Food Chemical Safety, Volume I: Contaminants (Woodhead Publishing in Food Science and Technology)

Food chemical safety Volume 1: Contaminants

Related titles from Woodhead’s food science, technology and nutrition list: Food chemical safety Volume 2: Additives (ISBN: 1 85573 563 6) This volume provides comprehensive information about additives in the food industry. The book opens with an explanation of risk analysis and analytical methods in relation to the use of additives in food products. This is followed by full details of relevant EU and USA regulations. The second part of the book provides information about specific subjects including flavourings, sweeteners and colourings. Making the most of HACCP (ISBN: 1 85573 504 0) Based on the experience of those who have successfully implemented HACCP systems, this book will meet the needs of food processing businesses at all stages of the HACCP system development. The collection is edited by two internationally recognised HACCP experts and includes chapters reflecting the experience of both major companies such as Cargill, Heinz and Sainsbury and the particular challenges facing SMEs. The scope of the book is truly international with experiences of HACCP implementation in countries such as Thailand, India, China and Poland. Foodborne pathogens (ISBN: 1 85573 454 0) A practical guide to identifying, understanding and controlling foodborne pathogens. This book relates current research to practical strategies for risk analysis and control and is designed for both microbiologists and non-specialists, particularly those concerned directly with food processing operations. The first part of the book examines specific microbiological hazards. This is followed by an examination of risk assessment and the concluding section provides a guide to controlling pathogens throughout the supply chain from farmer to consumer. Details of these books and a complete list of Woodhead’s food science, technology and nutrition titles can be obtained by: • visiting our web site at www.woodhead-publishing.com • contacting Customer services (email: [email protected]; fax: +44 (0) 1223 893694; tel.: +44 (0) 1223 891358 ext. 30; address: Woodhead Publishing Limited, Abington Hall, Abington, Cambridge CB1 6AH, England)

If you would like to receive information on forthcoming titles in this area, please send your address details to: Francis Dodds (address, tel. and fax as above; e-mail: [email protected]). Please confirm which subject areas you are interested in.

Food chemical safety Volume 1: Contaminants Edited by David H. Watson

Published by Woodhead Publishing Limited Abington Hall, Abington Cambridge CB1 6AH England www.woodhead-publishing.com Published in North and South America by CRC Press LLC 2000 Corporate Blvd, NW Boca Raton FL 33431 USA First published 2001, Woodhead Publishing Limited and CRC Press LLC ß 2001, Woodhead Publishing Limited The authors have asserted their moral rights. This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. Reasonable efforts have been made to publish reliable data and information, but the authors and the publishers cannot assume responsibility for the validity of all materials. Neither the authors nor the publishers, nor anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged to be caused by this book. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming and recording, or by any information storage or retrieval system, without permission in writing from the publishers. The consent of Woodhead Publishing Limited and CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from Woodhead Publishing Limited or CRC Press LLC for such copying. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress. Woodhead Publishing Limited ISBN 1 85573 462 1 CRC Press ISBN 0-8493-1210-8 CRC Press order number: WP1210 Cover design by The ColourStudio Project managed by Macfarlane Production Services, Markyate, Hertfordshire ([email protected]) Typeset by MHL Typesetting Limited, Coventry, Warwickshire Printed by TJ International, Padstow, Cornwall, England

Contents

List of contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

xi

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Watson, Food Standards Agency, London 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Pesticides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Veterinary drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Persistent environmental chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5 Processing contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.6 Migration from materials and articles in contact with food . . 1.7 Naturally occurring toxicants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.8 Control measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.9 Current and future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.10 Dedication and acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.11 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 4 6 7 8 8 9 11 11 11

Part I Analytical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13

2

Risk analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. R. Tennant, Consultant, UK 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Hazard identification in the food supply chain . . . . . . . . . . . . . . . 2.3 Dose-response characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Exposure analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Risk evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Methods for risk management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

15 15 16 19 21 28 29

vi

Contents 2.7 2.8 2.9

3

4

5

Future trends in risk analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sources of further information and advice . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32 35 35

Analytical methods: quality control and selection . . . . . . . . . . . . . . R. Wood, Food Standards Agency, London 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Legislative requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 FSA surveillance requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Laboratory accreditation and quality control . . . . . . . . . . . . . . . . . 3.5 Proficiency testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Analytical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7 Standardised methods of analysis for contaminants . . . . . . . . . . . 3.8 The future direction for methods of analysis . . . . . . . . . . . . . . . . 3.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix: Information for potential contractors on the analytical quality assurance requirements for food chemical surveillance exercises .......................................................

37 37 38 41 41 47 53 57 61 62 64

Molecular imprint-based sensors in contaminant analysis . . . . . P. D. Patel, Leatherhead Food Research Association 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 The principles of molecularly imprinted polymer-based techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 The development and application of MIP-based sensors . . . . . 4.4 Case studies: contaminant analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6 Sources of further information and advice . . . . . . . . . . . . . . . . . . . 4.7 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

71

Bioassays in contaminant analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L. A. P. Hoogenboom, State Institute for Quality Control of Agricultural Products (RIKILT), Wageningen 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Dioxins and the DR-CALUX bioassay . . . . . . . . . . . . . . . . . . . . . . 5.3 The use of bioassays for other groups of compounds . . . . . . . . 5.4 Future developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

91

71 73 76 79 84 86 87 88

91 92 100 102 102 102

Contents Part II Particular contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

7

8

9

vii 107

Veterinary drug residues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. N. Dixon, Food Standards Agency, London 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Control of veterinary products in the UK . . . . . . . . . . . . . . . . . . . . 6.3 Chemical substances commonly used in veterinary medicines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Surveillance for veterinary drug residues . . . . . . . . . . . . . . . . . . . . 6.5 Analytical methods employed in drug residues surveillance . 6.6 Results of surveillance for veterinary drug residues in the UK (1998) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7 Potential effects on human health of veterinary drug residues in food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.8 Current issues relating to residues of veterinary drugs in food in the UK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.10 Further reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

109

Inorganic contaminants in food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. Harrison, Food Standards Agency, London 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Metals and metalloids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 Nitrate and nitrite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

148

Environmental organic contaminants in food . . . . . . . . . . . . . . . . . . . N. Harrison, Food Standards Agency, London 8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Aromatic hydrocarbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Polycyclic aromatic hydrocarbons . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Dioxins and PCBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.5 Chlorinated hydrocarbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.6 Phthalic acid esters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.7 Endrocrine disrupters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

169

Chemical migration from food packaging . . . . . . . . . . . . . . . . . . . . . . . L. Castle, Ministry of Agriculture, Fisheries and Food, York 9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Chemical migration and the main factors that control it . . . . . 9.3 The range and sources of chemicals in food packaging that pose a potential risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4 Research on health issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5 Regulatory context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

193

109 112 117 132 134 138 143 144 146 147

148 150 163 165

169 171 172 175 182 184 185 186

193 195 199 205 206

viii

Contents

9.6 9.7 9.8

Migration testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Suggested further reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

209 212 217

10 Pesticides I. Shaw and R. Vannoort, Institute of Environmental Science and Research, Christchurch 10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Monitoring pesticides in food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 High risk groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4 The UK’s approach to pesticide surveillance . . . . . . . . . . . . . . . . 10.5 Findings from the UK pesticide monitoring scheme . . . . . . . . . 10.6 Human exposure monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.7 Should we ban pesticides? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

218

11 Mycotoxins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J. E. Smith, University of Strathclyde, Glasgow 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Health implications of mycotoxins . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 Analytical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.4 Application of HACCP systems to reduce mycotoxin presence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.5 Prevention and control of mycotoxins . . . . . . . . . . . . . . . . . . . . . . . 11.6 Conclusion and future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

238

Part III Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The international regulation of chemical contaminants in food ....................................................... T. Berg, Danish Veterinary and Food Administration, Soborg 12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2 The nature of international regulation: Codex Alimentarius . 12.3 Decision making and enforcement mechanisms . . . . . . . . . . . . . 12.4 The Codex General Standard on Contaminants and Toxins in Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.6 Sources of further information and advice . . . . . . . . . . . . . . . . . . . 12.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

218 225 228 229 231 234 236 236

238 241 246 250 253 256 257 261

12

263 263 265 268 271 274 276 277

Contents The regulation of chemical contaminants in foodstuffs in the European Union . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. A. Slorach, National Food Administration, Uppsala 13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2 Scientific advisory committees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.3 Pesticide residues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.4 Veterinary drug residues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5 Mercury and histamine in fishery products . . . . . . . . . . . . . . . . . . 13.6 Other chemical contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.7 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

13

Contaminant regulation and management in the United States: the case of pesticides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. K. Winter, University of California, Davis 14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.2 Pesticide regulation in the US . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.3 Regulatory monitoring of pesticides in the US . . . . . . . . . . . . . . . 14.4 Managing pesticides in foods in the US . . . . . . . . . . . . . . . . . . . . . 14.5 Improving the management of pesticides in foods . . . . . . . . . . . 14.6 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.7 Further information and advice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

279 279 280 282 284 287 288 289 290

14

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

295 295 297 298 301 305 310 311 311 314

This page intentionally left blank

Contributors

Chapter 1

Chapter 3

Dr David Watson Food Standards Agency Room 212 Ergon House 17 Smith Square London SW1P 3WG

Dr Roger Wood Food Standards Agency c/o Institute of Food Research Norwich Research Park Colney Norwich NR4 7UA

Tel: +44 (0)20 7238 6250 Fax: +44 (0)20 7238 6124 E-mail: david.Watson@food standards.gsi.gov.uk

Tel: +44 (0)1603 255298 Fax: +44 (0)1603 507723 E-mail: roger.wood@foodstandards. gsi.gov.uk

Chapter 2 Dr D. R. Tennant 14 St Mary’s Square Brighton East Sussex BN2 1FZ Fax: +44 (0)1273 276358 E-mail: [email protected]

Chapter 4 Dr Pradip Patel Leatherhead Food RA Randalls Road Leatherhead Surrey KT22 7RY Tel: +44 (0)1372 822200 Fax: +44 (0)1372 386228 E-mail: [email protected]

xii

Contributors

Chapter 5

Chapter 9

Dr Ron Hoogenboom Department of Food Safety and Health State Institute for Quality Control of Agricultural Products (RIKILT) PO Box 230 6708AE Wageningen The Netherlands

Dr Laurence Castle Ministry of Agriculture, Fisheries and Food Central Science Laboratory Sand Hutton York YO41 1LZ

Tel: +31 317 475623 Fax: +31 317 417717 E-mail: L.A.P.Hoogenboom@RIKILT. DLO.NL

Chapter 6 Dr S. N. Dixon Food Standards Agency PO Box 31037 Room 409b Ergon House 17 Smith Square London SW1P 3WG Tel: +44 (0)20 7238 6358 Fax: +44 (0)20 7238 6402 E-mail: Stephen.Dixon@food standards.gsi.gov.uk

Chapters 7 and 8 Dr N. Harrison Food Standards Agency Room 234 Ergon House 17 Smith Square London SW1P 3JR Fax: +44 (0)20 7238 5331 E-mail: Nigel.Harrison@food standards.gsi.gov.uk

Tel: +44 (0)1904 462 540 Fax: +44 (0)1904 462 133 E-mail: [email protected]

Chapter 10 Professor Ian Shaw and Dr Richard Vannort Food Safety Programme Manager Institute of Environmental Science and Research PO Box 29-181 Christchurch New Zealand Tel: +64 3 351 6019 Fax: +64 3 351 0010 E-mail: [email protected]

Chapter 11 Emeritus Professor J. E. Smith Department of Bioscience and Biotechnology University of Strathclyde Glasgow G1 1XW Fax: +44 (0)141 553 4124 (via secretary) E-mail: [email protected]

Contributors xiii

Chapter 12

Chapter 14

Dr Torsten Berg Division of Chemical Contaminants Ministry of Food, Agriculture and Fisheries Danish Veterinary and Food Administration 19 Morkhoj Bygade DK-2860 Soborg Denmark

Professor Carl K. Winter Department of Food Science and Technology University of California One Shields Avenue Davis CA 95616-8598 USA

Tel: +45 33 95 60 00 Fax: +45 33 95 60 01 E-mail: [email protected]

Chapter 13 Dr S. Slorach Swedish National Food Administration Box 622 75126 Uppsala Sweden E-mail: [email protected]

Tel: (1) 530 752 5448 Fax: (1) 530 752 3975 E-mail: [email protected]

This page intentionally left blank

1 Introduction D. Watson, Food Standards Agency, London

1.1

Background

Scientific knowledge about chemical contamination of food has grown considerably in recent years. Only a few years ago I wrote that ‘the study of chemical contaminants in food is still a relatively young science’ (Watson, 1993). Since then this area of science has continued to develop, in particular becoming an established part of regulatory reviews of food safety across the world. Chapters 12 to 14 of this book describe this major new development. Chapters 2 to 5 detail growing development of practical methods of detecting, monitoring and managing chemical contamination of food. Practical methods of working on these substances have developed steadily. The middle chapters (6 to 11) summarise and review information about the different types of chemical contaminants. The main groups of chemical contaminants that can be found in food share the following characteristics: • They are not intentionally added to food. • Contamination can happen at one or more stages in food production. • Illness is likely to result if consumers ingest enough of them.

The first of these points distinguishes chemical contaminants from other chemicals in food, e.g. vitamins and additives. The wide range of possible sources of chemical contamination has major resource implications, particularly in controlling chemicals that find a wide range of uses, for example pesticides as opposed to veterinary drugs. In order to ensure consumer and worker protection very careful attention must be given at all stages in food production, unless it is known that contamination with a particular chemical cannot occur at some stages.

2

Food chemical safety

We know most about residues in food of chemicals such as pesticides and veterinary drugs that are used in food production. Companies that produce these chemicals are generally required to convince the licensing authorities that there will not be unsafe levels of residues in food, if the products containing the chemicals are used as instructed. This requirement has generated a huge amount of information, some of it now in the public domain. Less is known about toxins that occur naturally in food, although there has been much chemical research on toxins produced by fungi (mycotoxins).

1.2

Pesticides

Many pesticides have been tested for in food. Organochlorine chemicals, such as DDT, have been included in surveys for residues of pesticides in food for many years. The detection of DDT and related compounds in the environment in the 1960s led to concern that their environmental persistence might cause widespread and lasting damage to ecosystems. Indeed their presence in the environment was linked with reduction in eggshell thickness and hence reduced breeding success of birds of prey. Such concerns led to surveillance for residues of persistent organochlorine pesticides in environmental and food samples. Group detection of these compounds by a common method of analysis aided this surveillance. It also generated data on chemically related compounds that are not pesticides, notably polychlorinated biphenyls (PCBs; see section 1.4), and demonstrated environmental contamination by phthalate esters (section 1.6). The latter esters probably contaminated food samples in the laboratory. PCBs proved to be very persistent contaminants in the environment. Environmentally persistent organochlorine pesticides such as DDT have largely been replaced as insecticides by organophosphorus compounds. This change has brought its own problems. Concerns about possible effects of organophosphorus pesticides on users is leading to their replacement by yet other pesticides such as pyrethroids that are thought to be less hazardous to man and the environment. In the case of pyrethroids their ‘green’ image derives from their origins in ancient usage of chrysanthemum (pyrethrum) products for a variety of purposes. Natural origin of chemicals in food does not mean that the chemicals are safer than residues of man-made ones in what we eat. The control of pesticide residues in food is done via the law in many countries. There is little variation between national laws in this area, but the way the law is applied differs between countries. As for pesticide residues, so also for other aspects of the chemical safety of food, global harmonisation of food law is a distant prospect. This is despite considerable effort in some groups of countries to achieve consistency, for example in the European Union, the MERCOSUR group (Argentina, Brazil, Paraguay and Uruguay) and the Australia New Zealand Food Authority. The FAO/WHO Codex Committee on Pesticide Residues (CCPR) and the Joint Meeting on Pesticide Residues (JMPR) have worked for many years towards globally accepted standards for pesticide

Introduction

3

residues in food, but only recently have these started to be adopted in countries. For all pesticides in food it is essential to check that any residues in food are within maximum residue limits or other standards provided these are based on extensive toxicological testing of the chemicals involved. Maximum residue limits for pesticide residues in food are set by many internationally recognised authorities, notably the European Union, the US Food and Drugs Agency, and the world bodies CCPR and JMPR. A wide range of practical steps can be taken to control pesticide residues in food, including the following: • Make information available to consumers so that they can make an informed choice. It is essential that consumers are properly informed about the food that they purchase. Apart from the fact that this is one of their basic rights, communication on food safety that is poor or absent can lead to ‘scare stories’ in the media with consequential concern for consumers and money lost by food suppliers. • Control the availability and usage of pesticides. A particularly effective method in theory, since it applies at source. This is not necessarily true in practice. If there is an urgent need to control a pest or even where the need is not pressing but the user is several stages divorced from the consumer, there is always the potential for more pesticide to be used than is actually needed, or for the pesticide to be misused (e.g. applied too close to harvest or slaughter). • Limit food contamination by pesticides present in the environment. It is important to recognise that the use of pesticides near to crops and farm animals, and in factories concerned with food production, can lead to residues in food. Obviously this can be particularly difficult to detect if surveillance for residues looks mainly for those pesticides used directly on crops or farm animals. The remedy is to extend surveillance and to remind users to be very careful to avoid adventitious contamination of food at all stages of production right through to marketing. • Police limits for pesticides in food. Global standard setting may be advancing but it is essential that there is effective surveillance. This must include action as well as monitoring. • Advice of the type noted above on avoiding adventitious contamination of food needs to be made both nationally and by management in companies involved with food. The control of pesticide residues in food could include HACCP (Chapter 11) to provide an important element of prevention. Other preventative methods include testing of incoming supplies of raw and other materials, and direction by retailers to suppliers that they must use only a defined list of pesticides in specified ways. • Halt the supply of contaminated food. This rather draconian measure can and has been used in extreme cases. In the UK there are measures in place to do this in the Food and Environment Protection Act 1985. • Apply an open and objective system of controlling the use, safety and availability of pesticides.

4

Food chemical safety

Only where all of these approaches are used is there likely to be the chance of convincing consumers that they are protected against unsafe levels of pesticides in food. Many consumers have now been influenced by the media to think that all residues of pesticides in food are unsafe. This has contributed to the growth in demand for organic food.

1.3

Veterinary drugs

Like pesticides, veterinary drugs have an important role to play in reducing disease and suffering but their use has been brought into question by concerns about residues of them in food. The fact that their use can lead to residues in the food supply was recognised rather later than for pesticide residues. Nevertheless, methods of analysis and surveillance for veterinary drug residues are now well established in many parts of the world. The main classes of veterinary drugs used in farm animals are: • Ectoparasiticides used to control flies, ticks and other skin parasites. These fall under the general heading of pesticides (section 1.2 above). • Antimicrobial agents which are used to treat and prevent diseases caused by bacteria and fungi. • Anthelmintic agents used against worms and flukes (hence the name – helminths include liver flukes). • Anabolic agents to promote growth. This group includes hormones and some antimicrobial substances. • Tranquillisers and beta-agonists which have been used to reduce the risk of harm to animals being taken to slaughter. • Coccidiostats to treat and prevent coccidial parasites in the GI tract.

Antimicrobial agents are of two general types – antibiotics and chemotherapeutic agents. The first of these inhibits microbial growth, the second kills the micro-organisms. Residues of antibiotics have been screened by testing of milk, meat or kidney samples to see if they inhibit microbial growth. This simple, direct approach has been used widely. The testers need to take account of possible false positive results if the sample contains a natural inhibitor, and the variable sensitivity of bacteria used in the test to different antibiotics. The use of chemical methods to measure residues of antibiotics has not been as popular, mainly because they are more labour intensive and hence more expensive than microbial inhibition tests. However, some antibiotics, such as sulphonamides, are not readily detected by bacterial tests. Chemical methods are essential for chemotherapeutic antimicrobial agents, such as chloramphenicol, since they are not detected effectively by bacterial inhibition. Residues of antimicrobial agents tend to be present at their highest levels in the liver and kidney, as well as the site of injection (if that is the route of application). The liver is the main site of biochemical modification of antimicrobial substances, as the body tries to convert them to less toxic

Introduction

5

compounds. Indeed, the body tries to do this to all substances that are foreign to the body. The kidney, as one of the two main sites of chemical excretion, the other being the GI tract, is therefore a good organ to test at slaughter for the presence of residues of veterinary drugs including antimicrobial agents. Residues of veterinary drugs are usually at lower levels in meat than in kidney or liver. Anthelmintic agents are a smaller, less diverse group than antimicrobial agents. They include levamisole and the benzimidazoles, and the newer avermectins. The benzimidazoles include thiabendazole which has also been used as a pesticide on plants and as a food preservative. Chemical methods have been used widely to quantify the residues of these substances in food, as part of national control programmes. They are an important sub-group of the antiparasitic agents that many consider to be essential for good animal husbandry. Anabolic agents are used to promote growth. They include hormones and some antimicrobial substances. The latter are used to inhibit the growth of harmful organisms in the GI tract. They are used at low levels in feeding stuffs and are thought unlikely to lead to residues. Hormonal anabolic agents are banned from use in the European Union. They are either synthetic and therefore clearly detectable, or natural and therefore more difficult to distinguish from hormones that are endogenously found in animal products. The use of immunoassay revolutionised the detection of synthetic hormone residues, much as it did in dope testing of athletes. The potential for consumer exposure to residues of tranquillisers and betaagonists has been tested intensively over the past twenty years. This followed evidence of illegal drug availability and usage. These substances could have immediate effect on health if consumed in large amounts. Action was taken to protect consumers. Surveillance continues to ensure that no detectable contamination of the food supply occurs with these substances (see, for example, VMD, 1995). Coccidiostats are used in both poultry and large farm animals to treat and prevent coccidial parasites in the GI tract. Residue testing for coccidiostats has required a lot of work to develop chemical methods of testing. Veterinary drug residues are controlled in similar ways to pesticide residues at national and international levels. The FAO/WHO Codex Committee on Residues of Veterinary Drugs in Food first met in 1986. It elaborates maximum limits, drawing on the toxicological advice of the Joint Expert Committee on Food Additives and Contaminants. National governments and the same international groupings noted above (section 1.2) have developed similar approaches to setting standards for veterinary drug residues in food, notably assessing residues data for tissues from animals treated under controlled conditions and applying surveillance to assess consumer exposure to residues. Where there are differences in approach from country to country, as for pesticide residues, they tend to be in how the law is applied. This makes it particularly important to monitor imported as well as homegrown animal products, although within supra-national alliances such as the

6

Food chemical safety

European Union the emphasis is on surveillance of home-grown products by each member state rather than checking each others’ exports. The sources of veterinary drug residues in food are clearly more limited in variety than those of pesticides. However, little attention seems to have been given to possible contamination of the environment and hence food by animal waste products containing residues of veterinary drugs. Residues of drugs and their metabolites in faeces could, in theory at least, lead to contamination of soil and hence crops following muck spreading.

1.4

Persistent environmental chemicals

Polychlorinated biphenyls (PCBs; section 1.2) and dioxins have been most widely studied in this category. PCBs were used in a wide variety of industrial applications, for example as dielectrics in transformers. But they are very persistent contaminants, both in the general environment and in human fat. In theory the routes of entry into food are: • uptake from the environment by food producing animals, particularly those with high fat content (as PCBs are lipophilic) • direct contamination of food or animal feed following an industrial accident • migration from packaging into food, just as other chemicals in packaging can migrate into food (section 1.6).

In practice there is little evidence for the migration of PCBs from packaging (JFSSG, 1999a). There is considerably more evidence for the other two routes leading to PCBs in food. Indeed, historical trends can be drawn up for residues of PCBs in human fat, breast milk and fish. There has been a very gradual decline in levels of these organochlorine compounds in the environment, food and human tissues. This is entirely consistent with the persistence of these compounds. There has been considerable work on the individual congeners of PCBs, so much so that analysis for these is now commonplace. Toxicological review of these substances has also been extensive. This has identified dioxin-like PCBs, probably the first time that such close parallels have been drawn between the toxicological and other properties of two major groups of chemical contaminants in food. Dioxins in food and the environment have been intensively studied over the past twenty years (see for example Steering Group on Chemical Aspects of Food Surveillance, 1992a). The term ‘dioxins’ has come to be used for polychlorinated dibenzo-p-dioxins and in some cases also polychlorinated dibenzofurans. Both of these are ubiquitous environmental contaminants. They are highly resistant to breakdown in the environment. They are particularly difficult to study because of the large number of substances involved and the very low levels of detection needed. Nevertheless, much surveillance work has been done on polychlorinated dibenzo-p-dioxins, particularly 2,3,7,8-tetrachlorodibenzo-p-

Introduction

7

dioxin (TCDD). A system of toxic equivalency has been developed to allow comparison of survey results with a complex and growing body of literature on the toxicology of these compounds. The sources of food contamination with dioxins are now known to be many more than was originally asserted when claims were made that incinerators were the main source of dioxins in the environment. Indeed known sources of dioxins now include vehicle exhausts, domestic coal fires, manufacture and use of organic chemicals, and metallurgical processors. Two main types of contamination of food appear to be involved: atmospheric deposition and spreading of sludge, in both cases on farmland. Other environmental contaminants in food include several metals and a yet to be defined number of the organic chemicals used by industry and by-products of industrial activity. Metal contamination of food can occur in a wide variety of ways, including environmental and other sources such as canning. There has been a huge amount of analytical work on metals such as lead in food. Indeed a large part of the periodic table has been covered. Early work on metals identified that analytical quality assurance is a key tool in the surveillance of food for chemical contaminants. It also led to the development of toxicological standards which can be used to define whether or not surveillance results show there is a hazard to consumer health. Both of these types of approaches are now standard in the best surveillance programmes, whether they are on contaminants or additives in food. Surveillance and other research on residues of industrial organic chemicals in food is at a much earlier stage than work on metals in food. From the list of about 50,000 such chemicals one needs to identify those that are likely to contaminate food and pose a risk to consumers. This is by no means an easy task. The sort of factors that one might consider are as follows (Steering Group on Food Surveillance, 1988): • production volume (e.g. exclude chemicals that are manufactured in very small amounts) • pattern of usage • potential for release into the environment • persistence in the food chain • toxicity following oral ingestion.

1.5

Processing contaminants

It has been very difficult to predict which chemicals might be formed during food processing and might pose a hazard to consumers. It cannot be assumed that this class of substances does not exist. There are already a few established examples: • There is evidence that carcinogenic N-nitrosamines can be formed during the production of alcoholic beverages, fermented foods and cured meats (Steering Group on Chemical Aspects of Food Surveillance, 1992b).

8

Food chemical safety

• Carcinogenic polycyclic aromatic hydrocarbons can contaminate smoked food (Bartle, 1991), although the main dietary sources of these compounds in the UK appear to be early in food production. • 3-Monochloropropane-1,2-diol (3-MCPD) and ethyl carbamate have both also proven to be unwanted contaminants that are formed during food processing (JFSSG, 1999b; Food Standards Agency, 2000).

Each of these process contaminants was discovered by studies on the foods in question, either by chance or after research on chemical contamination of the environment. A more systematic approach is needed to identify processing contaminants. If this can be achieved, and this is no easy task, work on the above processing contaminants has shown how they can be researched so that levels are reduced and hence consumers protected.

1.6

Migration from materials and articles in contact with food

Early work on phthalate esters (section 1.2) and several monomers such as styrene, which are used to make plastics, demonstrated that chemical migration can occur from packaging into food. There has been a huge amount of practical work on this over the last thirty years (Gilbert, 1997), much of it on plastics. Thus there are now in place detailed controls on this aspect of plastics in the European Union (EU) and the USA. The controls in the EU have been fully implemented in Great Britain (FCM Unit, 2000). Less is known about chemical migration from other packaging materials. Paper and board have been subjected to surveillance which so far has shown that some chemicals can migrate from it into food (e.g. diisopropylnaphthalenes; JFSSG, 1999c) whilst others probably do not (e.g. PCBs; JFSSG, 1999a). The reason that some chemicals migrate and others do not is unclear at present. It is not due to a layer of packaging between the food and the contaminated packaging. Nor does it appear to be due to a fundamental difference in volatility or other physical property of the chemical migrants. Research on chemical migration from other types or components of packaging material – e.g. glass, wood, cork, coatings, adhesives – has been carried out sporadically. There is now a concerted UK programme to study chemical migration from these so that problems can be identified and dealt with in a consistent way (Working Party on Chemical Contaminants from Food Contact Materials and Articles, 1999).

1.7

Naturally occurring toxicants

These are of three types:

Introduction

9

• toxins produced by microbial contamination of food and raw materials used in food production • toxins produced by crops (in some cases at least to protect the plants from insects) • toxins ingested by food-producing animals.

The first category includes toxins produced by fungi (mycotoxins) and bacteria. The second group includes a wide range of food-producing plants. The third is a small group of marine toxins, mostly produced by dinoflagellate algae, which find their way up the food chain and hence onto our plates. These algal toxins and those produced by bacteria are unusual chemical contaminants in that they are quick to take effect. Most other chemical contaminants, including mycotoxins and inherent toxins in crops, would take some time to cause illness. It is perhaps not surprising that more effort has been concentrated on these quick-acting toxins, than on many of the slower ones from plants or fungi. Considerable progress can be made in protecting consumers from many natural toxins in food by applying straightforward good agricultural practices and through the careful handling of food. For example, crop rotation can reduce mycotoxin contamination, as can keeping stored grain and seeds dry; and bacterial toxins are much less likely to be found in food if HACCP is applied in food production. More sophisticated measures are needed to protect consumers from inherent toxins in crops. Plant breeding can lead to higher as well as lower levels of toxins. Careful testing of new varieties for toxins associated with the plant family involved is a key measure that should be adopted. Some wild forms of food plant, for example potatoes, contain very high levels of toxins (glycoalkaloids in the case of potatoes) which have been reduced over the centuries by selective plant breeding. It is not inconceivable that consumer demand for more ‘natural’ foods might lead to reintroduction of high toxin levels by crossing existing cultivars with wild varieties. Similarly the main line of defence against the ‘red tides’ of algae that cause toxin contamination of bivalve molluscs is to assay the amount of toxin in recently gathered molluscs. If the level is unsafe consumers must be warned not to gather the respective species of shellfish. Much research has been done on the incidence of marine toxins around the world (Leftley and Hannah, 1998). It is now essential to use this and other information on natural toxins in food to protect consumers.

1.8

Control measures

As noted above, general good hygienic practice can go a long way in many cases towards minimising chemical contamination of food. It is not always necessary to resort to sophisticated, expensive methods of analysis to reduce consumer exposure to toxic chemicals in food, although knowing what is there does help considerably, and in most cases provides one with an edge over what can be a series of complex problems.

10

Food chemical safety

Much emphasis has been placed in this chapter on the use of survey information as a source of information in controlling chemical contamination of food. This has helped governments around the world to assess problems in this area. They have found that surveillance can stimulate action as well as press coverage. The key is to ensure that action is taken when problems are found. In no particular order the main options are: • • • • •

control the availability and usage of man-made contaminants limit or eliminate the source of contamination police limits advise halt the supply of contaminated food.

As noted in section 1.2 all of these can be applied to pesticides. There is less room for manoeuvre in the case of natural contaminants, where few limits have been set and advice can become very complicated. In between the extremes of advice and further legislative burden, most chemical contaminants can be controlled by using at least two or three of the above options. The key is defining as exactly as possible what the problem is and then taking the most appropriate action. In many cases a balance must be drawn between competing factors. For example, reduction in nitrosamine levels in cured meats must be done without prejudicing protection provided by nitrite against Clostridial contamination, although nitrite is the precursor of nitrosamines in cured meat. The slowly lengthening list of international limits for chemical contaminants in food promises to provide some more flexibility for government control, if not for food producers. International procedures have developed considerably over the past ten years at committees reporting to the Codex Alimentarius Commission and in other multinational trade arrangements, notably in MERCOSUR (the Southern Common Market; section 1.2), the European Union (EU) and the Australia–New Zealand Food Authority. There has been significant progress in agreeing international standards for residues of pesticides and veterinary drugs. Progress has been much slower worldwide on setting limits for mycotoxins and metals in food. Work has yet to begin in earnest on defining controls across the world on chemical migration from packaging, although this is well advanced for plastics in the EU and the USA. The respective committees in the Codex Alimentarius for chemical contaminants are as follows: • Pesticides: Codex Committee on Pesticide Residues (CCPR) and Joint Meeting on Pesticide Residues (JMPR). • Veterinary drug residues: Codex Committee on Residues of Veterinary Drugs in Food (CCRVDF) and the Joint Expert Committee on Food Additives and Contaminants (JECFA). • Other chemical contaminants including natural toxins: Codex Committee on Food Additives and Contaminants (CCFAC) and JECFA.

Introduction

11

The secretariats of these Codex Committees are based at the Food and Agriculture Organisation in Rome. The secretariat of JECFA is based in Geneva at WHO headquarters.

1.9

Current and future trends

The control of chemical contamination of food is clearly developing. If there is a key part of this process at present it is the international harmonisation of controls. The Codex Committee on Food Additives and Contaminants (CCFAC) is actively developing a Codex General Standard for Contaminants and Toxins in Food (Chapter 12). This followed a suggestion from the UK delegation, in 1991, that there was a need to develop a Codex philosophy on contaminants. Work on this Standard has accelerated the use of maximum limits for contaminants in food. It has also stimulated work on a general code of practice for source-directed measures to reduce contamination. Effort now needs to be applied to completing position papers and standards for specific contaminants. At the moment attention at the CCFAC is focused on a relatively small number of contaminants, notably a few of the better-studied mycotoxins (aflatoxins, ochratoxins, fumonisin, zearalenone and patulin), PCBs, dioxins and lead. Indeed this short list may prove rather too long if the respective reviews at CCFAC do not make progress. Delay in agreeing standards would send out the wrong signals about the utility of what is a potentially very valuable Codex Standard in a growing international debate on controlling chemical contaminants in food.

1.10

Dedication and acknowledgement

To Linda. I acknowledge colleagues and friends in the Food Standards Agency, and elsewhere. The views expressed in this chapter are those of the author and not of his employer.

1.11

References

(1991) Analysis and occurrence of polcyclic aromatic hydrocarbons in food, in Food contaminants: sources and surveillance, publ. RSC, Cambridge, UK. FCM UNIT (2000) Explanatory note on the legislation in Great Britain, available from the Food Standards Agency, Room 216, P.O. Box 31037, London SW1P 3WG, Tel. No. +44 (0)20 7238 6528, fax +44 (0)20 7238 6124. FOOD STANDARDS AGENCY (2000) Survey for ethyl carbamate in whisky, Food Surveillance Information Sheet 2/00, available from the Food Standards BARTLE, K.D.

12

Food chemical safety

Agency, Room 303b, P.O. Box 31037, London SW1P 3WG, Tel. No. +44 (0)20 7238 6245/6150, fax +44 (0)20 7238 6330 or email: information [email protected]. GILBERT J. (ed.) (1997) Food packaging: ensuring the safety and quality of foods, ed. J Gilbert, Food Additives and Contaminants 14, publ. Taylor and Francis, Basingstoke, Hants, UK. JFSSG (1999a) Survey of retail paper and board food packaging materials for polychlorinated biphenyls (PCBs), Food Surveillance Information Sheet 174, available from the Food Standards Agency, Room 303b, P.O. Box 31037, London SW1P 3WG, Tel. No. +44 (0)20 7238 6245/6150, fax +44 (0)20 7238 6330 or email: [email protected]. JFSSG (1999b) Survey of 3-monochloropropane-1,2-diol in soya sauce and similar products, Food Surveillance Information Sheet 187, available from the Food Standards Agency, Room 303b, P.O. Box 31037, London SW1P 3WG, Tel. No. +44 (0)20 7238 6245/6150, fax +44 (0)20 7238 6330 or email: [email protected]. JFSSG (1999c) Diisopropylnaphthalenes in food packaging made from recycled paper and board, Food Surveillance Information Sheet 169, available from the Food Standards Agency, Room 303b, P.O. Box 31037, London SW1P 3WG, Tel. No. +44 (0)20 7238 6245/6150, fax +44 (0)20 7238 6330 or email: [email protected] LEFTLEY, J.W. and HANNAH, F. (1998) Phycotoxins in seafood, pp. 182–224 in Natural toxicants in food, CRC/Sheffield Academic Press, ed. D. H. Watson, ISBN 1-85075-862-X. STEERING GROUP ON CHEMICAL ASPECTS OF FOOD SURVEILLANCE (1992a) Dioxins in food, Food Surveillance Paper No. 31, publ. HMSO, London. STEERING GROUP ON CHEMICAL ASPECTS OF FOOD SURVEILLANCE (1992b) Nitrate, nitrite and N-nitroso compounds in food: second report, Food Surveillance Paper No. 32, publ. HMSO, London. STEERING GROUP ON FOOD SURVEILLANCE (1988) Food surveillance 1985 to 1988, Food Surveillance Paper No. 24, publ. HMSO, London. VMD (THE VETERINARY MEDICINES DIRECTORATE) (1995) Annual report on surveillance for veterinary residues in 1995, publ. VMD, Weybridge, UK. WATSON, D.H. (1993) Preface to Safety of chemicals in food: chemical contaminants, Woodhead Publ., Cambridge, UK, ISBN 0-13-787862-1. WORKING PARTY ON CHEMICAL CONTAMINANTS FROM FOOD CONTACT MATERIALS

(1999) Review of current research projects, available from the Food Standards Agency, Room 216, P.O. Box 31037, London SW1P 3WG, Tel. No. +44 (0)20 7238 6528, fax +44 (0)20 7238 6124.

AND ARTICLES

Part I Analytical methods

This page intentionally left blank

2 Risk analysis D. R. Tennant, Consultant, UK

2.1

Introduction

We all share the expectation that food will be safe to eat. However, the opportunities for food to become contaminated by chemicals at some stage in its production are legion. Nevertheless, incidents of chemical contamination are very rare and this is testimony to systems for risk assessment and risk management that are applied by food producers, processors and retailers. Any reliable system for assessing and controlling chemical risks must contain six key elements whose relationships are described in Fig. 2.1: 1. 2.

Hazard identification. It is necessary to be aware of what chemical contaminants might occur in a particular foodstuff and the nature of the harmful consequences to human health that might be associated with them. Dose-response characterisation. Different chemicals will be associated with different toxicological end-points and the risk of any individual experiencing toxicity is related to the dose that they receive. Very often it is

Fig. 2.1

Food chemical risk assessment and risk management.

16

3.

4.

5.

6.

Food chemical safety possible to identify a dose level below which the probability of anyone experiencing an adverse effect is very low or zero. Exposure analysis. The amount of any chemical that an individual is exposed to will depend upon the levels that occur in food and the amounts of those foods that are consumed. Different population groups will often have different levels of exposure and it is therefore necessary to identify such sub-groups. Risk evaluation. If a dose level at which no adverse effects are experienced has been identified then it is necessary to identify any population subgroups whose exposure might exceed that level. Risk evaluation should aim to quantify the level of risk that any such populations are exposed to. Risk management. If any population sub-group has been identified as being potentially at risk then measures to control the risk must be assessed and introduced. Any benefits associated with the foods affected must be taken into account and the costs associated with alternative methods of control evaluated. Risk communication. Where there are potential risks associated with chemical contaminants in food other interested parties, in particular any sub-groups particularly affected, must be informed.

The remaining sections of this chapter will provide detailed information about each of these six steps. However, specific strategies for risk assessment and management may need additional elements, depending on the nature of the hazard, the foods in which it may occur and other specific conditions.

