

September 25, 2007

20:0

Trans Fatty Acids

3.6.1

Effects on insulin resistance and diabetes

The incidences of both diabetes and its forerunner impaired glucose tolerance are increasing worldwide and are beginning to affect younger populations (Belury, 2003; Belury and Vanden Heuvel, 1999; Houseknecht et al., 1998). The potential cost to the health services of treating this disease and its attendant morbidities such as macro- and microvascular disease, including coronary heart disease, retinopathies and gangrene, is enormous. Central to this disease is obesity, and modest lifestyle changes resulting in a small reduction in body weight (circa 7%) are associated with a significant reduction in the risk of developing diabetes in people at known risk (diabetes-prevention programme). Recent observations indicating that adipose tissue produces inflammatory cytokines as well as hormones like leptin and adiponectin involved in regulation of energy metabolism suggest that adipose tissue plays an important role in normal and aberrant energy homeostasis (Hotamisligil et al., 1995; Yamasaki et al., 2003). The increased inflammatory cytokine formation attributed to adipose tissue in obese people could explain the increased risk of diabetes and cardiovascular disease in this subpopulation, as cytokines increase inflammatory processes in vasculature and increase lipolysis which can lead to insulin resistance (Belury, 2003; Belury and Vanden Heuvel, 1999; Houseknecht et al., 1998). The effect of CLAs on adipose mass in animals and men has been discussed above. The results are controversial, particularly in men, suggesting that reduced adipocity may not be the mechanism that underlies the reported beneficial effect of CLAs on glucose tolerance and insulin resistance (see below). The observations that CLAs reduce adipocytesecreted leptin levels in plasma, a hormone that regulates food intake in diabetic Zucker rats, non-diabetic mice and in humans with type 2 diabetes through hypothalamic signalling pathways (Belury, 2003; Belury and Vanden Heuvel, 1999), may indicate regulation of the most important mechanism for diabetes prevention. Although CLAs reduce food intake in some reported animal experiments, they have no effect in others and no effect in man at a level of 3 g/day, suggesting that reduced food intake may not play a role in CLAs’ effects on glucose metabolism. The reduction in leptin levels with 3 g/day CLA was only transient (see above; Belury, 2003; Belury and Vanden Heuvel, 1999). Also, the known attenuation of inflammatory cytokine production by CLAs in immune cells (Yu et al., 2002), if occurring in adipocytes, could offer another explanation for the beneficial effects of CLAs on diabetes and insulin resistance on obese subjects. Increased insulin resistance appears to be associated with increased visceral adiposity, but a clear explanation why this should be the case has not been promulgated. Although still controversial, the reduction in adipose mass observed in some experiments with human volunteers receiving CLA supplements was most apparent in the visceral region that was also associated with a significant decrease in sagittal abdominal diameter (Riserus et al., 2001). The effects of CLAs on insulin resistance and glucose homeostasis are variable and dependent on species and the type of isomer present (Kelley and Erickson, 2003; Salas-Salvado et al., 2006). Zucker rats fed diets containing a 1.5 g mix of CLA isomers (circa 42% cis9, trans-11 and trans-10, cis-12) or 0.02 g/100 g troglitazone, an anti-diabetic drug and PPAR agonist, resulted in significant reductions in fasting plasma glucose, plasma insulin, triacylglycerols, unesterified fatty acids and leptin compared with control animals. These beneficial effects on impaired glucose metabolism and insulin resistance in Zucker rats were not due solely to reduced food intake (Belury, 2003; Belury and Vanden Heuvel, 1999). In contrast to effects in diabetic rats, chronic feeding of CLA to non-diabetic swine and mice resulted in slight increases in fasting serum glucose and/or insulin (Belury, 2003; Belury and Vanden Heuvel, 1999). Similar slightly adverse effects on insulin and glucose homeostasis

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects Table 3.6

81

Summary of effects of CLAs on insulin resistance and diabetes.

↓ adipocyte-secreted leptin levels in plasma ↓ fasting plasma glucose, plasma insulin, triacylglycerols, unesterified fatty acids and leptin ↑ fasting serum glucose and insulin in non-diabetic swine and mice ↑ serum insulin in healthy women and obese men

Belury (2003); Belury and Vanden Heuvel (1999); Houseknecht et al. (1998) Belury (2003); Belury and Vanden Heuvel (1999); Houseknecht et al. (1998) Belury (2003); Belury and Vanden Heuvel (1999); Houseknecht et al. (1998) Medina et al. (2000); Riserus et al. (2001)

were observed in human volunteers. Healthy women receiving 3.9 g/day CLA mix exhibited a 20% increase in serum insulin compared with controls (Medina et al., 2000), and obese men with metabolic syndrome had a threefold increase in serum insulin compared with a placebo-control group (Riserus et al., 2001). These reported, apparently adverse, effects of CLA in animals and men appear to be largely due to the presence of the trans-10, cis-12 isomer and contrast with the reported beneficial effects of the CLA mix in enhancing insulin sensitivity in rats (see above). Table 3.6 summarises some of the reported variable effects of CLAs on insulin resistance and diabetes.

3.6.2

Modulation of immune functions

Changes in different aspects of immune-inflammatory functions due to ageing, dietary deficiencies or excesses are implicated in the aetiology of common diseases. It is now well established that alterations in the levels of fatty acids have an impact on immune function and both physiological and pathophysiological states. Understanding the molecular interactions between different fatty acids and cell responses in the immune system, especially to differentiate between physical effects and the modulation of key molecules critical to cell function, has become important in our understanding of the balance between health and disease. The major route by which the fatty acid profile of blood lipids and cells can be altered is dietary. Thus, alteration of the dietary profile of fatty acids will be reflected in the fatty acids present in blood and those incorporated into cells. This is especially important in situations that result in profoundly increased levels of free fatty acids in blood, which include trauma, stress, intense exercise and diabetes. These disease states will in turn alter specific cell responses. An especially critical response that can acutely control the balance between health and disease is the immune response to any given situation. The effect of a wide variety of saturated and PUFAs on these immune processes has been extensively studied (Bhattacharya et al., 2006; Grimble, 1998; Harbige, 2003; Tricon et al., 2005; Wahle et al., 2004). However, very little information is available on the effects of the CLAs on immune function in animals and men. The present discussion will review the mechanisms by which CLAs may influence cell function with particular emphasis on immune responsiveness. As the immune response involves many diverse components that can be subdivided into discreet response phases, the discussion will evaluate the impact of CLAs on these individual response phases. 3.6.2.1

Acute phase or inflammatory immune functions

A potential site of interaction for fatty acids is the ‘acute phase response’. The initial response of the immune system to infectious agents, such as the outer cell wall endotoxin (lipopolysaccharide, LPS) of gram-negative bacteria, is the activation of circulating monocytes or

BLUK122-Dijkstra

82

September 25, 2007

20:0

Trans Fatty Acids

tissue-resident macrophages. Monocytes are thought to play a pivotal role during the acute phase response and initiate the repertoire of immunostimulatory responses; they are critical to the overall immune response in vivo. Activated monocytes release a variety of cytokines, including IL-1 and TNF-, which are the primary initiators of the inflammatory response of immune function. This phase is also important in the coordination and maintenance of subsequent phases of the immune response (Dinarello, 1997; Suffredini et al., 1999). IL-1 and TNF- can in turn, via activation of phospholipase A2 , induce the release of PUFA from membrane phospholipids and the formation of eicosanoids, i.e. prostaglandins and leukotrienes, which are formed in many tissues and are integral parts of inflammatory-type reactions. PGE2 is particularly important, as it produces the symptoms of inflammation, such as pain, oedema and fever in a systemic reaction (Dinarello, 1999). In this sequence of reactions the role of specific PUFA is particularly important; the major precursor of prostaglandins is ARA, unless other PUFAs, like eicosapentaenoic acid, have been incorporated into membrane phospholipids. The availability of ARA to the cyclooxygenase or lipoxygenase enzymes can then be altered depending on the presence and level of other, potentially competing, unsaturated fatty acids (Dinarello, 1999). The presence of these competing PUFAs is determined by dietary intake and membrane incorporation (Wahle and Rotondo, 1999). In this context it is intriguing that CLAs are reported to decrease the production of eicosanoids, particularly prostaglandins of the II series, in vivo and in cell cultures (Iwakiri et al., 2002; Ma et al., 2002; Nakanishi et al., 2003; Ogborn et al., 2003). This is not a universal effect, and some cell studies have reported an increase in prostaglandin formation in the presence of CLA (see below). It is not clear at present if the effect of CLAs in inhibiting prostaglandin formation is due entirely to their competition with ARA for substrate in the COX reaction. CLAs are apparently also able to suppress the gene expression of the COX enzymes, indicating a fundamental effect of these fatty acids on gene regulation (Iwakiri et al., 2002). Clearly, attenuation of PGE2 formation would reduce the symptoms of fever and pain accompanying inflammation. CLAs have also been reported to suppress the release of pro-inflammatory cytokines, particularly TNF-, in animals (Akahoshi et al., 2002, 2004; Yang and Cook, 2003). In the study of Akahoshi et al. (2004), a mixture of CLA isomers was able to reduce serum TNF- levels to about 50%, compared to the serum of mice given predominantly linoleic acid. In agreement with these observations, Yang and Cook (2003) showed that feeding CLAs, especially the cis-9, trans-11 isomer, suppressed the production of TNF- in vivo following the injection of LPS compared to a corn-oil-fed group of animals. The cis-9, trans-11 isomer was also able to suppress TNF- production directly when incubated with the macrophage cell-line RAW. This indicated that the observed dietary effects of CLAs may occur directly at the level of the macrophages/monocytes that are responsible for producing the majority of inflammatory cytokines. It is not certain how CLAs, especially the cis-9, trans-11 isomer, can lead to downregulation of pro-inflammatory cytokine production. However, the prime candidates would be those mechanisms that are known to be involved in the suppression of inflammatory cytokine responses. As indicated previously, PGE2 is formed at the end of the sequence of inflammatory mediators and it is also a potent immunosuppressive prostanoid. Pro-inflammatory cytokines appear to mutually induce their own synthesis. This positivefeedback loop appears to ensure mobilisation of the response cascade, possibly to prevent immune failure. The progression of this cytokine cascade can also be limited by the increased levels of PGE2 produced at the end stage of inflammation following sufficient activation of the inflammatory process and of phospholipase A2 (Davidson et al., 2001; Rotondo et al., 1988). Inflammatory immune reactions can thus be controlled and prevented from becoming detrimental to the host by PGE2 acting as a negative-feedback regulator formed at the end of

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

83

the cascade. This is a mechanism that may also occur in vivo. The circulating levels of PGE2 can increase almost tenfold during immunoactivation, indicating a massive upregulation of biosynthesis. LPS administered i.v. can lead to an increase in PGE2 levels to almost 450– 500 pg PGE2 /mL plasma compared to levels of 30–40 pg PGE2 /mL plasma in unstimulated animals given control saline treatment (Davidson et al., 2001; Rotondo et al., 1988). There are two key pieces of evidence that support this downregulatory action of PGE2 . Firstly, in the presence of exogenously added PGE2 , less TNF- and IL-1 is produced in response to LPS. Secondly, an increased production of these cytokines is observed in the presence of inhibitors of PGE2 synthesis, indicating that endogenous prostanoids are actively suppressing cytokine production. In order for CLAs to elicit their anti-inflammatory effects through this pathway, an increase in PGE2 or other inhibitory prostanoids would have to occur. However, it has been reported that CLAs, usually but not always, reduce prostaglandin production, particularly PGE2 , in a variety of metabolic conditions and in different tissues. In a polycystic kidney disease model in rats, dietary CLAs reduced the ex vivo production of PGE2 from interstitial cells (Ogborn et al., 2003). Similarly, CLAs were shown to directly reduce PGE2 release from mammary tumour cells in vitro, which was preceded by a reduction in the availability of ARA (Ma et al., 2002). This indicated a classical interference with the flux of ARA and thus its availability for PGE2 production. Most, but not all, studies to date have indicated a suppression of prostanoids by CLA (see above). Paradoxically, this would be expected to result in increased inflammatory cytokine production according to the mechanism described above. However, very few studies have investigated the effect of CLAs on cytokine-stimulated prostanoid production. One such study has confirmed that CLAs can directly reduce the production of PGI2 from HUVEC reticulated with IL-1 (Torres-Duarte and Vanderhoek, 2003). Antigen-induced PGE2 production from guinea-pig trachea derived from animals given dietary CLA was also reduced (Whigham et al., 2001). Thus, it would seem unlikely that the actions of CLAs, in reducing the release of pro-inflammatory cytokines, are mediated via an enhancement of prostanoid levels, and it would appear that the converse is true. This could more likely be indicative of an effect of CLAs on gene expression of inflammatory cytokines, either directly through fatty acid response elements in the promoter regions of the cytokine genes or through their known effect in activating PPARs that are transcription factors (see below; Wahle et al., 2003). However, a limited number of studies have reported an increase in prostanoid production in the presence of CLAs, which suggests that the prostanoid inhibitory feedback loop cannot be entirely precluded as a regulatory mechanism for cytokine production (Davidson et al., 2001; Ogborn et al., 2003; Rotondo et al., 1988). An explanation for this is not available at present but suggests the possibility that an increase in prostanoid production cannot be entirely precluded as a factor in downregulating inflammatory cytokine formation under certain circumstances. Another potential mechanism related to lipid-induced suppression of cytokine production is via PPARs, as mentioned above. PPARs are activated by a group of structurally diverse compounds, including fatty acids, eicosanoids and drugs, such as the hypolipidaemic fibrates (Kliewer et al., 1997). PPARs can also be activated by the thiazolidinediones, a class of anti-diabetic drugs (oral hypoglycaemics), thus providing a useful tool with which to characterise their distribution and the spectrum of actions that they modulate. Specific thiazolidinediones such as troglitazone and roglitazone have been used to elucidate the role of PPAR- -activated pathways and these compounds have been shown to bind directly to the PPARs (Lehmann et al., 1995). The discovery that PPARs were activated by direct ligand binding leads to a search for the physiological ligand regulator of PPARs, especially PPAR- . A potential candidate, on the basis of binding studies and the ability to functionally

BLUK122-Dijkstra

84

September 25, 2007

20:0

Trans Fatty Acids

activate PPAR- , was 15-deoxy-12,14-prostaglandin J2 (15d-PGJ2). It is now generally accepted that the naturally occurring activating ligand for PPAR- is 15d-PGJ2, which can be derived from prostaglandin D2 (PGD2 ). PGD2 metabolites are major products of ARA metabolism in macrophages, and PGD2 synthase, which is required for 15d-PGJ2 synthesis, is predominantly expressed in macrophages and other immune cells (Urade et al., 1989). Thus, PPAR- potentially possesses an important role in downregulating inflammatory responses by modulating the activity of macrophages/monocytes. Several studies on the role of PPAR- in macrophage functions have confirmed that its activation does indeed attenuate their activity. It has been shown that resting murine macrophages express low levels, whereas activated macrophages express high levels of PPAR- mRNA (Ricote et al., 1998). In addition, in the presence of either 15d-PGJ2 or the thiazolidinedione roglitazone, a specific PPAR- agonist, IFN- -activated macrophages retained the characteristics of resting cells (Ricote et al., 1998). This indicates that activation of PPAR- pathways is capable of halting macrophage activation and inflammation. PPAR- agonists have been shown to inhibit the production of monocyte inflammatory cytokines. In human peripheral blood mononuclear cells, the PMA-induced TNF- and IL-1 synthesis was potently inhibited by PGJ2 and 15d-PGJ2 (Jiang et al., 1998). mRNA for the inflammatory cytokines was also reduced, indicating that PPAR- pathways inhibit cytokine synthesis pre-translationally at the level of gene expression. PPARs act as transcription factors and certain genes contain PPAR response elements (PPAREs) in their promoter regions (Fruchart et al., 1999). It is therefore interesting in relation to their possible mechanisms of action that CLAs can decrease the production of both IL-1 and TNF- whilst activating PPAR- in the murine macrophage cell-line RAW (Yu et al., 2002). In the same study RAW cells were transfected with a mutation of the ligand-binding domain of PPAR- , resulting in a protein which does not lead to activation of downstream pathways. In cells transfected with this protein, CLAs failed to downregulate an inducible nitric oxide synthase (iNOS) reporter construct (Yu et al., 2002). This provides the strongest indication to date that the modulation of immune/inflammatory responses by CLA occurs via a PPAR- –dependent process. It is not clear at present if the effects of CLAs on PPAR- are direct or through some prostanoid metabolite of CLAs. Another major transcription factor that is also closely linked to the activation of the stressstimulated cell signal cascade and consequently to pro-inflammatory cytokine formation is NF-B, which exists in the cell cytosol as an inactivated complex with a specific inhibitor termed IB (Baldwin, 2001). Recent evidence from our laboratory shows that CLAs can also reduce the initial step in NF-B activation, the phosphorylation of the IB component of the inactive complex in cancer cells (Song et al., 2004). The initial IB phosphorylation leads to multiple ubiquitination of the NF-B/IB complex and its subsequent proteosomal degradation with the release of the free NF-B, which is then translocated to the nucleus where it binds to specific sites on the promoter regions of genes encoding for cytokines, adhesion molecules, heat shock proteins and other stress-induced molecules (Baldwin, 2001; Song et al., 2004; Zhang and Young, 2002). It has been suggested that PPARs and NF-B actions are integrated within the cell and that PPARs can inhibit NF-B activation by directly binding the free molecule in a similar manner to IB, thereby preventing its translocation to the nucleus (Delerive et al., 2001). However, it is not clear at present if the effects of CLAs on inflammatory cytokine formation are due to PPAR activation and subsequent NF-B sequestration and inhibition. The latter mechanism is supported by the fact that PPAR/NF-B complexes have been identified in the cytosol of PPAR-activated cells (Delerive et al., 2001). NF-B activation and binding of free NF-B to gene promoter regions is reportedly attenuated by an increase in the reductive state of cells and evidence from our laboratory indicated that

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

85

CLAs, like long-chain omega-3 PUFA, can increase the expression of redox enzymes in cells (Crosby et al., 1996; Sneddon et al., 2003). 3.6.2.2

Adaptive or specific immune functions

The initial acute phase of the immune response also involves an initial set of cytokine signals, which aids the development of a specific immune response over a period of days to weeks. This acute phase relates primarily to either T cells or B cells. These are the only cells within the body that can generate the diversity of gene combinations to provide the many antigen receptors (on T cells) or antibody variations (from B cells) to either transformed cells (infected or otherwise) or circulating foreign agents respectively. The generation of a T-cell response, i.e. the clonal expansion of a T cell whose receptor recognises an antigen presented to it via another cell, can also be linked in the majority of cases to antibody production (DeKruyff et al., 1993). This occurs as a result of the cytokines that are released from the T cell that may influence the isotype of antibody released from a B cell that can initially respond to an antigen and produce antibodies independently of other cells. However, the ‘maturation’ of an antibody response involves the actions of cytokines that can induce isotype switching within immunoglobulins, e.g. from low-affinity IgM to high-affinity IgG subclasses. Perhaps the best example of the influence of T-cell responses on antibody production is the development of B cells that produce IgE classically associated with allergy, i.e. type 1 hypersensitivity. The maturation of an IgE-producing B cell occurs in direct response to the presence of IL-4, the majority of which appears to derive from T cells which have developed a pattern of cytokine release, termed T-helper 2 (TH 2) responses (Del Prete, 1992, 1998). Much attention has focused on the development of T-helper cells that can polarise into a TH 1 response and into the pattern of cytokines that support the continuation of inflammatory responses. This is characterised by the production of IL-12 and IFN- or the production of IL-4 in a TH 2 response that is described above. It is now recognised that the presence of IgE is indicative of a TH 2 response since it is directly stimulated by IL-4 and suppressed by IFN- (Del Prete et al., 1988). Previously, it had been erroneously proposed that the inhibition of a TH 1 response occurred via IL-10 production by TH 2 cells. However, it is now recognised that much higher levels of IL-10 can be produced in response to inflammatory signals (resulting in attenuation of inflammation) in the absence of T-cell participation. It is also accepted that the definitive indicator of TH 2 polarisation is the control of IL-4 gene transcription that is spontaneously expressed in response to the transcription factor GATA-3 in the absence of inhibitory signals following at least two cell divisions (Reiner, 2001). Conversely, the presence of IL-12 inhibits IL-4 transcription and necessarily results in a TH 1 polarisation. How can CLAs influence these adaptive immune responses in a manner beneficial to human health? CLAs can decrease the ex vivo production of IL-4 from splenocytes stimulated with concanavalin A (Yang and Cook, 2003). This would tend to indicate a tendency towards a TH 1 response. This is confirmed by another study in which IgE levels in mice fed CLA were shown to be reduced (Sugano et al., 1998). However, it is unclear what effect CLAs have on the TH 1 profile of cytokines, i.e. IL-12 or indirectly IFN- . Thus, it is uncertain whether this merely reflects a generalised suppression of all immune cell activities or whether CLAs can specifically enhance particular actions. There is some evidence that CLAs can enhance some specific immune functions. For example, CLAs can elicit an enhanced production of protective immunoglobulins, i.e. IgA, IgG and IgM, whilst IgE is simultaneously reduced (Sugano et al., 1998). It has also been reported that CLAs can increase the proliferation of CD8+ T cells in pigs infected with type 2 porcine circovirus (Bassaganya-Riera et al., 2003).

BLUK122-Dijkstra

86

September 25, 2007

20:0

Trans Fatty Acids

These cells are important for an anti-viral and anti-cancer immune response and this increase in CD8+ cells correlates with a decreased morbidity in the CLA-treated animals. An increase in these cells has also been observed in the peripheral blood of pigs following vaccination; the level of CD8+ cells was greatly elevated in animals fed a CLA diet compared to those on a control diet of soya bean oil (Bassaganya-Riera et al., 2001). The authors also observed an increase in the level of natural killer (NK) cells. This suggests that CLAs can enhance the protective immune responses that eradicate intracellular pathogens, such as antibodies and cells (CD8+ and NK cells). But they suppress those responses which can be detrimental to the host animal, such as IgE production or the release of pro-inflammatory mediators. CLAs may, therefore, be useful as a supplement for specific disorders, requiring enhancement of antiviral immune mechanisms in humans. A recent randomised, double-blind, parallel dietary study was conducted with male volunteers who received the main isomers of CLA, cis-9, trans-11 or trans-10, cis-12 in their diets. The number of volunteers reaching protective antibody levels against hepatitis B was almost doubled in the group treated with CLAs in a 50:50 mixture or 80% 9/11 and 20% 10/12 mixture (Albers et al., 2003). However, none of the CLA treatments altered the ability of peripheral blood mononuclear cells to produce pro-inflammatory cytokines ex vivo following stimulation with bacterial endotoxin compared to a sunflower-oil reference group (Albers et al., 2003). This could be a reflection of the concentration of endotoxin used since a lack of cytokine response to such a stimulus could depend on the relative concentrations of both CLA and endotoxin. Another recent dietary study in humans was also not able to show any effect of CLAs on a wide variety of cytokines ex vivo (Kelley et al., 2001). This appears to indicate that the actions of CLAs on the human immune system are limited and that the effects on gross outcomes such as increased protective antibody levels may not be related to effects on cytokines. However, recent observations from our laboratory (Song et al., 2005) contrast with previous reports on the effects of CLAs on immune function in human volunteers in that the pro-inflammatory cytokines TNF- and IL-1 were significantly reduced in those receiving a 50:50 mix of CLAs compared to those receiving placebo. C-reactive protein (CRP) levels in plasma, an indicator of general inflammation, were not significantly altered by CLAs in this study, but it appeared as if the volunteers consisted of responders and non-responders, with some in the CLA group showing decreased CRP levels. IgE levels were also significantly reduced (Song et al., 2005). A lack of any stimulating effect of CLA on CRP has also been reported by others (Tricon et al., 2005), but this contrasts with the increased CRP when trans-10, cis-12 isomer was given to patients (Riserus et al., 2002). This would suggest that the responses in man were not dissimilar to those observed in some, but not all, animal experiments (see above). A study using different mixtures of the two isomers (50:50 and 80:20 of cis-9, trans-11 and trans-10, cis-12) reported a marked increase in proliferation of blood lymphocytes with the 50:50 mix and an opposite decrease with the 80:20 mix in response to phytohaemagglutinin stimulation. This suggested that the greater cis-9, trans-11 isomer content promoted the cellmediated immune response and the trans-10, cis-12 has the opposite effect (Roche et al., 2001). Tricon et al. (2005) reported a lack of effect of either of the common isomers of CLA on PBMNC subsets and inflammatory cytokine production ex vivo, but they showed a dose-dependent reduction in T-lymphocyte activation (assessed by CD69 surface expression, which was inversely proportional to the content of the isomers in the cell lipids). In contrast to the findings by Roche et al. (2001), these authors did not observe any isomer-specific effects on lymphocyte proliferation. CLAs appear to be able to alter immune functions, albeit to a small extent, in animals and men, although published results differ for both within and between species and whether both

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

87

or only one of the isomers is the effective modulator (Roche et al., 2001; Song et al., 2005; Tricon et al., 2005). In general, the two main isomers of CLAs appear to have similar effects on immune function, although isomer-specific effects have been reported (see above). This contrasts with their different and specific effects on body composition (Tricon et al., 2005). Their effects are not necessarily mediated via an alteration of ARA metabolism or consequent modification of eicosanoid production. CLAs appear to elicit their effects directly or by the modulation of intracellular regulatory pathways, such as PPARs (Roche et al., 2001; Song et al., 2005) and possibly NF-B and AP-1. 3.6.2.3

CLAs and inflammation

The previous section has mentioned the effects of CLAs on inflammatory processes. Again, the evidence in animals appears to be fairly strong but much less so in man (see above). Pharmacological PPAR- agonists inhibited the peripheral blood mononuclear cell synthesis of inflammatory cytokines TNF-, IL-1 and IL-6 at the level of mRNA expression. Since CLAs are known ligands for PPARs, it is conceivable that the observed reduction in inflammatory cytokines with CLAs in the animal studies (see above) and in the human volunteer study of Song et al. (2005) could be due to PPAR activation. However, the precise relationship between CLAs and cellular signal mechanisms in immune cell function requires further detailed investigation before the role of PPARs, other transcription factors and CLAs can be clearly defined. The effect of CLAs on CRP expression, a predictor of CVD, is also controversial (see above). Some authors report increased CRP, especially with the trans-10, cis-12 isomer, and others find no effect or a decrease (Bhattacharya et al., 2006; Riserus et al., 2002; Roche et al., 2001; Song et al., 2005; Tricon et al., 2005; Wahle et al., 2004). This may reflect differences in responsiveness to CLAs between young, healthy volunteers and older, obese subjects and the specific isomer used. Table 3.7 summarises some of the variable effects of CLAs on immune-inflammatory functions reported in the literature.

3.6.3

Effects of CLAs on biomarkers of cardiovascular disease

Cardiovascular disease is regarded as a complex disease with a strong lipid metabolic and inflammatory component targeted on the vascular endothelium. The inflammatory response of the cells in the vascular wall initiate the upregulation of adhesion molecules and the extravasation of lymphocytes and monocytes into the intimal space culminating in foam cell formation and atherogenesis (Bhattacharya et al., 2006; Tricon et al., 2004, 2005; Wahle et al., 2004; Wahle and Rotondo, 1999). Feeding a CLA mix or individual isomers has been shown to reduce the severity of cholesterol-induced atherosclerotic lesions in aortic arch and thoracic aorta in both rabbits and hamsters (Kritchevsky, 1999; Nicolosi et al., 1997). Even when fed at levels as low as 0.1% of the diet, the aortic arch and thoracic aorta atherosclerosis was reduced by 28 and 41% respectively. This was enhanced with increasing dosage of CLAs, such that 0.5% CLA lowered severity of aortic atherosclerosis by 60% and 56% in the arch and thorax respectively (Kritchevsky et al., 2002). The positive effects at the lower levels of dietary CLA intake indicate that levels of CLAs that may be beneficial in preventing heart disease can be achieved in human diets. It is not known what effect long-term ingestion of low levels of CLAs will have on incidence of atherosclerosis in man. The individual isomers of CLA

BLUK122-Dijkstra

88

September 25, 2007

20:0

Trans Fatty Acids

Table 3.7

Summary of effects of CLAs on immune functions.

Acute phase or inflammatory immune response ↓ pro-inflammatory cytokines release in animals ↓ PGE2 production in polycystic kidney disease model rats ↓ PGE2 production in mammary tumour cells ↓ PGE2 production in HUVEC ↓ PGE2 in guinea-pig trachea ↓ IL-1 and TNF- production whilst activating PPAR- in murine macrophages ↓ NF-B activation in cancer cells Adaptive or specific immune response ↓ ex vivo production of IL-4 from splenocytes ↓ IgE levels in mice ↑ production of protective Ig (IgA, IgG and IgM) ↑ proliferation of CD8+ T cells in pigs ↑ proliferation of CD8+ T-cells and NK-cells level ↑ level of protective antibody against hepatitis B in humans No effect on pro-inflammatory cytokines production ex vivo in humans ↓ TNF- , IL-1 production and IgE levels in human ↑ proliferation of blood lymphocytes with 50:50 mix and decrease with 80:20 No effect on PBMNC subsets and inflammatory cytokine production ex vivo Dose dependent ↓ in T-lymphocyte activation

Akahoshi et al. (2002, 2004); Yang and Cook (2003) Ogborn et al. (2003) Ma et al. (2002) Torres-Duarte and Vanderhoek (2003) Whigham et al. (2001) Yu et al. (2002) Song et al. (2004) Yang and Cook (2003) Sugano et al. (1998) Sugano et al. (1998) Bassaganya-Reira et al. (2003) Bassaganya-Reira et al. (2001) Albers et al. (2003) Albers et al. (2003); Kelley et al. (2001) Tricon et al. (2004) Roche et al. (2001) Song et al. (2005); Tricon et al. (2005) Song et al. (2005); Tricon et al. (2005)

had similar effects on atherosclerosis to that observed with the CLA mix that contrasts with the effects of individual isomers on body composition, insulin sensitivity (see above) and some, but not all, cancer studies (see below). An important aspect of the role of CLAs in atherogenesis is the observation that feeding rabbits with pre-established atherosclerotic plaques resulted in plaque regression, an important process that had not been previously reported for dietary fatty acids (Kritchevsky et al., 2004). A similar regression of pre-established atherosclerotic lesions was recently observed in apoE knockout mice fed CLAs (Toomey et al., 2006). This indicates that CLAs may be affecting the lipid transfer out of the plaque. They could, therefore, be of value in treatment of established vascular disease in man. To date, no clinical nutritional studies using CLA treatment have, to our knowledge, been conducted in human volunteers, either on plaque initiation or on regression. The positive, beneficial effects of CLA on atherogenesis described above contrast with the effects reported earlier in C57BL/6 mice fed an atherogenic diet (Munday et al., 1999). Despite dietary CLA eliciting a favourable lipoprotein profile by increasing the HDL:total cholesterol ratio and decreasing the triacylglycerol concentration in serum (both supposed beneficial effects with regard to CVD risk) compared to controls, the CLA increased the development of aortic fatty streaks. Earlier studies with cholesterol-fed rabbits also did not show any effect of feeding 0.5 g/day CLA on fatty streak lesions (Lee et al., 1994). The reasons for these contrasting findings are unclear but could relate to the type of animals used and/or the early types of CLA used (differing proportions of isomers). If the proportions of trans-10, cis-12 CLA were different in the various preparations, this could elicit contrasting results (see below).

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

89

Studies have shown that pharmacological PPAR activators attenuate the inflammatory response of endothelial cells and vascular adhesion molecule expression that would be expected to attenuate atherogenesis (Delerive et al., 1999; Houseknecht et al., 2002). PPAR- but not PPAR- agonists also inhibited the expression of IL-1, IL-6 and COX-2 in human aortic smooth muscle cells (Delerive et al., 1999). The PPAR agonists inhibited the NF-B signalling cascade that normally upregulates the gene transcription of inflammatory cytokines, adhesion molecules and COX genes, resulting in the attenuation of expression of these adverse responses (Delerive et al., 1999; Houseknecht et al., 2002; Meager, 1999; Su et al., 1999). CLAs are also known to attenuate expression of inflammatory cytokines in animals and men (see above), and it has recently been shown in our group that they inhibit the expression of cytokine-induced adhesion molecules on endothelial and smooth muscle cells at the level of mRNA and protein (Goua et al., 2003; Goua and Wahle, 2007; Masso-Welch et al., 2002; Sneddon et al., 2006). We also showed that CLAs can attenuate the activation of the cytosolic NF-B/IB complex by preventing the phosphorylation of the IB component of the complex in vascular cells (Goua, 2003; Masso-Welch et al., 2002). We also showed that the adhesion effects involved platelet-activating factor (Sneddon et al., 2006). In contrast to these apparent beneficial effects of CLA, a group in Germany has recently reported the precise opposite, stating that CLAs did not affect TNF-induced adhesion molecule expression or the expression of the chemokine MCP-1 involved in the extravasation of monocytes in human arterial endothelial cells (Schleser et al., 2006). Inhibition of the NF-B signalling pathway (Goua, 2003; Masso-Welch et al., 2002) would also be expected to reduce the transcription of the cytokine, adhesion molecule and COX genes involved in initiation and progression of atherogenesis. It is conceivable that CLAs are able to attenuate the atherosclerotic process through inhibition of the initiating inflammatory cytokines, as well as through inhibition of the stress signalling cascades these cytokines elicit. Inhibition of these processes in animals and in human vascular cells in culture appears to contradict the supposed pro-oxidative, proinflammatory role of CLAs reported in some animal and human volunteer (e.g. increased isoprostane formation) studies since oxidative stress would be expected to enhance adhesion molecule expression (see below). Attenuation of eicosanoid synthesis is generally regarded as a major factor in the beneficial, anti-disease effects elicited by CLAs in mammalian systems. In contrast to this tenet, our group observed an increase in 6-keto-PGF1 formation (stable form of PGI2 ) in HUVEC at high (100 M) CLA concentrations (mixture of isomers), with no significant effects at lower concentrations (Khaza’ai, 1997). Recent reports confirm and extend these findings and show that the CLA mix and cis-9, trans-11 isomer inhibit at both low and high concentrations, whilst the trans-10, cis-12 isomer actually increases eicosanoid formation (PGF2 ) at high concentrations in human saphenous vein endothelial cells (Urquhart et al., 2002); 6-keto-PGF1 formation was not investigated by these authors, which is unfortunate, as increased 6-keto-PGF1 could confer benefit in attenuating vascular disease (see above). These findings indicate an anti-inflammatory effect of CLAs at lower (50 M) concentrations that might possibly be achieved by dietary intakes. However, when HUVEC were pre-incubated with either trans-10, cis-12 or cis-9, trans-11 CLA isomers in the free state, they inhibited thrombin-induced 6-keto-PGF1 with I50s of 2.6 and 5.5 M respectively; cis,cis or trans, trans equivalent isoforms were without effect even at 60 M (Torres-Duarte and Vanderhoek, 2003). Inhibition of PGI2 would be expected to have adverse effects on vascular cell function and blood rheology. When the two CLAs (25 M) were allowed time to be incorporated into IL-1 -stimulated HUVEC membranes (18 h), the cis-9, trans-11 isomer increased 6keto-PGF1 formation eightfold, whereas the trans-10, cis-12 isomer increased it only by

BLUK122-Dijkstra

90

September 25, 2007

20:0

Trans Fatty Acids

Table 3.8

Summary of effects of CLAs on cardiovascular disease biomarkers.

↓ cholesterol-induced atherosclerotic lesions in rabbits and hamster Regression of atherosclerotic plaques in rabbits and mice ↑ aortic fatty streaks and ↑ lipoprotein profile ↓ expression of cytokine-induced adhesion molecules ↓ activation of cytosolic NF-B/IB complex No effect on TNF-induced adhesion molecule expression or chemokine MCP-1 expression ↑ 6- keto-PGF1 formation in HUVEC ↓ eicosanoid formation (PGF2 ) with CLA mix and c9,t11 ↓ thrombosin-induced 6-keto-PGF1 ↑ 6- keto-PGF1 in IL-1 -stimulated HUVEC if isomers incorporated in cell membrane

Kritchevsky (1999); Kritchevsky et al. (2002); Nicolosi et al. (1997) Kritchevsky et al. (2004); Toomey et al. (2006) Lee et al. (2004) Goua et al. (2003); Goua and Wahle (2007); Masso-Welch et al. (2002); Sneddon et al. (2006) Goua et al. (2003); Masso-Welch et al. (2003) Schleser et al. (2006) Khaza’ai (1997) Urquhart et al. (2002) Torres-Duarte and Vander Hoek (2003) Torres-Duarte and Vander Hoek (2003)

threefold. The importance of this observation is that both isomers increased the formation of a vital anti-thrombotic eicosanoid when incorporated into cell lipids, which contrasts with their effects when in the non-esterified state. Similar results were obtained with resting or stimulated platelets and the extent of thromboxane formation and suggest that this effect is general to cells containing COX-1 or/and COX-2 isoenzymes. The observations from these studies in vitro indicate that CLAs can elicit inhibitory or stimulatory effects on eicosanoid production in blood-vascular cells, depending on the specific isomer, whether the isomer is in the free or esterified form and whether the cells are in the resting or stimulated state. Clearly, these findings suggest that CLAs can have a number of complex effects on the blood-vascular system that can influence vascular homeostasis. The key finding is that the beneficial effects on blood eicosanoids are elicited at low concentrations of CLAs that are probably achievable through dietary intakes and that the more abundant dietary isomer, cis-9, trans-11, is the most effective. Long-term supplementation studies with different concentrations of CLAs (mix and individual isomers), in both nutritional and nutraceutical concentrations, in animals and men are required before we can be certain of the benefits of these fatty acids in the prevention of vascular disease. The more recent findings with known mixtures or individual isomers and their effects on vascular cell functions at low concentrations are intriguing and warrant further investigation. Table 3.8 summarises some of the reported effects of CLA on biomarkers of cardiovascular disease risk.

3.7

REPORTED ADVERSE HEALTH EFFECTS OF CLAs IN VIVO AND IN VITRO

The numerous health benefits that have previously been reported for various CLAs on the basis of studies with cells in culture, with animal models of disease and with human volunteers in vivo, are indicative, as mentioned previously, of CLAs being a veritable panacea (see Table 3.1). Unless the different disorders and diseases have a common, fundamental

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

91

aetiology, which cannot be precluded at this juncture, the diverse health effects of CLAs appear remarkable and warrant critical appraisal. In keeping with this suggestion, a number of apparent adverse effects of CLAs on biomarkers of disease have recently been reported which apparently counter the overwhelming number of beneficial reports to date. They appear to relate mainly to the effects of CLAs on oxidative processes, on eicosanoid production and on carcinogenesis and are largely, but not exclusively, observed in mice and with the trans-10, cis-12 but not the cis-9, trans-11 isomer or mixtures of both isomers. A group in Sweden reported that the trans-10, cis-12 isomer induced an increase in isoprostane secretion, an indicator of non-enzymatic eicosanoid formation and lipid peroxidation, in urine of healthy and obese volunteers (Smedman and Vessby, 2001; Riserus et al., 2001). However, the isoprostane excretion was initially reported not to be affected by -tocopherol, the major lipid-targeting antioxidant, which suggested the possibility that isoprostane formation was a non-oxidative process. However, subsequent observations with non-smokers subjected to passive smoke and healthy subjects ingesting vegetable soup identified a clear reduction in isoprostane formation with enhanced antioxidant intake, particularly vitamin C (Dietrich et al., 2003; Sanchez-Moreno et al., 2006). Enhanced production of isoprostanes, for whatever reason, might be cause for concern in human health terms if these products have deleterious effects in vivo, a fact that has not been clearly demonstrated despite the fact that these products are regarded as sensitive biomarkers of lipid (mainly, membranebound ARA) oxidation in situ. As mentioned above, these frequently reported observations of increased isoprostane formation contrast with the many effects of CLAs in cells and tissues indicating an antioxidative role, and consequently, they need to be substantiated before a clear consensus can be formed. Members of the Swedish group also observed an increase in distal colon carcinogenesis, but no effects in proximal colon or caecum in ApcMin mice fed the trans-10, cis-12 isomer. The authors suggested that this apparent detrimental effect was due to the pro-oxidative effects of this isomer (isoprostane formation) and subsequently to the oxidative activation of the transcription factor NF-B, although they did not detect any such activation (Rajakangas et al., 2003). Other examples of apparent deleterious effects of CLAs on eicosanoid production and carcinogenesis are described above. It is not clear at present if these adverse effects are due largely to the trans-10, cis-12 isomer and not the mix or cis-9, trans-11 isomer; recent observations from the Swedish group suggest that the cis-9, trans-11 isomer also induces isoprostane formation but to a lesser extent (see above). Deleterious effects have mainly been observed in transgenic rodents and normal mice, with some evidence in human studies and human cell studies. There are also contradictions regarding the effect of CLAs on glucose and insulin homeostasis in man, with recent reports suggesting that CLAs can increase insulin concentrations in plasma (Medina et al., 2000; Riserus et al., 2001; Smedman and Vessby, 2001; Tricon et al., 2004), whilst others show no effect on glucose concentrations or CRP, a biomarker of oxidation (Song et al., 2004; Tricon et al., 2004); some previous reports suggested that CLAs actually enhanced insulin sensitivity (see above). Clearly, at present the reports of adverse effects of CLAs on health are still in the minority but they cannot be disregarded and need to be critically appraised before a consensus regarding the safety and health benefits of CLAs can be arrived at.

3.8

CONCLUSIONS

It is clear from the foregoing that CLAs, as a mix or as individual isomers, elicit specific, often different, and sometimes opposing, metabolic and cell signalling effects in a variety of tissues and cells, in vivo and in vitro, which could explain their reported health benefits as well as

BLUK122-Dijkstra

92

September 25, 2007

20:0

Trans Fatty Acids

certain adverse effects in animals and men. These effects appear to be regulated by a number of diverse cellular signal systems that include various kinases, transcription factors and response elements in promoter regions of genes, including oncogenes, responsible for the cells reaction to stress factors. Whether the two main isomers of CLA elicit similar or opposing effects on cell mechanisms and cell functions appears to depend on a number of variables, including cell type, tissue type, in vivo or in vitro determinations, different mammalian species and CLA concentrations. This adds to the complexity of trying to evaluate the health benefits, or otherwise, of the different CLAs. The recently reported detrimental influences of CLAs on mammalian systems, particularly relating to the oxidative isoprostane formation from the trans-10, cis-12 isomer, but to a lesser extent also the cis-9, trans-11 isomer, give particular cause for concern in the light of suggested correlations with increased insulin resistance and hyperglycaemia. A clear link between isoprostanes and specific oxidative damage has first to be established. It would be an oversimplification to take the recent small number of detrimental findings as the absolute truth and disregard the wealth of information showing positive health benefits of CLAs. It is also interesting to note that no such detrimental effects on health have been reported for the more commonly employed 50:50 mixtures of the two main isomers of CLA or in general, the most abundant cis-9, trans-11 isomer, although the latter has recently also been reported to increase isoprostane production. Precisely, what this means in terms of adverse health initiation remains to be elucidated. Clearly, this area of lipid research must be progressed with caution and due diligence with respect to health safety issues until the specific effects of the two main CLA isomers and their interactions have been clearly characterised. Suffice it to say that observations with these fatty acids regarding their possible health effects, particularly their anti-inflammatory and anti-cancer effects, are encouraging, and their use in a variety of functional foods or as possible cancer preventatives and therapeutic agents is a distinct possibility that must be explored further.

ACKNOWLEDGEMENTS The authors acknowledge the support of Scottish Higher Education Funding Council (SHEFC), the Robert Gordon University Research and Development Initiative (RDI) and Aberdeen University, Department of Surgery.

REFERENCES Adams, A.L., Trout, J.R. & Karmali, R.A. (1990) Effect of n-3 fatty acids on spontaneous and experimental metastasis of rat mammary tumour 13762. Br J Cancer 61, 290–291. Adlof, R.O. (2003) Application of silver ion chromatography to the separation of conjugated linoleic acid isomers. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R.O. Adlof). AOCS Press, Champaign, IL, pp. 37–55. Akahoshi, A., Goto, Y., Murao, K. et al. (2002) Conjugated linoleic acid reduces body fats and cytokine levels of mice. Biosci Biotechnol Biochem 66, 916–920. Akahoshi, A., Koba, K., Ichinose, F. et al. (2004) Dietary protein modulates the effect of CLA on lipid metabolism in rats. Lipids 39, 25–30. Albers, R., van der Wielen, R.P., Brink, E.J., Hendriks, H.F., Dorovska-Taran, V.N. & Mohede, I.C. (2003) Effects of cis-9, trans-11 and trans-10, cis-12 conjugated linoleic acid (CLA) isomers on immune function in healthy men. Eur J Clin Nutr 57, 595–603.

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

93

Aro, A., Mannisto, S., Salminen, I., Ovaskainen, M.L., Kataja, V. & Uusitupa, M. (2000) Inverse association between dietary and serum conjugated linoleic acid and risk of breast cancer in postmenopausal women. Nutr Cancer 38, 151–157. Atkinson, R.L. (1999) Conjugated linoleic acid for altering body composition and treating obesity. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yurawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, pp. 348–353. Azain, M.J., Hausman, D.B., Sisk, M.B., Flatt, W.P. & Jewell, D.E. (2000) Dietary conjugated linoleic acid reduces rat adipose tissue cell size rather than cell number. J Nutr 130, 1548–1554. Baldwin, A.S. (2001) Series introduction: the transcription factor NF-B and human disease. J Clin Invest 107, 3–6. Banni, S., Heys, S.D. & Wahle, K.W.J. (2003) Conjugated linoleic acids (CLAs) in cancer prevention and treatment: animal studies and cellular mechanisms. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adlof). AOCS Press, Champaign, IL, pp. 267–282. Banni, S., Petroni, A., Blasevich, M. et al. (2004) Conjugated linoleic acids (CLA) as precursors of a distinct family of PUFA. Lipids 39, 1143–1146. Barber, M.D., Fearon, K.C., Tisdale, M.J., McMillan, D.C. & Ross, J.A. (2001) Effect of a fish oil-enriched nutritional supplement on metabolic mediators in patients with pancreatic cancer cachexia. Nutr Cancer 40, 118–124. Bassaganya-Riera, J., Hontecillas, R., Zimmerman, D.R. & Wannemuehler, M.J. (2001) Dietary conjugated linoleic acid modulates phenotype and effector functions of porcine CD8(+) lymphocytes. J Nutr 131, 2370–2377. Bassaganya-Riera, J., Pogranichniy, R.M., Jobgen, S.C. et al. (2003) Conjugated linoleic acid ameliorates viral infectivity in a pig model of virally induced immunosuppression. J Nutr 133, 3204–3214. Baumann, D.E. & Griinari, J.M. (2003) Nutritional regulation of milk fat synthesis. Ann Rev Nutr 23, 203–227. Baumgard, L.H., Matitashvili, E. & Corl, B.A. (2002) Trans-10, cis-12 conjugated linoleic acid decreases lipogenic rates and expression of genes involved in milk lipid synthesis in dairy cows. J Dairy Sci 85, 2155–2163. Belury, M.A. (2002a) Dietary conjugated linoleic acid in health: physiological effects and mechanisms of action. Ann Rev Nutr 22, 505–531. Belury, M.A. (2002b) Inhibition of carcinogenesis by conjugated linoleic acid: potential mechanisms of action. J Nutr 132, 2995–2998. Belury, M.A. (2003) Conjugated linoleic acids in type 2 diabetes mellitus: implications and potential mechanisms. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adolf). AOCS Press, Champaign, IL, pp. 302–315. Belury, M.A., Moya-Camarena, S.Y., Shi, L.L., Lu, M., Leesnitzer, L.M. & Blanchard, S.M. (2002) Conjugated linoleic acid activates peroxisome proliferator activated receptor (PPAR) gamma. Nutr Res 22, 817–824. Belury, M.A. & Vanden Heuvel, J.P. (1997) Protection against cancer and heart disease by the dietary fat, conjugated linoleic acid: potential mechanisms of action (Invited review). Nutr Dis Update J 1, 58–63. Belury, M.A. & Vanden Heuvel, J.P. (1999) Modulation of diabetes by conjugated linoleic acid. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yucawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, pp. 404–411. Bhattacharya, A., Banu, J., Rahman, M., Causey, J. & Fernandes, G. (2006) Biological effects of conjugated linoleic acids in health and disease. J Nutr Biochem 17, 789–810. Bingham, S.A., Luben, R., Welch, A., Wareham, N., Khaw, K.T. & Day, N. (2003) Are imprecise methods obscuring a relationship between fat and cancer. Lancet 326, 212–214. Blankson, H., Stakkestad, J.A., Fagertun, H., Thom, E., Wadstein, J. & Gudmundsen, O. (2000) Conjugated linoleic acid reduces body fat mass in overweight and obese humans. J Nutr 130, 2943–2948. Brodie, A.E., Manning, V.A., Ferguson, K.R., Jewell, D.E. & Hu, C.Y. (1999) Conjugated linoleic acid inhibits differentiation of pre- and post- confluent 3t3-l1 preadipocytes but inhibits cell proliferation only in preconfluent cells. J Nutr 129, 602–606. Cesano, A., Visonneau, S., Scimeca, J.A., Kritchevsky, D. & Santoli, D. (1998) Opposite effects of linoleic acid and conjugated linoleic acid on human prostatic cancer in SCID mice. Anticancer Res 18, 833–838. Chajes, V., Lavillonniere, F., Ferrari, P. et al. (2002) Conjugated linoleic acid content in breast adipose tissue is not associated with the relative risk of breast cancer in a population of French patients. Cancer Epidemiol Biomarkers Prev 11, 672–673.

BLUK122-Dijkstra

94

September 25, 2007

20:0

Trans Fatty Acids

Chajes, V., Lavillonniere, F., Maillard, V. et al. (2003) Conjugated linoleic acid content in breast adipose tissue of breast cancer patients and the risk of metastasis. Nutr Cancer 45, 17–23. Chilliard, Y., Ferlay, A. & Doreau, M. (2001) Effect of different types of forages, animal fat or marine oils in cow’s diet on milk fat secretion and composition, especially conjugated linoleic acid (CLA) and polyunsaturated fatty acids. Livest Prod Sci 70, 31–48. Cho, H.J., Kim, W.K., Jung, J.I. et al. (2005) Trans-10, cis-12, not cis-9, trans-11, conjugated linoleic acid decreases ErbB3 expression in HT-29 human colon cancer cells. World J Gastroenterol 11, 5142– 5150. Cho, H.J., Kim, W.K., Kim, E.J. et al. (2003) Conjugated linoleic acid inhibits cell proliferation and ErbB3 signaling in HT-29 human colon cell line. Am J Physiol Gastrointest Liver Physiol 284, G996–G1005. Choi, N., Kwon, D., Yun, S.H., Jung, M.Y. & Shin, H.K. (2004) Selectively hydrogenated soybean oil with conjugated linoleic acid modifies body composition and plasma lipids in rats. J Nutr Biochem 15, 411–4117. Christie, W.W. (2003) Analysis of conjugated linoleic acid – an overview. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R.O. Adlof). AOCS Press, Champaign, IL, pp. 1–12. Cohen, L.A., Zhao, Z., Pittman, B. & Scimeca, J. (2003) Effect of soy protein isolate and conjugated linoleic acid on the growth of Dunning R-3327-AT-1 rat prostate tumors. Prostate 54, 169–180. Collomb, M., Schmid, A., Sieber, R., Wechsler, D. & Ryhanen, E.L. (2006) Conjugated linoleic acid (CLA) in milk fat: variations and physiological effects. Rev Int Dairy J 16, 1347–1361. Cook, M.E., Butz, D., Li, G., Pariza, M., Whigham, L. & Yang, M. (2003) Conjugated linoleicacid enhances immune responses but protects against the collateral damage of immune events. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adolf). AOCS Press, Champaign, IL, pp. 283–291. Davidson, J., Abul, H.T., Milton, A.S. & Rotondo, D. (2001) Cytokines and cytokine inducers stimulate prostaglandin E2 entry into the brain. Pflugers Arch 442, 526–533. DeKruyff, R.H., Rizzo, L.V. & Umetsu, D.T. (1993) Induction of immunoglobulin synthesis by CD4+ T cell clones. Semin Immunol 5, 421–430. Del Prete, G. (1992) Human Th1 and Th2 lymphocytes: their role in the pathophysiology of atopy. Allergy 47, 450–455. Del Prete, G. (1998) The concept of type-1 and type-2 helper T cells and their cytokines in humans. Int Rev Immunol 16, 427–455. Del Prete, G., Maggi, E., Parronchi, P. et al. (1988) IL-4 is an essential factor for the IgE synthesis induced in vitro by human T cell clones and their supernatants. J Immunol 140, 4193–4198. Delany, J.P., Blohm, F., Truett, A.A., Scimeca, J.A. & West, D.B. (1999) Conjugated linoleic acid rapidly reduces body fat content in mice without affecting energy intake. Am J Physiol 276, R1172–R1179. Delerive, P., De Bosscher, K., Besnard, S. et al. (1999) Peroxisome proliferator-activated receptor alpha negatively regulates the vascular inflammatory gene response by negative cross-talk with transcription factors NF-kappa B and AP-1. J Biol Chem 274, 32048–32054. Delerive, P., Fruchart, J.C. & Staels, B. (2001) Peroxisome proliferator-activated receptors in inflammation control. J Endocrinol 169, 453–459. Dietrich, M., Block, G., Benowitz, N.L. et al. (2003) Vitamin C supplementation decreases oxidative stress biomarker F2-isoprostanes in plasma of non-smokers exposed to environmental tobacco smoke. Nutr Cancer 45, 176–184. Dinarello, C.A. (1997) Interleukin-1. Cytokine Growth Factor Rev 8, 253–265. Dinarello, C.A. (1999) Cytokines as endogenous pyrogens. J Infect Dis 179, S294–S304. Dobson, G. (2003) Gas chromatography-mass spectrometry of conjugated linoleic acid and metabolites. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R.O. Adlof). AOCS Press, Champaign, IL, pp. 13–36. Durgam, V.R. & Fernandes, G. (1997) The growth inhibitory effect of conjugated linoleic acid on MCF-7 cells is related to estrogen response system. Cancer Lett 116, 121–130. Eder, K., Schleser, S., Becker, K. & Korting, R. (2003) Conjugated linoleic acids lower the release of eicosanoids and nitric oxide from human aortic endothelial cells. J Nutr 133, 4083–4089. Faulconnier, Y., Arnal, M.A., Mirand, P.P., Chardigny, J.M. & Chilliard, Y. (2004) Isomers of conjugated linoleic acid decrease plasma lipids and stimulate adipose tissue lipogenesis without changing adipose weight in post-prandial adult sedentary or trained Wistar rat. J Nutr Biochem 15, 741–748. Fisher, S.M. (1995) Eicosanoids and tumor promotion. In: Skin Cancer: Mechanisms and Human Relevance (ed. H. Mukhtar). CRC Press, Boca Raton, FL, pp. 129–143.

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

95

Fritsche, J., Rickert, R. & Steinhart, H. (1999) Formation, contents, and estimation of daily intake of conjugated linoleic acid isomers and trans-fatty acids in foods. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P Yurawecz, M.M. Mossobo, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, 378–396. Fruchart, J.-C., Duriez, P. & Staels, B. (1999) Peroxisome proliferator-activated receptor-alpha activators regulate genes governing lipoprotein metabolism, vascular inflammation and atherosclerosis. Curr Opin Lipidol 10, 245–257. Gaullier, J.M., Halse, J., Hoye, K. et al. (2004) Conjugated linoleic acid supplementation for 1 y reduces body fat mass in healthy overweight humans. Am J Clin Nutr 79, 1118–1125. Gaullier, J.M., Halse, J., Hoye, K. et al. (2005) Supplementation with conjugated linoleic acid for 24 months is well tolerated by and reduces body fat mass in healthy, overweight humans. J Nutr 135, 778–784. Gavino, V.C., Gavino, G., Leblanc, M.-J. & Tuchweber, B. (2000) An isomeric mixture of conjugated linoleic acids but not pure cis-9, trans-11-octadecadienoic acid affects body weight gain and plasma lipids in hamsters. J Nutr 130, 27–29. Goua, M. (2003) Effects of polyunsaturated fatty acids on the regulation of eicosanoid enzymes and adhesion molecules in human endothelial and smooth muscle cells. Ph.D. Thesis, Department of Immunology, University of Strathclyde, Glasgow, Scotland. Goua, M., Sneddon, A.A., Rotondo, D. & Wahle, K.W.J. (2003) I B phosphorylation in response to CLA and n-3 PUFA in SMC and HUVEC: a link with adhesion molecule regulation [abstract]? In: 44th International Conference on the Bioscience of Lipids. Oxford, England, September 2003. Goua, M. & Wahle, K.W.J. (2007) Atherosclerosis: cell biology and lipoproteins. Curr Opin Lipid Res 18, 113–116. Greenberg, A.S., Nordan, R.P., McIntosh, J., Calvo, J.C., Scow, R.O. & Jablons, D. (1992) Interleukin 6 reduces lipoprotein lipase activity in adipose tissue of mice in vivo and in 3T3-L1 adipocytes: a possible role for interleukin 6 in cancer cachexia. Cancer Res 52, 4113–4116. Griinari, J.M. & Bauman, D.E. (1999) Biosynthesis of conjugated linoleic acid and its incorporation into meat and milk of ruminants. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yurawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, Chapter 13, pp. 180–200. Grimble, R.F. (1998) Nutritional modulation of cytokine biology. Nutrition 14, 634–640. Gulati, S.K., Kitessa, S.M., Ashes, J.R. et al. (2000). Protection of conjugated linoleic acids from ruminal hydrogenation and their incorporation into milk fat. Animal Feed Sci Technol 86, 139–148. Ha, H.H., Ahn, I.S., Byun, J.M. et al. (2003) Effects of conjugated linoleic acid on adipocyte secreted proteins in vitro. J Food Sci Nutr 8, 253–259. Ha, Y.L., Grimm, N.K. & Pariza, M.W. (1987) Anticarcinogens from fried ground beef: heat-altered derivatives of linoleic acid. Carcinogenesis 8, 1881–1887. Ha, Y.L., Storkson, J. & Pariza, M.W. (1990) Inhibition of benzo(a)pyreneinduced mouse forestomach neoplasia by conjugated dienoic derivatives of linoleic acid. Cancer Res 50, 1097–1101. Harbige, L.S. (2003) Fatty acids, the immune response, and autoimmunity: a question of n-6 essentiality and the balance between n-6 and n-3. Lipids 38, 323–341. Hibbeln, J.R., Ferguson, T.A. & Blasbalg, T.L. (2006) Omega-3 fatty acid deficiencies in neurodevelopment, aggression and autonomic dysregulation: opportunities for intervention. Int Rev Psychiatry 18, 107–118. Hotamisligil, G.S., Arner, P., Caro, J.F., Atkinson, R.L. & Spiegelman, B.M. (1995) Increased adipose tissue expression of tumor necrosis factor-alpha in human obesity and insulin resistance. J Clin Invest 95, 2409– 2415. Houseknecht, K.L., Cole, B.M. & Steele, P.J. (2002) Peroxisome proliferator-activated receptor gamma (PPARgamma) and its ligands: a review. Domest Anim Endocrinol 22, 1–23. Houseknecht, K.L., Vanden Heuvel, J.P., Moya-Camarena, S.Y. et al. (1998) Conjugated linoleic acid normalizes impaired glucose tolerance in the Zucker diabetic fatty fa/fa rat. Biochem Biophys Res Commun 244, 678–682. Howe, G.R., Hirohata, T., Hislop, T.G. et al. (1990) Dietary factors and risk of breast-cancer: combined analysis of 12 case control studies. J Natl Cancer Inst 82, 561–569. Hubbard, N.E., Lim, D., Summers, L. & Erickson, K.L. (2000) Reduction of murine mammary tumor metastasis by conjugated linoleic acid. Cancer Lett 150, 93–100. Hunter, D.J., Spiegelman, D., Adami, H.O. et al. (1996) Cohort studies of fat intake and the risk of breast cancer – a pooled analysis. N Engl J Med 334, 356–361.