2.2

Hazard identification in the food supply chain

Chemical contamination can occur at any point in the food chain from the field through to the point of consumption (Fig. 2.2). Once contamination has occurred it is usually expensive and technically difficult to remove it. Processing may further complicate the situation, by concentrating the contaminant in a particular fraction such as vegetable oil. It is therefore vital to identify all potential sources of contaminants in order to prevent contamination from occurring. 2.2.1 Primary food production In the field, soil can sometimes be a significant source of contamination. For example, in areas of mineralisation, heavy metals may occur naturally at elevated levels in soils. If there has been mining activity in the past this can exacerbate the problem by the spreading of mine spoil at the surface. Animals, particularly cattle, which tend to tug at forage rather than biting it, can ingest significant amounts of soil and metals can accumulate in tissues such as the liver. Plants may also take up elements from the soil that can accumulate in

Risk analysis 17

Fig. 2.2

Opportunities for chemical contamination in the food chain.

edible parts. Where mineral workings drain to the sea, metals can precipitate where chemical conditions change in estuarine environments. Shellfish caught from such estuaries may also have elevated levels of heavy metals. Soils may also become contaminated with industrial pollutants or with agricultural chemicals. For example, fields located close to industrial plants such as incinerators or metal smelters can gradually accumulate residues of combustion products and other chemicals from the fall-out from smoke plumes. Organo-chlorine pesticides, which are now largely banned, can persist in soils for many years and nitrates used in fertilisers can accumulate in soils which, under certain growth conditions, can result in high levels in certain crops. Animal feeds appear to be particularly vulnerable to chemical contamination. Chemical hazards associated with fungal toxins (mycotoxins) were first identified when poultry were adversely affected. Mycotoxins such as aflatoxins

18

Food chemical safety

and ochratoxins were found to be carcinogenic and to occur at low levels in a variety of plant and animal-derived foods. Many national authorities have taken steps to control the levels of mycotoxins in food. On occasions, animal feed has been suspected of deliberate contamination. Incidents involving contamination of animal feed by industrial by-products such as polycylic aromatic hydrocarbons (PAHs) and combustion products such as dioxins are not uncommon. A problem with animal feed is that there is sometimes inadequate control over the provenance of feed constituents. For example, spent cooking oil from food-processing plants is a legitimate feed component. Unfortunately, the temptation for the unscrupulous to dispose of other unwanted oils in this way is too great for some. In many cases such adulterants are probably diluted to such an extent that they are undetectable by conventional chemical analyses. Nevertheless, they may still represent a longterm cumulative hazard to consumers of products from animals fed on such material. 2.2.2 Food processing It is often necessary to process food before it is suitable for human consumption. Grain must be ground into flour, milk churned into butter, barley and hops brewed into beer, for example. Simple contamination might arise from direct contact with containers and tools used in food processing if they are not made from suitable materials. Machinery lubricants and coolants sometime leak and they can find their way into food. 2.2.3 Retailing and consumption Foods are frequently packaged before being put on sale and constituents of the packaging or inks, dyes and glues used in packaging present the potential for migration into food. During transport, food and other materials can become inadvertently intermingled thus presenting a further hazard. In the home there are further potential hazards such as contamination from storage vessels and cooking containers and utensils. The process of cooking can alter food so that new chemical substances are formed. Cooking meat so that it is well browned (for example by roasting, grilling or frying) can produce heterocyclic amines, which are potentially carcinogenic. Hazards can occur right up to the point of consumption. For example, ceramic glazes with a high lead content can be leached out by acidic foods such as wine. 2.2.4 Hazard characterisation Once potential hazards have been identified the nature of the hazard must be characterised. Initially the nature of any toxicological damage should be identified. For example, cadmium, which can be present at high levels in certain soils and sediments, can cause damage to the kidneys, whereas polycyclic

Risk analysis 19 aromatic hydrocarbons, which are pollutants produced by high-temperature combustion, are carcinogenic. The time-scale over which contaminants can exert an adverse effect is also of importance. For carcinogens such as PAHs and mycotoxins, it is the long-term cumulative dose that is most important. Conversely for organophosphorus pesticide residues, one meal might be sufficient to cause some inhibition of the cholinesterase enzyme.

2.3

Dose-response characterisation

The severity of any adverse effect associated with a chemical contaminant is usually directly related to the dose. Severity can be measured as the degree of damage to an individual or the probability of being affected, or a combination of these effects. For a substance that causes tissue damage such as cadmium, increasing the total dose will tend to increase the degree of damage to the kidneys as measured by the loss of proteins in the urine. For carcinogens, where just one molecule has the theoretical ability to induce a tumour, increasing the dose increases the probability that an individual will contract a tumour. In either case there may be some threshold level below which no effects are observed. 2.3.1 Thresholded end-points For many substances the body’s own mechanisms for de-toxification and repair mean that low doses of some chemicals can be tolerated without experiencing any adverse effects. However, once a certain threshold has been exceeded then the degree of adverse effect is related to the dose. The highest dose at which no adverse effects are observed in the most susceptible animal species is identified at the No Observed Adverse Effect Level (NOAEL). The NOAEL is used as the basis for setting human safety standards for contaminants, Provisional Tolerable Weekly Intakes (PTWIs) or Tolerable Daily Intakes (TDIs).1 The PTWI is defined as the amount that an individual can ingest weekly over a lifetime without appreciable health risk. It is related to the NOEAL so that: PTWI (mg/kg bw/week) = NOEAL (mg/kg bw/day)  UF1  UF2  7 Where: UF1 = Uncertainty factor to allow for extrapolation from animal species to humans. UF2 = Uncertainty factor to allow for inter-individual variability in humans. Uncertainty factors usually have a default value of 100 so that the PTWI is usually equal to the NOEAL  700. If human data are available then UF1 is usually taken to be one. Intakes that exceed the PTWI will not necessarily result in any adverse effect because the uncertainty factors are designed to be conservative. In practice it is

20

Food chemical safety

probable that most people could exceed the PTWI by a considerable margin before suffering any harm. Nevertheless, the probability that an individual will suffer harm (risk) increases once the PTWI is exceeded and so this must be balanced against the costs of control. The level of risk below the PTWI is never quite zero because there is always a residual risk that relates to the lack of absolute certainty in the methods used for toxicological testing. 2.3.2 Non-thresholded end-points Some chemical contaminants are believed to have no threshold below which toxic effects are observed. The most common group of hazards in this respect are genotoxic carcinogens. It is generally understood that the risk associated with a very low dose of a carcinogen is proportionate to the risk associated with a higher dose.2 However, we are all subject to continual low doses of carcinogens in our diet that are derived from natural constituents of plants and man-made chemical contaminants probably make up a very small fraction of our total carcinogen load. The indication is that some protective mechanism exists to neutralise small doses of genotoxic carcinogens. This may be a metabolic process or related to the consumption of natural protective chemicals in the diet. The mechanisms for this effect are still poorly understood but it is clear that if they did not exist then the incidence of diet-related cancers would be much higher than it is. Non-thresholded chemicals that are not carcinogens are less frequently identified. For many years lead was considered to be thresholded because its effects on haemoglobin synthesis were not seen at low doses. However, recent work into the effects of lead on mental development suggest that there may be no threshold for this end-point. Food is a relatively minor source of lead exposure compared with air and dust in urban environments. For chemicals that relate to toxicological end-points that do not show thresholds it is not possible to identify a NOAEL or PTWI. In such cases it is desirable to estimate the level of risk associated with a given level of exposure. 2.3.3 Quantitative Risk Assessment In recent years it has become increasingly apparent that for chemical contaminants that are abundant in the environment a more sophisticated approach to dose-response characterisation is required. There is increasing evidence that small but significant sub-populations are exposed to intakes that exceed PTWIs and most people are exposed to potential carcinogens through their diet. In such cases the PTWI concept is redundant because it is necessary to assess the actual levels of risk to which individuals are exposed in order to introduce proportionate control measures. Simply knowing that the hazard exists is not sufficient. Quantitative Risk Assessment (QRA) techniques were pioneered in the USA from the 1960s after introduction of the now repealed Delaney

Risk analysis 21 amendment of 1954 which effectively prohibited the presence of any carcinogen in food.3 Many carcinogenic chemicals are not genotoxic and can exhibit threshold effects. Such substances are demonstrably carcinogenic if sufficiently large doses are given to animals to overcome any defence mechanisms. However, under Delaney these mechanistic arguments could not be taken into account. US risk analysts responded by creating a QRA approach that could be extrapolated to estimate hypothetical risks at the very low doses likely to be encountered by the public. Early QRA methods aimed to define a mathematical relationship between dose and risk of carcinogenesis. They were criticised because they seemed to be forcing biological data into mathematical models rather than developing mathematical models that fitted the data. More recent developments such as ‘physiologically-based pharmacokinetic’ (PB-PK) models aim to take physiological and biological mechanisms better into account (for further information about these developments see section 2.6.1 of this chapter). QRA modelling is still under development and is not widely used outside the USA.

2.4

Exposure analysis

Having established some information about the relationship between dose and response it is necessary to determine the levels of actual doses to the human population. Two pieces of information are vital for this: • Occurrence data: the concentrations of the chemical in the foods of concern including, if relevant, the frequency of occurrence. • Food consumption data: the amounts of the affected foods eaten including, if necessary, consumption by sub-groups.

Intake can then be calculated using a relatively simple equation: Occurrence (mg/kg)  Consumption (kg/week) Intake (mg/kg bw/week) ˆ Bodyweight (kg) Many factors can influence the accuracy of intake estimates and it is of primary importance to ensure that the assumptions made and data used are relevant to the specific risk analysis.4 The selection of inappropriate data and methods can easily lead to estimates of intake that are orders of magnitude greater or less than real levels. A particular question is the selection of statistics to represent a particular population. In the past a population average figure would have been typically used but this approach could very easily underestimate intakes of consumers at the upper end of the distribution of intake levels. Modern guidelines demand that particular sub-groups (such as children) are considered and the intakes at the upper end of the intake range (commonly 90th or 97.5th percentiles). This means that the distributions of contaminant concentrations in food and food consumption patterns must be taken into account.

22

Food chemical safety

2.4.1 Occurrence data Comprehensive data on the levels and frequency of occurrence of chemical contaminants in food are extremely expensive to obtain and are thus relatively rare. Nevertheless there are some reliable data sources and it is usually possible to acquire some data. Unfortunately, it is usually only possible to obtain data on average levels from many published sources. Even if ranges are published with the mean, interpretation is extremely difficult. In the UK the Joint Food Science and Safety Group of the Department of Health and the Ministry of Agriculture, Fisheries and Food have published the results of many analyses for chemical contaminants in food carried out under their Food Surveillance Programme. In many cases the raw data from these surveys are available for analysis. Table 2.1 lists the results of analyses for lead in some samples of cow, sheep and pig kidney obtained in Scotland and England.5 There are clear differences between species and some evidence of differences between sampling locations. What is not clear is the extent to which the variability observed is due to real and consistent differences between species and location or to normal biological variation. For the purpose of risk analysis it is necessary to determine lead concentrations that can be used to represent the distribution of observed values in the above intake calculation. If the average of all the data is used it will overlook differences between species. If the average by species is used it may overlook inter-regional differences. For example, it is possible that the high concentrations in the data set represent cattle in geographic locations in Scotland and England where consumers are consistently exposed to such levels. However, if such high values were used in the risk assessment they would not represent UK consumers as a whole. This example is used to illustrate the difficulty of completing an exposure analysis unless the context of the analysis is fully understood. In this case the differences are relatively Table 2.1

Lead in animal kidneys in England and Scotland (mg/kg)

Location

Cow

Sheep

Pig

North England North England North England North England Scotland Scotland Scotland Scotland S/SE England S/SE England S/SE England S/SE England S/SE England Mean

0.14 0.16 0.08 0.1 0.09 0.35 0.12 0.22 0.07 0.19 0.11 0.35 0.11 0.16

0.08 0.04 ND 0.04 ND 0.21 0.09 ND 0.03 0.12 0.16 0.14 0.23 0.09

ND ND ND ND ND ND ND ND ND ND ND ND ND ND

Risk analysis 23 minor. In other circumstances an inappropriate choice of data could lead to considerable errors. 2.4.2 Food consumption data There are many sources of data on food consumption although not all are necessarily appropriate for risk assessment. The most readily available data are Food Balance Sheets (FBSs) which are prepared globally every year by the UN Food and Agriculture Organisation (FAO).6 These list the domestic production, imports, exports and non-food uses for major raw food commodities for each country together with the calculated per capita annual consumption. Such data are invaluable for making comparisons between national diets since they provide a good indication of the types of food being consumed in each country. They are of limited value for risk assessment since they give no indication of the range of food consumption patterns within the country. Food consumption surveys conducted at the household level provide more information about the distribution of consumption levels. If data on the composition of the household by age and sex are available, modelling can provide some basic information about consumption of individuals. However, for reliable estimates of food consumption by individuals the weighed diary method is probably the best. In this type of survey respondents are asked to weigh and record everything that they eat for the period of the survey. Subjects are usually selected from geographical regions and at different times of year so that the survey is as representative as possible. The principal disadvantage of this type of survey is that it can only cover a few consecutive days and so food consumption over longer time-scales cannot be determined without a supplementary questionnaire about frequency of consumption. The use of weighed dietary survey data can be illustrated by extending the example of lead in kidney used above. Tables 2.2 and 2.3 give the average weekly consumption of cows’, sheep and pigs’ kidney by adults7 and preschool children8 in the UK. These figures represent those individuals who reported consumption of one of these foods during the survey and the per capita average for all individuals in the survey, whether they consumed kidney or not. Since the proportion who reported consumption is less than Table 2.2

Consumption of kidney by UK adults Food consumption, g/day Consumers only Per capita % Mean 90th Mean 90th Consuming percentile percentile

Bovine kidney 14.5 Ovine kindey 1.4 Porcine kidney 1.9 All kidney 17.4

6.05 16.14 6.15 7.01

10.37 40.45 14.17 12.61

0.88 0.22 0.12 1.22

IC IC IC IC

24

Food chemical safety

Table 2.3

Consumption of kidney by UK pre-school children Food consumption, g/day Consumers only Per capita % Mean 90th Mean 90th Consuming percentile percentile

Bovine kidney Ovine kindey Porcine kidney All kidney

3.6 0.2 0.1 4.0

3.67 7.56 0.46 3.81

6.71 15.78 0.52 8.02

0.13 0.02 0.00 0.15

IC IC IC IC

20%, the per capita consumption figures will always be at least five times smaller. Similarly, children tend to eat less of certain foods, such as kidney, than adults. Because all of the data from the original survey are available, it is possible to determine percentiles of the distribution of possible values as well as averages. The choice of figure that is taken to represent kidney consumption will depend on the question being asked. The difference between the per capita average kidney consumption for pre-school children (0.15 g/person/week) and the 90th percentile for adults who are lamb’s kidney consumers (40.5 g/person/week) could have a significant effect on the result of any intake calculation. 2.4.3 Estimating intakes To calculate estimates of intake it is necessary to multiply levels in food with food consumption. As the foregoing sections indicate, this is not a simple matter since great care must be exercised in deciding which level in food to use and which level of consumption. Combining averages provides a simple solution but gives no information about the higher levels of intake to which individuals could be exposed. Combining average contaminant concentrations with high-level consumption might be satisfactory if distribution of the foods of concern is such that over a long time period any individual’s exposure will tend to average out. However, if contamination is localised then it may be necessary to consider using high-level contaminant data with high-level consumption. This worst-case approach would overestimate intakes for the vast majority of consumers. Tables 2.4 and 2.5 provide estimates of intake based on the average of the lead concentrations in Table 2.1 and kidney consumption distributions summarised in Tables 2.2 and 2.3. Intake figures tend to follow food consumption patterns although for pork kidney the intake is zero because no lead was detected in it. As expected, when calculated on a per person per week basis children have lower intakes of lead than do adults. However, when consumption is corrected for individual bodyweight, as is necessary for comparison with a PTWI, children’s intake of all lead is about three times that of

Risk analysis 25 Table 2.4

Potential intakes of lead from kidney by UK adults

Pb, mg/kg Bovine kidney 0.16 Ovine kindey 0.09 Porcine kidney 0.00 All kidney

Pb, mg/kg Bovine kidney Ovine kindey Porcine kidney All kidney

0.16 0.09 0.00 17.4%

Lead intake, ng/person/day Consumers only Per capita Mean 90th Mean 90th percentile percentile 968 1453 0 925

1659 3640 0 1747

141 20 0 161

IC IC IC IC

Lead intake, ng/kg bodyweight/day Consumers only Per capita Mean 90th Mean 90th percentile percentile 14 21 0 13

23 51 0 23

2 0 0 2

IC IC IC IC

adults. This could be significant if children’s intakes were a cause of particular concern. However, if cumulative intake over the longer term was a concern then elevated intake during childhood might be of lesser importance. There is clearly considerable scope for producing a wide variety of intake estimates for any given scenario. It is therefore vital that the underlying Table 2.5

Potential intakes of lead from kidney by UK pre-school children

Pb, mg/kg Bovine kidney Ovine kindey Porcine kidney All kidney

0.16 0.09 0.00 4.0%

Pb, mg/kg Bovine kidney Ovine kindey Porcine kidney All kidney

0.16 0.09 0.00 4.0%

Lead intake, ng/person/day Consumers only Per capita Mean 90th Mean 90th percentile percentile 588 681 0 576

1073 1420 0 1284

21 2 0 23

IC IC IC IC

Lead intake, ng/kg bodyweight/day Consumers only Per capita Mean 90th Mean 90th percentile percentile 42 51 0 41

68 106 0 83

2 0 0 2

IC IC IC IC

26

Food chemical safety

toxicological concerns and the nature of food consumption data are fully understood if an intake estimate that is relevant to the particular situation is to be provided.

2.4.4 Probabilistic intake modelling The approaches for estimating intakes described thus far have relied on being able to extract single figure statistics from data sets to represent those data. In reality all data will show a distribution of values and this distribution tends to be overlooked using conventional methods. This may be unimportant if longterm intakes are relevant to the risk assessment since in the long term data tend to average out. However, when acute intakes are of interest the problem is more complex. It would be possible for a very high level consumer of a particular foodstuff to select, unknowingly, a food item that contains the maximum amount of contaminant. This worst-case scenario may be very unlikely and it might be imprudent to base a risk assessment on it. Yet considering averages alone would underestimate the true situation which lies somewhere between these extremes. Probabilistic (also known as Monte Carlo) modelling provides a means of making use of all of the data from the distribution of contaminant levels and food consumption to predict the probability of a given level of intake occurring. Probabilistic modelling works by taking a random sample from the distribution of contaminant levels and combining it with a random sample from the distribution of food consumption levels. It is well illustrated by the estimation of acute intakes of an organophosphate pesticide from individual apples by children. Figure 2.3 shows the distribution of an organophosphate pesticide in individual apples harvested after treatment with the agent.9 The distribution is skewed so that the mean is 0.69 mg/kg whilst the median is 0.53 mg/kg. The highest value is 3.9 mg/ kg but this represents only one out of 54 samples. It would be possible to make an exposure estimate based on either the mean, median or highest value but this would not represent the real distribution of all possible values. Figure 2.4 is the distribution of apple consumption by occasion by UK preschool children aged 1½ to 4½ years. Only raw apple consumption is considered and it is assumed that each eating occasion represents one apple or less. The data show a typically skewed distribution where the mean value is 65 g whilst the median is 59 g. Interestingly, the modal (most frequently reported) value is exactly 100 g and represents nearly 10% of the 1299 data points. This is probably due to carelessness by the reporting adults and illustrates the need to maintain caution when interpreting such data. The maximum value is 217 g but since this survey is based on a sample of only four days it is unlikely that this level of consumption would be sustained over a longer period. A better estimate of high-level intake over the longer term might be the 97.5th percentile, which is 150 g. Nevertheless, for organophosphate pesticides the concern relates to possible inhibition of the cholinesterase

Risk analysis 27

Fig. 2.3

Incidence of pesticide residues in individual apples.

enzyme and this can occur very rapidly. In this case the amount of pesticide residue ingested on a single occasion or over one day is more relevant to the risk assessment. In the Monte Carlo analysis samples are drawn at random from the residue distribution and then from the apple consumption distribution to provide the data points for the intake distribution. This sequence is repeated several thousand times until a smooth intake distribution curve is produced. The intake distribution shown in Fig. 2.5 represents 20,000 samples drawn from the pesticide residues and apple consumption distributions shown in Figs 2.3 and 2.4. The bars represent the relative frequency of each intake level and the line is the cumulative frequency distribution. The distribution is very skewed and it can be seen that the cumulative frequency is nearing 100% when only the mid-point of the distribution is being approached. This means that very high intakes are relatively rare occurrences.

Fig. 2.4

Daily apple consumption by UK pre-school children.

28

Food chemical safety

Fig. 2.5

Potential intake of an OP pesticide by UK pre-school children.

The mean intake is 0.044 mg and the 97.5th percentile is 0.171 mg. The maximum possible value (0.85 mg) does not appear in the distribution indicating its relative improbability (1 in 20,000). The US Environmental Protection Agency currently uses the 99th percentile,10 which for this distribution is 0.491 mg. It should be noted that these figures relate only to consumers of apple and assume that all apples are treated with this particular product. If the proportion of the population who do not eat raw apple and the proportion of apples that are actually treated with this particular product are taken into account, then intakes will be correspondingly lower at any given percentile. Monte Carlo analysis does not in itself provide a solution for risk managers. All the method can do is to provide the best possible representation of the real situation. The acceptability of a given proportion of consumers being exposed to a given residue level will depend on such factors as the nature of the hazard and the size of any safety margins, if present. For example, occasional minor stomach upsets would be far more acceptable than seizure or sudden death.

2.5

Risk evaluation

Risk evaluation is an apparently simple task of comparing an estimate of intake with the PTWI. If intakes are below the PTWI then there is no risk whereas if they exceed the PTWI then some risk management action may be required. For non-thresholded contaminants risk is assumed to be proportional to intake and therefore intakes should be as low as practically achievable. In practice risk evaluation is a far less certain science. A vital and often over-looked aspect of risk evaluation is ensuring that the estimate of intake corresponds to the PTWI so that like is being compared with like. For example, toxicological end-points are frequently time-related. On rare

Risk analysis 29 occasions, such as in the case of the bacterial botulinum toxins, a single dose can rapidly lead to poisoning and even death. For other end-points, such as carcinogenicity or kidney damage caused by cadmium exposure, it is the cumulative dose over long periods that is significant. However, risk evaluations cannot be conveniently divided into acute and chronic scenarios. In the case of the neurological and behavioural consequences of lead exposure, sustained exposure during a particularly vulnerable phase of life (early childhood) is regarded as being critical. For teratogenic substances the mother’s intake during a particular stage of foetal development can be critical whilst exposure at any other time is insignificant.

2.6

Methods for risk management

The outputs from risk assessment will normally include information about the relationship between dose and risk and estimates of levels of doses and thus risks in the population. For contaminants that have a toxicological threshold the Provisional Tolerable Weekly Intake (PTWI) might be defined and the number of consumers who have the potential to exceed this level of intake quantified. If a PTWI cannot be established (such as for genotoxic carcinogens) then it may be possible to quantify the proportion of a population exposed to a given level of risk by using QRA methods. If QRA methods cannot be applied then a qualitative assessment can be made such as to reduce intake levels to as low as is reasonably practicable. In either case it is the function of risk management to identify an optimal course of action to minimise the risk to consumers. 2.6.1 Contaminants with toxicological thresholds In many cases the risk assessment might indicate that the proportion of consumers with the potential to exceed the PTWI is zero and that there is no need for any action to be taken to control risks. However, even in these circumstances other risk management activities, such as risk communication, might be appropriate. If there is the potential for consumers to exceed the PTWI then it will be necessary to consider measures to control risks. The simplest approach would be to identify the level of contamination in a foodstuff that would cause a highlevel consumer to exceed the PTWI and to introduce legislation or some other form of control to establish that as the maximum permitted concentration. In practice there may be multiple routes of potential exposure and so all potentially affected foods should be taken into account. Setting maximum permitted concentrations for all affected foods would then require that the PTWI be apportioned between foods according to the potential intake from each food. This method assumes that a high-level consumer always consumes foods that contain the contaminant at the maximum permitted level. In reality this is an

30

Food chemical safety

unlikely scenario and so the method would lead to unnecessarily conservative restrictions. Alternative methods for setting standards where several possible food routes are involved include probabilistic modelling which can take into account the proportion of a particular foodstuff that might be contaminated at a particular level. In making risk management decisions it is necessary to take into account information other than technical data. For example, if several different foods are affected by contamination and action needs to be taken to reduce intakes then it may be much cheaper to reduce levels in one food (perhaps by making small modifications to processing variables) than another (where, for example, only suspending supply would suffice). In such cases it would be necessary to take into account the cost-effectiveness of different control options before identifying an optimal risk management strategy. 2.6.2 Contaminants without toxicological thresholds Contaminants such as carcinogens that are assumed to have no toxicological threshold are usually managed by setting maximum permissible concentrations at the lowest levels practically achievable. In practice this is often interpreted as being the lowest concentration that can be readily detected using current analytical methods. This is because any lower level would be unenforceable. Whilst providing a practical solution the levels selected are actually quite arbitrary, depending as they do on the best analytical performance available at any given time. As analytical methods improve so maximum permitted levels would be expected to decrease. For non-thresholded contaminants some mechanism is required that will allow the benefits in terms of reduced risks and costs associated with control to be taken into account. The costs of control will include enforcement costs as well as costs to producers in reaching ever stricter standards. Ultimately these costs will be borne by consumers in taxes, increased prices or reduced choice. Economic theory dictates that there must be a point where the extra increase in the cost of control is not justified by the corresponding increase in benefit (reduction in risk). This optimal point will differ for each contaminant according to the technology needed to control it, the nature of the hazard, and the relationship between dose and risk. It is in this latter context that quantitative risk assessment (QRA) becomes critical (see section 2.3.4 of this chapter). 2.6.3 Consumer perceptions of risk In making risk management decisions it is important to take into account nontechnical factors in addition to scientific and economic information. Recent crises in the food industry have indicated that consumers’ perceptions about risks are driven factors that would not be considered in conventional risk assessments. Research has shown that factors such as whether sub-groups

Risk analysis 31 (particularly children) might be affected, whether the hazard is familiar, if there are effects on the environment or if risks and benefits are equitably shared can determine consumers’ reactions to an issue. Risk managers must be aware that in the event of a crisis consumers’ perceptions about risks can have as great or greater impact on the outcome than the real food safety issues. 2.6.4 Decision analysis Decisions about risk need to take into account a wide range of quantitative and qualitative factors if they are to reflect both the true nature of the risk and the social context in which it is expressed. It is particularly difficult to balance scientific facts against consumer demands. A current example of this problem underlies a trade dispute between the EU and USA. Independent experts and the WHO/FAO Joint Expert Committee on Food Additives and Contaminants have consistently advised that the use of certain naturally occurring hormones as growth promoters in animal production does not present a health hazard to consumers. The US Food and Drugs Administration endorses this view and has approved the use of these substances in American agriculture. In the EU the Medicines Control Agency has imposed a moratorium on the use of these substances, largely because of consumer objections.11 The consequence of this is that imports of hormone-treated meat from the USA are prohibited in Europe. The UN World Trade Organisation (WTO) has judged in favour of the US position because the WTO is allowed only to take scientific data into account in their decisions. More recently the EU Scientific Committee on Veterinary Measures Relating to Public Health has concluded that no threshold levels can be defined for the endocrine, developmental, immunological, neurobiological, immunotoxic, genotoxic and carcinogenic effects associated with hormone residues in bovine meat and meat and that the available data do not enable a quantitative estimate of the risk. It is yet to be known whether the WTO will be persuaded by this argument. In reality, decisions about public health are often made on the basis of politics rather than science. If this were not so, more resources would be committed to controlling food poisoning relating to micro-organisms and less to relatively minor health hazards such as pesticides and environmental contaminants. Political decision-making needs to balance the needs of public health against legitimate consumer expectations. Recent trends have seen more open acknowledgement of the need to balance scientific and social factors. For example, in the UK the Food Advisory Committee which advised food ministers on food safety issues is comprised of a wide range of expertise including toxicologists, chemists, food technologists, economists, and representatives of consumer organisations and the food industry. The Committee is thus able to provide a balanced view which takes all interests into account.

32

Food chemical safety

2.6.5 Risk communication An important part of the risk management process involves informing consumers, industry and other stakeholders of the decisions made by regulatory authorities. However, this is a narrow view of risk communication which does not take into account the potential for dialogue between interested parties that can result in better decision-making. Understanding how consumers view the risks associated with chemical contaminants in food can help to avoid either under- or over-regulation. There are more ways of controlling exposures to chemical contaminants in food than by simply imposing maximum permitted concentrations. Providing advice and information about routes of contamination may allow food producers to take simple precautions to control contamination. Whilst each producer might introduce similar methods they will not all necessarily be able to reduce contaminant levels to the same level. However, the overall effect on the food supply might be sufficient to lower consumers’ intakes to acceptable levels. This principle is enshrined in the Hazard Analysis Critical Control Points (HACCP) approach which is used to control microbiological hazards. Here the emphasis is on exchanging information about good practice rather than imposing absolute limits. Risk communication can of itself be a useful risk management tool. Regulatory authorities have for many years gathered information about the levels of chemical contamination in foods. However, such technical information is difficult to disseminate and many consumers and food producers are probably unaware that it takes place. Some authorities have taken the decision that the brand names associated with foods should be released along with the data. This ‘name and shame’ approach will allow consumers to take avoiding action if they wish. The consequence will probably be that those food producers and retailers with a good brand image to protect will make more stringent efforts to identify and control chemical contaminants in their products.

2.7

Future trends in risk analysis

Risk analysis, like most scientific disciplines, is subject to continual evolution. This means that methods and concepts that are in common use today may well become discredited and obsolete in the future. This presents a problem for risk managers because it creates the impression that everything done before was somehow ‘wrong’. In fact most changes are gradual and tend to take effect at the margins rather than overturning all previous assessments. An example of this is the introduction of acute risk assessments for pesticide residues. It must not be assumed that all pesticides now present acute risks for consumers. In fact only a minority even hold the potential for acute effects and of these probably only a small proportion will require any action to prevent the possibility of adverse events. Similarly the development of PB-PK methods in dose-response modelling and aggregate exposure estimates will probably affect only a small proportion of the chemical contaminants previously assessed.

Risk analysis 33 2.7.1 Physiologically-based pharmacokinetic (PB-PK) modelling In section 2.3 of this chapter the present approach to characterisation of doseresponse relationships was described. In most cases it is necessary to extrapolate from animal species that are used in testing to humans. It may also be necessary to extrapolate from experimental conditions to real human exposures. At the present time default assumptions (which are assumed to be conservative) are applied to convert experimental data into predictive human risk assessments. However, the rates at which a particular substance is adsorbed, distributed, metabolised and excreted can vary considerably between animal species and this can introduce considerable uncertainties into the risk assessment process. The aim of PB-PK models is to quantify these differences as far as possible and so to be able to make more reliable extrapolations. The PB-PK model is based on a mathematical representation of the physiological and biological structure of the species being described. Certain physiological parameters such as blood flow into an organ are specific for each species and so a multi-compartmental model can be assembled which can be used to predict the behaviour of a chemical over a wide range of conditions. Data from animal studies can be used to investigate the absorption, distribution, metabolism and elimination of a particular substance in a given species. Once the behaviour of the chemical in the animal model is fully understood the physiological parameters can be altered so that they represent a human. This will allow the behaviour of the chemical in the human system to be predicted. The use of biological markers such as levels of the chemical in blood and urine can be used to validate the model, particularly if human volunteer studies are available. PB-PK modelling allows further refinement of the dose-response evaluation by partitioning the relationship into pharmacokinetic (exposure vs. tissues dose) and pharmacodynamic (tissue dose vs. toxic response) components. This allows the uncertainties associated with each component to be assessed separately and adds accuracy to the overall animal to man extrapolation. Future developments of PB-PK modelling may allow specific sub-populations such as the newborn or individuals with metabolic variations to be taken into account. However, before this can be done there will need to be considerable growth in the amounts of physiological, pharmacokinetic and pharmacodynamic information available. 2.7.2 Aggregate exposure assessment Until very recently the risks associated with different types of chemicals such as food additives, pesticides, environmental contaminants and natural constituents of food were assessed and managed separately. However, a particular substance might fall into two or more of these categories and so the opportunity for simultaneous exposure might be overlooked. Furthermore, exposure to a chemical could occur through diet, drinking water, air pollution or dermal absorption. Aggregate exposure assessment aims to take all of the possible sources and routes of exposure into account in a realistic manner and thereby obtain a better overall estimate of risk. Initiatives have been set up in both the

34

Food chemical safety

consumer affairs directorate of the European Commission and the Environmental Protection Agency in the USA. Aggregate exposure assessment is naturally more complex than the methods used for dietary risk assessment. In the simplest analysis a worst case can be established for each source and exposure route and then summed to give a total exposure. If this were below any threshold of concern such as the PTWI then no further action would be required. However, if the total worst case exposure was above a PTWI then it is unlikely to reflect the real situation since the probability that any individual would be exposed to each source by each route at the maximum level is very remote. When the EPA considered exposures to insecticide residues in the home they identified at least six possible sources and routes; these are given in Table 2.6. Their original approach apportioned the acceptable daily intake (ADI) between the various routes but it soon became clear that this was unrealistic because an individual was unlikely to be exposed via all routes on any one day. The EPA’s present strategy is to develop an approach called micro-exposure event modelling. Micro-exposure event modelling is based on statistical data on the frequencies and levels of contamination of food, water, etc. and on behavioural information about the frequency of use of lawn/pet/timber treatments, etc. The combined data are assembled in a probabilistic model called ‘LIFELINE’ which is able to predict the frequency and level of exposure to a group of hypothetical individuals over their lifetime.12 The model is also able to take account of the relative proportions of different types of accommodation, the incidence of pet ownership or any other data that will affect real levels of exposure. The output from the LIFELINE model allows the exposures of individuals in a population to be modelled over any interval from a single occasion to a lifetime. Micro-exposure event modelling combined with probabilistic modelling provides a great opportunity to assess real aggregate exposures in the real world. However, the method is highly dependent on the availability of complete and accurate information about levels of contamination and human behaviour. Whilst some of this is available, particularly in the context of pesticides, for other chemicals it may be a long time before this approach can be employed.

Table 2.6 Source

Route

Residues in food Residues in drinking water Residues in water Pet treatments Lawn treatments Timber treatments

oral oral inhalation/dermal (volatilisation in showers, etc.) oral/inhalation/dermal oral/inhalation/dermal oral/inhalation/dermal

Risk analysis 35

2.8

Sources of further information and advice

World-Wide-Web resources The EU Commission Directorate General on Health and Consumer Protection World Health Organisation UK Pesticides Safety Directorate UK Food Standards Agency Joint FAO/WHO Expert Committee on Food Additives and Contaminants Institute of Food Science and Technology (IFST) International Society of Exposure Analysis (ISEA)

http://europa.eu.int/_omm./dgs/ health_consumer/index_en.htm http://www.who.int/ http://www.pesticides.gov.uk/ http://www.foodstandards.gov.uk/ http://www.inchem.org/ aboutjecfa.html http://www.ifst.org/ http://www.ISEAweb.org

Publications The Economics of Food Safety. Elsevier Applied Science Publishers Ltd (1991). ISBN 0-444-01614-7. TENNANT, D. R. Food Chemical Risk Analysis. Blackie Academic and Professional, Chapman and Hall, London (1997). ISBN 0-412-723107. CASWELL, J. A.

2.9 1. 2. 3.

4.

5. 6.

References and STEADMAN, J. H. (1990). Criteria for setting quantitative estimates of acceptable intakes of chemicals in food in the UK. Food Additives and Contaminants 7/3, 287–302. MCDONALD, A. L., FIELDER, R. J., DIGGLE, G. E., TENNANT, D. R. and FISCHER, C. E. (1996). Carcinogens in food: priorities for regulatory action. Human and Experimental Toxicology 15 739–46. LOVELL, D. P. and THOMAS, G. (1997). Quantitative risk assessment. pp. 57– 86 in TENNANT, D. R. (ed.) Food Chemical Risk Analysis. Blackie Academic and Professional, Chapman and Hall, London. ISBN 0-412723107. DOUGLASS, J. S. and TENNANT, D. R. (1997). Estimations of dietary intake of food chemicals. pp. 195–215 in TENNANT, D. R. (ed.) Food Chemical Risk Analysis. Blackie Academic and Professional, Chapman and Hall, London. ISBN 0-412-723107. MAFF UK Food Surveillance Information Sheets. Number 160, September 1998. Lead, cadmium copper and zinc in offals. Can be down-loaded from the FAO website at: http://apps.fao.org/lim500/ nph-wrap.pl?FoodBalanceSheet&Domain=FoodBalanceSheet RUBERY, E. D., BARLOW, S. M.

36 7. 8. 9. 10. 11.

12.

Food chemical safety and WISEMAN, M. (1990). The Dietary and Nutritional Survey of British Adults. HMSO, London. GREGORY, J. R., COLLINS, D. L., DAVIES, P. S. W., HUGHES, J. M. and CLARKE P. C. (1995). National Diet and Nutrition Survey: Children aged 1½ to 4½ years. HMSO, London. ISBN 0-11-691611-7. PESTICIDES SAFETY DIRECTORATE (1998). The Occurrence of Unit to Unit Variability of Pesticide Residues in Fruit and Vegetables. PSD, UK. OFFICE OF PESTICIDE PROGRAMS (1998). Guidance for Submission of Probabilistic Exposure Assessments to the Office of Pesticides Programs’ Health Effects Division. US Environmental Protection Agency. Opinion of the Scientific Committee on Veterinary Measures Relating to Public Health: Assessment of potential risks to human health from hormone residues in bovine meat and meat products (30 April 1999). http://europa.eu.int/comm/dg24/health/sc/scv/out21_en.html For more information contact the lifeline web page at Hampshire Research: http://www.hampshire.org/welcome.html GREGORY, J., TYLER, H.

3 Analytical methods: quality control and selection R. Wood, Food Standards Agency, London

3.1

Introduction

It is now universally recognised as being essential that a laboratory produces and reports data that are fit-for-purpose. For a laboratory to produce consistently reliable data it must implement an appropriate programme of quality assurance measures; such measures are now required by virtue of legislation for food control work or, in the case of the UK Food Standards Agency (FSA), in their requirements for contractors undertaking surveillance work. Thus customers now demand of providers of analytical data that their data meet established quality requirements. These are further described below. The significance of the measures identified are then described and some indications are given as to the future of analytical methodology within the food laboratory. These are then discussed. Methods of analysis have been prescribed by legislation for a number of foodstuffs since the UK acceded to the European Community in 1972. However, the Community now recognises that the quality of results from a laboratory is equally as important as the method used to obtain the results. This is best illustrated by consideration of the Council Directive on the Official Control of Foodstuffs (OCF) which was adopted by the Community in June, 1989.1 This, and the similar Codex Alimentarius Commission requirements, are described below. As a result of this general recognition there is a general move away from the need to prescribe all analytical methodology in detail towards the prescription of the general quality systems within which the laboratory must operate. This allows greater flexibility to the laboratory without detracting from the quality of results that it will produce.

38

Food chemical safety

3.2

Legislative requirements

3.2.1 The European Union For analytical laboratories in the food sector there are legislative requirements regarding analytical data which have been adopted by the European Union. In particular, methods of analysis have been prescribed by legislation for a number of foodstuffs since the UK acceded to the European Community in 1972. However, the Union now recognises that the competency of a laboratory (i.e. how well it can use a method) is equally as important as the ‘quality’ of the method used to obtain results. The Council Directive on the Official Control of Foodstuffs which was adopted by the Community in 19891 looks forward to the establishment of laboratory quality standards, by stating that ‘In order to ensure that the application of this Directive is uniform throughout the Member States, the Commission shall, within one year of its adoption, make a report to the European Parliament and to the Council on the possibility of establishing Community quality standards for all laboratories involved in inspection and sampling under this Directive’ (Article 13). Following that, in September 1990 the Commission produced a Report which recommended establishing Community quality standards for all laboratories involved in inspections and sampling under the OCF Directive. Proposals on this have now been adopted by the Community in the Directive on Additional Measures Concerning the Food Control of Foodstuffs (AMFC).2 In Article 3 of the AMFC Directive it states: 1.

2.

3.

Member States shall take all measures necessary to ensure that the laboratories referred to in Article 7 of Directive 89/397/EEC1 comply with the general criteria for the operation of testing laboratories laid down in European standard EN 450013 supplemented by Standard Operating Procedures and the random audit of their compliance by quality assurance personnel, in accordance with the OECD principles Nos. 2 and 7 of good laboratory practice as set out in Section II of Annex 2 of the Decision of the Council of the OECD of 12 Mar 1981 concerning the mutual acceptance of data in the assessment of chemicals.4 In assessing the laboratories referred to in Article 7 of Directive 89/ 397/EEC Member States shall: (a) apply the criteria laid down in European standard EN 45002;5 and (b) require the use of proficiency testing schemes as far as appropriate. Laboratories meeting the assessment criteria shall be presumed to fulfil the criteria referred to in paragraph 1. Laboratories which do not meet the assessment criteria shall not be considered as laboratories referred to in Article 7 of the said Directive. Member States shall designate bodies responsible for the assessment of laboratories as referred to in Article 7 of Directive 89/397/EEC. These bodies shall comply with the general criteria for laboratory accreditation bodies laid down in European Standard EN 45003.6

Analytical methods: quality control and selection 4.

39

The accreditation and assessment of testing laboratories referred to in this article may relate to individual tests or groups of tests. Any appropriate deviation in the way in which the standards referred to in paragraphs 1, 2 and 3 are applied shall be adopted in accordance with the procedure laid down in Article 8.

In Article 4, it states: Member States shall ensure that the validation of methods of analysis used within the context of official control of foodstuffs by the laboratories referred to in Article 7 of Directive 89/397/EEC comply whenever possible with the provisions of paragraphs 1 and 2 of the Annex to Council Directive 85/591/EEC of 23 December 1985 concerning the introduction of Community methods of sampling and analysis for the monitoring of foodstuffs intended for human consumption.7 As a result of the adoption of the above directives legislation is now in place to ensure that there is confidence not only in national laboratories but also those of the other Member States. As one of the objectives of the EU is to promote the concept of mutual recognition, this is being achieved in the laboratory area by the adoption of the AMFC directive. The effect of the AMFC Directive is that organisations must consider the following aspects within the laboratory: its organisation, how well it actually carries out analyses, and the methods of analysis used in the laboratory. All these aspects are inter-related, but in simple terms may be thought of as: • becoming accredited to an internationally recognised standard; such accreditation is aided by the use of internal quality control procedures • participating in proficiency schemes, and • using validated methods.

In addition it is important that there is dialogue and co-operation by the laboratory with its customers. This is also required by virtue of the EN 45001 Standard at paragraph 6, and will be emphasised even more in future with the adoption of ISO/IEC Guide 17025.8 The AMFC Directive requires that food control laboratories should be accredited to the EN 45000 series of standards as supplemented by some of the OECD GLP principles. In the UK, government departments have nominated the United Kingdom Accreditation Service (UKAS) to carry out the accreditation of official food control laboratories for all the aspects prescribed in the Directive. However, as the accreditation agency will also be required to comply with the EN 45003 Standard and to carry out assessments in accordance with the EN 45002 Standard, all accreditation agencies that are members of the European Cooperation for Accreditation of Laboratories (EA) may be asked to carry out the accreditation of a food control laboratory within the UK. Similar procedures will be followed in the other Member States, all having or developing equivalent organisations to UKAS. Details of the UK requirements for food control laboratories are described later in this chapter.

40

Food chemical safety

3.2.2

Codex Alimentarius Commission: guidelines for the assessment of the competence of testing laboratories involved in the import and export control of food The decisions of the Codex Alimentarius Commission are becoming of increasing importance because of the acceptance of Codex Standards in the World Trade Organisation agreements. They may be regarded as being semilegal in status. Thus, on a world-wide level, the establishment of the World Trade Organisation (WTO) and the formal acceptance of the Agreements on the Application of Sanitary and Phytosanitary Measures (SPS Agreement) and Technical Barriers to Trade (TBT Agreement) have dramatically increased the status of Codex as a body. As a result, Codex Standards are now seen as de facto international standards and are increasingly being adopted by reference into the food law of both developed and developing countries. Because of the status of the CAC described above, the work that it has carried out in the area of laboratory quality assurance must be carefully considered. One of the CAC Committees, the Codex Committee on Methods of Analysis and Sampling (CCMAS), has developed criteria for assessing the competence of testing laboratories involved in the official import and export control of foods. These were recommended by the Committee at its 21st Session in March 19979 and adopted by the Codex Alimentarius Commission at its 22nd Session in June 1997;10 they mirror the EU recommendations for laboratory quality standards and methods of analysis. The guidelines provide a framework for the implementation of quality assurance measures to ensure the competence of testing laboratories involved in the import and export control of foods. They are intended to assist countries in their fair trade in foodstuffs and to protect consumers. The criteria for laboratories involved in the import and export control of foods, now adopted by the Codex Alimentarius Commission, are: • to comply with the general criteria for testing laboratories laid down in ISO/ IEC Guide 25: 1990 ‘General requirements for the competence of calibration and testing laboratories’8 (i.e. effectively accreditation) • to participate in appropriate proficiency testing schemes for food analysis which conform to the requirements laid down in ‘The International Harmonised Protocol for the Proficiency Testing of (Chemical) Analytical Laboratories’11 (already adopted for Codex purposes by the CAC at its 21st Session in July 1995) • to use, whenever available, methods of analysis that have been validated according to the principles laid down by the CAC • to use internal quality control procedures, such as those described in the ‘Harmonised Guidelines for Internal Quality Control in Analytical Chemistry Laboratories’.12

In addition, the bodies assessing the laboratories should comply with the general criteria for laboratory accreditation, such as those laid down in the ISO/IEC Guide 58:1993: ‘Calibration and testing laboratory accreditation systems – General requirements for operation and recognition’.13

Analytical methods: quality control and selection

41

Thus, as for the European Union, the requirements are based on accreditation, proficiency testing, the use of validated methods of analysis and, in addition, the formal requirement to use internal quality control procedures which comply with the Harmonised Guidelines. Although the EU and Codex Alimentarius Commission refer to different sets of accreditation standards, the ISO/IEC Guide 25: 1990 and EN 45000 series of standards are similar in intent. It is only through these measures that international trade will be facilitated and the requirements to allow mutual recognition to be fulfilled will be achieved.

3.3

FSA surveillance requirements

The Food Standards Agency undertakes food chemical surveillance exercises. It has developed information for potential contractors on the analytical quality assurance requirements for food chemical surveillance exercises. These requirements are describe below but reproduced as an appendix to this chapter; they emphasise the need for a laboratory to produce and report data of appropriate quality. The requirements are divided into three parts dealing with: • Part A: quality assurance requirements for surveillance projects provided by potential contractors at the time that tender documents are completed and when commissioning a survey. Here information is sought on: – the formal quality system in the laboratory if third-party assessed (i.e. if UKAS accredited or GLP compliant) – the quality system if not accredited – proficiency testing – internal quality control – method validation. • Part B: information to be defined by the FSA customer once the contract has been awarded to a contractor, e.g. the sample storage conditions to be used, the methods to be used, the IQC procedures to be used, the required measurement limits (e.g. limit of detection (LOD), limit of determination/ quantification (LOQ), and the reporting limits) • Part C: information to be provided by the contractor on an on-going basis once contract is awarded – to be agreed with the customer to ensure that the contractor remains in ‘analytical control’.

3.4

Laboratory accreditation and quality control

Although the legislative requirements apply only to food-control laboratories, the effect of their adoption is that other food laboratories will be advised to achieve the same standard in order for their results to be recognised as equivalent and accepted for ‘due diligence’ purposes. In addition, the Codex

42

Food chemical safety

requirements affect all organisations involved in international trade and thus provide an important ‘quality umbrella’. As shown above, these include a laboratory to be third-party assessed to international accreditation standards, to demonstrate that it is in statistical control by using appropriate internal quality control procedures, to participate in proficiency testing schemes which provide an objective means of assessing and documenting the reliability of the data it is producing and to use methods of analysis that are ‘fit-for-purpose’. These requirements are summarised below and then described in greater detail later in this chapter. 3.4.1 Accreditation The AMFC Directive requires that food-control laboratories should be accredited to the EN 45000 series of standards as supplemented by some of the OECD GLP principles. In the UK, government departments will nominate the United Kingdom Accreditation Service (UKAS) to carry out the accreditation of official food-control laboratories for all the aspects prescribed in the Directive. However, as the accreditation agency will also be required to comply to EN 45003 Standard and to carry out assessments in accordance with the EN 45002 Standard, any other accreditation agencies that are members of the European Co-operation for Accreditation of Laboratories (EA) may also be nominated to carry out the accreditation. Similar procedures will be followed in the other Member States, all having or developing equivalent organisations to UKAS. It has been the normal practice for UKAS to accredit the scope of laboratories on a method-by-method basis. In the case of official food-control laboratories undertaking non-routine or investigative chemical analysis it is accepted that it is not practical to use an accredited fully documented method in the conventional sense, which specifies each sample type and analyte. In these cases a laboratory must have a protocol defining the approach to be adopted which includes the requirements for validation and quality control. Full details of procedures used, including instrumental parameters, must be recorded at the time of each analysis in order to enable the procedure to be repeated in the same manner at a later date. It is therefore recommended that for official food-control laboratories undertaking analysis, appropriate methods are accredited on a generic basis with such generic accreditation being underpinned where necessary by specific method accreditation. Food-control laboratories seeking to be accredited for the purposes of the Directive should include, as a minimum, the following techniques in generic protocols: HPLC, GC, atomic absorption and/or ICP (and microscopy). A further protocol on sample preparation procedures (including digestion and solvent dissolution procedures) should also be developed. Other protocols for generic methods which are acceptable to UKAS may also be developed. Proximate analyses should be addressed as a series of specific methods including moisture, fat, protein and ash determinations.

Analytical methods: quality control and selection

43

Where specific Regulations are in force then the methods associated with the Regulations shall be accredited if the control laboratory wishes to offer enforcement of the Regulations to customers. Examples of these are methods of analysis for aflatoxins and methods of analysis for specific and overall migration for food contact materials. By using the combination of specific method accreditation and generic accreditation it will be possible for laboratories to be accredited for all the analyses of which they are capable and competent to undertake. Method performance validation data demonstrating that the method was fit-for-purpose shall be demonstrated before the test result is released and method performance shall be monitored by on-going quality-control techniques where applicable. It will be necessary for laboratories to be able to demonstrate quality-control procedures to ensure compliance with the EN 45001 Standard,3 an example of which would be compliance with the ISO/AOAC/IUPAC Guidelines on Internal Quality Control in Analytical Chemistry Laboratories.12 3.4.2 Internal quality control (IQC) IQC is one of a number of concerted measures that analytical chemists can take to ensure that the data produced in the laboratory are of known quality and uncertainty. In practice this is determined by comparing the results achieved in the laboratory at a given time with a standard. IQC therefore comprises the routine practical procedures that enable the analyst to accept a result or group of results or reject the results and repeat the analysis. IQC is undertaken by the inclusion of particular reference materials, ‘control materials’, into the analytical sequence and by duplicate analysis. ISO, IUPAC and AOAC INTERNATIONAL have co-operated to produce agreed protocols on the ‘Design, Conduct and Interpretation of Collaborative Studies’14 and on the ‘Proficiency Testing of [Chemical] Analytical Laboratories’.11 The Working Group that produced these protocols has prepared a further protocol on the internal quality control of data produced in analytical laboratories. The document was finalised in 1994 and published in 1995 as the ‘Harmonised Guidelines For Internal Quality Control In Analytical Chemistry Laboratories’.12 The use of the procedures outlined in the Protocol should aid compliance with the accreditation requirements specified above. Basic concepts The protocol sets out guidelines for the implementation of internal quality control (IQC) in analytical laboratories. IQC is one of a number of concerted measures that analytical chemists can take to ensure that the data produced in the laboratory are fit for their intended purpose. In practice, fitness for purpose is determined by a comparison of the accuracy achieved in a laboratory at a given time with a required level of accuracy. Internal quality control therefore comprises the routine practical procedures that enable the analytical chemist to accept a result or group of results as fit-for-purpose, or

44

Food chemical safety

reject the results and repeat the analysis. As such, IQC is an important determinant of the quality of analytical data, and is recognised as such by accreditation agencies. Internal quality control is undertaken by the inclusion of particular reference materials, called ‘control materials’, into the analytical sequence and by duplicate analysis. The control materials should, wherever possible, be representative of the test materials under consideration in respect of matrix composition, the state of physical preparation and the concentration range of the analyte. As the control materials are treated in exactly the same way as the test materials, they are regarded as surrogates that can be used to characterise the performance of the analytical system, both at a specific time and over longer intervals. Internal quality control is a final check of the correct execution of all of the procedures (including calibration) that are prescribed in the analytical protocol and all of the other quality assurance measures that underlie good analytical practice. IQC is therefore necessarily retrospective. It is also required to be as far as possible independent of the analytical protocol, especially the calibration, that it is designed to test. Ideally both the control materials and those used to create the calibration should be traceable to appropriate certified reference materials or a recognised empirical reference method. When this is not possible, control materials should be traceable at least to a material of guaranteed purity or other well characterised material. However, the two paths of traceability must not become coincident at too late a stage in the analytical process. For instance, if control materials and calibration standards were prepared from a single stock solution of analyte, IQC would not detect any inaccuracy stemming from the incorrect preparation of the stock solution. In a typical analytical situation several, or perhaps many, similar test materials will be analysed together, and control materials will be included in the group. Often determinations will be duplicated by the analysis of separate test portions of the same material. Such a group of materials is referred to as an analytical ‘run’. (The words ‘set’, ‘series’ and ‘batch’ have also been used as synonyms for ‘run’.) Runs are regarded as being analysed under effectively constant conditions. The batches of reagents, the instrument settings, the analyst, and the laboratory environment will, under ideal conditions, remain unchanged during analysis of a run. Systematic errors should therefore remain constant during a run, as should the values of the parameters that describe random errors. As the monitoring of these errors is of concern, the run is the basic operational unit of IQC. A run is therefore regarded as being carried out under repeatability conditions, i.e. the random measurement errors are of a magnitude that would be encountered in a ‘short’ period of time. In practice the analysis of a run may occupy sufficient time for small systematic changes to occur. For example, reagents may degrade, instruments may drift, minor adjustments to instrumental settings may be called for, or the laboratory temperature may rise. However, these systematic effects are, for the purposes of IQC, subsumed into the

Analytical methods: quality control and selection

45

repeatability variations. Sorting the materials making up a run into a randomised order converts the effects of drift into random errors. Scope of the guidelines The guidelines are a harmonisation of IQC procedures that have evolved in various fields of analysis, notably clinical biochemistry, geochemistry, environmental studies, occupational hygiene and food analysis. There is much common ground in the procedures from these various fields. However, analytical chemistry comprises an even wider range of activities, and the basic principles of IQC should be able to encompass all of these. The guidelines will be applicable in the great majority of instances although there are a number of IQC practices that are restricted to individual sectors of the analytical community and so not included in the guidelines. In order to achieve harmonisation and provide basic guidance on IQC, some types of analytical activity have been excluded from the guidelines. Issues specifically excluded are as follows: • Quality control of sampling. While it is recognised that the quality of the analytical result can be no better than that of the sample, quality control of sampling is a separate subject and in many areas not yet fully developed. Moreover, in many instances analytical laboratories have no control over sampling practice and quality. • In-line analysis and continuous monitoring. In this style of analysis there is no possibility of repeating the measurement, so the concept of IQC as used in the guidelines is inapplicable. • Multivariate IQC. Multivariate methods in IQC are still the subject of research and cannot be regarded as sufficiently established for inclusion in the guidelines. The current document regards multianalyte data as requiring a series of univariate IQC tests. Caution is necessary in the interpretation of this type of data to avoid inappropriately frequent rejection of data. • Statutory and contractual requirements. • Quality assurance measures such as pre-analytical checks on instrumental stability, wavelength calibration, balance calibration, tests on resolution of chromatography columns, and problem diagnostics are not included. For present purposes they are regarded as part of the analytical protocol, and IQC tests their effectiveness together with the other aspects of the methodology.

Recommendations The following recommendations represent integrated approaches to IQC that are suitable for many types of analysis and applications areas. Managers of laboratory quality systems will have to adapt the recommendations to the demands of their own particular requirements. Such adoption could be implemented, for example, by adjusting the number of duplicates and control material inserted into a run, or by the inclusion of any additional measures favoured in the particular application area. The procedure finally chosen and its accompanying decision rules must be

46

Food chemical safety

codified in an IQC protocol that is separate from the analytical system protocol. The practical approach to quality control is determined by the frequency with which the measurement is carried out and the size and nature of each run. The following recommendations are therefore made. (The use of control charts and decision rules are covered in Appendix 1 to the guidelines.) In all of the following the order in the run in which the various materials are analysed should be randomised if possible. A failure to randomise may result in an underestimation of various components of error. Short (e.g. n < 20) frequent runs of similar materials Here the concentration range of the analyte in the run is relatively small, so a common value of standard deviation can be assumed. Insert a control material at least once per run. Plot either the individual values obtained, or the mean value, on an appropriate control chart. Analyse in duplicate at least half of the test materials, selected at random. Insert at least one blank determination Longer (e.g. n > 20) frequent runs of similar materials Again a common level of standard deviation is assumed. Insert the control material at an approximate frequency of one per ten test materials. If the run size is likely to vary from run to run it is easier to standardise on a fixed number of insertions per run and plot the mean value on a control chart of means. Otherwise plot individual values. Analyse in duplicate a minimum of five test materials selected at random. Insert one blank determination per ten test materials. Frequent runs containing similar materials but with a wide range of analyte concentration Here we cannot assume that a single value of standard deviation is applicable. Insert control materials in total numbers approximately as recommended above. However, there should be at least two levels of analyte represented, one close to the median level of typical test materials, and the other approximately at the upper or lower decile as appropriate. Enter values for the two control materials on separate control charts. Duplicate a minimum of five test materials, and insert one procedural blank per ten test materials. Ad hoc analysis Here the concept of statistical control is not applicable. It is assumed, however, that the materials in the run are of a single type. Carry out duplicate analysis on all of the test materials. Carry out spiking or recovery tests or use a formulated control material, with an appropriate number of insertions (see above), and with different concentrations of analyte if appropriate. Carry out blank determinations. As no control limits are available, compare the bias and precision with fitness-for-purpose limits or other established criteria. By following the above recommendations laboratories would introduce internal quality control measures which are an essential aspect of ensuring that data

Analytical methods: quality control and selection

47

released from a laboratory are fit-for-purpose. If properly executed, quality control methods can monitor the various aspects of data quality on a run-by-run basis. In runs where performance falls outside acceptable limits, the data produced can be rejected and, after remedial action on the analytical system, the analysis can be repeated. The guidelines stress, however, that internal quality control is not foolproof even when properly executed. Obviously it is subject to ‘errors of both kinds’, i.e. runs that are in control will occasionally be rejected and runs that are out of control occasionally accepted. Of more importance, IQC cannot usually identify sporadic gross errors or short-term disturbances in the analytical system that affect the results for individual test materials. Moreover, inferences based on IQC results are applicable only to test materials that fall within the scope of the analytical method validation. Despite these limitations, which professional experience and diligence can alleviate to a degree, internal quality control is the principal recourse available for ensuring that only data of appropriate quality are released from a laboratory. When properly executed it is very successful. The guidelines also stress that the perfunctory execution of any quality system will not guarantee the production of data of adequate quality. The correct procedures for feedback, remedial action and staff motivation must also be documented and acted upon. In other words, there must be a genuine commitment to quality within a laboratory for an internal quality control programme to succeed, i.e. the IQC must be part of a complete quality management system.

3.5

Proficiency testing

Participation in proficiency testing schemes provides laboratories with an objective means of assessing and documenting the reliability of the data they are producing. Although there are several types of proficiency testing schemes they all share a common feature: test results obtained by one laboratory are compared with those obtained by one or more testing laboratories. The proficiency testing schemes must provide a transparent interpretation and assessment of results. Laboratories wishing to demonstrate their proficiency should seek and participate in proficiency testing schemes relevant to their area of work. The need for laboratories carrying out analytical determinations to demonstrate that they are doing so competently has become paramount. It may well be necessary for such laboratories not only to become accredited and to use fully validated methods but also to participate successfully in proficiency testing schemes. Thus, proficiency testing has assumed a far greater importance than previously. 3.5.1 What is proficiency testing? A proficiency testing scheme is defined as a system for objectively checking laboratory results by an external agency. It includes comparison of a

48

Food chemical safety

laboratory’s results at intervals with those of other laboratories, the main object being the establishment of trueness. In addition, although various protocols for proficiency testing schemes have been produced the need now is for a harmonised protocol that will be universally accepted; the progress towards the preparation and adoption of an internationally recognised protocol is described below. Various terms have been used to describe schemes conforming to the protocol (e.g. external quality assessment, performance schemes, etc.), but the preferred term is ‘proficiency testing’. Proficiency testing schemes are based on the regular circulation of homogeneous samples by a co-ordinator, analysis of samples (normally by the laboratory’s method of choice) and an assessment of the results. However, although many organisations carry out such schemes, there has been no international agreement on how this should be done – in contrast to the collaborative trial situation. In order to rectify this, the same international group that drew up collaborative trial protocols was invited to prepare one for proficiency schemes (the first meeting to do so was held in April 1989). Other organisations, such as CEN, are also expected to address the problem. 3.5.2 Why proficiency testing is important Participation in proficiency testing schemes provides laboratories with a means of objectively assessing, and demonstrating, the reliability of the data they produce. Although there are several types of schemes, they all share a common feature of comparing test results obtained by one testing laboratory with those obtained by other testing laboratories. Schemes may be ‘open’ to any laboratory or participation may be invited. Schemes may set out to assess the competence of laboratories undertaking a very specific analysis (e.g. lead in blood) or more general analysis (e.g. food analysis). Although accreditation and proficiency testing are separate exercises, it is anticipated that accreditation assessments will increasingly use proficiency testing data. 3.5.3 Accreditation agencies It is now recommended by ISO Guide 25,8 the prime standard to which accreditation agencies operate, that such agencies require laboratories seeking accreditation to participate in an appropriate proficiency testing scheme before accreditation is gained. There is now an internationally recognised protocol to which proficiency testing schemes should comply; this is the IUPAC/AOAC/ ISO Harmonised Protocol described below. 3.5.4

ISO/IUPAC/AOAC International Harmonised Protocol For Proficiency Testing of (Chemical) Analytical Laboratories The International Standardising Organisations, AOAC, ISO and IUPAC, have co-operated to produce an agreed ‘International Harmonised Protocol For

Analytical methods: quality control and selection

49

Proficiency Testing of (Chemical) Analytical Laboratories’.11 That protocol is recognised within the food sector of the European Community and also by the Codex Alimentarius Commission. The protocol makes the following recommendations about the organisation of proficiency testing, all of which are important in the food sector. Framework Samples must be distributed regularly to participants who are to return results within a given time. The results will be statistically analysed by the organiser and participants will be notified of their performance. Advice will be available to poor performers and participants will be kept fully informed of the scheme’s progress. Participants will be identified by code only, to preserve confidentiality. The scheme’s structure for any one analyte or round in a series should be: • • • • • • • •

samples prepared samples distributed regularly participants analyse samples and report results results analysed and performance assessed participants notified of their performance advice available for poor performers, on request co-ordinator reviews performance of scheme next round commences.

Organisation The running of the scheme will be the responsibility of a co-ordinating laboratory/organisation. Sample preparation will either be contracted out or undertaken in house. The co-ordinating laboratory must be of high reputation in the type of analysis being tested. Overall management of the scheme should be in the hands of a small steering committee (Advisory Panel) having representatives from the co-ordinating laboratory (who should be practising laboratory scientists), contract laboratories (if any), appropriate professional bodies and ordinary participants. Samples The samples to be distributed must be generally similar in matrix to the unknown samples that are routinely analysed (in respect of matrix composition and analyte concentration range). It is essential they are of acceptable homogeneity and stability. The bulk material prepared must be effectively homogeneous so that all laboratories will receive samples that do not differ significantly in analyte concentration. The co-ordinating laboratory should also show the bulk sample is sufficiently stable to ensure it will not undergo significant change throughout the duration of the proficiency test. Thus, prior to sample distribution, matrix and analyte stability must be determined by analysis after appropriate storage. Ideally, the quality checks on samples referred should be performed by a different laboratory from that which prepared the sample,

50

Food chemical safety

although it is recognised that this would probably cause considerable difficulty to the co-ordinating laboratory. The number of samples to be distributed per round for each analyte should be no more than five. Frequency of sample distribution Sample distribution frequency in any one series should not be more than every two weeks and not less than every four months. A frequency greater than once every two weeks could lead to problems in turn-round of samples and results. If the period between distributions extends much beyond four months, there will be unacceptable delays in identifying analytical problems and the impact of the scheme on participants will be small. The frequency also relates to the field of application and amount of internal quality control that is required for that field. Thus, although the frequency range stated above should be adhered to, there may be circumstances where it is acceptable for a longer time scale between sample distribution, e.g. if sample throughput per annum is very low. Advice on this respect would be a function of the Advisory Panel. Estimating the assigned value (the `true' result) There are a number of possible approaches to determining the nominally ‘true’ result for a sample but only three are normally considered. The result may be established from the amount of analyte added to the samples by the laboratory preparing the sample; alternatively, a ‘reference’ laboratory (or group of such expert laboratories) may be asked to measure the concentration of the analyte using definitive methods or thirdly, the results obtained by the participating laboratories (or a substantial sub-group of these) may be used as the basis for the nominal ‘true’ result. The organisers of the scheme should provide the participants with a clear statement giving the basis for the assignment of reference values which should take into account the views of the Advisory Panel. Choice of analytical method Participants can use the analytical method of their choice except when otherwise instructed to adopt a specified method. It is recommended that all methods should be properly validated before use. In situations where the analytical result is method-dependent the true value will be assessed using those results obtained using a defined procedure. If participants use a method that is not ‘equivalent’ to the defining method, then an automatic bias in result will occur when their performance is assessed. Performance criteria For each analyte in a round a criterion for the performance score may be set, against which the score obtained by a laboratory can be judged. A ‘running score’ could be calculated to give an assessment of performance spread over a longer period of time.

Analytical methods: quality control and selection

51

Reporting results Reports issued to participants should include data on the results from all laboratories together with participants’ own performance score. The original results should be presented to enable participants to check correct data entry. Reports should be made available before the next sample distribution. Although all results should be reported, it may not be possible to do this in very extensive schemes (e.g. 800 participants determining 15 analytes in a round). Participants should, therefore, receive at least a clear report with the results of all laboratories in histogram form. Liaison with participants Participants should be provided with a detailed information pack on joining the scheme. Communication with participants should be by newsletter or annual report together with a periodic open meeting; participants should be advised of changes in scheme design. Advice should be available to poor performers. Feedback from laboratories should be encouraged so participants contribute to the scheme’s development. Participants should view it as their scheme rather than one imposed by a distant bureaucracy. Collusion and falsification of results Collusion might take place between laboratories so that independent data are not submitted. Proficiency testing schemes should be designed to ensure that there is as little collusion and falsification as possible. For example, alternative samples could be distributed within a round. Also instructions should make it clear that collusion is contrary to professional scientific conduct and serves only to nullify the benefits of proficiency testing. 3.5.5 Statistical procedure for the analysis of results The first stage in producing a score from a result x (a single measurement of analyte concentration in a test material) is to obtain an estimate of the bias, thus: bias ˆ x

X

where X is the true concentration or amount of analyte. The efficacy of any proficiency test depends on using a reliable value for X. Several methods are available for establishing a working estimate of X^ (i.e. the assigned value). Formation of a z-score Most proficiency testing schemes compare bias with a standard error. An obvious approach is to form the z-score given by: z ˆ …x X^ †= where  is a standard deviation.  could be either an estimate of the actual variation encountered in a particular round (~s) estimated from the laboratories’

52

Food chemical safety

results after outlier elimination or a target representing the maximum allowed variation consistent with valid data. A fixed target value for  is preferable and can be arrived at in several ways. It could be fixed arbitrarily, with a value based on a perception of how laboratories should perform. It could be an estimate of the precision required for a specific task of data interpretation.  could be derived from a model of precision, such as the ‘Horwitz Curve’.15 However, while this model provides a general picture of reproducibility, substantial deviation from it may be experienced for particular methods. Interpretation of z-scores If X^ and  are good estimates of the population mean and standard deviation then z will be approximately normally distributed with a mean of zero and unit standard deviation. An analytical result is described as ‘well behaved’ when it complies with this condition. An absolute value of z (jzj) greater than three suggests poor performance in terms of accuracy. This judgement depends on the assumption of the normal distribution, which, outliers apart, seems to be justified in practice. As z is standardised, it is comparable for all analytes and methods. Thus values of z can be combined to give a composite score for a laboratory in one round of a proficiency test. The z-scores can therefore be interpreted as follows: jzj < 2 ‘Satisfactory’: will occur in 95% of cases produced by ‘well behaved results’ 2 < jzj < 3 ‘Questionable’: but will occur in 5% of cases produced by ‘well behaved results’ jzj > 3 ‘Unsatisfactory’: will only occur in 0.1% of cases produced by ‘well behaved results’ Combination of results within a round of the trial There are several methods of combining the z-scores produced by a laboratory in one round of the proficiency test described in the Protocol. They are: The sum of scores, SZ = Rz The sum of squared scores, SSZ = Rz2 The sum of absolute values of the scores, SAZ = R jzj All should be used with caution, however. It is the individual z-scores that are the critical consideration when considering the proficiency of a laboratory. Calculation of running scores Similar considerations apply for running scores as apply to combination scores above.

Analytical methods: quality control and selection

3.6

53

Analytical methods

Analytical methods should be validated as fit-for-purpose before use by a laboratory. Laboratories should ensure that, as a minimum, the methods they use are fully documented, laboratory staff trained in their use and control mechanisms established to ensure that the procedures are under statistical control. The development of methods of analysis for incorporation into International Standards or into foodstuff legislation was, until comparatively recently, not systematic. However, the EU and Codex have requirements regarding methods of analysis and these are outlined below. They are followed by other International Standardising Organisations (e.g. AOAC International (AOACI) and the European Committee for Standardization (CEN)). 3.6.1 Codex Alimentarius Commission This was the first international organisation working at the government level in the food sector that laid down principles for the establishment of its methods. That it was necessary for such guidelines and principles to be laid down reflects the confused and unsatisfactory situation in the development of legislative methods of analysis that existed until the early 1980s in the food sector. The ‘Principles for the Establishment of Codex Methods of Analysis’16 are given below; other organisations which subsequently laid down procedures for the development of methods of analysis in their particular sector followed these principles to a significant degree. They require that preference should be given to methods of analysis the reliability of which have been established in respect of the following criteria, selected as appropriate: • specificity • accuracy • precision; repeatability intra-laboratory (within laboratory), reproducibility inter-laboratory (within laboratory and between laboratories) • limit of detection • sensitivity • practicability and applicability under normal laboratory conditions • other criteria which may be selected as required.

3.6.2 The European Union The Union is attempting to harmonise sampling and analysis procedures in an attempt to meet the current demands of the national and international enforcement agencies and the likely increased problems that the open market will bring. To aid this the Union issued a Directive on Sampling and Methods of Analysis.7 The Directive contains a technical annex, in which the need to carry out a collaborative trial before it can be adopted by the Community is emphasised. The criteria to which Community methods of analysis for foodstuffs should now conform are as stringent as those recommended by any international

54

Food chemical safety

organisation following adoption of the Directive. The requirements follow those described for Codex above, and are given in the Annex to the Directive. They are: 1.

2.

3. 4.

Methods of analysis which are to be considered for adoption under the provisions of the Directive shall be examined with respect to the following criteria: (i) specificity (ii) accuracy (iii) precision; repeatability intra-laboratory (within laboratory), reproducibility inter-laboratory (within laboratory and between laboratories) (iv) limit of detection (v) sensitivity (vi) practicability and applicability under normal laboratory conditions (vii) other criteria which may be selected as required. The precision values referred to in 1 (iii) shall be obtained from a collaborative trial which has been conducted in accordance with an internationally recognised protocol on collaborative trials (e.g. International Organisation of Standardization ‘Precision of Test Methods’).17 The repeatability and reproducibility values shall be expressed in an internationally recognised form (e.g. the 95% confidence intervals as defined by ISO 5725/1981). The results from the collaborative trial shall be published or be freely available. Methods of analysis which are applicable uniformly to various groups of commodities should be given preference over methods which apply to individual commodities. Methods of analysis adopted under this Directive should be edited in the standard layout for methods of analysis recommended by the International Organisations for Standardization.

3.6.3 Other organisations There are other international standardising organisations, most notably the European Committee for Standardization (CEN) and AOACI, which follow similar requirements. Although CEN methods are not prescribed by legislation, the European Commission places considerable importance on the work that CEN carries out in the development of specific methods in the food sector; CEN has been given direct mandates by the Commission to publish particular methods, e.g. those for the detection of food irradiation. Because of this some of the methods in the food sector being developed by CEN are described below. CEN, like the other organisations described above, has adopted a set of guidelines to which its Methods Technical Committees should conform when developing a method of analysis. The guidelines are:

Analytical methods: quality control and selection

55

Details of the interlaboratory test on the precision of the method are to be summarised in an annex to the method. It is to be stated that the values derived from the interlaboratory test may not be applicable to analyte concentration ranges and matrices other than given in annex. The precision clauses shall be worded as follows: Repeatability: The absolute difference between two single test results found on identical test materials by one operator using the same apparatus within the shortest feasible time interval will exceed the repeatability value r in not more than 5% of the cases. The value(s) is (are): . . . Reproducibility: The absolute difference between two single test results on identical test material reported by two laboratories will exceed the reproducibility, R, in not more than 5% of the cases. The value(s) is (are): . . . There shall be minimum requirements regarding the information to be given in an Informative Annex, this being: Year of interlaboratory test and reference to the test report (if available) Number of samples Number of laboratories retained after eliminating outliers Number of outliers (laboratories) Number of accepted results Mean value (with the respective unit) Repeatability standard deviation (sr) (with the respective unit) Repeatability relative standard deviation (RSD r) (%) Repeatability limit (r)w(with the respective units) Reproducibility relative standard deviation (sR) (with the respective unit) Reproducibility relative standard deviation (RSD R) (%) Reproducibility (R) (with the respective unit) Sample types clearly described Notes if further information is to be given.

3.6.4 Requirements of official bodies Consideration of the above requirements confirms that in future all methods must be fully validated if at all possible, i.e. have been subjected to a collaborative trial conforming to an internationally recognised protocol. In addition this, as described above, is now a legislative requirement in the food sector of the European Union. The concept of the valid analytical method in the food sector, and its requirements, is described below.

56

Food chemical safety

3.6.5 Requirements for valid methods of analysis It would be simple to say that any new method should be fully tested for the criteria given above. However, the most ‘difficult’ of these is obtaining the accuracy and precision performance criteria. Accuracy Accuracy is defined as the closeness of the agreement between the result of a measurement and a true value of the measureand.18 It may be assessed with the use of reference materials. However, in food analysis, there is a particular problem. In many instances, though not normally for food additives and contaminants, the numerical value of a characteristic (or criterion) in a Standard is dependent on the procedures used to ascertain its value. This illustrates the need for the (sampling and) analysis provisions in a Standard to be developed at the same time as the numerical value of the characteristics in the Standard are negotiated to ensure that the characteristics are related to the methodological procedures prescribed. Precision Precision is defined as the closeness of agreement between independent test results obtained under prescribed conditions.19 In a standard method the precision characteristics are obtained from a properly organised collaborative trial, i.e. a trial conforming to the requirements of an International Standard (the AOAC/ISO/IUPAC Harmonised Protocol or the ISO 5725 Standard). Because of the importance of collaborative trials, and the resource that is now being devoted to the assessment of precision characteristics of analytical methods before their acceptance, they are described in detail below. Collaborative trials As seen above, all ‘official’ methods of analysis are required to include precision data. These may be obtained by subjecting the method to a collaborative trial conforming to an internationally agreed protocol. A collaborative trial is a procedure whereby the precision of a method of analysis may be assessed and quantified. The precision of a method is usually expressed in terms of repeatability and reproducibility values. Accuracy is not the objective. Recently there has been progress towards a universal acceptance of collaboratively tested methods and collaborative trial results and methods, no matter by whom these trials are organised. This has been aided by the publication of the IUPAC/ISO/AOAC Harmonisation Protocol on Collaborative Studies.14 That Protocol was developed under the auspices of the International Union of Pure and Applied Chemists (IUPAC) aided by representatives from the major organisations interested in conducting collaborative studies. In particular, from the food sector, the AOAC International, the International Organisation for Standardisation (ISO), the International Dairy Federation (IDF), the Collaborative International Analytical Council for Pesticides (CIPAC), the Nordic Analytical Committee (NMKL), the Codex Committee on Methods of

Analytical methods: quality control and selection

57

Analysis and Sampling and the International Office of Cocoa and Chocolate were involved. The Protocol gives a series of 11 recommendations dealing with: • • • • • • • •

the components that make up a collaborative trial participants sample type sample homogeneity sample plan the method(s) to be tested pilot study/pre-trial the trial proper.

3.6.6 Statistical analysis It is important to appreciate that the statistical significance of the results is wholly dependent on the quality of the data obtained from the trial. Data that contain obvious gross errors should be removed prior to statistical analysis. It is essential that participants inform the trial co-ordinator of any gross error that they know has occurred during the analysis and also if any deviation from the method as written has taken place. The statistical parameters calculated and the outlier tests performed are those used in the internationally agreed Protocol for the Design, Conduct and Interpretation of Collaborative Studies.14

3.7

Standardised methods of analysis for contaminants

There are many organisations that publish standardised methods of analysis for contaminants, such methods normally having been validated through a collaborative trial organised to conform to one of the internationally accepted protocols described previously. Such organisations will include AOACI, the European Organisation for Standardisation (CEN) and the Nordic Committee for Food Analysis (NMKL). Within Europe, the most important of these international standardising organisations is probably CEN. CEN has a technical committee dealing with horizontal methods of analysis in which both additive and contaminant methods of analysis are discussed (TC 275). The methods of analysis for contaminants within its work programme are outlined below. This is given by Working Group. The titles under the Working Group heading refer to the work item (topic area) of that Working Group. The Working Groups not listed (e.g. 1, 2, etc.) are concerned with additive methods of analysis. Work programme of CEN TC 275 Working Group 3: Pesticides in Fatty Foods Work Item A: determination of pesticides and polychlorinated biphenyls (PCBs): Part 1: general considerations Part 2: extraction of fat, pesticides and PCBs and determination of fat content

58

Food chemical safety

Part 3: clean-up methods Part 4: determination, confirmatory tests, miscellaneous. Work programme of CEN TC 275 Working Group 4: Pesticides in Non-Fatty Foods Work Item A: multiresidue methods for the gas chromatographic determination of pesticide residues: Part 1: general considerations Part 2: methods for extraction and clean-up Part 3: determination and confirmatory tests. Work Item B: determination of dithiocarbamate and thiuram disulfide residues: Part 1 spectrometric method Part 2: gas chromatographic method Part 3: xanthogenate method. Work Item C: determination of bromide residues: Part 1: determination of total bromide as inorganic bromide Part 2: determination of bromide. Work Item D: determination of N-methyl carbamate residues. Work Item E: determination of benomyl, carbendazim, thiabendazole and thiophanate-methyl. Work programme of CEN TC 275 Working Group 5: Biotoxins Work Item A: determination of aflatoxin B1 and/or the sum of B1, B2, G1 and G2 in cereals, shell fruits and derived products – high-performance liquid chromatographic method with postcolumn derivatisation and immunoaffinity column. Work Item B: determination of ochratoxin A in cereals and cereal products: Part 1: HPLC method with silica gel clean-up Part 2: HPLC method with bicarbonate clean-up. Work Item C: determination of ochratoxin A in cereals and cereal products – HPLC method with immunoaffinity clean-up. Work Item D: determination of patulin content. Work Item E: determination of fumonisins. Work Item F: criteria of analytical methods for mycotoxins – CEN-Report.