BLUK122-Dijkstra

96

September 25, 2007

20:0

Trans Fatty Acids

Hwang, D.H. & Kinsella, J.E. (1979) Effects of trans, trans methyl linoelaidate on the concentration of prostaglandins and their precursors in rat. Prostaglandins 17, 543–559. Igarishi, M. & Miyazawa, T. (2001) The growth inhibitory effect of conjugated linoleic acid on a human hepatoma cell line, HepG2, is induced by a change in fatty acid metabolism, but not the facilitation of lipid peroxidation in the cells. Biochim Biophys Acta-Mol Cell Biol Lipids 1530, 162–171. Ip, C., Dong Y., Ip, M.M. et al. (2002) Conjugated linoleic acid isomers and mammary cancer prevention. Nutr Cancer 43, 52–58. Ip, C., Ip, M.M., Loftus, T., Shoemaker, S. & Shea-Eaton, W. (2000) Induction of apoptosis by conjugated linoleic acid in cultured mammary tumor cells and premalignant lesions of the rat mammary gland. Cancer Epidem Biomark Prev 9, 689–696. Ip, C., Scimeca, J.A. & Thompson, H. (1995) Effect of timing and duration of dietary conjugated linoleic acid on mammary cancer prevention. Nutr Cancer 24, 241–247. Ip, C., Singh, M., Thompson, H.J. & Scimeca, J.A. (1994) Conjugated linoleic acid suppresses mammary carcinogenesis and proliferative activity of the mammary gland in rat. Cancer Res 54, 1212–1215. Ip, M.M., Masso-Welch, P.A., Shoemaker, S.F., Shea-Eaton, W.K. & Ip, C. (1999) Conjugated linoleic acid inhibits proliferation and induces apoptosis of normal rat mammary epithelial cells in primary culture. Exp Cell Res 250, 22–34. Ip, M.M., McGee, S.O., Masso-Welch, P.A. et al. (2007) The t10, c12 isomer of conjugated linoleic acid stimulates mammary tumorigeneis in transgenic mice overexpressing erbB2 in mammary epithelium. Carcinogenesis [E pub ahead of print]. Iwakiri, Y., Sampson, D.A. & Allen, K.G.D. (2002) Suppression of cyclooxygenase-2 and inducible nitric oxide synthase expression by conjugated linoleic acid in murine macrophages. Prostagl Leukotr Essential Fatty Acids 67, 435–443. Jiang, C., Ting, A.T. & Seed, B. (1998) PPAR-gamma agonists inhibit production of monocyte inflammatory cytokines. Nature 391, 82–86. Kang, J.X. (2005) From fat to fat-1: a tale of omega-3 fatty acids. J Membr Biol 206, 165–172. Karmali, R.A., Marsh, J. & Fuchs C. (1984) Effect of omega-3 fatty acids on growth of a rat mammary tumor. J Natl Cancer Inst 73, 457–461. Katan, M.B., Zock, P.L. & Mensink, R.P. (1995) Trans fatty acids and their effect on lipoproteins in humans. Annu Rev Nutr 15, 473–493. Kavanaugh, C.J., Liu, K.L. & Belury, M.A. (1999) Effect of dietary conjugated linoleic acid on phorbol ester-induced PGE2 production and hyperplasia in mouse epidermis. Nutr Cancer 33, 132–138. Keim, N.L. (2003) Conjugated linoleic acid and body composition. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adlof). AOCS Press, Champaign, IL, pp. 316–324. Kelley, D.S. & Erickson, K.L. (2003) Modulation of body composition and immune cell functions by conjugated linoleic acid in humans and animal models: benefits vs. risks. Lipids 38, 377–386. Kelley, D.S., Simon, V.A., Taylor, P.C. et al. (2001) Dietary supplementation with conjugated linoleic acid increased its concentration in human peripheral blood mononuclear cells, but did not alter their function. Lipids 36, 669–674. Khaza’ai, H. (1997) Modulation of polyunsaturated fatty acid metabolism in human umbilical vein endothelial cell cultures by antioxidants. Ph.D. Thesis. Aberdeen University, Aberdeen, Scotland. Kim, E.J., Kang, I.J., Cho, H.J., Kim, W.K., Ha, Y.L. & Park, J.H. (2003) Conjugated linoleic acid downregulates insulin-like growth factor- I receptor levels in HT-29 human colon cancer cells. J Nutr 133, 2675–2681. Kim, M.R., Park, Y., Albright, K.J., Pariza, M.W. (2002) Differential responses of hamsters and rats fed high-fat or low-fat diets supplemented with conjugated linoleic acid. Nutr Res 22, 715–722. Kliewer, S.A., Sundseth, S.S., Jones, S.A. et al. (1997) Fatty acids and eicosanoids regulate gene expression through direct interactions with peroxisome proliferator-activated receptors alpha and gamma. Proc Natl Acad Sci USA, 94, 4318–4323. Klingenberg, M. & Huang, S.G. (1999) Structure and function of the uncoupling protein from brown adipose tissue. Biochim Biophys Acta 1415, 271–290. Knekt, P. & Jarvinnen, R. (1999) Intake of dairy products and breast cancer risk. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yucawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, pp. 444–468. Kritchevsky, D.T.S. (1999) Conjugated linoleic acid and experimental atherosclerosis in rabbits. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yucawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, pp. 397–403.

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

97

Kritchevsky, D., Tepper, S.A., Wright, S., Czarnecki, S.K., Wilson, T.A. & Nicolosi, R.J. (2004) Conjugated linoleic acid isomer effects in atherosclerosis: growth and regression of lesions. Lipids 39, 611–616. Kritchevsky, D.T.S., Wright, S. & Czarnecki, S.K. (2002) Influence of graded levels of conjugated linoleic acid (CLA) on experimental atherosclerosis in rabbits. Nutr Res 22, 1275–1279. Kuhnt, K., Kraft, J., Moeckel, P. & Jahreis, G. (2006) Trans-11 18:1 is more effectively delta-9 desaturated compared with trans-12-18;1 in humans. Br J Nutr 95, 752–761. Kuniyasu, H., Yoshida, K., Sasaki, T., Sasahira, T., Fujii, K. & Ohmori, H. (2006) Conjugated linoleic acid inhibits peritoneal metastasis in human gastrointestinal cancer cells. Int J Cancer 118, 571–576. Landa, M.C., Frago, N. & Tres, A. (1994). Diet and risk of breast cancer in Spain. Eur J Cancer Prev 3, 313–320. Larsson, S.C., Bergkvist, L. & Wolk, A. (2005) High-fat dairy food and conjugated linoleic acid intakes in relation to colorectal cancer incidence in the Swedish Mammography Cohort. Am J Clin Nutr 82, 894–900. Ledoux, M., Chardigny, J.M., Darbois, M., Soustre, Y., S´eb´edio, J.-L. & Laloux, L. (2003) Seasonal and regional variations of the levels of conjugated linoleic acid in French butters. Sci Aliments 23, 443–461. Ledoux M & Laloux L. (2006) Conjugated linoleic acids: occurrence in food and physiological properties. Rev Sci Aliments 26, 291–314. Lee, K.N., Kritchevsky, D. & Pariza, M.W. (1994) Conjugated linoleic acid and atherosclerosis in rabbits. Atherosclerosis 108, 19–25. Lee, S.H., Yamaguchi, K., Kim, J.S., Eling, T.E., Park, Y. & Baek S.J. (2006) Conjugated linoleic acid stimulates an anti-tumorigenic protein NAG-1 in an isomer specific manner. Carcinogenesis 27, 972–981. Lehmann, J.M., Moore, L.B., Smith-Oliver, T.A., Wilkison, W.O., Willson, T.M. & Kliewer, S.A. (1995) An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome proliferator-activated receptor gamma (PPAR gamma). J Biol Chem 270, 12953–12956. Liew, C., Schut, H.A., Chin, S.F., Pariza, M.W. & Dashwood, R.H. (1995) Protection of conjugated linoleic acids against 2-amino-3- methylimidazo [4,5-f]quinoline-induced colon carcinogenesis in the F344 rat: a study of inhibitory mechanism. Carcinogenesis 16, 3037–3043. Lim, D.Y., Tyner, A.L., Park, J.B., Lee, J.Y., Choi, Y.H. & Park, J.H. (2005) Inhibition of colon cancer cell proliferation by the dietary compound conjugated linoleic acid is mediated by the CDK inhibitor p21CIP1/WAF1. J Cell Physiol 205, 107–113. Lock, A.L., Tyburczy, C., Dwyer, D.A. et al. (2007) Trans-10 octadecenoic acid does not reduce milk fat synthesis in dairy cows. J Nutr 137, 71–76. Lu, M., Gottschling, B., Kamendulis, L., Klaunig, J.E. & Belury, M.A. (2000) Dietary CLA induces hepatocyte proliferation in F33 rats. FASEB J 14, A721. Ma, D.W., Field, C.J. & Clandinin, M.T. (2002) An enriched mixture of trans-10,cis-12-CLA inhibits linoleic acid metabolism and PGE2 synthesis in MDA-MB-231 cells. Nutr Cancer 44, 203–212. Majumder, B., Wahle, K.W., Moir, S. et al. (2002) Conjugated linoleic acids (CLAs) regulate the expression of key apoptotic genes in human breast cancer cells. FASEB J 16, 1447–1449. Masso-Welch, P.A., Winston, J.S., Edge, S. et al. (2001) Altered expression and localization of PKC eta in human breast tumors. Breast Cancer Res Treat 68, 211–223. Masso-Welch, P.A., Zangani, D., Ip, C. et al. (2002) Inhibition of angiogenesis by the cancer chemopreventive agent conjugated linoleic acid. Cancer Res 62(15), 4383–4389; 62(19):5624–5624. Masso-Welch, P.A., Zangani, D., Ip, C. et al. (2004) Isomers of conjugated linoleic acid differ in their effects on angiogenesis and survival of mouse mammary adipose vasculature. J Nutr 134, 299–307. McCann, S.E., Ip, C., Ip, M.M. et al. (2004) Dietary intake of conjugated linoleic acids and risk of premenopausal and postmenopausal breast cancer, Western New York Exposures and Breast Cancer Study (WEB Study). Cancer Epidemiol Biomarkers Prev 13, 1480–1484. McCarthy-Beckett, D.O. (2002) Dietary supplementation with conjugated linoleic acid does not improve nutritional status of tumor-bearing rats. Res Nursing Health 25, 49–57. McGuire, M.A., McGuire, M.K., Parodi, P.W. & Jensen, R.G. (1999) Conjugated linoleic acids in human milk. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yurawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, pp. 296–306. McGuire, M.K., McGuire, M.A., Ritzenthaler, K. & Shultz, T.D. (1999) Dietary sources and intakes of conjugated linoleic acid intake in humans. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yurawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, pp. 369–377. Meager, A. (1999) Cytokine regulation of cellular adhesion molecule expression in inflammation. Cytokine Growth Factor Rev 10, 27–39.

BLUK122-Dijkstra

98

September 25, 2007

20:0

Trans Fatty Acids

Medina, E.A., Horn, W.F., Keim, N.L. et al. (2000) Conjugated linoleic acid supplementation in humans: effects on circulating leptin concentrations and appetite. Lipids 35, 783–788. Miller, A., Stanton, C. & Devery, R. (2002) Cis 9, trans 11- and trans 10, cis12-conjugated linoleic acid isomers induce apoptosis in cultured SW480 cells. Anticancer Res 22, 3879–3887. Miner, J.L., Cederberg, C.A., Nielsen, M.K., Chen, X. & Baileet, C.A. (2001) Conjugated linoleic acid (CLA), body fat, and apoptosis. Obes Res 9, 129–134. Mirand, P.P., Arnal-Bagnard, M.A., Mosoni, L., Faulconnier, Y., Chardigny, J.M. & Chilliard, Y. (2004) Cis-9, trans-11 and trans-10, cis-12 conjugated linoleic acid isomers do not modify body composition in adult sedentary or exercised rats. J Nutr 134, 2263–2269. Moon, E.J., Lee, Y.M. & Kim, K.W. (2003) Anti-angiogenic activity of conjugated linoleic acid on basic fibroblast growth factor-induced angiogenesis. Oncol Rep 10, 617–621. Moseley, E.E., McGuire, H.K., Williams, J.E. & McGuire, M.A. (2006) Cis-9, trans-11 conjugated linoleic acid is synthesised from vaccenic acid in lactating women. J Nutr 136, 2297–2301. Mougios, V., Matsakas, A., Petridou, A. et al. (2001) Effect of supplementation with conjugated linoleic acid on human serum lipids and body fat. J Nutr Biochem 12, 585–594. Munday, J.S., Thompson, K.G. & James, K.A. (1999) Dietary conjugated linoleic acids promote fatty streak formation in the C57BL/6 mouse atherosclerosis model. Br J Nutr 81, 251–255. Nakanishi, T., Koutoku, T., Kawahara, S., Murai, A. & Furuse, M. (2003) Dietary conjugated linoleic acid reduces cerebral prostaglandin E2 in mice. Neurosci Lett 341, 135–138. Nicolosi, R.J., Rogers, E.J., Kritchevsky, D., Scimeca, J.A. & Huth, P.J. (1997) Dietary conjugated linoleicacid reduces plasma lipoproteins and early aortic atherosclerosis in hypercholesterolemic hamsters. Artery 22, 266–277. Ochoa, J.J., Farquharson, A.J., Grant, I., Moffat, L.E., Heys, S.D. & Wahle, K.W. (2004) Conjugated linoleic acids (CLAs) decrease prostate cancer cell proliferation: different molecular mechanisms for cis-9, trans-11 and trans-10, cis-12 isomers. Carcinogenesis 25, 1185–1191. Oettgen, P. (2001) Transcriptional regulation of vascular development. Circ Res 89, 380–388. Ogborn, M.R., Nitschmann, E., Bankovic-Calic, N., Weiler, H.A., Fitzpatrick-Wong, S. & Aukema, H.M. (2003) Dietary conjugated linoleic acid reduces PGE2 release and interstitial injury in rat polycystic kidney disease. Kidney Int 64, 1214–1221. Ohnuki, K., Haramizu, S., Oki, K., Ishihara, K. & Fushiki, T. (2001) A single oral administration of conjugated linoleic acid enhanced energy metabolism in mice. Lipids 36, 583–587. Palombo, J.D., Ganguly, A., Bistrian, B.R. & Menard, M.P. (2002) The antiproliferative effects of biologically active isomers of conjugated linoleic acid on human colorectal and prostatic cancer cells. Cancer Lett 177, 163–172. Pariza, M.W. & Hargreaves, W.A. (1985) A beef-derived mutagenesis modulator inhibits initiation of mouse epidermal tumors by 7,12-dimethylbenz[a]anthrazene. Carcinogensis 6, 591–593. Park, H.S., Cho, H.Y., Ha, Y.L. & Park, J.H. (2004) Dietary conjugated linoleic acid increases the mRNA ratio of Bax/Bcl-2 in the colonic mucosa of rats. J Nutr Biochem 15, 229–235. Park, H.S., Ryu, J.H., Ha, Y.L. & Park, J.H. (2001) Dietary conjugated linoleic acid (CLA) induces apoptosis of colonic mucosa in 1,2-dimethylhydrazine-treated rats: a possible mechanism of the anticarcinogenic effect by CLA. Br J Nutr 86, 549–555. Park, Y., Albright, K.J., Storkson, J.M., Liu, W., Cook M.E. & Pariza M.W. (1999) Changes in body composition in mice during feeding and withdrawal of conjugated linoleic acid. Lipids 34, 243–248. Park, Y. & Pariza, M.W. (2007) Mechanisms of body fat modulation by conjugated linoleic acid (CLA) (Review). Food Res Int 40, 311–323. Parodi, P.W. (1997) Cows’ milk fat components as potential anticarcinogenic agents. J Nutr 127, 1055–1060. Parodi, P.W. (2003) Conjugated linoleic acid in food. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adlof). AOCS Press, Champaign, IL, pp. 101–122. Parrish, F.C., Wiegand, B.R., Beitz, D.C., Ahn, D.U., Du, M. & Trenkle, A.H. (2003) Use of dietary CLA to improve composition and quality of animal-derived foods. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adlof). AOCS Press, Champaign, IL, pp. 189–217. Perfield, J.W., Delmonte, P., Lock, A.L., Yurawecz, M.P. & Baumann, D.E. (2006) Trans-10, trans-12 conjugated linoleic acid does not affect milk yield but reduces delta-9 desaturase index in dairy cows. J Dairy Sci 89, 2559–2566. Petrick, M.B.H., McEntee, M.F., Johnson, B.T., Obukowicz, M.G. & Whelan, J. (2000) Highly unsaturated (n-3) fatty acids, but not -linolenic, conjugated linoleic or -linolenic acids, reduce tumorigenesis in APCMin/+ mice. J Nutr 130, 2434–2443.

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

99

Petridou, A., Mougios, V. & Sagredos, A. (2003) Supplementation with CLA: isomer incorporation into serum lipids and effect on body fat of women. Lipids 38, 805–811. Rahman, S.M., Wang, Y.M., Yotsumoto, H. et al. (2001) Effects of conjugated linoleic acid on serum leptin concentration, body-fat accumulation, and beta-oxidation of fatty acid in OLETF rats. Nutrition 17, 385– 390. Rajakangas, J., Basu, S., Salminen, I. & Mutanen, M. (2003) Adenoma growth stimulation by the trans-10, cis-12 isomer of conjugated linoleic acid (CLA) is associated with changes in mucosal NF-B and cyclin D1 protein levels in the min mouse. J Nutr 133, 1943–1948. Reiner, S.L. (2001) Helper T cell differentiation, inside and out. Curr Opin Immunol 13, 351–355. Ricote, M., Li, A.C., Willson, T.M., Kelly, C.J. & Glass, C.K. (1998) The peroxisome proliferator-activated receptor-gamma is a negative regulator of macrophage activation. Nature 391, 79–82. Ringseis, R., Muller, A., Dusterloh, K., Schleser, S., Eder, K. & Steinhart, H. (2006) Formation of conjugated linoleic acid metabolites in human vascular endothelium. Biochim Biophys Acta 1761, 377–383. Riserus, U., Basu, S., Jovinge, S., Frederikson, G.N., Arnlov, J. & Vessby, B. (2002) Supplementation with conjugated linoleic acid causes isomer-dependent oxidative stress and elevated C-reactive protein. Circulation 106, 1925–1929. Riserus, U., Berglund, L. & Vessby, B. (2001) Conjugated linoleic acid (CLA) reduced abdominal adipose tissue in obese middle-aged men with signs of the metabolic syndrome: a randomised controlled trial. Int J Obes Relat Metab Disord 25, 1129–1135. Roche, H.M., Noone, E., Nugent, A. & Gibney, M.J. (2001) Conjugated linoleic acid: a novel therapeutic nutrient? Nutr Res Rev 14, 173–187. Rotondo, D., Abul, H.T., Milton, A.S. & Davidson, J. (1988) Pyrogenic immunomodulators increase the level of prostaglandin E2 in the blood simultaneously with the onset of fever. Eur J Pharmacol 154, 145–152. Saebo, A. (2003) Commercial synthesis of conjugated linoleate. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adlof). AOCS Press, Champaign, IL, pp. 71–81. Salamini, R., La Vecchia, C., Recarli, A. et al. (1984) Social factors, diet and breast cancer in a northern Italian population. Br J Cancer 49, 723–729. Salas-Salvado, J., Marquez-Sardoval, F. & Bullo, M. (2006) Conjugated linoleic acid intake in humans: a systematic review focussing on its effects on body composition, glucose and lipid metabolism. Crit Rev Food Sci Nutr 46, 479–488. Satory, D. & Smith, S.B. (1999) Conjugated linoleic acid inhibits proliferation but stimulates lipid filling of murine 3T3-L1 preadipocytes. J Nutr 129, 92–97. Sanchez-Moreno, C., Cano, M.P., De Aucos, B. et al. (2006) Mediterranean vegetable soup consumption increases plasma vitamin C and decreases F2-isoprostanes, prostaglandin E2 and monocyte chemotactic protein-1 in healthy humans. J Nutr Biochem 17, 183–189. Schleser, S., Ringseis, R. & Eder, K. (2006) Conjugated linoleic acids have no effect on TNF-induced adhesion molecule expression, U937 monocyte adhesion and chemokine release in human aortic endothelial cells. Atherosclerosis 186, 337–344. Scimeca, J.A. (1999) Cancer inhibition in animals. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yurawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, pp. 420–443. S´eb´edio, J.-L., Gnaedig, S. & Chardigny, J.-M. (1999) Recent advances in conjugated linoleic acid research. Curr opin Clin Nutr Metabol Care 2, 499–506. Sisk, M.B., Hausman, D.B., Martin, R.J. & Azain, M.J. (2001) Dietary conjugated linoleic acid reduces adiposity in lean but not obese Zucker rats. J Nutr 131, 1668–1674. Smedman, A. & Vessby, B. (2001) Conjugated linoleic acid supplementation in humans-metabolic effects. Lipids 36, 773–781. Sneddon, A.A., McLeod, E., Wahle, K.W.J. & Arthur, J.R. (2006) Cytokine-induced adhesion to endothelial cells involves platelet-activating factor: suppression by conjugated linoleic acid. Biochim Biophys Acta 1761, 793–801. Song, H.J., Barker, P.A., Faisant, P. et al. (2001) Conjugated linoleic acid (CLA) modulation of protein kinase C (PKC) in prostate cancer. In: Proceedings of 42nd ICBL. Bergen, Norway, pp. 2–12. Song, H.J, Grant, I., Rotondo, D. et al. (2005) Effect of CLA supplementation on immune function in young healthy volunteers. Eur J Clin Nutr Nat 59, 505–517. Song, H.J., Sneddon, A., Barker, P. et al. (2004) Conjugated linoleic acid inhibits proliferation and modulates protein kinase c in human prostate cancer cells. Nutr Cancer 49, 100–108.

BLUK122-Dijkstra

100

September 25, 2007

20:0

Trans Fatty Acids

Song, H.-J., Sneddon, A.A., Heys, S.D. & Wahle, K.W.J. (2006) Induction of apoptosis and inhibition of NF-B activation in human prostate cancer cells by the cis-9, trans-11 but not the trans-10, cis-12 isomer of conjugated linoleic acid. Prostate 66, 839–846. Sporn, M.B., Suh, N. & Mangelsdorf, D.J. (2001) Prospects for prevention and treatment of cancer with SPARMs (selective PPAR Modulators). Trends Mol Med 7, 395–400. Stanton, C., Murphy, J., McGrath, E. & Devery, R. (2003) Animal feeding strategies for conjugated linoleic acid enrichment of milk. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R.O. Adlof). AOCS Press, Champaign, IL, pp. 123–145. Su, C.G., Wen, X., Bailey, S.T. et al. (1999) A novel therapy for colitis utilizing PPAR- ligands to inhibit the epithelial inflammatory response. J Clin Invest 104, 383–389. Suffredini, A.F., Fantuzzi, G., Badolato, R., Oppenheim, J.J. & O’Grady, N.P. (1999) New insights into the biology of the acute phase response. J Clin Immunol 19, 203–214. Sugano, M., Tsujita, A., Yamasaki, M., Noguchi, M. & Yamada, K. (1998) Conjugated linoleic acid modulates tissue levels of chemical mediators and immunoglobulins in rats. Lipids 33, 521–527. Thom, E., Wadstein, J. & Gudmundsen, O. (2001) Conjugated linoleic acid reduces body fat in healthy exercising humans. J Int Med Res 29, 392–396. Toomey, S., Harhen, B., Roche, H.M., Fitzgerald, D. & Belton, O. (2006) Profound resolution of early atherosclerosis with conjugated linoleic acid. Atherosclerosis 187, 40–49. Torres-Duarte, A.P. & Vanderhoek, J.Y. (2003) Conjugated linoleic acid exhibits stimulatory and inhibitory effects on prostanoid production in human endothelial cells and platelets. Biochim Biophys Acta 1640, 69–76. Tricon, S., Burdge, G.C., Kew, S. et al. (2004) Effects of cis-9, trans-11 and trans-10, cis-12 conjugated linoleic acid on immune cell function in healthy humans. Am J Clin Nutr 80, 1626–1633. Tricon, S., Burdge, G.C., Williams, C.M., Calder, P.C. & Yaqoob, P. (2005) The effects of conjugated linoleic acid on human health-related outcomes. Proc Nutr Soc 64, 171–182. Tsuboyama-Kasaoka, N., Takahashi, M., Tanemura, K. et al. (2000) Conjugated linoleic acid supplementation reduces adipose tissue by apoptosis and develops lipodystrophy in mice. Diabetes 49, 1534–1542. Tsuzuki, T., Kawakami, Y., Nakagawa, K. & Miyazawa, T. (2006) Conjugated docosahexaenoic acid inhibits lipid accumulation in rats. J Nutr Biochem 17, 518–524. Turpeinen, A.M., Mutanen, M., Aro, A. et al. (2002) Bioconversion of vaccenic acid to conjugated linoleic acid in humans. Am J Clin Nutr 76, 504–510. Urade, Y., Ujihara, M., Horiguchi, Y., Ikai, K. & Hayaishi, O. (1989) The major source of endogenous prostaglandin D2 production is likely antigen-presenting cells. Localization of glutathione-requiring prostaglandin D synthetase in histiocytes, dendritic, and Kupffer cells in various rat tissues. J Immunol 143, 2982–2989. Urquhart, P., Parkin, S.M., Rogers, J.S., Bosley, J.A. & Nicoleaou, A. (2002). The effect of conjugated linoleic acid on arachidonic acid metabolism and eicosanoid production in human saphenous vein endothelial cells. Biochim Biophys Acta 1580, 150–160. Voorrips, L.E., Brants, H.A., Kardinaal, A.F., Hiddink, G.J., van den Brandt, R.A. & Goldbohm, R.A. (2002) Intake of conjugated linoleic acid, fat, and other fatty acids in relation to postmenopausal breast cancer: the Netherlands Cohort Study on Diet and Cancer. Am J Clin Nutr 76, 873–882. Wahle, K.W.J. (1994) Royal College Of General Practitioners Reference Book. Royal College of General Practitioners, London, pp. 541–543. Wahle, K.W.J. & Heys, S.D. (2002) Cell signal mechanisms, conjugated linoleic acids (CLAs) and antitumorigenesis. Pros Leuk Ess Fatty Acids 67, 183–186. Wahle, K.W.J., Heys, S.D. & Rotondo, D. (2004) Conjugated linoleic acids: are they beneficial or detrimental to health? Prog Lipid Res 43, 553–587. Wahle, K.W.J. & James, W.P.T. (1993) Isomeric fatty acids and human health. Eur J Clin Nutr 47, 828–839. Wahle, K.W.J. & Rotondo, D. (1999) Fatty acids and endothelial cell function: regulation of adhesion molecule and redox enzyme expression. Curr Opin Clin Nutr Metabolic Care 2, 109–115. Wahle, K.W.J., Rotondo, D. & Heys, S.D. (2003) Polyunsaturated fatty acids and gene expression in mammalian systems. Proc Nutr Soc 62, 349–360. Wang, Y.W. & Jones, P.J. (2004) Conjugated linoleic acid and obesity control: efficacy and mechanisms. Int J Obes Rel Metab Disord 104, 963–968. Warren, J.M., Simon, V.A., Bartolini, G., Erickson, K.L., Mackey, B.E. & Kelley, D.S. (2003). Trans-10, cis-12 CLA increases liver and decreases adipose tissue lipids in mice: possible roles of specific lipid

BLUK122-Dijkstra

September 25, 2007

20:0

Conjugated linoleic acid effects

101

metabolism genes. Lipids 38, 497–504. West, D.B., Blohm, F.Y., Truett, A.A. & Delaney, J.P. (2000) Conjugated linoleic acid persistently increases total energy expenditure in AKR/J mice without increasing uncoupling protein gene expression. J Nutr 130, 2471–2477. West, D.B., DeLany, J.P., Camet, P.M., Blohm, F., Truett, A.A. & Scimeca, J. (1998) Effects of conjugated linoleic acid on body fat and energy metabolism in the mouse. Am J Physiol 44, R667–R672. Whigham, L.D., Cook, E.B., Stahl, J.L. et al. (2001) CLA reduces antigen-induced histamine and PGE(2) release from sensitized guinea pig tracheae. Am J Physiol Regul Integr Comp Physiol 280, R908–R912. Wigmore, S.J., Barber, M.D., Ross, J.A., Tisdale, M.J. & Fearon, K.C. (2000) Effect of oral eicosapentaenoic acid on weight loss in patients with pancreatic cancer. Nutr Cancer 36, 177–184. Wong, M.W, Chew, B.P., Wong, T.S., Hosick, H.L., Boylston, T.D. & Shultz T.D. (1997) Effects of dietary conjugated linoleic acid on lymphocyte function and growth of mammary tumors in mice. Anticancer Res 17, 987–993. Yamasaki, M., Ikeda, A., Oji, M. et al. (2003) Modulation of body fat and serum leptin levels by dietary conjugated linoleic acid in Sprague-Dawley rats fed various fat level diets. Nutrition 19, 30–35. Yamasaki, M., Ikeda, A., Tanaka, Y. et al. (2002) Dose-dependent effect of dietary conjugated linoleic acid on the growth of rat hepatoma dRLh-84 cells in vivo. J Nutr Sci Vitaminol 48, 505–511. Yang, H., Holcroft, J., Glickman, B.W. & de Boer, J.G. (2003) Conjugated linoleic acid inhibits mutagenesis by 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine in the prostate of Big Blue(R) rats. Mutagenesis 18, 195–200. Yang, M. & Cook, M.E. (2003) Dietary conjugated linoleic acid decreased cachexia, macrophage tumor necrosis factor-alpha production, and modifies splenocyte cytokines production. Exp Biol Med 228, 51–58. Yu, Y., Correll, P.H. & Vanden Heuvel, J.P. (2002) Conjugated linoleic acid decreases production of proinflammatory products in macrophages: evidence for a PPAR gamma-dependent mechanism. Biochim Biophys Acta 1581, 89–99. Zambell, K.L., Horn, W.F. & Keim, N.L. (2001) Conjugated linoleic acid supplementation in humans: effects on fatty acid and glycerol kinetics. Lipids 36, 767–772. Zambell, K.L., Keim, N.L., Van Loan, M.D. et al. (2000) Conjugated linoleic acid supplementation in humans: effects on body composition and energy expenditure. Lipids 35, 777–782. Zhang, X. & Young, H.A. (2002) PPAR and immune system – what do we know? Int Immunopharmacol 2, 1029–1044.

BLUK122-Dijkstra

4

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

Jean-Louis S´ eb´ edio and W.M. Nimal Ratnayake

4.1

INTRODUCTION

Dietary trans fats include mono- and polyunsaturated fatty acids having methyleneinterrupted double bonds but also isomers having conjugated double bonds, such as conjugated isomers of linoleic acid or conjugated linoleic acid (CLA). These trans fats are naturally present in ruminant meat and milk as a result of biohydrogenation in the rumen (Precht et al., 2001; Wolff, 1995), but they are also formed during technological treatments, such as partial hydrogenation to produce fat blends for margarine and shortening production (Dutton, 1979), vacuum steam stripping or deep fat frying (Juaneda et al., 2003; S´eb´edio and Juaneda, 2007). Numerous studies (Mensink et al., 2003; Mozaffarian et al., 2006; Willett, 2006a) have reported the deleterious effects of the trans fatty acids. In recent years, Denmark, Canada and the USA have taken legislative steps to reduce the trans fat content in food products (Ratnayake and Zehaluk, 2005). A number of other countries are also contemplating taking similar measures to reduce the trans content in foods (Aro, 2006; Ratnayake and Zehaluk, 2005; Willett, 2006b). Recent reports indicate that the trans contents in margarines and several other food sources that previously contained high amounts of trans fats have decreased this content tremendously (Elias and Innis, 2002; Friesen and Innis, 2006; Leth et al., 2006; Minister of Health Canada, 2006; Precht and Molkentin, 2000; Ratnayake et al., 2007; Stender et al., 2006). However, one question still remains as to whether vaccenic acid (or t-11, 18:1) has the same cholesterol-raising effects as elaidic acid (or t-9, 18:1). The TRANSFACT study (Chardigny et al., 2006) that is looking at the effects of mixtures of trans 18:1 isomers representing either a hydrogenated oil or a milkfat enriched in trans fatty acids (about 22% total trans 18:1 in both fats) on the high-density lipoprotein (HDL) cholesterol in humans should enable this very important question to be elucidated.

4.2

ISOMERIC FATTY ACIDS IN THE HUMAN DIET

The trans-octadecenoic acid isomers (trans 18:1) present in both ruminant fats (RF) and partially hydrogenated vegetable oils (PHVO), quantitatively and nutritionally are the most important group of trans fatty acids in the human diet (Wolff, 1995). The position of the double bond of these dietary trans 18:1 isomers, counting from the carboxylic carbon, usually varies from 4 to 16 carbon atoms of the fatty acid molecule. Very often in PHVO, the trans 18:1 isomers form a Gaussian distribution that centres around the 9 or 10 double bond. The distribution depends on the fatty acid composition of the starting vegetable oils and the extent

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

103

Table 4.1 Distribution of positional trans 18:1 isomers (as % trans 18:1 isomers) of some partially hydrogenated canola oil (PHCO) and soya bean (PHSO) base stocks hydrogenated to different extents. PHCO

PHSO

trans 18:1 isomer

IV 92

IV 80

IV 64

IV 109

IV 86

IV 64

4t 5t 6t 7t 8t 9t 10t 11t 12t 13t 14t 15t 16t Total t18:1 (as % total fatty acids)

— — — — 6.0 38.6 20.8 14.9 8.6 5.3 2.6 2.2 1.0 18.9

— — 0.2 1.0 9.9 28.7 25.5 16.1 9.4 5.1 2.2 0.9 1.0 38.3

0.1 0.1 0.5 3.8 13.4 23.0 23.3 15.7 9.7 5.7 3.1 1.3 0.3 39.8

— — 0.1 1.2 8.5 21.1 23.7 23.4 9.7 7.5 3.4 1.4 — 13.8

0.1 0.1 1.0 1.2 7.6 18.6 26.9 23.8 9.9 6.3 2.8 1.1 0.6 27.0

0.1 0.1 0.6 1.4 8.8 22.5 21.8 18.2 12.5 8.1 3.5 1.4 1.0 36.7

Total trans fatty acids (as % total fatty acids)

25.3

40.1

40.4

17.7

30.6

37.4

of hydrogenation. Usually, in partially hydrogenated canola oil basestocks, elaidic acid (i.e. t-9, 18:1) and t-10, 18:1 are the major isomers followed by vaccenic acid (t-11, 18:1). In partially hydrogenated soya bean oil basestocks, these three isomers are present in almost equal proportions (Table 4.1). The trans 18:1 isomer distribution of ruminant fats is distinctly different from that of PHVO. In ruminant fats, 11t-18:1 is always the major isomer, whereas 9t-18:1 and 10t-18:1 isomers occur in relatively low amounts (Wolff et al., 1998a). Recent analysis of Canadian dairy products showed that vaccenic acid accounts for about 30–36% of total trans 18:1 (Table 4.2), whereas dairy fat from Germany and France appear to have a higher proportion of this isomer (Wolff et al., 1998a). Values averaging 58% were reported for Table 4.2 Distribution of positional 18:1 trans isomers (% total trans 18:1 isomers) in some dairy products purchased from retail stores in Canada in May and June 2006. 18:1t isomer 4t 5t 6t-8t 9t 10t 11t 12t 13t 14t 15t 16t Total t18:1 (as % total fat) Total trans fatty acids (as % total fat)

Milk ( n = 6)

Butter (n = 7)

Cheese (n = 9)

Cream (n = 2)

0.7 ± 0.1 0.5 ± 0.1 2.0 ± 0.7 4.7 ± 0.8 11.4 ± 1.3 33.4 ± 1.8 10.4 ± 0.4 10.4 ± 0.4 11.6 ± 0.5 7.7 ± 0.3 7.3 ± 1.2 3.7 ± 0.4

0.7 ± 0.04 0.5 ± 0.05 2.6 ± 0.4 5.5 ± 0.4 11.8 ± 0.7 35.8 ± 2.0 9.5 ± 0.3 9.6 ± 0.3 10.6 ± 0.6 7.6 ± 0.2 5.8 ± 0.7 3.6 ± 0.3

0.9 ± 0.1 0.6 ± 0.1 1.9 ± 0.4 4.7 ± 0.7 10.5 ± 1.1 32.8 ± 2.3 11.0 ± 0.6 9.9 ± 0.5 12.0 ± 0.9 8.4 ± 0.5 7.4 ± 1.8 3.5 ± 1.4

0.6 ± 0.2 0.6 ± 0.2 3.7 ± 0.7 6.7 ± 0.5 13.4 ± 0.2 30.2 ± 0.3 10.4 ± 0.2 10.7 ± 0.2 11.5 ± 0.6 7.8 ± 0.0 4.3 ± 0.2 3.8 ± 0.1

5.6 ± 0.6

5.2 ± 0.3

5.3 ± 1.3

5.4 ± 0.1

BLUK122-Dijkstra

104

September 25, 2007

20:1

Trans Fatty Acids

Table 4.3 Distribution of positional cis 18:1 isomers of some PHCO and PHSO base stocks hydrogenated to different iodine values. PHCO

PHSO

18:1 cis isomer

IV 92

IV 80

IV 64

IV 109

IV 86

IV 63

6c 7c 8c 9c 10c 11c 12c 13c 14c 15c 16c

1.0 — 2.6 78.0 5.6 7.2 4.2 0.5 0.3 2.2 0.6

0.2 0.4 2.5 74.0 5.2 7.3 7.4 1.0 0.5 0.6 0.9

1.4 3.6 11.1 37.7 18.2 13.1 8.2 3.6 2.1 0.7 0.3

0.1 0.1 1.2 72.2 3.4 7.6 12.6 1.3 0.7 0.8 —

0.4 0.4 2.3 59.2 7.2 8.5 18.3 0.5 1.6 1.2 0.4

1.3 0.5 2.1 42.8 11.6 11.6 18.9 4.9 2.7 2.4 1.3

Total c18:1

59.4

40.9

29.2

30.0

44.6

30.5

spring bovine milk and 48% for samples collected during winter months (Wolff et al., 1998a). In addition to the trans 18:1 isomers, PHVO contain several cis-octadecenoic isomers (cis 18:1), whose double bond positions generally range from 6 to 16 (Table 4.3). The naturally occurring cis isomer, namely oleic acid (i.e. 9c-18:1), is always the predominant isomer followed by 10c-18:1 and 11c-18:1. Ruminant fats also contain several cis 18:1 isomers (Table 4.4) but the isomer distribution is less complex than that for PHVO. Oleic acid (9c-18:1) is also the predominant isomer and it accounts for approximately 95% of the total isomers (Table 4.4). Cis-vaccenic acid (11c-18:1) is the second major isomer that is followed by 12c-18:1 and 13c-18:1. In addition to 18:1 isomers, dietary fats may contain a number of positional and geometrical isomers of linoleic and -linolenic acids, which are frequently present in low concentrations in both partially hydrogenated and non-hydrogenated dietary fats (Ratnayake, 2004). PHVO contain 15 or more isomers of linoleic acids; the major ones generally are 9c,13t-18:2, 9c,12t-18:2 and 9t,12c-18:2 (Table 4.5). These isomers are often detected in large quantities in mildly hydrogenated vegetable oils (up to 6% of total fatty acids), whereas they are hardly detectable in heavily hydrogenated oils. The linoleic and -linolenic acid isomers present in non-hydrogenated fats, or in many common foods, fats are the result of exposure of these polyunsaturated fatty acids to some form of heat treatment, such as steam deodorisation, Table 4.4 Distribution of positional cis 18:1 isomers (% total cis 18:1 isomers) in some dairy products purchased from retail stores in Canada in May and June 2006. 18:1 cis isomer

Milk (n = 6)

Butter (n = 7)

Cheese (n = 9)

Cream (n = 2)

9c 11c 12c 13c

95.3 ± 1.0 2.8 ± 0.2 1.6 ± 0.4 0.3 ± 0.4

95.7 ± 0.7 2.6 ± 1.8 1.3 ± 0.4 0.4 ± 0.1

95.9 ± 5.1 2.5 ± 0.4 1.3 ± 0.3 0.3 ± 0.1

95.3 ± 0.1 2.7 ± 0.1 1.7 ± 0.1 0.3 ± 0.1

Total cis 18:1 isomers (as % total fat)

20.6 ± 0.8

22.1 ± 4.5

18.8 ± 3.6

20.6 ± 0.2

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

105

Table 4.5 Composition (wt % of total fatty acids) of linoleic acid isomers in some hydrogenated canola oil base stocks.

18:2 isomer 9c,13t+8t,13c 9c,12t 8c,13t 9t,12c 10t.15ca +9t,15c ttb 9t,12t 8c,13ca 9c,13ca 9c,14ca 9c,15c a b

Mildly hydrogenated Moderately hydrogenated Heavily hydrogenated canola oil (iodine canola oil (iodine canola oil (iodine value 92) value 80) value 64) 1.94 1.09 0.54 1.0 1.11 0.31 0.36 0.31 0.06 0.03 1.27

0.71 0.0 0.23 0.0 0.17 0.47 0.28 0.27 0.0 0.0 0.48

0.08 0.0 0.0 0.0 0.0 0.1 0.0 0.1 0.0 0.0 0.1

Tentative identification. Sum of four different tt-18:2 isomers with unknown double bond positions.

stripping during refining of oils (Ackman et al., 1974) or simple heating in deep fat frying (Grandgirard et al., 1984). In these processes, the double bonds do not shift in position but are isomerised from cis to trans, resulting in the formation of small amounts of geometric trans isomers. -Linolenic acid is more prone to isomerisation than linoleic acid, whereas oleic acid is not isomerised at all. In many non-hydrogenated dietary fats, usually the two mono-trans isomers of linoleic acid (i.e. 9c,12t-18:2 and 9t,12c-18:2) are present at similar levels and very often higher than the all-trans isomer 9t,12t-18:2. Eight geometric isomers are possible for -linolenic acid, but usually only four are present in industrially refined oils or oils subject to mild heat treatments. They have been identified as 9t,12c,15t-18:3, 9c,12c,15t18:3, 9c,12t,15c-18:3 and 9t,12c,15c-18:3 (Ackman et al., 1974; Grandgirard et al., 1984). Some of the isomers formed may be substrates for desaturation and elongations to produce trans isomers of arachidonic and docosahexaenoic acids. These can then be incorporated into membrane phospholipids (Grandgirard et al., 1994) and affect some biological functions, such as lipoprotein metabolism, platelet aggregation or retinal function (Acar et al., 2002; S´eb´edio et al., 2000). Positional and geometrical isomers of linoleic acid (18:2n-6) having two conjugated double bonds (CLA isomers) are also present in the human diet (dairy products) even if these have been detected in PHVO and heated fats. During the past two decades, especially since the discovery of Ha et al. (1987) who reported that an extract from grilled ground beef later identified as CLA had biological effects, numerous studies have been carried out first on animals and in recent years in humans, which have demonstrated the possible beneficial health effects of CLA isomers. A comprehensive review has recently been published by Tricon et al. (2005), and a review by Wahle will be found in Chapter 3. In food, CLA is found in products from ruminants such as milk and meat products made from them, such as cheese and culture dairy products (Parodi, 2003). CLA has been shown to be produced in the rumen as a result of biohydrogenation from linoleic acid,-which is always present in small quantities in grass and other cattle feeds. Rumenic acid (i.e. 9c,11t-18:2) is the major CLA isomer produced. CLA has also been shown to be produced in the mammary gland by 9 desaturation of vaccenic acid, which passed via the circulatory system to the mammary gland and adipose tissue (Corl et al., 2001). The second most important naturally occurring

BLUK122-Dijkstra

106

September 25, 2007

20:1

Trans Fatty Acids

CLA isomer, the 7t,9c-18:1 has been shown to be only produced in the mammary gland (Corl et al., 2002). Linolenic acid present in cattle feed does not directly produce rumenic acid but contributes to its level via biohydrogenation to vaccenic acid which can then be desaturated to 9c,11t-18:1 isomer. This review focuses on the state of the art methodologies for the analysis of trans fatty acids. Emphasis will be placed on the analysis of the trans 18:1 isomers, as these are still the major trans isomers consumed by humans (Graig-Schmidt, 2006), and on conjugated fatty acids, as recent studies have shown that all the CLA isomers do not seem to have the same biological properties (Tricon et al., 2005). Consequently, it is very important to have methods capable of determining with precision the composition of CLA fractions present in ruminant fats and biological tissues of humans who have been subjected to CLA consumption.

4.3

GAS CHROMATOGRAPHY AND FOURIER TRANSFORM INFRARED SPECTROSCOPY

Fourier transform infrared spectroscopy is a useful technique to quantify the total amount of trans fatty acids in complex matrices. Infrared spectroscopy is based on the C–H out of plane deformation band at 966 cm−1 , which is characteristic of isolated trans double bonds. Many modifications have been proposed to improve the accuracy of the infrared determination, but in any case this method cannot be used to quantify individual trans fatty acid isomers. However, when coupled to gas chromatography (GC), this technique (gas chromatography and Fourier transform infrared spectroscopy, GC-FTIR) can readily determine the geometry of double bonds GC-FTIR and gas–liquid chromatography coupled to mass spectrometry (GC-MS) are complementary techniques. While GC-MS permits the localisation of the ethylenic bond on the carbon chain after proper derivatisation, GC-FTIR is a powerful tool to confirm the double bond configuration. In the case of CLA, GC-FTIR analysis distinguishes between cis,trans; trans,cis; cis,cis and trans,trans isomers. Assignments of the cis,trans and trans,cis isomers are not possible after direct GC-FTIR analysis but can be made using the GC retention times. This review describes only methods allowing quantification of single trans isomers; further details on this subject can be found in the comprehensive review published by Mossoba et al. (2004).

4.4 4.4.1

DIRECT GC ANALYSIS Analysis of monounsaturated isomers

Analysis of cis and trans isomers is best carried out by GC using 100-m long, flexible, fused silica capillary columns coated with highly polar cyanopolysiloxane stationary phases, containing various polar substituents. A variety of cyanopolysiloxane capillary columns are available from chromatographic suppliers and marketed under trade designations such as SP-2560, SP-2340, CP-Sil 88, BPX-70 and HP-88. These cyanopolysiloxane columns enable analysts to separate the geometric and positional isomers of fatty acids (as their fatty acid methyl esters (FAME)) that cannot be resolved by columns coated with non-polar or moderately polar phases.

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

107

9c 9t

12c

11t

11c

12t

22.5

25

min

Fig. 4.1 Gas chromatogram of a standard FAME mixture of 9t-, 11t-, 12t-, 9c-,11c- and 12c-18:1. Analysis on an SP-2560 fused silica capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 180◦ C. Hydrogen carrier gas, flow rate 1 mL/min.

In cyanopolysiloxane capillary columns, the trans isomer always elute before the corresponding cis isomer. For examples, 9t-18:1 elutes before 9c-18:1, 11t-18:1 before 11c-18:1 and 12t-18:1 before 12c-18:1 (Fig. 4.1). Furthermore, within the trans or cis isomeric group, the positional isomers with the double bond closest to the carboxylic group (i.e. low value isomers) elute before the positional isomer, with the double bond farthest away from the carboxylic group (i.e. the high isomers). For example, the three trans 18:1 isomers shown in Fig. 4.1 elute in the order 9t-18:1, 11t-18:1 and 12t-18:1. However, overlaps could occur between some cis and trans isomers with different double bond positions (Fig. 4.2). These overlaps are especially noted when analysing partially hydrogenated fats, ruminant fats and other matrices that contain the full range of the positional cis and trans 18:1 fatty acid isomers. As a general rule, with any type of cyanopolysiloxane capillary column, if operated at usual operating conditions, the isomers from 4t-18:1 to 11t-18:1 are readily separated from all the cis 18:1 isomers. But after 11t-18:1 (i.e. trans isomers from 12t-18:1 to 16t-18:1), cis,trans overlapping problems start to occur with some cyanosilicone capillary columns. The extent of cis,trans 18:1 overlaps primarily depends on two factors: the length of the capillary column and the GC operating parameters. The extent of overlaps is quite severe on shorter capillary columns. For example, it has been observed that in the analysis of a PHVO using a 60-m SP2340 capillary column with isothermal column temperature at either 160 or 185◦ C, the 13t-, 14t- and 15t-18:1 isomers were completely overlapped with the major cis 18:1 peak which included five isomers (6c to 10c-18:1; Ratnayake and Beare-Rogers, 1990). In addition, there

BLUK122-Dijkstra

108

September 25, 2007

20:1

Trans Fatty Acids 9c 9t

12c 7c

12t 7t

22.5

11t 11c

25

min

Fig. 4.2 Gas chromatogram of a standard FAME mixture of 7t-, 9t-, 11t-, 12t-, 7c-, 9c-, 11c- and 12c-18:1. Analysis on an SP-2560 fused silica capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 180◦ C. Hydrogen carrier gas, flow rate 1 mL/min.

were two other cis,trans overlaps; the 12t-18:1 isomer only partially separated with this major cis 18:1 peak and 16t-18:1 overlapped with 14c-18:1. Similar overlapping problems occur even when the analysis is performed using a column temperature programme (Ratnayake, 1998). These overlapping problems are not limited to 60-m SP-2340 capillary column but also to other short-capillary columns coated with cyanosilicone capillary columns (e.g. 50-m CP-Sil 88; Wolff et al., 1998). These cis,trans isomer overlaps place a constraint on accurate determination of the total trans fatty acid content and, consequently, on the determination of the fatty acid composition of partially hydrogenated oils, ruminant fats and human tissue samples by direct GC alone using 60-m cyanosilicone columns. It is evident that considering solely the trans isomer peak eluting before the oleic acid peak will lead to considerable underreporting of trans fatty acids. In some fatty acid analyses, the results for total trans would be lower by about 30% (corresponding to the masked 12t- to 16t-18:1 isomers; Wolff et al., 1998). Improved separation of the cis- and trans 18:1 isomers can be achieved using 100-m-long cyanosilicone capillary columns. With the 100-m capillary columns, the overlaps between the cis and trans isomers can be minimised, provided that the proper operating conditions are selected. A recent study by Ratnayake et al. (2006) has shown that on 100-m SP-2560 and CP-Sil 88 capillary columns, the best separation of fatty acids of PHVO is achieved when the column temperature is operated isothermally at 180◦ C, using hydrogen as the carrier gas with a flow rate of 1.0 mL/min. Figures 4.3 and 4.4 show the 18:1 and 18:2 regions of the chromatogram for a hard margarine fat sample analysed on 100-m SP-2560 and CP-Sil 88

BLUK122-Dijkstra

September 25, 2007

20:1

24

18:2n−6

9c−18:1 26

109

28

9c,15c−18:2

9c,12t−18:2 9t,12c−18:2

9c,13t+8t,12c−18:2

15c−18:1 9t,12t−18:2

tt−18:2

14c−18:1

16t−18:1

11c−18:1 12c−18:1

11t−18:1 (13+14)t+(6−8)c−18:1

9t−18:1

10c+15t−18:1

13c−18:1

4t−18:1

5t−18:1

12t−18:1

6t−8t−18:1

10t−18:1

18:0

Analysis of trans mono- and polyunsaturated fatty acids

30

min

Fig. 4.3 The C18 region of gas chromatogram of FAMEs from a margarine sample analysed on an SP2560 fused silica capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 180◦ C. Hydrogen carrier gas, flow rate 1 mL/min.

capillary columns, respectively. Note that the 12t-18:1 peak is readily resolved from the oleic acid peak (9c-18:1). In addition, the 13t-18:1 and 14t-18:1 isomer pair, which always elute together in any GC analysis, is sufficiently separated from oleic acid that it can be quantified. As shown in Fig. 4.3, there should always be a distinct valley between the 13t+14t-18:1 peak and the oleic acid peak, though it does not need to be a large valley to get correct results. Furthermore, the 16t-18:1 isomer is also separated from the cis 18:1 isomers. Usually, the 15t-18:1 is the only trans 18:1 isomer that remains unresolved. Quantitatively, this is not an important drawback, since 15t-18:1 is a minor dietary component. Its level very often is less than 2% of total trans 18:1 isomers (Table 4.1). It should be noted here that the peak for 13t+14t-18:1 contains three cis 18:1 isomers, namely 6c-18:1, 7c-18:1 and 8c-18:1. These overlaps are of minor importance in many partially hydrogenated oils, where the total amount of these three isomers usually account for less than 2% of total fatty acids. However, these isomers are important in some partially hydrogenated oils, e.g. the heavily hydrogenated canola oil base stocks (Table 4.1). For such samples, the complete fatty acid composition needs to be determined by combining GC analysis with another separation technique such as silver nitrate thin-layer chromatography (described below). The column operating temperature has a considerable influence on the cis,trans isomer separation (Ratnayake et al., 2002). As discussed above, with both 100-m SP-2560 and CPSil 88 capillary columns, isothermal operation at 180◦ C produced the fewest overlapping peaks of the cis and trans isomers. Isothermal operations of the column above or below 180◦ C or column temperature programming greatly affects the cis,trans isomer separation

BLUK122-Dijkstra

110

September 25, 2007

20:1

Trans Fatty Acids CP-Sil 88 (100 m × 0.25 mm) 180°C isothermal

Counts 14000 18:0

9

12000

10

10000

11 12 13 14

6

8000

5

20 15 19 21 16 18 17

8 6000

4

7

3

26 25 24

23 4000

22

2

27 28

34+20:1 29

33

30

31

20:0

32

1 2000

24

26

28

30

32

34

36

38

40

min

Fig. 4.4 The C18 region of a GC chromatogram of FAMEs from a margarine sample analysed using a CP-Sil 88 capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 180◦ C. Hydrogen carrier gas, flow rate 1 mL/min. Peak identification: (1) 4t-18:1; (2) 5t-18:1; (3) (6t-8t)-18:1; (4) 9t-18:1; (5) 10t-18:1; (6) 11t-18:1; (7) 12t-18:1; (8) (13t+14t)-18:1+(6c-8c)-18:1; (9) 9c-18:1; (10) (15t+10c)-18:1; (11) 11c-18:1; (12) 12c-18:1; (13) 13c-18:1; (14) 16t-18:1; (15) 14c-18:1; (16) 15c18:1; (17–20) tt-18:2; (21) 9t,12t-18:2; (22) (9c,13t+8t,12c)-18:2; (23) 9c,12t-18:2; (24) 8c,13c-18:2; (25) 16c-18:1; (26) 9t,12c-18:2; (27) (9t,15c+10t,15c)-18:2; (28) 9c,13c-18:2; (29) 9c,14c-18:2; (30) 9c,15c-18:2; (31) 9t,12c,15t-18:3; (32) 9c,12c,15t-18:3; (33) 9c,12t,15c-18:3; (34) 9t,12c,15c-18:3. (Reprinted from J AOAC Int 87, 523–539. Copyright 2004 AOAC International.)

and the relative elution order of some of the 18:1 isomers, as well as the 11c-20:1 and geometric isomers of -linolenic acid. For instance, isothermal operation at 190◦ C results in overlap of 16t-18:1 isomer with the 13c-18:1 isomer peak (peaks 13 and 14 in Figs. 4.5 and 4.6). However, a minor advantage is the improved separation of the isomer pair 13t+14t18:1 from the oleic acid peak, but the peak for 13t+14t-18:1 still contains the (6-8)c-18:1 peaks. Operating the column temperature below 180◦ C also creates a few extra overlapping problems. For example, with isothermal operation at 170◦ C, as shown in Figs. 4.7 and 4.8, the resolution of 13t+14t-18:1 (peak 8) from oleic acid (peak 9) is lost; there is no distinct valley between these two peaks and therefore accurate measurement of the areas of these two peaks is difficult. In addition, the 16t-18:1 (peak 14) now overlaps with the 14c-18:1 isomer (peak 15). The only advantage at lower operating temperature is that there is now improved separation of the individual trans 18:1 isomers compared to that at 180◦ C or other higher temperatures. These overlapping problems are also seen with column temperature programme analysis (Fig. 4.9).

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

111

Counts 12000

18:0

SP-2560 (100 m × 0.25 mm) 190°C isothermal 18:2n−6

10000

9+10

8000 5

6 18:3n−3

4

6000

3

8 7

12 11 13+14

26 19 27 25 18 28 17 20 24 16 2123 15 22

4000 12

34+11c−20:1 30 29

20:0

31

32

33

2000 20

22

24

26

28

min

Retention time (min) Fig. 4.5 The C18 region of a GC chromatogram of FAMEs from a margarine sample analysed using a SP-2560 capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 190◦ C. Hydrogen carrier gas, flow rate 1 mL/min. For peak identifications, see Fig. 4.4. (Reprinted from J AOAC Int 87, 523–539. Copyright 2004 AOAC International.)

Increasing the column length above 100 m has only a minimal influence on the cis,trans isomer separation. It was reported by Sidisky et al. (1989) that increasing the column length of SP-2560 and SP-2380 columns from 100 to 150 m resulted in only minimal improvements in cis–trans isomer separation but was coupled with increased analysis time. Overall, 100-m columns provide best resolution of geometrical and positional isomers.

4.4.2

Isomers of linoleic and -linolenic acids

In the analysis of trans fatty acids, the main focus has always been on trans isomers of oleic acid; however, for accurate determination of the fatty acid profile of dietary fats, the isomers of linoleic and -linolenic acids, which are very often present in low concentration in human diets, should also be taken into consideration. The three geometrical isomers of linoleic acid, namely 9t,12t-18:2, 9c,12t-18:2 and 9t,12c18:2, which are very often present in refined, unhydrogenated liquid vegetable oils, are readily separated in the order stated on capillary columns coated with polar cyanosilicone columns (Fig. 4.10). The four common geometric isomers of -linolenic acid, i.e. 9t,12c,15t18:3, 9c,12c,15t-18:3, 9c,12t,15c-18:3 and 9t,12c,15c-18:3, which are also very often present in refined liquid vegetable oils, give peaks that can be readily recognised in GC analyses

BLUK122-Dijkstra

112

September 25, 2007

20:1

Trans Fatty Acids

Counts 14000

18:0

18:2n−6

9+10

CP-Sil 88 (100 m × 0.25 mm) 190°C isothermal

12000

10000 6 5 8000 4 6000

3

4000 1

7 8

11 12 13+14

18:3n−3 28

15 16

27

17 18

26 21 25 20 19 24 23 22

2

30 29 34 11c−20:1 20:0

31

32

2000 20

22

24

26

28

33

30 min

Retention time (min) Fig. 4.6 The C18 region of a GC chromatogram of FAMEs from a margarine sample analysed using a CP-Sil 88 capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 190◦ C. Hydrogen carrier gas, flow rate 1 mL/min. For peak identifications, see Fig. 4.4. (Reprinted from J AOAC Int 87, 523–539. Copyright 2004 AOAC International.)

on cyanosilicone capillary columns and are eluted from the column in order stated above (Fig. 4.11). The 18:2 isomer group of PHVO may contain up to 15 other 18:2 isomers (Table 4.5). The identification of these isomers poses some difficulties, because commercial standards of these isomers are not available, except for 9t,12t-18:2, 9c,12t-18:2 and 9t,12c-18:2. Fortunately, the elution patterns of these isomers have been established for SP-2560 and CP-Sil 88 columns, and GC chromatograms that illustrate the 18:2 region can be found in the literature (AOCS, 2005; Ratnayake et al., 2006; Ratnayake and Pelletier, 1992). As for the 18:1 isomers, the best separation of the various 18:2 isomers is obtained when the column temperature (with both SP-2560 and CP-Sil 88) operated isothermally at 180◦ C. In general, these isomers elute in the order trans,trans < trans,cis < cis,trans followed by cis,cis (Fig. 4.12). Please note that 9c,13t-18:2, which is the major 18:2 isomer formed during partial hydrogenation of vegetable oils, has a retention time close to that of 9t,12t-18:2, and in some GC analyses (especially at column temperatures above 190◦ C and column temperature programmes) these two isomers are not well resolved; they elute together. In such instances, because of the non-availability of a commercial standard of 9c,13t-18:2, there is a tendency to misidentify this isomer as 9t,12t-18:2. It is recommended that a FAME mixture from a partially hydrogenated vegetable oil with well-established fatty acid composition be used as the secondary standard, especially

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

113

Counts 18:0 18:2n−6 7000

SP-2560 (100 m × 0.25 mm) 170°C isothermal

9

6000 5 4

5000 3 4000 1

2

18:3n−3+11c−20:1 6 8 10 14+15 21 11 7 20 28 12 19 13 26 18 17 25 16 24 27 23 22

20:0

29 30

31 32

33 34

3000

30

32.5

35

37.5

40

42.5

45

47.5

50

52.5

min

Retention time (min) Fig. 4.7 The C18 region of a GC chromatogram of FAMEs from a margarine sample analysed using a SP-2560 capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 170◦ C. Hydrogen carrier gas, flow rate 1 mL/min. For peak identifications, see Fig. 4.4. (Reprinted from J AOAC Int 87, 523–539. Copyright 2004 AOAC International.)

for the identification of the peaks in the 18:2 region of the chromatogram. Such a standard is now available from the American Oil Chemists’ Society (Champaign, IL, USA).

4.4.3

Resolution of eicosenoic and -linolenic acid isomers

11c-Eicosenoic acid (11c-20:1) is a natural monounsaturated fatty acid present in appreciable amounts in some vegetable oils, such as high-erucic rape seed oil, peanut oil and canola oil. Animal fats, especially lard, also contain this fatty acid but at very low levels. The 11c20:1 elutes in the 18:3 region of the chromatogram, and its relative retention time with respect to the 18:3 varies with the column temperature. Depending on the temperature of the column, the 11c-20:1 isomer may elute before, with or after -linolenic acid (9c,12c,15c18:3). Therefore, an understanding of these variations is critical for the correct identification of all the peaks in the 18:3 region of the chromatogram and for achieving correct fatty acid composition data. Generally, on the SP-2560 capillary column, as for the 18:1 cis and trans isomer separation, isothermal operation at 180◦ C gives the best separation of 11c20:1 from -linolenic acid and its geometric isomers (Ratnayake et al., 2002) (Fig. 4.11). At column temperature above 180◦ C, the 11c-20:1 isomer coelutes with 9t,12c,15c-18:3

BLUK122-Dijkstra

114

September 25, 2007

20:1

Trans Fatty Acids

Counts

9000

18:0

CP-Sil 88 (100 m × 0.25 mm) 170°C isothermal

9

18:2n−6

8000

7000 6

6000

18:3n−3

5 21 4

5000

8 3 7

4000 2 3000

10

25

20 14+15

11c−20:1

19 18

11 12

17 13 16

1

24 23 22

28 27 26

30 29

20:0

31 32

34 33

2000 35

40

45

50

55

min

Retention time (min) Fig. 4.8 The C18 region of a GC chromatogram of FAMEs from a margarine sample analysed using a CP-Sil 88 capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 170◦ C. Hydrogen carrier gas, flow rate 1 mL/min. For peak identifications, see Fig. 4.4. (Reprinted from J AOAC Int 87, 523–539. Copyright 2004 AOAC International.)

(Fig. 4.5), and below 180◦ C it overlaps with -linolenic acid (Fig. 4.7). In contrast, the CPSil 88 column operated at 180◦ C does not cleanly separates 11c-20:1 from the -linolenic isomers; at this operating temperature, 11c-20:1 overlaps with 9t,12c,15c-18:3 (Fig. 4.4). However, isothermal operations of CP-Sil 88 column at 10◦ C above or below 180◦ C allowed baseline separation of 11c-20:1 from -linolenic acid and its geometrical isomers. At 190◦ C, 11c-20:1 elutes before 9t,12c,15c-18:3 and 9c,12t,15c-18:3 isomers (Fig. 4.6), whereas at 170◦ C, 11c-20:1 elutes after 9t,12c,15c-18:3 but before -linolenic acid (Fig. 4.8). It is recommended that pure standard 11c-20:1 and -linolenic acid be used for correct location of these fatty acids in the analysis of dietary fat samples.

4.4.4

Effect of the type of carrier gas and flow rate on cis and trans isomer resolution and fatty acid quantification

Recently, Ratnayake et al. (2006) tested the performances of hydrogen and helium as a carrier at different flow rates on the separation and quantification of FAME of dietary fats containing trans fatty acids. Table 4.6 shows the normalised area percentages concentration for total trans,trans 18:1, trans 18:2, trans 18:3, CLA, 16:0 and 18:0 of a margarine

BLUK122-Dijkstra

September 25, 2007

20:1

10c−18:1+15t−18:1

(13+14)t−18:1+(6−8)c−18:1

11c−18:1

9c−18:1

6

15c−18:1

8

5t−18:1

4t−18:1

10

13c−18:1

12

14c−18:1

10t−18:1

14

115

12c−18:1

16

16t−18:1

18

12t−18:1

6t−8t−18:1

20

11t−18:1

18:0

pA

9t−18:1

Analysis of trans mono- and polyunsaturated fatty acids

4 36

40

42

44

Fig. 4.9 The C18 region of a GC chromatogram of FAMEs from a margarine sample analysed using a SP-2560 capillary column (100 m ×0.25 mm i.d. × 20 m film thickness). Analysis performed by the following column temperature programme: initial temperature 45◦ C (hold 4 min), programmed to 175◦ C at 13◦ C/min (hold at 175◦ C for 27 min), programmed to 215◦ C at 4◦ C/min (hold at 215◦ C for 3 min). Hydrogen carrier gas, flow rate 1 mL/min.

sample when analysed on different columns (100-m SP-2560 and CP-Sil 88) with different carrier gases (hydrogen and helium) and flow rates (0.6, 0.8 and 1.0 mL/min). The results were identical at all different conditions for both columns. Even though the quantification differences were minimal, their study demonstrated an advantage of using hydrogen as the carrier gas at a flow rate of 1 mL/min. Use of hydrogen at the same flow rate as helium results in shortening the GC run time by 20 min. For example, on the 100-m SP-2560 column with Table 4.6 Effect of helium and hydrogen carrier gas flow rates on the fatty acid composition (% total fatty acids) of a partially hydrogenated margarine sample on 100-m CP-Sil 88 and SP-2560 columns. CP-Sil 88 Fatty acid Total trans 18:1 trans 18:2 trans 18:3 trans CLA 16:0 18:0

SP-2560

H2 H2 H2 He H2 He 0.6 mL/min 0.8 mL/min 1.0 mL/min 1.0 mL/min 1.0 mL/min 1.0 mL/min 34.41 28.06 6.09 0.26 0.55 11.79 11.52

From Ratnayake et al. (2006).

34.33 28.01 6.07 0.25 0.51 11.76 11.51

34.37 28.04 6.08 0.25 0.56 11.78 11.51

34.30 27.91 6.09 0.30 0.54 11.65 11.08

33.98 27.92 5.87 0.19 0.32 11.91 10.74

34.09 27.85 6.02 0.22 0.41 11.96 10.98

BLUK122-Dijkstra

20:1

25

9c,12t−18:2 9t,12c−18:2

9t,12t−18:2

13t−18:1

11c−18:1

18:2n−6

Trans Fatty Acids

9c−18:1

116

September 25, 2007

30

min

Fig. 4.10 The 18:2 region of a GC chromatogram of FAMEs from a refined canola oil sample analysed using a SP-2560 capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 180◦ C. Hydrogen carrier gas, flow rate 1 mL/min.

hydrogen carrier and the flow rate set at 1 mL/min, the 24:0 peak, usually the last to elute in many dietary fats, comes off in about 60 min. With helium as the carrier gas with the same flow rate, the same peak elutes at over 80 min. This time-saving could be of benefit to most of the lipid analytical laboratories where run time is very important. In addition, as shown in Table 4.7, hydrogen carrier gas gives a better resolution of the critical pairs of 13t+14t-18:1 and 9c-18:1, 16t-18:1 and 14c-18:1 and of 11c-20:1 and 9c,12c,15c-18:3. Furthermore, hydrogen produced sharper peaks (calculated as the ratio of peak height to area) than helium, and 1.0 mL/min produced the sharpest peaks compared to flow rates of 0.6 and 0.8 mL/min (Table 4.8). It was also noted by Ratnayake et al. (2006) that there was no loss of resolution between flow rates of 0.6, 0.8 and 1.0 mL/min in the early eluting peaks for either hydrogen or helium. However, for the separation of the peaks in the 18:3 region, flow rate 1.0 mL/min provided the most satisfactory separation. These data suggest that hydrogen carrier gas at 1.0 mL/min and isothermal column temperature operation at 180◦ C are the ideal operating conditions for the analysis of fatty acid profiles of vegetable and animal fats and oils containing trans fatty acids.

4.4.5

Conjugated fatty acids

As previously mentioned, the major CLA isomer in ruminant fat is rumenic acid (9c,11t-18:2). This isomer represents more than 70% of the total CLA isomers but it is accompanied by a

BLUK122-Dijkstra

September 25, 2007

20:1

117

32.5

11c−20:1

9t,12c,15c−18:3

9c,12t,15c−18:3

9c,12c,15t−18:3

9t,12c,15t−18:3

20:0

18:3n−3

Analysis of trans mono- and polyunsaturated fatty acids

35

37.5

min

28

30

9c,15c−18:2

9c,14c−18:2

20:0

18:2n−6 9c,13c−18:2

9c,12t−18:2

16c−18:1 9t,12c−18:2 9t,15c+10t,15c−18:2

8c,13c−18:2

ct/tc−18:2 8c,13t−18:2

9c,13t−18:2 (major)+8t,12c−18:2 (minor)

9t,12t−18:2

tt−18:2

15c−18:1

16t−18:1 14c−18:1

13c−18:1

Fig. 4.11 The 18:3 region of a GC chromatogram of FAMEs from a refined canola oil sample analysed using a SP-2560 capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 180◦ C. Hydrogen carrier gas, flow rate 1 mL/min.

32

min

Fig. 4.12 The 18:2 region of a GC chromatogram of FAMEs from a margarine sample analysed using a SP-2560 capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 180◦ C. Hydrogen carrier gas, flow rate 1 mL/min.

BLUK122-Dijkstra

118

September 25, 2007

20:1

Trans Fatty Acids Table 4.7 Effect of type of carrier gas (hydrogen vs helium) on resolution factors for critical pairs of fatty acids at a flow rate of 1.0 mL/min at 180◦ C on both the CP-Sil 88 and SP-2560 columns. Resolution factor CP-Sil 88

SP-2560

Critical pair of fatty acids

Hydrogen

Helium

Hydrogen

Helium

(13+14)t-18:1 and 9c-18:1 9c-18:1 and 11c-18:1 16t-18:1 and 14c-18:1 11c-18:1 and 9c,12,15c-18:3

1.3 2.0 0.5 2.3

ND 2.0 ND 1.9

1.1 2.4 ND 1.3

1.0 2.3 0.3 1.0

ND, not determined. Data adapted from Ratnayake et al. (2006).

complex mixture of cis,cis; cis,trans and trans,trans isomers. Similarly, due to the number of scientific articles published on the potential beneficial effects of some CLA isomers, numerous CLA supplements are now available on the market. However, it should be noted that not all CLA isomers produce beneficial health effects; at least one CLA isomer, namely 10t,12c-18:2, has been shown to have adverse effects in some studies (Riserus et al., 2004). In analysing CLA samples from either natural or synthetic origins, it is therefore important to be able to identify and quantify all of the different isomers. Some earlier studies reported CLA content of dairy products but it is only since the development of sample preparation and sophisticated chromatographic methods (Christie, 2003) that problems of coeluting peaks and artefacts formations could be avoided. Furthermore, analysis of commercial CLA supplements that contain large quantities of isomers may be quite straightforward, while detection of minor CLA isomers in animal or human tissues may require a pre-fractionation step. Before fatty acids can be analysed by GC these have to be first converted to more volatile compounds, such as methyl esters. It is now well accepted that acid-catalysed esterification is undesirable, as this leads to the formation of trans,trans CLA isomer artefacts, as well as some other undesirable compounds (Yurawecz et al., 1999). Base-catalysed esterification using sodium methoxide in methanol has been demonstrated as the method of choice (Yurawecz et al., 1999), and a typical preparation scheme has been described by Christie et al. (2001). Capillary columns using a carbowax phase are of very little use for the analysis of complex CLA mixtures. Highly polar long capillary columns of length 100–120 m are usually preferred (Christie, 2003). An example is shown in Fig. 4.13. An elution order using natural and synthetic CLA has been published by Roach et al. (2002), using a 100-m CP-Sil 88 column Table 4.8 Peak sharpness of 9c-18:1 peak at different flow rates and with different carrier gases on the CP-Sil 88 and SP-2560 columns. Hydrogen 9c-18:1 Height Area Height/area

0.6 mL/min 0.8 mL/min 1.0 mL/min CP-Sil 88 CP-Sil 88 CP-Sil 88 103.4 1455 0.071

From Ratnayake et al. (2006).

132.3 1536 0.086

151.1 1518.1 0.0995

Helium 1.0 mL/min 1.0 mL/min 1.0 mL/min SP-2560 CP-Sil 88 SP-2560 152.0 1636.0 0.093

92.8 1212.9 0.077

147.3 2336.4 0.063

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

119

Table 4.9 Elution order of positional and geometrical CLA isomers on GLC (100-m CP-Sil 88 column) and on AgNO3 -HPLC. c/t-18:2

c,c-18:2

t,t-18:2

GLC

HPLC

GLC

HPLC

GLC

HPLC

7c,9t 8c,10t 7t,9c 9c,11t 8t,10c 10c,12t 9t,11c 11c,13t 10t,12c 12c,14t

12,14c/t 11t,13c 11c,13t 10,12c/t 9c,11t 9t,11c 8,10c/t 7t,9c

8,10 9,11 10,12 11,13 12,14

12,14 11,13 10,12 9,11 8,10 7,9

12,14 11,13 10,12 9,11 8,10 7,9

12,14 11,13 10,12 9,11 8,10 7,9

Adapted from Roach et al. (2002) and Kramer et al. (1999).

(Table 4.9). Under these conditions, the elution order is the cis,trans, followed by the cis,cis followed by the trans,trans isomers. However, overlap occurred between the cis,trans and the cis,cis isomers. The elution order decreased within the cis,trans isomers as the delta value of the cis ethylenic bond increased. When two cis,trans isomers have the same delta value for the cis ethylenic bond, the isomer having the lowest delta value for the trans bond elutes first. The elution order for the cis,cis isomers increases with increasing delta values, while that of the di-trans isomers decreases with increasing delta values. On this type of column, CLA isomers elute long after the non-conjugated dienes but a major challenge when analysing CLAs of natural origin is to avoid interferences of the minor CLA isomers, with other naturally occurring minor fatty acids often present in ruminant fats, such as 21:0 and 20:2 positional isomers (Kramer et al., 2004).

16:0

14:0

9c11t

CLA

18:1

trans-trans 18:0 10:0

9c11c

12:0 9t11c

6:0

0

10t12c

8:0 18:2

10

20

30

40

50 min

Fig. 4.13 GC analysis of the total fatty acid methyl esters of a milkfat sample on CP-Sil 88, 100 m × 0.25 mm i.d., film thickness 0.2 m, temperature programming 60–170◦ C at 20◦ C/min, using hydrogen as the carrier gas (S´ eb´ edio and Juaneda, 2007).

BLUK122-Dijkstra

120

September 25, 2007

20:1

Trans Fatty Acids 21:0 t10,c12

c9,t11 c11,t13

t8,t10− t10,t12

9,11

t8,c10

11,13

c,c

t11,t13

21:0

t9,c11

(a)

c9,c11 t11,c13

48

49

min

(b)

Fig. 4.14 GC analyses of (a), a commercial mixture of CLA isomers spiked with 21:0 and (b), a milkfat sample on CP-Sil 88, 100 m. (Adapted from Kramer et al., 2004. For experimental conditions see CruzHernandez et al., 2004.)

GC-MS is then the ideal tool to detect any non-conjugated fatty acid isomers, which may coelute with CLA isomers, as shown in Fig. 4.14. Among the different derivatives used for GC-MS analysis, dimethyloxazoline (DMOX) has been used extensively, as this gives spectra that are easier to interpret (Dobson, 2003). For example, as shown in Fig. 4.15, the mass spectra of 10t,12c-18:2 contain a very prominent molecular ion at m/z 333. In the saturated hydrocarbon part of the fatty acid chain, the difference between adjacent ions is 14 amu, corresponding to the loss of CH2 units. When a carbon–carbon double bond is present, this pattern is broken. The presence of a double bond between carbons at position n and n+1 corresponds to a gap of 12 amu between ions containing n-1 and n carbons (delta 10). Similarly, the gap of 12 amu between ions at 236 (position m-1) and 248 (position m) indicates a double bond at the delta 12 position. It is also interesting to note that ions at m+2 and m+3 are more prominent than the ions at m-1 and m. The m+2 and m+3 correspond to ions at 262 and 276, respectively, for the 9c,11t-18:1 isomer (Fig. 4.15). Roach et al. (2002) identified seven diagnostic DMOX fragment ions for each of the CLA isomers studied. They are listed in Table 4.10 for the isomers with double bond positions from C6 to C15 . The two pairs of ions at n-1, n and m-1, m can be used for identification purposes but the allylic cleavage ions at m+2 and m+3 which are of greater intensity are also very useful. For more details about relative intensities of the diagnostic ions, one can refer to the publication of Roach et al. (2002). Another technique is to use high-resolution selected ion monitoring on the fatty acid methyl esters (Roach et al., 2000). Another alternative approach that uses acetonitrile chemical ionisation mass spectrometry was recently described by Michaud et al. (2005). Chemical ionisation of acetonitrile produces an abundant ion at m/z 54, which adds covalently to a carbon–carbon double bond to give a

120

140

140

3

126

4

160

154

9,11−18:2

113

2

196

n−2

n 222 m−1 n−1 236 210 248

m

262

168

5

180

182

6

200

196

8

208

Δ9

182

7

196

208

220

222

9

234

Δ11 248

234

11

240

222

10

13

260

262

m+2

12

276

14

280

276

m+3

262

304

170 180 190 200 210 220 230 240 250 260 270 280 290 300

182

290

m+3

290

15

300

304

16

17

320

318

333

M+

333

18 CH3

340

Mass spectra of the dimethyloxazoline derivatives of (a) 10t,12c-18:2 and (b) 9t,11c-18:2 isomers (S´ eb´ edio and Juaneda, 2007).

113

126

O

1

0 m/z–>

400

800

1200 168

1600

276

m+2

Analysis of trans mono- and polyunsaturated fatty acids

Fig. 4.15

0 m/z–>

10000

20000

30000

40000

50000

60000

70000

N

10,12−18:2

2000

2400

2800

Abundance

September 25, 2007

(b)

(a)

BLUK122-Dijkstra 20:1

121

BLUK122-Dijkstra

122

September 25, 2007

20:1

Trans Fatty Acids Table 4.10 Diagnostic fragments for DMOX derivatives of CLA isomers with double bond positions from C6 to C15 . CLA isomers

n-2

n-1

n

m-1

m

6,8 7,9 8,10 9,11 10,12 11,13 12,14 13,15

140 154 168 182 196 210 224 238

154 168 182 196 210 224 238 252

166 180 194 208 222 236 250 264

180 194 208 222 236 250 264 278

192 206 220 234 248 262 276 290

m+2 m+3 220 234 248 262 276 290 304 318

Molecular ion

234 248 262 276 290 304 318 332

333 333 333 333 333 333 333 333

Adapted from Roach et al. (2002).

[M+54]+ ion. Under collisionally activated dissociation, the latter ion produces diagnostic ions ( and ) that correspond to the cleavage of the carbon–carbon bond. The relative abundance of these two ions depends on the geometry of the double bonds, and therefore permits assignment of the double bond geometry. In any case, great care must be taken when analysing samples containing natural CLA isomers, such as those from milkfat as minor fatty acids; for example, 21:0 does not always coelute with the same conjugated isomers. For example, Kramer et al. (2004) have shown that its retention time changes with the age of the column (Fig. 4.16). Consequently, GC analysis of mixture of natural CLAs and biological samples containing these conjugated fatty acids should be coupled to GC-MS analysis in order to confirm the structure of the components. For this purpose, DMOX are the preferred derivatives, as these are not only well resolved but their MS spectra are easy to interpret. 21:0 c11,t13 c9,t11

t,t

t10,c12 t8,c10

8,10 9,11 10,12

11,13

c,c 21:0

21:0

48

49

min

Fig. 4.16 GC analyses of a commercial CLA mixture spiked with 21:0 on different CP-Sil 88 columns, 100 m in length. (Adapted from Cruz-Hernandez et al., 2004.)

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

4.5

4.5.1

123

SILVER NITRATE THIN-LAYER AND HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY SEPARATION OF CIS AND TRANS ISOMERS Monounsaturated fatty acid isomers

It is apparent from the above discussion that direct GC analysis using capillary columns may produce a few minor overlaps of cis and trans 18:1 isomers (particularly 15t-18:1 with 9c-18:1 or 10c-18:1 isomers, and 13t+14t-18:1 with 6c+7c+8c-18:1) but the data should be accurate enough for many applications, including food labelling, nutrition studies and quality control work. However, if precise data on trans fats are required, then GC analysis has to be used in conjunction with another separation technique, for example silver ion chromatography, either in the thin-layer chromatography mode or high-performance liquid chromatography (HPLC) mode. Silver nitrate thin-layer chromatography (AgNO3 -TLC) is a very inexpensive technique, which fractionates fatty acids (as fatty acid methyl esters) on the basis of the number, the configuration and, to some extent, the position of the double bonds. In this way, saturated, monounsaturated and polyunsaturated fatty acids as well as their cis and trans isomers can be separated. Analyses of these fractions by capillary GC allow complete and accurate determination of cis and trans isomers and therefore, the fatty acid composition. The first step in AgNO3 -TLC is the preparation of the TLC plate with a uniform layer of silver nitrate. Various techniques are available for the preparation of silver nitrate TLC plates. The procedure routinely in use in one of the authors’ laboratory (W.M.N.R.) is described below. Commercial pre-coated silica gel TLC plate (20 cm × 20 cm, thickness 0.5 mm for preparative work) is pre-washed to remove dust and other particles by developing in a TLC developing tank containing either ethyl acetate or chloroform/methanol and then activated in an oven at 105–110◦ C for an hour. Allow the plate to come to room temperature and then place it horizontally (the side containing the silica gel facing downwards) in a glass tray containing 5% solution of silver nitrate in acetonitrile for 30 min. The plate is activated by heating in an oven at 110◦ C for about an hour. It is advisable to use the plate immediately after preparation; if not, it should be stored in a dessicator over drying agents in a dark place. About 4–5 mg of the FAME sample (dissolved in hexane) is applied to the plate using a TLC streaker or a disposable glass pipette and developed using either 100% toluene or a mixture of hexane and diethyl ether (90:10, vol/vol). The plates are normally developed at room temperature in a dark place to minimise oxidation of highly unsaturated fatty acids. The separated bands are made visible by spraying the plate with a 0.1% solution of 2 ,7 -dichlorofluorescein in ethanol and examining under UV light (234 nm). The various bands are scraped off and extracted with diethyl ether or a 1:1 mixture of hexane and chloroform, and then analysed by GC using 100-m SP-2560 or CP-Sil 88 column. The most important application of AgNO3 -TLC is for the determination of total trans 18:1 and total cis 18:1 isomers. This is possible because, as shown in Plate 4.1 (positioned within text of Chapter 9), AgNO3 -TLC gives a clean separation of the trans 18:1 isomers as a group from the cis 18:1 isomers and other fatty acids. Isolation and analysis of these two bands by GC permit quantification of the levels of trans and cis 18:1 isomers. Figure 4.17 shows the GC traces of the trans and cis 18:1 fractions of a margarine FAME obtained after AgNO3 TLC fractionation shown in Plate 4.1. In practice, a convenient means of quantification of the trans 18:1 fraction is to treat the trans 18:1 isomers, with the double bond close to the carboxyl group (from 6 to 11 ), as the internal standard. This is possible, because as discussed earlier, the 6t-18:1 to 11t-18:1 isomers are always well resolved from the cis 18:1

BLUK122-Dijkstra

124

September 25, 2007

20:1

Trans Fatty Acids 10t

11t

12t

13t+14t

9t

6t+7t+8t

18:0

15t 4t

16t

5t

24

26

min

9c

11c

10c

13c

6c+7c+8c

15c 12c

24

26

14c

28

min

Fig. 4.17 GC chromatograms of the trans and cis 18:1 AgNO3 -TLC bands of the margarine sample shown in Fig. 4.13. Analysis on a SP-2560 fused silica capillary column (100 m × 0.25 mm i.d. × 20 m film thickness), operated isothermally at 180◦ C. Hydrogen carrier gas, flow rate 1 mL/min.

on polar cyanosilicone capillary GC columns. In this method, the proportion of the trans 18:1 isomers from 12 to 16 that overlap with the cis 18:1 isomer peak on capillary GC is calculated by comparing the 18:1 region of the GC chromatogram of the isolated trans 18:1 with that of the parent fatty acid methyl esters prior to AgNO3 -TLC fractionation. This calculation is done with respect to the well-separated trans 18:1 isomers (i.e. sum of 6t-18:1 to 11t-18:1). The total trans 18:1 content is then calculated by summing the proportion of the trans 18:1 isomers (12t-18:1 to 16t-18:1) that overlap with the cis isomers with the wellseparated trans 18:1 isomers (from 6t-18:1 to 11t-18:1). This also allows calculation of the total cis 18:1 content. This approach eliminates the errors resulting from sample application, scraping losses, incomplete extraction, weighing of small quantities of internal standard and isolated bands. This procedure was used to determine the fatty acid composition, including the trans fatty acids of Canadian human milk (Chen et al., 1995a) and adipose tissue (Chen et al., 1995b) samples.

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids t10,c12

125

c9,t11

trans,trans t8,c10

9,11

10,12

8,10 7,9

8,10

12,14

HPLC

cis,cis 11,13

11,13

10,12 9,11

c11,t13

(a) t10,c12

c9,t11 t8,c10

10,12 11,13

8,10 9,11

cis,cis

trans,trans 10,12 11,13

c11,t13

GC

9,11 8,10

(b) Fig. 4.18 Chromatograms of a CLA preparation. (a) Silver ion HPLC separation using three columns in series and a mixture of hexane:diethyl ether:acetonitrile (99.4:0.5:0.1) as the mobile phase. (b) GC separation on a CP-Sil 88 column, 100 m in length. (Adapted from Kramer et al., 2004.)

Another approach for quantification of the trans 18:1 involves collection of the AgNO3 TLC bands corresponding to saturated and the trans 18:1 together followed by GC analysis of the combined fraction (Chardigny et al., 1996). This procedure allows the use of endogenous 16:0 or 18:0 as the internal standard. Comparison of the GC peak area ratios of 18:0 (or 16:0) to trans 18:1 before (i.e. the starting, unfractionated total FAME sample) and after AgNO3 -TLC fractionation permits calculation of the actual proportion of trans 18:1 in the starting FAME sample.

4.5.2

Conjugated fatty acids

As previously shown, GC analysis even on long polar columns permits only partial resolution of the complex mixture of CLA isomers present in either synthetic samples or milkfat and meat from ruminants. Considering the number of positional and geometrical isomers involved, methods using silver ion chromatography have been developed (Adlof, 2003; Christie, 2003; Cruz-Hernandez et al., 2004, Delmonte et al., 2004; Roach et al., 2002). Silver ion chromatography enables separation according to the number, configuration and position of double bonds in molecules. In much of the earlier work, it was carried out in conjunction with TLC but in recent years with HPLC. AgNO3 -HPLC methods developed over the years were based on the work of Christie (1995) that showed that a silica matrix containing chemically

BLUK122-Dijkstra

126

September 25, 2007

20:1

Trans Fatty Acids trans,trans CP-Sil 88, 100 m

9c11t

10c12t 9t11c

10t12c 9c11c 10c12c

8t10c

8c10c

11c13t

(a)

10t12t

9t11t

c/t + t/c 10,12

9,11

8t10t 11t13t

11,13

8,10

10c,12c 11c,13c

20

9c,11c 8c,10c

63 min (b)

Fig. 4.19 Chromatograms of the CLA fraction as methyl esters isolated from sunflower oil. (a) GC analysis on a CP-Sil 88 column, 100 m in length. (b) Silver ion HPLC on two chromspher lipids columns. (Adapted from Juaneda et al., 2003.)

bonded phenylsulphonic acid can exchange its protons for silver ions which will not elute with the solvent. Various AgNO3 -HPLC procedures are described in the literature, but in this chapter, we will review only the most efficient procedures for analysing mixtures of CLA isomers in complex matrices. For a more complete review on this subject, one may refer to the articles of Adlof (2003) and Kramer et al. (1999). The first attempt to analyse CLA mixtures as their methyl esters was initiated by Sehat et al. (1998) who reported that AgNO3 -HPLC carried out on one AgNO3 column using 0.1% acetonitrile in hexane separates all the geometrical and positional isomers, especially for the cis,trans, trans,cis isomers improved using three columns in a series. Further improvement was achieved using a solvent system hexane: diethyl ether: acetonitrile of 99.4:0.5:0.1 (Delmonte et al., 2004). This solvent system, as shown in Fig. 4.18, gives a baseline separation

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

127

16 14 12

C16:0

Fraction 1

10

C18:2 CLA

8 6 4

RP-HPLC

Fraction 2 C18:1 cis

Fraction 3 C18:1 trans

C18:3

C 18:0

2 0 20.0

12t14t

30.0

9c11t 12,14 13c13t

10.0

11t13t 10t12t 9t11t 8t10t 7t9t

0.0

40.0

50.0

(min)

7t9c AgNO3-HPLC Fraction 1 9c11c

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62

Fig. 4.20 Chromatograms of milk fatty acid methyl esters by reversed-phase HPLC (two Kromasil C18 colums, acetonitrile at 4 mL/min (top) and of fraction 1 isolated from total milkfat using silver ion HPLC on two chromspher columns. (Adated from Juaneda, 2002.)

of the various trans,trans isomers, whereas GC analysis of the same CLA mixture shows only two peaks for the trans,trans isomers. It is now very clear that if all of the 24 CLA isomers so far detected in dairy fats (Table 4.9) have to be resolved, one will have to use a combination of GC using 100-m highly polar capillary columns and silver ion HPLC (Delmonte et al., 2004). However, the utilisation of two columns in series is in most cases a good compromise, for example to analyse frying fats (Fig. 4.19) that may contain complex mixtures of CLA isomers due to heat treatment (Juaneda et al., 2003). While a better separation was obtained by GC for the cis,trans isomers, HPLC gave a detailed composition of the all trans CLAs.

4.6

UTILISATION OF PRE-FRACTIONATION STEPS PRIOR TO CHROMATOGRAPHIC ANALYSIS: THE CASE OF DAIRY FATS

It is sometimes mandatory to use pre-fractionation steps especially when fatty acids to be analysed are in small concentrations. This is the case for milk lipids for which it is impossible to get, in a single run without any pre-fractionation, the complete profile of both the complex mixture of CLAs and of trans 18:1 isomers. Recently, a method was developed to fractionate

BLUK122-Dijkstra

128

September 25, 2007

20:1

Trans Fatty Acids GC of Fraction 2

Δ9

cis Δ10 Δ7

36

37

Δ11 38

39

Δ12

Δ13 40

Δ11

Δ9

41

42

43

GC of Fraction 3

Δ10

Δ13

Δ14

trans

Δ12

Δ6–7–8

Δ15 36

Δ15

Δ14

37

38

Δ16 39

40

41

Δ17 42

43

Fig. 4.21 GC analyses of fractions 2 and 3 collected from milkfat (Fig. 7) on a BPX-70 column of 120 m using hydrogen as the carrier gas. (Adapted from Juaneda, 2002.)

milkfat lipids as fatty acid methyl esters by reversed-phase liquid chromatography (Juaneda, 2002) using two columns in series, as reported in Fig. 4.20. Using this technique, trans 18:1 isomers are completely separated from their cis counterparts that are all eluting as a single peak along with 16:0. The CLA isomers are eluting with the non-conjugated 18:2 isomers. This combined 18:2 fraction may be reanalysed by AgNO3 -HPLC, which gives a better separation compared to GC as previously described. On the other hand, GC analysis of the two 18:1 fractions (Fig. 4.21) using an internal standard permits quantification of the cis and trans 18:1 isomers.

4.7

CONCLUSION

Currently, detailed fatty acid composition data of fats containing trans and cis isomeric fatty acids can be obtained by GC analysis alone using 100-m capillary columns coated with cyanosilicone liquid phases and operated isothermally at 180◦ C using hydrogen as the carrier gas at a flow rate of 1 mL/min. These conditions guarantee optimal separation of all the fatty acids generally encountered in dietary fats and animal tissues. The only drawback is the minor overlaps of some cis and trans 18:1 isomers; however, these isomers could be easily quantified by combining GC analysis with AgNO3 -TLC fractionation. Analysis of CLA isomers either present as fractions used for nutritional intervention or in milkfat samples is a tedious problem, as these fatty acids are heat sensitive and can undergo isomerisation if great care is not taken for sample preparation and analysis. Precise isomer composition can only be obtained using complementary techniques, such as GC on long highly polar columns and AgNO3 -HPLC. Analysis may also require a pre-fractionation step.

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

129

Confirmation of the isomer structures is sometimes mandatory in order to detect the overlap with other minor fatty acids also present in the same samples.

REFERENCES Acar, N., Chardigny, J.M., Bonhomme, B., Almanza, S., Doly, M. & S´eb´edio, J.-L. (2002) Long term intake of trans n-3 polyunsaturated fatty acids reduces the b-wave amplitude of electroretinograms in rats. J Nutr 132, 3151–3154. Ackman, R.G., Hooper, S.N. & Hooper, D.L. (1974) Linolenic acid artifacts from the deodorization of the oils. J Am Oil Chem Soc 51, 42–49. Adlof, R.O. (2003) Application of silver-ion chromatography to the separation of conjugated linoleic acid isomers. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie, & R. Adlof). AOCS Press, Champaign, IL, pp. 37–55. AOCS. (2005) Determination of cis-, trans-, saturated, monounsaturated and polyunsaturated fatty acids in vegetable or non-ruminant animal oils and fats by capillary GLC method. In: Official and Recommended Practices of the AOCS, 5th edn. Revisions and Corrections. AOCS Press, Champaign, IL. Official Method Ce 1h-05, approved 2005, revised 2005. Aro, A. (2006) The scientific basis for trans fatty acid regulations-is it sufficient? A European perspective. Atheroscler Suppl 7, 67–68. Chardigny, J.M., Malpuech-Brugere, C., Dionisi, F. et al. (2006) Rationale and design of the TRANSFACT project phase I: a study to assess the effect of the two different dietary sources of trans fatty acids on cardiovascular risk factors in humans. Contemp Clin Trials 27, 364–373. Chardigny, J.-M., Wolff, R.L., Mager, E. et al. (1996) Fatty acid composition of French infant formulas with emphasis on the content and detailed profile of trans fatty acids. J Am Oil Chem Soc 73, 1595–1601. Chen, Z.Y., Pelletier, G., Hollywood, R. & Ratnayake, W.M.N. (1995a) Trans fatty acid isomers in Canadian human milk. Lipids 30, 15–21. Chen, Z.Y., Ratnayake, W.M.N., Fortier, L., Ross, R. & Cunnane, S.C. (1995b) Similar distribution of trans fatty acid isomers in partially hydrogenated vegetable oils and adipose tissue of Canadians. Can J Physiol Pharmacol 73, 718–723. Christie, W.W. (2003) Analysis of conjugated linoleic acid: an overview. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adlof). AOCS Press, Champaign, IL, pp. 1–12. Christie, W.W. (1995) Silver-ion high performance liquid chromatography. In: New Trends in Lipid and Lipoprotein Analyses (eds J.-L. S´eb´edio & E.G. Perkins). AOCS Press, Champaign, IL, pp. 59–74. Christie, W.W., S´eb´edio, J.-L. & Juaneda, P. (2001) A practical guide to the analysis of conjugated linoleic acid. Inform 12, 147–152. Corl, B.A., Baumgard, L.H., Dwyer, D.A., Griinari, J.M., Phillips, B.S. & Bauman, D.E. (2001) The role of delta 9-desaturase in the production of cis9,trans11 CLA’. J Nutr Biochem 12, 622–630. Corl, B.A., Baumgard, L.H., Griinari, J.M. et al. (2002) Trans7-cis 9 CLA is synthesized endogenously by delta 9-desaturase in dairy cows’. Lipids 37, 681–688. Cruz-Hernandez, C., Deng, Z., Zhou, J. et al. (2004) Methods for analysis of conjugated linoleic acids and trans 18:1 isomers in dairy fats by using a combination of gas chromatography, silver-ion thin layer chromatography/gas chromatography, and silver ion liquid chromatography. J AOAC Int 87, 545–562. Delmonte, P., Yurawecz, M.P., Mossoba, M.M., Cruz-Hernandez, C. & Kramer, J.K.G. (2004) Improved identification of conjugated linoleic acid isomers using silver-ion HPLC separations. J AOAC Int 87, 563–568. Dobson, G. (2003) Gas chromatography-mass spectrometry of conjugated linoleic acids and metabolites. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adlof). AOCS Press, Champaign, IL, pp. 13–36. Dutton, H.J. (1979) Hydrogenation of fats and its significance. In: Geometrical and Positional Fatty Acid Isomers (eds. E.A. Emken & H.J. Dutton). AOCS Press, Champaign, IL, pp. 1–16. Elias, S.L. & Innis, S.M. (2002) Bakery foods are the major dietary source of trans-fatty acids among pregnant women with diets providing 30 percent energy from fat. J Am Diet Assoc 102, 46–51.

BLUK122-Dijkstra

130

September 25, 2007

20:1

Trans Fatty Acids

Friesen, S. & Innis, S.M. (2006) Trans fatty acids in human milk in Canada declined with the introduction of trans fat food labeling. J Nutr 136, 2559–2561. Graig-Schmidt, M.C. (2006) World-wide consumption of trans fatty acids. Atheroscler Suppl 7, 1–4. Grandgirard, A., Bourre, J.M., Julliard, F. et al. (April 1994) Incorporation of trans long-chain n-3 polyunsaturated fatty acids in rat brain structures and retina. Lipids 29(4), 251–258. Grandgirard, A., S´eb´edio, J.-L. & Fleury, J. (1984) Geometrical isomerization of linolenic acid during heat treatment of vegetable oils. J Am Oil Chem Soc 61, 1563–1568. Ha, Y.L., Grimm, N.K. & Pariza, M.W. (1987) Anticarcinogens from fried ground beef: heat altered derivatives of linoleic acid. Carcinogenesis 8, 1881–1887. Juaneda, P. (2002) Utilisation of reversed-phase high performance liquid chromatography as an alternative to silver ion chromatography for the separation of cis and trans C18:1 fatty acid isomers. J Chromatogr A 954, 285–289. Juaneda, P., Brac de la Perriere, S., S´eb´edio, J.-L. & Gregoire, S. (2003) Influence of heat and refining on formation of CLA isomers in sunflower oil. J Am Oil Chem Soc 80, 937–940. Kramer, J.K.G., Cruz-Hernandez, C., Deng, Z., Zhou, J., Jahreis, G., & Dugan, M.E.R. (2004). Analysis of conjugated linoleic acid and trans 18:1 isomers in synthetic and animal products. Am J Clin Nutr 79, 1137S–1145S. Kramer, J.K.G., Sehat, N., Fritsche, J. et al. (1999) Separation of conjugated fatty acid isomers. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yurawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza, & G.J. Nelson). AOCS Press, Champaign, IL, pp. 83–109. Leth, T., Jensen, H.G., Mikkelsen, A.E. & Bysted, A. (2006) The effect of the regulation on trans fatty acid content in Danish food. Atheroscler Suppl 7, 53–56. Mensink, R.P., Zock, P.L., Kester, A.D.M. & Katan, M.B. (2003) Effects of dietary fatty acids and carbohydrates on the ratio of serum total to HDL cholesterol and on serum lipids and apolipoproteins: a meta analysis of 60 controlled trials. Am J Clin Nutr 77, 1146–1155. Michaud, A.L., Lawrence, P., Adlof, R. & Brenna, J.T. (2005) On the formation of conjugated linoleic acid diagnostic ions with acetonitrile chemical ionization tandem mass spectrometry. Rapid Commun Mass Spectrom 19, 363–368. Minister of Health Canada (2006) Transforming the food supply. Report of the Trans Fat Task Force submitted to the Minister of Health, June 2006. www.healthcanada.ca/transfat. Mossoba, M.M., Yurawecz, M.P., Delmonte, P. & Kramer, J.K. (2004) Overview of infrared methodologies for trans fat determination. J AOAC Int 87, 540–544. Mozaffarian, D., Katan, M.B., Ascherio, A., Stampfer, M.J. & Willett, W.C. (2006) Trans fatty acids and cardiovascular disease. N Engl J Med 354, 1601–1613. Parodi, P.W. (2003) Conjugated linoleic acid in food. In: Advances in Conjugated Linoleic Acid Research, Vol. 2 (eds J.-L. S´eb´edio, W.W. Christie & R. Adlof). AOCS Press, Champaign, IL, pp. 101–122. Precht, D. & Molkentin, J. (2000) Recent trends in the fatty acid composition of German sunflower margarines, shortening and cooking fats with emphasis on individual C16:1, C18:1, C18:2, C18:3 and C20:1 isomers. Nahrung 44, 222–228. Precht, D., Molkentin, J., Destaillats, F. & Wolff, R.L. (2001) Comparative studies on individual isomeric 18:1 acids in cow, goat, and ewe milk fats by low-temperature high-resolution capillary gas-liquid chromatography. Lipids 36, 827–832. Ratnayake, W.M.N. (1998) Analysis of trans fatty acids. In: Trans Fatty Acids in Human Nutrition (eds J.-L. S´eb´edio & W.W. Christie). Oily Press, Dundee, Scotland, pp. 115–161. Ratnayake, W.M.N., Hansen, S.L. & Kennedy, M.P. (2006) Evaluation of the CP-Sil 88 and SP-2560 GC columns used in the recently approved AOCS official method Ce 1h-05: determination of cis-, trans-, saturated, monounsaturated, and polyunsaturated fatty acids in vegetable or non-ruminant animal oils and fats by capillary GLC method. J Am Oil Chem Soc 83, 475–488. Ratnayake, W.M.N. (2004) Overview of methods for the determination of trans fatty acids by gas chromatography, silver-ion thin-layer chromatography, silver-ion liquid chromatography, and gas chromatography/mass spectrometry. J AOAC Int 87, 523–539. Ratnayake, W.M.N. & Beare-Rogers, J.L. (1990) Problems of analyzing C18 cis- and trans-fatty acids of margarine on the SP-2340 capillary column. J Chromatogr Sci 28, 633–639. Ratnayake, W.M.N., Gagnon, C., Dumais, L. et al. (2007) Trans fatty acid content of Canadian margarines prior to mandatory trans fat labelling. J Am Oil Chem Soc (in press). Ratnayake, W.M.N. & Pelletier, G. (1992) Positional and geometrical isomers of linoleic acid isomers in partially hydrogenated oils. J Am Oil Chem Soc 69, 95–105.

BLUK122-Dijkstra

September 25, 2007

20:1

Analysis of trans mono- and polyunsaturated fatty acids

131

Ratnayake, W.M.N., Plouffe, L.J., Pasquier, E. & Gagnon, C. (2002) Temperature-sensitive resolution of cisand trans-fatty acid isomers of partially hydrogenated vegetable oils on SP-2560 and CP-Sil 88 capillary columns. J AOAC Int 85, 1112–1118. Ratnayake, W.M.N. & Zehaluk, C. (2005) Trans fatty acids in foods and their labelling regulations. In: Healthful Lipids (eds C.C. Akoh & O.-M. Lai). AOCS Press, Champaign, IL, pp. 1–32. Riserus, U., Vessby, B., Arnlov, J. & Basu, S. (2004) Effects of cis-9,trans-11 conjugated linoleic acid supplementation on insulin sensitivity, lipid peroxidation, and proinflammatory markers in obese men. Am J Clin Nutr 80,279–283. Roach, J.A.G., Mossoba, M.M., Yurawecz, M.P. & Kramer, J.K.G. (2002) Chromatographic separation and identification of conjugated linoleic acid isomers. Anal Chim Acta 465, 207–226. Roach, J.A.G., Yurawecz, M.P., Kramer, J.K.G., Mossoba, M.M., Eulitz, K. & Ku, Y. (2000) Gas chromatography high resolution selected ion mass spectrometric identification of trace 21:0 and 20:2 fatty acids eluting with conjugated linoleic acid isomers. Lipids 35, 797–802. S´eb´edio, J.-L. & Juaneda, P. (2007) Isomeric and cyclic fatty acids as a result of frying. In: Deep Fat Frying (ed. M. Erickson). AOCS Press, Champaign, IL, pp. 57–86. S´eb´edio, J.-L., Vermunt, S.H., Chardigny, JM. et al. (2000) The effect of dietary trans alpha-linolenic acid on plasma lipids and platelet fatty acid composition: the TransLinE study. Eur J Clin Nutr 54, 104–413. Sehat, N., Yurawecz, M.P., Roach, J.A., Mossoba, M.M., Kramer, J.K.G. & Ku, Y. (1998) Silver-ion high performance liquid chromatographic separation and identification of conjugated linoleic acid isomers. Lipids 33, 217–221. Sidisky, L.M., Stormer, P.L., Nolan, L., Keeler, M.J. & Bartram, R.J. (1989) High temperature partially cross-linked cyanosilicone capillary column for general purpose gas chromatography. J Chromatogr Sci 26, 320–324. Stender, S., Dyerberg, J., Bysted, A., Leth, T. & Astrup A. (2006) A trans world journey. Atheroscler Suppl 7, 47–52. Tricon, S., Burge, G.C., Williams, C.M. & Calder, P.C. (2005) The effects of conjugated linoleic acid on human health-related outcomes. Proc Nutr Soc 64, 171–182. Willett, W.C. (2006a) Trans fatty acids and cardiovascular disease: epidemiological data. Atheroscler Suppl 7, 5–8. Willett, W.C. (2006b) The scientific basis for TFA regulations – is it sufficient? Comments from the USA. Atheroscler Suppl 7, 69–71. Wolff, RL. (1995) Content and distribution of trans-18:1 acids in ruminant milk and meat fats: their importance in European diets and their effect on human milk. J Am Oil Chem Soc 72, 259–272. Wolff, R.L., Combe, N.A., Precht, D., Molkentin, J. & Ratnayake, W.M.N. (1998b). Accurate determination of trans-18:1 isomers by capillary gas-liquid chromatography on cyanoalkyl polysiloxane stationary phases. Oleagineux Corps Gras Lipides 5, 295–300. Wolff, R.L., Precht, D. & Molkentin, J. (1998a) Occurrence and distribution profiles of trans 18:1 acids in edible fats of natural origin. In: Trans Fatty Acids in Human Nutrition (eds J.-L. S´eb´edio & W.W. Christie). Oily Press, Dundee, Scotland, pp. 1–33. Yurawecz, M.P., Kramer, J.K.G. & Ku, Y. (1999) Methylation procedures for conjugated linoleic acid. In: Advances in Conjugated Linoleic Acid Research, Vol. 1 (eds M.P. Yurawecz, M.M. Mossoba, J.K.G. Kramer, M.W. Pariza & G.J. Nelson). AOCS Press, Champaign, IL, pp. 64–82.

BLUK122-Dijkstra

5

September 25, 2007

20:1

Controlling physical and chemical properties of fat blends through their triglyceride compositions

Albert J. Dijkstra

5.1

INTRODUCTION

Traditional fatty products like butter, salad oil or lard shortening were originally produced without recourse to fat modification processes, such as blending, hydrogenation, interesterification and fractionation. Subsequently, substitute products made good use of the possibilities offered by these processes. As recounted by Van Alphen (1964), even the first margarine, a butter substitute, involved fractionation since it was based on a tallow olein fraction for which M`ege-Mouri`es invented a fractionation process (1869). Fat modification processes were also used to improve product properties. Salad oils were given a more balanced fatty acid composition by blending a number of different vegetable oils. In the USA, the keepability of soya bean oil was extended by brush hydrogenation followed by fractionation (or ‘winterisation’), and the plastic range of lard was extended by interesterification, which could be random (Dominick et al., 1953; Vander Wal and Van Akkeren, 1951) or directed (Hawley and Holman, 1956). Moreover, just as the use of tallow olein in margarine alleviated a butter shortage in France, the invention of the hydrogenation process (Normann, 1903) alleviated a shortage of solid fats used in margarine and shortening manufacture by enabling vegetable oils and also fish oil and whale oil to be converted into solid fats. Proper control of the modification processes listed above requires analytical support, and as indicated in the previous chapter, it is much easier to determine the fatty acid composition of a fatty sample than its triacylglycerol (‘triglyceride’, for short) composition. Consequently, much attention has been given to fatty acid compositions, despite the fact that fat properties follow from its triglyceride composition rather than from its fatty acid composition. A simple example can illustrate this distinction. Cocoa butter is a sharp-melting fat. Randomising the fatty acid distribution of its fatty acids over the glycerol moieties does not affect the fatty acid composition but greatly affects the triglyceride composition and, consequently, the physical properties. The randomised cocoa butter contains about 30% trisaturated triglycerides, has a much higher melting point and a totally different mouthfeel. It therefore cannot be used to produce anything like chocolate. As shown in Table 5.1, the various oil modification processes affect the analytical parameters that are related to the fatty acid composition much less than they affect the parameters that are related to the triglyceride composition. In an indirect way, this means that the oil modification processes given in Table 5.1 are more effective in arriving at the fat properties aimed for than blending alone. They generate novel triglycerides and/or triglyceride compositions, whereas blending only leads to a composition according to the weighted

BLUK122-Dijkstra

September 25, 2007

20:1

Controlling physical and chemical properties of fat blends Table 5.1

133

The effect of modification processes on analytical parameters. Fractionation

Analytical parameter

Hydrogenation Interesterification

Stearin

Fatty acid composition

More trans and saturates Decrease No change No change

No change

More saturates More unsaturates

No change No change Reflecting randomisation Reflecting randomisation

Decrease Increase Slight decrease Slight increase Slight change Slight change

Iodine value Saponification value Carbon number distribution Partition number distribution

Reflecting saturation

Change

Olein

Change

average of the blended components; these processes are also more expensive than blending (Dijkstra, 2002).

5.2

DEFINING TRIGLYCERIDE COMPOSITIONS

Triglyceride compositions can be described in various ways. There is the complete description listing every triglyceride that is present. For non-fractionated vegetable oils, such triglyceride compositions can be calculated by assuming the 1,3-random, 2-random distribution (Vander Wal, 1960) to be valid and determining the fatty acid composition of the 2-position and the overall fatty acid composition and calculating the fatty acid composition of the 1,3-position by difference. Since the saturation and the cis,trans isomerisation reactions during hydrogenation occur without positional selectivity (Beyens and Dijkstra, 1983), the triglyceride compositions of hydrogenated vegetable oils can also be calculated. For oils that have been randomised by interesterification, the overall fatty acid composition allows their triglyceride composition to be calculated. Consequently, the only products for which this complete triglyceride composition cannot be calculated on the basis of fatty acid compositions are fractions and mixtures. The latter include directed interesterification products and animal fats. The 1,3-random, 2-random approach is probably also valid for specified portions of animal fat, like for example suet, but in practice, trimmings are rendered as a mixture so that animal fat is also a mixture. This also holds for fish oil. For practical purposes, however, the complete triglyceride composition contains superfluous information. When developing a margarine fat blend, the temperature range of interest and in which the solid fat content (SFC) matters runs from refrigeration temperature to body temperature, and what SFC values the blend has at temperatures outside this range is immaterial. Accordingly, simplifying the triglyceride composition is fully justified and the literature describes a number of ways to do this. A method that is often used for describing margarine fat blends (Holemans et al., 1988; Schijf et al., 1983) is based upon grouping of fatty acids: H M L U

a saturated or trans monounsaturated fatty acid moiety with 16 or more carbon atoms; a saturated fatty acid moiety with 12–14 carbon atoms; a saturated carbon atom with less than 12 carbon atoms; and cis unsaturated fatty acids.

BLUK122-Dijkstra

134

September 25, 2007

20:1

Trans Fatty Acids

Table 5.2

melting points (◦ C) of triglycerides.

Monoacid triglycerides Structure

C12 C12 C12

C14 C14 C14

CN

36

42

Diacid triglycerides m.p.

35

47

C16 C16 C16

48

56

C18 C18 C18

54

65

Structure

CN

m.p.

C12 C10 C10 C10 C12 C10 C10 C12 C12 C14 C10 C10 C12 C10 C12 C10 C14 C10 C16 C10 C10 C10 C16 C10 C14 C12 C12 C10 C14 C14 C18 C10 C10 C12 C14 C12 C14 C10 C14 C10 C18 C10 C12 C14 C14 C16 C12 C12 C14 C12 C14 C12 C16 C12 C10 C16 C16 C18 C12 C12 C16 C10 C16 C12 C18 C12 C16 C14 C14 C12 C16 C16 C14 C16 C14 C16 C12 C16 C14 C16 C16 C18 C14 C14 C16 C14 C16 C14 C18 C14 C18 C10 C18 C12 C18 C18 C18 C12 C18 C18 C16 C16 C14 C18 C18 C18 C14 C18 C16 C18 C16 C16 C18 C18 C18 C16 C18

32 32 34 34 34 34 36 36 38 38 38 38 38 38 40 40 40 40 42 42 42 42 44 44 44 44 46 46 46 46 46 48 48 50 50 50 50 52 52

26 34 31 31 33 30 32 36 39 38 38 44 40 40 42 43 45 43 41 42 48 43 51 50 55 50 52 52 55 53 53 52 58 60 58 59 65 61 64

Triacid triglycerides Structure

CN

m.p.