Analytical methods: quality control and selection

59

Work Item G: determination of domoic acid in mussels. Work Item H: determination of aflatoxin B1 and total aflatoxins by immunoaffinity column clean-up and HPLC in fig paste, pistachios, peanut butter and paprika powder. Work Item I: determination of okadaic acid and dinophysis toxin in mussels by HPLC. Work Item J: determination of saxitoxin and dicarbamoyl saxitoxin in mussels by HPLC. Work Item K: determination of aflatoxin M1 in liquid milk. Work programme of CEN TC 275 Working Group 6: Microbiology Work Item A: enumeration of Staphylococcus aureus: Part 1: colony count technique with confirmation of colonies (ISO/DIS 68881: 1997) Part 2: colony count technique without confirmation of colonies (ISO/DIS 6888-2: 1997). Work Item B: horizontal method for the detection of coagulase positive Staphylococci (Staphylococci aureus and other species). Work Item C: horizontal method for the detection and enumeration of Listeria monocytogenes: Part 1: detection method Part 2: enumeration method. Work Item D: enumeration of Clostridium perfringens – colony count technique. Work Item E: horizontal method for the detection of Salmonella. Work Item F: general guidance for enumeration of Bacillus cereus – colony count technique at 30ºC. Work Item G: detection of thermotolerant Campylobacter. Work Item H: detection of Yersinia enterocolitica. Work Item I: preparation of the test sample, of initial suspension and of decimal dilutions, for microbiological examination: Part 1: general rules for the preparation of the initial suspension and of decimal dilutions

60

Food chemical safety

Part 2: specific rules for the preparation of the test samples and initial suspension of meat and meat products Part 3: specific rules for the preparation of the test samples and initial suspension of milk and milk products Part 4: specific rules for the preparation of the test samples and initial suspension of fish products Part 5: specific rules for the preparation of the test samples and initial suspension of products other than milk and milk products, meat and meat products and fish products. Work Item J: general guidance for microbiological examinations. Work Item K: validation of alternative microbiological methods. Work Item L: guidelines on quality assurance and performance testing of culture media: Part 1: quality assurance of culture media in the laboratory Part 2: performance testing Part 3: practical implementation of the general guideline on quality assurance of culture media in the laboratory. Part 4: performance testing of culture media. Work Item M: horizontal method for the detection of Escherichia coli O 157. Work Item N: horizontal method for the enumeration of Bacillus cereus. Work Item O: guidelines on quality assurance and performance testing of culture media (to be elaborated as European Prestandards C02/97, C03/97): Part 1: quality assurance of culture media in the laboratory Part 2: practical implementation of the general guidelines on quality assurance of culture media in the laboratory Part 3: performance testing. Work programme of CEN TC 275 Working Group 10: Determination of Trace Elements Work Item A: determination of trace elements – general considerations. Work Item B: determination of mercury by CVAAS after pressure digestion. Work Item C: determination of lead and cadmium by ETAAS after dry ashing. Work Item D: performance criteria and general considerations. Work Item E: pressure digestion.

Analytical methods: quality control and selection

61

Work Item F: determination of lead, cadmium, chromium and molybdenum by ETAAS after pressure digestion. Work Item G: determination of lead, cadmium, zinc, copper, iron, chromium and nickel after dry ashing. Work Item H: determination of lead and cadmium by ETAAS after microwave digestion. Work programme of CEN TC 275 Working Group 11: Genetically Modified Organisms Work Item A: detection of genetically modified organisms and derived products – sampling. Work Item B: detection of genetically modified organisms and derived products – nucleic acid extraction. Work Item C: detection of genetically modified organisms and derived products – qualitative nucleic acid based methods. Work Item D: detection of genetically modified organisms and derived products – protein-based methods.

3.8

The future direction for methods of analysis

There is current discussion on an international basis whereby the present procedure by which specific methods of analysis are incorporated into legislation are replaced by one in which method performance characteristics are specified. This is because by specifying a single method: • the analyst is denied freedom of choice and thus may be required to use an inappropriate method in some situations • the procedure inhibits the use of automation • it is administratively difficult to change a method found to be unsatisfactory or inferior to another currently available.

As a result the use of an alternative approach whereby a defined set of criteria to which methods should comply without specifically endorsing specific methods is being considered and slowly adopted in some sectors of food analysis. This approach will have a considerable impact on the food analytical laboratory. There are a number of issues that are of concern to the food analytical community of which analysts should be aware. These are outlined briefly below.

62

Food chemical safety

3.8.1 Measurement uncertainty It is increasingly being recognised both by laboratories and the customers of laboratories that any reported analytical result is an estimate only and the ‘true value’ will lie within a range around the reported result. The extent of the range for any analytical result may be derived in a number of different ways, e.g. using the results from method validation studies or determining the inherent variation through different components within the method, i.e. estimating these variances as standard deviations and developing an overall standard deviation for the method. There is some concern within the food analytical community as to the most appropriate way to estimate this variability.

3.8.2 In-house method validation There is concern in the food analytical community that although methods should ideally be validated by a collaborative trial, this is not always feasible for economic or practical reasons. As a result, IUPAC guidelines are being developed for in-house method validation to give information to analysts on the acceptable procedure in this area. These guidelines should be finalised by the end of 2001.

3.8.3 Recovery It is possible to determine the recovery that is obtained during an analytical run. Internationally harmonised guidelines have been prepared which indicate how recovery information should be handled. This is a contentious area amongst analytical chemists because some countries of the organisations require analytical methods to be corrected for recovery, whereas others do not. Food analysts should recognise that this issue has been addressed on an international basis.

3.9 1. 2. 3. 4.

References EUROPEAN UNION,

Council Directive 89/397/EEC on the Official Control of Foodstuffs, O.J. L186 of 30.6.1989. EUROPEAN UNION, Council Directive 93/99/EEC on the Subject of Additional Measures Concerning the Official Control of Foodstuffs, O.J. L290 of 24.11.1993. EUROPEAN COMMITTEE FOR STANDARDIZATION, General Criteria for the Operation of Testing Laboratories – European Standard EN 45001, Brussels, CEN/CENELEC, 1989. ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT , Decision of the Council of the OECD of 12 Mar 1981 Concerning the Mutual Acceptance of Data in the Assessment of Chemicals, Paris, OECD, 1981.

Analytical methods: quality control and selection 5. 6. 7.

8. 9. 10. 11.

12. 13. 14. 15. 16. 17. 18. 19.

EUROPEAN COMMITTEE FOR STANDARDIZATION,

63

General Criteria for the Assessment of Testing Laboratories – European Standard EN45002, Brussels, CEN/CENELEC, 1989. EUROPEAN COMMITTEE FOR STANDARDIZATION, General Criteria for Laboratory Accreditation Bodies – European Standard EN45003, Brussels, CEN/CENELEC, 1989. EURPEAN UNION , Council Directive 85/591/EEC Concerning the Introduction of Community Methods of Sampling and Analysis for the Monitoring of Foodstuffs Intended for Human Consumption, O.J. L372 of 31.12.1985. INTERNATIONAL ORGANIZATION FOR STANDARDIZATION , General Requirements for the Competence of Calibration and Testing Laboratories ISO/IEC 17025, Geneva, ISO, 1999. CODEX ALIMENTARIUS COMMISSION, Report of the 21st Session of the Codex Committee on Methods of Analysis and Sampling – ALINORM 97/ 23A, Rome, FAO, 1997. CODEX ALIMENTARIUS COMMISSION, Report of the 22nd Session of the Codex Alimentarius Commission – ALINORM 97/37, Rome, FAO, 1997. INTERNATIONAL UNION OF PURE AND APPLIED CHEMISTRY , The International Harmonised Protocol for the Proficiency Testing of (Chemical) Analytical Laboratories, ed. Thompson M and Wood R, Pure Appl. Chem., 1993 65 2123–2144 (also published in J. AOAC International, 1993 76 926–940). INTERNATIONAL UNION OF PURE AND APPLIED CHEMISTRY, Guidelines on Internal Quality Control in Analytical Chemistry Laboratories, ed. Thompson M and Wood R, Pure Appl. Chem., 1995 67 649–666. INTERNATIONAL ORGANIZATION FOR STANDARDIZATION, Calibration and Testing Laboratory Accreditation Systems – General Requirements for Operation and Recognition – ISO/IEC Guide 58, Geneva, ISO, 1993. HORWITZ W, ‘Protocol for the Design, Conduct and Interpretation of Method Performance Studies’, Pure Appl. Chem, 1988 60 855–864 (revision published 1995). HORWITZ W, ‘Evaluation of Analytical Methods for Regulation of Foods and Drugs’, Anal. Chem., 1982 54 67A-76A. CODEX ALIMENTARIUS COMMISSION, Procedural Manual of the Codex Alimentarius Commission – Tenth Edition, Rome, FAO, 1997. INTERNATIONAL ORGANIZATION FOR STANDARDIZATION, Precision of Test Methods – Standard 5725, Geneva, ISO, 1981 (revised 1986 with further revision in preparation). INTERNATIONAL ORGANIZATION FOR STANDARDIZATION, International Vocabulary for Basic and General Terms in Metrology – 2nd Edition, Geneva, ISO, 1993. INTERNATIONAL ORGANIZATION FOR STANDARDIZATION, Terms and Definitions used in Connections with Reference Materials – ISO Guide 30, Geneva, ISO, 1992.

64

Food chemical safety

Appendix: Information for potential contractors on the analytical quality assurance requirements for food chemical surveillance exercises Introduction The FSA undertakes surveillance exercises, the data for which are acquired from analytical determinations. The Agency will take measures to ensure that the analytical data produced by contractors are sufficient with respect to analytical quality, i.e. that the results obtained meet predetermined analytical quality requirements such as fitness-for-purpose, accuracy and reliability. Thus when inviting tenders FSA will ask potential contractors to provide information regarding the performance requirements of the methods to be used in the exercise, e.g. limit of detection, accuracy, precision etc., and the quality assurance measures used in their laboratories. When presenting tenders laboratories should confirm how they comply with these specifications and give the principles of the methods to be used. These requirements extend both to the laboratory as a whole and to the specific analytical determinations being required in the surveillance exercise. The requirements are described in three parts, namely: • Part A: Quality assurance requirements for surveillance projects provided by potential contractors at the time ROAMEs are completed and when commissioning a survey • Part B: Information to be defined by the FSA customer once the contract has been awarded – to be agreed with contractor • Part C: Information to be provided by the contractor on an on-going basis once contract is awarded – to be agreed with the customer.

Each of these considerations is addressed in detail below. Potential contractors are asked to provide the information requested in Part A of this document when submitting ROAME forms in order to aid the assessment of the relative merits of each project from the analytical/data quality point of view. This information is best supplied in tabular form, for example that outlined in Part A, but may be provided in another format if thought appropriate. The tables should be expanded as necessary. Parts B and C should not be completed when submitting completed ROAME forms. Explanation of Parts A, B and C of Document Part A Part A describes the information that is to be provided by potential contractors at the time that the ROAME Bs are completed for submission to the Group. Provision of this information will permit any FSA ‘Analytical Group’ and customers to make an informed assessment and comparison of the analytical quality of the results that will be obtained from the potential contractors bidding for the project. Previously potential contractors have not been given defined

Analytical methods: quality control and selection

65

guidance on the analytical quality assurance information required of them and this has made comparison between potential contractors difficult. Part A is supplied to potential contractors at the same time as further information about the project is supplied. The list has been constructed on the premise that contractors will use methods of analysis that are appropriate and accredited by a third party (normally UKAS), participate in and achieve satisfactory results in proficiency testing schemes and use formal internal quality control procedures. In addition, Parts B and C are made available to the potential contractors so that they are aware of what other demands will be made of them and can build the costs of providing the information into their bids. Part B This section defines the analytical considerations that must be addressed by both the customer and contractor before the exercise commences. Not all aspects may be relevant for all surveys, but each should be considered for relevancy. Agreement will signify a considerable understanding of both the analytical quality required and the significance of the results obtained. Part C This section outlines the information that must be provided by the contractor to a customer on an on-going basis throughout the project. The most critical aspect is the provision of Internal Quality Control (IQC) control charts thus ensuring that the customer has confidence that the contractor is in ‘analytical control’. By following the above the FSA customers will have confidence that the systems are in place in contractors with respect to analytical control and that they are being respected. It is appreciated that not all aspects outlined in Parts A, B and C will be appropriate for every contract but all should be at least considered as to their appropriateness. Contents of Parts A, B and C of document Part A Potential contractors should provide the information requested below. Please provide the information requested either in section 1 or in section 2 and then that in sections 3 to 5. Section 1: Formal quality system in the laboratory if third party assessed (i.e. if UKAS accredited or GLP compliant) Please describe the quality system in your laboratory by addressing the following aspects:

66

Food chemical safety

• To which scheme is your laboratory accredited or GLP compliant? • Please describe the scope of accreditation, by addressing: 1. the area that is accredited 2. for which matrices, and 3. for which analytes or supply a copy of your accreditation schedule. • Do you foresee any situation whereby you will lose accreditation status due to matters outside your immediate control, e.g. closure of the laboratory?

Section 2: Quality system if not accredited Please describe the quality system in your laboratory by addressing the following aspects: • Laboratory Organisation: 1. Management/supervision 2. Structure and organisational chart 3. Job descriptions if appropriate. • Staff: 1. Qualifications 2. Training records 3. Monitoring of the analytical competency of individual staff members. • Documentation 1. General lab procedures 2. Methods to be used (adequate/detailed enough to control consistent approach). • Sample Preparation 1. Location 2. Documented procedures 3. Homogenisation 4. Sub-sampling 5. Sample identification 6. Cross-contamination risk 7. Special requirements. • Equipment Calibration 1. Frequency 2. Who 3. Records 4. Marking. • Traceability 1. Who did what/when 2. Equipment – balances etc. 3. Sample storage/temperature 4. Calibration solutions: how prepared and stored. • Results/Reports 1. Calculation checks

Analytical methods: quality control and selection

67

2. Typographic checks 3. Security/confidentiality of data 4. Software usage/control 5. Job title of approved signatory. • Laboratory Information management System Please outline the system employed. • Internal Audits 1. Audit plan 2. Frequency 3. Who carries out the audit? 4. Are internal audit reports available? 5. What are the non-compliance follow-up procedures? • Sub-contracting 1. In what circumstances is sub-contracting carried out? 2. How is such sub-contracting controlled and audited?

Section 3: Proficiency testing Please describe the arrangements for external proficiency testing in your laboratory by addressing the following aspects: • Do you participate in proficiency testing schemes? If so, which schemes? • Which analyte/matrices of the above schemes do you participate in? • What are your individual proficiency scores and their classification, (e.g. zscores or equivalent), over the past two years, for the analyte/matrices of relevance to this proposal? • What remedial action do you take if you should get unsatisfactory results?

Section 4: Internal quality control Please describe the IQC measures adopted in your laboratory by addressing the following: • • • • • • • •

What control samples do you use in an analytical run? Do you follow the Harmonised Guidelines?1 What IQC procedures are in place? Do you use Certified Reference Materials (CRMs), and if so, how? For example, specify the concentration(s) matrix type(s) etc. Which appropriate CRMs do you use? Do you use In-House Reference Materials (IHRM) and how are they obtained? For example, specify the concentration(s) matrix type(s). Are they traceable? For example, to CRM, a reference method, interlaboratory comparison, or other. What criteria do you have regarding reagent blanks?

1 ‘Guidelines on Internal Quality Control in Analytical Chemistry Laboratories’, ed. M. Thompson and R. Wood, Pure Appl. Chem., 1995, 67, 649–66.

68

Food chemical safety

• • • •

What action/warning limits are applied for control charts? What action do you take if the limits are exceeded? Do you check new control materials and calibration standards? If so, how? Can we see the audit of previous results – what actions have been taken or trends observed? • Do you make use of duplicate data as an IQC procedure? • How frequently are control materials (CRMs, blanks, IHRM etc.) incorporated in the analytical run? • Do you randomise your samples in an analytical run? (including duplicates). Section 5: Method validation Please describe the characteristics of the method of analysis you propose to use in the survey by addressing the following: • What methods do you have to cover the matrix and analyte combinations required? • Do you routinely use the method? • Is the method accredited? • Has the method been validated by collaborative trial (i.e. externally)? • Has the method been validated through any In-House Protocol? • Is it a Standard (i.e. published in the literature or by a Standards Organisation) Method? • Please identify the performance characteristics of the methods, i.e. 1. LOQ 2. LOD 3. Blanks 4. Precision values over the relevant concentration range expressed as relative standard deviations 5. Bias and recovery characteristics including relevant information on traceability. • Do you estimate measurement uncertainty/reliability? • Do you normally give a measurement uncertainty/reliability when reporting results to your customer?

Part B The FSA customer is to consider and then define the following in consultation with the contractor: 1. 2. 3. 4.

What analysis is required for what matrices. The sample storage conditions to be used. Are stability checks for specific analytes undertaken? The methods to be used and a copy of Standard Operating Procedures (SOPs) where accredited, including any sampling and sample preparation protocols, to be supplied to the customer. The IQC procedures to be used. In particular the following should be considered:

Analytical methods: quality control and selection

69

• • • •

5. 6. 7.

8. 9. 10. 11. 12. 13. 14.

the use of the International Harmonised Guidelines for IQC the use of control charts randomisation within the run the composition of the analytical run (e.g. the number of control samples, and in particular the number of blanks, spikes, IHRMs etc.) • the reference materials to be used • the determination of recoveries on each batch using procedures as described in the International Harmonised Guidelines with all results to be corrected for recovery except where otherwise specified (i.e. for pesticides) and for the recovery data quoted to be reported. The measurement limits (i.e. limit of detection (LOD): limit of determination/quantification (LOQ), and reporting limits, etc.). The maximum acceptable measurement reliability (also known as measurement uncertainty) for each analytical result. The treatment of individual results with respect to uncertainty, reliability, i.e. as (a) x  y g/kg where y is the measurement reliability (i.e. as if the sample were to be a ‘historic’ surveillance result), or (b) not less/more than x g/kg where x is the analytical result determined less the measurement reliability (i.e. as if the sample were to be an ‘enforcement style’ result) when assessing compliance with a (maximum) limit. Whether there are to be action limits whereby the customer is immediately notified of ‘abnormal’ results. The procedures to be used for confirmation of ‘abnormal’ results, e.g. those that exceed any defined statutory limit. The procedures to be used if qualitative analysis is to be undertaken. The consistent way of expressing results, e.g. (a) on a wet (as is) basis, on a dry weight basis or on a fat weight basis, and (b) the reporting units for specific analytes to be used throughout survey (i.e. mg/kg etc.). The time interval for customer visits (e.g. once every three months, or as otherwise appropriate) and for submission of control charts. Whether there are any possibilities of developing integrated databases between customer and major contractors. If not, the customer to provide reporting guidelines. The procedure for logging in of samples and traceability of sample in the laboratory. The security of samples within the laboratory.

Part C The following are to be provided by the contractor on an on-going basis throughout the contract to confirm that the contractor remains in ‘analytical control’. 1. Copies of the control charts and duplicate value control charts or other agreed measures to monitor IQC.

70

Food chemical safety

2.

Records of action taken to remedy out-of-control situations to be provided at the same time with control charts. Where action limits have been identified in Part B (see para. 8), the results of samples that exceed the action limits are to be sent to the customer as soon as available. Any relevant proficiency testing scheme results obtained during the course of the survey.

3. 4.

4 Molecular imprint-based sensors in contaminant analysis P. D. Patel, Leatherhead Food Research Association

4.1

Introduction

Sample extraction and preparation remain the most time-consuming and errorprone steps in the analytical process, but these are crucial procedures because food scientists need to isolate and concentrate a wide variety of analytes from complex and varied matrices. Advances in sample extraction and preparation in chemical analysis have only in the past several years been given critical consideration as an important component in obtaining reliable and robust analytical results. This situation is also true of analyses carried out in other industrial sectors (e.g. chemical and microbiological contaminants in food, agriculture and environment). A typical analytical process for PCBs and dioxins involves extraction of the contaminants from a bulk matrix into a solvent (e.g. liquid-liquid extraction with hexane), sample clean-up to remove the bulk co-extractants and allow separation and concentration of the target analyte (e.g. chromatography using fluorisil, silica, alumina and activated carbon) and, finally, identification and quantification (e.g. by GC-electron capture detector or GC-MS). The general steps in the analytical process are equally applicable to other contaminants (e.g. pesticides, antibiotics, b-agonists and mycotoxins). The selected analytical procedures are based on the requirements that must be met for the analytes in question. This includes high sensitivity and low limits of detection (i.e. ppt/subppt levels, owing to the extreme toxicities of some of these compounds), high selectivity (i.e. distinction required between PCBs and dioxin from other coextractants and, possibly, interfering compounds present at much higher concentrations), high specificity (a differentiation of various congeners is desired (e.g. 17 2,3,7,8 substituted PCDDs) and high accuracy and precision (i.e.

72

Food chemical safety

the determination should give a valid estimate of true concentration in the sample). However, in practice, many of the classical techniques can be extremely labour-intensive (and hence costly), cumbersome, lengthy and not wholly reliable. With the current strict regulatory framework (encompassing modern QA systems and product specifications that include trace level measurements of contaminants) under which a variety of foods is produced, there is a dire need for analytical techniques that are rapid, quantitative, reliable, robust and cost-effective. Recently, solid phase extraction (SPE), based on a solid/liquid partition in polypropylene or polystyrene columns or SPE discs containing one or more sorbents, has been widely adopted, replacing many conventional methods of extraction, isolation and concentration of numerous contaminants, including antibiotics, -agonists, mycotoxins, pesticides and tainting compounds (Pimbley and Patel, 2000). Their major advantages include high speed (e.g. using SPE discs), better reproducibility than conventional adsorbents, versatility (applicable to wide-ranging analytes, including hydrophobic, hydrophilic and ionic types), low cost through mass production techniques, price competition and reduction in solvent usage. The SPE systems have been made highly specific and selective for analytes by introduction of antibody-linked sorbents (immunoaffinity). The major advantages of this system include reduction in coextraction of potential cross-reacting or interfering species, concentration of the test analyte on the sorbent, which results in increased sensitivity, and the fact that the subsequent detector need not be highly specific. The immunoaffinity systems have been widely used in selective areas of chemical and microbiological contamination, including analysis of mycotoxins (Dragacci and Fremy, 1996), b-agonists (Haines and Patel, 1998) and microbial pathogens (e.g. Salmonella, Listeria and E. coli O157; Cudjoe et al., 1993). An alternative to the immunoaffinity systems is the use of molecular imprinted polymers (MIPs). Unlike biological antibodies in the immunoaffinity systems, MIPs are based on chemical synthetic polymers, which are highly stable (i.e. do not degrade by enzymic action, processing conditions such as heat and low pH, and organic solvents), can be regenerated extensively, require lowcost reagents to prepare MIPs and can be reproducibly prepared in larger quantities (Owens et al., 1999). This chapter on MIPs addresses the area related to SPE as part of the overall analytical process and covers: (i) some basic principles of MIPs and MIP-based techniques in which MIPs are largely used as sorbents packed in chromatographic columns; (ii) examples of the usage of MIPs as physicochemical recognition elements in sensors with a number of transduction elements (e.g. optical and electrochemical); (iii) some case-study examples of the application of MIPs in contaminant analysis (e.g. antibiotics and clenbuterol), largely based on results obtained in the EC collaborative project (FAIR CT96-1219) co-ordinated by the Leatherhead Food RA Ltd; and (iv) some perspectives on future developments in this field, together with further sources of information for the readers.

Molecular imprint-based sensors in contaminant analysis 73

4.2 The principles of molecularly imprinted polymer-based techniques Molecular imprinting involves preparation of a ‘mould’ comprising polymerisable functional monomers around a print molecule (e.g. Penicillin V, see Fig. 4.1). Initially, the monomer (methacrylic acid) is allowed to establish bond formation with the print molecule and the resulting complexes or adducts are then co-polymerised with cross-linkers (e.g. ethylene glycol dimethacrylate) into a rigid polymer. The print molecule is extracted to leave specific recognition sites in the polymer that are structurally and functionally complementary to the print molecule. The resulting so-called ‘plastic antibodies’ can be functionally as effective as the biological antibodies. For details regarding the structural basis of the imprinting process, refer to Yu and Mosbach (1998). In practice, the molecular imprinting process involves the following steps. 1.

The print molecule (e.g. clenbuterol) is dissolved in a porogen (e.g. acetonitrile) together with either one or two monomers (e.g. methacrylic acid and 4-vinylpyridine). This allows non-covalent complexes to form

Fig. 4.1

Schematic representation of a molecular imprinting process.

Table 4.1

Examples of solid-phase MIP particles for contaminant analysis

Contaminant

Functional monomer

Test format

Sensitivity

Reference

Diaminonaphthalenes in solution

Acrylic acid

NS

Matsui et al., 1993

4-Nitrophenol in river water

10 ug/l

Masque´ et al., 2000

Atrazine herbicide in solution

4-Vinylpyridine (4-VP) MAA

1 ug/ml

Muldoon and Stanker, 1995

Atrazine in bovine liver

MAA

0.02 ppm in meat

Muldoon and Stanker, 1997

Atrazine in solution Atrazine in solution

MAA MAA

0.1 mM 0.25 uM

Matsui et al., 1995 Siemann et al., 1996

Triazine in solution

DEAEM

0.01 mM

Piletsky et al., 1997

Chlorotriazine in environmental samples (natural waters and sediment samples)

MAA

In-situ prepared MIPpacked HPLC column MIP-based SPE clean-up, then RP-HPLC MIP-based competitive inhibition assay (molecularly imprinted sorbent assay) MIP-based SPE clean-up, then HPLC MIP-based HPLC column MIP-based HPLC and radio-ligand assay MIP-based competitive fluorescent assay MIP-based SPE cartridge with LC diode array detection

0.05–0.2 ug/L depending on the type of pesticide

Ferrer et al., 2000

2,4-dichlorophenoxyacetic acid herbicide in solution Simazine herbicide in water

4-VP

Ametryn and other triazines herbicides in tap water Chloramphenicol (CL) in serum

NS

Penicillin V, Penicillin G and oxacillin in solution Clenbuterol in solution

MAA and 4-VP

Listeria monocytogenes cells partitioned between organic/ aqueous medium

Poly(allylamine) used in presence of diacid chloride

NS = Not specified.

MAA

Diethylamino-ethyl methacrylate (DEAEM)

MAA

MIP-based radioligand assay MIP-based SPE and HPLC MIP-based cartridge Competitive displacement of CL-methyl red dye conjugate from CLimprinted polymer packed in HPLC column MIP-packed HPLC column MIP-packed HPLC column Polymeric microcapsules (bacteria-mediated lithography)

30 ng/ml

Haupt et al., 1998

NS

Matsui et al., 1997

< 1 ug/L. However, low recovery (10–40%) 3 ug/ml

Ferrer and Barcelo´, 1999

NS

Skudar et al., 1999

NS

Crescenzi et al., 1998

NS

Aherne et al., 1996 Whitcombe et al., 1997

Levi et al., 1997

76

2.

3.

4.

Food chemical safety between the chemical functionalities of the monomers and complementary ionic groups of the print molecule. In the next step, an excess of cross-linking monomer (e.g. trimethylolpropane trimethacrylate or ethylene glycol dimethacrylate) is added together with an initiator (e.g. 2,2’-azoisobutyronitrile), which induces the polymerisation process. Under nitrogen and high temperature, the polymerisation process results in the formation of a rigid mass of polymer. The mass of polymer is then ground and wet-sieved, normally through 25 m or less mesh with acetone. The collected particles are sonicated and allowed to settle in acetone to eliminate any fines. The solvent is then evaporated, leaving behind particles of size 10–25 m. In most reported applications, MIP particles are then packed into stainless steel columns for HPLC analysis (Yoshizako et al., 1998). However, it should be noted that MIPs have also been used in other applications, including specific sorbents in SPE (Ferrer and Barcelo´, 1999; Stevenson, 1999), support for electrophoretic separations, as membranes and as synthetic models of biological antibodies in sensors and affinity assays (Owens et al., 1999; Surugiu et al., 1999; Takeuchi and Haginaka, 1999). The final step is to extract the print molecule before the MIP particles can be used as solid phase for separation and concentration studies. A range of extractants have been used in this step, e.g. methanol containing triethylamine or acetic acid are added until a steady baseline is obtained.

Molecular imprinting is a relatively new technique for generating solid phase molecular imprinted polymer (MIP)-based affinity systems, but already many applications have been demonstrated (particularly involving organically soluble low molecular weight compounds), including drugs, herbicides, sugars, nucleotides, amino acids and proteins (Bru¨ggemann et al., 2000). Table 4.1 shows examples of MIPs reported for contaminant analysis.

4.3

The development and application of MIP-based sensors

The term sensor has been defined as a device or system – including control and processing electronics, software and interconnection networks – that responds to a physical or chemical quantity to produce an output that is a measure of that quantity. A biosensor comprises two distinct elements: a biological recognition element (e.g. antibodies, enzymes, lectins, receptors and microbial cells) and, in close contact, a signal transduction element (e.g. optical, amperometric, acoustic and electrochemical) connected to a data acquisition and processing system. Thus, the signal generated as a result of analyte interaction with the biological element is converted to a quantifiable signal, e.g. optical or electrochemical. For a review of biosensor applications, the reader is referred to Patel (2000). MIPs are increasingly being used in combination with a range of transducer elements

Molecular imprint-based sensors in contaminant analysis 77 (fibre optic and electrochemical) for the rapid detection and quantification of analytes (Dickert and Hayden, 1999; Yano and Karube, 1999). Some examples of reported applications of MIP-based sensors, also referred to as biomimetic sensors, for contaminant analysis are shown in Table 4.2. Piezoelectric transducer-based MIP sensors are devices based on materials such as quartz crystals (quartz cystal microbalance, QCM) that resonate on application of an external alternating electric field. The frequency of the resulting oscillation is a function of the mass of the crystal. Thus, interaction of an analyte in a sample with the corresponding MIP, previously immobilised to a quartz crystal, will increase the overall mass that is measured as a change in the frequency of oscillation. In the surface acoustic wave (SAW) device, an acoustic wave travels over the surface of the crystal that is coated with the specific MIP. A change in the frequency will occur as the analyte in a sample binds to the corresponding MIP. These MIP-based sensors have been reported for the determination of a range of polycyclic aromatic hydrocarbons (e.g. anthracene, chrysene, pyrene, perylene and benzoperylene) and o-xylene in water (Dickert et al., 1999; Dickert and Hayden, 1999). Two types of electrochemical MIP-based sensor have been reported for the analysis of herbicides atrazine (Piletsky et al., 1995) and 2,4-dichlorophenoxyacetic acid (2,4-D; Kro¨ger et al., 1999). In the atrazine sensor, polymeric membranes containing molecular recognition sites for atrazine were prepared by a molecular imprinting process comprising radical polymerisation of diethyl aminoethyl methacrylate and ethylene glycol dimethacrylate in the presence of atrazine as template. The subsequent electrochemical measurement was carried out in a cell with two platinum electrodes separated by the imprinted membrane. A change in membrane electroresistance was measured by applying a small amplitude alternating voltage with a varying frequency generated by a low-frequency wave form generator to the electrodes, both with and without atrazine, and recorded as a function of time. Atrazine was detected in the range 0.01–0.50 mg/l with no interference from the herbicide simazine. In the 2,4-D sensor, a suspension of 2,4-D imprinted polymer or control polymer particles was air-dried onto the working electrode of the screen-printed electrodes system (electrodes were printed onto polyester sheets). The particles were immobilised using an over-layer of agarose gel. The measurement of 2,4-D was carried out by a competition assay as follows. The electrodes were placed in solutions containing the electroactive probe (homogentisic acid) and the analyte (2,4-D). The non-related electroactive probe had previously been shown to bind effectively to the analyte-imprinted particles, but not to the control particles. After incubation, the excess reactants were washed off the electrodes and the bound probe then quantified by differential-pulse voltammetry using an electrochemical analyser. The peak current generated was directly proportional to the concentration of probe bound to the electrode and inversely proportional to the analyte concentration in the sample. 2,4-D was detected in the range approximately 0.1–100 M, taking into account the background response due to the control particles.

Table 4.2

Examples of MIP-based sensors for contaminant analysis

Contaminant

Functional monomer

Test format

Sensitivity

Reference

o-xylene in solution

NS Methacrylic acid (MAA) or DEAEM 4-VP

QCM: 1 ng SAW: 1 pg 0.01 mg/L

Dickert and Hayden, 1999

Atrazine in solution

MIP-coupled QCM or SAW MIP-based conductimetric sensor MIP-based electrochemical detection on disposable screenprinted electrode MIP-based sensors (QCM, SAW and optical transducers) MIP-based electrochemical sensor

1 uM

Kro¨ger et al., 1999

30 ng/L (ppt level)

Dickert et al., 1999

0.5 ug/ml

Unpublished

2,4-dichlorophenoxy-acetic acid in solution Polycyclic aromatic hydrocarbons (PAHs) in water

Phloroglucinol and triisocyanate

Clenbuterol in solution

MAA

QCM: Quartz crystal microbalance, SAW: Surface acoustic wave.

Piletsky et al., 1995

Molecular imprint-based sensors in contaminant analysis 79

4.4

Case studies: contaminant analysis

Most of the work reported in the literature is related to academic development of MIPs and MIP-based sensors for wide-ranging analytes covering a number of major industrial sectors, including pharmaceutical, agriculture, food and environment. Application of MIPs and MIP-based sensors to real industrial matrices is beginning to increase in order to realise their true potential for commercial exploitation. In this context, MIP-based SPE techniques have been applied to several types of biological sample, including detection of atrazine in chloroform extracts of bovine liver homogenate, propranolol in dog plasma, rat bile and human urine, tamoxifen in human plasma and urine, 7hydroxycoumarine in urine, and nicotine in chewing gum (Owens et al., 1999; Sellergren, 1999). An MIP-based bulk acoustic wave (BAW) sensor has been developed for the estimation of caffeine in human serum and urine (Liang et al., 1999). In this section, two examples of recent application of MIPs to contaminant analysis in the food industry, developed as part of the EC project FAIR CT961219, are considered. 4.4.1 MIP-based extraction of clenbuterol from food The details regarding preparation of clenbuterol imprinted polymers, HPLC columns and detection have been described previously (Crescenzi et al., 1998). A typical chromatogram showing the resolution of clenbuterol and timolol from a mixture at pH values 2.0 and 3.4 is shown in Fig. 4.2. In terms of the selectivity of the stationary phase, expressed as separation factor , the values at pH 2.0 and pH 3.4 were 3.1 and 14.4, respectively. For control particles, the  value was 1. The following procedure was used to determine the potential value of the MIP for extraction of clenbuterol from foodstuffs. Each of beef and lamb kidney samples were spiked with a high level (500 g/ml) of clenbuterol. Unspiked food samples were used as negative controls. Briefly, a sample (1 g) of finely chopped food was mixed with 3 ml of the extraction buffer (acetonitrile:0.001M phosphate buffer pH 3.4, 70:30) containing 1 mg of a non-specific protease. The mixture was homogenised and allowed to incubate for 1 h at 55ºC with shaking. The mixture was cooled to 4ºC, vortexed for 30 s and centrifuged at room temperature for 15 min at 4000 rpm. A portion (2.4 ml) of the supernatant was carefully transferred into a clean vial and mixed with 0.2 ml of the protease inhibitor solution. Finally, the solution was made up to 3.2 ml using the extraction buffer, giving a final concentration of 0.25 g/ml. The extracted samples (in mobile phase: acetonitrile:0.001M phosphate buffer pH 3.4, 70:30) were injected (20 l per sample, 1 ml/min flow rate) onto the clenbuterol-MIP and blank polymer packed HPLC columns. The bound fraction was eluted in the mobile phase and the elution profile monitored at 208– 300 nm.

80

Food chemical safety

Fig. 4.2 Chromatograms of timolol (a) and clenbuterol (b) mixture on polymer imprinted against clenbuterol (EC project FAIR CT96-1219).

Figures 4.3 and 4.4 show the elution profiles of the spiked and control food extracts in the clenbuterol and blank polymer columns. With the blank polymer, the food matrix and clenbuterol were eluted from the HPLC column within 6 min of the sample injection, compared with retention of clenbuterol up to 18 min after sample injection. The retention times for the clenbuterol spiked in both beef and lamb kidney were 18 min 14 s and 18 min 48 s, respectively, indicating reproducible performance of the MIP column. 4.4.2 MIP-based extraction of -lactam antibiotics from milk The details regarding preparation of oxacillin imprinted polymers, HPLC columns and detection have been described previously (Bru¨ggemann et al.,

Molecular imprint-based sensors in contaminant analysis 81

Fig. 4.3 Clenbuterol MIP (a) and control blank MIP (b) based HPLC profiles showing resolution of clenbuterol from spiked (—) and unspiked (- - -) kidney samples (EC project FAIR CT96-1219).

82

Food chemical safety

Fig 4.4 Clenbuterol MIP (a) and control blank MIP (b) based HPLC profiles showing resolution of clenbuterol from spiked (—) and unspiked (- - -) beef samples (EC project FAIR CT96-1219).

Molecular imprint-based sensors in contaminant analysis 83 2000). A typical chromatogram showing the resolution of various -lactam antibiotics from a mixture at pH 3.5 is shown in Fig. 4.5. In terms of the selectivity of the stationary phase, the separation factors  were 1, 2.15 and 1.54 for oxacillin, penicillin G and penicillin V, respectively. The procedure for extraction of the antibiotic oxacillin using oxacillin imprinted polymer involved the following steps: (i) spiked milk samples

Fig. 4.5 Chromatogram representing print molecule (oxacillin) and two other -lactam antibiotics (penicillin G and V) on oxacillin MIP (a) compared with the control blank MIP (b). Reprinted from J. Chromatogr. A, 889, Bru¨ggemann et al. ‘New configurations and applications of molecularly imprinted polymers’, 15–24 (2000), with permission from Elsevier Science.

84

Food chemical safety

containing 0 and 50g/ml oxacillin were centrifuged at 11,000 rpm for 20 min and the pellet discarded; (ii) the supernatants were adjusted to pH 4.5 with 5N hydrochloric acid and then centrifuged as before; (iii) the pH of the supernatants were raised to pH 7.5 with 5N NaOH and methanol added (2 parts to 1 part aqueous phase); (iv) the supernatants were extracted twice with chloroform, each for 10 min under slight agitation; (v) the top aqueous/methanol phases were dried under a stream of nitrogen; (vi) the dried extracts were reconstituted in the mobile phase acetonitrile:0.001 M phosphate buffer, pH 3.5 (50:50) and then injected (20 l per sample, 0.5 ml/min flow rate) onto the HPLC column; and (vii) the bound fraction was eluted in the mobile phase and the elution profile monitored at 208–300 nm. Figure 4.6 shows the elution profiles of the unspiked milk sample (a) and milk spiked at 50 g/ml (b) using oxacillin polymer packed columns. The bulk milk was eluted from the column within 15 min, whereas the oxacillin was retained until after approximately 38 min from the time of the sample injection.

4.5

Future trends

Overall, MIPs are not as yet widely used for routine applications in analytical laboratories because of several significant limitations. These are as follows: (i) unless the template (print material) is inexpensive, preparation of MIPs can be very costly since only about 20% of the starting polymer is employed as HPLC stationary phase, owing to losses during grinding and sieving processes; (ii) robustness and sensitivity are still not adequate, owing to inefficient total removal of the print molecules during MIP preparation, which can result in gradual leaching of the template; (iii) the MIP-based sensors also have lengthy response times and are not particularly easy to prepare reproducibly; and (iv) MIPs are produced at the laboratory level and procedures now need to be developed for scaling up to commercial production. The combination of MIPs with a range of tranducers to produce sensors is expected to be a powerful realtime analytical technology. Already, the potential of this has been demonstrated as described previously, but their true value will be realised only when the basic limitations described previously are satisfactorily addressed. The majority of applications of MIPs are directed to low molecular weight, organically soluble analytes of interest in the pharmaceutical and medicare industries (e.g. chiral drug separation or controlled release of drug, Allender et al., 2000). Applications of MIPs in other industrial sectors will increase, e.g. agriculture, environment and food. In addition, several newer approaches to MIP production (e.g. surface imprinting) are expected to allow MIPs to be prepared in aqueous media for aqueous-soluble analytes, and extend the scope to macromolecules (e.g. proteins, DNA, polysaccharides) and particles (e.g. microorganisms). An interesting new approach to generating synthetic materials capable of specific recognition of macromolecules (e.g. proteins) for use in similar areas of application to conventional MIPs has been reported by Shi and

Molecular imprint-based sensors in contaminant analysis 85

Fig. 4.6 Oxacillin MIP-based HPLC profiles showing resolution of oxacillin from spiked (a) and unspiked (b) milk samples (EC project FAIR CT96-1219).

Ratner (2000). In that study, protein-binding nanocavities were prepared on a polysaccharide-like surface using a novel radiofrequency plasma deposition of thin films. The proteins included albumin, immunoglobulin, fibrinogen, lysozyme and alpha-lactalbulin. Protein-specific nanometer-sized ‘pits’ were clearly demonstrated using techniques such as electron microscopy, mass spectrometry and radioligand binding. It was observed, however, that the structurally unstable protein alpha-lactalbumin exhibited weaker template recognition than the ‘robust’ proteins (e.g. lysozyme). Some other future developments in this field are described below.