C14 C12 C10

36

34

C18 C10 C12 C18 C12 C10

40 40

42 40

C18 C10 C14 C18 C14 C10 C16 C14 C12

42 42 42

50 42 44

C18 C18 C18 C18 C18 C18

C16 C14 C12 C10 C16 C12

44 44 44 44 46 46

54 52 46 47 56 52

C18 C14 C16 C18 C16 C14

48 48

56 56

C10 C12 C14 C16 C12 C16

CN, carbon number; m.p., slip melting point (◦ C). From Bailey, 1950, tables 21 and 22.

Grouping fatty acids according to chain length permits triglycerides to be grouped according to their carbon number, and as shown in Table 5.2 for saturated triglycerides, this carbon number is strongly indicative of the triglyceride melting points as pure compounds (Bailey, 1950). As shown in Table 5.3, the triglycerides with the highest melting points (up to 65◦ C) are the H3 group. According to Schijf et al. (1983), elaidic acid is regarded as an H-fatty acid.

BLUK122-Dijkstra

September 25, 2007

20:1

Controlling physical and chemical properties of fat blends Table 5.3

135

’ melting points (◦ C) of triglyceride groups.

Triglyceride group

Carbon number range

Melting point (◦ C)

H3 H2 M HM2 ; H2 L HML; M3 HL2 ; M2 L ML2 L3

48–54 44–50 40–46 36–42 32–38 28–34 24–30

59 54 48 41 35 25 <10

Given the melting points of 58◦ C for elaidodistearin and 52◦ C for elaidodipalmitin, it can be concluded that introducing an elaidic acid moiety into a triglyceride affects its melting point by about the same amount as the introduction of a myristic acid or palmitic acid. Its melting point contribution is therefore somewhat lower than that of stearic acid and this constitutes a major advantage of elaidic acid; it provides consistency to products without causing a sticky mouthfeel. Table 5.2 allows average melting points to be calculated for the various triglyceride groups. The calculation results are summarised in Table 5.3. In addition to the triglycerides comprising only saturated fatty acids and elaidic acid, some triglycerides may also comprise one or more cis unsaturated fatty acids and these can be monounsaturated or polyunsaturated. By introducing cis unsaturated fatty acids as a separate fatty acid group (U), the lower melting points of triglycerides containing cis unsaturated fatty acids can also be taken into account. Given the melting points of oleodistearin, oleodipalmitin and oleodimyristin of 37, 29 and 19◦ C respectively, oleic acid can be given a nominal carbon number of 1–2 whereby this nominal carbon number can be used to estimate the melting point of the triglyceride concerned. Since triglycerides comprising polyunsaturated fatty acid moieties have even lower melting points, it is evident that few triglycerides containing unsaturated acids will contribute to the consistency of the fat blend. Oleodistearin is the only triglyceride that can contribute to the SFC but only when its concentration is sufficiently high and the temperature is sufficiently low. Accordingly, the grouping of fatty acids permits triglycerides to be grouped according to their melting points and this latter grouping facilitates understanding what happens during fat modification processes. It also permits the definition of some of the criteria a fat blend has to meet, such as SFC profile and mouthfeel. This grouping is therefore a valuable product development tool but as will be explained in Chapter 9, the processing of the fat blends also has other aspects that have to be taken into account.

5.3

MELTING POINTS AND SFC

Like almost every compound, pure triglycerides have sharp melting points. Below this melting point, the SFC equals 100%, and above this point, the triglyceride is fully molten. In nature – and thus in the food industry – oils and fats are always mixtures of different triglycerides, which means that they have a melting range. Cocoa butter, which consists of almost 80% of symmetrical monounsaturated triglycerides that contain only palmitic, stearic and oleic acid, has a relatively narrow melting range but fats that comprise both H3 and U3 triglycerides have a very wide melting range.

BLUK122-Dijkstra

136

September 25, 2007

20:1

Trans Fatty Acids

According to Gunstone and Harwood (2007), palm oil contains 5.4% tripalmitin (m.p. 56◦ C) and 1.0% dipalmitostearin (m.p. 59◦ C), both H3 triglycerides. What does this mean for the melting point of palm oil? This is much lower at 36.7◦ C (Gunstone and Harwood, 2007) because liquid oil can dissolve some trisaturated triglycerides. How much the oil can dissolve depends upon the temperature. At the melting temperature of the trisaturated triglycerides, liquid oil and molten trisaturates are fully miscible. When the temperature is lowered, the solubility decreases and in the case of palm oil, it has apparently decreased to some 6.4% at 36.7◦ C. Palm oil on its own has a somewhat sticky mouthfeel. To ensure a good meltdown and mouthfeel, margarine fat blend specifications could therefore list that the H3 content of the blend should not exceed 2 or perhaps 3%. If an oil blend comprises only H3 and U3 triglycerides, it would have a melting point that depends upon its H3 content and below this melting point, the SFC would gradually increase and finally equal the H3 content. In practice, oil blends also comprise other types of triglycerides. When the temperature is lowered, the solubility of these other types of triglycerides will also be exceeded at some stage, so that they will also start to contribute to the SFC of the oil blend. Accordingly, it is conceivable that an SFC profile be derived from the triglyceride composition as expressed in triglyceride groups and that, conversely, limits can be specified for these triglyceride groups so that the resulting oil blend meets its SFC specification. Specifying such limits not only ensures that the resulting oil blend is within specification, it also permits a cost optimisation of these blends by the linear programming technique. This technique can also accommodate other criteria (constraints), such as a minimum linoleic acid content, a maximum linolenic acid content and/or a maximum trans content. However, in linear programming, the introduction of a further constraint (such as maximum trans content) never leads to a lower blend cost and usually to a higher blend cost. Limiting the trans content of fat blends therefore increases their cost.

5.4

THE EFFECT OF OIL PROCESSING ON TRIGLYCERIDE GROUPS

Because of the way triglyceride groups have been defined, the effect of oil modification processes will be discussed in terms of these groups. Topics to be discussed are therefore to what extent hydrogenation, fractionation and interesterification will affect those triglyceride groups that turn out to be critical in linear programming of fat blends. After all, meeting the critical criteria is what often necessitates the use of more expensive blend components and where, hopefully, some money can be saved by adjusting the oil modification process, so that blend component variants result that are more suited to the linear programming demands. In this context, ‘more suited’ means that the use of these variants leads to a lower blend cost. After the individual modification processes have been discussed, the benefits of combined use will be illustrated with some examples from the literature.

5.4.1

Hydrogenation

Hydrogenation leads to the saturation and isomerisation of double bonds. Accordingly, this modification process causes some of the cis unsaturated fatty acids (U) originally present to be

BLUK122-Dijkstra

September 25, 2007

20:1

Controlling physical and chemical properties of fat blends

137

converted into either saturated or trans unsaturated fatty acids (H). Since the unsaturated acids involved invariably have a chain length of more than 16 carbon atoms, the hydrogenation process converts U to H. This conversion is not purely random. Because the suggestion (Bailey and Fisher, 1946) of a ‘common fatty acid pool’ turns out to be invalid, a U moiety in H2 U has a higher probability to react than the same U moiety in HU2 (Dijkstra, 1997). Therefore, the triglyceride group composition must be arrived at by a determination of the fatty acid compositions of the 2-position and the 1,3-positions and subsequent calculation in accordance with the 1,3-random, 2-random fatty acid distribution (Vander Wal, 1960). As will be explained in Chapter 6, partial hydrogenation invariably leads to the formation of trans isomers. Since industrial hydrogenation processes are characterised by a limited hydrogen supply to the catalyst surface, they tend to be ‘selective’, which means that little stearic acid is formed as long as there is a minimum amount of linoleic acid left; this amount can be some 15%. When the linoleic acid concentration falls below this level, monounsaturated fatty acids start to get saturated and this can lead to the formation of trisaturated triglycerides. These trisaturated triglycerides are classified as H3 but elaidodipalmitin is also classified as H3 . A calculation example can illustrate this H3 formation. Soya bean oil normally contains some 12 wt % palmitic acid and 4 wt % stearic acid. The total saturates content is therefore 16 wt % but since the saturated fatty acids are positioned only on the 1- and 3-position of the glycerol moiety, the concentration on those positions is 24 wt %. On a molar basis, this amounts to 25.64%. Accordingly, the triglyceride group composition of soya bean oil is as follows: H3 H2 U HU2 U3

= 0% = (0.2564)2 × 100% = 6.6 mol % = 2 × 0.2564 × 0.7436 × 100% = 38.1 mol % = (0.7436)2 × 100% = 55.3 mol %

By the time the linoleic acid content has decreased to 15%, all linolenic acid will have been saturated, so that a fatty acid composition can be estimated that still lists 12% C16:0 and 4% C18:0 and in addition, 69% C18:1 and 15% C18:2 ; this corresponds to an iodine value (IV) of 85. Depending upon the hydrogenation conditions, a trans content of some 30% may well have been realised by then; this corresponds to an isomerisation index of 0.64 (%/IV). Pursuing the hydrogenation a bit further to an IV of 70 causes the linoleic acid content to decrease further to 5 wt % and the stearic acid content to increase from the original 4 to 12 wt %. At the same time the trans content increases to some 40%. Given the original fatty acid compositions of the 1,3- and 2-positions (see Table 3.122 in Padley et al., 1994), the triglyceride group compositions can be calculated after estimates have been made of how the fatty acid compositions of the various positions will change in the course of the hydrogenation. The calculation results have been summarised in Table 5.4. The data in Table 5.4 are only approximate. They do not take minor amounts of fatty acids into account nor the fact that the fatty acid composition of the 1-position differs slightly from that of the 3-position. Because the spread in relative molecular mass between the various fatty acids is small, the weight percentages have not been converted to molar fractions and back either. Nevertheless, they clearly show how soya bean oil changes from a liquid oil without H3 triglycerides to a fat with increasing amounts of H3 . They also show that as and when H3 is being created, the concentration of H2 U increases even more. These latter triglycerides provide a margarine fat blend with some consistency, without contributing to a sticky mouthfeel.

BLUK122-Dijkstra

138

September 25, 2007

20:1

Trans Fatty Acids

Table 5.4 bean oil.

Change of fatty acid composition and triglyceride groups during hydrogenation of soya

Iodine value

132

85

70

Trans content

0

30

40

Position

overall

C16:0 C18:0 C18:1 cis C18:1 trans C18:2 C18:3 H U H3 HUH HHU H2 U UHU HUU HU2 U3

12 4 22 0 54 8 16 84

1,3 18 6 22 0 45 9 24 76 0.0 5.7 0.0 5.7 0.0 36.5 36.5 57.8

2 0 0 22 0 72 6 0 100

overall

1,3

12 4 39 30 15 0 46 54

18 6 36 27 13 0 51 49 9.4 16.6 18.0 34.6 8.6 32.0 40.6 15.4

2 0 0 45 36 19 0 36 64

overall

1,3

12 12 31 40 5 0 64 36

18 14 27 37 4 0 69 31 25.7 21.9 23.1 45.0 5.2 19.7 24.9 4.4

2 0 8 39 46 7 0 54 46

Hydrogenation therefore does a number of things. It forms H3 , which must be limited for most applications (mouthfeel). It also forms some trans isomers and when their formation is reduced, the formation of stearic acid tends to be increased so that trans isomer reduction still leads to H3 formation. The process also forms H2 U-type triglycerides, and finally, it reduces the polyunsaturated fatty acid content in general and the linolenic acid content in particular. This latter aspect is considered to be more important in the United States than in Europe (Dijkstra, 2005). Accordingly, the entire fat product used in margarine, for instance, tended to be hydrogenated whereas in Europe, as will be explained in Chapter 8, fat products tend to be blends of liquid unmodified oils and one or more hardstocks.

5.4.2

Fractionation

As will be discussed in more detail in Chapter 7, the fractionation process constitutes a powerful tool for the modification of fat properties. In this process, a fat is allowed to form crystals by cooling, after which these crystals are separated from the mother liquor. The usual separation method is filtration and the resulting filter cake (the stearin fraction) will contain varying amounts of entrained mother liquor (the olein fraction). Consequently, the properties of the stearin depend very much on the way the fractionation process is conducted. Nevertheless, some general conclusions can be made with respect to the triglyceride groups. During cooling, the triglycerides with the highest melting points will start to crystallise first. If present, these are the H3 triglycerides. With the lowering of the temperature, more and more H3 triglycerides will crystallise and gradually, some H2 U triglycerides will start to crystallise as well. After filtration, the cake will therefore be enriched in H3 and H2 U triglycerides, which implies that the olein will be enriched in the other types of triglycerides. These enrichments reveal themselves in the properties of the fractions. Non-fractionated palm oil has an IV of 51–53 and a melting point of some 37◦ C; its H3 content is some 5.4 wt % (Gunstone and Harwood, 2007). Palm stearin has a lower IV, for instance 32–36 and a much

BLUK122-Dijkstra

September 25, 2007

20:1

Controlling physical and chemical properties of fat blends Table 5.5

Selected properties of palm oil and its fractions. Stearin

Iodine value Drop point C14:0 C16:0 C16:1 C18:0 C18:1 C18:2 C18:3 C20:0 H U

139

Olein

Palm oil

First

Second

First

Second

51–53 37–39 1.1 43.7 0.2 4.4 39.9 9.7 0.2 0.4 48.5 49.8

32–36 54–56 1.3 60.8 0.1 5.4 26.1 5.4 0.1 0.4 66.6 31.6

17–21 61–63 1.1 77.4 0.1 4.7 13.4 2.9 — 0.3 82.4 16.3

57–59 18.0 1.0 38.8 0.2 4.1 43.8 10.8 0.2 0.4 42.9 54.8

64–66 1.1 33.0 0.3 3.6 46.4 14.2 0.3 0.3 36.6 60.9

Adapted from Deffense, 1995.

higher drop melting point (48–50◦ C). As illustrated by Table 5.5, fractionation generates products with quite different properties. This table shows the H and U contents of the various fractions but does not show their H3 contents. The triglyceride group contents of fractions cannot be calculated from the H and U contents but can only be arrived at by analysis. On a qualitative basis, it is clear that the H3 content of an olein corresponds to the solubility of these H3 triglycerides at filtration temperature and that most of the unsaturated fatty acids present in a stearin will stem from entrained mother liquor (olein). The increased H3 contents of stearin fractions are illustrated by their melting points. Triglycerides containing one or more elaidic acid moieties are to some extent concentrated in the stearin in that a H2 O or HMO triglyceride (O, oleic acid) is less likely to crystallise than a H2 E or HME triglyceride (E, elaidic acid). For low-melting triglycerides, like U2 O and U2 E, the presence of a trans isomer makes no difference, since these triglycerides will remain liquid anyway.

5.4.3

Interesterification

Randomisation of triglycerides allows the triglyceride composition of the reaction product to be calculated from its fatty acid composition after this has been converted from a weight basis to a molar basis. Consequently, the composition according to triglyceride groups can also be calculated. To what extent the composition according to triglyceride groups changes on interesterification depends very much on the raw material. As mentioned above, cocoa butter shows dramatic changes in triglyceride groups and thus in physical properties on interesterification. On the other hand, the randomisation of a liquid oil, like sunflower seed oil, causes it to change only slightly. Redistribution of the saturated fatty acids will cause some H3 to be formed but at some 2.5 wt %, this is too little to provide any consistency. It will only cause the oil to become cloudy when cooled to refrigeration temperature. Similarly, the randomisation of tallow hardly causes any changes, either in melting point or in triglyceride groups. Consequently, tallow is not interesterified on an industrial scale.

BLUK122-Dijkstra

140

September 25, 2007

20:1

Trans Fatty Acids

Table 5.6

Fatty acid compositions of interesterification products and their starting materials. CN-H

PO-H

Fatty acid

wt %

wt %

wt %

mol %

wt %

mol %

wt %

mol %

C8:0 C10:0 C12:0 C14:0 C16:0 C18:0 H M L H3 H2 M HM2 +H2 L HML+M3 HL2 +M2 L ML2 L3

8 8 48 16 9 11

46 54

6.0 6.0 36.0 12.0 18.3 21.7

9.1 7.6 39.4 11.5 16.7 15.7 32.4 50.9 16.7 3.4 16.0 30.4 29.7 15.7 4.3 0.5

4.0 4.0 24.0 8.0 27.5 32.5

6.5 5.4 28.0 8.2 25.1 26.8 51.9 36.2 11.9 14.0 29.2 30.0 18.1 6.9 1.6 0.2

2.0 2.0 12.0 4.0 36.7 43.3

3.5 2.9 15.1 4.4 35.9 38.2 74.1 19.5 6.4 40.7 32.1 19.0 6.3 1.6 0.2 —

0.5 5.2 22.1 38.7 25.7 7.1 0.7

100

75 CN/25 PO

4.1 18.0 31.8 28.4 13.9 3.4 0.4

50 CN/50 PO

16.0 30.9 29.7 16.4 5.7 1.2 0.1

25 CN/75 PO

43.7 31.9 17.6 5.4 1.3 0.1 —

CN-H, fully hydrogenated coconut oil; PO-H, fully hydrogenated palm oil.

Large changes in triglyceride groups are observed when mixtures of oils, such as a lauric oil and a non-lauric oil, are interesterified. Lauric oils, like coconut oil and palm kernel oil, have high M3 and M2 L contents and if such oils are interesterified with products such as palm stearin having a high H3 content, the M3 and H3 contents in the reaction mixture are much reduced so that many more H2 M and HM2 triglycerides are formed. This has been illustrated in Table 5.6. The figures in this table have been arrived at as follows:

r r r r r r

First of all, the fatty acid compositions of fully hydrogenated coconut oil and palm oil are listed as wt %. Then the fatty acid compositions of their 75/25, 50/50 and 25/75 mixtures are calculated, again as wt %. The weight percent figures of these mixtures are then normalised to mol %. Because of the wide spread in the relative molecular masses of the fatty acids concerned, there is quite a difference between the weight percentages and the molar percentages as shown by the adjacent columns. The fatty acids of these mixtures are then grouped according to H, M and L by adding the molar percentages concerned. Now, the triglyceride composition of the mixtures after interesterification can be calculated on a molar basis, as has already been illustrated for soya bean oil in Section 5.4.1. Finally, the triglyceride group composition of the interesterification products can be reverted to a weight basis by estimating average relative molecular masses per group.

Table 5.7 highlights the dramatic changes in group composition on interesterification. It shows that a fully hydrogenated 50/50 mixture of palm oil and coconut oil contains just over half H3 -type triglycerides and will thus cause a poor mouthfeel. On interesterification, the H3 content drops to 16 wt % or less than a third. At the same time, there is an increase in the triglyceride groups H2 M, HM2 and H2 L, which provide some consistency without

BLUK122-Dijkstra

September 25, 2007

20:1

Controlling physical and chemical properties of fat blends Table 5.7

Changes in group composition on interesterification. 75 CN/25 PO

H3 H2 M HM2 +H2 L HML+M3 HL2 +M2 L ML2 L3

141

50 CN/50 PO

25 CN/75 PO

Before

After

Before

After

Before

After

25.1 3.9 16.6 29.0 19.3 5.3 0.5

4.1 18.0 31.8 28.4 13.9 3.4 0.4

50.3 2.6 11.1 19.3 12.7 3.6 0.4

16.0 30.9 29.7 16.4 5.7 1.2 0.1

75.1 1.3 5.5 9.7 6.4 1.8 0.2

43.7 31.9 17.6 5.4 1.3 0.1 —

causing a poor mouthfeel. Moreover and as illustrated by Table 5.7, shifting the balance between the lauric and the non-lauric component permits the emphasis between these two properties to be shifted. Increasing the lauric component at the expense of the non-lauric one leads to a decrease of the H3 content and an increase of the M3 and M2 L contents of the interesterification product. When used as a hardstock in margarine fat blends, these blends will have an improved mouthfeel but also have an increased saturated fatty acid content.

5.4.4

Other oil treatments

Amongst the other oil treatments, degumming and chemical neutralisation do not affect the triglyceride group distribution, but bleaching is another matter. Because bleaching earths are acid catalysts, they can also protonate double bonds, and this reaction, being reversible, can lead to both positional and geometrical isomerisation. Ney (1964) has studied both types of isomerisation and concludes that below 150◦ C, bleaching earth does not cause such isomers to be formed. During deodorisation and physical refining of edible oils, the temperature tends to be higher and indeed, isomerisation of the double bond can occur and interesterification has also been observed during these vacuum stripping processes. Physically refined palm oil was observed to have changed its crystallisation behaviour (Willems and Padley, 1985), and this change has been linked to interesterification. Similarly, cocoa butter that has been deodorised at 260–264◦ C will have undergone so much interesterification that it can no longer be used for chocolate manufacture (Stage, 1984). To what extent free fatty acids, present at the early stages of the physical refining process, affect/catalyse the interesterification has not been investigated. It is also not known to what extent the ester interchange takes place between separate triglyceride molecules (interesterification) or only within the triglyceride molecules (intraesterification). In the latter case, the triglyceride groups will of course not change. Ester interchange during vacuum steam stripping can be avoided by setting a maximum to the process temperature whereby the level of this maximum depends upon the sensitivity of the product. In this context, cocoa butter can be considered a highly sensitive product because small changes in the composition of its triglycerides have a considerable effect upon its crystallisation behaviour. Its deodorisation temperature is therefore limited to 150◦ C or perhaps 170◦ C. When palm oil is to be used as raw material for the production of confectionery fats, its temperature during physical refining can be limited to perhaps some 220◦ C. Since physical

BLUK122-Dijkstra

142

September 25, 2007

20:1

Trans Fatty Acids

refining is thermally more intensive than just deodorisation, some producers of palm-based confectionery fats prefer to neutralise the crude oil chemically and remove colour by adsorptive bleaching so that no heat bleaching is required, and a light deodorisation treatment will suffice to obtain a light-coloured and bland-tasting product. This product is of course considerably more expensive than physically refined palm oil. Consequently, a physical refining process comprising a short high-temperature stage in a counter-current packed column is now the preferred method of neutralisation. Selecting crude oil on the basis of its low free fatty acid content is also practised. Thermally induced isomerisation of double bonds has been found not to involve positional shifts but to be geometrical only (Ackman et al., 1974). The authors also conclude that amongst the common unsaturated fatty acids, linolenic acid is primarily responsible for the formation of trans isomers that predominantly comprise the cis-9,cis-12,trans-15 and the trans-9,cis-12,cis-15 isomers of linolenic acid; this was subsequently confirmed (Devinat et al., 1980; Grandgirard et al., 1984). In this respect, thermal isomerisation of double bonds differs strongly from the isomerisation of double bonds during hydrogenation. In this latter process, the monounsaturated oleic acid is the main fatty acid to isomerise and this isomerisation can also involve positional isomerisation. As only to be expected, the rate of thermally induced isomerisation of linolenic acid depends strongly upon the temperature. According to Wolff (1993), this rate is first order in linolenic acid and on the basis of his kinetic data, an activation energy of 148 kJ/mol has been worked out (Dijkstra, 2007). This value is quite close to the values of 146 kJ/mol (O’Keele et al., 1993), 144 kJ/mol (K¨ov´ari et al., 1997) and 136 kJ/mol (H´enon et al., 1997). This sizeable activation energy illustrates that the formation of trans isomers during vacuum steam stripping can easily be suppressed by slightly lowering the processing temperature. This will of course also lower the vapour pressure of the volatiles to be removed during the stripping process, and thereby increase the stripping medium requirement. Blowing more steam through the oil requires more time and thus causes the oil to be exposed to the processing temperature for a longer period of time. However, since the isomerisation depends more strongly upon the temperature than the vapour pressure does, fewer trans isomers result when the oil is deodorised at a lower temperature for a longer time. Lowering the temperature also has the advantage that it causes more tocopherols to be retained, and this may well contribute to the oxidative stability of the oil (Verleyen et al., 2002). It has the disadvantage of increased processing cost. In industrial practice, corn-germ oil that is deodorised in a tray-type deodoriser for 80 min at 240◦ C contains less than 0.5% trans fatty acids, which is below a purchase specification in force. Because of the higher reactivity of linolenic acid, soya bean oil and rape seed oil are preferably not heated above 230 or 240◦ C in a tray-type deodoriser, but when heated for a short time for treatment in a packed column, temperatures below 260◦ C do not cause the trans fatty acid content to be outside the specification of 1%. Isomerising some double bonds to their trans configuration will hardly affect the triglyceride group distribution, but it may cause a haze in liquid oils when these are stored in a refrigerator (Billek, 1992). The trans formation is more an aspect to be included in the linear programming of fat blends. Commercially available linear programming software also provides data on the cost of adhering to the specifications used to define the fat blends concerned. By including a range of the potential blend components with different properties (such as trans content) in the database, assuming realistic prices for these components and calculating the lowest cost blends for a range of blend specifications, an impression can be gained of what prescribing a low trans content is costing.

BLUK122-Dijkstra

September 25, 2007

20:1

Controlling physical and chemical properties of fat blends

143

A likely outcome of the above exercise is that lowering the trans content by a small proportion costs relatively little but reducing it to almost negligible levels causes a significant increase in the cost of the fat blend.

5.5

USING TRIGLYCERIDE GROUPS IN PRODUCT DEVELOPMENT

It has already been mentioned that triglyceride groups can indicate what mouthfeel a margarine fat blend will have. In addition, they can also be used for the calculation of appropriate fat blends by linear programming. The easiest way to go about this is by starting with existing fat blends that meet the requirements, work out their triglyceride groups composition, decide which limiting values are likely to be critical and quantify these values. Table 5.8 shows typical SFC profiles for various margarines to be discussed in more detail in Chapter 8. The SFC profile of these blends constitutes one of these requirements, but as will be explained in Chapter 9, these profiles hardly take the processing and post-hardening of the fat blends into account, which means that further parameters indicative of these properties may also be required. In addition, other fat blend properties such as the maximum linolenic acid content, the minimum linoleic acid content and/or the maximum trans isomer content can be added to the linear programming criteria. In fact, functions of component properties can also be included as criteria. It is, for instance, possible to define a criterion for stability. This should be related to the fatty acid chain-length heterogeneity of the triglycerides partaking in the crystallisation. Before illustrating what the triglyceride group concept can contribute to the design of fat blends in general and fat blends used in labelled products in particular, another question has to be discussed: what are the product property priorities?

r r

Saturated fatty acid content: Labels have to mention this product property and the ‘consumer’ generally regards these saturated fatty acids as bad. (The ‘consumer’ also dislikes ‘thickening agents’ since he or she regards these as a cause of obesity.) Label-friendly products should therefore aim at minimising the saturated fatty acid content. Saturated fatty acid types: Labels do not mention which saturated fatty acids are present in a product, thus not recognising differences between the various saturated fatty acids (Dijkstra, 2006). Stearic acid is readily dehydrogenated by the body to oleic acid, which is regarded as nutritionally innocuous. Longer chain saturated fatty acids are not absorbed Table 5.8

SFC for various margarine types.

Product type

10◦ C

20◦ C

25◦ C

30◦ C

35◦ C

Health margarine Soft table margarine Table margarine Wrapper margarine Cream margarine Cake margarine Croissant margarine Puff pastry margarine Hard puff pastry margarine

13–16 16–22 28–34 24–30 46–52 46–54 50–58 58–66 62–70

8–12 10–16 20–28 16–24 20–26 22–30 32–38 38–46 42–50

5–10 8–14 10–16 8–12 10–14 14–20 20–26 24–30 26–32

0–4 4–8 6–10 3–6 4–8 8–12 14–20 16–22 18–24

<1 <1 0–3 <1 <3 4–8 8–12 10–16 14–20

BLUK122-Dijkstra

144

r r

r

September 25, 2007

20:1

Trans Fatty Acids

by the body and therefore the poor image of saturated fatty acids must stem from C16 and below; however, not too much below, since really short-chain saturated fatty acids follow a different digestion route and should therefore not be lumped together with other saturated fatty acids. Accordingly, insight into the nutritional effect of the various saturated fatty acids conflicts with the marketing target of minimising saturates. Trans content: Again, the labelling requirements do not distinguish between monounsaturated trans acids and polyunsaturated trans acids, such as, for instance, conjugated linoleic acid (see Chapter 3). Again, only a total trans content is mentioned on the label and insight into the nutritional effect of the various trans fatty acids is lacking. Processing: As will be illustrated in Chapter 9, lowering the trans content of fat blends may slow down the fat crystallisation during processing because of the persistency of the polymorph formed initially in the scraped-surface heat exchanger. This causes final product properties to deteriorate by post-stiffening but the question to be answered is if this is a disadvantage to be remedied. Cost: In linear programming, additional constraints tend to raise the cost of the fat blend. Moreover, processing cost may increase since deodorisation at a lower temperature takes longer, and subsequent processing of these blends in scraped-surface heat exchangers may also necessitate a throughput reduction. Replacing the linolenic acid containing soya bean oil or rape seed oil by linolenic acid free oils is also likely to cost money. The question of which of the many possible constraints related to raw material selection and process conditions should be taken into consideration when formulating a product is a matter for commercial judgement guided by scientific information.

Answering the above and similar questions is outside the scope of the present work. Besides, it is more political than technical, let alone scientific. Accordingly, future nutritional insights that contravene current labelling regulations may well face an unhealthily long waiting time before being implemented. Producing trans-free margarines is nothing new. In 1969 a patent was published, mentioning Fondu and Willems as its inventors. It describes a hardstock that is obtained by interesterifying palm stearin with a lauric oil. Neither of these components was hydrogenated, and consequently, the interesterification product contains quite a large proportion of triglycerides (MLU, M2 U, LU2 etc.) that do not contribute to blend consistency but nevertheless introduce saturated fatty acids. This was improved by fully hydrogenating the lauric and non-lauric components (Delfosse, 1971; Graffelman, 1971; Ward, 1982). This choice of ingredients eliminated U-containing triglycerides. As a result, far less hardstock is required to achieve a set SFC and, as explained above and illustrated by Table 5.7, the mouthfeel of blends containing such hardstocks can, to some extent, be controlled by adjusting the ratio of the components to be interesterified. Instead of interesterifying hydrogenated components, it is also possible to interesterify nonhydrogenated components and hydrogenate the interesterification product; an identical final product will result. Even though the above hardstocks do not contain low-melting, U-containing triglycerides, they do contain triglycerides with only L and M fatty acids and these triglycerides do not contribute to product consistency, but do contribute to the saturates that have to be mentioned on the label. There are two ways of improving this situation. One way involves using a hardstock with a relatively high content of H3 triglycerides. Such a hardstock results from interesterifying somewhat less fully hydrogenated lauric fat with somewhat more fully hydrogenated

BLUK122-Dijkstra

September 25, 2007

20:1

Controlling physical and chemical properties of fat blends

145

non-lauric fat and then removing the excess H3 triglycerides by fractionation (Kattenberg and Verhagen, 1976). Another way would involve using a hardstock with a relative excess of low-melting triglycerides and removing those by fractionation. This method has the disadvantage that it requires solvent or multiple dry fractionation and is therefore more expensive. Producing a mid-fraction of an interesterification product and using this as hardstock for a fat blend with minimal saturated fatty acid content has also been claimed (Holemans et al., 1988). The by-products (stearin and second olein) can, to some extent, be recycled but a certain purge will be required to compensate for the difference in fatty acid compositions of the mid-fraction and the raw materials blend. Multiple fractionation and by-product disposal will all add to the cost of this hardstock, which is already quite high because of the hydrogenation and interesterification processes involved. In a more recent publication (Sahasranamam, 2004), the presence of hydrogenated fats is avoided by interesterifying a super palm stearin with a lauric fat stearin. The use of stearins minimises the unsaturated fatty acid content of the interesterification product and thereby the amount of triglyceride groups that do not contribute to the functionality of the hardstock. Accordingly, only a small amount is required as hardstock so that the ensuing margarine fat blend also has a low saturated fatty acid content.

REFERENCES Ackman, R.G., Hooper, S.N. & Hooper, D.L. (1974) Linolenic acid artifacts from the deodorization of oils. J Am Oil Chem Soc 51 (3), 42–49. Bailey, A.E. (1950) Melting and Solidification of Fats. Interscience, New York. Bailey, A.E. & Fisher, G.S. (1946) Modifications of vegetable oils. V: relative reactivities toward hydrogenation of the mono- di- and triethenoid acids in certain oils. Oil & Soap 23 (1), 14–18. Beyens, Y. & Dijkstra, A.J. (1983) Positional and triglyceride selectivity of hydrogenation of triglyceride oils. In: Fat Science 1983, Proceedings of 16th ISF congress, Budapest (ed. J. Holl´o). Akad´emiai Kiad´o, Budapest, pp. 425–432. Billek, G. (1992) Die Ver¨anderungen von Nahrungsfetten bei h¨oheren Temperaturen. Fett Wiss Technol 94, 161–172. Deffense, E.M.J. (1995) Dry multiple fractionation: trends in products and applications. Lipid Technol 7, 34–38. Delfosse, J.K.F. (1971) Margarine fat composition. British Patent 1 244 868, assigned to Unilever. Devinat, G., Scamaroni, L. & Naudet, M. (1980) Isom´erisation de l’acide linol´enique durant la d´esodorisation des huiles de colza et de soja. Rev fr des Corps Gras 27 (6), 283–287. Dijkstra, A.J. (1997) Hydrogenation revisited. Inform 8 (11), 1150–1158. Dijkstra, A.J. (2002) Hydrogenation and fractionation. In:Fats in Food Technology (ed. K.K. Rajah). Sheffield Academic Press, Sheffield, pp. 123–158. Dijkstra, A.J. (2005) No solution yet to the linolenic acid mystery. Inform 16 (5), 283. Dijkstra, A.J. (2006) Revisiting the formation of trans isomers during the partial hydrogenation of triglyceride oils. Eur J Lipid Sci Technol 108 (3), 249–264. Dijkstra, A.J. (2007) Vacuum stripping of oils and fats. In: The Lipid Handbook, 3rd edn (eds F.D. Gunstone, J.L. Harwood & A.J. Dijkstra). Taylor & Francis Group, LLC, Boca Raton, FL, pp. 236–253. Dominick, W.E., Nelson, D.W. & Mattil, K.F. (1953) Lard crystal modification. US Patent 2,625,484, assigned to Swift & Company. Fondu, M.P.V. & Willems, M.A.G. (1969) Margarine fat formulation. British Patent 1,245,539, assigned to Unilever. Graffelman, H.A. (1971) Margarine fat and process for preparing same. US Patent 3,617,308, assigned to Unilever. Grandgirard, A., S´eb´edio, J.-L. & Fleury, J. (1984) Geometrical isomerization of linolenic acid during heat treatment of vegetable oils. J Am Oil Chem Soc 61 (10), 1563–1568.

BLUK122-Dijkstra

146

September 25, 2007

20:1

Trans Fatty Acids

Gunstone, F.D. & Harwood, J.L. (2007) Occurrence and characteristion of oils and fats. In: The Lipid Handbook, 3rd edn (eds F.D. Gunstone, J.L. Harwood & A.J. Dijkstra). Taylor & Francis Group, LLC, Boca Raton, FL, pp. 37–142. Hawley, H.K. & Holman, G.W. (1956) Directed interesterification as a new processing tool for lard. J Am Oil Chem Soc 33 (1), 29–35. H´enon, G., Kem´eny, Z., Recseg, K., Zwobada, F. & K¨ov´ari, K. (1997) Degradation of -linolenic acid during heating. J Am Oil Chem Soc 74 (12), 1615–1617. Holemans, P.M.J., Schijf, R., van Putte, K.P.A.M. & de Man, T. (1988) Fat and edible emulsions with a high content of cis-polyunsaturated fatty acids. US Patent 4,791,000, assigned to Unilever. Kattenberg, H.R. & Verhagen, L.A.M. (1976) Margarine fat. US Patent 3,956,522, assigned to Unilever. K¨ov´ari, K., Denise, J., Zwobada, F. et al. (1997) Kinetic of trans isomer fatty acid formation during heating. Paper presented at the 22nd ISF World Congress, Kuala Lumpur. M`ege, H. (1869) Production de certains corps gras d’origine animale. French Patent 86 480. Ney, K.H. (1964) Einfluss der Bleicherde-Behandlung bei verschiedenen Temperaturen auf die Lage der Doppelbindungen von Fetts¨auren. Fette Seifen Anstrichmittel 66 (7), 512–517. Normann, W. (1903) Process for converting unsaturated fatty acids or their glycerides into saturated com¨ pounds. British Patent 1 515, assigned to Herforder Maschinenfett-und Olfabrik Leprince und Siveke. O’Keele, S.F., Wiley, V.A. & Wright, D. (1993) Effect of temperature on linolenic acid loss and 18:3 delta 9-cis, delta 12-cis, delta 15-trans formation in soybean oil. J Am Oil Chem Soc 70 (9), 915–917. Padley, F.B., Gunstone, F.D. & Harwood, J.L. (1994) Occurrence and characteristics of oils and fats. In: The Lipid Handbook, 2nd edn (eds F.D. Gunstone, J.L. Harwood & F.B. Padley). Chapman & Hall, London, pp. 47–223. Sahasranamam, U.R. (2004) Trans free hard structural fat for margarine blend and spreads. US Patent 6,808,737, assigned to Premium Vegetable Oils Berhad. Schijf, R., Trommelen, A.M. & Lansbergen, G.J.T. (1983) Margarine fat blend, and process for producing said fat blend. European Patent 0 089 082, assigned to Unilever. Stage, H. (1984) Ents¨auerung und Desodorierung von Speise¨olen und Fetten in ein-oder mehrstufig betriebenen Gegenstrom-Filmanlagen-Grundlagen und Wirtschaftlichkeit. Fette Seifen Anstrichmittel 86 (7), 255–264. van Alphen, J. (1964) De Chevreul a` M`ege Mouri`es. Ol´eagineux 19 (8–9), 525–528. Vander Wal, R.J. (1960) Calculation of the distribution of the saturated and unsaturated acyl groups in fats, from pancreatic lipase hydrolysis data. J Am Oil Chem Soc 37 (1), 18–20. Vander Wal, R.J. & Van Akkeren, L.A. (1951) Modified lard and process of producing same. US Patent 2,571,315, assigned to Armour and Company. Verleyen, T., Kamal-Eldin, A., Mozuraityte, R. et al. (2002) Oxidation at elevated temperatures: competition between -tocopherol and unsaturated triacylglycerols. Eur J Lipid Sci Technol 104 (4), 228–233. Ward, J. (1982) Edible fat product. US Patent 4,341,812, assigned to Nabisco Brands Inc. Willems, M.A.G. & Padley, F.B. (1985) Palm oil: quality requirements from a customer’s point of view. J Am Oil Chem Soc 62 (2), 454–459. Wolff, R.L. (1993) Heat-induced geometrical isomerization of -linolenic acid: effect of temperature and heating time on the appearance of individual isomers. J Am Oil Chem Soc 70 (4), 425–430.

BLUK122-Dijkstra

September 25, 2007

6

20:1

Trans isomer control in hydrogenation of edible oils

Annemarie Beers, Rob Ariaansz and Douglas Okonek

6.1

INTRODUCTION

Oils and fats derived from vegetable and animal sources are essential ingredients for a variety of food products. Naturally occurring oils are primarily triacylglycerols having long-carbonchain linkages interrupted in certain positions by double bonds. The carbon chain lengths, degree of unsaturation and the position of carbon–carbon double bonds dictate the physical properties of the triglycerides. In many cases, oils can be used in their natural state without processing other than refining, bleaching and deodorisation steps. In many other applications, the oils require additional modification for certain end uses.

6.1.1

Hydrogenation process

Catalytic hydrogenation is an important oil modification technique for altering the physical and chemical properties of the oils. Over the past century, catalytic hydrogenation has become one of the most versatile processes available to modify fats and oils. It is estimated that worldwide at least 6 million tonnes of oil is hydrogenated per annum. Since hydrogenated oils are often blended, a much larger quantity of oils and fats contains hydrogenated oils.

6.1.2

History of hydrogenation

The discovery of catalytic hydrogenation is generally attributed to Sabatier and Senderens at the end of the nineteenth century. Liquid-phase hydrogenation was reported in a patent issued to Normann (1903) and liquid-phase commercial processes were developed around 1905–1908 (Weber and Alsberg, 1934). In Europe, Joseph Crosfields & Sons in Warrington, UK, began industrial hydrogenation in 1906 (Kaufmann, 1939) and in the US, The Procter & Gamble Company began commercial production of hydrogenated cottonseed oil in 1911 (Buehler, 1989), signalling the beginning of the modern vegetable oil industry. Reviews of the early history of fats and oils hydrogenation have been published (Ellis, 1930; Hilditch, 1941).

6.1.3

Reasons for hydrogenation

Improving the oxidative stability of oils is one of the goals of catalytic hydrogenation. It has been shown that certain polyunsaturated components of natural fats and oils are prone to

BLUK122-Dijkstra

148

September 25, 2007

20:1

Trans Fatty Acids

oxidise in contact with air, especially when these oils are heated (e.g. in a frying application). Air oxidation of oils is undesirable for food applications since the resulting oxygen-containing compounds can ultimately impart an off-flavour or rancidity to the edible oil. Shelf-life, higher temperature usage and flavour properties can all be compromised if appreciable air oxidation has occurred. Selective hydrogenation to reduce the amount of polyunsaturation is a technique utilised to improve oil stability (Frankel, 1980; Mounts, 1980). Partial hydrogenation can be performed in such a way as to maintain the fluid properties of the oil. In many cases, a pourable, stable liquid oil is desired. Hydrogenation is also performed to increase the melting temperature of the oil (Latondress, 1980; Sipos and Szuhaj, 1996). End-usage convenience is usually the goal (e.g. spreadability, coating properties, shortening consistency, etc.). In the manufacturing of fats that are solid or semi-solid at room temperature, meeting melting point specifications is a primary goal. The physical consistency and melting properties of the fat (i.e. the solid fat profile) have become equally important. Each individual type of feedstock oil contains triglycerides with various combinations of carbon chain lengths and numbers of carbon–carbon double bonds. This complex mixture of triglyceride molecules determines the melting properties of the oil. Elimination of carbon–carbon double bonds present in natural oils via catalytic hydrogenation results in an increase of the melting temperature, the magnitude of the increase being dependent upon the degree of hydrogenation or double bond saturation and the selective nature of the hydrogenation process. The degree of geometric isomerisation also has a major influence on melting properties. Selectivity and degree of isomerisation are dictated by several factors, including reaction conditions (e.g. temperature, hydrogen pressure and agitation speed), feedstock characteristics, reactor configuration and choice of catalyst system.

6.2 6.2.1

ISOMERISATION Geometric and positional isomerisation

During partial hydrogenation, the carbon–carbon double bonds remaining in the oil molecules may also have reacted. Isomerisation of some of the non-hydrogenated double bonds occurs as a parallel reaction. This includes positional isomerisation in which the location of the unreacted double bond in the carbon chain shifts from the original position. Positional isomerisation causes only minor changes in the physical properties of the oil. Under certain circumstances, remaining double bonds can also conjugate. In geometric isomerisation, the stereoisomeric form of the double bond changes. Most naturally occurring double bonds are in the stereoisomeric cis form in which alkyl side chains bonded to adjacent double-bond carbon atoms extend spatially on the same side of the carbon–carbon double bond. During hydrogenation, the cis bonds can isomerise to the trans configuration. This conversion is accompanied by a significant increase of the melting temperature of the fatty acid moieties. Therefore, during hydrogenation, the number of positional and/or geometric isomers in the triglycerides increases in comparison with those in the starting oil and then decreases to very low numbers as double-bond saturation continues. Since different catalysts and different hydrogenation conditions result in different fatty acid compositions, selective hydrogenation provides a useful tool for controlling the properties of the fat.

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

6.2.2

149

Controlling isomerisation

Some control over the degree of geometric isomerisation is possible by the selection of appropriate reaction conditions similar to those used to control the selectivity of the hydrogenation of double bonds. The degree of isomerisation is influenced in a similar fashion to double-bond hydrogenation selectivity by reaction conditions (temperature, hydrogen pressure, agitation, etc.), feedstock characteristics, reactor configuration and choice of catalyst system. It is important to note that heterogeneous catalytic hydrogenation is accompanied by some degree of geometric isomerisation. Consumption of trans isomers has been the subject of several health studies and it has been concluded (Katan et al., 1995) that trans isomers can change the content of blood serum cholesterol in a way generally believed to be associated with the incidence of cardiovascular disease. However, the direct scientific evidence of the relationship of trans isomer consumption and cardiovascular health problems has been questioned (Ravnskov, 2000). Nevertheless, control of geometric isomerisation (trans isomer formation) has become a major focus of edible oil processors. Replacement of the melting functionality of trans isomers with non- or low-trans oil components in food oil formulations has also become a goal of the edible oil industry.

6.3 6.3.1

REACTION MECHANISM ‘Half-hydrogenated’ intermediate

It is generally accepted that the hydrogenation and isomerisation mechanism involves a set of reactions in which hydrogen atoms are added to and removed from the fatty acid moieties in the triglyceride oil molecule in a stepwise fashion, as described in the classical mechanism proposed by Horiuti and Polanyi (1934). A review of the mechanism of triglyceride oil hydrogenation has recently been published (Dijkstra, 2006). Initial steps in the mechanism include the dissolution of hydrogen in the liquid oil, adsorption of hydrogen and oil molecules containing carbon–carbon double bonds on the catalyst surface and dissociation of hydrogen molecules.

6.3.2

Saturation, positional and geometric isomerisation

The addition of a hydrogen atom to the carbon–carbon double bond (sometimes referred to as ‘half hydrogenation’) shown below creates an intermediate. H

H C

R

C R'

H

H C

+ H R

C

H

R'

This intermediate then undergoes subsequent reactions:

r

Free rotation around the carbon carbon bond axis followed by dissociation of a hydrogen atom back to the catalyst surface to re-form the carbon–carbon double bond. Because of the free rotation illustrated below, the re-formed double-bond molecule can be in the

BLUK122-Dijkstra

150

September 25, 2007

20:1

Trans Fatty Acids

cis or trans geometric configuration. The geometrically isomerised molecule (trans or cis configuration) can desorb from the catalyst surface back into the bulk of the oil. H

H C

C

R

H

H C

+ H R

R'

R'

H

C

H

C R

R'

C

H

H

H

+

H

H trans

Addition of a second hydrogen atom to either of the two intermediates illustrated above to create a saturated carbon carbon bond (saturation). Both reactions give rise to the same saturated moiety and both reactions are irreversible. The saturated molecule then desorbs from the catalyst surface back into the bulk of the oil. H

H C R

C

H

+ H H H

R' R R'

H C R

r

C

R

cis

r

R' C

C

C

C

R'

H H H

+ H

H

Dissociation of a hydrogen atom from a carbon atom adjacent to the site of the first hydrogen atom addition. As shown below, this re-forms the carbon–carbon double bond but in a different position on the carbon chain (positional isomerisation). The positionally isomerised molecule can then desorb from the catalyst surface back into the bulk of the oil. H

H C

R

C H2

C R'

H

H + H

C R

C H2

C

H H

R'

R

H

C C H CH R'

+

H

The various adsorption–desorption steps are reversible. The concentration of hydrogen in the bulk oil (and ultimately the concentration of hydrogen at the catalyst surface) can be the determining factor in dictating the relative extents of saturation, positional and geometric isomerisation. The above description is simplified since the actual hydrogenation reaction is more complex and involves multiple simultaneous adsorption–desorption steps. The above reaction equations provide simplified representations of geometric isomerisation, saturation and positional isomerisation respectively.

6.4 6.4.1

INDUSTRIAL HYDROGENATION Batch process

The industrial hydrogenation of triglyceride feedstocks for edible products is conventionally carried out in the slurry phase (i.e. using a heterogeneous catalyst in a tank reactor). With only a few rare exceptions, this reaction is typically carried out as a batch process. The catalyst is suspended in the feedstock in powder form, hence the term ‘slurry reactor’.

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

151

PIC Internal H2 recirculation Hollow shaft

Self-inducing turbine

Primary disperser

H2

Fig. 6.1

Dead-end tank reactor.

6.4.2

Reactor types and features

Two types of slurry reactors are used industrially. The most widely used is the agitated tank type, which is generally referred to as a ‘dead-end’ reactor. The hydrogen is typically introduced at the bottom of the reactor in the form of tiny bubbles and dispersed into the feedstock by means of a centrally mounted agitator. In the other type of reactor the reaction mixture is pumped around in an external loop (hence the name ‘loop reactor’). At one point in the loop, the mixture passes at high velocity through a high shear area, such as a nozzle at high velocity, creating a more intense mixing between hydrogen and feedstock than that in the dead-end reactor (Leuteritz, 1969). Figures 6.1 and 6.2 illustrate these reactors. The reaction will proceed at an appreciable rate only after it has reached a certain kick-off temperature. Therefore, the reactor needs to be equipped with a heat exchanger of sufficient capacity that will allow the oil to be heated to the desired reaction temperature (Hastert, 1996). As the reaction is exothermic, the reactor should also have the capability to maintain the temperature at a desired level by cooling when necessary.

6.4.3

Reaction parameters

6.4.3.1 Catalyst type and features Nickel is the most frequently used catalytic metal. To provide a large catalytic surface, the nickel is deposited on a high-surface-area support, commonly a silica, alumina or

BLUK122-Dijkstra

152

September 25, 2007

20:1

Trans Fatty Acids

Hydrogen

Reaction mixer Reaction heat exchanger

Reaction autoclave

Reactants with catalyst suspension

Reaction pump Fig. 6.2

Loop Venturi reactor.

60 50

Fish oil Soya bean oil Palm oil

% trans

40 30 20 10 0 150

100 IV

Fig. 6.3

Trans isomer formation in three different oils.

50

0

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

153

silica–alumina. The support is a fine powder with a particle size of about 2–10 m. The high porosity of the support allows easy transport of the molecules from the bulk phase to the nickel and back to the bulk oil. Nickel possesses catalytic properties only in the reduced (metallic) state. Therefore, suppliers of nickel catalysts provide products in the reduced state in a form ready for use. High surface area reduced nickel is pyrophoric when exposed to air. To protect the catalyst from air oxidation, it is suspended by the manufacturer in a high melting fat, e.g. fully hardened palm or soya bean oil. The suspension is cooled and solidified in the form of pellets (‘droplets’). Since the catalyst is a powder, it must be removed after the hydrogenation reaction by filtration, carried out in an external filter. 6.4.3.2 Catalyst reuse Some hydrogenators prefer to reuse the catalyst. However, during subsequent uses, the catalyst may absorb impurities that influence its performance in terms of selectivity and trans isomer formation. For this reason, an increasing number of companies prefer single use whereby the minimum amount of catalyst is used. After use, the spent catalyst is sent to specialised companies that reclaim the nickel from the catalyst.

6.4.4

Influence of feedstock on trans

In principle, hydrogenation is the prerequisite for the formation of trans isomers. Therefore the rate of formation of trans isomers can best be described as a function of the progress of the hydrogenation reaction, i.e. as a function of the change in iodine value (IV). During hydrogenation, an increasing percentage of the unreacted cis double bonds will be converted into trans isomers until the cis,trans equilibrium has been reached. At the same time, the number of double bonds will be decreasing with decreasing IV. Obviously, at IV = 0, there will be no trans isomers left. A graphic representation of the presence of trans fatty acids in oil as a function of the IV assumes the shape of an almost parabolic curve (see Fig. 6.3) (BASF Catalysts LLC, unpublished observations). The height of the maximum depends, amongst other things, on the level of unsaturation of the original feedstock. As illustrated in Fig. 6.3, the higher the IV at the outset, the higher the level of trans fatty acids at any point in the course of the hydrogenation process. A number of naturally occurring impurities in the feedstock can interfere with the hydrogenation reaction by altering the properties of the catalyst. This has consequences for preferential selectivity as well as for the level of trans isomers at the desired end point of the reaction. Two common catalyst poisons are sulphur and phosphorus (Patterson, 1983). 6.4.4.1 Sulphur compounds Sulphur compounds occur in several vegetable oils, as well as in most oils and fats from animal sources. Nickel sulphide formation significantly reduces the activity of the catalyst. It also affects the Ni-catalysed dissociation of hydrogen molecules that results in a much lower concentration of hydrogen atoms on the surface of the nickel catalyst crystallites. The consecutive saturation reaction, involving the addition of a second hydrogen atom, is dependent on both the reduced concentration of hydrogen atoms and the reduced concentration of the half-hydrogenated intermediate. The rate of saturation is therefore reduced to a greater extent than the rate of formation of the intermediate and, consequently, than the rate of isomerisation. This results in an increase of the formation of trans isomers. Trans isomers have higher melting points than the corresponding cis isomers. Catalysts that are purposely sulphided

BLUK122-Dijkstra

154

September 25, 2007

20:1

Trans Fatty Acids 70 0.04% Ni 0.06% Ni

60

% trans

50 40 30 20 10 0 195

Fig. 6.4

175

155

135

115

95 IV

75

55

35

15

0

Hydrogenation of fish oil – effect of catalyst concentration.

by the manufacturer are therefore used to produce ‘steep melting fats’, such as cocoa butter replacers for confectionery applications. If the level of trans isomers becomes too high due to sulphur poisoning, there are several options to reduce trans:

r r r

r

r

Improve oil purification by better bleaching (more active clay, higher dosage) or even by introducing a deodorisation step to remove catalyst poisons prior to hydrogenation. Do not reuse the catalyst. Sulphur adsorbs irreversibly on the catalyst. The catalyst will get more severely poisoned each time it is reused. Use optimal reaction conditions directed towards lower trans. Increase of the catalyst concentration will have the effect of ‘diluting the sulphur’ distributed over the nickel. Therefore, the trans fatty acid yield will drop (see Fig. 6.4) (BASF Catalysts LLC, unpublished observations) with increasing catalyst dosage (illustrated for fish oil that contains sulphur) until the increasing hydrogen demand will become the dominant factor, resulting in an increase in trans fatty acids. Use a high dispersion catalyst (see Fig. 6.5 illustrating hydrogenation of fish oil-containing sulphur). A high surface area of the catalyst will lead to a lower percentage of poisoned nickel surface. Or stated differently, a larger fraction of the nickel surface will give a normal reaction with normal trans levels. The result will still be higher trans than without the sulphur but lower than with a low dispersion catalyst. Apply a two-stage catalyst dosage technique. Even when a catalyst is so severely poisoned that no residual activity can be observed, the catalyst can still continue to adsorb sulphur. Two-stage dosage enables the first charge of catalyst to adsorb most of the sulphur to well beyond the point of extinction of its hydrogenation activity. The second dose can then provide a surface for hydrogenation with little interference from the sulphur.

6.4.4.2

Phosphorus compounds

Another major poison for nickel catalysts is phosphorus-containing compounds. Phosphorus occurs in the form of phosphatides, predominantly in bean and seed oils. If the non-hydratable phosphatides are not sufficiently removed in refining (to maximum 2–5 ppm phosphorus),

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

155

Low Ni surface area High Ni surface area

70 60

% trans

50 40 30 20 10 0 195

175

155

135

115

95

75

55

35

15

0

IV Fig. 6.5

Hydrogenation of fish oil – effect of catalyst metal surface area.

they would primarily affect the preferential selectivity of the catalyst. An extreme example (see Fig. 6.6) of the effect of a phosphorus compound, purposely added in the form of lecithin, has been reported (Ottesen and Jensen, 1980). It may be surprising that the trans selectivity is reduced as well (see Fig. 6.7). However, with the theory about the selectivity in mind, an explanation can be offered. Trans isomers are formed only when a double bond adsorbs on the nickel surface, half hydrogenates so that

80 C18:0 C18:1

70

C18:2

60

% fatty acid

50

40

30

20

10

0 0

Fig. 6.6

4 ppm P

Effect of phosphorus on soya bean oil hydrogenation.

8

BLUK122-Dijkstra

156

September 25, 2007

20:1

Trans Fatty Acids 60

50

% trans

40

30

20

10

0 0

Fig. 6.7

4 ppm P

8

Trans isomers versus ppm phosphorus.

bond rotation can and does occur and finally desorbs without being hydrogenated. Four types of molecules can be identified as contributing to the IV:

r r r r

Those that have never entered the catalyst pores. Those that have entered the pores and been fully saturated. Those that have entered the catalyst pores but remained unchanged. Those that have entered the catalyst pores and in which at least one double bond has been converted into a trans isomer. Some of these molecules will have been partially hydrogenated in the process.

A contribution to the IV reduction comes only from the second and the last category. At a given IV, a catalyst poisoned with phosphatides will result in a product with considerably higher levels of fully saturated fatty acids as well as higher levels of relatively unsaturated fatty acids. At the same IV, in the case of the poisoned catalyst, the percentage of molecules in the first two categories must be higher than that with an unpoisoned catalyst, at the expense of molecules of the last three categories. Hence at the same IV, the level of trans isomers is reduced when using the poisoned catalyst. In the case of phosphatides, catalyst reuse magnifies the effects. To summarise, sulphur deactivates part of the metal surface due to inhibition of hydrogen adsorption, causing saturation to slow down to a greater extent than isomerisation, eventually leading to an increase in trans isomer content. Non-hydratable phosphatides, on the other hand, tend to block some of the catalyst pores, primarily resulting in a reduction of the preferential selectivity with a reduction of the trans isomer content as a secondary consequence.

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

6.4.5 6.4.5.1

157

Influence of reaction conditions on trans Reaction pressure

To explain the effect of hydrogen pressure, it is important to focus on the hydrogen concentration on the nickel surface and revisit the mechanism of hydrogenation. The hydrogen gas bubbles are introduced into the autoclave and dissolved into the oil. By mixing and diffusion, the dissolved hydrogen will travel through the oil to the catalyst particle. At sufficiently high pressure, hydrogen can be transported from the gas phase to the catalyst to nearly completely cover the Ni surface. In a relatively large number of cases the half-hydrogenated intermediate will find sufficient hydrogen atoms chemisorbed on the nickel surface in the immediate vicinity to react with another hydrogen atom at the carbon atom adjacent to the one to which the first hydrogen atom was added. This is likely to happen before the molecule has had a chance to rotate around the C C bond that was previously a C C double bond. Consequently, fewer trans isomers will be formed under high-pressure conditions. Alternatively, at low pressures leading to a hydrogen-deficient Ni surface, the halfhydrogenated intermediate will not always be in proximity to another hydrogen atom. In that situation, a higher percentage of the intermediates can rotate 180◦ around the C C single-bond axis and potentially form a trans isomer that desorbs back into the bulk of the oil. With a low concentration of hydrogen on the nickel surface, there is less likelihood that during the same encounter with the catalyst particle, a second double bond in the same molecule is hydrogenated in addition to the first one. Therefore, at low hydrogen pressure, more molecules will leave the catalyst with only one double bond hydrogenated than at a high pressure. Similarly, at a high pressure, the possibility of a second, third or even more hydrogen additions is more likely. Hence, at high pressure, more double-bond saturation is likely. To summarise, preferential selectivity (the selectivity controlling hydrogenation of triunsaturated vs diunsaturated vs monounsaturated fatty acids) is directionally inversely proportional to the pressure. Another explanation has been discussed (Dijkstra, 1997, 2002a). It is suggested that the preferential selectivity deteriorates as the pressure increases due to the fact that the reaction rate of monoene saturation has a higher order in hydrogen than the saturation of a double bond in a polyene. The reduction of trans can then be seen as a secondary effect of the lower preferential selectivity (see also Section 6.4.4.2).

6.4.5.2 Reaction temperature Increasing the reaction temperature affects the reaction kinetics and increases reaction rate constants. In large industrial reactors (e.g. 5–40 tonnes per batch), the supply of hydrogen is constrained by some level of mass transport limitation due to poor mixing relative to the demand for hydrogen dictated by the reaction chemistry. Due to the faster reaction rate at a higher temperature, the concentration of hydrogen on the surface of the catalyst will decrease as the temperature increases. As a result, the level of trans isomers as well as the preferential selectivity will increase along with the temperature. When the concentration of the catalyst is increased, the demand for hydrogen in the reactor will also increase, leading to a further depletion of hydrogen on the catalyst surface. Similarly

BLUK122-Dijkstra

158

September 25, 2007

20:1

Trans Fatty Acids

Table 6.1

Effect of reaction conditions on hydrogenation. Effect on H2 supply

Increase of Pressure Agitation Catalyst concentration Temperature Catalyst activity Degree of oil unsaturation

H2 concentration Preferential Cis–trans on Ni selectivity isomerisation + + − − − −

Larger supply Larger supply Larger demand Larger demand Larger demand Larger demand

− − + + + +

− − + + + +

to the situation described above, the formation of trans isomers and the preferential selectivity will increase as the catalyst concentration is increased. The effect of the reaction conditions on the hydrogenation of soya bean oil is illustrated by Table 6.1 and Figs. 6.8–6.10 (BASF Catalysts LLC, unpublished observations).

6.4.6

Influence of catalyst on trans

The differences in preferential selectivity referred to earlier have indirect consequences on the percentage of trans isomers in the product. If hydrogenation to a specified melting point is desired, a ‘selective’ catalyst would require achieving a lower IV than a ‘non-selective’ catalyst because of less saturate formation (and less contribution to an increasing melting point). Under selective reaction conditions, differences between nickel catalysts’ ability to form trans isomers are very small. Differences become significant if ‘non-selective’ (high-pressure, low-temperature) conditions are selected. This is illustrated in Figs. 6.9 and 6.10 (BASF Catalysts LLC, unpublished observations), showing the hydrogenation of soya bean oil with

40 35

Composition (wt%)

30 25 Saturates (wt%) 20

Trans isomers (wt%)

15 10 5 0 0

20

40

60

Hydrogen pressure (bar) Fig. 6.8 Effect of hydrogen pressure on the selectivity of a Ni catalyst in the hydrogenation of soya bean oil to 70 IV at 140◦ C.

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

159

50 45 40

% trans

35 30 25

Selective catalyst

20

Unselective catalyst

15 10 5 0 130

110

90

70

50

30

IV Fig. 6.9

Trans versus IV under selective conditions; comparison of a selective and non-selective catalyst.

selective and non-selective nickel catalysts under very different conditions. The catalysts differ in porosity. At the same IV, non-selective catalysts will produce lower levels of trans isomers than selective catalysts when used under non-selective conditions. The effect is similar to that of phosphatides.

30

25

% trans

20 Selective catalyst

15

Unselective catalyst

10

5

0 130

110

90

70

50

30

IV Fig. 6.10 catalyst.

Trans versus IV under non-selective conditions; comparison of a selective and non-selective

BLUK122-Dijkstra

160

September 25, 2007

20:1

Trans Fatty Acids

In conclusion, the differences in trans isomer formation between various nickel catalysts are only the consequence of differences in preferential selectivity and degree of hydrogenation.

6.4.7

Influence of reactor design on trans

The mixing intensity in an industrial autoclave (e.g. the number of revolutions per minute of the agitator in a ‘dead-end’ reactor) is usually fixed. This means that the degree of mixing of hydrogen and oil should be regarded as an important but uncontrollable parameter in the reaction. Generally, the hydrogen supply to the catalyst does not satisfy the chemical demand (i.e. the consumption of hydrogen on and inside the catalyst particles). This is particularly true at the onset of the reaction. The rate of hydrogen uptake by the oil rH (mol/m3 s) can be described by the equation (Bern et al., 1976; Chen et al., 1981; Koetsier, 1997): rH = kL · a · (cH0 − cH ) kL liquid-side mass-transfer coefficient (m/s) in the stagnant film of liquid surrounding the bubbles; a specific interfacial area (m2 /m3oil ), separating the oil from the gas in the bubbles and the head space of the autoclave. ‘a’ depends primarily on reactor design; cH0 hydrogen concentration (mol/m3 ) in the gas–oil interface (i.e. solubility of hydrogen in oil under the given reaction conditions); cH hydrogen concentration (mol/m3 ) in the bulk oil. In practice, large differences can occur between kL · a values (s−1 ) of industrial reactors. It is estimated that commercial dead-end autoclaves may have kL · a values ranging from about 0.05 to >0.5 (s−1 ). Recent developments in agitator design (e.g. with internal gas circulation through the hollow agitator shaft) have further improved gas supply to the oil. Laboratory autoclaves are known to have kL · a values of up to 2.0 (s−1 ) (Bern et al., 1976). Each reactor will have a maximum rate of hydrogen uptake that can be defined as: rH (max.) = kL · a · cH0 This explains why the reaction may proceed differently in different reactor systems. Consequently, observed catalyst activity (in terms of IV drop per minute at a given catalyst concentration), selectivity and cis–trans isomerisation will differ from reactor to reactor. Reactors with efficient mixing will tend to give lower preferential selectivity and yield less trans isomers than reactors with poor mixing.

6.4.8

Trans isomer control

6.4.8.1 High trans isomer formation Trans isomer formation in conventional hydrogenation can be maximised by using a sulphurpromoted nickel catalyst. This catalyst can be made in situ, for instance, by adding flowers of sulphur to a slurry of conventional nickel catalyst and oil. Another possibility is to reuse a catalyst that has previously been used in a feedstock that contained sulphur. The

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

161

great disadvantage of this latter method is the lack of reproducibility. High trans applications are usually the first step in the production of sophisticated speciality fats for the bakery and confectionery industry. Variability of catalyst performance is highly undesirable since the functionality of these products depends upon a reproducible level of trans fatty acids. The best alternative is to use a pre-sulphided catalyst from a catalyst supplier. These types of products are available from both BASF (Nysosel 210) and Johnson Matthey (Pricat 9908). Process conditions have much less influence on the hydrogenation with these catalyst systems compared to a conventional nickel catalyst. The deficiency of hydrogen on the surface of the catalyst is created by the sulphur rather than the process conditions, and this effect is less influenced by supplying more or less hydrogen to the catalyst particle. The reaction is approximately ten to twenty times slower than with a conventional catalyst at equal catalyst dosages. This means that five to ten times more catalyst must be added to obtain a reasonable reaction rate. Since the rate of the reaction is determined by the transport of hydrogen to the catalyst particle, higher pressures (0.2–0.4 MPa) and temperatures (180–200◦ C) are generally used (Okonek, 1987). The chemisorption of sulphur is generally regarded as an irreversible process. In catalyst reuse, sulphur originating from the feedstock will therefore accumulate during subsequent reuses, causing an increase of trans isomers, which is usually not desirable. 6.4.8.2 Minimising trans isomer formation Minimising trans fatty acids using conventional catalyst systems is much more of a challenge. First, process conditions must be considered and chosen such that the nickel surface is well saturated with hydrogen. In conventional processes, a reaction initiation temperature of 110– 140◦ C is typically used. The lower starting temperature limitation is the result of a low reaction rate of the catalyst at low reaction temperatures combined with the cooling capabilities of the reactor. To achieve full benefits from low-temperature hydrogenation, the catalyst must be of maximum activity and taken from a previously unopened container so that prior air exposure is minimised. A higher pressure not only lowers trans but also partly compensates for the disadvantage of a low starting temperature in terms of reaction rate. In practice, most conventional industrial reactors are limited to 0.3–0.5 MPa maximum hydrogen pressure. In recent years, reactors have been built that are capable of up to 2.0–2.5 MPa. In most reactors the rate of agitation cannot be changed, but a limited number have the capability of increasing the rotation speed of the agitator. If designed properly, this will improve transport of hydrogen from the gas phase to the catalyst. Ideally, catalyst concentration should be minimised to allow maximum hydrogen contact with the catalyst particles. However, if the feedstock contains sulphur, an increased catalyst concentration may reduce the formation of trans fatty acids. The choice of catalyst can also be of significant importance. If the linolenic and linoleic acid selectivities are not critical, a less selective catalyst will generally produce less trans. Finally, the hydrogenation procedure itself influences trans isomer formation. As previously discussed, catalyst reuse can influence the trans fatty acid level. Some hydrogenators that are concerned with the selectivity under trans-limiting conditions use two-stage hydrogenation, with trans-suppressing conditions in the first stage. In the second stage (e.g. after 75% reaction completion), the temperature is increased and the pressure is reduced to remove polyunsaturates and limit the formation of saturates.

BLUK122-Dijkstra

162

September 25, 2007

20:1

Trans Fatty Acids

6.5

NEW DEVELOPMENTS IN LOW TRANS HYDROGENATION

With the recent emphasis on trans isomer content in edible oils, studies have been made reporting improved trans isomer control during hydrogenation. Some of these reports will be reviewed and discussed.

6.5.1

Alternative reaction conditions

As discussed in the previous section, trans isomer formation can be controlled by changing reaction conditions, like temperature, pressure and mixing efficiency. Reaction temperatures of 140–200◦ C and hydrogen pressures of maximum 0.5 MPa are typically used. With conventional nickel edible oil hydrogenation catalysts, a relatively high amount of trans (up to 45% depending upon final IV) can be formed, using conventional conditions. 6.5.1.1

Temperature

Temperature is the most important factor in trans isomer control. Lowering the temperature below the conventional levels of 140◦ C and higher results in a significant trans isomer reduction. Figure 6.11 shows the selectivities with respect to trans isomers, saturates and linolenic acid (18:3) plotted against reaction temperature for lightly hydrogenated soya bean oil (105 IV). These studies used a specially developed low-trans nickel catalyst (Beers, 2006) at below typical hydrogenation temperatures (40–100◦ C). Lowering the temperature results in lower trans isomer fatty acid (TFA) formation while saturate levels increase. At a hydrogenation temperature of 40–60◦ C, a TFA level of 6% is achieved with a nickel catalyst. The 18:3 18 16

Selectivity (%)

14 12 Trans-isomer formation (wt%) Saturate formation (%) Linolenic acid (%)

10 8 6 4 2 0 0

50

100

150

200

Temperature (°C) Fig. 6.11 Effect of the hydrogenation temperature on the selectivity of a Ni catalyst in the hydrogenation of soya bean oil to 105 IV at 2–4 bars hydrogen pressure.

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

163

levels (known to affect the oxidative stability of the oil product) are hardly influenced by the reaction temperature. However, the use of lower temperatures decreases the catalytic activity. It is therefore essential to use a catalyst that is capable of saturating the double bonds at these lower non-conventional temperatures. These types of catalysts will be described in the following sections on catalysts. Conventional hydrogenation plants are configured for typical reaction initiation temperatures in the 110–150◦ C range. Performing the hydrogenation reaction at lower temperatures means that the heat-exchange capacity of the existing edible oil plants must be capable of sufficient heat removal from the exothermic reaction to ensure that lower reaction temperatures are maintained. 6.5.1.2

Hydrogen pressure

Increasing the hydrogen pressure promotes the saturation of the catalyst surface with hydrogen and suppresses the isomerisation reaction (cis–trans isomers). Figure 6.8 shows the effect of hydrogen pressure on the formation of saturates and trans isomers for soya bean oil hydrogenation at 140◦ C for a product with an IV of 70, using a standard Ni catalyst. At a typical (non-selective) hydrogenation pressure of 0.2 MPa, about 30–40% of TFA is formed and saturate levels (18:0) of about 15% are observed. At increasing hydrogen pressure, less TFA formation is observed, but at the expense of higher saturate levels which leads to an oil composition with higher levels of solid fats. To achieve a low TFA level of less than 10%, atypically high pressures need to be applied (>6 MPa). Most existing edible-oil–processing plants have a maximum pressure rating of approximately less than 0.5 MPa. As a consequence, higher pressure hydrogenation does not appear to be a viable option with existing equipment. 6.5.1.3

Mixing efficiency

Another option for increasing the hydrogen transport from the oil to the catalyst surface to suppress trans isomer formation and promote saturation of double bonds is to consider the mixing efficiency of the reactor system. If mixing is not efficient, transport of hydrogen through the oil to the catalyst surface may be slow. This situation will lead to a lower hydrogen concentration on the catalyst surface and higher trans isomer formation than desired. Industrial-scale hydrogenation plants have some level of hydrogen mass transport limitation. Increasing the agitation speed or performing the hydrogenation in a high-efficiency reactor system, such as a loop reactor, may change selectivity in the desired direction. Laboratory reactors have generally been designed for optimal mixing efficiency. It is often difficult to compare the selectivity results of a laboratory reactor with an industrial hydrogenation plant of lower mixing efficiency. Figures 6.12 and 6.13 illustrate the effects of reaction temperature and pressure on trans isomer, saturate and linolenic content in hydrogenated oil.

6.5.2

Alternative hydrogenation processes

In the edible oil industry, the most common type of hydrogenation reactor has been the batch slurry reactor. The results of hydrogenation reactions can be altered by using different types of reactors in both batch and continuous mode. Several additional new reactors and processes in which the option of increasing the activity and/or improving selectivity towards mono- and cis unsaturated fats have been studied. Examples include the use of solvents (in particular, supercritical solvents), ultrasonic devices, membrane and monolithic reactors.

BLUK122-Dijkstra

164

September 25, 2007

20:1

Trans Fatty Acids

Additional trans-isomer formation (%)

50 45 40 35 Ni (60–140°C) Ni (175–200°C) Pd Pt Pt modified

30 25 20 15 10 5 0 0

5

10

15

20

25

Additional saturates (%) Fig. 6.12 Trans isomer against saturate formation for the hydrogenation of soya bean oil to 70 IV (solid product) at hydrogen pressures of maximum 0.5 MPa with various catalysts and at various temperatures (<200◦ C).

18

Additional trans-isomers (%)

16 14 12 Ni 175°C Ni 100°C Ni <100°C Pt 50°C

10 8 6 4 2 0 0

2

4

6

8

10

Additional saturates (%) Fig. 6.13 Trans isomer against saturate formation for the hydrogenation of soya bean oil to 105– 110 IV (liquid product) at hydrogen pressures of maximum 0.5 MPa with various catalysts and at various temperatures (<200◦ C).

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

6.5.2.1

165

Fixed-bed reactors

A fixed-bed reactor can be defined as a vessel that contains a stationary, large particle catalyst (the ‘fixed bed’) through which feedstock and hydrogen pass. The reactants are converted into the reaction product while passing through the catalyst bed. This is a continuous process that is operated until product specifications are no longer met or the catalyst is exhausted and requires replacement. Process parameters can be adjusted to make the product meet the specifications. These parameters include temperature, pressure, feedstock flow rate (liquid hourly space velocity) and hydrogen flow rate (gas hourly space velocity). Fixed-bed processes in general have certain advantages and disadvantages when compared to conventional batch processes. The advantages include:

r r r r r

elimination of the need for a catalyst filtration step and catalyst filter; suitability for large-volume operation; ease of automation; low operating cost and constant product quality.

Disadvantages include:

r r r r r

not suitable for small-volume operation; not suitable for fluctuations in product demand. Generally, the design does not allow the rate of throughput to be reduced too much below the design capacity, without affecting the product properties. Also, operating a fixed bed intermittently at times of reduced product demand is generally not economically feasible; not suitable for frequent product-type changes due to product mixing during changeovers; not suitable for variable feedstock quality and large IV drop applications (like edible oil hydrogenation) will cause a bed temperature gradient and potential temperature-control issues.

When soya bean oil is hydrogenated in fixed-catalyst-bed (e.g. trickle-phase) reactors, both the isomerisation selectivity (% TFA/IV) and the linolenate and linoleate selectivities have been shown be lower than for hydrogenations performed in a comparable capacity batch reactor. When conventional, commercially available, fixed-bed catalysts are used, the lower selectivities are likely to result from longer residence times of triglyceride molecules in the catalyst pores (Boger et al., 2004; Edvardsson and Irandoust, 1994). 6.5.2.2

Membranes

Novel catalytically active, porous membranes have also been developed for the hydrogenation of edible oil in a membrane reactor. In one study (Veldsink, 2001), the catalysts were prepared by impregnating a tubular porous membrane with palladium on one side to provide a reactive interface between gaseous hydrogen and oil. The hydrogen was introduced from the opposite, inactive side of the tube to the activated palladium surface to contact the oil. The ratio of catalyst active site to oil was found to be low, and the catalyst was easily deactivated. In another study (Ilinitch et al., 2001), a palladium/polymeric membrane was tested for edible oil hydrogenation and a decrease in trans isomers was reported at comparable hydrogenation levels when compared to a conventional slurry reaction with palladium/carbon as the catalyst.

BLUK122-Dijkstra

166

September 25, 2007

20:1

Trans Fatty Acids

The improvement of the selectivity is attributed to the increased mass transfer in the pores of the membrane. In comparison with the above results, the work of Fritsch and Bengtson (2006) showed relatively high amounts of trans isomers with palladium (35% TFA) and platinum membranes (22–25% TFA) for sunflower oil hydrogenation. 6.5.2.3

Monoliths

The use of monolithic reactors in edible oil hydrogenation has been studied because of the higher mass-transfer characteristics of these reactor systems, as demonstrated for other chemical reactions. When comparing the performance of a palladium-coated monolith with a powdered palladium catalyst used in a slurry system, hydrogenated soya bean oil product compositions were similar at 90 IV and lower trans isomer formation was observed for the monolith. These trans isomer levels were observed to increase for both monolith (32.5% TFA) and slurry reactor (40% TFA) when soya bean oil was hydrogenated to approximately 80 IV (Boger et al., 2004). 6.5.2.4

Use of solvents and supercritical systems

Various studies can be found in the literature in which the oil to be hydrogenated was dissolved in a solvent. The purpose of the solvent is to enhance the solubility of the hydrogen, resulting in an improved coverage of the catalytic surface and a lower formation of trans isomers. Several potential disadvantages of such a solvent process can be identified:

r r r r r

The solvent could be a highly flammable organic liquid (e.g. a short-chain hydrocarbon), requiring special equipment and permits for use on a commercial scale. The higher volume of oil plus solvent requires a larger volume of the reactor equipment to achieve a similar oil throughput to that of a non-solvent system. Due to the volatility of the solvent, the system pressure requirements are higher than in conventional hydrogenation. Following hydrogenation, the solvent must be removed quantitatively from the oil. Economics dictate that the solvent should be recycled and solvent losses should be minimised.

In another approach using the solvent concept, several investigators worked under supercritical conditions. Under these conditions, with very specific combinations of pressure, temperature and oil/solvent ratio, the constituents of the reaction mixture are completely miscible. It is claimed to achieve near-complete removal of limitations to the hydrogen transfer to the surface of the catalyst. Under certain conditions this may result in hydrogenation rates that are up to 1000 times the rates observed in conventional hydrogenation. This could potentially result in more efficient catalyst usage (lower consumption) and it could obviate the need for a larger reactor. The efficiency advantage applies only if the optimal conditions allow the processor to benefit from the high reaction rates. Additional disadvantages of supercritical systems compared to other hydrocarbon solvent hydrogenation systems are that the pressures are even higher (e.g. 4–10 MPa) and the concentration of oil in solvent is limited (typically, 2–30%).

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils Table 6.2

167

Hydrogenation of soya bean oil using a supercritical process. Fatty acid composition

IV

18:0

18:t

18:c

ct

cc

cct

ccc

3.8 15.4 27.3

0.9 1.2

23.7 41.5 53.5

1.1 0.7

54.9 27.5 6.3

0.4

6.9 1.3 0.1

133 91 59

Selectivity SII

SLn

SLo

4 3

1.7 1.9

2.2 4.7

SII , -d(trans)/d(IV)100; SLn , k1 /k2 (linolenic acid selectivity); SLo , k2 /k3 (linoleic acid selectivity).

Hydrogenation of edible oil in a fixed-bed reactor was studied (Tacke, 1995), using supercritical carbon dioxide as a solvent. Palladium on carbon, palladium on Deloxan® , and platinum on Deloxan® fixed-bed catalysts were compared. With platinum/Deloxan® , it was observed that a relatively small amount of trans isomers was formed with good linoleate selectivity. It is interesting to note that platinum catalysts were observed to be more selective than palladium. Recent studies (Piqueras et al., 2006a,b) used a batch reactor consisting of a small cell under a piston, in which the conditions could be changed quickly while using sunflower oil as the substrate and propane as the solvent. Powdered palladium catalysts were used in which the catalytic metal was deposited on two types of alumina support. The catalysts differed in degree of metal dispersion. Differences in the tendency to form trans isomers were reported but data do not suggest that low trans isomer levels could be achieved in combination with a reasonable linoleic acid selectivity. The tendency to form trans isomers with the Pd dispersion of the catalysts is linked with the assumed electronic properties of the Pd. Since low trans was associated with low linoleic acid selectivity, low trans may be the consequence of low selectivity. The effect of the catalyst support (porosity, particle size and acidity) and metal location on the support (edge or interior deposition) were not discussed. Studies have also been reported using supercritical conditions (H¨arr¨od et al., 2005). The catalyst used in these studies has 100-m particles and is mounted in a fixed bed to allow continuous operation. In earlier work (H¨arr¨od and Møller, 2001), it was shown that, under certain single-phase conditions, less than half the amount of trans isomers is formed compared with conventional slurry hydrogenations. This can be explained by the effects of particle size and increased hydrogen supply to the catalytic surface, which are known effects in conventional hydrogenation. Poor double-bond selectivity was also observed. As demonstrated in Table 6.2 (H¨arr¨od, 2006) by optimising the process, reduced trans (2–3%) is observed, while reasonable preferential selectivity is maintained. However, the economic feasibility of supercritical hydrogenation of edible oils on a commercial scale remains a question which requires further investigation (Dijkstra, 2006). 6.5.2.5 Electrochemical process An electrochemical hydrogenation process for vegetable oils has been described that allows hydrogenation at low temperature with significantly lower production of trans isomers (Lalvani and Mondal, 2004). This method involves the reaction of unsaturated triglyceride oils with hydrogen in the presence of a formate electrocatalyst. An electrochemical cell is used where hydrogen is generated at the cathode. The formate electrocatalyst participates

BLUK122-Dijkstra

168

September 25, 2007

20:1

Trans Fatty Acids

as a mediator, providing hydrogen atoms to the carbon–carbon double bonds of the oil. The use of the electrochemical cell process allows the reaction to occur at a significantly lower reaction temperature (25–75◦ C) than conventional hydrogenation (140–200◦ C). Soya bean oil hydrogenated to 110 IV was shown to contain approximately 6% trans isomers, using the mediator-assisted electrochemical hydrogenation technique, compared to approximately 25% trans isomers, using a conventional commercial process. Use of this method would require separation of the hydrogenated oil from the aqueous emulsion and washing to remove water-soluble additives followed by drying of the oil. The process claims a method to reduce trans isomer formation during hydrogenation that allows the use of low reaction temperatures, low amounts of inexpensive chemicals and with low operating costs. Another electrocatalytic process has been described (Pintauro, 1993) that discusses the advantages of lower temperature hydrogenation of oils compared to commercial processes. Hydrogenation in an H cell is described, in which oil, water, salt, emulsifier and solvent are in contact with an anode and a low-hydrogen overvoltage catalytic cathode, consisting of granular or powdered catalyst like Raney metal or alloy. Other catalytic metals like platinum, palladium or ruthenium can also be used as the catalytic cathode. Atomic hydrogen is generated at the cathode in amounts sufficient to hydrogenate the double bonds. This reaction takes place at 15–75◦ C compared to a typical range of 150–200◦ C in commercial processes. Soya bean oil hydrogenated to 104 IV using this technique had a trans isomer content of 5.8%. The isomerisation index (ratio of % trans isomers per IV unit decrease) was 0.166 for the electrocatalytic technique compared to values of 0.36–1.79 for conventional higher temperature hydrogenation processes. The electrocatalytic reaction appears to produce higher levels of saturates (e.g. 27.1% at 74 IV for soya bean oil) compared to conventional processing techniques. The possibility of hydrogenating at appreciable rates and low temperatures to keep trans isomer content low was cited as advantage of this technology. A further advantage of both electrocatalytic processes is the elimination of mass-transfer resistances of molecular hydrogen, leading to higher reaction rates, since atomic and not molecular hydrogen is formed and used in the process. A novel electrochemical hydrogenation method utilising a solid polymer electrolyte reactor was described by An et al. (1998). Hydrogenation can take place at moderate temperatures (60◦ C) and atmospheric pressure. It operates without a solvent or emulsifier, and water is the source of hydrogen. The process utilises a Pd-black cathode. A review of electrocatalytic hydrogenation studies related to low trans isomer formation has been published by Jang et al. (2005). 6.5.2.6

Ultrasonic process

Another innovative reactor design that resulted in lower trans isomer formation uses an ultrasonic device in a continuous hydrogenation system. A slurry of feedstock oil and catalyst is pumped through a continuous processing cell in contact with a device supplying ultrasonic energy. Reaction rates can be increased with such a device by producing extreme localised heat and pressure under the proper conditions (Moulton et al., 1983). The use of ultrasonic energy resulted in over a 100-fold higher rates of copper-chromite-catalysed hydrogenation of soya bean oil in a continuous system. The selectivity that is reported is lower, but trans isomer formation is described as lower than that in batch hydrogenation. Values of % trans/delta IV were in the range of 0.56–0.60.

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

6.5.3 6.5.3.1

169

Hydrogenation additives Nitrogen compounds

During the 1980s, several patents were issued claiming improved selectivity for hydrogenation of edible oils and decrease in trans isomer formation by the addition of basic nitrogencontaining compounds to nickel and precious metal (Pd, Pt and Rh) catalysts (Cahen, 1979, 1980; Kuiper, 1980, 1981a,b). Hydrogenation experiments were performed in a solvent system. The addition of a nitrogen-containing compound to a Ni catalyst is expected to decrease activity because the active sites on the nickel catalyst will be partially blocked. The effect of adding amines to precious metals (Pd) was also studied for the hydrogenation of fatty acid methyl esters of sunflower oil (Nohair et al., 2005). A lower trans isomer formation was observed when both methyl and butyl amines were added. Activity decreased with methyl amine, but butyl amine addition caused no deactivation. 6.5.3.2

Other additives

A patent application (Higgins, 2004) claims that trans isomer formation is minimised when a nickel catalyst is used that is conditioned with agents like organic acid phosphates and phosphoric acid. The purpose of these additives was to decrease the activity of the catalyst by poisoning, thus inhibiting the formation of trans isomers. When an acid phosphate was added (e.g. phosphated mono- and diglycerides), a low trans hydrogenated oil could be prepared that contained <10% trans fatty acids after hydrogenation at relatively low reaction temperatures (132◦ C). In another patent (Qualeatti, 1985), the addition of a phosphorus-containing compound to a nickel catalyst is described for the hydrogenation of edible oil. In this case, it is claimed that selectivity increases without activity loss. However, the levels of trans isomers were high (approximately, 46–49%) at 75 IV for soya bean oil. In this work, nickel is modified in various ways, including the treatment of zero-valent nickel with phosphoric acid compounds. This material is reduced in a hydrogen atmosphere at relatively high temperatures (500–700◦ C) after the treatment. The high-temperature reduction step may explain the relatively high trans isomer formation.

6.5.4 6.5.4.1

Alternate catalysts Modified nickel

Nickel catalysts are widely used for edible oil hydrogenation and several types are available. When hydrogenating under typical reaction conditions (e.g. 140–200◦ C and <0.5 MPa of hydrogen), standard catalysts produce relatively large amounts of trans isomers (see Fig. 6.12 for hydrogenation of soya bean oil to a solid product at 70 IV). Conventional nickel catalysts produce about 30–50% trans isomers under these conditions, while about 7–20% additional saturates (18:0 for soya bean oil) are also formed. It is well known that the reaction temperature has a substantial impact on formation of trans isomers in the hydrogenation of edible oil. Lowering the temperature causes a decrease in trans isomer formation accompanied by a small increase in saturate level (see Fig. 6.12 showing the effect of temperature on TFA and saturates) for soya bean oil hydrogenation to 70 IV. This plot indicates that trans isomer formation decreased from 30–40% (at hydrogenation temperatures of 175–200◦ C) to

BLUK122-Dijkstra

170

September 25, 2007

20:1

Trans Fatty Acids 70 60 Trans fatty acids

Solids (%)

50

Ni-catalysed standard hydrogenation

40

Low % trans, higher % saturates

30

Low % trans, low % saturates

20 10 Saturates 0 0

10

20

30

40

50

60

Temperature (°C) Fig. 6.14

Solid fat curves for fats with varying amounts of trans and saturates.

15–20% (at 60–140◦ C). For light hydrogenation (see Fig. 6.13), the trans isomer formation decreases from about 15% (at hydrogenation temperatures of 175–200◦ C) to less than 8% (at 80–100◦ C), depending on the conditions and catalyst. There is a link between trans isomer and saturate formation: lowering trans isomer levels increases saturate levels. This link is shown in both Figs. 6.12 and 6.13. Lowering the temperature will lead to lower reaction rates, especially when the nickel catalyst has been partially deactivated by exposure to air. The reduced nickel metal will oxidise, which will impair its catalytic activity, especially at lower reaction temperatures. Use of this partially oxidised nickel catalyst at low hydrogenation temperatures requires a pre-reduction treatment to convert the oxidised nickel back into its zero-valent state. The literature (van Toor et al., 2005) describes a method in which a nickel-based conventional edible oil hydrogenation catalyst is heated in the presence of hydrogen. During this treatment, the fat (in which the catalyst droplet is supplied) is melted, while the nickel catalyst is rereduced, making it active at lower temperatures. An oil product can be added during this treatment. Improved nickel catalyst systems for low trans isomer formation are the subject of continuing investigations. The hydrogenation performance of a catalyst at lower reaction temperatures (see Figs. 6.12–6.14) is shown for specially developed nickel catalysts that are active at temperatures lower than 100◦ C (Beers, 2006). If the aim is to produce a lightly hydrogenated, stable liquid frying or salad oil with low amounts of trans isomers, solid fats (saturates) could be potentially removed via winterisation or fractionation processes. 6.5.4.2 Palladium catalysts Other types of hydrogenation catalysts include precious metals like palladium, platinum, ruthenium and rhodium supported on inert materials like carbon, alumina, silica, titania or zirconia. Precious metal catalysts show higher catalytic activity at much lower temperatures

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

171

(e.g. 50◦ C) than conventional nickel catalysts at equal dosage (metal basis). In the 1960s and 1970s, it was reported that catalytic precious metals are active in the hydrogenation of edible oils according to the following sequence: Pd > Rh > Pt >> Ir > Ru > Os, whereby the activity of the latter three catalysts is too low to be economically attractive (Rylander, 1970). Each type of precious metal has different selectivity characteristics for saturation and isomerisation (see Figs. 6.12 and 6.13). In another study (Zajcew, 1960a), the order of trans isomer formation was defined as follows: Pd > Rh > Ru > Ir > Pt. It was also reported that Pd forms more trans isomers than Ni (Zajcew, 1960b). The observation that palladium does not lead to lower trans isomer formation than nickel was also reported (Ray, 1985). A reaction temperature study on a Pd catalyst indicated a strong dependency of trans isomer formation on the hydrogenation temperature (Makaryan et al., 2000). The trans isomer formation decreases from 22% at 90◦ C to 15% at 50◦ C in the hydrogenation of soya bean oil to 92–97 IV. In Fig. 6.12, trans isomer formation has been plotted against saturate formation for standard nickel and various precious metal catalysts. These laboratory results were obtained at hydrogen pressures (< 0.5 MPa) that can be achieved in most typical edible oil plants. Palladiumbased catalysts show similar selectivity results to those of nickel, while the ruthenium- and rhodium-based catalysts show higher saturate formation with a similar trans isomer formation to those of Ni or Pd. 6.5.4.3 Platinum and nanoplatinum Platinum catalysts demonstrate very low formation (as low as 2.0%) of trans isomers (see Fig. 6.12). However, the amount of saturates that are formed in these studies is high. For a product with a higher IV, in which a liquid oil product is desired (e.g. frying or salad oil), saturates might possibly be removed with a fractionation or winterisation step after the hydrogenation. The result would be a liquid fraction with low trans isomer content. For a more solid, more fully hydrogenated product (e.g. shortening), a dramatic change in melting characteristics is observed. Melting properties are frequently expressed as plots of solid fat content (SFC) against temperature. Hydrogenation with platinum (high saturates and low trans) produces a flatter SFC curve in comparison with that from a product obtained with a standard nickel catalyst (lower saturates and higher trans isomers). Figure 6.14 shows SFC curves for products obtained with Ni- and Pt-catalysed hydrogenation. Another type of Pt catalyst was prepared by modifying the surface with nitrogen-containing compounds, so that low amounts of trans isomers and saturates (see Fig. 6.12) are produced. In the conventional process where higher levels of trans isomers are formed, a higher SFC is found at low (10◦ C) temperature. However, oils produced with this modified catalyst result in lower trans isomer levels and SFC curves that indicate very low amounts of solids at low temperature (see markers in Fig. 6.14). This type of SFC curve is characteristic of a soft product that may be too soft for use in many food applications, in which a harder product is desired. Moreover, catalyst activity decreases dramatically by the inhibition of the nitrogencontaining compound. The effects of adding a basic nitrogen-containing compound to both nickel and precious metal catalysts have been discussed in more detail in Section 6.5.3.1. A new type of precious metal catalyst has been synthesised in such a way that the active metal is present in the form of nanoparticles (Beers and Berben, 2006). The morphology of the catalyst is an aggregate of three components: precious metal particles, support and a polymer. The precious metal nanoparticles are present in the form of clustered elementary nanoparticles (1–12 nm). The first step in the preparation is the formation of precious metal nanoparticles in the presence of a polymer into clusters of polymer-bonded nanoparticles.

BLUK122-Dijkstra

172

September 25, 2007

20:1

Trans Fatty Acids

The nature of the clusters is not entirely understood, but one theory is that the particles are bonded together by the polymer. It may be possible that the polymer provides some sort of coating or shielding between the nanoparticles to produce the clustered nanoparticles. The size of these clusters is generally 12–40 nm. In the second step, the clustered nanoparticles are combined with the polymer and the support. This aggregate forms a stable heterogeneous catalyst which can be used for the hydrogenation of edible oils. As an example, a nano Pt catalyst with polyvinyl pyrrolidone as the polymer and silica as the support was used to hydrogenate soya bean oil. Hydrogenation was carried out to 70 IV, and only 4.5% total trans isomers and 25.9% C18 saturates were observed using 50◦ C and 0.4 MPa hydrogen pressure reaction conditions. Considering that the feedstock initially contains about 1.5% trans isomers and 4.3% C18:0 , only 3% additional trans isomers are formed with this catalyst system. Under these same reaction conditions, a standard Pt catalyst produces about twice as much trans isomers and more saturates. Precious metal catalysts like platinum require different process-handling techniques from those used with conventional nickel catalysts. The higher cost of precious metals compared with nickel makes it necessary to optimise the use of the catalyst and recover the precious metal as efficiently as possible after hydrogenation to optimise process economics. Precious metal catalyst recovery is common in many industries, including the fatty acid industry. The adoption of precious metal catalysts in the edible oil industry to take advantage of the low trans isomer forming potential would require special consideration. Reuse and efficient spent catalyst recovery followed by economic metal recovery are essential if this type of catalyst is to be used instead of the more conventional nickel. 6.5.4.4 Gold catalysts The catalytic activity of supported gold catalysts has been reported (Caceres et al., 1985) for the hydrogenation of canola oil at reaction temperatures of 150–250◦ C and hydrogen pressures of up to 5.6 MPa. Despite the higher pressure, the activity was significantly lower than that observed for other types of catalyst systems. Higher levels of trans isomers were formed when compared to standard nickel catalysts. Gold catalysts do not appear to have economic or technical advantages for reducing trans isomer levels during hydrogenation. 6.5.4.5 Copper catalysts Copper catalysts are widely used in hydrogenation chemistry for other applications outside the field of edible oils. Application examples in the oleochemical area include the hydrogenolysis of fatty acids, methyl esters of fatty acids and fatty wax esters to fatty alcohols. Catalysts used for these processes are typically copper deposited on an inert support, copper chromite and copper–zinc. Reduction of the carboxyl group to the hydroxyl group requires high pressures (4–30 MPa). In a conventional liquid-phase process, in the pressure range of about 10 MPa, the predominant reaction is the saturation of the carbon–carbon double bonds in the fatty acid chain. At these lower pressures, copper can be used for partial hydrogenation of edible oils. Since copper is less active than nickel, higher dosages and higher pressures are required for the saturation reaction compared to those currently used in edible oil hardening. The mechanism of the reaction is different from that with nickel (Dijkstra, 2002b). At low-to-moderate pressures, copper will not hydrogenate isolated double bonds. Conjugated polyenes are first formed through hydrogen abstraction, followed by hydrogenation of one of the conjugated double bonds. Under hydrogenation pressures normally employed in edible oil hydrogenation, little stearic acid is formed (De Jonge et al., 1965; Okkerse et al.,

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

173

25 Cu/Cr catalyst Ni catalyst

% trans fatty acid

20

15

trans

10

C18:0

5 C18:3 0 133

128

123

118

113

108

103

98

IV Fig. 6.15

Hydrogenation with Cu and Ni catalysts.

1967). In addition, it is likely that copper shows the same behaviour as other metals in that trans isomers are reduced under enhanced hydrogen supply conditions (M¨unzing et al., 1986). Not all literature is consistent with this generalisation. At high pressure (ca. 6 MPa), high levels of trans isomers were found (Koritala, 1980). This contrasts with findings at atmospheric pressure where geometric isomerisation was negligible (Koritala and Scholfield, 1970). In both cases, methyl esters rather than triglycerides were studied. Excluding the very high pressures used in the conventional fatty alcohol processes (20–30 MPa), the high linoleate selectivity of copper catalysts makes it possible to use relatively high pressures to obtain low trans levels while maintaining excellent selectivity (see Fig. 6.15, BASF Catalysts LLC, unpublished observations). Even though low trans hydrogenation would probably require higher pressure equipment than is typically used for conventional edible oil hydrogenation, this idea warrants further investigation. 6.5.4.6 Homogeneous catalysts Homogeneous catalysts have been studied for catalytic hydrogenation of edible oils. These are metal complexes that dissolve in the feedstock oil. These types of complexes are noted for superior selectivities in various chemical reactions. However, these complexes are costly, making recovery essential to make processes economically viable. Their solubility in edible oil makes catalyst removal and recovery difficult but these are necessary for food safety requirements. Several metal carbonyl complexes (Fe, Cr, Co and Mn), metal acetyl acetonate compounds, and several nickel and platinum metal complexes have been studied. Chromium carbonyl complexes were reported to saturate the double bonds of the triglyceride very selectively

BLUK122-Dijkstra

174

September 25, 2007

20:1

Trans Fatty Acids

while keeping trans isomer formation at a minimum (Awl et al., 1978; Bernstein et al., 1989; Frankel et al., 1979). The RuCl2 (CO)2 (PPh3 )2 complex is reported to show high activity in the hydrogenation of edible oils (Bello et al., 1985). Although the researchers concluded that this Ru complex has the potential to produce partially hydrogenated oils with low trans isomer concentrations, the level of trans isomers observed for canola oil hydrogenation was about 67% at 74 IV. In another study, a Ru complex gave a trans isomer formation of 7% for canola oil hydrogenated to 60 IV at 110◦ C (Wright et al., 2003a). An attempt was made to control activity, selectivity and trans isomerisation by mixing conventional heterogeneous nickel catalysts with homogeneous Cr, Pd and Ru complexes (Wright et al., 2003a,b). While the combination of Cr and Pd complexes with Ni catalyst did not improve performance, the addition of a Ni catalyst to a Ru complex resulted in increased activity while producing a trans isomer content as low as 9% for oil hydrogenated to 60 IV at 110◦ C (Wright et al., 2003a). Because homogeneous catalysts are soluble in the reaction medium, removal and recovery after hydrogenation is difficult. A way to overcome this problem is to immobilise the active catalyst on a support which makes them heterogeneous and removable via filtration. This was done for a PdCl4 complex on an ion-exchange resin. Although trans isomer formation was reportedly lower than that with a copper catalyst, 22% TFA was formed in the hydrogenation of cottonseed oil from 98 to 58 IV at 40◦ C while using ethanol as solvent (Hinze, 1975). Many homogeneously catalysed edible oil hydrogenation studies were performed under conditions that can be considered impractical for the edible oil industry because of high pressures (up to 20 MPa), high catalyst dosages and use of solvents. Metal carbonyl complexes such as Co2 (CO)8 , Fe(CO)5 , Fe3 (CO)12 and Cr(CO)6 were tested for the hydrogenation of soya bean oil under reaction conditions that are more suitable for the edible oil industry (no solvent, maximum hydrogen pressure of 0.5 MPa and moderate temperatures of 50–200◦ C). These complexes showed little hydrogenation activity (BASF Catalysts LLC, unpublished observations). Precious metal complexes such as RuCl2 (PPh3 )3 , Rh(H)(CO)(PPh)3 and RhCl(PPh)3 , which are well known for their activity and selectivity in other hydrogenation reactions, have been tested for hydrogenation of soya bean oil to 105 IV (no solvent). Only RhCl(PPh)3 (Wilkinson catalyst) showed hydrogenation activity. This complex shows a high formation of trans isomers (up to 60%) for a light hydrogenation of soya bean oil to 105 IV (BASF Catalysts LLC, unpublished observations). A high isomerisation index has also been observed for platinum–tin complexes during a hydrogenation of soya bean oil, which showed a strong preference for the hydrogenation of linoleic acid (Van’t Hof, 1970; Van’t Hof and Linsen, 1967). The reaction was performed with a solvent and high trans isomer levels were observed (33% at 102 IV and 60% at 70 IV). Besides the practical problems of catalyst recovery, reusability and high cost for edible oil hydrogenation, homogeneous catalyst systems do not provide ideal selectivity for the saturation of multiple double bonds like 18:3, low saturate formation and minimal formation of trans isomers. It should be noted that a homogeneous complex can be reduced to a heterogeneous metallic compound under certain hydrogenation conditions and that the complexes are not always dissolved in the reaction medium. 6.5.4.7

Zeolite supports

The use of zeolites as an alternative inorganic support has been studied (Jacobs et al., 2001). Zeolites are crystalline aluminosilicate structures that possess a fixed dimensional framework

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

175

with cavities and pores of uniform size and shape. Due to this unique structure, a special type of reaction selectivity may be possible. This is sometimes termed ‘shape selectivity’, in which some molecules with typical sizes or shapes can be preferentially admitted to the structured zeolite pores. For edible oil hydrogenation, zeolite structures can theoretically be selected in which the linear trans unsaturated fatty acid chains of a triglyceride can enter the pores, while the non-linear cis unsaturated chains are excluded. The active catalytic metal (e.g. Ni or precious metal) needs to be deposited only within the pore structure of the zeolite. By competitive ion exchange of a ZSM-5 zeolite with a Pt precursor and a salt, it is claimed (Jacobs et al., 2001) that an edible oil hydrogenation catalyst results that will selectively hydrogenate the trans isomers, while the cis isomers are not affected. This is an elegant way to hydrogenate very selectively, but the study was mainly carried out on methyl esters and little information is given on the activity of such a system for triglyceride oils. Since the majority of the active sites must be located inside the pores, the question remains of how to achieve sufficiently high metal loadings of the active catalyst inside the pores. Continuing work on the zeolite catalyst system has led to improvements, but has not yet been published (Dijkstra, 2006).

6.6

SUMMARY

Hydrogenation of edible oils to improve oxidative stability and increase SFC without significant increase in saturate levels has served the food industry well for many years. The formation of trans isomers as a by-product of the hydrogenation reaction has provided an opportunity to achieve desirable melting properties for usage convenience (e.g. margarines, shortenings, coating fats, etc.). The edible oil industry has designed processes to take advantage of the physical properties of the geometric isomers and has also found methods to control unwanted levels of these oil components (e.g. winterisation or fractionation). In recent years, the beneficial properties of trans isomers have been offset somewhat by the desire to minimise or eliminate them from products to address the demand for healthier food oils. The industry has responded to reduction of trans isomer levels in food products in a variety of ways. These include:

r r r r

Use of an alternative processing technique like interesterification, in which trans isomers are not formed. Interesterification feedstocks can include fully hydrogenated oils that do not contain trans isomers. Reformulation of products to incorporate non-hydrogenated oils (no trans content). Use of unhydrogenated, high-saturate, high-solid-content oils like palm oil or palm oil fractions for imparting higher melting properties. Use of more oxidatively stable oils produced from genetically modified special oilseeds.

The demand for lower trans isomer levels has presented a new challenge to the industry to develop alternative processing methods to achieve more stringent control of the hydrogenation process. While the use of hydrogenation in the edible oil industry has declined somewhat, this process remains widely practised. The hydrogenation reaction has come under increasing scrutiny because of the formation of the trans isomer by-product. Some level of control over trans isomer formation is possible by modification of reaction conditions. Using conventional existing hydrogenation equipment, careful control of parameters like temperature, hydrogen pressure, mixing speed, catalyst type

BLUK122-Dijkstra

176

September 25, 2007

20:1

Trans Fatty Acids

and catalyst dosage can help achieve some reduction of trans isomer levels. In many cases, controlling the formation of trans isomers in conventional processes by adjusting reaction conditions leads to increases in formation of saturates. This necessitates a compromise and balancing reduced trans, saturate and polyunsaturated levels to achieve the best results. Modifying reaction conditions within the capabilities of the conventional hydrogenation equipment is well known and has been practised to varying degrees for many years. Modification of reaction conditions or other process parameters beyond the capabilities of the conventional equipment and process leads to new challenges. For example, it has been shown that hydrogenations performed at below conventional reaction initiation temperatures can produce a significant reduction in trans isomer levels, especially if measured as trans isomer content per unit IV decrease. However, lower reaction temperatures imply the ability to reduce oil temperatures below those of storage conditions or the level at which the oil exits upstream refining and bleaching processes. Maintaining conditions at or near this low starting temperature during the exothermic hydrogenation reaction requires adequate heat-removal capability in the reactor. Process changes that do not allow recovery of the exothermic reaction heat via heat-exchanging equipment will not allow the heating of incoming oil possibly leading to increased energy costs. These capabilities are not always part of the design of existing, conventional commercial-scale hydrogenation equipment. The commercial hydrogenator could be facing some capital expense to achieve adequate temperature-control and heat-recovery capability. Alternative processing techniques like fixed bed, electrochemical, ultrasonic or supercritical solvent hydrogenation also present equipment considerations. While these techniques may offer the opportunity for better trans isomer control, they would require non-conventional equipment and modification or replacement of existing equipment, which would entail capital expense. For example, the use of solvents would require larger vessels suitable for higher pressures, solvent removal and recovery processes and equipment. Consideration of operating rates and costs also need to be considered for any new hydrogenation process. The advantages of lower trans isomer levels must be studied in the context of throughputs, costs, energy consumption and overall process economics to establish whether they are of commercial value. Alternative catalyst systems also present opportunities for improved trans isomer control. Current technology on non-conventional catalyst systems does not indicate the possibility to hydrogenate, without at least some formation of trans isomers. Some alternative catalysts require lower reaction temperatures to achieve a reduction in trans isomer levels and present the same temperature-control challenges, as noted above. Other catalyst systems suggest new considerations for catalyst handling. Efficient recovery of precious metal catalysts after use is essential to prevent losses of high-value materials and ensure maximised process economics. Copper catalysts would require complete removal, so that trace levels of this metal would not promote post-hydrogenation oxidation reactions. More exotic catalyst systems may increase the costs of the hydrogenation process. Homogeneous catalyst systems would require special processing techniques to remove the catalyst from food oils. Modifications of conventional nickel catalysts could necessitate an extra processing step. Of the various catalytic metals that have been studied for edible oil hydrogenation, platinum seems to represent the most favourable opportunity for minimising trans isomer formation. Research studies have indicated that platinum catalysts used at low reaction temperatures provide trans isomer contents that are significantly lower than traditional levels. The drawbacks to commercial acceptance of this technology have been process cost, saturate selectivity, lowtemperature reaction/heat-removal issues and spent catalyst recovery considerations. While

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

177

platinum catalysis is used for other chemical processes, their use for edible oil hydrogenation to achieve lower trans levels would pose different process, catalyst handling and recovery considerations from conventional processing. Reduction in trans isomer levels in more completely hydrogenated edible oil products also presents a dilemma to the processor. When the higher melting temperature trans isomer components are removed or decreased, they must be replaced with saturated components to achieve the desired SFC properties and meet the specifications of the original product. The use of saturated fats for this purpose can lead to crystallisation and texture issues in edible oil hard fats, like shortenings and margarines. Major reformulations may be required. Most alternative catalyst systems and reaction conditions that have been studied to lower trans produce higher levels of saturates at comparable IV. As the demand for lower trans isomer contents in food products continues, the industry is at a crossroads and faces major challenges to meet new dramatic changes in product specifications. The scientific effort continues in many areas to meet these challenges. As seen in this review, alternate processing schemes, catalyst systems and processing equipment have been made available and continue to be studied to address the newly changing needs of the food industry. The expertise and perseverance of the industry and the technical support staff will lead to new and improved methods of food oil production.

REFERENCES An, W., Hong, J.K., Pintauro, P.N., Warner, K. & Neff, W.E. (1998) The electrochemical hydrogenation of edible oils in a solid polymer electrolyte reactor. I: reactor design and operations. J Am Oil Chem Soc 75 (8), 917–925. Awl, R.A., Frankel, E.N., Friedrich, J.P. & Pryde, E.H. (1978) Cis-Bond-producing hydrogenation of polyunsaturates catalyzed by polymer-complexed Cr(CO)3 catalysts. J Am Oil Chem Soc 55 (7), 577–582. Beers, A.E.W. (2006) Catalyst technology for the low trans hydrogenation of edible oils. Paper presented at the AOCS World Conference and Exhibition on Oilseed and Vegetable Oil Utilization, Istanbul, Turkey. Beers, A.E.W. & Berben, P.H. (2006) Process for the hydrogenation of unsaturated triglycerides. PCT Patent Application 2006/121320, assigned to Engelhard Corporation. Bello, C., Diosady, L.L., Graydon, W.F. & Rubin, L.J. (1985) Homogeneous catalytic hydrogenation of canola oil using a ruthenium catalyst. J Am Oil Chem Soc 62 (11), 1587–1592. Bengtson, G. & Fritsch, D. (2006) Catalytic membrane reactor for the selective hydrogenation of edible oil: platinum versus palladium catalyst. Desalination 200, 666–667. Bern, L., Lidefelt, J.-O. & Sch¨oo¨ n, N.-H. (1976) Mass transfer and scale-up in fat hydrogenation. J Am Oil Chem Soc 53 (7), 463–466. Bernstein, P.A., Graydon, W.F. & Diosady, L.L. (1989) Hydrogenation of canola oil using chromium catalysts. J Am Oil Chem Soc 66 (5), 680–684. Boger, T., Zieverink, M.M.P., Kreutzer, M.T. et al. (2004) Monolithic catalysts as an alternative to slurry systems: hydrogenation of edible oil. Ind Eng Chem Res 43, 2337–2344. Buehler, G. (1989) Hydrogenation of fatty acids. In: Fatty Acids in Industry (eds R.W. Johnson & E. Fritz). Marcel Dekker Inc, New York and Basel, pp. 113–138. Caceres, L., Diosady, L.L., Graydon, W.F. & Rubin, L.J. (1985) Supported gold catalysts in the hydrogenation of canola oil. J Am Oil Chem Soc 62 (5), 906–910. Cahen, R.M. (1979) Hydrogenation process. US Patent 4,161,483, assigned to Labofina S.A. Cahen, R.M. (1980) Hydrogenation catalyst and hydrogenation process. US Patent 4,229,361, assigned to Labofina S.A. Chen, A.H., McIntire, D.D., Gibson, P. & Covey, J.E. (1981) Investigation and modeling of mass transfer in soybean hydrogenation. Paper presented at the 72nd AOCS Annual Meeting and Expo, New Orleans. de Jonge, A., Coenen, J.W.E. & Okkerse, C. (1965) Selective hydrogenation of linolenate groups in soybean oil. Nature 206 (4984), 573–574.

BLUK122-Dijkstra

178

September 25, 2007

20:1

Trans Fatty Acids

Dijkstra, A.J. (1997) Hydrogenation revisited. Inform 8 (11), 1150–1158. Dijkstra, A.J. (2002a) Hydrogenation and fractionation. In: Fats in Food Technology (ed. K.K. Rajah). Sheffield Academic Press, Sheffield, pp. 123–158. Dijkstra, A.J. (2002b) On the mechanism of the copper-catalysed hydrogenation; a reinterpretation of published data. Eur J Lipid Sci Technol 104 (1), 29–35. Dijkstra, A.J. (2006) Revisiting the formation of trans isomers during the partial hydrogenation of triglyceride oils. Eur J Lipid Sci Technol 108 (3), 249–264. Edvardsson, J. & Irandoust, S. (1994) Reactors for hydrogenation of edible oils. J Am Oil Chem Soc 71 (3), 235–242. Ellis, C. (1930) Hydrogenation of fatty oils. In: Hydrogenation of Organic Substances, 3rd edn. D. Van Nostrand Co, Inc, New York, pp. 313–316. Frankel, E.N. (1980) Soybean oil flavor stability. In: Handbook of Soy Oil Processing and Utilization (eds D.R. Erickson, E.H. Pryde, O.L. Brekke, T.L. Mounts & R.A. Falb). American Soybean Association and American Oil Chemists’ Society, St Louis and Champaign, pp. 229–241. Frankel, E.N., Awl, R.A. & Friedrich, J.P. (1979) Cis-unsaturated fatty acid products by hydrogenation with chromium hexacarbonyl. J Am Oil Chem Soc 56 (12), 965–969. Fritsch, D. & Bengtson, G. (2006) Development of catalytically reactive porous membranes for the selective hydrogenation of sunflower oil. Catal Today 118, 121–127. H¨arr¨od, M. (2006) Low trans frying oils by selective supercritical single phase hydrogenation. Paper presented at the 97th AOCS Annual Meeting & Expo, St Louis. H¨arr¨od, M., Holmqvist, A. & van der Hark, S. (2005) Selective hydrogenation of functional groups in substrates and partially hydrogenated fatty acids and fatty acid derivatives. PCT Patent Application WO 2005/095306, assigned to H¨arr¨od Research AB. H¨arr¨od, M. & Møller, P. (2001) Partially hydrogenated fatty substances with a low content of trans fatty acids. US Patent 6,265,596, assigned to Poul Møller Ledelses- og Ingeni¨orr˚adgivning APS. Hastert, R.C. (1996) Hydrogenation. In: Bailey’s Industrial Oil and Fat Products, 5th edn (ed. Y.H. Hui). John Wiley & Sons, Inc, New York, pp. 213–300. Higgins, N.W. (2004) Low trans-stereoisomer shortening systems. US Patent Application Publication 2004/0146626 A1. Hilditch, T.P. (1941) The Industrial Chemistry of Fats and Waxes. D. Van Nostrand Co, Inc, New York. Hinze, A.G. (1975) Hydrogenation of fatty acids and their triglycerides using a Pd(II) catalyst in an ionexchange resin. US Patent 3,865,853, assigned to Unilever. Horiuti, I. & Polanyi, M. (1934) Exchange reactions of hydrogen on metallic surfaces. Trans Faraday Soc 30, 1164. Ilinitch, O.M., Simonov, P.A. & Cuperus, F.P. (2001) Nanosize palladium loaded catalytic membrane: preparation and cis-trans selectivity in hydrogenation of sunflower oil. Stud Surf Sci Catal 118, 55–61. Jacobs, P.A., Maes, P.J., Paulussen, S.J. et al. (2001) Elimination of trans-unsaturated fatty acid compounds by selective adsorption with zeolites. US Patent 6,229,032 B1, assigned to K.U. Leuven Research & Development. Jang, E.S., Jung, M.Y. & Min, D.B. (2005) Hydrogenation for Low trans and high conjugated fatty acids. Comp Rev Food Sci Saf 1 (22), 24. Katan, M.B., Zock, P.L. & Mensink, R.P. (1995) Trans fatty acids and their effect upon lipoproteins in humans.Annu Rev Nutr 15, 473–493. Kaufmann, H.P. (1939) Wilhelm Normann, zum Ged¨achtnis. Fette und Seifen 46 (5), 259–264. Koetsier, W.T. (1997) Hydrogenation of edible oils, technology and applications. In: Lipid Technologies and Applications (eds F.D. Gunstone & F.B. Padley). Marcel Dekker, Inc, New York, Basel, Hong Kong, pp. 265–303. Koritala, S. (1980) Selective hydrogenation with copper catalysts: V. Kinetics and mechanism at high pressure. J Am Oil Chem Soc 57 (9), 293–298. Koritala, S. & Scholfield, C.R. (1970) Selective hydrogenation with copper catalysts: I. Isolation and identification of isomers formed during hydrogenation of linoleate. J Am Oil Chem Soc 47 (8), 262–265. Kuiper, J. (1980) Selective hydrogenation. US Patent 4,228,088, assigned to Unilever. Kuiper, J. (1981a) Process for the selective hydrogenation of triglyceride oils with a metallic catalyst in the presence of a diamine. US Patent 4,307,026, assigned to Unilever. Kuiper, J. (1981b) Process for the selective hydrogenation of triglyceride oils with a metallic catalyst in the presence of ammonia. US Patent 4,278,609, assigned to Unilever. Lalvani, S.B. & Mondal, K. (2004) Electrochemical hydrogenation of vegetable oils. US Patent 6,825,367 B2, assigned to Board of Trustees of Southern Illinois University.