86

Food chemical safety

4.5.1 Magnetic MIPs Immunomagnetic particles have been widely used in molecular biology and an increasing number of products are becoming available for food and clinical microbiologists (Cudjoe et al., 1993). Recently, magnetic MIPs have been developed for the beta-blocker (S)-propranolol in which the MIPs retained high binding capability to the print molecule (S)-propranolol compared with control non-imprinted polymer whilst exhibiting superparamagnetic properties (Ansell and Mosbach, 1998). 4.5.2 Catalytic MIPs (also referred to as enzyme mimics or plastizyme) Since MIPs are highly stable and can be sterilised, they are valuable for use in biotransformation processes (Ramstro¨m and Mosbach, 1999). The application of MIP in catalytic reaction has been demonstrated with reference to the enzymic condensation of Z-L-aspartic acid with L-phenylalanine methyl ester to give Zaspartame (Ye et al., 1999). In this study, when the product-imprinted polymer was present, a considerable increase (40%) in product yield was found. 4.5.3 MIP-based sensors and molecularly imprinted sorbent assays These types of MIP-based assay allow direct analyte recognition and estimation from a test sample. To date, several applications of non-radioligand-based competitive assays and MIP-based sensors for contaminant analysis have been reported using MIPs as alternatives to antibodies (Tables 4.1 and 4.2). Already, several other transducers have been combined with MIPs for sensor development, including a surface plasmon resonance sensor for theophylline, caffeine and xanthan (Lai et al., 1998). However, most of these applications have been restricted to assay development using standard solutions and determination of assay characteristics (e.g. sensitivity, selectivity, specificity and cross-reactivity). To be of commercial value, application of the MIP-based sensors and ligand binding assays will need to be demonstrated in real matrices (e.g. foodstuffs, agricultural products, potable and river waters, serum and urine). In addition, the MIP-based systems would have to be at least equivalent to the modern analytical techniques (e.g. HPLC, GC-MS and ELISAs) in terms of performance characteristics and cost of analysis.

4.6

Sources of further information and advice

There is a vast amount of published information available for anyone interested in molecular imprinting, from historical perspective to current and future research and development in the field. This is complemented with proactive workshops, conferences and discussions groups around the world. Much of the information is also available on the Internet, together with other electronic products (e.g. CD-ROMs). In this context, further information can be obtained in

Molecular imprint-based sensors in contaminant analysis 87 the form of reviews (Anon., 1999; Liu et al., 1999; Ramstro¨m and Mosbach, 1999; Mosbach and Haupt, 1998) and books (Reid et al., 1998; Sellergren, 2000). The proceedings of a recent MIP 2000 workshop (first international workshop on molecularly imprinted polymers), held at the Cardiff University on 3–5 July 2000, will be published in a special issue of Analytica Chimica Acta. This will be a valuable resource to find out the current state of the art in the field, major international groups working in the area and developments for real-world use. A range of research and interest groups also exist which proactively encourage and promote molecularly imprinted polymer and related fields internationally. The Molecularly Imprinted Materials for Integrated Chemical Sensors (MIMICS) is a project that is funded by the European Community (Brite-Euram BE-95-1745), with eight pan-European organisations. Details can be found at the following Web site: http://inn7201.casaccia.enea.it/index.html. The site also provides links to major international groups working in the molecular imprinting field (e.g. Lund University, Cambridge University, University of California and University of Kalmar) as well as many references related to the MIP technology. The Society for Molecular Imprinting (c/o Pure and Applied Biochemistry, Lund University, P.O. Box 124, SE-221 00, Lund, Sweden; http:// www.ng.hik.se/~SMI) provides the latest in upcoming international conferences, workshops and venues on molecular imprinting and also new articles written in the field, current job postings, international members and addresses. The LFRA Ltd is an international independent membership-based organisation providing research in all aspects of food science and technology, and information including food legislation and trends in marketing world-wide. Further information on the activities and membership details can be found on our Web sites (www.lfra.co.uk and www.foodindustryweb.com). The author’s multidisciplinary team has considerable expertise in innovative analytical technologies covering many areas of food chemistry, food microbiology, immunology, sensor technology and miniaturisation. For full details and published references, the reader is referred to the following Web site: www.foodconsulting-lfra.com.

4.7

Acknowledgements

All technical work reported on the development of MIPs for clenbuterol and lactam antibiotics was carried out on a recent European-funded project (FAIR CT96-1219), involving partners from University of Lund, University of Rome, Merck Eurolab (Prolabo), UNIR Association and Leatherhead Food Research Association Ltd (LFRA, Co-ordinator). The technical work on application of clenbuterol and -lactam-MIPs to foods was carried out by Mr John Haines and Miss Francesca Aulanta of LFRA Ltd.

88

4.8

Food chemical safety

References

and VULFSON E N (1996), ‘Bacteria-mediated lithography of polymer surfaces’, J Am Chem Soc, 118, 8771–8772. ALLENDER C J, RICHARDSON C, WOODHOUSE B, HEARD C M and BRAIN K R (2000), ‘Pharmaceutical applications for molecularly imprinted polymers’, Int J Pharm, 195, 39–43. ANON. (1999), ‘Several review articles on MIP technology’, Trends Anal Chem, 18, 137–204. ANSELL R J and MOSBACH K (1998), ‘Magnetic molecularly imprinted polymer beads for drug radioligand binding assay’, Analyst, 123, 1611–1616. ¨ GGEMANN O, HAUPT K, LEI Y, YILMAZ E and MOSBACH K (2000), ‘New BRU configurations and applications of molecularly imprinted polymers, Review’, J Chromatogr A, 889, 15–24. CRESCENZI V, MASCI G, FONSI, M and CASATI G (1998), ‘Molecularly imprinted polymers binding clenbuterol’, Presentation at the American Chemical Society meeting, held in Boston, USA (August 1998). CUDJOE K S, PATEL P D, OLSEN E, SKJERVE E and OLSVIK Ø (1993), ‘Immunomagnetic separation techniques for the detection of pathogenic bacteria in foods’, in Kroll R G, Gilmour A and Sussman M, New Techniques in Food and Beverage Microbiology, Oxford, Blackwell Scientific Publications, 17–31. DICKERT F L and HAYDEN O (1999), ‘Molecular imprinting in chemical sensing’, Trends Anal Chem, 18, 192–199. DICKERT F L, TORTSHANOFF M, BULST W E and FISCHERAUER G (1999), ‘Molecularly imprinted sensor layers for the detection of polycyclic aromatic hydrocarbons in water’, Anal Chem, 71, 4559–4563. DRAGACCI S and FREMY J M (1996), ‘Application of immunoaffinity column cleanup to aflatoxin M1’, J Food Protect, 59, 1011–1013. ´ D (1999), ‘Validation of new solid-phase extraction FERRER I and BARCELO materials for the selective enrichment of organic contaminants from environmental samples’ Trends Anal Chem, 18, 180–192. FERRER I, LANZA F, TOLOKAN A, HORVATH V, SELLERGREN B, HORVAI G and ´ D (2000), ‘Selective trace enrichment of chlorotriazine pesticides BARCELO from natural waters and sediment samples using terbuthylazine molecularly imprinted polymers’, Anal Chem, 72, 3934–3941. HAINES J H and PATEL P D (1998), ‘A brief evaluation of the performance of three commercial ELISAs for the analysis of -agonists in foods’, Leatherhead Food RA Scientific and Tech. Notes No. 124. HAUPT K, DZGOEV A and MASBACH K (1998), ‘Assay system for the herbicide 2, 4–dichlorophenoxyacetic acid using a molecularly imprinted polymer as artificial recognition element’, Anal Chem, 70, 628–631. ¨ GER S, TURNER A P, MOSBACH K and HAUPT K (1999), ‘Imprinted polymerKRO based sensor system for herbicides using differential-pulse voltametry on screen-printed electrodes’, Anal Chem, 71, 3698–3702. AHERNE A, ALEXANDER C, PAYNE M J, PEREZ N

Molecular imprint-based sensors in contaminant analysis 89 and POLSKY B (1998), ‘Surface plasmon resonance sensors using molecularly imprinted polymers for sorbent assay theophylline, caffeine and xanthan’, Can J Chem, 76, 265. LEVI R, MCNIVEN S, PILETSKY S, CHEONG S H, YANO K and KARUBE I (1997), ‘Optical detection of chloramphenicol using molecularly imprinted polymers’, Anal Chem, 69, 2017–2021. LIANG C, PENG H, BAO X, NIE L and YAO S (1999), ‘Study of a molecular imprinting polymer coated BAW bio-mimic sensor and its application to the determination of caffeine in human serum and urine’, Analyst, 124, 1781–1785. LIU Q, ZHOU Y X and LIU Y T (1999), ‘Recent progresses in research on molecular imprinting sensors’, Chin J Anal Chem, 27, 1341–1347. MASQUE´ N, MARCE´ R M, BORULL F, CORMACK P A G and SHERRINGTON D C (2000), ‘Synthesis and evaluation of a molecularly imprinted polymer for selective on-line solid-phase extraction of 4-nitrophenol from environmental water’, Anal Chem, 72, 4122–4126. MATSUI J, KATO T, TAKEUCHI T, SUZUKI M, YOKOYAMA K, TAMIYA E and KARUBE I (1993), ‘Molecular recognition in continuous polymer rods prepared by a molecular imprinting technique’, Anal Chem, 65, 2223–2224. MATSUI J, MIYOSHI Y, DOBLHOFF-DIER O and TAKEUCHI T (1995), ‘A molecularly imprinted synthetic polymer receptor selective for atrazine’, Anal Chem, 67, 4404–4408. MATSUI M, OKADA M, TSURUOKA T and TAKEUCHI T (1997), ‘Solid-phase extraction of a triazine herbicide using a molecularly imprinted synthetic receptor’, Anal Commun, 34, 85–89. MOSBACH K and HAUPT K (1998), ‘Some new developments and challenges in non-covalent molecular imprinting technology’, J Mol Recognit, 11, 62– 68. MULDOON M T and STANKER L H (1995), ‘Polymer synthesis and characterisation of a molecularly imprinted sorbent assay for atrazine’, J Agric Food Chem, 43, 1424–1427. MULDOON M T and STANKER L H (1997), ‘Molecularly imprinted solid phase extraction of atrazine from beef liver extracts’, Anal Chem, 69, 803. OWENS P K, KARLSSON L, LUTZ E S M and ANDERSSON L I (1999), ‘Molecular imprinting for bio- and pharmaceutical analysis’, Trends Anal Chem, 18, 146–154. PATEL P D (2000), ‘(Bio)sensors for measurement of analytes implicated in food safety: A review’, Leatherhead Food RA Scientific and Tech. Notes No. 125. PILETSKY S A, PILETSKAYA E V, ELGERSMA A V, YANO K and KARUBE I (1995), ‘Atrazine sensing by molecularly imprinted membranes’, Biosens Bioelectro, 10, 959–964. PILETSKY S A, PILETSKAYA E V, EL’SKAYA A V, LEVI R, YANO K and KARUBE I (1997), ‘Optical detection system for triazine based on molecularlyimprinted polymers’, Anal Lett, 30, 445–455. LAI E P C, FARFARA A, VANDERNOOT V A, KONO M

90

Food chemical safety

and PATEL P D (2000), ‘Agrifood applications of solid-phase extraction: A review’, Leatherhead Food RA Scientific and Tech. Notes No. 197. ¨ M O and MOSBACH K (1999), ‘Synthesis and catalysis by molecularly RAMSTRO imprinted materials’, Curr Opin Chem Biol, 3, 759–764. REID E, HILL H M and WILSON I D (1998), Drug development assay approaches, including molecular imprinting and biomarkers – SP 226. London, Royal Society of Chemistry. SELLERGREN B (1999), ‘Polymer- and template-related factors influencing the efficiency in molecularly imprinted solid-phase extractions’, Trends Anal Chem, 18, 164–175. SELLERGREN B (2000), Molecularly imprinted polymers. Man-made mimics of antibodies and their application in analytical chemistry. London, Elsevier Science. SHI H and RATNER B D (2000), ‘Template recognition of protein-imprinted polymer surfaces’, J Biomed Mater Res, 49, 1–11. SIEMANN M, ANDERSSON L I and MOSBACH K (1996), ‘Selective recognition of the herbicide atrazine by noncovalent molecularly imprinted polymers’, J Agric Food Chem, 44, 141–145. ¨ M O (1999), ¨ GGEMANN O, WITTLESBERGER A and RAMSTRO SKUDAR K, BRU ‘Selective recognition and separation of -lactam antibiotics using molecularly imprinted polymers’, Anal Commun, 36, 327–332. STEVENSON D (1999), ‘Molecular imprinted polymers for solid-phase extraction’, Trends Anal Chem, 18, 154–159. SURUGIU I, YE L, YILMAZ E, DZGOEV A, DANIELSSON B, MOSBACH K and HAUPT K (1999), ‘An enzyme-linked molecularly imprinted sorbent assay’, Analyst, 125, 13–16. TAKEUCHI T and HAGINAKA J (1999), ‘Separation and sensing based on molecular recognition using molecularly imprinted polymers’, J Chromatogr B Biomed Sci Appl, 728, 1–20. WHITCOMBE M J, ALEXANDER C and VULFSON E N (1997), ‘Smart polymers for the food industry’, Trends Food Sci and Technol, 8, 140–145. YANO K and KARUBE I (1999), ‘Molecularly imprinted polymers for biosensor applications’, Trends Anal Chem, 18, 199–204. ˚ NSSON M O and MOSBACH K (1999), ‘Use of ¨ M O, ANSELL R J, MA YE L, RAMSTRO molecularly imprinted polymers in a biotransformation process’, Biotechnol Bioeng, 64, 650–655. YOSHIZAKO K, HOSOYA K, IWAKOSHI Y, KIMATA K and TANAKA N (1998), ‘Porogen imprinting effects’, Anal Chem, 70, 386. YU C and MOSBACH K (1998), ‘Insights into the origins of binding and the recognition properties of molecularly imprinted polymers prepared using an amide as the hydrogen bonding functional group’, J Mol Recognit, 11, 69–74. PIMBLEY D

5 Bioassays in contaminant analysis L. A. P. Hoogenboom, State Institute for Quality Control of Agricultural Products (RIKILT), Wageningen

5.1

Introduction

Since ancient history, humankind has relied on bioassays to determine the safety of food and environment. In medieval times, food tasters were employed to ensure that food was free of poisons. Miners used small birds to detect the possible presence of toxic gases in mining tunnels. With increasing knowledge about the responsible toxicants, improvements in analytical chemistry, combined with the need to reduce animal experiments, we now rely on chemical methods aimed at the detection of compounds by their physicochemical properties. The use of animal bioassays is more or less restricted to the testing of the safety of specific substances, thereby supported by in vitro models with mammalian and prokaryotic cells. However, even today, bioassays with mice and rats are still the only reliable way to detect paralytic and diuretic poisons in shellfish,1,2 and the neurotoxins produced by Clostridium botulinum.3 Fish assays are widely used for testing the quality of drinking water. However, despite the rapid improvements in analytical chemistry, at the same time we start to realize that these methods may no longer be sufficient to deal with the often very complex mixtures of chemicals or ever changing chemical structures of toxicants present as residues in our food chain. Furthermore, there is a strong need for rapid screening assays that can be used for extensive monitoring programmes. Bioassays with pro- or eukaryotic cells capable of detecting compounds based on their effects, offer a possible solution. For the detection of antibiotics in milk and meat, a number of different tests are used for the screening4,5 and in many cases, chemical identification of the responsible substances is no longer required. Recent advances in cell biology and in particular biotechnology have

92

Food chemical safety

allowed the development of a new generation of bioassays, based on the possibilities to introduce specific properties and reporter genes into stable cellular systems. This chapter will describe this new generation of bioassays and demonstrate their advantages, especially when used in combination with sensitive analytical methods. This will be demonstrated by the experiences obtained with the so-called DR-CALUX assay, a bioassay used for the detection of dioxins. The inclusion of bioassays in modern test strategies will allow rapid screening and detection of new, possibly unknown, agonists and help to evaluate the possible health hazards involved with the presence of such compounds in the food chain.

5.2

Dioxins and the DR-CALUX bioassay

5.2.1 Development of the assay After the discovery of dioxins in the food chain, it became clear that it would be impossible to set up large monitoring programmes for this group of compounds. The major reason for this was the very expensive and laborious analytical procedure required to detect 17 different 2,3,7,8-chlorinated dibenzo-p-dioxins (PCDDs) or furans (PCDFs) at the pg/g level. This can only be achieved after extensive clean-up and by using a high resolution mass spectrometer. A set of so-called toxic equivalency factors (TEFs), ranging from 1 to 0.0001, has been developed in order to express the concentrations of each of the congeners into one figure, which represents the group as if it was only the most toxic congener 2,3,7,8-tetrachloro-dibenzo-p-dioxin (TCDD).6 A typical limit for food is the maximum residue limit of 6 pg i-TEQ/g fat for dairy products in the Netherlands. Even lower limits or action levels are used in several other countries. Levels like this are required to meet the tolerable daily intake of 1–4 pg per kg body weight as set by the WHO. It has now become clear that many other substances like the planar non-ortho and mono-ortho PCBs and most likely also some of the brominated polyaromatic hydrocarbons should be included in these limits. New analytical methods are required to include these compounds in the analysis. In response to the limited analytical capacity, bioassays with mammalian cells have been developed, initially based on the known effects of these compounds. Receptor assays have been developed based on the binding of dioxins to a specific receptor (Ah-receptor) in the cells. The so-called ERODassay measures the deethylation of ethoxy-resorufin by certain cytochrome P450 enzymes, following the binding of dioxins to the cytosolic Ah-receptor, the binding of the complex to specific sites (DREs) in the DNA and the increased transcription of the gene encoding for the enzyme. Assays based on this principle are used for determining the levels of dioxin-like compounds in environmental samples like sewage sludge7 and sediments.8 A major drawback of this system is the possible inhibition of the enzyme by many different compounds, including natural occurring substances. The specificity of the test

Bioassays in contaminant analysis 93

Fig. 5.1 Principle behind the CALUX bioassay for Ah-receptor agonists. Following binding of the agonist to the Ah-receptor, the complex will be transported to the nucleus and bind to a so-called dioxin responsive element, resulting in the increased transcription of the luciferase gene and production of luciferase. Following incubation this enzyme can subsequently be measured in cell lysates by a light producing reaction.

was therefore tremendously increased by the development of a cell-line which contains the reporter gene luciferase under control of a murine DRE.9, 10, 11 In response to dioxins, this H4IIE rat hepatoma cell-line will synthesize luciferase in a dose-dependent way, which can subsequently be quantified by an enzymatic light producing reaction (Fig. 5.1). Figure 5.2 presents a typical dose-response curve, showing an increased luciferase production at concentrations as low as 0.5 pM. Since the test can be performed in 96 well-plates, a response is obtained with less than 50 fg TCDD. In principle, the amount of dioxins can be quantified by comparison of the response in the test with the calibration curve for TCDD. Several other dioxin and PCB congeners have been tested and were shown to give a response that reflects the differences in the TEF values (Fig. 5.3). However, congeners with a low TEF value showed a relatively low response in the test. This is similary true for 1,2,3,7,8-PeCDD which TEF value was recently adjusted from 0.5 to 1, and which is often a relatively important contributor to the total dioxin content. As a result the test may underestimate the total TEQ content, if calculations were based on the calibration curve for TCDD. However, in general it is evident that the bioassay obeys the TEQ principle and that the result will reflect the total TEQ content of the sample. 5.2.2 Validation for milk fat Following the succesful development of the cells, a rapid clean-up procedure for fat samples was developed, based on the use of an acid silica column. Using

94

Food chemical safety

Fig. 5.2 Dose-response curve for 2,3,7,8-tetrachloro-p-dibenzodioxin (TCDD) in the CALUX bioassay. The concentration (expressed as TCDD) can subsequently be determined by comparing the response obtained with a sample extract with the calibration curve.

Fig. 5.3 Comparison of the relative response of a number of dioxins and non-ortho (#126, 169) and mono-ortho (#105, 118 and 156) PCBs in the CALUX bioassay and the TEF values established by the WHO.6 Compounds were selected based on their relative importance (contribution to total TEQ levels) in food samples. A major difference between the WHO-TEF values and the i-TEF values used in previous studies is the increase of the TEF for 1,2,3,7,8-penta-PCDD from 0.5 to 1, which is not supported by the CALUX assay.

Bioassays in contaminant analysis 95 Table 5.1 Reproducibility of the CALUX assay with milk fat samples. Spiked samples were tested singly in three independent test series (adapted from 12) Sample number

GC/MS CALUX determined dioxin content* determined (pg/g) level series 1 series 2 series 3 Mean (pg i±SD TEQ/g)

CV (%)

(%)

1 2 3 4 5 6

1 3 6 9 12 15

97 4 54 10 27 11

80 110 75 83 78 87

0.0 3.5 6.9 7.1 12.3 14.6

1.0 3.2 4.6 7.2 8.1 11.8

1.3 3.2 2.1 8.3 7.8 12.8

0.8±0.7 3.3±0.2 4.5±2.4 7.5±0.7 9.4±2.5 13.1±1.4

Recovery

*

CALUX determined levels were corrected for the blank sample being respectively 6.3, 2.0 and 3.5 pg/g fat for series 1, 2 and 3 respectively. In addition values were corrected for the difference between relative responses in the CALUX assay and i-TEF values.

dimethylsulphoxide as intermediate, the extracted dioxins are transfered to the tissue culture medium and subsequently added to the cells. After exposure for 24 hours the luciferase concentration in the cells is determined. The test was validated for milk fat using a number of samples spiked at 1 to 15 pg i-TEQ/g fat (1.2–17.5 pg WHO-TEQ/g) with a mix containing the 17 congeners at equal amounts.12 Table 5.1 shows the reproducibility obtained in three independent tests with these samples. Concentrations were calculated based on the TCDD calibration curve, and subsequently corrected for the 15% difference between CALUX and i-TEF values. The calculated limit of detection was around 1 pg iTEQ/g fat, explaining the high variation obtained with the lowest sample. When calculated in i-TEFs, the recovery varied between 70% and 103%. These results demonstrate the suitability of the test to screen milk fat samples. Another important conclusion from these studies was the need to include reference samples with levels around 0 and the residue limit, in order to control for possible impurities introduced with the chemicals, recovery losses and differences in TEF values. Based on this approach, dioxin-like compounds were measured in oil obtained from a large number of different fish and shellfish products. Levels up to 100 pg i-TEQ were detected but based on general agreements these should be corrected for the sometimes very low oil levels in, for example, shellfish. As shown in Fig. 5.4, a good correlation was obtained with the combined dioxin and non-ortho PCB contents in these oils, although in a few cases relatively large differences were observed. Although this might be caused by high levels of mono-ortho PCBs (not included in GC/MS measurements), it cannot be excluded that other compounds are responsible for this effect.

96

Food chemical safety

Fig. 5.4 Comparison of CALUX determined dioxin levels and combined GC/MS determined levels of dioxins and non-ortho PCBs (77, 126 and 169) in fish and shellfish oil. Since oil levels vary widely in these samples, dioxin levels are normally expressed on a pg/g product base.

5.2.3 Citrus pulp incident Following the succesful validation of the test for milk fat, the bioassay was first used in the food and feed area during the Brazilian citrus pulp incident. Increasing milk levels in German cows were traced back to the use of citrus pulp that had been mixed with contaminated lime. Pulp samples of 5 g were extracted and cleaned by the same procedure as used for the milk fat. A rapid comparison between CALUX and GC/MS data showed that the assay was capable of selecting the highly contaminated samples, using a cut-off value of 5000 pg iTEQ/kg. Most samples contained levels higher than this limit and required GC/ MS confirmation. At the end of the crisis the limit was officially set at 500 pg iTEQ/kg, based on the detection limit of the GC/MS method. The test procedure was subsequently optimized and validated. Based on the consideration that an increased response is not necessarily caused by dioxins or dioxin-like PCBs, and that samples with an increased response would still have to be confirmed by GC/ MS, it was decided to switch to a screening approach. This approach is based on the comparison of the response obtained with test sample with that of a reference sample, containing 400 pg i-TEQ/kg. Table 5.2 shows the results obtained with 71 citrus pulp samples containing GC/MS determined levels between <250 and 6800 pg i-TEQ/kg. From 41 samples with a level below 500 pg i-TEQ/kg, 38 were negative and 3 (7%) showed an elevated response. From 30 samples with a level above 500 pg i-TEQ/kg, 27 (90%) showed an elevated response and no

Bioassays in contaminant analysis 97 Table 5.2 Evaluation of field samples citrus pulp and citrus pulp containing animal feed samples* measured by GC/MS and CALUX Content (pg i-TEQ/kg)

Total

Negative

Suspected

500 or lower 500–6800**

41 30***

38 0

3 27

* This series included ten pulp-containing animal feeds, three low and seven high samples. ** Ten samples between 500 and 2000 pg i-TEQ/kg and 20 between 2000 and 6800 pg i-TEQ/kg. *** Three samples showed a cytotoxic response and could not be evaluated.

false-negatives were obtained. Three samples caused a cytotoxic effect, but following further fivefold dilution they showed a positive effect. This shows that the test performs extremely well even at these low residue limits. 5.2.4 Belgian dioxin incident Following the analysis of 100–1000-fold increased dioxin levels in three chicken feed, fat and egg samples in spring 1999, it soon became clear that a major food incident had happened. Due to poor tracebility of the contaminated feed, many food samples became suspected and required testing. During the first month of the crisis, hundreds of samples, particularly milk fat, were screened with the bioassay. Later in the year, this was followed by feed samples due to contaminated kaolinic clay, fish meal, dried grass, and breadmeal. By the end of September, four months after the start of the crisis, almost 1400 samples had been screened (Table 5.3). Fat samples were screened by comparison with a milk fat sample containing 5 pg i-TEQ (2.7 pg i-TEQ dioxins and 2.3 pg WHOTEQ non-ortho PCBs), feed samples with a citrus pulp sample containing 400 pg i-TEQ dioxins. About 10% of these samples showed an elevated response. For various reasons only 38% of these samples was investigated by GC/MS and 55% of the samples was confirmed to contain elevated dioxin levels. Another 20% of the elevated samples could be explained by the presence of elevated levels of PCBs, which are not included in the residue limits. The remaining 25% of suspected samples could not be explained. The majority of the samples (88%) showed a negative response and were reported as such. A routine control Table 5.3 Numbers of samples analysed with the CALUX bioassay in June–September 1999, including the number and fraction of GC/MS analysed negative, toxic and suspected samples. Method

Analysed

Toxic

Negative

Suspected

CALUX GC/MS % GC/MS

1380 157 11

28 28 100

1213 82 7

139 53 38

98

Food chemical safety

programme selected 82 samples for GC/MS analysis. One of these samples, a feed sample containing kaolinic clay, was slightly above the limit, 767 versus 500 pg i-TEQ/kg. This could be explained by the lowered extractability of dioxins from the clay. Starting in 2000 the CALUX bioassay was introduced into monitoring programmes for dioxins in feed and feed ingredients in the Netherlands and in 2001 this will be further extended to meat, eggs, fish, milk and other food samples. 5.2.5 Use of the CALUX assay for other types of samples Other food samples like egg, animal fat and fish oil appear to behave very similarly to milk fat and extensive validation has not been carried out thus far. The suitability was, however, demonstrated by inclusion of positive or spiked samples. The assay was validated for blood samples from wildlife species for high concentrations of dioxins and dioxin-like PCBs.13 A special clean-up procedure was developed and validated for sediment, pore water and other environmental samples, allowing the use of the assay for official testing of these type of samples.14 5.2.6 Specificity In addition to a number of dioxin and PCB congeners, several polyaromatic hydrocarbons15 as well as - and -naphtoflavone, known agonists of the Ahreceptor, were shown to give a response in the test. A similar result was observed with a number of benzimidazole drugs,16 used as fungicides and anthelmintic drugs. The latter compounds showed negative results in the EROD assay with H4IIE cells,17 confirming the insensitivity of the CALUX assay for false-negative results. A typical effect was the slightly elevated response obtained with corticosteroids, in particular dexamethasone.16 These compounds also stimulated the response obtained with TCDD and are suspected to act indirectly on the steps involved in the pathway leading to the response of the cells. This observation has consequences for the direct analysis of plasma samples, which contain varying concentrations of corticosteroids. Very typical was the strong response observed with a hexane extract of blank citrus pulp, which was not further cleaned with acid silica. Translated to TEQ levels, these probably natural compounds would amount to more than 500 ng TEQ/kg, being more than 1000-fold over the limit for dioxins. Several compounds have been reported as potent antagonists of the Ahreceptor pathway and such compounds might in theory cause false-negative effects. On the other hand such compounds could be used to investigate whether an increased signal was caused by a real Ah-receptor agonist. Resveratrol, naphtoflavone and 4-amino-3-methoxyflavone were tested but failed to show the expected effect. The latter two compounds actually caused a positive effect themselves. At present the only clear cause of false-negative effects may be

Bioassays in contaminant analysis 99 compounds causing cytotoxic effects and thus prevent the cells from responding to the agonists. However, in general cytotoxic effects are clearly recognized by visual control of the cells after the exposure and a decreased response in comparison to the control. Although in principle any compound causing a positive response may be regarded as a possibly dioxin-like and therefore health-threatening compound, it may be very difficult to regulate bioactive compounds in food on this basis. Limits for food are normally the result of risk management, taking into account the use of a number of possibly conservative uncertainty factors and sometimes the ALARA (as low as reasonably achievable) principle. As a result, a level above the limit does not necessarily comprise a direct health risk. This is especially clear for certain naturally occurring substances, which would prevent the consumption of many food commodities if treated as food contaminants. Furthermore, the bioassays do not take into account factors like absorption, distribution, metabolism and excretion of compounds, which to a large extent are responsible for the toxicity of the accumulating and persistent dioxins. For this reason it is questionable whether a response in the test should be expressed in dioxin levels, or whether a sample should be regarded as suspected rather than positive when showing an elevated response in the test. However, the specificity of the test can be largely improved by the use of a selective clean-up process like the acid-silica (33% H2SO4) procedure, which is likely to destroy many compounds. Furthermore, in the case of the CALUX assay, the metabolic capacity of the cells and the use of long exposure periods may help to further improve the specificity of the test. In particular some of the polyaromatic hydrocarbons, like benzo(a)pyrene were shown to give a response in the test but only when incubated for a short period. At longer exposure times, the cells were able to metabolize both the active compound and the luciferase produced during the first hours of incubation. In practice, however, samples spiked with benzo(a)pyrene failed to show any effect even after short incubation, when purified on acid silica columns. Future investigations will have to reveal which compounds may actually interfere with the test result and whether these compounds should be included in the TEQ principle and legislation. Ideally, these bioassays should be supported by databanks with compounds that may act as agonists or antagonists. 5.2.7 Future developments Regarding the strong need for rapid and high-throughput screening tests for dioxins, it is essential that the bioassay will be evaluated in an international validation project. In addition a data bank should be established with known agonists in the test but also possible antagonists. This should include possible information on the behaviour of the compounds in the clean-up procedure. In addition, methods should be developed for the identification of unknown agonists which may be of toxicological concern.

100

Food chemical safety

5.3

The use of bioassays for other groups of compounds

5.3.1 Estrogen assays For a number of years there has been an increased awareness of possible adverse effects due to hormonal activities (endocrine disruption) of compounds that were previously considered harmless. Many different compounds, including synthetic hormones, natural plant ingredients, pesticides and plasticizers, were shown to possess estrogenic and antiestrogenic activity. In order to test compounds for their estrogenic potency and to detect such compounds in the environment, a number of bioassays have been developed,18 initially based on the proliferating effects of estrogens on breast cell-lines.19 The use of the E-screen, a test based on the increased proliferation of MCF-7 breast tumor cells, by accident resulted in the recognition of the estrogenic activity of p-nonyl-phenol due to its introduction into plastic used for tissue culture tubes.20 The estrogenic potency was later confirmed in the more classical rodent uterotrophic assay,21 based on an increased weight gain of the uterus in immature rats or mice. The discovery of the estrogenic potency of this widely used class of plastic additives and surfactants by the E-screen demonstrates one of the advantages of the inclusion of bioassays as screening tools into monitoring programmes. Using molecular biological techniques, several reporter gene assays were developed, most of them using yeast cells.22, 23 This was achieved by introducing both the gene encoding for the human estrogen receptor, as well as the reporter gene under control of an estrogen responsive element. Yeast-based assays have been used to measure estrogen activity in environmental samples, and on bovine24 and human plasma samples.25 The latter study revealed that serum estrogen levels and as a result the endogenous production in young children may thus far have been overestimated by the use of too insensitive immunoassays. This much lower endogenous production is an important issue in the safety evaluation of estrogens in food. In analogy with the DR-CALUX assay, the ER-CALUX assay, a new cellline has been developed which produces luciferase in response to estrogens.26 Again the luc-gene under control of estrogen responsive elements (ERE) was introduced into the T47D human breast tumor cell-line, which responds to 17ßestradiol at concentrations below 1 pM (50 fg/well). This sensitivity should in theory allow the detection of estradiol in meat and blood at the pg/g levels. Thus far the test has only been used on environmental samples and requires further investigation in the food area. Similar assays have been developed using other cell-lines, like human HeLa cells.27 5.3.2 Acetyl-choline esterase inhibitors An extremely interesting area for bioassays is the group of acetylcholineesterase inhibitors, the organophosphate and carbamate pesticides. First of all this is based on the need for rapid screening tests, allowing on-site testing of fruits and vegetables. Secondly there is the awareness that the cumulative effects

Bioassays in contaminant analysis 101 of these groups of compounds should be taken into account when evaluating the safe use of these pesticides.28, 29 Several enzyme inhibition assays have been developed based on the use of crude enzyme preparations and the enzymatic conversion of a colourless compound into a coloured substance. Tejada et al.30 developed an assay presented by Chiu et al.31 into an on-site test, based on the use of a fly head or pig liver extract. Filter papers were partly dipped into an acetone extract of the fruit ot vegetable, subsequently sprayed with the enzyme preparation and then treated with a solution of indoxyl acetate, which upon enzymatic conversion to indoxyl turns into the blue dye indigoid. In the case of the presence of an acetylcholinesterase inhibitor, the filter half dipped into the extract will remain white. A similar assay based on purified acetylcholinesterase from electric eel or a crude homogenate of honeybee heads was developed by Hamers et al.32 and used to measure the presence of these type of compounds in water. The sensitivity for different pesticides varies with the type of enzyme extract, which may partly be explained by the need for an enzymatic activation of the compound into its active metabolite. A possible solution may be a chemical or biochemical conversion of the pesticides prior to the testing. Procedures using bromine or rat liver microsomes were developed by Barber et al.33 who used a bioassay for organophosphates based on human neuroblastoma cells. Further optimization and validation should be focused on these facts and ideally the response of different pesticides in the test should be in line with future possibly species-specific TEF values, based on dose-response curves in insects or mammalian organisms. 5.3.3 Shellfish poisons Numerous attempts have been made to replace the bioassays with rats and mice by chemical analytical or immunological assays. In general these methods are unable to detect all the different toxic compounds.34 At present over 21 saxitoxin congeners (paralytic shellfish poisoning), nine different brevitoxins (neurotoxic shellfish poisoning), eight congeners of okadaic acid (diarethic shellfish poisoning) and seven congeners of domoic acid (amnesic shellfish poisoning) have been identified.35 The first two types of compounds are capable of binding to specific sites on the voltage-dependent sodium channel, in the first case resulting in a blockade of neuronal activity, in the second case in a persistent activation. The latter effect is also observed with the Ciguatera toxins, which are structurally related to the brevitoxins. DSPs are inhibitors of ser/thr protein phosphatases and exposure is thought to result in hyperphosphorylation of ion channel proteins in the GI tract, thereby causing impaired water balance and loss of fluids. ASPs are capable of binding to glutamate receptors, resulting in activation of voltage-dependent calcium channels, elevated intracellular calcium levels and eventually neuronal cell death. Regarding the large number of different active compounds, bioassays based on, for example, the use of mouse neuroblastoma cells might be more promising than the existing mouse bioassays and chemical analytical methods. One of these assays, initially developed for

102

Food chemical safety

sodium channel blocking agents (PSP) is based on the reduced cytotoxicity of ouabain and veratridine, which in the presence of, for example, saxitoxins are no longer able to cause the toxic sodium influx in the mouse neuroblastoma cells used in the test.36, 37, 38 A similar test, using cytotoxicity in mouse neuroblastoma cells, was developed for tetrodotoxins.39 A shipable kit for PSP called MIST (Maritime In Vitro Shellfish Test) has been developed and is also based on the same principles. Fairey et al.40 transformed the neuroblastoma cell test into a reporter gene assay by introducing a c-fos-luciferase construct into the cells. In addition to brevitoxin the test was sensitive to ciguatoxins but also to saxitoxins. The latter was achieved by treating the cells with brevitoxin, following treatment with saxitoxins. By blocking the ion channels, the latter compounds reduced the increase in luciferase production caused by brevitoxin. Domoic acid did not show any effect in the test, demonstrating that these cells do not possess the receptors for this group of compounds. Since the mechanism underlying DSP is quite different, assays based on competition for the active site of protein phosphatase have been developed.41, 42

5.4

Future developments

The increasing knowledge on the effects of compounds on signal transduction pathways in cells, including the transcription of specific genes, should lead to the development of many new bioassays. New high throughput techniques may eventually allow rapid screening of many samples for many different types of activities in a relatively short period of time. This will not be possible without the development of rapid but specific clean-up procedures which at present are the rate-limiting step. Areas that are of great interest include hormonal effects like ßagonist and corticosteroid activity, but also neurotoxic effects. Another major area is the shellfish poisons where at present bioassays with animals are still required.

5.5

Acknowledgements

Transfected H4IIE cells were developed by the Department of Toxicology at the Agricultural University in Wageningen, The Netherlands, in cooperation with the University of California in Davis, USA. At present these cells are commercially available from Biodetection Systems (BDS) in Amsterdam, The Netherlands.

5.6 1.

References HUNGERFORD J M,

‘AOAC Official Method 959.08. Paralytic shellfish poison’, Official Methods of Analysis, 16th edition 1995 (Arlington, VA: AOAC International), Chapter 35.1.37.

Bioassays in contaminant analysis 103 2.

LEDOUX M

and HALL S, ‘Proficiency testing of eight French laboratories in using the AOAC mouse bioassay for paralytic shellfish poisoning: interlaboratory collaborative study’, J. AOAC Int 2000 83 305–10.

3.

GALEY F D, TERRA R, WALKER R, ADASKA J, ETCHEBARNE M A, PUSCHNER B,

and TOR E, ‘Type C botulism in dairy cattle from feed contaminated with a dead cat’, J Vet Diagn Investig 2000 12 204–9. NOUWS J F M, OEFFEN G, SCHOUTEN J, VANEGMOND H, KEUKENS H and STEGEMAN H, ‘Testing of raw milk for tetracycline residues’, J Dairy Sci 1998 81 2341–5. NOUWS J F M, BROEX N J G, DEN HARTOG J M P and DRIESENS F, ‘The new Dutch kidney test’, Arch Lebensmittelhyg 1988 39 135–8. FISHER E, WHITLOCK R H, ROCKE T, WILLOUGHBY D

4. 5. 6.

¨ M B, COOK P, VAN DEN BERG M, BIRNBAUM L, BOSVELD B T C, BRUNSTRO FEELEY M, GIESY J P, HANBERG A, HASEGAWA R, KENNEDY S W, KUBIAK T, LARSEN J C, VAN LEEUWEN F X R, LIEM A K D, NOLT C, PETERSON R E, POELLINGER L, SAFE S, SCHRENK D, TILLITT D, TYSKLIND M, YOUNES M,

and ZACHEREWSKI T, ‘Toxic Equivalency Factors (TEFs) for PCBs, PCDDs, PCDFs for humans and wildlife’, Environm Health Persp 1998 106 775–92. ¨ M B, NA ¨ F C and HJELM K, ‘Levels of dioxin-like ENGWALL M, BRUNSTRO compounds in sewage sludge determined with a bioassay based on EROD induction in chicken embryo liver cultures’, Chemosphere 1999 38 2327– 43. GALE R W, LONG E R, SCHWARTZ T R and TILLETT D E, ‘Evaluation of planar halogenated and polycyclic aromatic hydrocarbons in estuarine sediments using ethoxyresorufin-O-deethylase induction of H4IIE cells’, Environm Toxic Chem 2000 19 1348–59. AARTS J M M J G, DENISON M S, DE HAAN L H J, SCHALK J A C, COX M A and BROUWER A, ‘Ah receptor-mediated luciferase expression: a tool for monitoring dioxin-like toxicity’, Organohal Comp 1993 13 361–4.

WÆRN F

7.

8.

9. 10.

AARTS J M M J G, DENISON M S, COX M A, SCHALK A C, GARRISON P A, TULLIS

and BROUWER A, ‘Species-specific antagonism of Ah receptor action by 2,2’,5,5’-tetrachloro- and 2,2’,3,3’,4,4’hexachlorobiphenyl’, Eur J Pharm Environ Tox 1995 293 463–74. SANDERSON J T, AARTS J M M J G , BROUWER A, FROESE K L , DENISON M S and GIESY J P, ‘Comparison of Ah-receptor-mediated luciferase and ethoxyresorufin O-deethylase induction in H4IIE cells: implications for their use as bioanalytical tools for the detection of polyhalogenated aromatic compounds’, Toxicol Appl Pharmacol 1996 137 316–25. K, DE HAAN L H J

11.

12.

BOVEE T F H, HOOGENBOOM L A P, HAMERS A R M, AARTS J M M J G, BROUWER

and KUIPER H A, ‘Validation and use of the CALUX-bioassay for the detection of dioxins and coplanar PCBs in bovine milk’, Fd Add Contam 1998 15 863–75.

A

13.

MURK A J, LEONARDS P E G, BULDER A S, JONAS A S, ROZEMEIER M J C, DENISON M S, KOEMAN J H

and BROUWER A, ‘The CALUX (Chemical-

104

Food chemical safety

14.

Activated Luciferase gene Expression) assay adapted and validated for measuring TCDD equivalents in blood plasma’, Environ Toxic Chem 1997 16 1583–7. MURK A J, LEGLER J, DENISON M S, GIESY J P, VAN DE GUCHTE C and BROUWER A, ‘Chemical-Activated Luciferase gene Expression (CALUX): a novel in vitro bioassay for Ah-receptor active compounds in sediments and pore water’, Fund Appl Toxic 1996 33 149–60.

15.

BOVEE T F H, HOOGENBOOM L A P, TRAAG W A, ZUIDEMA T, HORSTMAN J H J,

and KUIPER H A, ‘Biological screening of Ah receptor agonist activity in butter fat and coconut oil by means of chemical-activated luciferase expression in a genetically engineered cell line (CALUX)’, Organohal Comp 1996 27 303–7. HOOGENBOOM L A P, HAMERS A R M and BOVEE T F H, ‘Bioassays for the detection of growth-promoting agents, veterinary drugs and environmental contaminants in food’, The Analyst 1999 124 79–85. HOOGENBOOM L A P and HAMERS A R M, ‘Effects of oxfendazole on the Ah receptor-mediated induction of ethoxyresorufin-O-deethylase and luciferase activity by 2,3,7,8–tetrachlorodibenzo-p-dioxin in Hepa-1c1c7 and H4IIE cell-lines’, Organohal Comp 1995 25 53–6. AARTS J M M J G, MURK T J, BROUWER A, DENISON M S

16. 17.

18.

ANDERSEN H R, ANDERSSON A-M, ARNOLD S F, AUTRUP H, BARFOED M, BERESFORD N A, BJERREGAARD P, CHRISTIANSEN L B, GISSEL B, HUMMEL R, BONEFELD JØRGENSEN E, KORSGAARD B, LE GUEVEL R, LEFFERS H, MCLACLAN J, MØLLER A, NIELSEN J B, OLEA N, OLES-KARASKO A, PAKDEL F, PEDERSEN K L, PEREZ P, SKAKKEBÆK N E, SONNENSCHEIN C, SOTO A M,

and GRANDJEAN P, ‘Comparison of short-term estrogenicity tests for identification of hormone-disrupting chemicals’, Environ Health Persp 1999 107 (suppl. 1) 89–108. SOTO A M, SONNENSCEIN C, CHUNG K L, FERNANDEZ M F, OLEA N and OLEASERRANO F, ‘The E-screen assay as a tool to identify estrogens: an update on estrogenic environmental pollutants’, Environm Health Persp 1995 103 (suppl. 7) 113–22. SOTO A M, JUSTICIA H, WRAY J W and SONNENSCEIN C, ‘p-Nonyl-phenol: an estrogenic xenobiotic released from ‘modified’ polysterene’, Environm Health Perp 1991 92 167–73. SUMPTER J P, THORPE S M

19.

20. 21.

ODUM J, LEFEVRE P A, TITTENSOR S, PATON D, ROUTLEDGE E J, BERESFORD N

and ASHBY J, ‘The rodent uterotrophic assay: critical protocol features, studies with nonyl phenols, and comparison with a yeast estrogenic assay’, Regul Toxic Pharmacol 1997 25 176–88. PHAM T A, HWUNG Y P, SANTISO M D, MCDONNELL D P and O’MALLEY B W, ‘Ligand-dependant and independent function of the transactivation regions of the human estrogen receptor in yeast’, Mol Endocr 1992 6 1043–50. ROUTHLEDGE E J and SUMPTER J P, ‘Structural features of alkylphenolic chemicals associated with estrogenic activity’, J Biol Chem 1997 272 3280–8. A, SUMPTER J P

22. 23.

Bioassays in contaminant analysis 105 24.

25. 26.

and BLEACH E C L, ‘Determination of oestrogen concentrations in bovine plasma by a recombinant oestrogen receptor-reporter gene yeast bioassay’, Analyst, 1998 123 2585–8. KLEIN K O, BARON J, COLLI M J, MCDONNELL D P and CUTLER G B, ‘Estrogen levels in childhood determined by an ultrasensitive recombinant cell bioassay’, J Clin Invest 1994 94 2475–80. BURDGE G C, COLDHAM N G, DAVE H, SAUER M J

LEGLER J, VAN DEN BRINK C E, BROUWER A, MURK A J, VAN DER SAAG P T,

and VAN DER BURG B, ‘Development of a stably transfected estrogen receptor-mediated luciferase reporter gene assay in the human T47D breast cancer cell line’, Toxicol Sci 1999 48 55–66. BALAGUER P, JOYEAX A, DENISON M S, VINCENT R, GILLESBY B E and ZACHEREWSKI T, ‘Assessing the estrogenic and dioxin-like activities of chemicals and complex mixtures using in vitro recombinant receptorreporter gene assays’, Can J Physiol Pharmacol 1996 74 216–22. NATIONAL RESEARCH COUNCIL, ‘Pesticides in the diets of infants and children, committee on pesticides in the diets of infants and children, board on agriculture and board on environmental studies and toxicology, Commission on Life Sciences’, National Academic Press, Washington DC. VETHAAK A D

27.

28.

29.

MILESON B E, CHAMBERS J E , CHEN W L, DETTBARN W, EHRICH M, ELDEFRAWI A T, GAYLOR D W, HAMERNIK K, HODGSON E, KARCZMAR A G, PADILLA S, POPE C N, RICHARDSON R J, SAUNDERS D R, SHEETS L P, SULTATOS

and WALLACE K B, ‘Common mechanism of toxicity: a case study of organophosphorus pesticides’, Toxic Sci 1998 41 8–20. TEJADA F R, MADAMBA L S P and TEJADA A W, ‘Development of a rapid method for the detection of organophosphate and carbamate insecticides using enzyme inhibition technique’, The Philippine Agriculturist 1996 79 217–23. CHIU C S, KAO C H and CHENG E Y, ‘Rapid bioassay of pesticide residues (RBPR) on fruits and vegetables’, J Agric Res China, 1991 40 186–203. HAMERS T, MOLIN K R J, KOEMAN J H and MURK A J, ‘A small-volume bioassay for quantification of the esterase inhibiting potency of mixtures of organophosphate and carbamate insecticides in rainwater: development and optimization’, Toxic Sci 2000 58 60–7. BARBER D, CORRELL L and EHRICH M, ‘Comparison of two in vitro activation systems for protoxicant organophosphorous esterase inhibitors’, Toxic Sci 1999 47 16–22. KASUGA F, HARAKUDO Y and MACHII K, ‘Evaluation of enzyme-linked immunosorbent assay (ELISA) kit for paralytic shellfish poisoning toxins’, J Fd Hygienic Soc Japan 1996 37 407–10. VAN DOLAH F M, ‘Marine algal toxins: origins, health effects, and their increased occurrence’, Environm Health Persp 2000 108 (suppl. 1) 133–41. GALLACHER S and BIRKBECK T H, ‘A tissue culture assay for the direct detection of sodium channel blocking toxins in bacterial culture supernates’, FEMS Microbiol Lett 1992 92 101–8. LG

30.

31. 32.

33. 34. 35. 36.

106

Food chemical safety

37.

and ‘Paralytic shellfish poison bioassays: automated end-point determination and standardization of the in vitro tissue culture bioassay, and comparison with the standard mouse bioassay’, Toxicon 1993 30 1143–56. GALLACHER S, MACKINTOSH F, SHANKS A M, O’NEILL S, RIDDOCH I and HOWARD F G, ‘Monitoring for paralytic shellfish poisons in Scotland and progress in research to replace the use of the mouse bioassay’, J Shellfish Res, 1998 17 1647–51. HAMASAKI K, KOGURE K and OHWADA K, ‘An improved method of tissue culture bioassay for tetrodotoxin’, Fish Sci 1996 62 825–9. FAIREY E R, EDMUNDS J S G and RAMSDELL J S, ‘A cell-based assay for brevitoxins, saxitoxins, and ciguatoxins using a stably expressed c-fosluciferase reporter gene’, Anal Biochem 1997 251 129–32. SERRES M H, FLADMARK K E and DOSKELAND S O, ‘An ultrasensitive assay for the detection of toxins affecting protein phosphatases’, Toxicon 2000 38 347–60. MOUNTFORD D O, SUZUKI T and TRUMAN P, ‘Protein phosphatase inhibition assay adapted for determination of total DSP in contaminated mussels’, Toxicon 2000 39 383–90. JELLET J F, MARKS L J, STEWART J E, DOREY M L, WATSON-WRIGHT W LAWRENCE J F,

38.

39. 40. 41. 42.

Part II Particular contaminants

This page intentionally left blank

6 Veterinary drug residues S. N. Dixon, Food Standards Agency, London

6.1

Introduction

This chapter will review the way in which veterinary pharmaceutical drugs are used and the controls placed on their use. It will also explain how information from veterinary drug residues surveillance programmes coupled with knowledge of the potential risk they pose can be used to assess the hazards they may present if allowed to enter the human food chain. The development of human pharmaceutical medicines has progressed significantly over the last twenty-five years. This has also been accompanied by a similar expansion in the development and range of pharmaceutical products available to the veterinary profession for the treatment and control of animal diseases. As with human medicines, veterinary medicines have to undergo extensive evaluation for efficacy and safety (target species, operator and consumer) to ensure that they are both effective and safe to use. The residues of these drugs that may be present in edible tissues, for some time following treatment, are of concern for those involved in consumer protection. In 1996 in the UK there were approximately 12 million cattle, 42 million sheep and 126 million poultry. In addition there has been an increase in other areas of production, particularly farmed fish, over the same period. Economic pressure on farmers has necessitated greater productivity, and this has resulted in animal production becoming more intensive with production units containing large numbers of animals. This has further resulted in an increasing need for effective therapeutic and prophylactic (preventative) means for controlling animal diseases. Veterinary drugs also play a vital role in the control of disease transmission from animals to humans (zoonoses), and when used as growth promoters, the efficiency of livestock production systems.

110

Food chemical safety

6.1.1 Therapeutic agents Therapeutic agents are medicines that are used to control infectious diseases (human and animal pathogenic micro-organisms, ecto- and endo-parasites and fungi) in farm and domestic animals. The therapeutic doses used are intended to rid the target animal of the disease-causing agent without adversely effecting the long-term health of the animal. The drugs are usually administered to individual animals by injection or alternatively when larger groups of animals are affected they can be administered orally as a feed additive or in drinking water. However, the latter route does not provide any control over the dose administered and arguably is less effective. 6.1.2 Prophylactic agents As intensified animal production continues to increase, the genetic selection of animals more suited to these production systems together with the prophylactic use of drugs as a precaution against infection has also increased. In particular with the intensification of poultry meat and pig meat production systems where there is a high risk of infection that can be rapidly transmitted, drugs are often administered as a preventative measure. The drugs are given in order to prevent outbreaks of diseases and control parasitic infections (endo- and ecto-parasites), for example at a certain time of year. Although producers are often accused of using prophylactic drug treatment as a substitute for good husbandry, including animal health and welfare, it should be noted that the development of intensive animal production systems would not have been possible without the prophylactic use of drugs. As with all herd or flock management and treatment, the most effective means of administration is in feed (drinking water, feed stuffs) or by dipping animals. However, the major drawback to these methods of administration is the difficulty in controlling the amount of drug received by each individual animal. As a result this has the greatest potential for creating problems with drug residues in the edible tissues of individual animals being presented for slaughter. 6.1.3 Growth-promoting agents Growth promoters fall into two classes: antimicrobial and anabolic agents. Antimicrobial compounds are mixed with feed at sub-therapeutic concentrations and are designed to alter the microbial flora in the gut of the animal by suppressing the activity of some of the naturally occurring bacteria present in the animal’s intestinal tract. The intention is to make more of the feed nutrients available to the animal and not the gut bacteria. This leads to increased feed conversion ratios which result in animals gaining weight more rapidly. However, there is increasing concern over bacteria developing resistance to antibiotics used in this manner and the implications for human health should the resistance to drugs used in animal production be transferred to drugs used in human medicine. In response to scientific evidence supporting these concerns,

Veterinary drug residues 111 the EU has recently banned the use of several antimicrobial feed additives (bacitracin zinc, spiramycin, tylosin phosphate, virginiamycin) and will be reviewing the use of other similar compounds. Anabolic growth promoters exert their effects via the animal’s metabolism. These compounds are usually administered as a pellet implanted in the animal’s ear, the implant releasing the growth promoters at a constant rate over a period of time (70–90 days). As a condition of the product licence animals can be slaughtered only after an acceptable withdrawal period. This, together with discarding the animal’s ear, is considered significantly to reduce any risk of contaminated food entering the human food chain. The use of these compounds is banned in the United Kingdom and the rest of the EU, although there is some evidence that they continue to be used illegally in some EU member states. However, their use is allowed in some other countries (e.g. USA, Canada, South Africa, Australia). The main reason for the ban on their use in the EU is based largely on the hazard these compounds are perceived to present to the consumer. However, the evidence supporting this view is hotly disputed and as result a trade dispute between countries where these compounds are allowed and the EU has developed. 6.1.4 Herd and flock management Hormonal drugs are used to control reproduction in many farm animals, by regulating fertility, in breeding programmes and to control parturition. Animals would not normally be slaughtered after such treatment. However, in the absence of a withdrawal period (the period between administering the drug to the animal, and edible meat or other produce entering the human food chain) it is possible for dairy cattle prior to calving to produce milk that may have contained veterinary drug residues. Withdrawal periods, which will be dealt with later in this chapter, are therefore defined when drugs are licensed. Tranquillisers may be given to animals to reduce excitement and stress or to control aggressive behaviour. Such drugs have the potential to be misused to control stress during transportation to slaughterhouses, and residues of drugs used for this purpose would remain at high concentrations in edible tissues. It is essential that veterinarians and producers observe the instructions for using these and other drugs to avoid this problem. A number of drugs can be employed for more than one purpose, and some examples of these are discussed later in this chapter. However, to reinforce a point already made, the use of therapeutic antimicrobials as growth promoters should be restricted to reduce the risk of pathogenic microorganisms developing resistance to these drugs. Otherwise the antimicrobial agent would be of little or no value in controlling outbreaks of disease caused by the resistant microorganism in either animals or humans.

112

Food chemical safety

6.2

Control of veterinary products in the UK

6.2.1 The Veterinary Products Committee In the United Kingdom, under the Medicines Act 1968, the licensing authority for veterinary drugs is the Ministry of Agriculture, Fisheries and Food (MAFF). The Veterinary Medicines Directorate (VMD) is responsible for advising ministers on the licensing of veterinary medicines. MAFF (VMD) is advised by an independent body of experts, the Veterinary Products Committee (VPC). The VPC includes expertise from a wide variety of disciplines including veterinarians, medical clinicians, toxicologists, agricultural specialists and environmental toxicologists. The inclusion of environmental toxicologists reflects increased concern over the impact of veterinary medicines on the environment, in particular from the fish farming industry where medicines are often released into the environment. The VPC can also seek advice from other expert bodies from within MAFF, the Department of Health and elsewhere. Pharmaceutical companies wishing to register new products must submit a dossier of data, relating to trials and investigations on the product, to MAFF. When MAFF (VMD) is satisfied that the data are complete and the research has been properly conducted to internationally accepted standards of competence (Good Laboratory Practice – GLP), the dossiers are passed to the VPC. The VPC then assesses the quality, efficacy and safety of the product. The safety assessment on the product includes the target species, human (operator, consumer) and environmental impact. If there are any substantial doubts, the licence may be refused and/or the company advised to carry out further investigations. In addressing consumer safety the VPC achieves this by consideration of the toxicological data and the no effect level (NEL = No observed Effect Level) for that substance in experimental animals. This, together with an additional safety factor to allow for any inter-species and intra-species variability, is used to calculate the acceptable daily intake (ADI). For an adult human the ADI is calculated from the NEL by using the following formula for a 60 kg person: NEL  60 ADI ˆ n where n = safety factor. A safety factor of 100 is usually used for an NEL based on the combined toxic effects of a drug residue. Where teratogenic effects (birth defects) are involved a safety factor of 1000 is employed. When licensing veterinary drugs the ADI is compared with the total amount of the substance that would be consumed in the daily diet (edible tissues, milk, etc.). If the daily intake of an extreme consumer of the substance is less than the ADI, a product licence can be granted. If the intake is greater than the ADI or the VPC was dissatisfied for some other reason about the concentration of residues present in the edible tissues, the pharmaceutical company would be advised to increase the recommended

Veterinary drug residues 113 withdrawal period so that residues can reduce to concentrations below the maximum residue limits (MRL). It is the responsibility of the licence applicant to demonstrate that residues concentrations have declined, through the processes of metabolism and excretion, to a safe concentration within the proposed withdrawal period. The product data sheet, which accompanies the veterinary product, is required to state the withdrawal period clearly. Veterinarians and farmers must observe these withdrawal periods if safe concentrations of residues are not to be exceeded. The withdrawal periods are binding in law and observation of them will ensure that the MRLs for veterinary drugs are not exceeded. This, together with residues surveillance, provides a further safeguard for the consumer. 6.2.2 Determining drug withdrawal periods When a veterinary drug is administered to animals on a continual basis the concentration in the animal’s blood, urine, faeces and edible tissues will tend to rise with time until a plateau is reached. At this point the rate at which the drug is being absorbed is equal to the combined effects of metabolism and excretion in eliminating the drug from the animal. The concentrations of the drug in different tissues will not be equal as different types of tissue (muscle, fat, kidney and liver) will have differing rates in the metabolism of the drug and these are reflected in the level and nature of the residues present. In general the concentrations of residues are often higher in the liver, because it is the principal organ of metabolism, and in the kidneys, because of its role in excretion. Drugs are sometimes designed to ‘target’ particular tissues if they are the sites of infection. When administration of the drug ceases, concentrations in tissues gradually decline as the drug is eliminated. Theoretically this decline will be near to exponential. However, in practice considerable variations can occur because tissues with a high affinity for the particular drug retain residues longer, and because individual animals and different species can have different metabolic pathways and rates of excretion. Drugs that have a radioactive isotope incorporated into their structures are invariably used in studies to establish safe withdrawal periods. The presence of the radioactive isotope is harmless to the animal, but offers a highly sensitive means of determining the biological half-life of the drug and the total residue concentrations in different tissues in different animals. The withdrawal period is set at the time taken for residues to fall to a concentration below the maximum residue limit (MRL) in each tissue – where an MRL has been set for a particular tissue and animal species, and allows a reasonable margin of error to allow for individual variations. For drugs where no safe concentration can be defined, for example in the case of some carcinogens, the limit of determination of the most sensitive analytical method available is usually used.

114

Food chemical safety

6.2.3 The influence of EU controls United Kingdom legislation on veterinary drug residues is governed by the regulations and directives that are agreed for the whole of the European Union (EU). The introduction into UK legislation of MRLs for some veterinary drugs follows similar legislation introduced throughout the EU. The EU legislation (Council Regulation No. 2377/90 and supplementary annexes which extend the regulation) sets out maximum residue limits in target tissues for veterinary drugs. The compounds are listed in four annexes (I–IV): Annex I Annex II Annex III Annex IV

Pharmacologically active substances for which an MRL has been fixed Substances not subject to a maximum residue limit Pharmacologically active substances used in veterinary medicine for which a provisional maximum residue limit has been fixed Pharmacologically active substances for which no maximum residue limit can be fixed.

Under UK legislation, it is an offence to sell or supply for slaughter an animal containing residues in excess of the prescribed MRL. (This also applies to animals submitted for slaughter within the specified withdrawal period for a particular veterinary drug and also to meat and meat products.) EC Council Directive 86/469/EEC was concerned with the examination of animals and fresh meat for the presence of residues of veterinary medicines, contaminants and other substances. To standardise the monitoring of veterinary drug residues throughout the EU this Directive was adopted in all EU member states; legislation was passed implementing the Directive in the UK in July 1988. The Directive acts in two ways. Firstly, producers are required to keep adequate records of drugs administered to animals in their charge. These records must include the name of the drug, its dosage and the dates on which it was applied. Secondly, the Directive stipulates sampling schemes for all cattle, pigs, sheep, goats and horses arriving at slaughterhouses and which are destined for human consumption. This Directive has been revised. In the UK statutory surveillance for veterinary drug residues is carried out under the revised Directive (Directive 96/ 23/EC), which extends surveillance to include poultry, salmon and trout. There are also changes in the sampling regime for red meat which require a higher proportion of samples to be taken on farms with a consequent reduction in sampling at slaughterhouses. Other commodities (eggs, milk, honey, farmed and wild game) were included in the statutory surveillance programme from 1 January 1999. The number of samples all EU member states are obliged to test are set out in Annex IV of the revised Directive. The numbers of samples taken from animals and animal producers is a fixed proportion of the animals that each member state forecasts will be slaughtered. In the UK, statutory surveillance is undertaken for the Laboratory of the Government Chemist and the Department of Agriculture for Northern Ireland’s

Veterinary drug residues 115 Science Service laboratories under the National Surveillance Scheme (NSS). These analyses are carried out in fulfilment of the requirements of EC Directives 96/23/EC. It is an offence in UK law under the Medicines (Hormone Growth Promoters) (Prohibition of Use) Regulations 1986 (S.I. [1986] No. 1876) to use any of the substances in Group A1 of the regulation. Any residues of these substances that are detected will be traced back to the farm of origin and, if possible, prosecutions brought. Similar EC Directives extend monitoring to products imported from non-EU countries and cover other farmed species – poultry, farmed fish, game, milk and honey. To assist the implementation of these Directives in the EU, the European Commission has defined and published the minimum requirements for analytical methods to confirm the presence of residues of veterinary drugs together with guidelines for screening methods (Heitzman R. J. 1994). The requirements for these methods are rigorous and ensure that the Directives can be effectively implemented. 6.2.4

The EC Committee for Veterinary Medicinal Products (CVMP) – working group on the safety of residues of veterinary medicines The EU Committee for Veterinary Medicinal Products (CVMP) was established to harmonise licensing procedures for veterinary medicinal products throughout the European Union. The committee recognised that residues of veterinary medicines could lead to barriers to trade in animal products if different standards were adopted in each EU Member State. A working group on the Safety of Residues of Veterinary Medicines has been established in order to agree common safety standards and prevent barriers to trade in meat and animal products within the EU. This working group recommends, to the CVMP, MRLs for veterinary drug residues in meat and other animal products and the results of their deliberations are published in annexes to Commission Regulation 2377/90. All drugs used in veterinary medicines are considered by the working group and where possible an MRL is agreed and recommended for ratification by the CVMP. Where no MRL can be recommended the compound will be either banned from use in the EU, or the sponsoring company will be offered the opportunity to provide data to enable and support the establishment of an MRL. 6.2.5

Other international organisations: Codex Committee on Residues of Veterinary Drugs in Food (CCRVDF) A large proportion of many nations’ animal produce is destined for international trade. In order for consignments to be acceptable to importing countries, common safety standards must be adopted between exporters and importers. The United Nations Food and Agriculture Organisation (FAO) and World Health Organisation (WHO) have collaborated to agree common standards for food additives and pesticides residues and other food contaminants. Historically, a joint FAO/WHO Expert Consultation on Residues of Veterinary Drugs in Foods

116

Food chemical safety

met in Rome in 1984 and on their advice the UN Codex Alimentarius Commission established the Codex Committee on Residues of Veterinary Drugs in Food (CCRVDF). The CCRVDF is comprised of experts on all aspects of veterinary drugs, from UN member nations. The Committee takes expert advice from the Joint FAO/WHO Expert Committee on Food Additives (JECFA) on potential public health hazards arising from veterinary drug residues in food. The CCRVDF works with JECFA to provide recommendations on the use of MRLs and ADIs for veterinary drugs. For example, when the CCRVDF met for the first time in Washington in 1986 it identified some priority substances for JECFA to focus on. In response, JECFA met in 1987 to re-evaluate trenbolone acetate and zeranol, and to evaluate chloramphenical, oestradiol, progesterone and testosterone. After assessing patterns of use in farm animals, the metabolism and pharmacokinetics, toxicological data, residue depletion (under field conditions) and analytical criteria for each compound, JECFA recommends acceptable daily intakes (ADIs) for them. Where JECFA is unable to establish a no effect level (NEL) for the toxic effect of a drug and hence is not able to define an ADI or acceptable residue concentrations (e.g. chloramphenicol), it recommends that the use of this drug should be prohibited. The Committee considered that it was also unnecessary to set an ADI for natural hormones as they are produced endogenously in human beings and show great variations in concentration according to human age, sex and physiological status. It was concluded that residues arising from the use of these compounds, in accordance with good animal husbandry practice, were unlikely to pose a hazard to human health. However, the use of these compounds by injection or subcutaneous implant at the base of the animal’s ear remains a potential hazard should the implant or injection site enter the human food chain. In 1987 the CCRVDF Committee began the process of developing maximum residue limits (MRLs) and has agreed and adopted the following definition of an MRL: MAXIMUM RESIDUE LIMIT (MRL) – is the maximum concentration of residue resulting from the use of a veterinary drug that is recommended by the Codex Alimentarius Commission to be legally permitted or recognised as acceptable in or on a food. The MRL is based on the type and amount of residue considered to be without any direct or indirect toxicological hazard for human health. It is established on the basis of an ADI or, where this is not possible because of insufficient scientific knowledge, on the basis of a temporary ADI that includes an additional safety factor. It takes into account factors such as the development of resistance to the drug, allergenic potential and other undesirable side effects, which may have either a direct or indirect effect on human health. The MRL may also need to be reduced to accommodate residues that may be present in other food products and/or the environment.

Veterinary drug residues 117 A definition of ‘good practice’ in the use of veterinary drugs has also been adopted: GOOD PRACTICE IN THE USE OF VETERINARY DRUGS (GPVD) is the officially recommended or authorised usage approved by national authorities, of veterinary drugs under practical conditions in a manner that leaves toxicologically acceptable residues of the smallest amounts practicable. The CCRVDF has adopted the MRL recommendations of JECFA for a range of veterinary drugs.

6.3

Chemical substances commonly used in veterinary medicines

6.3.1 Antimicrobial and antibiotic agents These compounds are chemicals that selectively inhibit the growth of pathogenic microorganisms, particularly bacteria. Derived either from bacterial metabolism or by chemical synthesis, they may be used therapeutically to treat disease, prophylactically to prevent disease or to promote growth. Growth-promoting antimicrobials were considered to be safe for inclusion in animal feeds at low concentrations, as the quantities used are considerably lower than those required therapeutically; as a result there was proportionately less concern over their residues. However, recent concern has focused on strong evidence indicating the development of bacterial resistance to these compounds. This has led to a ban on the use in animal feed as growth promoters of certain compounds (bacitracin zinc, spiramycin, tylosin phosphate, virginiamycin), the reason being the risk to human (and animal) health from resistant bacteria (Salmonella, Campylobacter, etc.). There is also increased and similar concern over drugs that are used therapeutically in animals and humans (e.g. fluoroquinolones). Sulphonamides (Fig. 6.1) The sulphonamide drugs (Fig. 6.1) are a family of substances derived from sulphanilamide and developed for the treatment of systemic bacterial diseases in human medicine. The core chemical structure of sulphonamide drugs of which there are several thousand individual compounds, strongly resembles p-amino benzoic acid (Fig. 6.1), a key component of the vitamin B complex which is necessary for bacterial survival. Although their use in human medicine has been gradually replaced by more modern antibiotics, they are still widely used therapeutically in animals because of their low cost, convenient (oral) administration and effectiveness. The use of this group of drugs in pig farming is particularly widespread for the control of pneumonia resulting from Haemophilus infections, a disease that can cause severe losses when pigs are kept in enclosed conditions during winter months.

118

Food chemical safety

Fig. 6.1

Sulphonamide drugs.

-Lactams (penicillins, Fig. 6.2a; -Lactamase inhibitors, Fig. 6.2b; cephalosporins, Fig. 6.2c) Isolated originally by Flemming (1929), this group of antibiotics is arguably the foundation that has provided the stimulus for the development of all antibiotics. The chemical structure of the penicillins and cephalosporins are based on the core -lactam ring to which either a five-membered thiazolidine ring (penicillin) or a six-membered thiazine ring (cephalosporin) is bonded. Used extensively to treat bacterial infections (e.g. mastitis) these compounds disrupt the development of the bacterial cell wall (by inhibition of transpeptidase activity) and as they only act on growing cells the greatest effect is when bacterial multiplication is at its greatest. Tetracyclines (Fig. 6.3) The tetracyclines (Fig. 6.3) are a large family of related compounds. They were developed after the penicillins and sulphonamides. The first of these

Veterinary drug residues 119

Fig. 6.2

-lactams. (a) penicillins (b) -lactamase inhibitors (c) cephalosporins.

120

Food chemical safety

Fig. 6.2

Continued.

Veterinary drug residues 121

Fig. 6.3

Tetracyclines.

compounds, chlortetracycline, was isolated from the actinomycete Streptomyces aureofaciens. Derivatives of this parent compound are widely used therapeutically in humans, animals and fish. They are broad spectrum antibiotics which are active against both Gram-positive and Gram-negative bacteria. Their mode of action is on bacterial protein synthesis where they block the attachment of amino-acyl transfer RNA ribosomes causing bacteriostasis. These compounds are used at sub-therapeutic doses as growth-promoting feed additives. Until recently considerable interest has focused on their use and the residues in fish meat resulting from their use in fish farming, as relatively large doses may be required to control the spread of infection among salmon and trout kept in close confinement on fish farms. Their use under these conditions also raises the question of the development of antibiotic resistance by bacteria present in the environment. However, their use in this industry has declined dramatically with the development and introduction of vaccines to control fish diseases. Aminoglycosides (Fig. 6.4) Isolated from fungal sources (Streptomyces griseus), Streptomycin was the first compound isolated and was the foundation for the development of other chemical related compounds. These substances are comprised of aminosugars linked by glycoside bridges. They all have similar antibacterial properties acting mainly against Gram-negative bacteria by interfering with the synthesis of bacterial cell protein and although developed initially for use in human medicine are now used widely to treat animal diseases. Macrolides (Fig. 6.5) The first of these compounds to be used in animal medicines, Tylosin, was isolated from Streptomyces fradiae. Its chemical structure which is typical of this group of compounds, is based on a lactone ring to which carbohydrate

122

Food chemical safety

Fig. 6.4

Aminoglycosides.

Fig. 6.5 Macrolides.

Veterinary drug residues 123 (sugars) are attached. They are active against Gram-positive bacteria, and have anti-mycoplasma and anti-treponema activity. Quinolones and fluoroquinolones (Fig. 6.6) These compounds, with the exception of Oxolinic acid and Naladixic acid, have broad specificity antibacterial activity (Gram +ve, Gram ve bacteria). Their mode of action is by the inhibition of bacterial DNA-gyrase (topoisomerase II), which is responsible for maintaining the topography of DNA. The action is primarily on the bacterial DNA-gyrase with insignificant inhibition of the corresponding mammalian enzyme. Ciprofloxacin, Sarafloxacin, Danofloxacin, Enrofloxacin and Marbofloxacin are widely used to control and treat bacterial and mycoplasma infections in all farmed species, including fish. Despite their widespread use in agriculture the use of these compounds has been increasingly questioned in recent years in relation to the development of resistance to them by human pathogenic bacteria. 6.3.2 Anabolic agents (Fig. 6.7) Anabolic agents used in veterinary medicine are naturally occurring steroids. They are no longer permitted for use as growth promoters in the European

Fig. 6.6

Quinolones and fluoroquinolones.

124

Food chemical safety

Fig. 6.7

Anabolic agents.

Community. The administration of hormone growth promoters was prohibited because of fears about health effects from residues, although there is a need to standardise conditions of trade in animal products from countries outside the EU where these compounds are still licensed for use. The use of one particular group of synthetic anabolic agents with oestrogenic activity, the stilbene compounds,

Veterinary drug residues 125 was prohibited in 1982 because of fears about adverse health effects from residues. The synthetic steroid hormones used are either identical to endogenous (natural) male and female sex hormones (oestradiol, testosterone, progesterone) or have similar structures. Hormones play an important role in growing animals in the development of muscle tissues. Administered as an implant to the base of the animal’s ear the effect on the animal’s metabolism results in a more efficient use of feed and hence more rapid growth. At slaughter the ear is discarded to prevent contamination of food with residues of the drug remaining in the implant. Both natural and synthetic hormonal growth promoters were once in widespread use in beef production in the UK and Continental Europe, to counter the effects of castrating bulls. The castrated male (steer) is deficient in endogenous hormones (androgens and oestrogens) and the administered hormones were used to stimulate normal growth without inducing the steer’s aggressive tendencies. Female cattle (heifers) were also given hormones (androgens) to increase their growth rate towards that of intact male animals. The use of oestrogens in the intact male is used to counter aggressive tendencies and increase growth rates. In healthy, non-castrated animals concentrations of naturally occurring hormones are of the same order of magnitude as those observed in animals that have received hormone implants. However, it is generally accepted that concentrations of residues of hormones in the tissues of treated castrate animals are lower than those that occur naturally in bulls, provided that the hormones had been administered as an implant and the appropriate withdrawal period observed. Otherwise the concentrations of natural hormones in treated animals are usually within the range of concentrations found for the same hormones in untreated animals. The normal concentrations of endogenous hormones in the tissues of treated animals make the detection of illegal use particularly difficult. For this reason sophisticated analytical techniques based on the detection of abnormal ratios of hormones to precursors or metabolites coupled with stable isotope analysis have to be employed to detect illegal use of these compounds. The synthetic hormones, like the natural substances, can be administered in preparations either singly or in combination, for example an androgen (male hormone) and an oestrogen (female hormone), in order to obtain a combination of adult male and female growth characteristics. The stilbenes and zeranol are oestrogenic whilst trenbolone derivatives are androgenic. The total androgen and oestrogen concentrations in muscle from animals treated with combinations of these substances have been estimated to be lower than the combined levels of these hormones found in bulls or pregnant cows. However, detection of illegal use is simpler because the compounds do not occur naturally and thus the presence of any residue is evidence of illegal use. Progestogens, such as progesterone, and oestrogens have also been used in animal husbandry for herd or flock management to synchronise animals coming into ‘season’, so that breeding programmes can be planned, and to induce milk production.

126

Food chemical safety

6.3.3 Anthelmintic agents Cattle and sheep rely to a large extent on grazing of pastures to meet their dietary needs, as do pigs to perhaps a lesser degree. In doing so these animals may ingest extraneous material or return to feed or root on pastures that have already been grazed or contaminated with faeces. The life cycle of most helminths usually involves the excretion of the eggs or larvae in faeces of animals and the subsequent uptake from the pasture by grazing animals. The three major types of helminths are tapeworms, roundworms and flukes. Anthelmintic drugs can act in a variety of ways, by influencing the metabolism of the parasite (e.g. disruption of glucose and glycogen metabolism) or neuromuscular effects (e.g. inhibition of acetyl cholinesterase – organophosphates). The compounds are usually administered to animals orally, by injections, as feed additives or in the form of pour-on preparations. The most widely used compounds, in animals, were levamisole and the benzimidazoles (Fig. 6.8). However these drugs have been largely replaced by ivermectin (Fig. 6.10) which has gained a large share of the animal health market in many countries. Ivermectin has a broad spectrum of activity against parasites and is used in cattle, pigs and sheep. Anthelmintics can be frequently use prophylactically when animals are turned out on pasture which may harbour dormant phases of parasitic organisms from the previous season. However, there is considerable concern and evidence that helminths are developing resistance to these compounds.

Fig. 6.8

Benzimidazoles.

Veterinary drug residues 127

Fig. 6.9

Tetrahydro-imidathiazoles and Tetrahydro-pyrimidines.

Benzimidazoles (Fig. 6.8) This group of compounds has played an important role in the control of worm infestations of cattle, sheep and horses, with particular activity against the larval stages of worms. Prior to their introduction there was no effective treatment of larval stages of worms. The benzimidazoles are also active against tapeworms and fluke. The main compounds in use are thiabendazole, cambendazole, fenbendazole, oxfendazole, mebendazole and albendazole. The mode of action of these compounds is through their effect on the uptake of nutrients by the helminth, which results in a reduction in glycogen and subsequent starvation. Tetrahydro-imidathiazoles and Tetrahydro-pyrimidines (Fig. 6.9) Levamisole (Fig. 6.9) is a tetrahydro-imidathiazole that is widely used and is particularly effective for the control of stomach, intestinal and lung worm in

Fig. 6.10

Ivermectin.

128

Food chemical safety

sheep and cattle. The mode of action of this compound is as a ganglion stimulant which leads to sustained muscular contraction, paralysis and ejection of the nematode. Morantel (Fig. 6.9) is a tetrahydro-pyrimidine which is active against stomach and intestinal roundworms. Its mode of action is as a neuromuscular blocking agent that results in paralysis of the nematode. These compounds can be administered orally, by injection or as a pour-on. Avermectins (Fig. 6.10) Compounds in this group of chemicals are active against both ecto- and endoparasites. The parent compound avermectin is a fungal metabolite of Streptomyces avermitilis. However, hydrogenation of this compound yields ivermectin which is a considerably more effective anthelmintic; indeed this compound has been found to be more potent than most other anthelmintics. Ivermectin is active against nematodes and arthropods (ticks and mites). This group of compounds exert their effect by blocking the GABA ( -amino-butyric acid) neurotransmitter. Anilides and substituted phenols (Fig. 6.11) These chemicals resulted from the search for alternative and better treatments for animals with liver fluke. The first of these, oxyclozanide, and rafoxanide are active against adult fluke. Diamphenethide was developed with specific activity against immature fluke. These substances are administered orally, although the introduction of nitroxynil as an injectable medicine for the control of fluke in sheep and cattle provided a longer acting formulation and hence was of greater benefit. 6.3.4 Coccidiostats (Fig. 6.12) Coccidia are intracellular protazoal parasites that are frequently found in the intestinal epithelial cells of animals; they are also found in other tissues (liver), and are usually transmitted by faecal infection. The potential for infection is at its highest when young animals are brought together in intensive housing systems (e.g. poultry–broiler production). As a result coccidiosis is a major issue for the poultry industry throughout the world. The majority of cocciodiostats appear to exert their influence during the first or second asexual cycle of the parasite. The main coccidiostatic compounds are the sulphonamides, nitrofurans, carbanilides (nicarbazine), pyrimidines (amprolium), 4-hydroxyquinolones (decoquinate), pyridinols (clopidol) and the ionophores (monensin, salinomycin, narasin, lasalocid). The ionophores are probably the largest group of compounds used for the control of coccidiosis, and are used widely to control coccidial gut parasites in young poultry. As the birds mature they develop a natural resistance to the parasites. This group of compounds are also used as growth promoters in feed for beef cattle.

Veterinary drug residues 129

Fig. 6.11

6.3.5

Anilides and substituted phenols.

Tranquillisers (Fig. 6.13) and beta-adrenergic agonists ( -agonists) (Fig. 6.14) Tranquillisers and beta-adrenergic blockers have been used illegally to control stress in animals being transported and while awaiting slaughter. Pigs are particularly sensitive to sudden changes in their environment and the metabolic consequences and stress can adversely affect meat quality. The tranquillisers most likely to have been used for this purpose are azaperone, azaperol and propriopromazine. The -agonist, clenbuterol, is also a tranquilliser. This group of compounds disrupt the uptake of adrenal hormones by nerve cells and stimulate the cardiovascular system. When they are used over a prolonged period, three to four months, they also induce a redistribution of fat to muscle tissue. This group of compounds has been used illegally as feed additives in some European member states as an alternative to the banned hormonal growth promoters. However, the quality of the meat produced where these compounds have been used, is of a poor quality.