BLUK122-Dijkstra

September 25, 2007

20:1

Trans isomer control in hydrogenation of edible oils

179

Latondress, E.G. (1980) Shortenings and margarine: base stock preparation and formulation. In: Handbook of Soy Oil Processing and Utilization (eds D.R. Erickson, E.H. Pryde, O.L. Brekke, T.L. Mounts & R.A. Falb). American Soybean Association and American Oil Chemists’ Society, St Louis and Champaign, pp. 145–154. ¨ Leuteritz, G. (1969) Die kontinuierliche Hydrierung von Olen und Fetten. Fette Seifen Anstrichmittel 71 (6), 441–445. Makaryan, I.A., Matveeva, O.V., Davydova, G.I. & Savchenko, V.I. (2000) Lowering the trans-isomer content in hydrogenation of triglycerides of unsaturated fatty acids at ambient temperatures. Stud Surf Sci Catal 130, 2039–2044. Moulton Sr., K.J., Koritala, S. & Frankel, E.N. (1983) Ultrasonic hydrogenation of soybean oil. J Am Oil Chem Soc 60 (7), 1257–1258. Mounts, T.L. (1980) Hydrogenation practices. In: Handbook of Soy Oil Processing and Utilization (eds D.R. Erickson, E.H. Pryde, O.L. Brekke, T.L. Mounts & R.A. Falb). American Soybean Association and American Oil Chemists’ Society, St Louis and Champaign, pp. 131–144. M¨unzing, M., Kut, O.M. & Gut, G. (1986) Kinetik der Fetth¨artung und Vergleich verschiedener Katalysatoren. Fette Seifen Anstrichmittel 88 (10), 387–391. Nohair, B., Especel, C., Lafaye, G. et al. (2005) Palladium supported catalysts for the selective hydrogenation of sunflower oil. J Mol Catal A Chem 229, 117–126. Normann, W. (1903) Process for converting unsaturated fatty acids or their glycerides into saturated com¨ pounds. British Patent 1 515, assigned to Herforder Maschinenfett-und Olfabrik Leprince und Siveke. Okkerse, C., de Jonge, A., Coenen, J.W.E. & Rozendaal, A. (1967) Selective hydrogenation of soybean oil in the presence of copper catalysts. J Am Oil Chem Soc 44 (2), 152–156. Okonek, D.V. (1987) Nickel-sulfur catalysts for edible oil hydrogenation. In: Hydrogenation: Proceedings of an AOCS Colloquium (ed. R.C. Hastert). American Oil Chemists’ Society, Champaign, IL, pp. 65– 68. Ottesen, I. & Jensen, B.H. (1980) Influence of remaining phosphatides during hydrogenation of soybean oil. Paper presented at the 71st AOCS Annual Meeting and Expo, New York. Patterson, H.B.W. (1983) Hydrogenation of Fats and Oils. Applied Science Publishers, New York. Pintauro, P.N. (1993) Electrocatalytic process for the hydrogenation of edible and non-edible oils and fatty acids. US Patent 5,225,581, assigned to Tulane Educational Fund. Piqueras, C., Bottini, S. & Damiani, D. (2006a) Sunflower hydrogenation on Pd/Al2O3 catalysts in singlephase conditions using supercritical propane. Appl Catal 313, 177–188. Piqueras, C., Fernandez, M., Tonnetto, G., Bottini, S. & Damiani, D. (2006b) Hydrogenation of sunflower oil on Pd catalysts in supercritical conditions: effect of particle size. Catal Commun 7, 344–347. Qualeatti, G.M. (1985) Selective reduction of edible fats and oils using phosphorus-modified nickel catalysts. US Patent 4,547,319, assigned to UOP Inc. Ravnskov, U. (2000) The Cholesterol Myths Exposing the Fallacy That Saturated Fat and Cholesterol Cause Heart Disease. New Trends Publishing, Inc, Washington, DC. Ray, J.D. (1985) Behavior of hydrogenation catalysts: I. Hydrogenation of soybean oil with palladium. J Am Oil Chem Soc 62 (8), 1213–1217. Rylander, P.N. (1970) Hydrogenation of natural oils with platinum metal group catalysts. J Am Oil Chem Soc 47 (12), 482–486. Sipos, E. & Szuhaj, B.F. (1996) Soybean oil. In: Bailey’s Industral Oil and Fat Products, 5th edn (ed. Y.H. Hui). John Wiley & Sons, Inc, New York, pp. 497–601. Tacke, T. (1995) Fetth¨aitung mit Festbettkatalysatoren. CAV Chemie – Anlagen und Verfahren, 11, 18–22. van Toor, H., van Rossum, G.J. & Kruidenberg, M. (2005) Low trans-fatty acid compositions; low-temperature hydrogenation, e.g. of edible oils. US Patent Application Publication 2005/0027136 A1, assigned to Cargill Incorporated. van’t Hof, L.P. (1970) Hydrogenation of unsaturated aliphatic compounds. US Patent 3,489,778, assigned to Unilever. van’t Hof, L.P. & Linsen, B.G. (1967) Homogeneous hydrogenation with platinum-tin chloride complexes as catalysts. J Catal 7, 295–297. Veldsink, J.W. (2001) Selective hydrogenation of sunflower seed oil in a three-phase catalytic membrane reactor. J Am Oil Chem Soc 78 (5), 443–446. Weber, G.M. & Alsberg, C.L. (1934) The American Vegetable Shortening Industry. Food Research Institute, Stanford University, Stanford. Wright, A.J., Mihele, A.L. & Diosady, L.L. (2003a) Ni-catalyst promotion of a cis selective Pd catalyst for canola oil. Food Res Int 36, 1069–1072.

BLUK122-Dijkstra

180

September 25, 2007

20:1

Trans Fatty Acids

Wright, A.J., Wong, A. & Diosady, L.L. (2003b) Cis selectivity of mixed catalysts systems in canola oil hydrogenation. Food Res Int 36, 797–804. Zajcew, M. (1960a) Hydrogenation of fatty oils with palladium catalysts: products of the tall oil industry. J Am Oil Chem Soc 37 (10), 473–475. Zajcew, M. (1960b) The hydrogenation of fatty oils with palladium catalyst: hydrogenation of fatty oils for shortening stock. J Am Oil Chem Soc 37 (1), 11–14.

BLUK122-Dijkstra

September 25, 2007

7

20:2

Fractionation and interesterification

Wim De Greyt and Albert J. Dijkstra

7.1

INTRODUCTION

In addition to the hydrogenation process discussed in the previous chapter, the processes used to provide an oil blend with the required physical and chemical properties are blending of different oils and fats, single- or multi-stage fractionation, interesterification and especially combinations of these processes. A trans-free margarine fat blend (Schijf et al., 1984) can, for example, be produced by:

r r r r r

blending a lauric oil and a non-lauric oil; fully hydrogenating the mixture; randomising the fully hydrogenated fat by interesterification; fractionating the interesterified product to eliminate high-melting triglycerides and/or lowmelting triglycerides and blending the olein or mid-fraction with a liquid oil.

This liquid oil can also have been subjected to a directed interesterification process (Holemans et al., 1988) to further reduce the saturated fatty acid content of the fat blend. The various modification steps involved in producing such a margarine fat blend markedly increase its cost price. The operating costs of these steps can be summarised in order of increasing costs as follows (Kellens, 2000):

r r

r

Blending is the cheapest way of modifying oil and fat properties. On the other hand, its scope is limited by the properties of the blend components and to attain certain blend properties, it may be necessary to incorporate expensive components into the blend. Dry fractionation has the advantage of not requiring auxiliary materials or entailing a yield loss; consequently, the operating costs are low. However, fractionation leads to co-products that have also to be valorised and if this can only be done at a discount, the main product has to bear this cost. On the other hand, it is often possible to recycle certain fractions in a multi-stage fractionation process and thereby increase the yield of the most expensive product. The investment required for the interesterification process is somewhat lower than that for fractionation but because of the yield loss resulting from the chemical interesterification process, its operating costs are somewhat higher than those of the fractionation process. In the case of enzyme-catalysed interesterification, the enzyme costs are quite significant.

BLUK122-Dijkstra

182

r

September 25, 2007

20:2

Trans Fatty Acids

Hydrogenation is the most expensive modification process, with respect to not only fixed costs but especially variable costs: catalyst and hydrogen. Partial hydrogenation requiring less time and hydrogen is of course cheaper than full hydrogenation.

7.2 7.2.1

FRACTIONATION Historical

During the fractionation process, the oil or fat being fractionated is partially crystallised and subsequently, the crystals formed are separated from the mother liquor. The crystals are isolated as a filter cake that is referred to as the stearin fraction and the filtrate is called the olein fraction. Although dry fractionation (crystallisation from the melt) was the first fractionation process to be used on an industrial scale (M`ege, 1869), it took over a century before the process started to make a real impact. The likely reason is that early separation systems like vacuum drum filters and later band filters (Tirtiaux, 1976) left quite an amount of olein in the stearin filter cake. In general, olein properties are determined by the cooling profile and slurry temperature at filtration. Stearin properties and the yield of both fractions depend on the filtration temperature but are also affected by the olein entrainment in the stearin filter cake (Dijkstra, 2002). Accordingly, a higher specificity of the fractionation could only be realised by diluting the olein in the cake with a solvent. Moreover, washing the filter cake with fresh solvent further reduced its olein content. Therefore, solvent fractionation could achieve what dry fractionation could not and this justified its continued use (Hamm, 1986), despite the expense. To avoid the use of flammable solvents, the detergent fractionation process (Seug´e and Vinconneau, 1975) was developed. In this process, originally described by Fratelli Lanza of Turin, Italy (Anonymous, 1907), and later developed by Alfa Laval as the Lipofrac® process, a surfactant solution in water is used to bring the crystallised fat into the aqueous phase which is then separated from the olein by using centrifugal separators. With improved separation technology resulting from the use of a conical sieve centrifuge fitted with a co-rotating scroll (Maes and Dijkstra, 1985), a high-pressure membrane filter press (Willner et al., 1989, 1990) or a decanter (Deffense, 2005), the dry fractionation process has completely ousted the detergent process. The fact that the olein content (at 35–52%) of the stearin resulting from the detergent process is no better than what can be achieved with a membrane press (Hamm and Timms, 2006), plus the cost of the centrifugal separators and the surfactant and the resulting water-disposal problem all contribute to explaining why new detergent fractionation processes are being installed for only very specific applications (e.g. fractionation of fatty acids). No new solvent fractionation plants are being built either, and the few plants that are still running are used for high-added-value specialities. As pointed out recently by Harris (2005), solvent fractionation is the only process that can handle a feed stream that is 80% crystallised. It can thus be used to remove a small amount of olein from cocoa butter and thereby upgrade this natural and therefore variable product. In addition, solvent fractionation has an advantage over dry fractionation in that crystallisation is much faster. With this fast crystallisation, smaller vessels would suffice for a given throughput but the use of dilute solutions requires larger vessels. Given the developments outlined above (Hamm, 1995), the present chapter will be limited to the various dry fractionation processes in current use. It will discuss both stationary

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

183

crystallisation processes and the partial crystallisation of an agitated melt, the various separation processes and the equipment required and the products obtained by these processes.

7.2.2

Fat crystallisation theory

Fat crystallisation affects a large number of food products and processes. It should for instance provide chocolate with a snap on breaking and should prevent margarine from oiling out. On the other hand, the crystals in ghee should sink to the bottom and leave a clear oily supernatant. In puff pastry margarine, the fat crystals should provide the product with plasticity; in physically ripened cream, the crystals should facilitate churning; and in dry fractionation, the crystals should permit the olein to be separated from the stearin. These various demands can only be met by different crystal morphologies, and arriving at these different morphologies necessitates using different crystallisation techniques: tempering for chocolate (Padley, 1997), scraped heat exchangers for margarine (Poot and Biernoth, 1994), a slow cooling for ghee (Podmore, 2002, section 9.5), patience for cream (Robinson and Rajah, 2002, section 6.4.3) and an artistic talent for dry fractionation (Tirtiaux, 1990). The situation is complicated by the fact that different fats behave differently. If molten butter is allowed to cool down without being agitated, a deposit of filterable crystals is formed eventually, as observed in ghee. If lard is treated the same way, a grainy plastic solid results. Apparently, the triglyceride composition of the fat being crystallised affects the resulting crystal morphology but that is not the only factor. Non-triglyceride components such as phosphatides (Smith, 2000) and mono- and diglycerides can also have an influence. Highmelting diglycerides can act as crystal initiators because their poor solubility causes them to crystallise at relatively high temperatures. On the other hand, diglycerides with a somewhat lower melting point may attach themselves at a crystal growth point, thereby disturbing the regularity of the crystal lattice and preventing it from growing further until the diglyceride has dissolved away again. Accordingly, some diglycerides can be used as crystallisation inhibitors and are therefore used as additives to improve cold stability of liquid olein fractions. A very important factor in fat crystallisation is the temperature. This parameter is therefore widely used to control the crystallisation and the ensuing crystal morphology. When producing margarine, which should contain many small crystals to provide plasticity and retain the liquid oil, the emulsion is passed through a scraped-surface heat exchanger that causes the fat present in the emulsion to crystallise in the unstable polymorph which subsequently recrystallises into the more stable polymorph. There is an even more stable polymorph () but in margarine, its formation is undesirable. Because of its high stability, its melting point is higher than that of the and polymorphs. Its crystals tend to be large and thus provide the margarine with a rough surface and a poor mouthfeel referred to as sandy. For dry fractionation, large crystals could well facilitate crystal separation and might thus be an advantage. However, not all fats crystallise in a polymorph. Its dense packing can be attained only when the crystallising triglycerides are very similar. Accordingly, tristearin happily crystallises in the form but milkfat does not. In this context, only the crystallising triglycerides should be considered: tristearate dissolved in a liquid oil will still form crystals, be it at lower temperature. Crystallisation comprises several steps (Foubert, 2007). If the thermodynamic driving force for crystallisation is sufficiently large, crystal nuclei will appear. They can be formed by homogeneous primary nucleation but this requires undercooling by >30 K (Kloek, 1998) and in industrial practice, heterogeneous primary nucleation by dust particles constitutes the main nucleation mechanism, at least in the beginning of the crystallisation process (Timms, 1991).

BLUK122-Dijkstra

184

September 25, 2007

20:2

Trans Fatty Acids

In a multi-stage fractionation of palm oil, the first olein may contain very low levels of trisaturated triglycerides (<0.5% tripalmitin) and so few of these dust particles that both homogenous and heterogeneous primary nucleation will be impeded. To facilitate the crystal initiation of palm olein, it is therefore common practice today in industry to add small amounts of palm oil (5–10%), containing 6–8% trisaturated triglycerides. This beneficial effect was described earlier by Maes et al. (1995), who found that introducing high-melting triglycerides by the gradual addition of a small amount of palm oil to the olein while this was being cooled down and crystallised increased the rate of crystallisation and caused the resulting crystals to be readily filterable. Inoculation with seed crystals is another possibility (Deffense, 1998; von Rappard and Plonis, 1980). When crystallisation is under way, new nuclei may also be formed by a process of heterogeneous secondary nucleation, for which phenomenon Walstra (1998) has given a tentative explanation. He assumed that clusters of more or less oriented molecules diffuse away from the crystal and subsequently form a new nucleus. This requires the crystal growth rate to be rather slow so that the clusters have the opportunity to diffuse away before being incorporated into the original crystal lattice. In the industrial dry fractionation process, secondary nucleation is considered to be responsible for the formation of additional small crystals that badly affect the filtration characteristics of the stearin cake and thus make the process less selective (Timms, 1997). The well-known difficulty of maintaining crystallisation characteristics on scaling up may well arise because secondary nucleation gains in importance with increasing vessel size and linear agitator speed (M. Kellens, personal communication). After nuclei have emerged, they will grow and form crystals, provided the temperature of the melt is lowered. The presence of solvents has very little influence on the triglyceride composition of these crystals (Coenen, 1974; Hamm, 1986; Timms, 1997); this means that diluting the fat to be fractionated with a liquid oil (Van Putte and Muller, 1987) does not lead to different crystals either. If only one species of high-melting triglycerides is present, this is the species that will crystallise, but natural products processed industrially normally contain many different triglyceride species. This raises the question of which species will crystallise and when. The different species present in natural fats will form solid solutions. They will not crystallise as separate, pure compounds but as mixtures of slightly different triglycerides. Therefore, they affect each others’ solubility. In this respect, they behave quite differently than, for instance, sugar and salt behave when they are dissolved in water together. The solubility of the salt is not affected by the sugar concentration and vice versa. The salt starts to crystallise as and when the solution becomes supersaturated in salt, irrespective of the amount of sugar present. When cooling a molten fat, the first crystals will be formed long before the melt becomes supersaturated in the triglycerides with the highest melting point. Similar but different triglycerides with a somewhat lower melting point will form mixed crystals with the higher melting triglycerides, and their total concentration determines when the melt becomes supersaturated. On further cooling, the triglyceride composition of newly formed crystalline matter changes in that it starts to contain more and more triglycerides with lower and lower melting points. There are cases known where different crystal phases can form a eutectic mixture (Smith, 2001). In palm oil, the higher melting phase is mainly PPP and POP, while the lower melting phase consists principally of POO. As only to be expected, the study of model systems shows that triglyceride interactions can lead to a more complex crystallisation behaviour, but in general, this requires higher concentrations of particular triglycerides or a greater substrate purity than encountered in the industrial environment. Moreover, the latter has to cope with variable partial glyceride concentrations that tend to have a larger effect on the crystallisation process than the interactions mentioned earlier.

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

7.2.3

185

Fat crystallisation practice

Two main systems are in use for the industrial dry fractional crystallisation of fats, which is always carried out as a batch process. One system employs large crystallisers provided with cooling means and agitators for heat-transfer purposes; it is also referred to as ‘suspension crystallisation’ (Peters-Erjawetz et al., 1999). The other system uses shallow trays (Yoneda et al., 1997) or narrow vertical crystallisation chambers (Hendrix and Kellens, 2003) that permit stationary crystallisation without agitation. A third system (Deffense, 1999; Tirtiaux and Tan, 1997) causes the fat to crystallise by mixing it with cold water, which is then separated from the partially crystallised fat and then can be treated in a filter press; this system has not reached industrial use. Lauric oils like palm kernel oil, which on cooling and crystallisation form crystal networks rather than separate crystals in suspension, were traditionally fractionated by artisanal and labour-intensive panning and pressing technology. In this process, crystallisation occurs in trays that are placed in cold rooms. After the crystallisation, the solid blocks are wrapped in filter cloths and loaded into a hydraulic filter press (Rossell, 1985). More recently, a new static crystalliser, named Statoliser, was developed for the crystallisation of these oils. The Statoliser technology combines a two-stage crystallisation (dynamic pre-crystallisation followed by final ‘static’ crystallisation) with high-pressure membrane press filtration (Hendrix and Kellens, 2003). Pre-crystallised oil can also be pumped into a filter press where it is allowed to crystallise further before being pressed (Higuchi et al., 1989). Oils that on crystallisation form crystals in suspension are generally cooled and crystallised in crystalliser vessels fitted with a variable or multi-speed agitator. However, before being cooled, they are generally heated to some 20◦ C above their cloud point to erase any crystal memory. This can be outside the crystalliser in a heat exchanger or in the crystalliser itself while running the agitator at its maximum speed. Subsequent cooling speeds depend upon the crystallisation equipment and reflect the approach of its supplier. To avoid incrustation of the cooling surface, its temperature must be close to the oil temperature, which means low heat transfer. In addition, large vessels provided with cooling coils or fins have a cooling surface of only a few m2 /m3 , so that the rate of cooling is low. However, this slow cooling is claimed to lead to easily filterable crystals (Ricci-Rossi and Deffense, 1984). Increasing the rate of cooling requires more cooling surface per unit of volume as in the concentric and tubular crystallisers (Kellens, 1994). These types of crystallisers have a higher cooling surface (approximately 7 m2 /m3 ). They tend to generate crystals that are less easy to filter. They are almost impossible to filter on a vacuum belt filter but pose no problems in a membrane filter press. Another way to increase the rate of heat transfer is by increasing the heat-transfer coefficient. This is done in the STAR crystalliser (Weber et al., 1998) that is fitted with an eccentrically rotating cooling coil bundle; this attains a heat-transfer coefficient of 300 W/m2 K, which is about double the value observed for agitated vessels fitted with stationary cooling coils.

7.2.4

Separation processes

The first type of filter to be used to separate the stearin from the olein was a rotary vacuum drum filter (Bernardini and Bernardini, 1975) but this was ousted by the successful vacuum band filter as developed by Tirtiaux (1976) because the belt provides more time for the oil to drain away from the cake (Timms, 2005). Even so, this filter had two disadvantages: it can

BLUK122-Dijkstra

186

September 25, 2007

20:2

Trans Fatty Acids Screen bowl Feed

Conveyor

Gearbox Rinse

Mother liquor Torque overload Fig. 7.1

Rinse

Solids

Cross-section of conical sieve centrifuge.

handle only well-developed crystals and even with those, the filter cake still contains between 50 and 70% entrained olein. Accordingly, further development was required to reduce the amount of entrained olein. The first development in this direction was by Maes and Dijkstra (1985), who used a conical sieve centrifuge fitted with a co-rotating scroll (Dijkstra, 1998); Figure 7.1 shows a crosssection of this type of centrifuge, which has also been described by Deffense (2000). It allows the olein content of the filter cake to be reduced to 30% and cocoa butter equivalents (CBE) and cocoa butter replacers (CBR) to be produced but only when the crystals exceed a certain size. If they are too small, they pass through the screen and if they block the holes in the screen, they reduce its permeability with the result that the feed flows over the screen without being filtered. However, with the right kind of crystals, a dry cake results and moreover, the ex-proof version of the equipment offers the possibility of washing the filter cake with a solvent and thereby lowering the olein content even further, as suggested by Timms (2006). Just washing the cake with a solvent would require much less solvent than crystallising from that solvent and would therefore be much cheaper than full solvent fractionation while still yielding comparable results. A separation process that is less sensitive to crystal size and morphology and that, therefore, has been widely adopted uses the membrane filter press (Dijkstra, 2002; Tirtiaux and Gibon, 1996; Willner et al., 1989, 1990; Willner and Weber, 1994). Early presses suffered from high maintenance costs but this problem has been overcome. The pressure has also been increased from 5 to 30 bar, which necessitates the use of hydraulic fluids. Higher squeezing pressures often allow the production of very specific stearin fractions (lower iodine value (IV), higher melting point and steeper solid fat content (SFC) curves) and result in a higher overall olein yield. The use of a nozzle centrifuge for separating the stearin from the olein has also been reported (Wilp, 2000). The process is continuous and involves a closed system, both of

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

187

which are advantages, but the disadvantage of the process is that the nozzles have a fixed discharge capacity for stearin crystals. If these are supplied at a rate that exceeds this discharge capacity, they leave the separator with the olein, and if they are supplied at a lower rate, olein leaves the machine together with the stearin (Dijkstra, 2007). The use of a decanter, on the other hand, does not suffer from this disadvantage. Its use had already been suggested (Maes and Dijkstra, 1985, example 1) and good results were obtained on a pilot scale recently (Deffense, 2005).

7.2.5

Fractionation products

Most early dry fractionation products resulted from a single fractionation process which was rather non-selective because of olein entrainment. Examples are the production of tallow olein by the removal of high-melting triglycerides (M`ege, 1869). Since then, significant progress has been made in the development of efficient crystallisers and high-pressure membrane press filters. In this way, dry fractionation has become a very versatile modification technology for the production of a broad range of food oils and fats with specific physicochemical properties. For beef tallow, improved separation and filtration processes have permitted novel products to be developed. The data provided by Deffense (2001), which involve multi-stage processes, have been summarised in Table 7.1. In this table, S-36 stands for a first stearin fraction that has been obtained by filtration at 36◦ C. Similarly, OS-20 stands for the stearin fraction obtained by fractionating a first olein fraction and filtering at 20◦ C. The SS and SO fractions obtained from the low-IV tallow with melting point 46.2◦ C were fractionated at 48◦ C, and not at 43◦ C like the high-IV tallow. Compound yields of fractions have been indicated in brackets. The table shows that tallow is a variable product and also that the membrane filter press leaves less entrained olein in the filter cake as exemplified by the lower yields and higher melting points of the stearin fractions. Table 7.2 illustrates the effect of the filtration temperature on the melting points, iodine values and yields of beef tallow fractions; a membrane filter press was used to separate the two fractions. As in Table 7.1, the melting points of the olein fraction are close to the filtration temperature and the trends induced by changes in the filtration temperature are well illustrated. When this temperature is lowered, the melting point of the stearin lowers and its yield increases. The olein yield decreases, but its iodine value increases.

Table 7.1

Melting points and yields of beef tallow fractions. Method of filtration Vacuum belt

Vacuum belt

Membrane press

Product

MP (◦ C)

Yield (%)

MP (◦ C)

Yield (%)

MP (◦ C)

Yield (%)

Tallow S-36 O-36 SS-43 SO-43 OS-20 OO-20

42.6 48.7 34.7 53.5 43.0 43.7 19.7

(IV = 51.1) 30 70 35 (10.5) 65 (19.5) 30 (21.0) 70 (49.0)

46.2 51.8 36.2 55.7 47.5 41.8 19.5.

(IV = 45.0) 34 66 33 (11.2) 67 (22.8) 48 (31.7) 52 (34.3)

44.8 52.9 34.0

(IV = 46.0) 24 76

43.1 20.0

27 (20.5) 73 (55.5)

Data from Deffense, 2001.

BLUK122-Dijkstra

188

September 25, 2007

20:2

Trans Fatty Acids Table 7.2

The influence of filtration temperature.

Filtration temperature (◦ C) Stearin

Olein

MP (◦ C) IV Yield (%) MP (◦ C) IV Yield (%)

38

33

28

22

53 30.0 15 38 50.0 85

51 34.5 30 33 53.0 70

50 36.0 40 28 54.5 60

48 38.5 55 22 56.5 45

Data from Deffense, 2001.

The trans isomer content of beef tallow varies between 1.3 and 6.6% (Rossell, 2001). Because these isomers raise the melting point of the triglycerides with respect to oleic acid, it is to be expected that the stearin fractions will have a somewhat higher trans content than the oleins. However, since this trans content is not part of the Codex standard for premier jus, it is not measured on a routine basis. No data on the trans content of the various fractions have been encountered in the literature. Palm kernel oil (IV 17–19) is subjected to a single-stage fractionation to obtain a stearin fraction (IV 7), which is used as cocoa butter substitute (CBS) after hydrogenation (Rossell, 1985). Typical properties of different industrial palm kernel oil fractions are given in Table 7.3. A two-stage dry static fractionation of palm kernel oil has been described by Calliauw et al. (2005). In the first stage, a palm kernel stearin fraction is obtained with IV 5 that can be used as a CBS without further hydrogenation. The corresponding olein fraction is then fractionated again, yielding a second palm kernel stearin (IV 7) that is suitable as a CBS after full hydrogenation. Rather than the higher stearin yield, the reduced hydrogenation capacity is probably the most important benefit of the two-stage process. Of the different oils that are fractionated, by far the largest tonnages are of palm oil. Today, industrial installations exist that fractionate up to 2000 tonnes of palm oil per day. Both crude and refined palm oil are fractionated, the latter being the most predominant. Single-stage fractionation yields an olein fraction with a cloud point <10◦ C. It is used as a substitute for soft oils in frying or cooking, or it is fractionated further. In the latter case, some palm oil may be added to the olein (Maes et al., 1995) to facilitate nucleation and crystallisation, as indicated by the dotted line in Fig. 7.2. Multi-stage fractionation of palm oil is applied to produce high-IV superoleins (IV >65) and soft-palm mid-fractions (S-PMF with IV 45–47). High-IV superoleins combine a low cloud point with good oxidative stability and are therefore useful as frying oil and salad oil. S-PMF is increasingly used as trans-free hardstock in margarines and shortenings. The flexibility of the dry fractionation process for the production of different qualities of superolein and S-PMF fractions has been described by De Greyt et al. (2003). S-PMF can also be used as feedstock for the production of hard-palm mid-fractions (H-PMF). Typical Table 7.3

Typical properties of industrial palm kernel oil fractions.

Parameters Iodine value Clear melting point (◦ C) Yield (%)

Palm kernel oil

Palm kernel stearin

Palm kernel olein

17–19 28

≤7 32 35–40

25–27 24 60–65

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

189

Palm oil (IV 51–53) 40%

60%

Hard stearin (IV 32–36) (vanaspati)

32%

Super stearin (IV 17–21) (animal feed)

First olein (IV 57–59) (deep frying oil)

28%

68%

Soft stearin (IV 40–42)

Soft mid-fraction (IV 42–48) (margarine)

72%

Super olein (IV 64–66) (cooking oil)

35%

Palm mid-stearin IV 32-36 (confectionery) Fig. 7.2

65%

Intermediate oleins

Top olein IV 70–72 (salad oil)

Multi-stage dry fractionation of palm oil. (Adapted from Deffense, 1995.)

H-PMF properties to serve as a good ingredient of CBE include an IV between 34 and 38, a PPP content of maximum 2% and a POP content higher than 60%. The latter two specifications ensure the typical and desired steep melting curve. As indicated in Fig. 7.2 by a dotted line, intermediate olein fractions can be recycled at least partially, so that their POP content is more fully utilised in the mid-fraction production. The fractionation process aims at the production of fractions with a higher value than that of the starting material, but there is always a co-product, which often has a lower value than that of the starting material. Because of their low price, outlets for these co-products have been and are being developed. For the production of a trans-free margarine, a palm stearin obtained by dry fractionation and with an IV of 32–36 was used (Fondu and Willems, 1972). Much later, Sahasramanam (2004) reduced the saturated fatty acid content of a margarine that also contains a liquid oil and an interesterified hardstock by using a palm stearin with a C16 level of >75% and preferably >83%; this is a kind of superstearin obtained by multi-stage fractionation which now commands a premium in this particular application. Although a lot of research is going on in this field, very little information is published on this topic. Know-how for the production of high-added-value speciality fats is available in most of the specialized companies and is obtained through long and laborious (fundamental and empirical) research. Some H-PMF fractions and corresponding mid-olein fractions obtained by dry fractionation of S-PMF on a pilot scale are given in Table 7.4 (De Greyt et al., 2003). Dry fractionation is also used for the modification of other vegetable oils (cottonseed oil, partially hydrogenated soya bean oil, etc.) and animal fat (lard, fish oil, etc.). In the

25

37 3.6 62.5 90.1 81.7 68.3 29.6 7.9

34

b

Applied squeezing pressure: 25 bar. Topped soft PMF. c Not detectable. d Determined by pulse NMR according to IUPAC Method 2.150. e SFC of the slurry just before filtration.

a

Fractionation process data Cooling curve (h) SFC slurrye (%) Slurry temperature (◦ C) Yield (%)

46–47 12.0–12.5

Iodine value Cloud point (◦ C) Melting point (◦ C) PPP (%) POP (%) SFCd at 10◦ C (%) SFC at 20◦ C (%) SFC at 25◦ C (%) SFC at 30◦ C (%) SFC at 35◦ C (%)

Hard PMF

10 20 20 75

N.D.c 40.5 61.5 33.9 0.0

50 6.5

Mid-olein

30

37 2.3 64.0 88.4 79.5 65.9 29.7 3.7

35

Hard PMF

12 22 23 70

N.D. 37.9 60.1 29.3 0.0

51

Mid-olein

Trial 2

35

36.5 2.0 63.5 87.1 76.4 61.9 25.2 1.5

36

Hard PMF

13 25 23 65

N.D. 37.8 57.6 21.8 0.0

51.5

Mid-olein

Trial 3

40

36.5 1.8 62.5 88.4 77.6 54.4 32.2 1.7

37

Hard PMF

13 30 20 60

N.D. 36.4 59.6 27.6 0.0

52.5 6.4

Mid-olein

Trial 4

September 25, 2007

0.7–0.8 45

Feedstockb

Parameters

Trial 1

Production of hard PMF and mid-olein fractions by dry fractionation of soft PMF.a

190

Table 7.4

BLUK122-Dijkstra 20:2

Trans Fatty Acids

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification Table 7.5

191

Dry fractionation of tuna oil. Trial 1

Parameter Slurry temperature SFC slurry Yield Cold test 0◦ C Cloud point IV EPA DHA

Tuna oil

Olein

(◦ C) (%) (%) (h) (◦ C) (%) (%)

Stearin

Trial 2 Olein

2.5 14.3 100 14.2 180.6 4.8 23.0

68.0 >4.5 −2.0 188.0 5.1 26.1

Stearin

Trial 3 Olein

2.4 32.0

164.0 3.5 17.6

69.9 >5.0 −1.9 188.0 4.9 25.3

Stearin 2.3

30.1

163.1 3.6 18.3

64.4 >5.5 −1.8 190.0 5.2 26.1

35.6

163.5 3.6 19.0

case of fish oil, the main objective of the dry fractionation is actually the production of a cold-stable olein fraction at the highest possible yield. An important and desired side effect is the potential increase of the omega-3 fatty acids (like eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)) in the fish olein. This increase originates from the fact that triglycerides containing polyunsaturated omega-3 fatty acid moieties hardly crystallise. Their relative increase in the olein fraction is therefore limited and in relative terms, cannot be higher than the SFC of the slurry at filtration. As an illustration, fractionation of tuna oil (SFC slurry: 14% and 27.8% EPA+DHA) results in an olein fraction with good cold stability (cold test at 0◦ C: 5 h) and a slightly increased EPA+DHA content (30–31%, Table 7.5) (W.F.J. De Greyt, internal communication)

7.3 7.3.1

INTERESTERIFICATON Historical

During the interesterification process, the fatty acid moieties that form part of the triglyceride molecules are ‘reshuffled’ over the glycerol moieties. Contrary to the fractionation process, interesterification is a non-reversible process resulting in one ‘randomised’ product with a fatty acid composition that is identical to that of the starting materials and with no formation of trans isomers. However, the triglyceride composition, as for instance illustrated by the carbon-number distribution, changes on interesterification. This also affects the physical properties of the product. Interesterification, or ester interchange, can be induced thermally, as observed by Friedel and Crafts in 1865. They heated a mixture of ethyl benzoate and amyl acetate to 300◦ C in a sealed tube and noted the formation of amyl benzoate and ethyl acetate. For edible oils and fats, originally two different two-step processes were investigated. In the first one (Gr¨un, 1922), the oil was allowed to react with glycerol to form partial glycerides, which were subsequently esterified with free fatty acids (FFA). The second process (Normann, 1924) was the opposite in that the first step entailed a partial acidolysis with FFA, which was then followed by the esterification of the FFA with glycerol. True ester interchange between triglyceride molecules was developed only later by Van Loon (1926), who used a variety of catalysts: heavy metals and their salts or hydroxides, alkali or alkaline earth compounds, organic sulphonic acids etc. He employed elevated temperatures of some 200◦ C even when using sodium ethylate as the catalyst. Subsequently, Eckey (1945b) discovered that heating a fat with some water to 235–240◦ C and subsequently flashing off

BLUK122-Dijkstra

192

September 25, 2007

20:2

Trans Fatty Acids

this water also led to interesterification. Sodium methylate was also used at high temperatures (Eckey, 1945a), but soon after it was found (Eckey, 1948) that sodium methylate was also an active interesterification catalyst at low temperatures, i.e. just above the melting point of the reaction mixture to be interesterified. The simultaneous development of alkali-catalysed, low-temperature alcoholysis processes (Bradshaw and Meuley, 1942) may have acted as an inspiration for the low-temperature interesterification process. Subsequently, a variety of low-temperature alkaline catalysts were reported: alkali metals (Eckey, 1948), sodium hydride (Eckey, 1951), sodium amide (Nelson and Mattil, 1953), a sodium–potassium alloy (Hawley and Dobson, 1956) and for a high-temperature interesterification, the condensation product of glycerol and sodium hydroxide was found to have catalytic activity (Burgers et al., 1965). Caustic soda is used in a batch process (Keulemans and Rozendaal, 1984) or a continuous interesterification process (Keulemans and Smits, 1986). These processes involve heating a dispersion in oil of a (50%) solution of caustic soda in glycerol and eliminating all water from this dispersion by spraying it in vacuo. To effect this drying, a fairly high temperature (125–140◦ C) is needed. Jakubowski (1971) uses a slightly different procedure. He adds the caustic glycerine at a temperature below 80◦ C and evacuates. When all water has evaporated, he heats it to 140◦ C, at which temperature the interesterification is complete. Originally, the interesterification reaction has also been used to replace low-boiling fatty acids in coconut oil by longer chain fatty acids (Barsky, 1939; Eckey, 1945c) and obtain products with improved consistency. This can be regarded as a kind of directed interesterification, but this term is more commonly used for the process whereby the highest melting triglycerides are withdrawn from the reaction mixture by fractional crystallisation (Eckey, 1948). This crystallisation disturbs the randomisation equilibrium with the result that more high-melting triglycerides are formed. This directed interesterification process has been used to elongate the plastic range of lard (Hawley and Holman, 1956) and to provide a margarine fat based upon sunflower seed oil with a low (<14%) saturated fatty acid content and a high (>66%) polyunsaturated fatty acid content (De Lathauwer et al., 1980). However, these processes are no longer in use and the only chemically catalysed interesterification processes in current industrial use are randomisation processes, using either a low-temperature sodium alkanolate catalyst or a high-temperature caustic soda condensation catalyst. In addition, the enzymatic catalysis of the interesterification reaction is also used industrially. Originally, this type of catalyst (extracellular microbial lipase, EC 3.1.1.3) was limited to processes where the 1,3-specificity of the ester interchange was essential, such as the production of CBE with a high content of symmetrical monounsaturated triglycerides (Coleman and Macrae, 1977; Macrae, 1983; Matsuo et al., 1981). Subsequently, more general-purpose continuous enzymatic interesterification processes have been developed for the production of margarine and shortening hardstocks. For these applications, enzymatic interesterification also gives rise to fully randomised products with physicochemical properties that are almost identical to the products obtained by chemical interesterification (Fig. 7.3). Enzymatically interesterified fats generally contain less partial glycerides and have a higher natural tocopherol content and lighter colour than chemically interesterified fats (De Greyt, 2005).

7.3.2

Interesterification mechanism

When discussing the interesterification process, Eckey (1956) suggested that the actual catalytic material presumably consists of anions of some sort formed in the fat by the added

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

193

50 45 40 Feedstock

SFC (%)

35 EIE FAT

30 25

CIE FAT

20 15

FH Soya bean oil/ Soya bean oil

10 5 0 0

10

20

30

40

50

60

Temperature (°C) Fig. 7.3 Solid fat content profile of a chemical (CIE) and enzymatic (EIE) fat blend consisting of fully hydrogenated soya bean oil/soya bean oil (27/73).

alkaline precursor. Subsequently, Baltes (1960) suggested that this anion is the glycerolate anion, and this has been generally accepted for quite some time. However, in 2004, Liu reported that -substituted fatty acids do not interchange in the presence of sodium methylate and that apparently, an active -hydrogen is an essential prerequisite for the reaction to proceed; this requirement does not follow from the ‘glycerolate’ mechanism. More anomalies of the ‘glycerolate’ mechanism were reported and a new mechanism, to be referred to as the ‘enolate’ mechanism, was proposed (Dijkstra, 2004); independent support for this mechanism has since been provided (Dijkstra et al., 2005). According to this mechanism, the actual catalytic intermediate is an enolate anion that is formed either directly by the action of a methylate anion added as sodium methylate according to:

O H O

C

O H R

O

O H R

O

R

H

H

O

CH3

O CH3

or indirectly, when a glycerolate anion abstracts a hydrogen ion according to: O H O

C H O

R

O O

H

O R

O

H R +

H O

BLUK122-Dijkstra

194

September 25, 2007

20:2

Trans Fatty Acids

This glycerolate anion can originate from various reactions, such as the initiation reactions of an alkali metal with a partial glyceride under evolution of hydrogen (Amat Guerri and Cosme Jim´enez, 1974) or the abstraction of a hydrogen from a partial glyceride by a methanolate anion and from the propagation reactions of the enolate anion with a hydroxyl compound like methanol or a partial glyceride. The latter reaction constitutes the interesterification according to: R

R

H

C

O

CH O

H

H O

O

R

C

H

H

C

C

O

O

C

O

H O

O

In this reaction, the enolate anion reacts with a partial glyceride with the result that the fatty acid moiety in the enolate anion shifts towards the partial glyceride and generates a glycerolate anion, which can then react with a glyceride to regenerate the enolate anion. If the enolate anion were to react with methanol, according to: R H C

R H

O CH O CH3

R

H C

H

H C

O C

O CH3

O C

O

O

H O CH3

O

a glycerolate anion is also formed, which again can regenerate the enolate anion. At the same time, a fatty acid methyl ester (FAME) molecule is formed so that the enolate mechanism satisfactorily explains how free methanol is quantitatively converted into FAME. If the enolate anion reacts with water according to: R H

C

R H

O CH O H O

H C O C O

R

R H O H

H C

H

H C

H

O C

OH

O C

O

O

OH

an FFA moiety is inevitably formed. In this respect, the enolate mechanism provides a logical explanation for the yield loss observed for the sodium-methanolate–catalysed interesterification process, an aspect which the formerly suggested mechanisms fail to explain. In accordance with experimental observations, such as loss of catalytic activity, when the reaction temperature is increased (Dijkstra et al., 2005), the enolate anion is assumed to be

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

195

thermally unstable and react to form a -keto ester anion according to: O O C

C R

O C

H . O CO CH2.R2

C

O C

R1

C

C

CH2.R2

H

H O C CH .R2 2 O O CO.CH .R3

CO.CH2.R3

O

O R1 O

O 1

O O CO.CH2.R3

2

O R1 O O C

C

CH2.R2

OH O

CO.CH2.R3

This -keto ester anion has been observed to be formed gradually (Heldal and Mørk, 1981) and this observation has been used as argument against its being the catalytically active material. However, it may well be that at quite high temperatures (for instance, above 150◦ C), it has regained some activity and is therefore responsible for the very slow interesterification observed under those conditions (Naudet, 1947). The possibility should also be considered that the process involving the high-temperature caustic soda condensation catalyst (Rozendaal, 1990) ultimately leads to this -keto ester anion as catalytically active material. With respect to the enzyme-catalysed interesterification, it has been assumed that the first step is a partial hydrolysis of a triglyceride molecule by water that is present in/on the enzyme (Kreye et al., 1994). This step involves the formation of a lipase/triglyceride complex followed by hydrolysis that generates a lipase/FFA complex and a free diglyceride molecule. If another free diglyceride molecule then joins the complex, the reverse reaction leads to ester interchange (Luck and Bauer, 1991). Under hydrolysis circumstances, when water is available ad libitum, the lipase/FFA complex can also dissociate in unoccupied lipase and an FFA. If this unoccupied lipase then takes up a further water molecule and combines with a diglyceride, this may be hydrolysed to form a monoglyceride. Accordingly, some water is necessary for the enzyme to be active and effect interesterification, but more water leads to hydrolysis rather than to interesterification. Moreover, partial glycerides tend to isomerise, whereby the esterification of the 1- or 3-position is favoured over the 2-position to the extent of some 4 kJ/mol (De Groot, 1974; Yang et al., 2004). +S O O S

O S S

−O

O O S

O S

S +O

O S O

BLUK122-Dijkstra

196

September 25, 2007

20:2

Trans Fatty Acids

As illustrated above, the 1,2- and 2,3-diglycerides formed will isomerise to 1,3diglycerides, which can be further hydrolysed to form monoglycerides which can isomerise and then be esterified to yield triglycerides. However, these triglycerides show a loss of 1,3specificity of the interesterification, and the inevitable presence of partial glycerides affects the crystallisation behaviour of CBE.

7.3.3

Interesterification practice

In the past, chemical interesterification was regarded as a risky process due to the use of quite unstable catalysts (e.g. Na–K alloys). In addition, it was also characterised by its high oil losses (up to 5%). For this reason, chemical interesterification was largely replaced by partial hydrogenation for the production of margarine fats and shortenings. At that time, enzymatic interesterification was suitable for the production of only high-added-value fats (CBE, medium-chain triglycerides and structured lipids) because the available lipases were very expensive and had a poor stability and low activity. Today, this situation has totally changed. The increasing trend to exclude trans fatty acids from food fats and the increasing demand for low- and zero-trans fats combined with improvements to the chemical interesterification process has meant that this process has replaced partial hydrogenation almost totally. Cheaper, more stable and more active enzymes have also become available, which makes enzymatic interesterification a cost-effective process for the production of commodity food fats. In this section, only randomisation processes will be discussed because hardly any chemically catalysed directed interesterification is performed industrially today, and enzymecatalysed, 1,3-specific interesterification has lost importance now that CBE produced by that route are excluded from confectionery products in Europe if these are still to be called chocolate (Stewart and Kristott, 2004). Besides, these latter processes are predominantly proprietary.

7.3.3.1

Chemical interesterification

In the chemically catalysed interesterification process, batch processes normally employ sodium methylate or ethylate as the catalyst. This catalyst is inactivated by water, FFA or peroxides so that a dry and neutral feedstock is mandatory. This can be assured by adding some caustic soda to the raw material after it has been heated to reaction temperature (80– 100◦ C) and removing the water by evacuating the agitated batch (Laning, 1985). Provided the batch is sufficiently dry and neutral, an amount of 0.05% by weight of sodium methylate suffices to ensure complete randomisation in a few minutes. To be on the safe side, a reaction time of some 30 min is usually provided. After the interesterification equilibrium has been reached, the catalyst is inactivated by the addition of either water or an aqueous solution of citric acid. Water addition leads to the formation of soaps, whereas the addition of acidified water leads to FFA formation. Soaps can be removed by water washing followed if necessary by a treatment with silica hydrogel, whereas FFA are removed during the subsequent deodorisation step, which is required anyway to remove the FAME formed after catalyst addition. Bleaching the interesterification product before deodorisation is recommended since the activation of the catalyst leads to a marked colour development. According to Liu and Lampert (2001), this development can be used to control the extent of interesterification but the merits of this control have been shown to

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

197

be spurious, since partially interesterified products can be more cheaply produced by mixing fully randomised products with their raw materials (Dijkstra, 2000). The main drawback of the chemically catalysed interesterification process is its yield loss, which is directly proportional to the amount of catalyst used. The addition of sodium methylate (MW = 54) leads to the formation of an equivalent amount of FAME (MW=∼295), and catalyst inactivation leads to a further formation of an equivalent amount of FFA (MW=∼280). In addition, neutral oil is lost on removing the soaps or FFA. Accordingly, the use of 0.1% by weight of sodium methylate will lead to the formation of approximately 1.0% FAME and FFA. In that case, total oil losses (including losses during post-bleaching and deodorisation) can rise up to 1.5%. If sodium methylate consumption can be limited to 0.05%, overall oil losses can be reduced to 0.8% (Kellens and De Greyt, 2005). Apart from the oil losses, there are other reasons to limit the amount of catalyst to maximum 0.1%. Higher catalyst concentrations will result in too high diglyceride levels and can give formation of unwanted side products, like ketones (Verh´e et al., 2006). Even at half the catalyst dosage, yield loss is responsible for about one-third of the total variable cost (Kellens, 2000). An increase of the diglyceride content is also observed when the catalytically active intermediate is prepared by the condensation of glycerol and sodium hydroxide (Keulemans and Smits, 1986). How much sodium hydroxide is used in industrial practice has not been published and the patent literature mentions only broad ranges (0.03–0.15% by weight) for the amount of sodium hydroxide to be used. This would correspond to a yield loss of 0.2–1.0% if the neutral oil loss is not taken into account. 7.3.3.2

Enzymatic interesterification

Apart from the 1,3-specific enzyme-catalysed interesterification that is used for CBE and infant formulae (Kavanagh, 1997; Quinlan and Moore, 1993), the application that is growing in importance is the enzyme-catalysed randomisation of fat blends to provide trans-free hardstock (Cowan and Husum, 2004). Unlike the chemical interesterification process, the enzymatic interesterification process currently carried out on industrial scale is continuous. A pre-treated blend is pumped through a number of packed bed reactors (usually three or four) placed in series and kept at a temperature of around 70◦ C. A typical flow rate is 1–2 kg oil per kg enzyme per hour. As a post-treatment, enzymatically interesterified oil does not need bleaching, but only a mild deodorisation to remove some FFA and off-flavours (De Greyt, 2004). Heat-stable lipases of microbial origin are used as the biocatalyst and to improve their stability, they have been immobilized on a silica support (Cowan and Husum, 2004). The most critical factor in enzymatic interesterification is the enzyme productivity. In general, enzymes are quite sensitive to external factors like pH and temperature and also to other unknown factors that may cause a sudden loss of activity (Diks, 2002). If freshly deodorised oil blends are used as raw material, 2.5–4.0 tonnes of oil can be interesterified with 1 kg of immobilized enzyme preparation. To obtain this productivity at the typical flow rate, the enzyme has to remain active for 2500–4000 h. With this enzyme productivity and based on today’s enzyme price, the enzyme cost will vary around 25 US dollars per tonne of oil. If an amount of sodium methylate catalyst equal to 0.1% is used in the chemical interesterification process, the total costs of both interesterification processes are about equal (Kellens and De Greyt, 2005). Using less chemical catalyst, e.g. 0.5 wt%, is often possible and saves on catalyst cost and especially on oil loss. On the other hand, enzyme productivity may also improve.

BLUK122-Dijkstra

198

September 25, 2007

20:2

Trans Fatty Acids

Moreover, the raw material for enzymatic interesterification requires much more extensive purification than is necessary for chemical interesterification. The latter only demands that the feedstock is dry and neutral, and this can be assured in situ by the addition of some caustic followed by drying. Oil to be interesterified enzymatically must be neutral and also bleached and freshly deodorised to prevent early enzyme inactivation; this may entail a second deodorisation treatment just prior to the interesterification. After the reaction, the chemically interesterified product still needs to be bleached and deodorised, whereas enzymatically interesterified oil may only require a final deodorisation step. When comparing the two processes, these cost aspects should also be taken into account.

7.3.4

Interesterification products

In the US, the main interesterification product has for a long time been lard based. It could be randomised (Vander Wal and Van Akkeren, 1951) and its properties could be further improved by the addition of some fully hydrogenated fat to the randomised lard (Going, 1967). It could also be subjected to a directed interesterification process (Hawley and Holman, 1956). These products improved aeration and cake volume when used as shortening. Soon after, hard butters (Cochran and Ott, 1957) and hardstocks (Babayan, 1961) were developed by interesterifying mixtures of lauric and non-lauric fats. In Europe, this approach was pursued by Graffelman (1971) and Delfosse (1971), who used palm kernel oil and coconut oil as lauric fat respectively and fully hydrogenated the product. Their hardstocks could therefore be used in health margarines with a high polyunsaturated fatty acid content and a low saturated fatty acid content by mixing these hardstocks with, for instance, sunflower seed oil. The use of babassu oil (Ward, 1982) and a high-lauric rapeseed oil (Sassen and Wesdorp, 2001) was described later. The use of a stearin fraction of such hardstocks (Galenkamp, 1965) has been disclosed even earlier. These margarines were virtually free from trans fatty acids but hydrogenation was still involved in their production. Accordingly, Fondu and Willems describe a hardstock (1972) that is made by interesterifying palm stearin and a lauric oil. Subsequently, a stearin fraction of the interesterification of palm kernel oil stearin and palm stearin is disclosed (Huizinga et al., 1999). To prevent margarines with an oil phase consisting of liquid sunflower seed oil and partially hydrogenated sunflower seed oil from developing sandiness, the partially hydrogenated sunflower seed oil can be interesterified on its own or with a small amount of liquid oil (Gander et al., 1966). Similarly, partially hydrogenated canola, which is even more prone to recrystallisation into the sandy polymorph, can be interesterified with a fat rich in palmitic acid, such as palm oil or its fractions (Gerschel and Helme, 1986). However, these margarines contain trans fatty acids; interesterifying palm stearin with fully hydrogenated high-erucicacid rapeseed oil (Lansbergen and Schijf, 1996) is claimed to provide a hardstock that allows trans-isomer-free margarines to be made that also have a relatively low content of saturated fatty acids.

REFERENCES Amat Guerri, F. & Cosme Jim´enez, J.L. (1974) Estudios sobre transesterificaci´on V. Hidr´ogeno producido en la reacci´on NaK-grasa. Grasas y Aceites 25 (1), 6–9. ¨ Anonymous (1907) Verfahren zur Zerlegung des aus Fetten und fetten Olen gewonnenen Fetts¨auregemisches ¨ aure und feste Fetts¨auren. German Patent 191238, assigned to Fratelli Lanza. in Ols¨

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

199

Babayan, V.K. (1961) Fat composition for margarine and other purposes. US Patent 3,006,771, assigned to I.E.F. Drew & Co. Baltes, J. (1960) Reaktionsmechanismus und Kinetik der Umesterung von Fetten. Die Nahrung 4 (1), 1–16. Barsky, G. (1939) Method of reconstructing glyceride oils. US Patent 2,182,332, assigned to Wecoline Products Inc. Bernardini, E. & Bernardini, M. (1975) Palm oil fractionation and refining using the C.M.B. process. Ol´eagineux 30 (3), 121–128. Bradshaw, G.B. & Meuley, W.C. (1942) Process of making pure soaps. US Patent 2,271,619, assigned to E.I. du Pont de Nemours and Company. Burgers, J., Mott, C.W. & Seiden, P. (1965) Interesterification process. US Patent 3,170,798, assigned to the Procter & Gamble Company. Calliauw, G., Foubert, I., De Greyt, W.F.J. et al. (2005) Production of cocoa butter substitutes via two-stage static fractionation of palm kernel oil. J Am Oil Chem Soc 82 (11), 783–789. Cochran, W.M. & Ott, M.L. (1957) Process for preparing hard butter. US Patent 2,783,151, assigned to the Glidden Company. Coenen, J.W.E. (1974) Fractionnement et interest´erification des corps gras dans la perspective du march´e mondial des mati`eres premi`eres et des produits finis I – fractionnement. Rev franc¸ Corps Gras 21 (6), 343–349. Coleman, M.H. & Macrae, A.R. (1977) Composition de mati`eres grasse et proc´ed´e pour sa fabrication. French Patent 2 340 979, assigned to Unilever. Cowan, W.D. & Husum, T.L. (2004) Enzymatic interesterification: process advantage and product benefits. Inform 15 (3), 150–151. de Groot, W.T. (1974) Isomerization of 1,2-diglycerides to 1,3-diglycerides. US Patent 3,845,087, assigned to Unilever. De Greyt, W.F.J. (2004) Chemical and enzymatic interesterification. Practical Short Course on Edible Oil Refining and Processing. Edinburgh. De Greyt, W.F.J. (2005) Enzymatic interesterification for the production of margarine fats. In: Enzymes in Lipid Technology. Forum organised by Lipidforum, Copenhagen. De Greyt, W.F.J., Kellens, M.J. & Hendrix, M. (2003) New developments in the dry fractionation of palm and palm kernel oil. In: Proceedings of the PIPOC 2003 Chemistry and Technology Conference, Putrajaya, Malaysia. De Lathauwer, R., Van Opstal, M. & Dijkstra, A.J. (1980) Process for the directed interesterification of a triglyceride oil or oil mixture. European Patent 0 009 207, assigned to N.V. Vandemoortele International. Deffense, E.M.J. (March 1995) Dry multiple fractionation: trends in products and applications. Lipid Technol 7, 34–38. Deffense, E.M.J. (1998) Dry fractionation and selectivity. Ol´eagineux Corps Gras Lipides 5 (5), 391– 395. Deffense, E.M.J. (1999) Method for producing microcrystals of vegetable and animal fats. PCT Patent Application WO 99/42545, assigned to Crystallisation & Degumming sprl. Deffense, E.M.J. (2000) Dry fractionation technology in 2000. Eur J Lipid Sci Technol 1 (3), 234–236. Deffense, E.M.J. (2001) Fractionation of animal carcass fats into hard and soft fats. In: Animal Carcass Fats (ed. J.B. Rossell). Leatherhead Publishing, Leatherhead, pp. 149–173. Deffense, E.M.J. (2005) Fractionation of fish, milk fat and other oils. Paper presented at SCI Symposium: Fractionation – current status and future prospects in a low-trans world, Ghent. www.soci.org. Delfosse, J.K.F. (1971) Margarine fat composition. British Patent 1 244 868, assigned to Unilever. Dijkstra, A.J. (1998) Alternativas a la hidrogenaci´on. Revista Aceites y Grasas 8 (32), 356–357. Dijkstra, A.J. (2000) Triglyceride composition of a partially randomized reaction mixture. J Am Oil Chem Soc 77 (3), 333. Dijkstra, A.J. (2002) Hydrogenation and fractionation. In: Fats in Food Technology (ed. K.K. Rajah). Sheffield Academic Press, Sheffield, pp. 123–158. Dijkstra, A.J. (2004) The interesterification mechanism revisited. Paper presented at the 3rd EuroFed Lipid Congress, Edinburgh, p. 86 in the book of abstracts. Dijkstra, A.J. (2007) Fractionation. In: The Lipid Handbook, 3rd edn (eds F.D. Gunstone, J.L. Harwood & A.J. Dijkstra). Taylor & Francis Group, LLC, Boca Raton, FL, pp. 302–317. Dijkstra, A.J., T˜oke, E.R., Kolonits, P. et al. (2005) The base-catalyzed, low-temperature interesterification mechanism revisited. Eur J Lipid Sci Technol 107 (12), 912–921. Diks, R.M.M. (2002) Lipase stability in oil. Lipid Technol 14, 10–14.