130

Food chemical safety

Fig. 6.12

Coccidiostats.

6.3.6 Non-hormonal growth promoters (Fig. 6.15) The quinoxaline di-N-oxides, carbadox and olaquindox, are used as growth promoters in pigs. As both of these compounds are rapidly metabolised

Veterinary drug residues 131

Fig. 6.13

Fig. 6.14

Tranquillisers.

-agonists.

132

Food chemical safety

Fig. 6.15 Non-hormonal growth promoters.

surveillance for residues must include their principal metabolites, quinoxaline carboxylic acid (QCA) and methyl-QCA.

6.4

Surveillance for veterinary drug residues

6.4.1 Introduction The growing concern of consumers and food safety specialists over the potential harmful effects of residues of veterinary drugs in meat and meat products has led to a growth in laboratories that offer services for the surveillance of residues of these compounds in food products. The surveillance of veterinary drug residues has now become well established both within Europe and the developing world. However, despite this there remains considerable scope to continue to develop new and improved analytical methods and to establish laboratory proficiency testing programmes that meet internationally accepted standards. These, together with appropriate legislation, provide effective mechanisms for assessment and control, and reassurance for the consumer. 6.4.2

The UK national surveillance programmes for residues of veterinary drugs in meat and animal products In the UK responsibility for advising government ministers on the safety of the national food supply, with regard to residues of veterinary drugs, rests with the Veterinary Medicines Directorate (VMD). The VMD has delegated responsibility for detailed work on veterinary drug residues to its Advisory Group on Veterinary Residues in Animal Products (AGVR). The AGVR comprises government scientists and administrators, representatives from consumer organisations, other independent experts and representatives from residues surveillance laboratories and industry. The advisory group co-ordinates practical work to determine the incidence and concentrations of veterinary drugs in food, and through the VMD provides advice to government ministers. The results of the surveillance programmes are published periodically in the VMD annual report. The veterinary drugs residues surveillance programme that is co-

Veterinary drug residues 133 ordinated by AGVR is comprised of two facets, statutory and non-statutory surveillance. The Statutory Surveillance Programme implements European Union (EU) legislation in accordance with the provisions set out in Directive 96/23/EEC. This Directive sets out the sampling regime and the veterinary drugs that must be monitored by each EU member state. The Non-Statutory Surveillance Programme is designed to supplement and complement the Statutory Surveillance Programme by covering food products and veterinary drugs residues that are not included in the Statutory Surveillance Programme (e.g. canned meats, prawns, paˆte´, baby food). The Statutory Surveillance Programme, conducted by the State Veterinary Service, concentrates on residues that may pose a potential hazard to health, and protects consumers against undue exposure to such residues. In other nations similar bodies exist for the surveillance of veterinary drug residues in food. Any country that is not a member of the EU, wishing to export meat or animal products into the EU, must have a residues surveillance programme that is compatable with EU legislation. 6.4.3 Surveillance for drug residues in food Surveillance for veterinary drug residues is undertaken with more than one objective in mind. These objectives are: • testing for compliance with national or international food safety standards • establishing the effectiveness of licensing and other control procedures for licensed veterinary drugs • testing for residues or evidence for the illegal use of banned substances • estimating the exposure of consumers to veterinary drug residues in their diet.

These objectives are not comprehensive but the approach taken in surveillance will depend largely upon the desired objective. In the UK the AGVR aims to produce surveillance data that can be used to confirm that there are no harmful residues concentrations in meat or animal products, that veterinary medicines are being used correctly and that no substances are being used illegally. This is achieved by adopting a targeted approach. Consultation with toxicologists and veterinary experts identifies drug/food combinations where residues are likely to occur and which might theoretically present a potential hazard to human health. In such cases ‘pilot’, short-term, surveys are undertaken to estimate the extent of the potential problem. If residues are detected then more extensive surveillance is undertaken to quantify the risk. The AGVR also recommends research to identify potential problem areas. Trials are commissioned to indicate the concentrations of residues that would occur in animals slaughtered with or without the withdrawal period being observed, after administering the drug at the recommended dose. The effects of different types of cooking on residues of veterinary drugs are also considered because while cooking may destroy some

134

Food chemical safety

residues, heat treatment may also release others that have become bound to tissues. The number of samples that needs to be analysed to determine the incidence of residues depends upon the objective of the surveillance and the nature of the potential harmful effects. For veterinary drug residues surveillance it is possible to define the numbers of samples that need to be analysed. The number of samples required to detect residues with a given degree of certainty increases as the incidence of positives declines. For an incidence of 1% with a confidence limit of 95% the minimum number of samples required, for compliance purposes, is 300 per year taken at random. When surveillance is undertaken for food safety purposes the number of samples required will vary. For example, for a compound where a single acute exposure could prove harmful many more samples will be required to confirm the safety of the food supply than for a compound where a long-term chronic exposure is needed to produce harmful effects. In the latter case an estimate of the average exposure is required. Despite this, sampling programmes of less than 100 samples are rarely considered adequate. For food safety purposes the overriding aim is that food contamination should be reduced to the lowest practicable level, bearing in mind the potential costs and benefits involved. Since it is difficult to establish cause and effect relationships following long-term (chronic) exposure at low concentrations, it may be necessary to base action on prudence rather than on proven harm to health. However, if this approach is to maintain the confidence of both consumers and producers of food, a rational evaluation of all relevant information is required so that the balance between the risks and benefits of veterinary drugs can be assessed. Information on the incidence of potentially harmful drug residues is fundamental to this cost-benefit analysis; so too is the consumption of the commodities involved (particularly for susceptible consumers or those consumers who eat more). Account must also be taken of the potential fall in food production if a drug is controlled or prohibited, and also the animal health and welfare implications that may result from the restriction of an animal medicine for which there may be no effective alternative.

6.5

Analytical methods employed in drug residues surveillance

6.5.1 Introduction Veterinary drug residues tend to be present in edible tissues at very low concentrations (ug/kg = parts per billion – ppb). It is therefore important that the presence of a detected residue is confirmed by a different analytical technique from the one used to detect the residue. Although the majority of samples analysed may give negative results, when a positive result is obtained a rigorous analytical approach must be adopted to establish or confirm it. A two-tier analytical approach is adopted by the AGVR in its surveillance programmes in the UK. Screening is

Veterinary drug residues 135 used to detect suspect positive samples which are then confirmed by re-analysing and quantifying by a different and more specific technique. 6.5.2 Screening analyses Ideally, screening techniques should be relatively inexpensive and rapid, and permit a large number of samples to be analysed. For veterinary drug residues analysis the basic criteria that the screening method must meet are: • a limit of determination below the MRL, or as low as possible where no MRL has been set • a low incidence of false negative results • a high level of repeatability and reproducibility.

It is also desirable that a screening test should have a low incidence of false positive results. (False positives will occur if the test is sensitive to other, similar compounds – such as natural substances in the tissue – or metabolites.) A high incidence of false positives will limit the cost-effectiveness of the screening method because unnecessary and often costly confirmatory tests will need to be performed. In practice, screening methods often give only qualitative (i.e. positive/ negative) or semi-quantitative (i.e. high/medium/low/negative) results. They are sometimes sensitive to more than one compound in a related group, giving either separate indications for each residue or a total for the group. 6.5.3 Confirmatory analyses Confirmatory techniques must conform to more stringent criteria than screening techniques and have a: • • • • •

low limit of detection high level of accuracy high level of precision high level of specificity to the drug in question high level of reproducibility and repeatability.

It is important that there is a low level of doubt about the identity of the compound being measured and a high level of certainty that the quantity determined is a true reflection of the amount actually present. To achieve this, confirmatory techniques usually employ complex separation procedures to isolate the compound of interest and require calibration procedures that involve adding known amounts of the compound to uncontaminated specimens of the material under analysis. As a final check on the validity of the method it is recommended that the method should be evaluated in a collaborative trial in a minimum of three independent laboratories against a standard reference material.

136

Food chemical safety

Standard reference materials are tissue specimens to which the compound has been added or result from slaughtering animals treated with the drug and sampling the same tissues as those being analysed. Such reference materials will have been analysed with fully validated methods in different laboratories, to produce a consensus estimate of the concentration of the drug. The value obtained from an analysis of these samples can be used to confirm the validity of results achieved in an individual laboratory. The EU has commissioned a programme for the preparation of certified standard reference materials for use in veterinary drug residue analysis (European Commission Bureau of Community Reference Materials – BCR). In addition, where possible, laboratories undertaking residues analysis should demonstrate their competence by participating in proficiency testing programmes. 6.5.4 Selection of tissues for analysis Selection of a method for determining given veterinary drug residues requires some knowledge of the metabolism of the drug and the resulting toxicological load. It is unlikely that all metabolites present in the tissues will be equally toxic and some may have longer biological half-lives than others. In addition, the most toxic metabolite may be transitory and present only in small quantities. The analytical method selected must be able to detect the residue in whatever form it is most prevalent and/or toxic. For example, carbadox (Fig. 6.15) is rapidly metabolised via desoxycarbadox to quinoxaline carboxylic acid (QCA); thus surveillance for residues of carbadox is achieved by analysis for QCA. The analyst should also be alert to the fact that some drug residues become bound to the tissues of the animal and that this may have the effect of deactivating the toxic potential of the residue. However, determining the total concentration of the parent compound and/or its major metabolites generally ensures the greatest margin of safety. A knowledge of drug metabolism will also give an indication of which tissues to target for analysis. In addition, tissues do not accumulate residues to the same degree, for example the role of the liver and the kidney in metabolism and elimination of a drug respectively means that these tissues frequently have the highest residue concentrations and are thus often selected for analysis. However, bile has also been used for monitoring residues of anabolic hormones for regulatory purposes, because the bile duct is the major excretory route for hormones that tend to concentrate in the liver. Any detectable residue of synthetic hormones in bile is evidence of a breach of the prohibition of the use of such substances. The picture for natural (endogenous) hormones is less clear as they will always be present and administered natural hormones tend to have no effect on the concentrations that lie within the normal range. There are indications that determining a range of hormones to give a ‘hormone profile’ may provide an indication of natural hormone administration. This is because the administration of hormones may have an effect on the animal’s secretion of endogenous hormones by influencing natural feedback mechanisms.

Veterinary drug residues 137 6.5.5 Analytical methods used to detect residues Microbial inhibition tests are widely employed in screening for antimicrobial residues. Such tests detect the antimicrobial activity of residues present in a tissue sample. The Four-Plate Test (FPT), also known as the Frontier Post Test, is typical. In the FPT, discs of tissue are placed on four agar plates inoculated with microorganisms. The plates are incubated under different conditions to allow inhibition of growth by a variety of antimicrobial drugs. A positive result is indicated by complete inhibition of growth on the surface of the medium in a zone not less than 2 mm wide around the tissue disc. If a positive is confirmed then the identity of the residue must be established by a second technique. The FPT is relevant for residues only where the toxic risk to the consumer is related to the antimicrobial activity. The presence of toxic metabolites or bound residues which have no antimicrobial activity will not be detected. This could present a particular problem with residues such as the penicillin metabolite penicilloyl-protein complex which can cause allergic reactions but lacks any antibacterial activity. The FPT also generates false positive results which are thought to be due to non-specific inhibitory effects of natural constituents present in the tissue. Another shortcoming of the FPT is its sensitivity towards some of the antimicrobial agents that are used in foodproducing animals. Despite this the FPT has some value in indicating the degree of compliance with recommended withdrawal periods for antimicrobial agents, and has limited value in assessing health risks from residues of such agents. Because of the deficiencies of the FPT, the AGVR places great emphasis on the development of quantitative chemical methods for the analysis of residues of antimicrobial residues and this is reflected in the priority the VMD has attached to research to develop improved methods of analysis. The chemical methods developed for drug residues usually require the homogenisation of the tissue followed by an often laborious extraction procedure for the compound to be determined. The constituents of the sample extract are usually separated by chromatographic techniques with high-performance liquid chromatography (HPLC) the most commonly used technique. The detection methods used are usually based on ultraviolet and fluorescence spectroscopy or electrochemical detection, although mass spectrometry (MS) is being increasingly used as it also provides a unique fingerprint of the chemical composition of the residue present in the tissue. Chromatographic analyses are often able to detect a range of chemical-related drug residues. This is particularly the case with ‘families’ such as the sulphonamides and tetracyclines. Such multi-residue techniques have obvious cost advantages and are increasingly employed as screening tests wherever possible. Immunoassay techniques are also used widely to determine residues of antimicrobial agents. Indeed the development of rapid dipstick tests offers the potential for the tests to be used to monitor for the presence of drug residues in liquid media, such as bile, serum or urine, on site and may be used for regulatory purposes at slaughterhouses. Residues of hormonal growth promoters in tissues are usually present at very low concentrations and many chromatographic methods are too insensitive.

138

Food chemical safety

Although thin-layer chromatography has been used to detect diethylstiboestrol and trenbolone in fluids, enzyme linked immunoassays (ELISAs) are the most widely used screening methods. For other drugs, where the concentration of residues is usually higher, chromatographic techniques are more widely used. Despite the many shortcomings of this technique, there is still much interest in the development of ELISA and immunoaffinity columns for the detection of anabolic agents, anthelmintic drugs and other medicines. However, considerable resources must be invested in developing such techniques and it is likely to be some time before they become widely available for drug residue analysis. 6.5.6 Sample pre-treatment The most time-consuming step in assays of tissues for drug residues, either by chemical analysis or immunoassay, is the preparation of a relatively pure extract that is suitable for analysis. The reason is that many natural constituents of tissues can interfere with the very sensitive analytical techniques employed for the end point determination of the residue. The introduction of immunoaffinity clean-up columns has been a major advance in this area. The basis of this technique is that impure aqueous solution of the tissue extract is passed through a chromatographic column containing immobilised antibodies to the drug residues to be analysed. The drug residues bind to the antibodies and all other materials are rinsed away. The drug residues are then desorbed using a different solvent and analysed by using one of the methods described above. The column can be reused many times and the process is suitable for automation. Immunoaffinity columns can be made to select a range of different or related compounds, which can in some cases be desorbed sequentially by altering the solvent.

6.6 Results of surveillance for veterinary drug residues in the UK (1998) 6.6.1 National statutory surveillance (GB) Red meat Under the UK statutory surveillance (VMD Annual Report 1999), during 1998 some 25,028 samples were tested for residues of veterinary drugs and other contaminants (pesticides, heavy metals etc). Of these, 24,998 (99.9%) were free of detectable residues, 17 (0.07%) contained detectable residues below the MRL for the compound and 13 (0.05%) contained residues above the MRL or ‘action level’. There was no evidence to suggest that either synthetic or natural hormones or -agonists were being administered illegally as growth promoters in live animals. The results by group of compounds are given below. Antimicrobials Out of 7,962 cattle, sheep and pig kidney samples analysed for the presence of antimicrobial compounds 7,960 were free of residues. The two positive samples

Veterinary drug residues 139 were from pigs and contained residue concentrations above (1,820 g/kg) the MRL and one below (300 g/kg) the MRL. Sulphonamides Kidney samples from cattle (125), sheep (128) and pigs (940) were analysed for the presence of this group of veterinary drugs. No residues were detected in cattle or sheep kidney. Of the 940 samples of pig kidneys tested 20 samples (2.1%) were found to have detectable residues with residues in four (0.4%) samples exceeding (1,700 g/kg, 230 g/kg, 230 g/kg, and 180 g/kg the MRL of 100 g/kg). Dimetridazole Under Annex IV of EU Council Regulation 2377/90 no MRL can be assigned to this compound and as such marketing authorisation for this compound has been withdrawn. However, under Directive 70/524EEC the compound is authorised for use only in turkeys and guinea fowl. No residues of this compound were detected in the 125 pig kidney samples analysed. However, dimetridazole was detected in pig feed samples at concentrations of 110 g/kg, 130 g/kg, 1,300 g/kg and 2,200 g/kg. Benzimidazoles Liver samples from cattle (292) and pigs (394) were found to be free of any detectable residues of these compounds. Of the 529 sheep liver samples analysed only one was found to have detectable residues (3,500 g/kg, albendazole) which are significantly above the MRL for this compound (1000 g/kg). Ionophores Out of 329 sheep liver samples tested only two were found to contain residues of monensin sodium (21 g/kg, and 40 g/kg). No ADI or MRL has been set for this compound, but informed scientific opinion suggests that the residues present were about 300 times lower than the reported no-effect level observed in rat studies and would therefore not represent a significant hazard to the consumer. This compound was listed under Annex I of the EU Feed Additives Directive 70/ 524 and permitted for use as a feed additive until the recent amendment to the Directive under which such products are no longer licensed as veterinary medicines. Poultry meat In the same survey a total of 8,155 samples of poultry meat were tested for residues of veterinary drugs and other chemical contaminants. Of these 8,055 (98.8%) contained no detectable residues. Residues below the MRL or action level were detected in four (0.05%) samples and above the MRL or action level in 96 (1.2%).

140

Food chemical safety

Antimicrobials No residues of antimicrobial compounds were detected in kidney samples obtained from 1,109 broilers and 71 ducks. However, four turkey kidney samples out of 299 sampled and one laying hen sampled of 110 were found to contain residues of chlortetracycline above the MRL (results: turkey – 610 g/ kg, 700 g/kg, 790 g/kg and 1,470 g/kg; laying hens – 970 g/kg). Dimetridazole This compound is currently licensed for use only in turkeys and guinea fowl as a feed additive and must not be used in laying hens. No residues of this compound were detected in liver samples (784) obtained from broilers, hens, ducks and turkeys. However, residues were found to be present in 25 broiler, three duck and one hen feed sample. In the future it is unlikely that this compound will continue to be authorised for use as a feed additive in any poultry species. Nicarbazin Authorised for use under Feed Additive Directive 70/524, nicarbazin may be used on its own at a level of 100–125mg/kg for broilers (<4 weeks age) or in combination with narasin (50%) at a combined level of 80–100mg/kg up to slaughter (~ 6 weeks). Poultry liver samples (249) were tested for the presence of this compound and 58 (23.3%) found to contain residues at concentrations ranging from 100 to 7,200 g/kg. Using the WHO, Codex-CCRVDF, MRL for this compound which is 200 g/kg, the number of positive samples reduces to 44. Against the ADI for this compound of 400 g/kg body weight, equivalent to 24,000 g/person/day, the single dose contained in 100 g of liver (based on the standard WHO diet) would be 720 g which is well below the ADI. Investigation of these results revealed that improved feed production practice and transport would significantly improve matters. Ionophores One liver sample from 252 samples of poultry liver that were tested was found to contain residues of monensin (5 g/kg). Three samples from 237 samples of poultry (broiler) liver tested for the presence of this compound contained lasalocid at concentrations of 140 g/kg, 63 g/kg and 62 g/kg. In both cases (monensin and lasalocid) the problem appeared to be caused by feed manufacturing, transport and storage procedures. Farmed fish Salmon Residues of veterinary drugs were detected in four samples out of 814 muscle samples analysed, and contained residues of tetracyclines above (290 to 680 g/ kg) the MRL. Subsequent investigation revealed that three of the samples were obtained from fish that were under withdrawal for the drug prior to slaughter and should not have been sampled.

Veterinary drug residues 141 Trout Muscle samples from 163 trout were found to contain no detectable residues of veterinary drugs. However, malachite green was detected at 8 g/kg in one sample. This compound is not licensed for medicinal purposes in fish and has historically been used by the fish farming industry for other purposes. Eggs A total of 512 egg samples were collected from battery, free range and perchery egg production systems. Each sample is comprised of 12 eggs on which a total of 1,212 analyses were carried out. No residues were detected in 499 (97.5%) of the samples, but 13 (2.5%) of the samples contained concentrations of residues above the action level. Dimetridazole Two samples contained residues of this drug at concentrations of 8 g/kg and 77 g/kg. The presence of these residues are of concern as this compound is not licensed for use in laying hens (see above). Nicarbazin The use of this compound is restricted to broilers and over the last three years the incidence in eggs has dropped 10.7% (1996) to 4.0% in 1997. Residues were detected in seven samples, but the concentrations present were not considered by toxicologists to represent a hazard for the consumer. Again, it appears that the source of these residues is the feed. This issue is currently being addressed by the industry and the licensing authorities. Lasalocid This compound is licensed for use in broilers and replacement laying hens up to 16 weeks of age as a feed additive in granular form as problems of crosscontamination of feed batches had been encountered when the compound had been used as a powder. This has resulted in a reduction in the proportion of samples testing positive dropping from 10.7% (1994) to 1.1% (1998) – two samples from a total of 175. Milk A total of 1,007 milk samples were collected, of which 1,005 (99.8%) were free of detectable residues. These samples were subjected to 2,134 analyses with 2,128 containing no detectable residues. Two samples contained penicillin G (8 g/kg and 24 g/kg) above the MRL. 6.6.2 Non-statutory surveillance (GB) During 1998 a total of 11,863 analyses were carried out in the UK on 1,965 samples. In 11,856 (99.9%) of these samples no residues above the action level

142

Food chemical safety

were detected (compared with 99.5% in 1997). Of all the samples only three contained residues, and in two samples more than one compound was detected. The sampling plan included samples that are based on popular preparations of meat and animal products. Ham One ham sample contained residues of the hydroxy- metabolite which is common to dimetridazole and ronidazole. Both these products are listed in Annex IV of Council Regulation 2377/90, i.e. compounds for which no MRL can be set and for which marketing authorisation has been withdrawn. Tiger prawns Of the 40 tiger prawn samples that were tested, multiple residues were detected in two imported samples (from outside the EU). One sample contained oxolinic acid (68 g/kg), sulphamethazine (160 g/kg), oxytetracycline (1,270 g/kg) and chlortetracycline (140 g/kg). Another sample was found to contain oxolinic acid (54 g/kg), oxytetracycline (170 g/kg) and chlortetracycline (63 g/kg). In neither case was it considered that these residues would represent a significant hazard to the consumer. 6.6.3

National statutory and non-statutory surveillance (Northern Ireland – 1998) A proportion of the UK sampling is allocated to Northern Ireland (NI). However, as the region traditionally exports a large proportion of its meat and animal products (~80%), the authorities operate an intensive and targeted surveillance programme of which ~66% of samples are targeted. The targeted sampling includes intensive sampling for residues of antimicrobial compounds, based on veterinary inspections and intelligence regarding possible illegal use of veterinary drugs. In 1998, 3,678 samples were obtained from slaughterhouses (cattle, 1,171; pigs, 587; sheep, 242; poultry, 608); in addition feed samples (137) together with blood serum (185) and urine (285) were obtained from farms. Also in accordance with the revised EU Council Directive 96/23/EC fresh milk (210), eggs (150) and fish (5) samples were analysed. The result of this sampling and analysis programme was that 99.1% of samples were found not to contain detectable residues of veterinary drugs. In total 33 samples contained residues in excess of the MRL or action level. Cattle The three residue violations were in one animal which tested positive for progesterone (0.9 g/kg), and two that tested positive for clenbuterol ( agonist). In the latter case the residue was detected in the retina and liver of the animals.

Veterinary drug residues 143 Pigs Out of a total of 22 positive samples, 19 bile samples from the 56 that were analysed contained residues of 17 -19Nortestosterone ranging in concentration from 8.8 g/l to 125 g/l. These concentrations are above the action level of 2 g/l, but as this steroid is endogenous in the male pig it was considered that the presence of this compound was not due to illegal use. Two kidney samples contained residues of sulphadiazine (310 g/kg and 450 g/kg) in excess of the MRL. One further kidney sample contained chlortetracycline (1,050 g/kg) in excess of the MRL. Poultry Coccidiostats A total of two poultry liver samples tested positive for the presence of monensin (2.9 g/kg and 5.9 g/kg) and two tested positive for nicarbazin (322.3 g/kg and 3,693 g/kg). Dimetridazole (500 g/kg) was also detected in one poultry feed sample. Eggs Residues of the coccidiostat lasalocid were also detected in three samples (9.7 g/kg, 26.0 g/kg and 29.2 g/kg). Fish No detectable residues of veterinary drugs were found in any fish samples. Milk None of the 210 samples analysed was found to contain any residues of veterinary drugs.

6.7

Potential effects on human health of veterinary drug residues in food

The most recently reported UK results on surveillance for veterinary drug residues in meat and animal products show that traces of these compounds can, and sometimes do, arise in food. As all of these compounds are biologically potent in order to be effective in use, it is necessary to ensure that any residual activity in a food product does not present a risk to the consumer. The use of veterinary medicines inevitably leads to the presence of trace residues in food and the purpose of toxicological safety evaluation is to determine at what concentration the residues of a particular compound becomes a cause for concern with regard to human health. Thus, doseresponse relationships have to be established and used to determine the concentration of a drug at which the risks to human health become acceptable and are outweighed by the benefits from the use of the drug. This is in essence the process involved in the setting of Acceptable Daily Intakes (ADIs) and

144

Food chemical safety

recommending withdrawal periods described earlier. Detailed discussion of toxicity and risk-benefit analysis are topics that are beyond the scope of this chapter. The presence of residues tends to be isolated to relatively infrequent occurrences and the risk they represent when present in food may be primarily limited to a very small minority of susceptible individuals. However, despite this and to protect the health of all consumers, it is the duty of the authorities and industry to monitor the safety of the food supply by ensuring that adequate surveillance is conducted and controls enforced.

6.8

Current issues relating to residues of veterinary drugs in food in the UK

6.8.1 Introduction Over a period of about twenty years the results of surveillance for veterinary drug residues in meat and animal products in the UK have provided reassurance that residues of veterinary drugs occur at very low concentrations and generally at low frequencies. Over this period repeated surveillance and improved communication and awareness of the producer and feed industry has resulted in a downward trend in the level and incidence of residues detected. However, contamination ‘hot spots’ remain (e.g. animal medicated feed) and will continue to be monitored in the future. In addition, as new products are introduced into the market they will be evaluated during the licensing process to establish any risk from their residues present in human food. The UK also imports a significant amount of animal-derived food and the presence and risk to the UK consumer of residues in these products must also be considered and appropriate measures taken to protect the UK consumer. 6.8.1 Antimicrobial residues Microbiological inhibition tests are generally considered to be an unreliable method for detecting residues of antimicrobial agents in meat. The development of more specific, sensitive, reproducible, repeatable and robust chemical-based methods of analysis will offer methods capable of detecting residues previously undetectable by biological assays. Particular effort has been devoted to controlling the incidence of residues of sulphonamide drugs in pig meat. In the surveys that have been conducted, the incidence and concentration of detectable residues detected have shown a continued decline. The application of in-feed medication to control disease in pigs (e.g. sulphonamides to control respiratory infections) presents problems for the producer. In particular, the practical application of withdrawal periods is difficult to administer in large pig units where all the animals are receiving their feed through a common, mechanised feeding system. The drug withdrawal period should allow sufficient time for residues of the drug and its metabolites to

Veterinary drug residues 145 be eliminated from the the animal’s body. Despite this it is widely believed that the presence of drug residues in pig meat reflects a failure of producers to observe withdrawal periods. However, there remains strong evidence that the feed compounding industry can play a leading role in reducing the incidence of violations due to excess feed incorporation rates and cross-contamination. It is encouraging to note that major food retailers are exerting their influence on the producers and are increasingly stipulating the animal production, husbandry, health and welfare standards that are used to rear pigs and other food-producing species. 6.8.2 Anabolic hormone residues and -agonists In 1986 (1 December) the use of anabolic hormones for growth promotion purposes in farm animals was prohibited in the UK. However, there is considerable evidence indicating that illegal use of these compounds continues in some EU member states. The main hormones used are the natural (endogenous) hormones and -agonists (clenbuterol). Hormones are usually administered by injection in oily suspensions of hormone ‘cocktails’ designed to have the slow-release characteristics of the previously licensed implants. Such injections are not confined to tissues such as the ears, which are normally discarded at slaughter and thus the uncontrolled and illegal use of anabolic hormones may result in injection sites containing very high concentrations of hormones entering the food supply. There is a thriving ‘black market’ for such preparations and a continued temptation for the majority of scrupulous farmers to use these illegal products when they believe themselves to be at an economic disadvantage to those that do. The unfortunate consequence of the hormones ban in the EU is that the consumer is arguably at greater risk from hormone residues now than when these products were licensed and properly administered. However, it is reassuring that the results of UK surveillance over the period has found no evidence of the illegal use of hormonal growth promoters. -agonists are usually administered as feed additives and as previously noted there is no evidence of the illegal use of these substances as growth promoters in the UK. There has, however, been considerable evidence indicating the illegal use of these substances, in particular clenbuterol, in the Benelux countries, Spain and the Republic of Ireland. However, these countries are now taking appropriate measures to stop this illegal use. 6.8.3 Veterinary drug residues in imported animal products Surveillance for veterinary drug residues in animal products imported into the UK has indicated that the incidences and concentrations of residues in such food are in general similar to those in animals produced on UK farms. However, some reports of surveillance in other countries suggest that this may not always be the case. This may have been due in part to licensing procedures differing from

146

Food chemical safety

country to country or a reflection of less effective surveillance for residues. The recent establishment of the European Medicines Evaluation Agency (EMEA) has meant that licensing and monitoring procedures have now become harmonised throughout Europe. 6.8.4 Conclusions The majority of farm animals are likely to receive some medication during their lifetime to control a disease or parasitic infection. The veterinarian has a considerable range of chemical substances that are licensed for use in veterinary medicines. Their use is essential for the health and welfare of food-producing animals. Surveillance for residues in the UK, by the Veterinary Medicines Directorate (AGVR), is targeted towards substances that may present a potential hazard to the consumer if present at high enough concentrations in food. The results of recent surveillance have confirmed the results obtained in previous years that residue concentrations of veterinary drugs present in UK meat and animal products are generally low and do not present a significant risk to the consumer. The implementation of legislation in the UK has ensured more effective control over residues than has previously been possible. This will help to deter an isolated number of farmers who do not observe the recommended conditions of use for veterinary products and who are therefore probably responsible for the low number of residue violations recorded. Refinements in analytical methodology, to improve the speed and reliability of residue detection, will help assist the expansion of the UK veterinary drug residues surveillance programme.

6.9

Summary

A large number of different veterinary products are licensed for use on farm animals in the UK and many other countries. These products include antimicrobials, anthelmintics, coccidiostats and other therapeutic and prophylactic agents, or those used for growth promotion purposes. The use of anabolic growth promoters is prohibited throughout the EC. In the UK the licensing of veterinary products is carried out under the guidance of the Veterinary Products Committee. The statutory and non-statutory surveillance programmes are carried out by the Veterinary Medicines Directorate and the responsibility for and co-ordination of these programmes are delegated to the Advisory Group on Veterinary Residues. The WHO, Codex Committee for Residues of Veterinary Drugs in Food (CCRVDF) establishes safety limits for residues of veterinary drugs to facilitate trade between nations. The setting up of the EMEA has increased harmonisation of veterinary product licensing procedure. Surveillance for residues is undertaken in the UK, Continental Europe and other countries throughout the world. As evidence for potential human health

Veterinary drug residues 147 risks emerges, suitably sensitive analytical methods are developed to address these needs. The results of surveillance for veterinary drug residues in the UK clearly demonstrate that meat and other animal products of UK origin are generally very safe. Although particular problems have been encountered in the past with respect to some drug residues (e.g. sulphonamides in pig meat), effective legislation and advice to the industry coupled with adequate controls have seen a dramatic reduction in violations.

6.10

Further reading

The Animals, Meat and Meat Products (Examination for Residues and Maximum Residue Limits) Regulations 1991 (Statutory Instrument [1991] No. 2843), HMSO, London. Chemicals for Animal Health Control. Ed. G.C. Brander, Taylor & Francis (1986). CORDLE, M. K. J. Animal Science, 66, 413–33 (1988). CROSSBY, N. T. (ed.) Determination of veterinary residues in food. Ellis Horwood (1991). Evaluation of certain veterinary drug residues in food. 32nd Report of the Joint FAO/WHO Expert Committee on Food Additives. World Health Organization, Geneva (1988). Food Safety Act 1990. HMSO, London. Food Surveillance Paper No. 22. HMSO, London (1987). Food Surveillance Paper No. 33. HMSO, London (1992). HEITZMAN, R. J. (ed.) Residues in food-producing animals and their products: reference materials and methods. EC, Luxembourg (1992). HEITZMAN, R. J. Veterinary drug residues – residues in food-producing animals and their products: Reference materials and methods. Second edition, ISBN 0-632-03786-5 (1994). The hormone issue 1980–1987. National Office of Animal Health, London (1987). MADDOX, J. G. The Veterinary Record, 122, 161 (1988). RICO, A. G. (ed.) Drug residues in animals. Academic Press, London (1986). Veterinary Medicines Directorate Annual Report on Surveillance for Veterinary Residues in 1998. Veterinary Medicines Directorate 1999. ISBN 09531234-2-1.

7 Inorganic contaminants in food N. Harrison, Food Standards Agency, London

7.1

Introduction

The presence of elements known to have adverse health effects in humans such as lead and arsenic is obviously undesirable in food. Environmental sources are the main contributors to contamination of food with most metals and other elements. Some elements (e.g. arsenic) are present naturally but the major sources of other elements (e.g. lead) in the environment are from pollution from industrial and other human activities. The presence of metals and other elements in food can also be the result of contamination from certain agricultural practices (e.g. cadmium from phosphate fertilisers) or manufacturing processes (e.g. tin in canned foods). Statutory legislation to control the levels of such substances in food has been introduced in the UK and elsewhere. In more recent years, other potentially toxic elements have come into focus. Lead, cadmium and mercury have been the subject of much monitoring of the food chain and other metals, in particular aluminium, are continuing to attract attention. Nitrate and nitrite in food from food additive use is regulated across the European Union, but its presence in food crops has raised concerns. A comprehensive programme of work for the surveillance of the food supply has been established for a long time in the United Kingdom to ensure that contamination of the food supply by inorganic contaminants is kept to a minimum.1 The surveys under this programme have provided a considerable amount of data that are of value for use in estimating the dietary exposure to the various contaminants. In order to estimate dietary exposure, data on concentrations of heavy metals are combined with data on consumption of foodstuffs. Among the methods

Inorganic contaminants in food

149

available for obtaining consumption data, three have been used extensively. These are: the total diet study; the diary study; and the duplicate diet study. The UK Total Diet Study (TDS) relies on nationally representative information about the average food consumption by individual households researched in the UK National Food Survey (based on a survey of approximately 7000 households).2, 3 Typical diets are constructed based on these data. Foodstuffs are purchased from retail outlets, then prepared and cooked in the normal manner. The individual foodstuffs are then usually combined into various groups of similar foods – for example cereals, green vegetables and fish – in the proportions eaten on average by consumers. Population dietary exposures can then be calculated using data from the TDS samples. Analysis of the food groups from various locations around the UK yields figures that, when combined with the consumption data, give the total exposure to the contaminant by the average person for that year. These are compared with similarly derived estimates from previous TDSs to follow trends in dietary exposure of the general UK population to inorganic contaminants. The use of the TDS in this way provides information on the food groups, if any, that are major contributors to dietary exposure. For example, most of the exposure to arsenic and mercury is from the fish group, whereas lead is essentially an ubiquitous contaminant. 7.1.1 Population exposure estimates The quantities of foods that make up the TDS and the relative proportions of each food are largely based on consumption data from the National Food Survey and are updated annually. Multiplying the amounts of foods consumed (based on consumption data from the appropriate year of the National Food Survey) by the corresponding mean concentrations of the various inorganic contaminants detected in each TDS food group gives an estimate of population average exposure (covering both adults and children) for that year. These estimates can be used to follow trends in exposure as they take into account changes in both consumption of the various foods making up the general UK diet and concentrations of inorganic contaminants in these foods. 7.1.2 Consumer exposure estimates These are estimates of dietary exposure to inorganic contaminants for individuals who eat average amounts of food (i.e. mean consumers) and those who eat more than average (i.e. upper range (97.5th percentile) consumers) and are based on consumption data from the UK National Adult Dietary Survey (NADS).4 They are calculated using the mean upper bound concentrations of specific contaminants in each food group and the consumption data from the NADS. Consumer exposure estimates are less suitable for following trends in exposure than population estimates as they are based on consumption data from the NADS which was carried out only once in 1986 and 1987 and is not updated

150

Food chemical safety

annually. They do not therefore take into account changes in consumption patterns and only reflect changes in concentrations of the various contaminants. However, consumer exposure estimates do take into account exposures by individuals rather than the population as a whole, and also consider those who consume above average amounts of food (upper range (97.5 percentile) consumers). These estimates therefore provide more accurate assessment of dietary exposure of individual consumers than population exposure estimates for comparison with any safety levels. 7.1.3 Diary studies Diary studies are used to determine in detail the consumption of a particular part of a diet. A population consuming above-average amounts of food that provides the main source of exposure to a contaminant can be identified using questionnaires. A record of the type and weight of food eaten, and the source, is then kept in a purpose-made diary by participants in the study. Representative samples of foods eaten are then analysed and the data combined. An extension of this approach is exemplified by the duplicate diet study, in which as exact a replicate as possible of all food consumed is collected for analysis. 7.1.4 Assessing the risks The risk to health from chemicals in food can be assessed by comparing estimates of dietary exposure with recommended safe levels of exposure. For most metals and other elements, these are the Provisional Tolerable Weekly Intakes (PTWIs) and the Provisional Tolerable Daily Intakes (PTDIs) recommended by the Joint Expert Committee on Food Additives of the Food and Agricultural Organisation of the United Nations and the World Health Organisation International Programme on Chemical Safety (JECFA). The European Commission’s Scientific Committee on Food has established other relevant safe levels. These are Acceptable Daily Intakes (ADIs) for chemicals added to food, and Tolerable Daily Intakes (TDIs) for chemical contaminants. The use of the term ‘tolerable’ implies permissibility rather than acceptability. All the above recommendations are estimates of the amount of substance that can be ingested over a lifetime without appreciable risk, expressed on a daily or weekly basis as appropriate.