BLUK122-Dijkstra

200

September 25, 2007

20:2

Trans Fatty Acids

Eckey, E.W. (1945a) Process for modifying the molecular structure of oils and fats. US Patent 2,378,007, assigned to the Procter & Gamble Company. Eckey, E.W. (1945b) Process for treating fats and fatty oils. US Patent 2,378,005, assigned to the Procter & Gamble Company. Eckey, E.W. (1945c) Process for treating fats and fatty oils. US Patent 2,378,006, assigned to the Procter & Gamble Company. Eckey, E.W. (1948) Process for treating fats and fatty oils. US Patent 2,442,531, assigned to the Procter & Gamble Company. Eckey, E.W. (1951) Catalytic processes for ester-ester interchange. US Patent 2,558,547. Eckey, E.W. (1956) Esterification and Interesterification. J Am Oil Chem Soc 33 (11), 575–579. Fondu, M.P.V. & Willems, M.A.G. (1972) Margarine fat containing randomized fat component. US Patent 3,634,100, assigned to Unilever. Foubert, I. (2007) Crystallisation and melting. In: The Lipids Handbook, 3rd edn (eds F.D. Gunstone, J.L. Harwood & A.J. Dijkstra). Taylor & Francis Group, LLC, Boca Raton, FL, pp. 476–495. Friedel, Ch. & Crafts, J.R. (1865) Ann., 133, 207–211. Galenkamp, H. (1965) Process of preparing a fat product which after plasticizing can be used as a spreading, baking and frying fat, and a process of preparing a margarine using this fat product. US Patent 3,210,197, assigned to N.V. Verenigde Textiel- & Oliefabrieken Afd. Crok & Laan. Gander, K.F., Hannewijk, J. & Haighton, A.J. (1966) Proc´ed´e de fabrication de margarine obtenue principalement a` partir de l’huile de tournesol. French Patent 1.457.751, assigned to Unilever. Gerschel, D. & Helme, J.-P. (1986) Phase grasse pour produits alimentaires, proc´ed´e pour sa pr´eparation et produits alimentaires contenant ladite phase grasse. French Patent 2 570 388, Danial Gorschol applied for the patent. Going, L.H. (1967) Interesterification products and processes. J Am Oil Chem Soc 44 (9), 414A– 456A. Graffelman, H.A. (1971) Margarine fat and process for preparing same. US Patent 3,617,308, assigned to Unilever. Gr¨un, A. (1922) Improvements in the manufacture of nutritious fats. British Patent 160,840, assigned to A.G. Georg Schicht. Hamm, W. (1986) Fractionation – with or without solvents? Fette Seifen Anstrichmittel 88, 533– 537. Hamm, W. (1995) Trends in edible oil fractionation. Trends Food Sci Technol 6, 121–126. Hamm, W. & Timms, R.E. (2006) Dry fractionation of oils and fats: entrainment – are we making progress? Lipid Technol 18 (1), 11–15. Harris, J. (2005) Why solvent fractionation? Paper presented at SCI Symposium: Fractionation – Current status and future prospects in a low-trans world, Ghent. www.soci.org. Hawley, H.K. & Dobson, R.D. (1956) Molecular rearrangement process. US Patent 2,733,251, assigned to the Procter & Gamble Company. Hawley, H.K. & Holman, G.W. (1956) Directed interesterification as a new processing tool for lard. J Am Oil Chem Soc 33 (1), 29–35. Heldal, J.A. & Mørk, P.C. (1981) On the mechanism of the base catalyzed interesterification of fats. Paper presented at the 11th Scandinavian Symposium on Lipids, Bergen, Norway, pp. 147–151. Hendrix, M. & Kellens, M.J. (2003) Process and installation for dry fractionation. European Patent 1 281 749 A1, assigned to De Smet Engineering N.V. Higuchi, K., Maeda, H., Iwasa, H. et al. (1989) Method for dry fractionation of fatty material. US Patent 4,795,569, assigned to Fuji Oil Company Ltd. Holemans, P.M.J., Schijf, R., van Putte, K.P.A.M. & de Man, T. (1988) Fat and edible emulsions with a high content of cis-polyunsaturated fatty acids. US Patent 4,791,000, assigned to Unilever. Huizinga, H., Sassen, C.L., Vermaas, L.F. & Schur, P. (1999) Process for making a margarine hardstock. US Patent 5,858,445, assigned to Unilever. Jakubowski, A. (1971) L’interest´erification entre corps gras animaux et huiles v´eg´etales. Rev franc¸ Corps Gras 18 (7), 429–437. Kavanagh, A.R. (1997) A breakthrough in infant formula fats. Ol´eagineux Corps Gras Lipides 4 (3), 165–168. Kellens, M.J. (1994) Development in fat fractionation technology. Paper no 0042 presented at the SCI Symposium on Fractional Crystallisation of Fats. Kellens, M.J. (2000) Oil modification processes. In: Edible Oil Processing (eds W. Hamm & R.J. Hamilton). Sheffield Academic Press, Sheffield, pp. 127–173.

BLUK122-Dijkstra

September 25, 2007

20:2

Fractionation and interesterification

201

Kellens, M.J. & De Greyt, W.F.J. (2005) Chemical and enzymatic interesterification. In: Short Course Refining and Modification. XIth Latin American Congress on Fats and Oils, Rosario, Argentina. Keulemans, C.N.M. & Rozendaal, A. (1984) Process and apparatus for the interesterification of a triglyceride oil and products therefrom. European Patent 0 121 440, assigned to Unilever. Keulemans, C.N.M. & Smits, G. (1986) Interesterification process and apparatus. US Patent 4,585,593, assigned to Unilever. Kloek, W. (1998) Mechanical properties of fats in their relation to their crystallisation. Ph.D. Thesis. Landbouwuniversiteit Wageningen, The Netherlands. Kreye, L., Herar, A., Bornscheuer, U.T. & Scheper, Th. (1994) Enzymatische Umsetzungen in organischen L¨osungsmitteln – diskontinuierliche und kontinuierliche Systeme. Fett Wissenschaft Technologie 96 (7), 246–251. Laning, S.J. (1985) Chemical interesterification of palm, palm kernel and coconut oils. J Am Oil Chem Soc 62 (2), 400–405. Lansbergen, G.J.T. & Schijf, R. (1996) Edible fats. US Patent 5,547,698, assigned to Unilever. Liu, L. (2004) How is chemical interesterification initiated: nucleophilic substitution or -proton abstraction? J Am Oil Chem Soc 81 (4), 331–337. Liu, L. & Lampert, D.S. (2001) Partial interesterification of triacylglycerols. US Patent 6,238,926 B1, assigned to Cargill Incorporated. Luck, T. & Bauer, W. (1991) Lipasekatalysierte Triglyceridumesterung in einem l¨osungsmittelfreien Prozess I: Analytik und Umesterungskinetik. Fett Wissenschaft Technologie 93 (2), 41–49. Macrae, A.R. (1983) Lipase-catalysed interesterification of oils and fats. J Am Oil Chem Soc 60 (2), 291– 294. Maes, P.J. & Dijkstra, A.J. (1985) Process for separating solids from oils. US Patent 4,542,036, assigned to N.V. Vandemoortele International. Maes, P.J., Dijkstra, A.J. & Seynaeve, P. (1995) Method for dry fractionation of fatty substances. European Patent 0 651 046, assigned to N.V. Vandemoortele International. Matsuo, T., Sawamura, N., Hashimoto, Y. & Hashida, W. (1981) Method for enzymatic transesterification of lipid and enzyme used therein. European Patent 0 035 883, assigned to Fuji Oil Company Ltd. M`ege, H. (1869) Production de certains corps gras d’origine animale. French Patent 86 480. Naudet, M. (1947) Contribution a` l’´etude des migrations d’acyles entre triglycerides. Ph.D. Thesis. Universit´e de Marseilles, France. Nelson, D.W. & Mattil, K.F. (1953) Modification of lard. US Patent 2,625,487, assigned to Swift & Company. Normann, W. (1924) Verfahren zur Herstellung von gemischten Glyceriden. German Patent 407180. Padley, F.B. (1997) Chocolate and confectionary fats. In: Lipid Technologies and Applications (eds F.D. Gunstone & F.B. Padley). Marcel Dekker, New York, pp. 391–432. Peters-Erjawetz, S., Ulrich, J., Tiedtke, M. & Hartel, R.W. (1999) Milk fat fractionation by solid-layer melt crystallisation. J Am Oil Chem Soc 76 (5), 579–584. Podmore, J. (2002) Culinary fats: solid and liquid frying oils and speciality oils. In: Fats in Food Technology (ed. K.K. Rajah). Sheffield Academic Press, Sheffield, pp. 333–359. Poot, C. & Biernoth, G. (1994) Margarine and butter production. In: The Lipid Handbook, 2nd edn (eds F.D. Gunstone, J.L. Harwood & F.B. Padley). Chapman & Hall, London, pp. 288–295. Quinlan, P. & Moore, S. (1993) Modification of triglycerides by lipases: process technology and its application to the production of nutritionally improved fats. Inform 4 (5), 580–585. Ricci-Rossi, G. & Deffense, E.M.J. (1984) Erfahrungen mit der Fraktionierung von Fetten nach dem TirtiauxVerfahren. Fette Seifen Anstrichmittel, 86 (Sonderheft), 500–505. Robinson, D.J. & Rajah, K.K. (2002) Spreadable products. In: Fats in Food Technology (ed. K.K. Rajah). Sheffield Academic Press, Sheffield, pp. 192–227. Rossell, J.B. (1985) Fractionation of lauric oils. J Am Oil Chem Soc 62 (2), 385–390. Rossell, J.B. (2001) Origins and chemical properties. In: Animal Carcass Fats (ed. J.B. Rossell). Leatherhead Publishing, Leatherhead, UK, pp. 1–31. Rozendaal, A. (1990) Interesterification of oils and fats. In: Edible Fats and Oils Processing: Basic Principles and Modern Practices (ed. D.R. Erickson). AOCS Press, Champaign, IL, pp. 152–157. Sassen, C.L. & Wesdorp, L.H. (2001) Edible fat spread. US Patent 6,238,723, assigned to Unilever. Schijf, R., Trommelen, A.M. & Lansbergen, G.J.T. (1984) Margarine fat blend, and a process for producing said fat blend. US Patent 4,486,457, assigned to Unilever. Seug´e, J.P. & Vinconneau, H.F. (1975) Le proc´ed´e Lipofrac Alfa-Laval pour le fractionnement et la frig´elisation en continu des graisses comestibles. Ol´eagineux 30 (1), 25–30.

BLUK122-Dijkstra

202

September 25, 2007

20:2

Trans Fatty Acids

Smith, K.W. (2001) Crystallisation of palm oil and its fractions. In: Crystallization processes in Fats and Lipid Systems (eds N. Garti & K. Sato). Marcel Dekker Inc, New York, pp. 357–380. Smith, P.R. (2000) The effects of phospholipids on crystallisation and crystal habit in triglycerides. Eur J Lipid Sci Technol 1 (2), 122–127. Stewart, I.M. & Kristott, J. (2004) European Union Chocolate Directive defines vegetable fats for chocolate. Lipid Technol 16 (1), 11–14. Timms, R.E. (1991) Crystallisation of fats. Chem Ind 10, 342–345. Timms, R.E. (1997) Fractionation. In: Lipid Technologies and Applications (eds F.D. Gunstone & F.B. Padley). Marcel Dekker, New York, pp. 199–222. Timms, R.E. (2005) Fractional crystallisation – the fat modification process for the 21st century. Eur J Lipid Sci Technol 107 (1), 48–57. Timms, R.E. (2006) Fractionation of palm oil: current status, future possibilities. In: Timothy L. Mounts Award Address presented at the AOCS World Conference and Exhibition on Oilseed and Vegetable Oil Utilization. Istanbul. Tirtiaux, A. (1990) Dry fractionation: a technique and an art. In: Edible Fats and Oils Processing: Basic Principles and Modern Practices (ed. D.R. Erickson). American Oil Chemists’ Society, Champaign IL, pp. 136–141. Tirtiaux, A. & Gibon, V. (1996) Dry fractionation: the boost goes on. Paper presented at the AOCS World Conference, Istanbul. Tirtiaux, A. & Tan, C.H. (1997) Fat crystallisation method and apparatus therefore. PCT Patent Application WO97/14777. Tirtiaux, F. (1976) Le fractionnement industriel des corps gras par crystallisation dirig´ee – proc´ed´e Tirtiaux. Ol´eagineux 31 (6), 279–285. van Loon, C. (1926) An improved process for the conversion of neutral or nearly neutral triglycerides, mutually or with other esters. British Patent 249,916. van Putte, K.P.A.M. & Muller, J.J. (1987) Fractionation of fat blends. European Patent 0 249 282, assigned to Unilever. Vander Wal, R.J. & Van Akkeren, L.A. (1951) Modified lard and process of producing same. US Patent 2,571,315, assigned to Armour and Company. Verh´e, R., Van Hoed, V. & De Greyt, W.F.J. (2006) Detection of alkyl ketones during chemical interesterification of lipids. Paper presented at 97th AOCS Annual Meeting & Expo, St Louis. von Rappard, G. & Plonis, G.F. (1980) Process for recovering triglycerides. British Patent 2 048 928, assigned to Walter Rau Lebensmittelwerke. Walstra, P. (1998) Secondary nucleation in triglyceride crystallization. Prog Colloid Polym Sci 108, 4–8. Ward, J. (1982) Edible fat product. US Patent 4,341,812, assigned to Nabisco Brands Inc. Weber, K., Homann, T. & Willner, T. (1998) Fat crystallizers with stirring surfaces: theory and practice. Ol´eagineux Corps Gras Lipides 5 (5), 381–384. Willner, T., Sitzmann, W. & M¨unch, E.-W. (1989) Herstellung von Kakaobutterersatz durch fraktionierte Speise¨olkristallisation. Fett Wissenschaft Technologie 91, 586–592. Willner, T., Sitzmann, W. & M¨unch, E.-W. (1990) Production of cocoa butter replacers by fractionation of edible oils and fats. In: Edible Fats and Oils Processing: Basic Principles and Modern Practicess, (ed. D.R. Erickson). American Oil Chemists’ Society, Champaign, IL, pp. 239–245. Willner, T. & Weber, K. (1994) High-pressure dry fractionation for confectionary fat production. Lipid Technol 6, 57–60. Wilp, Ch. (2000) Dry fractionation of fats and oils by means of centrifugation. Paper presented at the 91st AOCS Annual Meeting & Expo, San Diego. Yang, T., Zhang, H., Mu, H., Sinclair, A.J. & Xu, X. (2004) Diacylglycerols from butterfat: production by glycerolysis and short-path distillation and analysis of physical properties. J Am Oil Chem Soc 81 (10), 979–987. Yoneda, S., Higuchi, K., Taniguchi, A. & Kuwabara, Y. (1997) Process for dry fractionation of fats and oils. European Patent Application 0 798 369 A2, assigned to Fuji Oil Company Ltd.

BLUK122-Dijkstra

8

September 25, 2007

20:2

Food applications of trans fatty acids

John Podmore

8.1

INTRODUCTION

Among in current use the fat modification processes, hydrogenation is the process that generates major quantities of trans fatty acids in food fats. The hydrogenation process was invented by Wilhelm Normann in 1902 (Kaufmann, 1939) and was a great boon to the food industry, particularly the margarine and cooking fat industry. Edible oils are natural products and their physical properties are dictated by their animal or agricultural origin. So to meet the requirements of the food manufacturer for specific physical and chemical characteristics, the processor must use physical and chemical techniques to modify the oils available to him or her. The main techniques are blending, hydrogenation, fractionation and interesterification (Dijkstra, 2002). At the time of its introduction, hydrogenation was found to be of particular value as the only modification process available that could raise the melting point of a triglyceride by saturating the double bonds in the component unsaturated fatty acids of that triglyceride. It was found that not only did hydrogenation convert liquid oils to solid fats (leading to the process being referred to as ‘hardening’) but it also improved resistance to rancidity by saturating the double bonds present in the starting oil. Other benefits were the removal of the odour of the natural oil; for example whale and fish oils lose their natural odour when hardened, and oils become paler in colour when hydrogenated. On the basis of these benefits and in spite of the high processing cost per tonne of oil and the dangers inherent in handling hydrogen, it was quickly adopted by processors internationally. The use of partial hydrogenation was so successful that it could be seen as the saviour of the margarine industry by removing its reliance on animal fats, which were only available as by-products of the meat and dairy industries. Hydrogenation led processors to exploit the possibility of using whale and fish oils that had previously been felt to be too oxidatively unstable, thereby keeping their product costs down and ensuring the continued development of the margarine industry (Schwitzer, 1956). The mechanism and kinetics of the hydrogenation process are still not fully understood, but the reaction soon became sufficiently well defined for processors to be able to control industrial hydrogenation. Geometric isomerism is invariably associated with the hydrogenation process (Coenen, 1976), and the greater the selectivity, the greater will be the production of trans fatty acids. The condition of the catalyst, whether poisoned or not, also strongly influences the selectivity and the rate of the hardening process. Table 8.1 shows how sulphur poisoning of the catalyst has an effect on the melting profile due to an increase in the concentration of trans fatty acids. The ability to suppress or promote the generation of trans fatty acids was quickly identified as a valuable tool by the processor. An example discussed by Andersen and Williams (1965b)

BLUK122-Dijkstra

204

September 25, 2007

20:2

Trans Fatty Acids Table 8.1

The influence of catalyst poisoning on melting behaviour. Catalyst type

Property Iodine value Solid fat index at

Standard

Sulphur poisoned

68.9 60.8 48.8 26.3 8.3

68.9 59.0 45.8 21.6 3.0

10.0◦ C (50 F) 21.1◦ C (70 F) 33.3◦ C (92 F) 40.0◦ C (104 F)

is the hydrogenation of groundnut oil to reduce the iodine value from 90 to 72. When trans suppressive conditions were used, it gave rise to a soft semi-fluid fat that could be used to replace part of the liquid portion of the margarine oil blend. In contrast, by applying transpromoting conditions a firm plastic fat with a slip melting point of 33◦ C resulted that could be used as up to 80% of a margarine oil blend. This simple example shows that there were considerable technical and economic advantages to be obtained in controlling the level of trans fatty acids in hydrogenated oils. Table 8.2 shows the change in fatty acid composition of soya bean oil hydrogenated to a range of iodine values, using selective hydrogenating conditions. It can be seen that as the iodine value falls to a value of 76 (corresponding to a fall by 53 units), the level of saturated fatty acids increases only by 3%, whilst more than 50% of the fatty acids become trans isomers, which have a significant influence on the physical properties of the fat. Early concerns were that the generation of trans isomers increased the melting point of a partially hydrogenated oil by reducing significantly only the polyunsaturated fatty acid content and introducing a monounsaturated isomeric structure into the oil that was not naturally present. However, the technological and economic advantages were overwhelming, and the ability to place some control on the level of trans isomers in a partially hydrogenated oil was quickly accepted, particularly by shortening and margarine manufacturers. By manipulation of the hydrogenation conditions, the processor could produce fats with the same degree of saturation but with differing physical characteristics due to varying levels of trans isomers. This benefit complemented by the greater complexity of fatty acids present in the triglycerides of a partially hydrogenated fat was found to influence the crystallising and melting behaviour of the partially hardened fat and its blends with other oils or fats. Some of the food applications taking advantage of enhanced levels of trans fatty acids in partially hardened oils will be discussed in the following sections. Table 8.2

Fatty acid composition of selectively hydrogenated soya bean oil.

Iodine value Fatty acid C16:0 C18:0 C18:1 cis C18:1 trans C18:2 C18:3 Melting point (◦ C) From Weiss (1983).

129

107

87

76

5

Fatty acid composition (%) 11 4 27 0 50 8

11 4 27 21 34 3 27

11 5 26 41 16 1 33

11 7 24 52 6 0 38

11 83 0 6 0 0 63

BLUK122-Dijkstra

September 25, 2007

20:2

Food applications of trans fatty acids

8.2

205

MARGARINE

The manufacture of margarine benefited greatly from the introduction of hydrogenation, and the presence of trans isomers had a significant impact on the melting characteristics of the product such that its use was expanded from being just a spread into use as a bakery fat for the manufacture of cake, cream and pastry.

8.2.1

Table margarine

The product was designed for spreading on bread, crackers, biscuits, etc., and was targeted to have a texture and consistency similar to that of butter with cool eating qualities and no residual waxiness on the palate. Subsequent developments led to the requirement that margarine could be stored in a refrigerator while maintaining its plastic character and that it should not significantly lose its consistency when in use outside a refrigerator. The value of generating trans isomers for this type of product is that sufficient solids can be present at room temperature, and the steep melting curve ensures that there are no residual solids at body temperature. The drawback is that the steep melting curve gives a narrow melting range and hence a brittle character. Blending strategies were introduced to overcome this problem. In Europe a large number of blend formulations have been used as different oils became available on an economic basis. The principle generally applied in order to avoid brittleness in the margarine was to have a relatively high melting partially hardened component that melted between 40 and 45◦ C, a mid-melting component that melted between 28 and 35◦ C, which again could be a partially hydrogenated oil, and finally a liquid vegetable oil component. The range of components in the oil blend with differing melting points was designed to ensure a minimum change of the solid/liquid triglyceride ratio with a change in temperature and so extended the plastic character of the margarine. The high melting and mid-melting components were hydrogenated in a trans-promoting manner to give trans fatty acid levels of between 30 and 40%. This gave the desired steep melting profile, and the presence of the trans isomers also led to an even greater mix of triglycerides in the product. In the period when whale oil and fish oils were readily available, they were popular for use in table margarine, not only because of their low cost but also because the wide range of fatty acids present led to a wide range of mixed triglycerides, which was increased even further by the presence of trans fatty acids resulting from partial hydrogenation to melting points between 30 and 50◦ C. There was also the additional advantage that the partially hydrogenated whale and fish oils preferentially crystallised in the phase. Blends of these hardened oils and their steep melting curve meant that the use of the more costly vegetable oil in the formulation could be minimised. Examples of blends are shown in Table 8.3, where it can be seen that alternative mid-melting oils can be introduced largely on an economic and availability basis, but in the case of the blends that are not all vegetable, the steep-melting partially hardened fish oil gives the product its desired melting behaviour. In the case of the all-vegetable blends, the presence of hydrogenated palm oil ensures a fine, smooth texture due to the development of a crystalline matrix, and the steep-melting hydrogenated rapeseed oil keeps the melting point below body temperature. In the USA, the manufacture of table margarine made use of oils like cottonseed oil and more predominantly soya bean oil. The blends were made up of soya bean oil hardened to

BLUK122-Dijkstra

206

September 25, 2007

20:2

Trans Fatty Acids

Table 8.3

Typical table margarine oil blends. Formulations (as percentages)

Components Non-all-vegetable Hydrogenated fish oil (m.p. 46◦ C) Hydrogenated fish oil (m.p. 30◦ C) Palm oil Coconut oil Lard Liquid vegetable oil All-vegetable Hydrogenated palm oil (m.p. 44◦ C) Hydrogenated canolaa (m.p. 43◦ C) Hydrogenated groundnut oil (m.p. 32◦ C) Palm oil Coconut oil Liquid vegetable oil a

1

2

3

4

5

3 60 27

13 44 9 19

2 65

8 17 20 15

6 42

10

15

20 13

45

52

15 45

10 45

5 15

40

20 10 15

10 70

70 10 20

Canola is a rapeseed oil with low erucic acid content.

a range of iodine values to provide a range of fatty acid compositions, as shown earlier in Table 8.2. The table shows that high levels of trans fatty acids were generated. The difficulty faced by the margarine processor of blends based solely on soya bean oil is the fact that partially hydrogenated soya bean oil crystallises preferentially in the form (see Table 8.4), which can lead to a grainy texture in the finished product. The preferred approach was to use partially hydrogenated cottonseed or palm oil as the highest melting blend component, to encourage the total blend to crystallise in the form (Merker, 1958). Latondress (1981) describes the preparation of hydrogenated feedstocks based entirely on soya bean oil for the manufacture of table margarines, in which all the components were hydrogenated using selective conditions. Table 8.5 shows the hydrogenation conditions, whereby end points were detected by iodine value; the resulting blend formulations that are used to make the margarines have been summarised in Table 8.6. Latondress also noted that the tendency for margarines to form crystals is accelerated by an increase in temperature and that better temperature control in transport and storage allowed Table 8.4 Crystal forms of various hydrogenated fats in their most stable state. form Palm oil Tallow Butterfat

form Natural, non-modified oils and fats Coconut oil Palm kernel oil Lard

Hydrogenated oils Cotton seed oil High-erucic-acid rapeseed oil Herring oil Whale oil From Wiedermann (1968).

Soya bean oil Sunflower seed oil Groundnut oil Canola

BLUK122-Dijkstra

September 25, 2007

20:2

Food applications of trans fatty acids Table 8.5

207

Hydrogenation conditions used for the margarine oil stocks.

Component Starting temperature (◦ C) Hydrogenation temperature (◦ C) Pressure (bar gauge) Nickel catalyst (wt %) Final iodine value Solid fat index at 10.0◦ C 21.1◦ C 33.3◦ C

1

2

3

150 175 1.0 0.2 106–108 4 2

150 220 0.34 0.2 73–76 36–38 19–21 2 max

150 220 0.34 0.2 64–68 58–61 42–46 21 max

cottonseed oil and palm oil to be dropped from margarine formulations. This confirmed earlier work that had shown that a mixture of hardened soya bean oils including a wide range of triglycerides further enhanced by the presence of trans isomers could be processed into a smooth-textured margarine. It was important that the oil blend was then processed, so that rapid nucleation took place at −10◦ C in the presence of considerable supercooling and that the product was then homogenised at 18◦ C to give a smooth texture and a character. The introduction of tub margarines designed for spreading directly after storage in a refrigerator changed the demands placed on the oil blend and its hydrogenated component. Since the margarine had a high liquid vegetable oil content (ca. 60–70%) and solid fats that give an acceptable consistency over a wide range of temperatures, i.e. 4–20◦ C, it could be regarded as having a dual plastic range. The margarine would have a wide plastic range between refrigerator and room temperature, converting to a sharp melting point in the mouth. The oil blend will of necessity have a ‘flat’ melting curve, with the hardstock providing approximately 25% of the solid triglycerides. In order to ensure that the hardstock of the oil blend quickly builds a crystalline network when being processed to support the high volume of liquid oil, a small proportion of a highmelting hydrogenated oil was introduced in order to initiate nucleation early in the process. The balance of the hardstock is made up of a lower melting hardened oil that has been selectively hardened to provide a steep melting curve due to the presence of a high level of trans isomers. The high solids at low temperatures encourage rapid crystallisation, while the low solids content at body temperature ensures the absence of palate waxiness. This type of formulation produced a smooth consistent product that did not ‘oil off’ in storage. Table 8.6

Fat blend compositions of US margarines.

Type of margarine Component 1 Component 2 Component 3 Liquid oil Solid fat index at 10.0 ◦ C 21.1 ◦ C 33.3 ◦ C Linolenic acid content (%) From Latondress (1981).

Soft stick

50 50 20–24 12–15 2–4 0

Stick Blend composition (%) 42 20 38 Blend properties 27–30 >17.5 2.5–3.5 <2

Tub 80 20

10–14 6–9 1–2 <3

BLUK122-Dijkstra

208

September 25, 2007

20:2

Trans Fatty Acids

As shown in Table 8.4, hydrogenated sunflower seed oil also tends to crystallise in the form. This has implications for margarines, in which sunflower seed oil is the only or almost the only oil used; they might recrystallise after packaging and become sandy. By interesterifying some or most of the partially hydrogenated sunflower seed oil with liquid sunflower seed oil, -stable margarine fat blends can be obtained (Gander et al., 1966).

8.2.2

Cake margarine

Investigations into the mechanisms by which fat functions in cake have shown that cake manufacture is highly dependent on the fat for proper aeration. In addition, fat contributes to the crumb structure and eating qualities (Shepherd and Yoell, 1976). Traditionally, with plastic fats, the fat and sugar were creamed together as the first stage in batter preparation and air bubbles were distributed in the fat phase. This method has now been superseded by all-in methods, where the batter preparation is completed in one step coupled with the use of highspeed mixers. Continuous mixers where a loose slurry of ingredients is delivered to a mixing head where air is injected into the batter are popular in large bakeries. Examination of batters made by all these methods show that the air is held initially in the fat phase. The finer the distribution of the fat and air in the batter, the better the final cake volume and crumb structure. The fat used to make a cake, either margarine or shortening, must be plastic, so that it has the ability to incorporate air and retain it; thus, in the creaming process, there must be enough liquid oil available to envelop the incorporated air bubble and sufficient crystalline fat to stabilise the system. The small needle-like crystals of the form have been found to be most effective in stabilising the air bubble in a cake batter (Brooker, 1993). It can be seen from the comments made above that plastic fats must be easily whipped into the cake batter and yet retain sufficient structure in order to retain the incorporated air. Traditionally, oil blends with flat melting curves were preferred, because they retained a consistent texture over a wide range of temperatures and withstood the work input during cake preparation without becoming unduly soft; that is, the margarine or shortening was not too firm at low temperatures and there was sufficient crystalline material available at higher temperatures. These oil blends contained high levels of liquid oil, and the hardstock was based on oils hardened in trans suppressive conditions to produce stearic acid and give a flat melting curve. Although these blends worked well, improved automation and temperature control in bakeries, as well as a desire by the processor to standardise his or her partially hydrogenated feedstocks, led to the possibility of supplying the baker with margarine and shortening with oil blends based on selectively hardened oils normally used in table margarines. Oil blends based on selectively hardened partially hydrogenated oils containing significant levels of trans isomers were developed. The major step was to introduce a larger quantity of higher melting hydrogenated fat and a greater quantity of liquid vegetable oil, which made the oil blend more plastic and reduced the risk of brittleness. It was found that the steep-melting characteristics improved creaming power and resulted in cakes with a finer crumb structure. In Table 8.7, some of the formulations used are shown. In the non-all-vegetable section, blend 1 is relatively intolerant to temperature change because of the very high content of hardened fish oil with a high trans fatty acid content. Blends 2 and 3 have improved temperature tolerance due to the introduction of other mid-melting-point oils and a greater quantity of liquid oil. Blend 4 contains a high proportion of coconut oil for applications discussed in the following paragraph. The all-vegetable blends (Table 8.7) again show how increased liquid oil and the introduction of alternative mid-melting fats allow the use of trans-promoting hardening of the partially hardened oil components.

BLUK122-Dijkstra

September 25, 2007

20:2

Food applications of trans fatty acids Table 8.7

209

Typical cake margarine oil blends. Formulation (as percentages)

Blend component Non-all-vegetable Hydrogenated fish oil, m.p. 46◦ C Hydrogenated fish oil, m.p. 32◦ C Lard Palm oil Coconut oil Liquid vegetable oil All-vegetable Hydrogenated palm oil, m.p. 46◦ C Hydrogenated soya bean oil, m.p. 32◦ C Palm oil Coconut oil Liquid vegetable oil

1

2

3

4

5 85

10 50 22

10 50

20 30

30 20

10

18

20 10 10

7 60 18

10 35 20

50 25

15

35

25

5 50 18 10 17

The use of trans-promoted partially hydrogenated fats in bakery margarines opened up the possibility of the manufacture of specialist products, for example margarines dedicated to the preparation of butter creams, filling creams and toppings. The creams needed to be sharp melting with no palate waxiness and high creaming power to produce smooth, stiff, high-volume creams. The desired performance characteristic could be achieved by amending the levels of the selectively hydrogenated hardstocks. The performance could be enhanced even further by the inclusion of coconut oil into the blend or by the addition of emulsifiers.

8.2.3

Pastry margarine

There seems to be relatively little scope for the use of selectively partially hydrogenated fats in puff pastry margarine oil blends. The final margarine is designed to form thin coherent films in the dough, so the margarine must be of a smooth texture and firm, tough consistency and must display a good resistance to worksoftening, i.e. it should withstand the heavy mechanical working and rolling without crumbling and softening, which would cause it to be absorbed by the dough. Traditionally, the fat blend for a pastry margarine contained a high proportion of a highmelting fat like oleostearine or premier jus, but both of these have now been replaced by an oil hydrogenated to a melting point between 46 and 50◦ C. The blend could also contain a mid-melting-point fat and a high proportion of liquid vegetable oil. These blends were designed to give good volume increase in the baked product and were successfully processed on a rotating chilling drum (Andersen and Williams, 1965a). The development of tubular chillers for processing pastry margarines opened the way for some changes in oil blend formulation, and this coupled with the consumer’s demand for a lower melting pastry margarine led to the introduction of fats containing high levels of trans isomers. In spite of their narrow melting range, blends using fats containing high levels of trans isomers give the margarine a lower melting point and sufficient solids at the working temperature. The inclusion into the blend of fats like partially hydrogenated palm oil or stearin fractions of palm oil improves the plasticity and texture of the margarine. The drawback with these formulations is that they have less resistance to worksoftening when compared with

BLUK122-Dijkstra

210

September 25, 2007

20:2

Trans Fatty Acids Table 8.8

Typical pastry margarine oil blends. Formulation (as percentages)

Blend component Non-all-vegetable Premier jus, m.p. 46◦ C Hydrogenated fish oil, m.p. 48◦ C Hydrogenated fish oil, m.p. 33◦ C Tallow, m.p. 40◦ C Liquid vegetable oil All-vegetable Hydrogenated palm oil, m.p. 46◦ C Hydrogenated palm oil, m.p. 40◦ C Hydrogenated rapeseed oil, m.p. 34◦ C Palm oil Liquid vegetable oil

1

2

3

4

30 25

60

65 25

45

30 10

55 20 10 15

20 70

10 25 45

10

20

15 35 40 10

10 10 25 50 15

the traditional ‘drum’ product; however, the bakers have adjusted to this limitation and goodquality products are being made with these margarines. Blends used for pastry margarine are shown in Table 8.8.

8.3

BISCUIT FATS

Fats for biscuit manufacture will be considered here in the major categories of dough fats and filling cream fats (Manley, 1983). The fats for biscuit coatings will not be discussed in this section, as confectionery fats will be dealt with later in the chapter. In biscuits, dough and cream fats have very different functional requirements. The major dough fat can be considered as a shortening in the most popular biscuit type, which is the short dough biscuit, where the fat interrupts the gluten development in the biscuit dough and so shortens and texturises the finished biscuit. Another type of dough fat is used for laminated biscuits. These are puff-type biscuits, where fat is folded into a tough, extensible dough as fine lumps or flakes to cause flakiness in the finished biscuit. Biscuit-filling-cream fats are used in simple fat/sugar creams and must melt sharply to release the sugar and added flavour and be firm enough at ambient temperatures to hold the biscuit shells together and not squeeze out of the biscuit when it is eaten.

8.3.1

Dough fats – short dough biscuits

The most critical requirement for a biscuit dough fat is oxidative stability, as biscuits are long-shelf-life products; hence, oils with high levels of unsaturation are usually avoided. Traditionally, fats like lard and butter were favoured, but now vegetable oils are preferred and these are often hydrogenated. The melting curve of the oil blend is targeted to give negligible solid fat content at body temperature and sufficient solids at dough temperature to allow the fat to be smeared through the dough to prevent gluten development. Blends that are now popular are based on hydrogenated vegetable oil and palm oil. The blends show good oxidative stability and can be readily processed in a tubular chiller before being used in the biscuit dough, but hardened oils with a steep melting curve are said to lead to the

BLUK122-Dijkstra

September 25, 2007

20:2

Food applications of trans fatty acids Table 8.9

211

Biscuit dough fat blends. Formulation (as percentages)

Blend component

1 42◦ C

Hydrogenated palm oil, m.p. Hydrogenated rapeseed oil, m.p. 34◦ C Hydrogenated soya bean oil, m.p. 34◦ C Palm oil Coconut oil

2

7 93 40 60

3

4

21

10 30

70 9

55 5

development of fat bloom. Experience has shown that partially hardened oils with trans fatty acids in excess of 25% are acceptable. Some typical blends are shown in Table 8.9.

8.3.2

Dough fat – laminated biscuits

The fat popularly used in this application is palm kernel oil, which can be delivered as flakes or small lumps onto a dough that is free of fat and that is then folded and rolled to create a layered structure. For economic reasons, palm kernel oil can be replaced by partially hydrogenated rapeseed or soya bean oil that has been hydrogenated to a melting point of 32–34◦ C, using trans-promoting conditions to give a trans fatty acid content of approximately 35% and a steep melting curve in order to mimic the behaviour of palm kernel oil.

8.3.3

Cream filling fat

Blends based on hydrogenated lauric oils are favoured for this application, as they give the characteristics required, which are:

r r r

The cream must solidify rapidly after spreading, so that the biscuit shells are held together during handling and transport. The cream must be firm enough at ambient temperature to hold the two biscuit shells together and avoid the cream being squeezed out of the sandwich. When eaten, the cream must give a firm bite and then melt sharply to release the sugar and flavourings.

The high cost and shortage of lauric oils led to the consideration of introducing partially hydrogenated vegetable oils. Blends were developed that used rapeseed oil hydrogenated to a melting point of 36◦ C in trans-promoting conditions to give a steep melting curve due to a trans fatty acid content of approximately 40%. Even when the hardened oil is blended with hardened lauric fat and the solids content is lowered due to the eutectic effect, the fat can still meet the rigorous requirements. Table 8.10 shows some typical blends.

8.4

FATS FOR CHOCOLATE CONFECTIONERY

Confectionery fats are fats used as a substitute for cocoa butter in chocolate formulations, either totally or partially. The product made with these alternative fats is often referred to as

BLUK122-Dijkstra

212

September 25, 2007

20:2

Trans Fatty Acids Table 8.10

Biscuit cream fat blends. Formulation (as percentages)

Component

1 42◦ C

Hydrogenated palm oil, m.p. Hydrogenated rapeseed oil, m.p. 36◦ C Hydrogenated palm kernel oil, m.p. 35◦ C Coconut oil Palm oil

7 73 20

2 50 45 5

3

4

7 93 50 30 20

compound chocolate in order to distinguish it from real chocolate, which is made exclusively with cocoa butter. Before considering these confectionery fats, it is necessary to have some understanding of the chemistry of cocoa butter. The fatty acid composition of cocoa butter is almost entirely made up of three fatty acids. Palmitic acid, stearic acid and oleic acid together make up more than 95% of the fatty acids present in the fat, and they are nearly equally distributed, which in turn gives a very specific triglyceride composition to the fat. The triglyceride group formed is Sat – O – Sat, where O is oleic acid and is in the sn-2 position and Sat indicates either palmitic acid or stearic acid and these are in the sn-1 and sn-3 positions. This leads to the presence of only three triglycerides – POS, POP and SOS, which have very similar chemical structures, so that cocoa butter can form mixed crystals and thus exhibit a sharp melting point. Confectionery fats are made from vegetable oils and fall into two categories: those made from lauric oils, the so-called cocoa butter substitutes (CBS), and those based on hydrogenated vegetable oils like cottonseed oil, soya bean oil, rapeseed oil and palm olein, which are referred to as cocoa butter replacers (CBR). Both categories are stable in the polymorph and are therefore ultimately incompatible with cocoa butter, which exhibits six polymorphs of which three are important commercially (forms IV, V and VI); after correct tempering, form V, a polymorph, is produced. The lauric CBS are made by fractionation of palm kernel oil or coconut oil and subsequent hydrogenation of the stearin fraction. They can exhibit characteristics of mouthfeel, hardness and flavour release similar to those found in cocoa butter. Their major drawback is that their triglyceride composition differs so much from that of cocoa butter that there is little compatibility, which leads to a strong depression in the melting behaviour for a wide range of mixing ratios due to the formation of strong eutectics. Figure 8.1 is a diagram of isotherms for cocoa butter and a lauric substitute, which shows a significant softening effect due to eutectic formation. The non-lauric CBR, in contrast, contain palmitic, stearic and oleic acids, suggesting that they should be much more compatible with cocoa butter. However, the arrangement of these fatty acids in their triglycerides is much more random, so the structure differs from that of cocoa butter. In order to achieve melting characteristics similar to those of cocoa butter, the starting oils are hydrogenated selectively, so that a relatively high level of trans double bonds is present in the hardened oil. Levels of over 45% trans isomers are not unusual. The hydrogenation step can be followed by fractionation to improve the physical properties further. The presence of the trans isomers makes the arrangement of the fatty acids in the triglycerides even more complex than that found in cocoa butter. A diagram of the isotherms of cocoa butter and a non-lauric CBR shows the improved compatibility compared with lauric CBS, though it is still very limited (Fig. 8.2).

BLUK122-Dijkstra

September 25, 2007

20:2

Food applications of trans fatty acids

213

80

NMR solid (%)

60 20°C 25°C 30°C 32.5°C 35°C

40

20

0 0

20

40

60

80

100

Lauric cocoa butter replacer (%) Fig. 8.1 Isotherms for lauric confectionery fats (CBS) and cocoa butter. (Reproduced with permission from Taylor & Francis from Smith, K.W. (2001) Cocoa butter and cocoa butter equivalents. In: Structured and Modified Lipids (ed. F.D. Gunstone). Dekker, New York, pp. 401–422.)

Because of the greater tolerance of the non-lauric CBR to cocoa butter, cocoa butter can tolerate up to 7% replacement with non-lauric CBR in a chocolate recipe, without affecting the hardness or gloss. In the reverse case, the non-lauric CBR in a compound chocolate for a coating can tolerate up to 25% cocoa butter (Talbot, 1997). This greater tolerance to cocoa butter means that in a compound chocolate recipe, cocoa liquor can be used instead of cocoa

80

NMR solid (%)

60 20°C 25°C 30°C 32.5°C 35°C

40

20

0 0

20

40

60

80

100

Non-lauric cocoa butter replacer (%) Fig. 8.2 Isotherms for non-lauric confectionery fats (CBR) and cocoa butter. (Reproduced with permission from Taylor & Francis from Smith, K.W. (2001) Cocoa butter and cocoa butter equivalents. In: Structured and Modified Lipids (ed. F.D. Gunstone). Dekker, New York, pp. 401–422.)

BLUK122-Dijkstra

214

September 25, 2007

20:2

Trans Fatty Acids Table 8.11

Compositions of toffee and fudge in parts by weight.

Ingredient

Toffee

Fudge

170 140 115

100 100 180 100 20 30 2 15

Glucose syrup (42 DE) Condensed milk Brown sugar Fondant Hydrogenated oil Egg frappe Salt Glycerine Flavour

75 3 1

From Lees and Jackson (1973).

powder. This gives a fuller and rounder flavour than that can be achieved with a lauric CBS, which necessitates the use of cocoa powder. The non-tempering nature of both types of confectionery fats means that processing compound chocolate is easier than processing real chocolate, and the product can be used immediately, i.e. without tempering for enrobing and moulding. An additional feature is that compound chocolate has a lower viscosity than real chocolate due to the absence of crystals, which facilitates the enrobing process control.

8.5

FATS FOR SUGAR CONFECTIONERY

Fat is used in only a limited number of sugar confectionery items; these are toffee, caramel and fudge. Toffee and caramel can be regarded as virtually the same, since the differences in the products relates to the difference in hardness, which depends on the residual moisture level. The basic ingredients used in toffee are sugar, glucose syrup, milk protein, fat, salt and water. The sugar and milk protein are the major ingredients. However, where fat is introduced, it makes a significant contribution to the texture, chewing characteristics, colour and flavour of the toffee; for example, when the fat content is low, a sticky caramel results and a high fat content causes oiling on the surface of a caramel. Fudge basically has a toffee formulation, but it contains higher proportions of sugar and milk and is processed to develop a grainy texture. General recipes are shown in Table 8.11. The characteristics required in the fat used will be considered here. The fats used in the manufacture of toffee and fudge are solid at ambient temperature and are virtually completely melted at body temperature. The fat originally favoured for toffee manufacture was butter because of its flavour and melting characteristics. Many years ago the butter in toffee was wholly or partially replaced by vegetable fat, the preferred one being partially hydrogenated palm kernel oil that melts between 30 and 34◦ C. Hardened palm kernel oil exhibits the characteristics desired in toffee, that is a narrow melting range and a brittle texture. Moreover, hardened palm kernel oil has excellent resistance to oxidation to prevent the development of off-flavours in the toffee. Other vegetable oils replaced palm kernel oil because of shortages and fluctuating prices. The vegetable oils that have been most popularly used are soya bean oil, rapeseed oil, palm oil and groundnut oil. In order that they most closely mimic the characteristics of hardened palm kernel oil, they are hydrogenated under selective conditions and the level of trans isomers

BLUK122-Dijkstra

September 25, 2007

20:2

Food applications of trans fatty acids

215

is maximised. This ensures a steep melting curve and narrow melting range. The oils were found to give toffee with a more than adequate shelf-life. These hardened vegetable oils can be used as single components or in blends with each other or with palm oil. They cannot be blended with hydrogenated palm kernel oil or other lauric fat due to the formation of eutectics, causing the mixture to become very soft.

8.6

VANASPATI

Vanaspati or vegetable ghee is a substitute for natural ghee. In the early part of the last century the economic situation in India caused the demand for animal fat to exceed production, resulting in a price increase that put natural ghee beyond the reach of the general population, which led to the development of a cheaper alternative vegetable fat to look like ghee and replace it in the household. The first vanaspati was made from hydrogenated vegetable fat imported from the Netherlands and was used as a substitute for ghee by bulk users, such as restaurants and manufacturers of sweetmeats. The first production of vanaspati in India was in 1930 and 10 years after this, imports ceased due to the rapid installation of factories in many provinces, indicating the growing popularity of the product. The quality standards of vanaspati in India are strictly regulated by the Vegetable Oil Products Control Order (1942), Vanaspati Manufacturers Association of India, New Delhi, Statistical data, 1994. Some important parameters are:

r r r r r

moisture content, 0.25% max; slip melting point, 31–41◦ C; free fatty acid (as oleic), 0.25% max; unsaponifiable matter, 2.0% max; nickel content 1.5 ppm max.

Colours and flavourings are not permitted, but the addition of sesame oil and vitamin A is mandatory. The oils permitted for use in vanaspati are under regular review. Among those currently permitted are soya bean oil, sunflower seed oil, rapeseed oil (high erucic acid), maize oil, palm oil, palm olein and rice bran oil. The early vanaspati was developed on the basis of groundnut oil, but this is no longer permitted because of its demand for use as a liquid oil. For the manufacture of vanaspati, the conditions to be applied in the hydrogenation of the vegetable oil and how they are varied for individual oils have been closely investigated (Achaya, 1997). Selective hydrogenating conditions are used so that a high level of trans isomers is formed to give a steep melting curve. It has been shown that as the hardened oil is cooled to 50◦ C, the triglycerides containing trans isomers nucleate to initiate crystallisation. Photomicrographs of vanaspati show the presence of large granules of radially arranged needle-like crystals. The crystal network that is formed contains large saturated triglycerides crystals to cause granularity, and also the liquid phase is trapped so that separation is prevented. The earlier production method to produce a granular product follows the crystallisation process described. The production process involved, holding the oil blend in a jacketed stirred tank and cooling it to 10◦ C above the melting point. Following this, the oil was filled into containers, which were stored in a cool room at 15◦ C to allow solidification. This slow cooling

BLUK122-Dijkstra

216

September 25, 2007

20:2

Trans Fatty Acids

leads to the formation of large crystals to give a grainy texture. The use of scraped-surface heat exchangers has superseded this method and gives more textural options.

8.7 8.7.1

SYNTHETIC CREAMS Whipped toppings

These products are designed to be all-vegetable alternatives to natural dairy cream and are based on vegetable fats and an added emulsifier system; their major application is as whipped toppings. The situation is more complicated than simply combining a fat and emulsifier with a protein solution and then whipping the mixture. Not only must the fat have an appropriate solid fat content profile, but the nature of the fat also appears to be crucial. The generation of a good whipped product involves the controlled de-emulsification combined with the crystallisation of the fat to stabilise the air cells. In order that it whips properly, a natural cream must contain sufficient fat: 30% minimum. When a cream is first whipped, a foam is created in which the air bubbles are stabilised by agglomerated fat globules at the air–water interface. It is important that the fat does not coalesce, which can lead to churning, and cooling the cream ensures there is sufficient solid fat present to prevent coalescence. Continued mixing shears the incorporated air bubbles increasing their number, which is stabilised by more agglomerated fat globules from the aqueous phase until a network is formed in the system that leads to an increase in viscosity. In the case of non-dairy whipped toppings, the fat must have the correct melting profile to ensure that the appropriate amount of solid fat is available to stabilise the incorporated air bubbles. During manufacture, the emulsion is homogenised and cooled. It is suggested that in the homogenised cream, the hydrophobic portions of the protein penetrate the fat globules and inhibit crystallisation, making the cream act as if the fat is in a supercooled state. Therefore, as the protein is desorbed during whipping, the fat at the air–water interface rapidly crystallises and the air bubbles are stabilised by the solid fat as it agglomerates. The nature of the fat used in this application is important, and the best results are obtained by using hydrogenated lauric fats. Synthetic creams give poor results when made with partially hydrogenated oils based on soya bean, sunflower seed and rapeseed oils, particularly if they do not exhibit good supercooling characteristics. Accordingly, the performance seems to be associated with C12 fatty acid (Barfod et al., 1989), though it is also important that the melting point of the fat is high enough to ensure crystals are present at the whipping temperature. It has been found that the inclusion of a partially hydrogenated fat containing a high proportion of trans C18 fatty acids with the partially hydrogenated lauric fat enhances the whipping performance in non-dairy whipped toppings. It is suggested that the additional fat gives better control of the crystallisation process.

8.7.2

Coffee whiteners

Coffee whiteners are used in coffee shops or fast food service outlets as an alternative to milk or cream for coffee. The product is available in either a liquid or a spray-dried form. A typical formulation is shown in Table 8.12. The level of fat is equivalent to that of a whipping cream and the fat used is invariably a hydrogenated vegetable fat. The product is an emulsion made as a liquid cream as described above. The hardened fat used has a melting

BLUK122-Dijkstra

September 25, 2007

20:2

Food applications of trans fatty acids Table 8.12

217

Typical coffee whitener recipe.

Ingredient Hydrogenated vegetable oil, m.p. 38◦ C Corn syrup solids Sodium caseinate Potassium hydrogen phosphate Glycerol monostearate

Weight % 37 56 5 1.6 0.4

From Lees and Jackson (1973).

point of approximately 37◦ C and has been selectively hydrogenated to promote the generation of trans isomers and hence a steep melting curve. The steep melting curve ensures there are sufficient solids present at the homogenisation temperature to give a stable emulsion and prevents the presence of an oily palate sensation.

8.8

CONCLUDING REMARKS

The technological value of trans isomers formed by hydrogenation has been illustrated in a range of products. The production of trans isomers is almost a by-product of hydrogenation, since the principal objectives are to turn liquid oils into solid fats and to produce a more stable product by eliminating or reducing polyunsaturation in the starting oil, but their level in a partially hydrogenated oil significantly influences its melting and solidification behaviour. Processors took advantage of the behaviour of trans fatty acids produced in partially hardened oils on both economic and technological grounds. Processors were able to use cheaper and more readily available oils and thus reduce their final product costs without significant loss in performance. This is most clearly illustrated by the replacement of hydrogenated lauric oils by other partially hardened vegetable oils with high trans fatty acid contents in a range of products in periods when lauric oils were overpriced and in short supply. The range of oils that became available to the margarine and shortening manufacturer with the advent of hydrogenation was considerable, and the ability to manipulate the levels of trans isomers was a great benefit in maintaining a consistent performance with this expanded range of oils. The work of Willett et al. (1993), which suggests a link between trans fatty acids and coronary heart disease, has meant that the use of partially hydrogenated oils has sharply declined and hence the food processor has lost the performance characteristics offered by trans fatty acids. Alternatives to partial hydrogenation and the attendant risks of generating trans fatty acid have been found using other forms of fat modification, like fractionation and interesterification as discussed in Chapter 7, often combined with the use of fully hardened oils. The drawback of these changes, however, is that they lead to a reduction in the range of oils commercially available to the processor and thus increases product cost. Other approaches and solutions will be discussed in the next chapter. The possibility of genetic modification of oilseed crops to produce oils with many of the desired characteristics already present in the oil has yet to be fully realised on a commercial basis, especially since the general public is not yet willing to accept these products, at least in Europe. When these oils are available and permitted, they will simplify processing and avoid the consequences of chemical and physical modification resulting from the use of hydrogenation and the generation of trans isomers.

BLUK122-Dijkstra

218

September 25, 2007

20:2

Trans Fatty Acids

REFERENCES Achaya, K.T. (1997) Ghee, vanaspati and special fats in India. In: Lipid Technologies and Applications (eds F.D. Gunstone & F.B. Padley). Marcel Dekker Inc, New York, pp. 369–390. Andersen, A.J.C. & Williams, P.N. (1965a) Process: cooling and crystallisation of emulsion. In: Margarine, 2nd edn. Pergamon Press Ltd, Oxford, pp. 196–208. Andersen, A.J.C. & Williams, P.N. (1965b) Raw materials. In: Margarine, 2nd edn. Pergamon Press Ltd, Oxford, p. 48. Barfod, N.M., Krog, N. & Buchheim, W. (1989) Lipid-protein-emulsifier-water interactions in whippable emulsions. In: Food Protein. Part 1: Structure and Functional Relationships (ed. J.E. Kinsella). American Oil Chemists Society, Champaign, IL, pp. 144–158. Brooker, B.E. (1993) The stabilisation of air in cake batters: the role of fat. Food Struct 12, 285–296. Coenen, J.W.E. (1976) Hydrogenation of edible oils. J Am Oil Chem Soc 53 (6), 382–389. Dijkstra, A.J. (2002) Hydrogenation and fractionation. In: Fats in Food Technology (ed. K.K. Rajah). Sheffield Academic Press, Sheffield, pp. 123–158. Gander, K.F., Hannewijk, J. & Haighton, A.J. (1966) Proc´ed´e de fabrication de margarine obtenue principalement a` partir de l’huile de tournesol. French Patent 1.457.751, assigned to Unilever. Kaufmann, H.P. (1939) Wilhelm Normann, zum Ged¨achtnis. Fette und Seifen 46 (5), 259–264. Latondress, E.G. (1981) Formulation of products from soybean oil. J Am Oil Chem Soc 58, 185–187. Lees, R. & Jackson, E.B. (1973) Caramels, toffee and fudge. In: Sugar Confectionery and Chocolate Manufacture. Leonard Hill Books, London, pp. 191–206. Manley, D.J.R. (1983) Classification of biscuits. In: Technology of Biscuits, Crackers and Cookies, Part II, Types of Biscuits. Ellis Horwood, Chichester, UK, pp. 152–158. Merker, D.R., Brown, L.C. & Wiedermann, L.H. (1958) The relationship of polymorphism to the texture of margarine containing soyabean oil and cottonseed oil. J Am Oil Chem Soc 35, 130–133. Schwitzer, M.K. (1956) Margarine and cooking fat: the history and world trade. In: Margarine and Other Food Fats. Leonard Hill Books Ltd, London, pp. 59–79. Shepherd, L.S. & Yoell, R.W. (1976) Cake emulsions. In: Food Emulsions (ed. S.E. Friberg). Marcel Dekker Inc, New York, pp. 216–275. Talbot, G. (1997) Vegetable fats. In: Industrial Chocolate Manufacture and Use (ed. S.T. Beckett). Blackwell Science Publishers, Oxford, pp. 319–321. Weiss, T.J. (1983) Basic processing of oils and fats. In: Food Oils and Their Uses, 2nd edn. Ellis Horwood, Chichester, UK, pp. 76–81. Wiedermann, L.H. (1968) Margarine formulation and control. J Am Oil Chem Soc 45, 515A, 520A–522A, 560A. Willett, W.C., Stampfer, M.J. & Mansen, J.E. (1993) Intake of trans fatty acids and the risk of coronary heart disease among women. Lancet 341, 581–583.

BLUK122-Dijkstra

9

September 25, 2007

20:3

Food products without trans fatty acids

Pernille Gerstenberg Kirkeby

9.1

INTRODUCTION

Margarine and related products traditionally contain a relatively high proportion of partially hydrogenated fats and thereby exhibit a high content of trans fatty acids (TFA). Scientific evidence for the supposed link between the consumption of TFA and the risk of coronary heart disease continues to be published (Stanley, 2007). In the previous chapters, the implications of TFA on health have been described in detail but also the benefits of TFA on product consistency and shelf-life, such as improved plasticity and low oxidation rate. These topics will therefore not be covered here. Margarine and related products like spreads contain a water phase and a fat phase, and they are characterised as water-in-oil (W/O) emulsions, in which the water phase is finely dispersed as droplets in the continuous fat phase. Accordingly, the raw materials used in the emulsion preparation prior to processing of margarine form part of either the fat phase or the water phase. For full-fat products, the fat phase is the main contributor to the functionality of the final margarine product. The impact of the water phase on the functionality of the margarine depends on the fat content, and thus the nature of the water phase affects only the final product significantly in lower (<40%) fat content products. When producing margarine and related products, the road to success and to high-quality products involves both their composition as prescribed by the recipe and the processing parameters during the production process. The recipe defines the composition of both the fat phase (including emulsifiers and other minor ingredients in the fat phase) and the water phase, which together should correspond to the final application of the product. In most modern processing facilities the recipe is linked to specific processing parameters in the control system in order to ensure that the same quality is consistently produced. Variations or modifications in the recipe generally necessitate changes in processing parameters in order to prevent the quality of the final product from altering. When switching from traditional margarine recipes to TFA-free recipes, different processing parameters are needed. Recommendations of how to achieve high-quality trans-free margarine products will be discussed in the following sections.

9.2

FAT PHASE

The main ingredient in the fat phase, the fat blend, normally consists of a blend of different fats and oils. When formulating a margarine blend, care should be taken that the fat blend provides the margarine with the desired characteristics and functionality. The objective is to

BLUK122-Dijkstra

220

September 25, 2007

20:3

Trans Fatty Acids

70

Puff pastry margarine

60

Table margarine

50 40

Frying margarine

30

Soft table margarine

20 10

Low-fat margarine

0 10°C Fig. 9.1

20°C

30°C

40°C

Solid fat content profiles for various margarine types.

obtain a defined solid fat content (SFC) at various temperatures, typically ranging from 5 to 40◦ C. As shown in Fig. 9.1, the SFC profile will vary according to the type of product; soft table margarine and low-fat spreads contain least solids at a given temperature and puff pastry the highest amounts of solid fats (Berger, 1989; Haighton, 1976). Variations occur within each product category, since different margarine manufacturers emphasise different product properties and thus arrive at different product specifications. In addition, when margarine manufacturers supply products to various countries with different climatic conditions, the manufacturer will generally adapt the oil blend formulation by supplying a harder formulation to hot areas and a softer one to colder areas. Blends adaptations can also be due to variations in geography, e.g. valley and mountain product, and thus a higher melting or a lower melting product, respectively. Adjusting the fat properties in accordance with seasonal variations was formerly quite common but has been abandoned when margarines started being stored in refrigerators. On the other hand, a margarine manufacturer may offer a range of croissant and puff pastry margarines with different solids contents to satisfy customers working in different bakery environments. The ingredients used to formulate the fat blend can be derived from any animal, vegetable or marine oil source. The choice of the ingredients will depend on legislation, economics, keepability, functionality and marketing demands. The latter can limit the interchangeability of fats if, e.g. a 100% vegetable or a low-TFA margarine product is in market demand. In addition, the availability of fats can be limited due to legislation, religious prohibitions or trade barriers (import restrictions). For a better overview as shown in Table 9.1, the fat ingredients can be divided into three main categories: liquid oils, semi-solid fats and hardstocks. A

Table 9.1

Fat categories with their ingredients.

Liquid oils

Semi-solid fats

Hardstocks

Unsaturated vegetable oils Palm olein

Vegetable oils hydrogenated to a melting point of 32–34◦ C Marine oils hydrogenated to a melting point of 32–34◦ C Palm oil Lard

Partially/fully hydrogenated oils and fats, melting point 40◦ C Interesterified fats

Palm kernel oil Coconut oil

Edible beef tallow Palm stearin

BLUK122-Dijkstra

September 25, 2007

20:3

Food products without trans fatty acids

221

suitable blend can be made by combining one or more ingredients from each of the three groups (Berger, 1989). Due to the modern refining processes, liquid oils such as soya bean oil, low-erucic-acid rapeseed oil (canola) or sunflower seed oil are today considered to be completely interchangeable as components of margarine and shortening blends. Fat modification processes, such as fractionation, hydrogenation and interesterification, enable a high degree of interchangeability among fats. The modification processes produce, individually or in combination, the full range of fatty intermediates used in the manufacture of all types of margarine and related products, and the processes enable fats to become almost completely interchangeable. The agricultural origin of the fat being used has therefore become a matter of cost price, as determined by the cost of crude oil or fat and the modification processes involved. A particular product specification can be met by a large number of alternative formulations. In times of shortage or high cost, the producer has the flexibility to interchange the best available and cheapest raw materials in an ever-changing market situation. This ensures that consumers can always buy a product with consistent quality and at more or less the same price. However, certain technical limitations exist when combining one or more ingredients from the three categories given in Table 9.1. The crystallisation habit, or polymorphism, of the fat blend used can set limits on the proportion of a particular fat used in the blend. Fats tend to crystallise in various forms, showing different melting points. Each of these crystalline forms with their respective melting points is called a polymorph and the phenomenon is called polymorphism (Timms, 1984). The different polymorphs can coexist in the fat, and their occurrence depends on the cooling and heating history of the fat. Due to the so-called crystal memory, some form of crystalline structure is preserved even when a fat is melted. This structure affects the crystallisation directly, especially if the rate of cooling is high. The polymorphic changes may lead to a grainy structure in margarine, which is also referred to as ‘sandiness’ (Johansson and Bergenst˚ahl, 1995; Merker et al., 1958; Timms, 1984). With some exceptions, fats exhibit three basic crystalline forms designated alpha (), beta prime ( ) and beta (), whereby the form is preferred in margarine. The structure showing the least amount of crystalline order is . It has the lowest heat of melting, the lowest melting point and is the least stable as well. The form has the most compact crystalline structure, is the most stable thermodynamically and shows the highest melting point. In general, the transition from one crystal form to another takes place in the order of → → , and the mechanism involves the melting/dissolution of the least stable form and the crystallisation from the melt/solution of the more stable form. The driving force in this process is the lower solubility of the more stable form. The transition may also involve a degree of fractionation since more highly ordered crystals have higher homogeneity of triglycerides than less ordered crystal forms. These transitions are irreversible, and converting a more stable form into a less stable form requires complete melting and heating to well above the melting point of the most stable form followed by crystallisation under circumstances that promote the formation of the crystal form aimed for. When cooling the fat blend, crystals are generally formed, but this form is never stable for long and is followed by the transformation to the lower energy level and form. In most cases, crystals are relatively slowly transformed to the stable form (Lutton, 1972), but this requires a certain homogeneity of the triglycerides for instance with respect to chain length. Moreover, the recrystallisation can be postponed by the addition of sorbitan tristearate or slowed down by the addition of diglycerides with a sufficiently high melting point. As the crystal form is preferred for margarine products, it is of interest to manufacture products

BLUK122-Dijkstra

222

September 25, 2007

20:3

Trans Fatty Acids

exhibiting this crystal form (Wiedermann, 1978). Please refer to Chapter 8, Table 8.4 for the list of various fats and their crystal habit. Crystallisation is theoretically considered to occur in a three-stage process, involving supersaturation, followed by nucleation and finally crystal growth (Grall and Hartel, 1992; Timms, 1995). Nucleation depends on temperature, time and fat composition. Nucleation is a process of molecular packing, and two mechanisms called primary and secondary nucleation (Ng, 1989) can be distinguished. Primary nucleation involves the formation of the first crystals called nuclei. During crystallisation, molecules aggregate resulting in energy liberation because the solid phase has a lower free energy than the liquid phase. This favours the process of nucleation. In practice, nucleation and crystal growth occur simultaneously. Secondary nucleation occurs when small pieces of the growing crystal are separated from the crystal surface; it is promoted by seeding with higher melting triglycerides or by stirring (Timms, 1995). As secondary nucleation creates a larger number of crystals and these can form intercrystal bonds, secondary nucleation is considered to be desirable in margarine production. It is not only the polymorphic form of the fat that is of interest when producing margarine and related products, but the speed at which the fat blend will crystallise (the crystallisation rate) is also of interest. The crystallisation rate of the fat blend has to be taken into account when configuring the crystallisation line, as the rate affects the amount of residence or holding time in the equipment that the emulsion needs in order to be ready for packing or filling at the end of the processing line. As examples of different crystallisation rates, coconut oil needs only 3 min at 20◦ C, whereas palm oil needs 27 min at 10◦ C (Gerstenberg, 1996, p. 27). The crystallisation of melted fat is a process of phase transition of molecules from liquid to solid state, the driving force being the difference in melting point of the crystals and the actual solution temperature (Grall and Hartel, 1992). In practice, the triglycerides interact to form solid solutions (Timms, 1984). The crystallisation rate can be measured in various ways. Danisco A/S uses NMR equipment to measure the increase in SFC over time at constant temperature (Danisco, 2006b). Differential scanning calorimetry can also be used in order to determine crystallisation rates of fats and fat blends (Gerstenberg, 1996, pp. 57–59). In industry, the crystallisation rate is reckoned with but it is rarely measured; however, it is generally accepted that palm oil crystallises at a relatively low rate, and thus the necessary processing actions are taken when blends containing only palm oil or high amounts of palm oil are used. If such actions are not taken, the product may be insufficiently crystallised when it leaves the crystallisation equipment. It may be too soft to be wrapped, and when filled in tubs, it may become unacceptably hard by further crystallisation and crystal sintering as result of formation of a three-dimensional network. Bakery margarines may lose some of their functionality because of this post-packaging crystallisation.

9.3

MARGARINE AND RELATED PRODUCTS

On the world market today, a great variety of margarine and related partially crystallised products can be found. Often a distinction is made between the consumer product segment and the industrial product segment. The industrial product segment also covers food service products. Table 9.2 shows some examples of traditional fat blends for margarine and shortening products, most of which contain TFA.

BLUK122-Dijkstra

September 25, 2007

20:3

Food products without trans fatty acids Table 9.2

223

Examples of margarine and shortening fat blends.

Fats Fully hydrogenated vegetable oil Palm stearin m.p. 51◦ C Partially hydrogenated palm oil m.p. 42–43◦ C Partially hydrogenated vegetable oil m.p. 41–42◦ C Partially hydrogenated vegetable oil m.p. 34–35◦ C Palm oil Coconut oil Liquid oil Usage temperature (◦ C) SFC values at 10◦ C 20◦ C 30◦ C 40◦ C Approximate trans content (%)

Soft table Table

Cake

Puff Liquid Pumpable All-purpose pastry shortening shortening shortening 5 5 50

25

20

20

20

40

60

15 25

50

25

40

30 75 5 19 9 2 0 4–8

15

5–10 20–25 20–25 28 45 48 14 22 28 3 4 20 0 0 4 10–13 16–20 8–13

95

50

35

Ambient 5 2 1 0 <3

Ambient 39 24 9 0 10–17

20–25 45 27 8 0 23–26

It can be noted that traditional margarine blends are high in partially hydrogenated fats that offer the advantages of oxidative stability, relatively fast crystallisation rate and excellent plasticity in the final product. The partially hydrogenated fats can be replaced by either fractionated fats, interesterified blends or combinations of fractionated fats blended with liquid oil and/or interesterified blends. Interesterified blends typically comprise either tropical oils (e.g. palm oil and especially hard palm stearin (Sahasranamam, 2005) mixed with palm kernel oil and/or coconut oil) or a fully hydrogenated oil, such as soya bean oil mixed with liquid oils in various ratios. The interesterification products based on tropical oils can also be fully hydrogenated (Delfosse, 1971; Graffelman, 1971). This increases their SFC so that less hardstock is required and more liquid oil can be included in the fat blend; this is especially relevant for fat blends that aim at a high content of polyunsaturated fatty acids. Soya bean oil is predominantly used in the US and tropical oils are used in Europe and elsewhere. It has been reported that tub margarine with suitable spreadability and emulsion stability can be produced from interesterified soya bean oil (List et al., 1995). Interesterified fully saturated soya bean oil mixed with soya bean oil in the ratio 50:50 has also been suggested for low-TFA margarine products (Lee Kok et al., 1999). When replacing a fat in a blend, it is important to consider the physical and chemical properties of the fat as described in Chapter 5, but in industry, the substitution often only entails interchanging a fat from the same group, as shown in Table 9.1, in a proportion that will result in an unchanged SFC profile of the blend concerned. It could be expected that products with similar SFC profiles should be processed the same way, but this is rarely the case. Often completely different processing parameters will apply when major changes in a fat blend have been made (Kirkeby, 2006). Table 9.3 shows examples of fat blends for trans-free margarine and related products.

BLUK122-Dijkstra

224 Table 9.3

September 25, 2007

Trans Fatty Acids Examples of trans-free margarine and shortening fat blends.

Fats Fully hydrogenated vegetable oil Palm stearin m.p. 51◦ C Interesterified blend PK4a Palm oil Coconut oil Liquid oil Usage temperature (◦ C) SFC values at 10◦ C 20◦ C 30◦ C 40◦ C Approximate trans content (%) a

20:3

Soft table Table

Cake

Puff Liquid Pumpable All-purpose pastry shortening shortening shortening 5

25

20 15

50 15 25

5

20

60

80

75

65

25 35 15 25

10

95

35

5

5–10

20–25

20–25

Ambient

Ambient

25–30

19 13 6 1 <3

25 14 6 1 <3

47 26 10 2 <3

50 30 21 8 <3

5 2 1 0 <3

34 22 9 0 <3

48 23 15 4 <3

PK4 is an interesterification product of palm stearin and coconut oil as supplied by Cargill, Harburg, Germany.

It is also reported in the literature that TFA-free vanaspati can be produced by ternary blends of palm oil, a palm oil fraction and palm kernel olein (Idris et al., 1999). Vanaspati was originally developed in the 1930s in India as an all-vegetable alternative to ghee, the traditional cooking fat made from milk fat, from either cow or buffalo milk. The typical vanaspati is high in TFA since its formulation is based on a single partially hydrogenated oil, e.g. soya bean, rapeseed, cottonseed or palm oil. The ternary test formulations resulting in grainy vanaspati products consisted of either palm oil, palm stearin and palm olein (industrially, these ternary blends would be produced more cheaply by blending a higher proportion of palm oil with just palm stearin) or palm oil, palm stearin and palm kernel olein in various ratios. Both sets of formulations resulted in grainy vanaspati products. Vanaspati is used as an all-round household fat for baking and frying. A uniform granular structure is expected. However, the texture depends on the local tradition, e.g. a granular texture with no separation is preferred in India, whereas in Pakistan the product is expected to exhibit larger, softer granules that can be dispersed in its supernatant liquid (Podmore, 2002). On the market today, vanaspati made of 100% palm oil is found, and ghee made of 100% anhydrous milk fat is being exported from Europe to the Middle East. Table 9.3 shows that an interesterified blend such as PK4 can replace the partially hydrogenated vegetable oils with a melting point of 41◦ C, which are the hardstocks of the traditional blends. The resulting triglyceride composition is quite different since the partially hydrogenated vegetable oil has a high content of TFA, whereas the interesterified blend PK4 has less than 2% TFA content. On the other hand, since PK4 is based on palm stearin and lauric oil (Fondu and Willems, 1972), it has a high (∼74%) content of saturated fatty acids compared with the partially hydrogenated hard stock. As listed in Table 9.4, the PK4 has a melting point range of 37–42◦ C and it can also be used for replacing fats in the semi-soft fat group. In that case, a smaller amount of the PK4 compared with the partially hydrogenated vegetable oil that must be used as the PK4 shows a significantly higher SFC value at 30◦ C compared with the SFC at 30◦ C of a semi-solid hydrogenated soya bean oil with a melting point of 35◦ C. However, the SFC values given in Tables 9.2–9.4 must be regarded as only

BLUK122-Dijkstra

September 25, 2007

20:3

Food products without trans fatty acids Table 9.4

SFC profiles of selected raw materials.

SFC (%) 10 20 30 40 Slip melting point (◦ C) a

225

Hydrogenated soya bean oil

Hydrogenated palm oila

PK4

77 49 14 6 35

93 81 51 11 42

72–80 49–55 23–28 3–7 37–42

Fractionated and hydrogenated palm oil.

indicative since the SFC values for the fats may differ for products from one refinery to another. There are other interesterified blends for margarine and shortening production available on the market, including those produced by Aarhus Karlshamn that supplies a range of interesterified blends under the name Cisao, ADM provides a wide product group of lowtrans solutions under the umbrella of NovaLipid, Loders Croklaan has the Sanstrans range and Bunge Oils offer a broad range of trans-reduction options (Kodali and List, 2005). In many different areas of the world, low-trans blends are compounded by using only palm oil and palm oil fractions. Of course, in the palm-oil-growing area the use of 100% palm oil blends is natural but 100% palm oil products are also on the market in Europe. Unfortunately, 100% palm oil blends require special processing parameters due to the slow crystallisation rate of palm oil. It has been reported that the transition time from to is extremely long for palm oil compared with other fats (Timms, 1994).

9.4

MANUFACTURING PROCESS

The manufacturing process for crystallised fat products covers every step from the recipe to final tempering, if applied. For crystallised fat products, the heart of the processing line is the crystallisation equipment, which includes scraped-surface heat exchangers (SSHE), pin rotor machines and a resting tube. Apart from the crystallisation equipment, a modern processing facility for margarine and related products, as illustrated in Plate 9.1, will typically include various tanks for (a) oil storage, (b) emulsifiers, (c) water phase and (d) emulsion preparation; the size and number of tanks are calculated in accordance with the capacity of the plant and its product portfolio. The facility also includes (e) a pasteurisation unit and (f) a remelting facility. Thus, the manufacturing process can in general be divided into the following subprocesses, as shown in Plate 9.1: 1. 2. 3. 4. 5.

preparation of the water phase and fat phase (zone 1); emulsion preparation (zone 2); pasteurisation (zone 3); chilling, crystallisation and kneading (zone 4); packing and remelting (zone 5).

The aqueous phase (zone 1) is often prepared batchwise in the water-phase tank (c). If milk constituents are to be part of the aqueous phase, the dairy product concerned is preferably

BLUK122-Dijkstra

226

September 25, 2007

20:3

Trans Fatty Acids

reconstituted by dispersing and dissolving spray-dried material, as indicated in Plate 9.1. The water to be used for the preparation of the aqueous phase should be of good drinking quality. If drinking quality water cannot be guaranteed, the water can be pre-treated with, e.g. UV irradiation or a filter system. Apart from the water, the water phase may comprise salt or brine, milk proteins (table margarine and low-fat spreads), sugar (puff pastry), stabilisers or thickening agents (reduced and low-fat spreads), acidulants like citric acid, chelating agents like ethylene diamine tetra acetate (EDTA), preservatives like potassium sorbate and watersoluble flavours. The various fats and oils for the fat phase are stored in oil-storage tanks (a), as either fat blends or single oils (fat blend components). The storage tanks are typically located outside the production facility, and they are kept at a constant temperature slightly above the melting point of the fat and agitated in order to avoid fractionation of the fat and facilitate handling. Nitrogen blanketing is recommended. Apart from the fat blend, the fat phase may typically comprise minor fat-soluble ingredients, such as emulsifiers, lecithin, flavours, colouring compounds (-carotene) and possibly antioxidants. These minor ingredients are dissolved in the fat blend before the water phase is added, thus before the emulsification process. The emulsion is prepared (zone 2) by transferring various oils and fats or fat blends to the emulsion tank (d). If the fat blend is prepared in situ, the high-melting fats or fat blends are usually added first, followed by the lower melting fats and the liquid oil. To complete the preparation of the fat phase, the emulsifier and other oil-soluble minor ingredients from the emulsifier tanks (b) are added to the fat blend. When all the ingredients for the fat phase have been properly mixed, the aqueous phase is added and the emulsion is formed under intensive but controlled mixing in the emulsion tank (d). For low-fat spreads containing more water than fat, the aqueous phase can be sprinkled onto the fat surface and the agitation will be quite gentle. Typically, a two-tank batch system is used for preparing the emulsion in order to be able to run the crystallisation line continuously. The two tanks (d) are discharged in turn. Different systems can be used for metering the various ingredients for the emulsion; two of these systems, the flow meter and weighing tank systems, operate batchwise. A continuous in-line emulsification system is a less preferred solution that is used in, e.g. high-capacity lines where limited space for emulsion tanks is available. This system employs dosing pumps and mass flow meters to control the ratio of the phases added into a small emulsion tank. This tanks acts as a buffer tank and also receives the emulsion originating from the remelting system (f). The flow meter system is for batchwise emulsion preparation, in which the various phases and ingredients are measured by mass flow meters (l) when transferred from the various phase preparation tanks into the emulsion tank. The accuracy of this system is ±0.3%. This system has the advantage of being insensitive to outside influences like vibrations and dirt. Like the flow meter system, the weighing tank system is also for batchwise emulsion preparation. Here the amounts of ingredients and phases are added directly to the emulsion tank that is mounted on load cells that control the amounts added to the tank. Accurate dosage of the minor ingredients is ensured by weighing on a separate balance or volumetrically in a measuring cylinder. The emulsion is normally pumped continuously from the buffer tank through either a plate heat exchanger (PHE) or a scraped-surface heat exchanger (SSHE) (e) for pasteurisation prior to entering the crystallisation line (zone 3). For full-fat products, a PHE comprising heating and cooling sections is typically used. For lower fat versions, where the emulsion is expected to exhibit a relatively high viscosity, and for heat-sensitive emulsions (e.g. emulsions with

BLUK122-Dijkstra

September 25, 2007

20:3

Food products without trans fatty acids

Fig. 9.2

227

High-pressure pasteuriser.

high protein content), the SSHE solution is recommended (e). The high-pressure pasteuriser solution is shown in Fig. 9.2. The pasteurisation process has several advantages. It inhibits growth of bacteria and yeasts, thus improving the microbiological stability of the emulsion and the final product. Pasteurisation of the water phase only is also a possibility, but pasteurisation of the complete emulsion is preferred since this will minimise the time lag between the pasteurisation of the emulsion and the filling or packing of the final product. Also, since the product is processed in-line from pasteurisation to filling or packing of the final product, pasteurisation of any rework material is ensured when the complete emulsion is pasteurised. The pasteurisation of the complete emulsion also ensures that the emulsion is fed to the crystallisation line at a constant temperature and thus permits constant processing parameters, product temperatures and thus a consistent final product texture. In addition, the occurrence of an emulsion exhibiting crystalline memory being fed to the crystallisation equipment can be prevented by ensuring that the emulsion is properly pasteurised and fed to the high-pressure pump at a temperature 5–10◦ C higher than the melting point of the fat phase. The emulsion is pumped to the crystallisation line (zone 4) by means of a high-pressure piston pump (g). The crystallisation line for the production of margarine and related products typically consists of a high-pressure SSHE (h), which is cooled by ammonia or a freon-type cooling medium (hydrogen–chlorine–fluorine–carbon (HCFC, also called transitional/service refrigerants) and halogen–fluorine–chlorine (HFC, also called chlorine-free refrigerants)). Figure 9.3 shows a Kombinator 250L, and Fig. 9.4 shows a Perfector 125. Pin rotor machine(s) (i) are normally included in the line in order to add extra kneading intensity and residence time for crystallisation, as shown in Fig. 9.5. A resting tube (j) is the final step of the crystallisation line and it is included only if the product is to be packed in blocks or slabs (k).

BLUK122-Dijkstra

228

September 25, 2007

20:3

Trans Fatty Acids

Fig. 9.3

Kombinator 250L.

The heart of the crystallisation line is the high-pressure SSHE, in which the warm emulsion is supercooled and its fatty phase is crystallised on the inner surface of the chilling tube. Because of this deep supercooling, the polymorph is the predominant form of crystallisation. The partially crystallised emulsion is efficiently scraped off by the rotating knives so that the very cold crystals are mixed with warmer emulsion, which causes them to melt and allow crystals to grow. On a macroscopic scale, the emulsion is chilled and kneaded simultaneously.

Fig. 9.4

Perfector 125.

i

Plate 4.1 AgNO3 -TLC fractionation of FAMEs from a margarine. Development in toluene. Bands were visualised after spraying with 0.1% solution of 2 ,7 -dichlorofluorescein in ethanol and examining under UV light (234 nm).

Origin

cis,cis,cis-18:3

trans-18:3

cis,cis-18:2

cis,trans-18:2

September 28, 2007

cis-18:1

trans-18:1

Saturates

BLUK122-Dijkstra 19:17

ii

5

f

Plate 9.1

2

Process flow sheet.

l

l

g

b

e

1

j

k

3

5

h

c

j

l

1

4

k

i

l

September 28, 2007

d

a

BLUK122-Dijkstra 19:17

BLUK122-Dijkstra

September 25, 2007

20:3

Food products without trans fatty acids

Fig. 9.5

229

Pin rotor machine.

When the fat in the emulsion crystallises, the fat crystals form a three-dimensional network entrapping the water droplets and the liquid oil, resulting in products of a plastic, semi-solid nature. Depending on the type of product to be manufactured and the type of fats used for the particular product, the configuration of the crystallisation line (i.e. the order of the chilling tubes and the pin rotor machines) can be adjusted to provide the optimum configuration for the particular product. After the product is chilled in the SSHE, it enters the pin rotor machine(s) in which it is kneaded for a certain period of time and with a certain intensity in order to assist the promotion of the three-dimensional network, which on the macroscopic level is the plastic structure consisting of primary and secondary bonds (Haighton, 1976). Primary bonds are strong and are not readily re-established when broken by mechanical treatment. Secondary bonds are weak and readily re-established by applying mechanical treatment. It is generally accepted that insufficient mechanical treatment results in products exhibiting less plasticity than products manufactured where sufficient mechanical treatment has been applied. The mechanical treatment results not only from the scraping and kneading in the SSHE but also from the kneading in the pin rotor machine(s). If the product is to be distributed as a wrapped product, it will be subjected to a further SSHE treatment before it settles in the resting tube prior to wrapping. If the product is filled into tubs, no resting tube is included in the crystallisation line. If the texture is not optimal or if packaging has to be temporarily interrupted due to mechanical hitch of the packing machine, the product will be diverted to the remelting system, melted and added to the buffer tank for reprocessing. This ensures that the crystallisation plant

BLUK122-Dijkstra

230

September 25, 2007

20:3

Trans Fatty Acids Table 9.5

Recommended residence times and SSHE area requirements.

Puff pastry margarine Approx. time (min) Table margarine Soft table margarine Low-fat spreads Shortenings

Pin rotor machine (s)

Resting tube (s)

Surface area at 1000 kg/h (m2 )

60–180 1–3 36–72 110–145 110–145 145

360–540 6–9 36–72 — — —

0.8–1.4 0.56 0.56 0.62–1.12 0.37–0.56

continues to run while the temporary problem is being fixed. Different remelting systems are available but the most widely used systems are PHE or low-pressure SSHE (f) (zone 5). In many factories, margarine, like other food products, is today produced under strict traceability procedures. These procedures typically cover the ingredients, the production and the final product. They result not only in enhanced food safety but also in a more constant food quality. Traceability requirements can be built into the (automated) control system of the factory and various systems have been developed and designed to control, record and document important conditions and parameters concerning the complete manufacturing process. CIP cleaning plants (CIP, cleaning in place) are also part of a modern margarine facility since margarine production plants should be cleaned on a regular basis. For traditional margarine products, once a week is a normal cleaning frequency. However, for sensitive products like low-fat (high water content) and/or high-protein-containing products, shorter intervals between the CIP are recommended, and in some plants it is every 48 h (Kirkeby, 2006).

9.5

OPTIMAL PROCESSING CONDITIONS

The optimal processing conditions depend on the crystallisation process, i.e. the processing parameters, which include the residence time in the SSHE, the pin rotor machines and the resting tube, if present. In order to successfully use the plant recommendations given in Table 9.5, it is presumed that the SFC profile for the blend corresponds to the application, the correct type and amount of emulsifier have been chosen and the crystallisation plant has the correct configuration, since all these factors have a great influence on the characteristics of the final margarine and related products. The capacity of the line is determined by the cooling surface of the SSHE. Different-size SSHE are on the market today, and they are chosen according to the capacity requirements and the range of products to be manufactured. The surface area that should be available for producing 1000 kg/h is listed for various products. It is not a fixed cooling surface area in square metres but a range, as certain formulations within the same product group require more or less cooling surface than other formulations in order to produce 1000 kg/h. The typical configuration for producing puff pastry margarine is a sequence of cooling in an SSHE, followed by kneading in a pin rotor machine and final cooling in a second SSHE before the product enters the resting tube prior to packing. For example, in order to produce 1000 kg/h of traditional puff pastry margarine containing TFA, the general recommendations would require 1 m2 of cooling surface area. The residence time in the pin rotor machine should be 3 min, which corresponds to a volume of 50 L, and 7 min in the resting tube,

BLUK122-Dijkstra

September 25, 2007

20:3

Food products without trans fatty acids

231

which corresponds to a volume 120 L for a rate of production of 1000 kg/h. In general, maximum cooling is applied when puff pastry margarine is produced, often resulting in a product temperature ranging from 12 to 18◦ C before it enters the resting tube. The product temperature depends among other things on the SFC profile of the blend: a higher SFC profile blend results in a higher product temperature prior to entering the resting tube. However, if the puff pastry margarine is based on a 100% palm oil blend, it is recommended to decrease the capacity to as low as 70% of the traditional puff pastry margarine processing parameters. This increases the residence time in the SSHE, the pin rotor machine and the resting tube. This will mean that the line will produce only 700 kg/h on the 1 m2 cooling surface area available. In some cases, especially for high-palm-stearin blends, a change in the configuration of the crystallisation line has proven to be an advantage resulting in a sequence of cooling in the SSHE before kneading in the pin rotor machine prior to entering the resting tube. In this case no intermediate kneading is applied, but the texture of the product is according to the requirements. Products with good plasticity have been achieved by limited cooling in the first section of the crystallisation line and more in the last section in order to give the slowly crystallising blends maximum cooling at the end of the line. On a pilot plant scale, palm-oil-based puff pastry margarine with good plasticity has been produced by including some intermediate resting time by the incorporation of a resting tube between the SSHE units. Because of the slow crystallisation rate of the palm oil, the palm oil blend starts to crystallise after the initial cooling in this resting tube, before further cooling is applied subsequently. For puff pastry margarines based on interesterified blends, it is generally accepted that the margarine has a lower temperature in order to have a packable texture. To achieve this lower product temperature, the obvious process adjustment would be to decrease the capacity. But interesterified blends are sensitive to mechanical treatment, and thus more intensive cooling is recommended instead. In addition, less mechanical treatment in the pin rotor machine is recommended in order to avoid overworking of the product, a loss in plasticity and a greasy consistency when kneaded are typical signs of an overworked product. The volume of the pin rotor machines can be reduced if the crystallisation line includes more than one pin rotor machine by physically bypassing the pin rotor machine, or the rotation per minute (rpm) of the shaft of the pin rotor machine can be decreased to minimum in order to minimise mechanical treatment. Resting tubes are delivered with sieves, the number of which depends on the size of the resting tube and the crystallisation behaviour of the products to be manufactured on the line. The first sieve is placed at the inlet of the resting tube to ensure an even distribution of the crystallised product into the resting tube and also to avoid channelling of product in the resting tube. The resting tube is equipped with a jacket for tempered water to avoid the product sticking to the wall and concomitant channelling, which results in a spread in residence time in the resting tube and often softer product on packing. A second sieve (or more) is often placed in the middle section of the resting tube to apply final kneading to the crystal network prior to packing. However, some blends, e.g. blends containing lauric oils and interesterified blends, can be sensitive to this final mechanical treatment, which is often detected on packing, when the product seems overworked and mechanical treatment earlier in the crystallisation process has been minimised. In that case, it is recommended to remove the second sieve from the resting tube. Unfortunately, the residence time in the resting tube differs from blend to blend and it is difficult to make firm distinctions. Naturally, it is of interest from the manufacturer’s point of view to have a short resting time, as this correlates with the amount of product held in the resting tube. This enables the manufacturer to limit

BLUK122-Dijkstra

232

September 25, 2007

20:3

Trans Fatty Acids

the amount of product waste when changing from one product to another or if cleaning the resting tube. However, the residence time in the resting tube is often higher than necessary in order to have the flexibility to change from fast crystallising blends to slowly crystallising blends. The above recommendations also apply to other margarines and related products; however, if the product is filled in tubs or bag-in-box, no resting tube is needed in the line. The processing of TFA-free blends can be improved by using specific emulsifiers that have been developed to facilitate and speed up the crystallisation of slow-crystallising blends containing for instance palm oil fractions or interesterification products. Danisco A/S has, for example, launched a range of emulsifiers under the name of Grindsted® Crystallizer emulsifier blends that are designed to increase the margarine processing speed by acting as a pro-crystalliser. The Grindsted® Crystallizer 400 is a mixture of mono- and diglycerides based on rape seed and/or palm oil and polyglycerol ester of edible vegetable fatty acids (Danisco, 2006a); it has been designed for high-performance puff pastry margarine production. By adding this emulsifier to a TFA-free blend, faster crystallisation, increased plasticity and easier lamination are achieved compared with a product containing a standard emulsifier for puff pastry margarine. The rate of crystallisation was measured by NMR at 28◦ C and the maximum crystallisation level expressed by SFC was reached in 10 min when Grindsted® Crystallizer 400 was used in the blend, compared to 14 min at the same SFC level with the standard emulsifier. The improvement in lamination resulted in a higher lift and a more even lamination structure of the baked samples (Danisco, 2006b). Tempering is the final treatment of the crystallisation process, and the correct tempering temperature and time can improve the quality of the product. This applies not only to industrial products like puff pastry margarine, cake margarine and shortenings but also to consumer products like soft table margarines and spreads. If the latter products are kept at too high a temperature or fluctuations in temperature appear in the storage room or during distribution, this can have a negative impact on the texture and shelf-life. Therefore, products that are sold at refrigerator temperatures should also be stored and distributed at these temperatures while also ensuring a limited fluctuation of product temperature. For industrial products, the tempering temperature and time are often based on tradition in the actual production facility, but the tempering procedure usually prescribes temperatures between 18 and 23◦ C for 5–7 days. For puff pastry margarine, some blends can be very sensitive to the tempering temperature and a difference of only a few degrees can change a good-quality product into a bad one. The plasticity of the product can change into brittleness if tempered at too low a temperature and into stickiness if tempered at too high a temperature. Puff pastry margarine based on blends with a high palm stearin content or interesterification products is particularly sensitive to a less than optimal tempering procedure, but unfortunately finding the optimal procedure is a process of trial and error for every blend composition. Puff pastry margarine exhibiting excellent plasticity has been achieved in a pilot plant by tempering high-palm-stearin blends at 23◦ C for 2 days and then transferring the product to 18◦ C for 5 more days. Puff pastry margarine based on interesterified blends that has been shown to require longer tempering periods at a lower temperature, up to 2 weeks at 16–18◦ C, before the crystallisation of the puff pastry margarine was acceptable has been reported from the market. The choice of tempering procedure also depends on the final application of the puff pastry. If the puff pastry is used in a small bakery, the initial plasticity and texture are very important for the baker for easy handling of the puff pastry margarine. If the same product is used in an industrial facility where the puff pastry margarine is fed to a pump before the lamination procedure, initial brittleness can be an advantage, as the pump will

BLUK122-Dijkstra

September 25, 2007

20:3

Food products without trans fatty acids

233

provide additional intensive mechanical treatment that will affect the structure of the puff pastry margarine. A plastic product without brittleness might not be able to withstand the mechanical treatment from the pump and a greasy product fed to the laminating line will be a result.

9.6

FINAL REMARKS

When producing margarine and related products, it is important to keep in mind that it is not only the ingredients like the oils and fats used or the recipe of the product that determine the quality of the final product but also the configuration of the plant, the processing parameters and the state of the plant. If the line or the equipment is not well maintained, there is a risk that the line will not perform efficiently. Therefore, to produce high-quality products, a wellfunctioning plant is a must, but the choice of fat blend with characteristics that correspond to the final application of the product is also important, as well as a correct configuration and choice of processing parameters of the plant. Last but not least, the final product must be temperature treated as appropriate to the final use.

REFERENCES Berger, K.G. (1989) Functionality and interchangeability of fats. Lipid Technol 1, 40–43. Danisco (2006a) Product description – PD 216862-2.OEN. Grindsted® Crystalliser 400 Emulsifier Blend. Danisco (2006b) Puff up a trans-free pastry. Technical information on Grindsted® Crystalliser 400 Emulsifier Blend for high performance production. Delfosse, J.K.F. (1971) Margarine fat composition. British Patent 1 244 868, assigned to Unilever. Fondu, M.P.V. & Willems, M.A.G. (1972) Margarine fat containing randomized fat component. US Patent 3,634,100, assigned to Unilever. Gerstenberg, P. (1996) Production of puff pastry margarine with a low content of trans fatty acids. M.Sc. Thesis. Department of Dairy and Food Science, Royal Veterinary and Agricultural University, Copenhagen, Denmark. Graffelman, H.A. (1971) Margarine fat and process for preparing same. US Patent 3,617,308, assigned to Unilever. Grall, D.S. & Hartel, R.W. (1992) Kinetics of butterfat crystallization. J Am Oil Chem Soc 69 (8), 741–747. Haighton, A.J. (1976) Blending, chilling, and tempering of margarines and shortenings. J Am Oil Chem Soc 53 (6), 397–399. Idris, N.A., Che Maimon, C.H., Hanirah, H., Zawiah, S. & Che Man, Y.B. (1999) Trans-free vanaspati containing ternary blends of palm oil-palm stearin-palm olein and palm oil-palm stearin-palm kernel olein. J Am Oil Chem Soc 76 (5), 643–648. Johansson, D. & Bergenst˚ahl, B.A. (1995) Sintering of fat crystal networks in oil during post-crystallization processes. J Am Oil Chem Soc 72, 911. Kirkeby, P.G. (2006) Margarine production – technology and process. Technical information from Gerstenberg-Schr¨oder, www.gs-as.com. Kodali, D.R. & List, G.R. (2005) Trans Fats Alternatives. AOCS Press, Champaign, IL. Lee Kok, L., Fehr, W.R., Hammond, E.G. & White, P.J. (1999) Trans-free margarine from highly saturated soybean oil. J Am Oil Chem Soc 76 (10), 1175–1181. List, G.R., Pelloso, T.A., Orthoefer, F.T., Chrysam, M.M. & Mounts, T.L. (1995) Preparation and properties of zero trans soybean oil margarines. J Am Oil Chem Soc 72 (3), 383–384. Lutton, E.S. (1972) Lipid structures. J Am Oil Chem Soc 49 (1), 1–9. Merker, D.R., Brown, L.C. & Wiedermann, L.H. (1958) The relationship of polymorphism to the texture of margarine containing soyabean oil and cottonseed oil. J Am Oil Chem Soc 35, 130–133. Ng, W.L. (1989) Nucleation behaviour of tripalmitin from triolein solution. J Am Oil Chem Soc 66, 1103–1106.

BLUK122-Dijkstra

234

September 25, 2007

20:3

Trans Fatty Acids

Podmore, J. (2002) Culinary fats: solid and liquid frying oils and speciality oils. In: Fats in Food Technology (ed. K.K. Rajah). Sheffield Academic Press, Sheffield, UK, pp. 348–351. Sahasranamam, U.R. (2005) Trans-free hard structured fat for margarine blend and spread. European Patent 1 552 751 A1, assigned to Premium Vegetable Oils Berhad. Stanley, J. (2007) The implications of recent research on trans fatty acids. Lipid Technol 19 (1), 16–17. Timms, R.E. (1984) Phase behaviour of fats and their mixtures. Prog Lipid Res 23, 1–38. Timms, R.E. (1994) Physical chemistry of fats. In: Fats in Food Products (eds D.P.J. Moran & K.K. Rajah). Blachie Academic & Professional, London. Timms, R.E. (1995) Crystallisation of fats. In: Developments of Fats and Oils (ed. R.J. Hamilton). Blackie Academic & Professional, London, pp. 204–233. Wiedermann, L.H. (1978) Margarine and margarine oil, formulation and control. J Am Oil Chem Soc 55, 823.

BLUK122-Dijkstra

October 5, 2007

9:14

Index

1,2-dimethylhydrazide (DMH) 64, 65, 69 1,3-random, 2-random distribution 133, 137 11-cis-vaccenyl ACP 12 11-octadecenoic acid 5 13c-docosenoic acid 5 2 ,7 -dichlorofluorescein 123 2-amino-1-methyl-6-phenylimidazol[4,5-b] (PhIP) 63 2-amino-3-methylimidazo[4,5-f]quinoline 64 3-ketoacyl-ACP 12 5c-tetradecenoic acid 5 9c-tetradecenoic acid 5 9-palmitoleoyl ACP 12 9t-octadecenoic acid 5 polymorph 144, 221, 225, 228 form 134, 135, 205, 206, 207, 208, 212, 221, 225 form 206, 207, 212, 221 -carotene 226 acetic acid 12 acidolysis 191 acrylamide 45 acyl carrier protein 12 adipocytes 74, 75, 79 adipose tissue 8, 30, 39, 42, 44, 47, 61, 62, 77, 78, 80, 105, 124 alcoholysis 192 allylic cleavage 120 alpha parinaric acid 8 alpha-eleostearic acid 18, 20 alpha-linolenic acid 6, 7, 15, 67, 104, 105, 106, 110, 111, 113, 114 alveoli 68 angiogenesis 63, 67, 72, 73, 74 anhydrous milk fat 7 anti-atherosclerotic benefits 55 anti-cancer agents 59 anti-cancer benefits 55 anti-obesity benefits 55 anti-tumourigenic effects 62 apoptosis 63, 67, 68, 69, 70, 72, 78, 79

arachidic acid 3 arachidonic acid 15, 105 Aralia elata 8 artefacts 116 ATBC Cancer Prevention Study 34, 36, 40 atherosclerosis 25, 73, 87 Australia trans fatty acid intake 35 autoclave 160 babassu oil 198 band filters 182, 185 batter 208 beef 8, 37, 54, 105 behenic acid 3 beta-eleostearic acid 8, 18, 20 beta-Carotene and Retinol Efficacy Trial 45 beta-hydroxydecenoyl ACP 12 beta-parinaric acid 8 biohydrogenation 27, 57 biomarkers 30, 37, 41, 54, 61, 90 biosynthesis 12 biscuit fat 210 biscuits 27 bleaching 141, 142, 154 bleaching earths 141 blending 132, 181, 205 blends 219, 220, 221, 222, 223, 226 Brazil trans fatty acid intake 35 bread 27, 30, 32 breakfast cereals 27, 29 breast cancer 60, 61, 68, 69 brownies 28 Bupleurum falcatum 8 butanoic acid 2, 7 butter 28, 37, 38, 61, 77, 103, 104, 210, 214 butterfat 7, 8, 43, 206 butyric acid 2, 7, 61 cachexia 66, 68, 74 cake margarine 208, 223, 224, 232 cakes 27, 205 calendic acid 8 cancer 44, 59

BLUK122-Dijkstra

236

October 5, 2007

9:14

Index

candy 27 canola oil 30, 103, 105, 109, 116, 117, 206 capric acid 3 caproic acid 3 caprylic acid 7 capsules 59 caramel 214 carbon number 134, 135 cardiac arrhythmias 43 cashew 7 catalyst 147, 148, 203 catalyst poisons 153 cell proliferation 63 cell regulation 63 cell signalling mechanisms 72 centrifugal separators 182 ceramide 61 cheese 28, 58, 103, 104, 105 cheeseburgers 30 chelating agents 225 chemical ionisation mass spectometry 120 chemical shifts 21 chicken nuggets 37 Chinese dishes 30 chocolate 36, 183 cis-9, trans-11octadecadienoic acid 54, 56, 57, 63, 66, 67, 70, 72, 74, 76, 78, 79, 80, 82, 86, 91, 92 citric acid 196, 226 cleaning plants 230 cloud point 185 cocoa butter 132, 135, 141, 182, 211, 212, 213 cocoa butter equivalents 186, 189, 192 cocoa butter replacers 154, 186, 212, 213 cocoa butter substitutes 212, 214 coconut oil 140, 141, 192, 198, 206, 208, 209, 211, 220, 222, 223, 224 coffee whiteners 216, 217 columbinic acid 58 composite diets 30 concanavalin 85 confectionery fats 141, 211 conical sieve centrifuges 186 Conioselenium unvittatum 8 Conium maculatum 8 conjugated linoleic acid 9, 40, 54, 116, 144 conjugated linoleic acid dietary intake 59 cookies 29 copper 172, 173 copper chromite 168, 172 corn oil 30, 104 coronary heart disease 25, 39, 40, 41, 58

cost optimisation 136 Costa Rica trans fatty acid intake 35 cottonseed oil 30, 189, 205, 206, 207, 212 cow’s milk 57, 61 COX-2 63, 66, 70, 82, 90 crackers 27, 29 cream 103, 104, 205, 209, 210, 211 creams synthetic 216 critical pairs 116, 118 croissants 27, 220 crumb structure 208 crystallisers 185 cyanopolysiloxane stationary phases 106, 107 dairy fat 37, 59, 103, 127 dairy products 32, 36, 39, 102, 105, 118 dead-end tank reactor 151 decanoic acid 3 delta-9 desaturase SCD1 57 deodorisation 141, 144, 196 desaturase 14 desorption 9, 10, 11, 150 desserts 31 diabetes 42, 43, 45, 73, 80, 81 differential scanning calorimetry 222 diglycerides 183, 195, 196, 221 dimethyloxazoline (DMOX) 120, 121, 122 dimyristooleate 135 dipalmitoelaidate 135 dipamitooleate 135 distearoelaidate 135 distearooleate 135 docosahexaenoic acid 105, 191 docosanoic acid 3 dodecanoic acid 3 dough fats 210 doughnuts 27 drum filters 185 dry fractionation 181, 182, 185, 187 E-configuration 4 eggs 27, 30 eicosanoic acid 3 eicosanoid 63, 64, 65, 67, 73, 74, 82, 87 eicosapentaenoic acid (EPA) 191 eicosenoic acid 113 elaidic acid 4, 5, 19, 57, 103, 134, 139 eleostearic acid 9c,11t,13t 8, 18 ELISA 69 emulsifier 216, 225, 232 emulsion 226

BLUK122-Dijkstra

October 5, 2007

9:14

Index

endothelial dysfunction 42 enolate anion 193, 194 enrobing 214 enzyme-catalysed interesterification 181 enzymic interesterification 192, 197 epidemiology 25 epidermal cancer 66 erucic acid 5, 198 erythrocyte membranes 47 Escherichia coli 12 essential fatty acids 44 Euramic study 41, 44, 46 eutectics 212, 214 fat blends 132 fat crystallisation theory 183 fatty acid nomenclature 2 fatty acid synthetase 12 feedstock flow rate 165 fetal growth 40 fibroblasts 73 filter cake 138 Finland 61 fish 58 fish oil 43, 57, 64, 133, 152, 154, 155, 189, 203, 205, 206, 208, 209, 210 fixed bed reactors 165 food applications 203 food disappearance data 30 Fourier Transform Infrared Spectroscopy 106 fractionation 136, 138, 145, 181 French fries 27, 29, 36, 37, 45 fried chicken 45 fried fish 45 fried potatoes 27 fruits 31 frying fats 127 frying margarine 220 fudge 214 gamma-linolenic acid 6, 7, 15 gas chromatography-mass spectometry (GC-MS) 106 gas chromatography-Fourier transform infrared spectroscopy (GC-FTIR) 106 gastrointestinal cancer 64, 68 genetic engineering 58 geometrical isomerisation 12 Germany trans fatty acid intake 35 ghee 38, 189, 215, 224 glucose 80, 214 glycerol monostearate 217

237

gold 172 gravy 29 ground nut oil 214, 215 half hydrogenated intermediate 149, 153, 157 hard stocks 138, 141, 144, 207, 208 Health Professionals Follow up Study 41, 44 Hedera japonica 8 helium 115, 118 hepatic cancer 66 Heracleum nipponicum 8 herring oil 206 heterogeneous catalysts 150 hexadecanoic acid 3 hexanoic acid 3 high performance liquid chromatography 123 high-pressure pasteuriser 227 homogeneous catalysts 173, 174 Horiuti-Polanyi mechanism 9, 149 human umbilical vein endothelial cells 65, 67, 73, 83, 88, 89 hydrogen 9 hydrogen flow rate 165 hydrogenation 11, 102, 136, 138, 147, 203 hydrogenation brush 132 hyperglycaemia 73, 92 immune function 54, 55, 81, 85, 86, 87 India trans fatty acid intake 35 inflammation 87 infrared spectra 20 insulin 43, 80, 81, 91 insulin-like growth factor (IGF-II) 66 interesterification 133, 134, 136, 139, 141, 181, 191 interesterification, directed 192 interesterified blends 231 interleukin-6 42, 74 intraesterification 141 iodine value (IV) 103, 104, 105, 133, 137, 153, 165, 186 isoform 69 isomerisation index 137, 168, 174 isoprostane 65 isotherms 212, jacaric acid 8 Japan trans fatty acid intake 35 kangaroos 56

BLUK122-Dijkstra

238

October 5, 2007

9:14

Index

labelling 23 laminated biscuits 210, 211 lard 43, 113, 183, 189, 198, 206, 209, 210 lauric acid 3, 39, 141, 144 lauric fats 216 lauric oil 140, 181, 211, 217, 224 lecithin 29, 155, 226, legislation 23 leptin 80 leukotrienes 70 Ligusticum acutifolium 8 linear programming technique 136, 142, 143, 144 linoleic acid 2, 6, 7, 15, 42, 62, 67, 69, 104, 105, 111, 137, 143, 161 linolenic acid 16, 137, 142, 143, 161, 100 lipase 192, 195, Lipofrac process 182 loop reactors 151, 152, 163 low-density lipoprotein 42, 43, 45, 58 low-fat margarine 220 LOX-5 63, 70 maize oil 215 malonyl CoA:ACP transacylase 12 margarine 23, 27, 28, 29, 36, 37, 41, 42, 102, 108, 109, 110, 111, 113, 114, 115, 117, 132, 133, 143, 183, 192, 203, 205 market share data 30 meat 31, 47, 58, 61, 125 meat products 32 medium-chain fatty acids 38 medium-chain triglycerides 196 melting points 17, 18, 19 membrane press filter 187 metal carbonyls 173, 174 metal catalysts 25 metastasis 73 methyl amine additives 169 methyl elaidate 19 methyl oleate 19 methyl petroselaidate 19 Mexican dishes 30 milk 28, 30, 38, 39, 47, 61, 64, 77, 102, 103, 104, 124, 127 milk fat 58, 102, 119, 120, 122, 125, 128, 183, 224 milk proteins 225 Minnesota Heart Study 32, 33 monocytes 81, 82 morphology 183 moulding 214

muffins 27 mutagenesis 63 myocardial infarction 38, 41, 42 myristic acid 3, 39, 135 nanoparticles 171, 172 New Zealand trans fatty acid intake 35 nickel 9, 151, 158, 160, 162, 170, 174 nitrogen blanketing 226 noodles 28 Normative Aging Study 33 nuclear magnetic resonance 22, 222 nucleation 183, 222 nucleation, secondary 184 Nurses Health Study 32, 33, 40, 42, 43, 44 nutraceuticals 59, 62 nutrition facts label 22 obesity 30, 75, 76, 77, 79, 80, 91 octadecanoic acid 2, 3, 4, 5 octanoic acid 7 oil consumption per capita 25, 26 oleic acid 2, 4, 5, 7, 9 , 15, 19, 39, 42, 104, 105, 109, 135, 212 olein 133, 138, 139, 181, 182, 187, 212, 220 oleostearine 209 olive oil 7 omega-3 fatty acids 191 omega-3 polyunsaturated fatty acids 44, 58, 64, 70, 73, 74, 85 omega-6 fatty acids 45 oncogenes 60, 63, 64, 65, 69 ornithine decarboxylase 67 Osmorhiza aristata 8 oxidative rancidity 9 palladium 165 166 167 168 169 170 171 palm kernel oil 26, 140, 185, 188, 198, 214, 220, 223 palm mid-fraction PMF 188, 189, 190 palm oil 25, 37, 43, 136, 139, 140, 141, 152, 184, 188, 206, 207, 209, 210, 211, 214, 215, 222, 223, 224, 225, 231 palmitelaidic acid 8 palmitic acid 2, 3, 12, 42, 135, 137, 212 palmitoleic acid 5, 14 Panax schinseng 8 pancreatic cancer 66, 74 panning and pressing 185 partial hydrogenated vegetable oils (PHVO) 25, 102 , 104, 105, 107, 108, 110, 112 pasteurisation unit 225

BLUK122-Dijkstra

October 5, 2007

9:14

Index

pastry margarines 209, 210, 220, 223, 224, 230, 231, 232 peanut butter 28 peanut oil 113 petroselaidic acid 19 petroselenic acid 4, 8, 19 Phelopterus litoralis 8 phorbol esters 67 phosphatide 156, 159, 183 phosphorus 153, 154, 169 phosphorylation 65, 71 physical refining 141 pies 27, 28 pig 8, 85, 86 pin rotor machines 225, 227, 229, 230, 231 pistachio 7 pizza 30 plasma phospholipid 37, 47 plate heat exchanger 226 platinum 166, 167, 168, 169, 171, 172, 176 polymorph 183, 198, 221, polymorphism 17, 221 polyunsaturated fatty acids 14, 56, 82, 223 popcorn 36, 37 post-menopausal women 61, 62 potassium sorbate 226 potato chips 27, 28, 29, 30 prediction errors 78 preferential selectivity 157, 158, 162 pre-fractionation 118, 127 pressure 157 prostaglandin 84, 89, 90 prostaglandin-E2 64, 82, 83 prostate cancer 62, 63, 66, 68, 69, 70 pro-tumourigenesis 60, 66 puff pastry 183 punicic acid 8 quantification 114 rancidity 148 randomising 132, 133, 139, 191 Raney metal 168 rapeseed oil 26, 113, 142, 198, 205, 206, 210, 211, 212, 214, 215, 221, 224 reactor 148 red meat 30, 39 resting tube 225, 227, 229, 230, 231 rice bran oil 215 rumen 8, 56 rumenic acid 6, 8, 57, 105, 106, 116

239

ruminant meats 27, 36, 103 ruthenium 168, 169, 171, 174 salad dressing 27, 31, 32, 170 sandiness 221 sandy mouthfeel 183, 208 saponification value 133 saturated acid 2 saturated fatty acid content 143 saturation 150 sauces 29 savoury snacks 41 SCID mice 62 Scotland trans fatty acid intake 35 Scottish Heart Health Study 41 scraped surface heat exchanger (SSHE) 225, 226, 227, 228, 229, 230, 231 sea food 31 sesame oil 215 Seven Countries Study 40 sheep 8 short chain fatty acids 38 shortening 23, 27, 28, 29, 171, 204, 221, 230 shortening, pumpable 223, 224 sieve 231 silver nitrate thin layer chromatography 109, 123 slip melting point 134 slurry reactors 150, 151, 163 sn-1 position 212, sn-2 position 212 sn-3 position 212, snack food 32 sodium amide 192 sodium ethylate 196 sodium hydride 192 sodium methylate 196, 197 sodium-potassium alloy 192, 196 solid fat content (SFC) 133, 135, 136, 143, 170, 171, 186, 193, 220, 222, 224, 231 solid solutions 184 sorbitan tristearate 221 soya bean oil, interesterified 223 soya bean oil 25, 26, 38, 42, 43, 103, 104, 132, 142, 152, 158, 164, 166, 193, 204, 206, 209, 211, 212, 214, 215, 221, 223, 225 speciality fats 161 spent catalyst 153 sphingomyelin 61 spreadability 148 spreads 219 stearic acid 2, 3, 4, 39, 42, 135, 137, 208, 212

BLUK122-Dijkstra

240

October 5, 2007

9:14

Index

stearin 133, 138, 139, 140, 182, 220, 223, 232 stick margarines 207 suet 133 sugar 225 sulphur 153, 160, 161, 204 sunflower seed oil 26, 139, 192, 198, 208, 215, 221 supercritical systems 166, 167 suspension crystallisation 185 Sweden trans fatty acid intake 35 Swedish Mammography Cohort Study 64 T cells 85 tallow 132, 139, 187, 206, 210, 220 tamoxifen 64 taxane 64 tempering 183, 232 tetradecanoic acid 3 thiazolidinediones 83 thrombosis 47 thromboxane B2 64 tin 174 tocopherols 142, 192 toffee 214 tonalin, natural 76 traceability 230 Trans Fat Task Force 32 trans-10, cis-12 octadecadienoic acid 54, 56, 63, 64, 65, 66, 67, 70, 72, 74, 76, 78, 79, 80, 86, 87, 88, 91 trans-11 octadecenoic acid 57 trans-2-decenoyl ACP 12 transcription factor 91, 92 TRANSFACT study 102 Transfair study 27, 34, 36, 61 triacylglycerols 1, 2, 78, 80, 132 trickle phase reactors 165 trielaidin 18, 19, 21 trierucin 18

triolein 18, 19, 20, 21 tripetroselaidin 18 tripetroselenin 18 tristearin 21 tri-trans-13-docosenoin 18 troglitazone 80, 83 tropical oils 223, tub margarines 207, 223, 232, tubular chillers 209, 210, tumour necrosis factor alpha receptors 42 tumour necrosis factor (TNF) 42, 74, 82, 89, 90 tuna oil 191 t-vaccenic acid 8, 37 two-stage dosage 154 ultrasonic process 168 ultraviolet spectra 18 USA trans fatty acid intake 33 vaccenic acid 5, 19, 57, 102, 103, 104, 105 vacuum drum filters 182, 185 vanaspati 38, 189, 215, 224 vascular cell adhesion molecules 42 vascular endothelial growth factor 72 vegetables 31 wallabies 56, 58 waxiness 205, 207, 209 whale oil 203, 205, 206 whipped toppings 216, winterisation 132 Women’s Health Study 45, 46 yoghurt 58 Z-configuration 4 zeolite 174 Zutphen Elderly Study 34, 40