7.2

Metals and metalloids

7.2.1 Lead Lead exists in small quantities in the Earth’s crust, but is well known as it is extracted easily from its ores. It is a grey, ductile, malleable metal which has been used extensively by man from earliest times, with references to it being found in Egyptian hieroglyphics of around 1500 BC. Amongst other uses, it was

Inorganic contaminants in food

151

known to have been employed by the Romans for lining aqueducts and water mains. The possible hazards arising from contamination of food have long been recognised, lead poisoning being well known as ‘saturnism’ to the early Romans. Lead still has a number of important uses in the present day; from sheets for roofing to screens for X-rays and radioactive emissions. The main sources of lead in food today are the past history of lead technology, including mining and the use of alkyl-lead compounds as petrol additives. Like many other inorganic contaminants, lead is ubiquitous and can be found naturally occurring in many different foods as metallic lead, inorganic ions and salts. Lead has no essential function in man, but has a number of adverse effects and young children and the developing foetus are considered to be at most risk from its toxic effects.5 The exposure to lead in food by the general population in the UK is well within international tolerable limits. Results from the TDS indicate that during the period 1976 to 1997 the dietary exposures over the whole population fell from 0.11 to 0.026 mg per person/day (Table 7.1).6 This excludes any contribution from drinking water, which is likely to be higher in areas with soft water. The PTWI recommended by JECFA is 0.025 mg/kg bodyweight, equivalent to 0.21 mg/day for a 60 kg adult. The dietary exposures for mean and 97.5th percentile consumers in 1997 were 0.024 mg/day and 0.043 mg/day respectively, well below the PTWI. These dietary exposures to lead in the UK are similar to those in Canada (0.024 mg/day),7 The Netherlands (0.01 to 0.032 mg/day),8 and the USA (0.015 mg/day).9 Although dietary lead intakes in the UK are currently well within recommended intakes, it is the UK Government’s policy to ensure that exposure to lead is reduced wherever practicable and, more specifically, to reduce blood lead levels in children to below 10 g/dl. Food is one of the major sources of lead exposure in the UK; the others are air (mainly lead dust originating from petrol) and drinking water. Exposure from all of these sources has been reduced, as demonstrated by the reduction in blood levels over the past 15 years.10 The decrease in dietary exposure reflects the success of the measures taken to reduce lead exposure and contamination of food, such as the use of lead-free petrol, welded food cans, and the banning of tin-coated lead capsules for wine bottles. Lead concentrations in most foods in the 1997 UK TDS were generally low,10 with all but a few samples containing lead at levels below the limits defined in the Lead in Food Regulations 1979, as amended.11 This includes a general limit of 1 mg/kg and limits for specific foods as defined in Schedules to the Regulations. Table 7.1

Population dietary exposures to lead from UK Total Diet Studies (mg/day)

1976

1979

1982

1985

1988

1991

1994

1997

0.11

0.09

0.069

0.066

0.06

0.028

0.024

0.026

152

Food chemical safety

Duplicate diet studies of people living in areas where exposure to lead from other sources is known to be low have provided some valuable information. The mean dietary lead exposure to women was estimated to be 0.31 mg/week and the mean lead exposure to children as 0.11 mg/week.12 These studies included the contribution from drinking water. In areas with elevated levels of lead in tap water, estimated lead intakes of both adults and children are found to be higher and, in a small percentage of cases, above the PTWI. For regular consumers of alcoholic beverages, wine and beer may make a significant contribution to dietary lead intake and this is reflected in the higher blood lead levels found for drinkers. Lead in beer is also absorbed more readily from the gastrointestinal tract than lead from the rest of the diet. A survey of lead in wines and beers from lead-capped bottles in 1982/1983 showed that about 90% of canned and bottled beers sampled contained < 10 g/l and that nearly half the draught beers sampled contained > 10 g/l; 4% contained > 100 g/l. All wines sampled directly from the bottle contained < 250 g/l.12 The contribution from deposition of airborne lead on soil and crops to lead in diets is estimated to be between 13% and 31% for children. For individual plants a high percentage of lead may derive from aerial deposition (40–100%). Where crops are contaminated by lead from the air and soil, much of this may be removed by washing and other normal culinary practices. 7.2.2 Cadmium The use of cadmium by man is relatively recent and it is only with its increasing technological use in the last few decades that serious consideration has been given to cadmium as a possible food contaminant. Cadmium is naturally present in the environment: in soils, sediments and even in unpolluted seawater. It is closely related chemically to zinc and is found wherever zinc is also found. Thus most commercially available zinc compounds will contain cadmium at low levels. Cadmium and its compounds have been used widely in industry with concomitant environmental pollution. Cadmium is emitted to air by mines, metal smelters and industries using cadmium compounds for alloys, batteries, pigments and in plastics, although many countries have stringent controls in place on such emissions. Within the European Union, as part of a Community Action Programme to combat environmental pollution by cadmium, Directive 89/677/EEC came into force in 1992.13 This banned the use of cadmium in pigments, stabilisers and plating, except in cases where no suitable alternative is available or where cadmium is used for safety reasons such as in the nuclear industry. The results of studies on animals show that cadmium is an extremely toxic metal. Cadmium is poorly excreted by the human body and although only 5– 10% of that ingested is absorbed, it does accumulate in the body over time with renal damage being caused by long-term exposure.14 One sign of this damage is proteinuria (the appearance of increased levels of unaltered proteins in the

Inorganic contaminants in food

153

urine). Cadmium and its compounds can also give rise to carcinogenic effects in animals, although the evidence for similar effects in humans is not conclusive. As cadmium is ubiquitous in the environment, all food is exposed to and contains cadmium. The principal route of absorption for non-industrially exposed populations is from food, although cigarette smoking can be an important source. The main sources of cadmium contamination of the UK food supply are phosphate fertilisers, atmospheric deposition and sewage sludge. There are currently no UK limits for cadmium in food. However, the existing recommendation of the UK Food Advisory Committee is that food containing levels of cadmium not acceptable in its country of origin should not be admitted to the UK. Cadmium is present at low concentrations in most foods, with those that are consumed in larger quantities making the largest contribution to the population dietary exposure. Although the concentrations of cadmium in food in the UK are generally low, some foods of minor dietary importance such as shellfish or kidney often contain levels in excess of 0.5 mg/kg. Plant-based foods rarely contain more than 0.2 mg/kg on a fresh weight basis, although some root crops such as carrots and parsnip, and some leafy crops, such as spinach and lettuce, tend to contain more cadmium than other plant foods. This is also true of cereals, which indicates that plants tend to take up cadmium from the soil, unlike lead. Most fish taken around the UK coast contain little cadmium, with the average being less than 0.2 mg/kg. The results of a routine surveillance exercise are given in Table 7.2.15 The 144 samples analysed were collected between 1995 and 1997. It was found that marine fish had concentrations of less than the LOD of 0.01 mg/kg, while shellfish contained higher concentrations than do most other foods, but, with the exception of lobster, whelks and crabs, shellfish from unpolluted waters rarely have an average cadmium concentration greater than 1 mg/kg. However, the body meat of crabs (brown meat) may often contain higher concentrations. The level of cadmium in animal tissues other than offal is usually low, average concentrations being less than 0.05 mg/kg. Individual samples of kidneys may contain levels higher than 0.5 mg/kg. Animals grazing on land contaminated with cadmium will often have meat with normal levels of cadmium, while the level in offal would usually be significantly higher.16 Table 7.3 presents the results from the TDS for cadmium between 1976 and 1997. The results show that for the average person in the UK, the dietary exposure to cadmium has remained roughly the same over this period at approximately 0.018–0.02 mg/day, but with a slight downward trend.10 These are upper-bound estimates, that is in the estimation of dietary intake it is assumed that levels in food that are less than the limit of detection are equal to the limit of detection. These estimates therefore overestimate the actual intake to some extent. For mean and 97.5th percentile adult consumers, the dietary exposures from the 1997 UK TDS were 0.014 mg/day and 0.024 mg/day respectively. These exposures to cadmium from food and other sources (e.g. air, water and smoking)

154

Food chemical safety Table 7.2 Cadmium in fish and shellfish (mg/ kg fresh weight of edible portion) Fish

Mean concentration

Haddock Cod Herring Mackerel Plaice Whiting Red fish Cod fish fingers Brown shrimps Cockles Crab Lobster Mussels Pink shrimps Queen scallops Squid Scallops Scampi Winkles

0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.04 0.07 1.5 0.12 0.19 0.05 0.17 0.01 0.54 0.06 0.19

Note: Data from reference 15.

can be compared with the JECFA PTWI of 0.007 mg/kg bodyweight, which is equivalent to 0.06 mg/day for a 60 kg person. Although the margin of safety between the PTWI and the dietary exposure is smaller than with most other metals, it is reassuring that the trend is for lower exposures. The most recent data for dietary exposures to cadmium are similar to those found elsewhere. For example, a study carried out by the European Union in 1995 and 1996 found that dietary exposures in the 15 member States of the Union ranged from 0.007 mg/ day to 0.057 mg/day.17 Other comparable dietary exposures are 0.015 mg/day for the USA,9 0.024 mg/day for Canada,7 and 0.028 mg/day for New Zealand.18 However, the results of duplicate diet studies have shown that there are localised high intakes by consumers in certain areas or by consumers of certain foods. In the old mining village of Shipham in Somerset, UK, where the cadmium levels in some vegetable samples were more than 1 mg/kg, the dietary exposure to cadmium of the study population was about double the average weekly dietary intake and some individuals exceeded the PTWI for cadmium.19 Table 7.3 day)

Population dietary exposures to cadmium from UK Total Diet Studies (mg/

1976

1979

1982

1985

1988

1991

1994

1997

0.02

0.017

0.018

0.018

0.019

0.018

0.014

0.012

Inorganic contaminants in food

155

There was no evidence that any of the residents of the village had suffered adverse health effects related to cadmium. 7.2.3 Arsenic Arsenic is a metalloid, but it is generally included in work on metals in food. It is rarely found as the free element in the natural environment, but more commonly as a component of sulphur-containing ores in which it occurs as metal arsenides. Arsenic is present in rocks, soils, water and living organisms in concentrations in the parts per million (mg/kg) range. Interest in the contamination of food with arsenic in the UK arose from a serious outbreak of arsenical poisoning in northern England in 1900 owing to beer made with glucose. It was established that the outbreak arose chiefly as a result of the use of arsenical pyrites to manufacture sulphuric acid used in production of the glucose. A Royal Commission set up in 1903 to consider the case concluded that maximum levels of arsenic in food and liquids should be established. In the UK, the general limit for food is 1 mg/kg, and 0.1 mg/kg for ready-to-drink beverages.20 However, naturally present arsenic in fish and edible seaweed or products containing fish or edible seaweed is specifically excluded, as are hops or hop concentrates intended for commercial brewing and any food with an arsenic level limited by other regulations. Arsenic is present in food in different forms (species) which vary in toxicity, with inorganic forms considered to be the most toxic. This is reflected in the JECFA PTWI of 0.015 mg/kg bodyweight (equivalent to 0.12 mg/day for a 60 kg adult), which applies to inorganic arsenic only.21 Most arsenic in the diet is present in the less toxic organic forms. However, it can be difficult to distinguish analytically between the different forms of arsenic and for this reason most surveys have measured total arsenic. The weekly dietary exposure to arsenic can be calculated from an analysis of the various food groups of the TDS and the most recent data cover 1976–97.10 The major source of arsenic in the diet is fish which, in 1997, accounted for 94% of the total dietary exposure. An important consideration when assessing the significance of these dietary exposures is the finding that arsenic is almost entirely present in fish as the organic chemical arsenobetaine, which has been found not to be metabolised in man.22, 23 Fish are known to accumulate arsenic and a study carried out in 1998 found appreciable quantities of total arsenic in all samples analysed.15 The mean concentration of arsenic in samples of fresh marine fish landed in UK ports in 1995–1997 ranged between 1.9 mg/kg and 8.4 mg/kg. An earlier survey found that fish that live on or close to the sea bed, such as plaice, dabs, flounders and skate, tend to have higher levels of arsenic than other fish,24 and this was confirmed by the later work, where the highest level was found in plaice. Arsenic levels in shellfish show more variation, ranging from 1.3 mg/kg to 30 mg/kg. High levels are frequently found in crab, in which the white meat generally contains more arsenic than the brown meat. Lobsters contained similar levels of arsenic to crabs, with the highest levels found in pink shrimps.

156

Food chemical safety

Table 7.4

Population dietary exposures to arsenic from UK Total Diet Studies (mg/day)

1976

1979

1982

1985

1988

1991

1994

1997

0.075

n.d.

0.09

n.d.

n.d.

0.07

0.063

0.065

Note: n.d. ˆ not determined

The total arsenic, that is inorganic plus organic, content of seaweed-based dietary supplements has been determined.25 Samples of kelp powder and tablets were analysed in 1987 for both total arsenic and inorganic arsenic (unstable organic species were also included). Levels from 7 to 45 mg/kg of total arsenic were found, with an outlying value of 120 mg/kg. The inorganic arsenic content was between 0.01 to 0.45 mg/kg, with the same outlying sample containing 50 mg/kg. A more recent study on a range of dietary supplements carried out between 1995 to 1998 found total arsenic levels ranging from below the LOD of 0.005 mg/kg to 7.3 mg/kg in a sample of a supplement produced from greenlipped mussels.26 It can be seen from Table 7.4 that the population dietary exposure to total arsenic for UK consumers has remained fairly constant. The dietary exposures for mean and 97.5th percentile adult UK consumers in 1997 were 0.12 mg/day and 0.42 mg/day respectively. Although these exposures are similar or greater than the JECFA Provisional Tolerable Daily Intake (PTDI) for inorganic arsenic of 0.002 mg/kg bodyweight, which is equivalent to 0.12 mg/day for a 60 kg person, these are estimated for total arsenic. As has been stated before, most of the arsenic in the diet is from fish and most of the arsenic in fish is in the less toxic organic forms. 7.2.4 Mercury Mercury is the only metallic element that is liquid at room temperature. It is ubiquitous throughout the environment. It occurs mainly in the form of the ore cinnabar (mercury (II) sulphide), but there are also at least thirty minerals in which the metal is found. Exposure to mercury vapour over periods of months or years can result in chronic poisoning. Exposure to inorganic mercury compounds can result in bleeding from the gastrointestinal tract and kidney damage, followed by death from uraemia. Organomercury compounds have a wide variety of effects, and range in value from therapeutic agents to lethal chemicals. Of the organomercury compounds, the alkylmercurials are considered to be the most toxic. There have been a number of incidents involving alkylmercurials, of which methylmercury is the most intensively studied as it is thought to be the main agent of Minimata disease. In this incident, children of consumers of fish from Minimata Bay in Japan suffered from damage to the central nervous system, resulting in disturbances in vision, hearing, muscle function and mentality.27 It was later determined that there had been discharges of methylmercury in effluent from a chemical plant to the bay.

Inorganic contaminants in food

157

The main sources of exposure to mercury for the general population are from the diet and dental amalgam.21, 28 The main dietary source of mercury is fish and this has led to interest in potential exposure to mercury on the neurological development of children from populations with high fish consumption.29 There is no statutory control of mercury in most food in the UK, but the levels of mercury in fish are controlled by European Commission Decision 93/351/EEC which sets an average limit for mercury in fish of 0.5 mg/kg or 0.5 ppm (part per million).30 This average limit is, however, increased to 1.0 mg/kg or 1 ppm for the edible parts of the predatory and bottom-dwelling species listed in the Annex to the Decision. Investigations into the presence of mercury in food have been carried out in a comparatively small number of foods in the UK since 1966. Mercury is included in the analytes measured in the samples of the Total Diet Study. The estimated dietary exposures to total mercury (organic and inorganic) for the general population as determined from the UK TDS (Table 7.5) have remained fairly constant between 0.002 and 0.005 mg/day. The dietary exposures of mean and 97.5th percentile consumers in the UK in 1997 were 0.0031 mg/day and 0.0064mg/day.10 These may be compared with the JECFA PTWI for mercury of 0.005 mg/kg bodyweight/week (of which no more than two-thirds should be methyl mercury),31 which is equivalent to 0.043 mg/day for a 60 kg adult. The dietary exposures are similar to those in the USA (0.008 mg/day)8 and the Netherlands (0.002 mg/day)7 but lower than New Zealand (0.013 mg/day).18 The highest concentrations of mercury are found in the fish group of the TDS, which contributes about 33% to the overall dietary exposure. Higher than usual concentrations of mercury can sometimes be found in the cereals and the meat groups, which indicates that contamination of these foods can occur. Although the use of organomercury compounds as seed dressings is now no longer permitted in the UK, it has been known for treated grain to be milled for human consumption in the past and there have been incidents in Iraq, Pakistan and Guatemala. A survey by the Working Party on the Monitoring of Foodstuffs for Heavy Metals in 1984 found that of 17 samples of the bread and cereals group analysed, none contained mercury in excess of 0.001 mg/kg.17 A hundred samples of grain analysed at the same time contained concentrations of less than 0.005 mg/kg in most cases, although four samples contained between 0.013 to 0.017 mg/kg total mercury. It is not known for sure why these levels were found,

Table 7.5 day)

Population dietary exposures to mercury from UK Total Diet Studies (mg/

1976

1979

1982

1985

1988

1991

1994

1997

0.005

0.004

0.003

n.d.

n.d.

0.002

0.004

0.003

Note: n.d. ˆ not determined.

158

Food chemical safety

but similar concentrations have been known to accumulate in grain due to atmospheric contamination. Mercury has been shown to accumulate in eggs when organomercury compounds are present in cereal used as feed. Egg white normally contains higher levels of total and organic mercury than the yolk owing to an association of mercury with the protein ovalbumin. In 1984, 75 samples of retail chicken eggs were analysed for total mercury and inorganic mercury.17 The mean mercury concentration was about 0.004 mg/kg (range < 0.0005 to 0.029 mg/kg). An extreme consumption of eggs would be about 500 g/week (about one-and-ahalf eggs a day) and taking the mean level results in a mercury exposure to 0.01 mg/week. Mercury is normally present in fish as it is ubiquitous in the environment. Fish and shellfish caught off the coasts of England, Wales and Scotland have been regularly monitored for mercury since 1973; the latest data are for 1996.32 These data are used to meet UK obligations in the European Community and to the Oslo and Paris Conventions. Examples of these data are those collected for Liverpool Bay, into which the Mersey flows. The Mersey is subject to discharges (from chlor-alkali plants) some of which can contain mercury. Data are submitted annually to the EC to indicate compliance with the Environmental Quality Standard (EQS) of 0.3 mg/kg mercury wet weight in fish flesh. Between 1977 and 1984 there was a significant reduction in the amount of mercury discharged into the marine environment: the data show a 52% reduction in mean mercury concentrations in all species of fish in the Liverpool Bay area. This decrease has been continued in recent years with the mean concentration decreasing from 0.23 mg/kg to 0.14 mg/kg between 1985 and 1994. Thus, it is apparent that, overall, levels of mercury in fish in Liverpool Bay are decreasing and easily comply with the EQS. As it is known that fish is a major source of mercury in the diet, extensive surveillance work has been carried out. Predatory species such as tuna and swordfish tend to accumulate relatively high levels. During 1973–74, a survey conducted by MAFF determined that mean mercury concentrations in imported fish were in general less than 0.1 mg/kg, although halibut and tuna were exceptions with mean concentrations of 0.16 mg/kg and 0.12 mg/kg respectively.27 Most of the species analysed had concentrations of less than 0.05 mg/kg. A further survey of frozen, tinned or dried imported fish in 1983 showed that with the exception of tuna and one sample of sardines, levels were all below 0.05 mg/kg, confirming the earlier findings. The mean level of tuna (tinned) was 0.12 mg/kg (range: 0.04 mg/kg to 0.22 mg/kg). Although this level was higher than those for the other species of fish, it is similar to the concentration found in the earlier survey indicating no large change in mercury concentrations. A recent survey on marine fish landed in UK ports found a range of 0.016 mg/kg to 0.14 mg/kg.15 Similarly, levels in shellfish ranged between 0.01 mg/kg and 0.29 mg/kg.

Inorganic contaminants in food

159

Table 7.6

Population dietary exposures to tin from UK Total Diet Studies (mg/day)

1976

1979

1982

1985

1988

1991

1994

1997

4.4

3.2

3.1

1.7

n.d.

5.3

2.4

1.8

Note: n.d. ˆ not determined.

7.2.5 Tin Tin is widely distributed in nature at low concentrations in plants, both marine and land, and animals. It has extensive uses in industry, where organotin compounds are used as heat stabilisers in plastics, salts in glazes in porcelain and oxide coatings used to reduce abrasion of glass containers. The main source of tin in the diet is from tin-plated steel used in the manufacture of cans for foods and beverages. High tin concentrations in food may cause short-term acute health effects in some people, including stomach upsets, abdominal cramps, nausea and/or diarrhoea. These short-term effects may occur in some individuals at concentrations above 200 mg/kg.33 The UK Tin in Food Regulations 1992 limit the maximum amount of tin in food sold in the UK to 200 mg/kg.34 Fortunately, concentrations of tin in most foods are well below 10 mg/kg, although canned foods may contain higher concentrations as a result of slow dissolution of the tin coating used on some cans to protect the steel body of the can from corrosion. Tomato-based products tend to have high levels of tin as nitrate in the food accelerates corrosion of the tin. A survey of canned tomato products sold in the UK in 1998/1999 found that 98% of the products tested were below the 200 mg/kg limit.35 The JECFA PTWI for tin is 14 mg/kg bodyweight, equivalent to 120 mg/day for a 60 kg adult. Average dietary exposure has remained in the range 1.7–5.3 mg/day (Table 7.6). The dietary exposure estimates for mean and 97.5th percentile consumers were 1.9 mg/day and 6.3 mg/day respectively, well below the PTWI. 7.2.6 Aluminium Aluminium is the third most abundant element in the earth’s crust and is used widely in the manufacture of construction materials, wiring, packaging materials and cookware. The metal and its compounds are used in the paper, glass and textile industries as well as in food additives. Despite the abundance of the metal, its chemical nature effectively excludes it from normal metabolic processes. This is due largely to the low solubility of aluminium silicates, phosphates and oxides that result in the aluminium being chemically unavailable. However, it can cause toxic effects when there are raised concentrations of aluminium in water used for renal dialysis. These effects are not seen when aluminium is at the concentrations usually present in drinking water. There is currently much activity to examine the factors that influence uptake of aluminium from the diet.

160

Food chemical safety

Table 7.7 day)

Population dietary exposures to aluminium from UK Total Diet Studies (mg/

1976

1979

1982

1985

1988

1991

1994

1997

n.d.

n.d.

n.d.

n.d.

3.9.

10

11

3.4

Note: n.d. ˆ not determined.

In the 1997 TDS, the highest concentrations were found in the bread (6.6 mg/ kg) and fish (6.1 mg/kg) food groups.10 The concentration found in bread may reflect the use of permitted aluminium-containing additives. The largest contributions to dietary exposure were from beverages, bread and cereal products; beverages owing to their high consumption and cereal products owing to the use again of aluminium additives. Aluminium has been measured in a range of crops.36 A large variation in levels was apparent for the various species. The highest concentrations were found in lettuce (6.2–810 mg/kg), parsnips (6.0–82 mg/kg) and Brussels tops (7.7–116 mg/kg). Aluminium has been included in the UK TDS only since 1988, owing to the difficulty of obtaining reliable analytical data in the presence of environmental levels of aluminium (Table 7.7). Dietary exposures for mean and 97.5th percentile for UK consumers in 1997 were 3.2 mg/day and 5.7 mg/day respectively,10 similar to the population exposure. All the estimated exposures were well below the JECFA PTWI of 7 mg/kg bodyweight, which is equivalent to 60 mg/day for a 60 kg adult. The population estimate is similar to dietary exposures reported for other countries. A dietary exposure in the USA of 11.5 mg/day was reported for 14–16-year old males,8 while dietary exposures in an Italian TDS were between 2.3–6.3 mg/day.37 7.2.7 Copper Copper is an essential element for all plants and animals, but can be toxic at high levels of exposure. It is widely distributed and always present in food. Animal offal, which is the major contributor to dietary exposure to copper, various shellfish and nuts contain, on average, more than 20 mg/kg. This was the limit recommended for food by the Food Standards Committee in 1956. Milk contains little copper, usually less than 0.1 mg/kg, and fresh fish and alcoholic drinks contain less than 1 mg/kg of copper. Table 7.8

Population dietary exposures to copper from UK Total Diet Studies (mg/day)

1976

1979

1982

1985

1988

1991

1994

1997

1.8

n.d.

1.3

1.3

n.d.

1.4

1.2

1.2

Note: n.d. ˆ not determined.

Inorganic contaminants in food

161

The population daily exposure to copper from UK Total Diet Study (TDS) samples (Table 7.8) has remained fairly constant, being 1.8 mg/person in 1976 and 1.2 mg/person in 1997.10 The mean and 97.5th percentile exposures were 1.4 mg/ day and 3.2 mg/day respectively. These latter values are well below the JECFA PTDI of 0.5 mg/kg bodyweight, equivalent to 30 mg/day for a 60 kg adult.38 7.2.8 Other elements Most of the surveys on metals and metalloids in food have concentrated on those elements that are known to be toxic, or where there are possible concerns about their levels in food. In the course of collecting the data, information on other metals is often collected in addition. Other metals that have been included in the UK Government’s surveillance are zinc, antimony, chromium, cobalt, indium, nickel, thallium and tin. Zinc As with copper, zinc is an essential element for all plants and animals. It is necessary for the correct function of various enzyme systems. However, excessive intakes of zinc can have long-term effects as they can interfere with absorption of copper and iron in the diet, and may result in anaemia.39 JECFA has recommended a PTDI of 1 mg/kg bodyweight, equivalent to 60 mg/day for a 60 kg adult.38 In food, the major contributors to the diet are meat and its products, from which zinc is readily absorbed. Liver, with concentrations of around 62 mg/kg, contains the highest levels of any meat products, with other tissues having values of a half to a third of this figure. The second greatest source of zinc is cereals. Concentrations of zinc in whole cereal products are similar to those in meat. The mean and high-level dietary exposures to zinc in the UK in 1997 were 11 and 20 mg/day respectively.10 The time trend for the population dietary exposure is given in Table 7.9. Antimony Compounds of antimony are used as fire retardants in plastics and paper, and for veterinary purposes. The metal is found in specialised alloys such as white metal bearings and pewter, which is an alloy of tin, antimony (up to 7.5%) and copper. Concentrations in food are low, generally in the range < 0.01 to 0.08 mg/kg, but have been found to be higher in samples of aspic jelly and cream of tartar.40 The Table 7.9

Population dietary exposures to zinc from UK Total Diet Studies (mg/day)

1976

1979

1982

1985

1988

1991

1994

1997

10

n.d.

10

10

n.d.

10

8.4

8.4

Note: n.d. ˆ not determined.

162

Food chemical safety

estimated population dietary exposure from the UK TDS was estimated to be 0.002 mg/day in 1985. In comparison, the average and mean dietary exposures in the 1994 TDS were both 0.003 mg/day. The corresponding high level (97.5th percentile) dietary exposure estimate was 0.004 mg/day.41 Chromium Chromium is used in the manufacture of stainless steel and other specialist steels, and non-ferrous alloys. Chromate salts are used as tanning agents, pigments, catalysts, corrosion inhibitors and in electroplating solutions. Although stainless steel is inert, the dissolution of chromium from this steel is likely to be the major source of chromium in food. Chromium is an essential element for man. The minimal requirement for man is estimated to be about 1 g/day. As the absorption of inorganic chromium (Cr III) is about 0.5% of a given dose, and the absorption of organically bound chromium is even higher, a dietary exposure to 200 g chromium/day will provide the estimated requirement. The US National Academy of Sciences has recommended a dietary intake of 50 to 200 mg/day for adults.42 From the 1997 UK TDS, the mean and 97.5th percentile consumer dietary exposures were 0.1 mg/day and 0.17 mg/day respectively. There is no JECFA PTWI or PTDI for chromium. Cobalt Cobalt is used in the manufacture of alloys and in nuclear technology. Cobalt compounds are also included in trace element supplement preparations for ruminants. The cobalt concentrations in vegetables and other foods have been found to be between < 0.01 and 0.83 mg/kg, while levels in milk are between 0.0002 and 0.06 mg/kg.40 Cobalt concentrations in dietary supplements ranged between <0.005 mg/kg to 4.1 mg/kg.26 The mean and 97.5th percentile dietary exposures to cobalt for adult consumers in 1994 were 0.11 mg/day and 0.019 mg/day respectively.10 Nickel Nickel salts are widely used in industry for plating and as pigments. Alloys of nickel are used in storage batteries, coins, cooking utensils and other products. Fats and oils are hydrogenated using a nickel catalyst. Nickel has not been shown to be an essential nutrient for humans, but is considered to be a normal constituent of the diet. However, there are adverse health effects associated with nickel, for example long-term contact with jewellery containing nickel can cause dermatitis in sensitive individuals. The highest concentrations of nickel in individual foods occur in tea, soya protein and herbs.40 Other foods with elevated levels are pulses, cocoa products and some nuts. Nickel has been found in crops at levels of <0.01 to 1.5 mg/kg, with the highest concentrations in broad beans and Brussels sprouts. Studies of nickel in marine fish and shellfish have shown that concentrations are low in fish in general, but that levels in shellfish average 0.61 mg/kg, with higher mean

Inorganic contaminants in food Table 7.10

163

Population dietary exposures to nickel (mg/day)

1976

1979

1982

1985

1988

1991

1994

1997

0.33

n.d.

0.15

0.14

n.d.

0.17

0.13

0.13

Note: n.d. ˆ not determined.

values in cockles (4.6 mg/kg), winkles (0.4 mg/kg), mussels (0.4 mg/kg) and crab (0.31 mg/kg).15 Surveys of dietary supplements have found nickel concentrations ranging from < 0.05 mg/kg to 10 mg/kg.26 The 1997 dietary exposure estimates for mean and 97.5th percentile adult consumers were 0.12 mg/day and 0.21 mg/day respectively.10 These can be compared with the World Health Organization (WHO) Tolerable Daily Intake (TDI) of 5 g/kg bodyweight (equivalent to 0.3 mg/day for a 60 kg person). The time trend for population dietary exposures is given in Table 7.10.

7.3

Nitrate and nitrite

Nitrate is a water-soluble ion that occurs naturally in the environment as a consequence of nitrogen fixation and is available to plants through its uptake from soil and subsequently to animals. Plants synthesise all their protein and other organic nitrogen compounds from inorganic nitrogen taken up as nitrate or ammonia.43 Most comes from the soil although in legumes, bacteria in root nodules use atmospheric nitrogen to make nitrate. If extra nitrate is available to the plant it may take it up and, if synthetic processes are unable to use it all, it may be retained. The nitrate content of vegetable products is a result of their cultivation and the fixation of atmospheric nitrogen. Nitrogen is added to the soils in organic forms in crop residues and manures and in synthetic forms (ammonium, nitrate and urea) in fertilisers. The organic nitrogen in soil is slowly mineralised by biological processes, the final stages being to ammonium, nitrate and nitrite.43 The conversion of ammonium to nitrite in soils is rapid, while the conversion of nitrite to nitrate is so rapid that nitrite is rarely detectable in soils. Nitrate is also present in some water supplies but usually at low levels. The nitrate ion is therefore likely to be present in almost everything we eat or drink and humans are likely to have been exposed to this ion throughout their evolution. Most foods contain nitrate and nitrite. This is present naturally or may be present as a result of the use of fertilisers on crops, or from its use as a preservative.44 Both nitrate and nitrite have been used extensively for the traditional preservation or curing of certain meats, for example bacon, hams and some sausages. The use of saltpetre (potassium nitrate) was known to the Sumerians, Greeks, Romans and Chinese more than one thousand years ago. The mechanism of this effect was not understood until 1891, when it was shown that the nitrate ion

164

Food chemical safety

is transformed into the nitrite by microbiological and biochemical processes occurring in meat. Nitrite has a bacteriostatic or growth-inhibiting effect on bacteria, and is particularly effective at inhibiting the spores of Clostridium botulinum and hence preventing the production of its toxin.43 It was not until the early 20th century that it was shown that the red coloration of meat is preserved by the formation of nitrosomyoglobin and nitrosomyochromogen. There have been health concerns about the presence of nitrate in food as it can be metabolised to potentially carcinogenic N-nitroso compounds. For this reason, the European Commission’s Scientific Committee on Food considered the implications for human health of nitrate in food in 1990 and set an Acceptable Daily Intake (ADI) for the nitrate ion of 0–3.65 mg/kg body weight (equivalent to 219 mg/day for a 60 kg person) and an ADI for the nitrite ion of 0–0.07 mg/kg bodyweight.45 These ADIs were reviewed in 1995, when the ADI was confirmed and a new ADI of 0–0.06 mg/kg bodyweight was set.46 7.3.1 Nitrate in food Vegetables, and in particular green leafy vegetables, naturally contain higher concentrations of nitrate than other types of food. As such they make the main contribution to total dietary exposure. A number of factors have been found to influence the levels of nitrate in food crops. These include the season in which the crops are grown, irrigation of the crops, storage, cooking and time of harvesting.47 It can be seen from Table 7.11 that spinach, lettuce and beetroot had significantly higher nitrate contents than the other types of vegetables analysed. Tomatoes contained the lowest concentrations. Seasonal differences in nitrate concentrations were observed for potatoes, lettuce, carrot, cauliflower and onion samples. Nitrate concentrations in potatoes were lowest in the summer period Table 7.11 Vegetable

Beetroot Cabbage Cauliflower Carrot Lettuce Onion Potato Spinach Sprouts Swede Tomato

Nitrate concentrations in raw vegetables No. of samples

22 16 8 20 22 8 180 13 4 4 4

Note: Data from reference 47.

Nitrate concentration (mg/kg) Mean

Range

1211 338 86 97 1051 48 155 1631 59 118 17

224–1877 26–1523 34–164 11–566 85–3857 13–192 3–1077 266–2634 30–80 39–239 4–42

Inorganic contaminants in food

165

compared with the other periods. For lettuce, nitrate concentrations were lowest in the spring period, as they also were for carrots. Conversely, nitrate levels were higher in cauliflowers grown in spring and summer periods than for the rest of the year. Nitrate levels in onions were higher in the autumn. This study also considered the effect of cooking on the nitrate levels in the vegetables. Boiling reduced the nitrate concentrations in cabbages, carrots, cauliflowers, potatoes, spinach and swedes, but did not significantly reduce concentrations in Brussels sprouts and beetroot. Frying increased concentrations in onions (probably due to water loss), but did not change concentrations in potatoes and tomatoes. Baking did not cause any significant changes in concentrations in potatoes. 7.3.2 Dietary exposure As with the metals and metalloids mentioned earlier, the dietary exposure to nitrate in the UK has been monitored for a number of years. A Total Diet Study (TDS) carried out for samples purchased in 1997 found that dietary exposures for the general population were lower than for the comparable study carried out in 1994.48 The population average dietary exposure for 1997 was 52 mg/day compared with 68 mg/day in 1994, the decrease being partly explained by lower nitrate concentrations found in green vegetables. From the same study, the mean and 97.5th percentile exposures were 57 mg/day and 136 mg day respectively. The estimate for a mean consumer in 1997 was considerably lower than the ADI for nitrate of 219 mg/day for a 60 kg person, even allowing for additional exposure from water and beer, giving a total of 88 mg/day. 7.3.3 Nitrite in food There have been few studies on nitrite concentrations in individual foods. The 1997 UK TDS48 found that the population average dietary exposure was 1.3 mg/ day, as compared with 1.7 mg/day in 1994. The mean consumer had a slightly higher exposure at 1.4 mg/day, while the 97.5th percentile had an exposure of 2.3 mg/day. Allowing for a minimal additional exposure from water, the total mean exposure of 1.4 mg/day is well below the ADI for nitrite of 3.6 mg/day for a 60 kg person.

7.4 1. 2. 3.

References Food Chemical Surveillance Annual Report – 1999, MAFF Publications, London, 2000. PEATTIE M E, BUSS D H, LINDSAY D G and SMART G O, ‘Reorganization of the British Total Diet Study for monitoring food constituents from 1981’, Food and Chemical Toxicology, 1983 21 503–7. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘The British diet: finding

166

4. 5. 6.

Food chemical safety the facts’, Food Surveillance Paper No. 40, London, The Stationery Office, 1994. GREGORY J, FOSTER K, TYLER H and WISEMAN M, ‘The Dietary and Nutritional Survey of British Adults’, London, The Stationery Office, 1990. WORLD HEALTH ORGANIZATION, ‘Inorganic lead’, Environmental Health Criteria 165, Geneva, World Health Organization, 1995. YSART G E, MILLER P F, CREWS H, ROBB P, BAXTER M, DE L’ARGY C,

and HARRISON N, ‘1997 UK total diet study – dietary exposures to aluminium, arsenic, cadmium, chromium, copper, lead, mercury, nickel, selenium, tin and zinc, Food Additives and Contaminants, 2000 17 775–86. DABEKA R W and MCKENZIE A D, ‘Survey of lead, cadmium, fluoride, nickel, fluoride, nickel and cobalt in food composites and estimation of dietary intakes of these elements by Canadians in 1986–1988’, Journal of AOAC International, 1995 78 897–909. ELLEN G, EGMOND E, VAN LOON J W, SAHERTIAN E T and TOLSMA K, ‘Dietary intakes of some essential and non-essential trace elements, nitrate, nitrite and N-nitrosamines by Dutch adults: estimated by a 24-hour duplicate portion study’, Food Additives and Contaminants, 1990 7 207–22. MACINTOSH D L, SPENGLER J D, OZKAYNAK H, TSAI L and RYAN B, ‘Dietary exposures to selected metals and pesticides’, Environmental Health Perspectives, 1996 104 202–9. LOFTHOUSE S, SARGENT C

7.

8.

9. 10.

DELVES H T, DIAPER S J, OPPERT S, PRESCOTT-CLARKE P, PERIAM J, DONG W,

and GOMPERTZ D, ‘Blood levels in the United Kingdom have fallen substantially since 1984’, British Medical Journal, 1996 313 883–4. The Lead in Food Regulations (S.I. 1979 No. 1254), as amended. London, The Stationery Office, 1979. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Lead in food: progress report’, Food Surveillance Paper No. 27, London, The Stationery Office, 1989. EUROPEAN COMMUNITY, ‘Council Directive of 18 June 1991 amending for the tenth time Directive 76/769/EEC on the approximation of the laws, regulations, and administrative provisions of the Member States relating to restrictions on the marketing and use of certain dangerous substances and preparations’, Official Journal of the European Communities L186 59–63. WORLD HEALTH ORGANIZATION, ‘Evaluation of certain food additives and contaminants’, WHO Technical Report Series Number 837, Geneva, World Health Organization, 1993. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Concentrations of metals and other elements in marine fish and shellfish’, Food Surveillance Information Sheet No. 151, 1998. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Lead, cadmium, copper and zinc in offals’, Food Surveillance Information Sheet No. 160, 1998. EUROPEAN COMMISSION, ‘Report on tasks for scientific co-operation: COLHOUN H

11. 12. 13.

14. 15. 16. 17.

Inorganic contaminants in food

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

167

dietary exposure to cadmium’, Food Science and Techniques, EUR 17257, Luxembourg, Office for Official Publications of the European Communities, 1997. VANNOORT R W, HANNAH M L and PICKSTON L, ‘1990/1991 New Zealand Total Diet Study. Part 2: Contaminants elements’, ESR: Health, Wellington, 1995. MORGAN H (ed), ‘The Shipham Report: An investigation into cadmium contamination and its implications for human health’, The Science of the Total Environment 1988 75 (1). The Arsenic in Food Regulations (S.I. 1959 No. 831), as amended. London, The Stationery Office, 1959. WORLD HEALTH ORGANIZATION, ‘Evaluation of certain food additives and contaminants’, WHO Technical Report Series Number 776, Geneva, World Health Organization, 1989. EDMONDS J S and FRANCESCONI K A, ‘Arsenic in seafoods: human health aspects and regulations’, Marine Pollution, 1993 26 665–74. BUCHET J P, PAUWELS J and LAUWERYS R, ‘Assessment of exposure to inorganic arsenic following ingestion of marine organisms by volunteers’, Environmental Research, 1994 66 44–51. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Lead, arsenic and other metals in food’, Food Surveillance Paper No. 52, London, The Stationery Office, 1998. NORMAN J A, PICKFORD C J, SANDERS T W and WALLER W, ‘Arsenic and iodine in kelp-based dietary supplements’, Food Additives and Contaminants, 1988 5 103–9. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Metals and other elements in dietary supplements and licensed medicinal products’, Food Surveillance Information Sheet No. 156, 1998. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Survey of mercury in food: second supplementary report’, Food Surveillance Paper No. 17, London, The Stationery Office, 1987. DEPARTMENT OF HEALTH, COMMITTEE ON TOXICITY OF CHEMICALS IN FOOD, CONSUMER PRODUCTS AND THE ENVIRONMENT,

‘Statement on the toxicity

of dental amalgam’, 1997. 29.

30. 31. 32.

GRANDJEAN P, WEIHE P, WHITE R F, DEDES F, ARAKI S, YOKOHAMA K, MURATA K, SORENSEN H R and JORGENSEN P J, ‘Cognitive deficit in 7-year

old children with prenatal exposure to methylmercury’, Neurotoxicity and Teratology 1997 19 417–28. EUROPEAN COMMUNITY, ‘Council Decision 93/315/EEC of 19 May 1993 on a limit for the total mercury content of the edible parts of fresh fish’, Official Journal of the European Communities L144 23–4. WORLD HEALTH ORGANIZATION, ‘Toxicological evaluation of certain food additives and contaminants’, WHO Food Additives Series Number 24, Geneva, World Health Organization, 1989. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Monitoring and

168

33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

Food chemical safety surveillance of non-radioactive contaminants in the aquatic environment and activities regulating the disposal of wastes at sea, 1995 and 1996’, Aquatic Environment Monitoring Report No. 51, London, MAFF, 1998. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Cadmium, mercury and other metals in food’, Food Surveillance Paper No. 53, London, The Stationery Office, 1998. The Tin in Food Regulations (S.I. 1992 No. 496), as amended. London, The Stationery Office, 1992. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Tin in canned tomato products’, Food Surveillance Information Sheet No. 179, 1999. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Aluminium in food’, Food Surveillance Paper No. 39, London, The Stationery Office, 1993. GRAMICCIONI L, INGRAO G, MILANA M R, SANTORONI P and TOMASSI G, ‘Aluminium levels in Italian diets and in selected foods from aluminium utensils’, Food Additives and Contaminants, 1996 13 767–74. WORLD HEALTH ORGANIZATION, ‘Toxicological evaluation of certain food additives’, WHO Food Additives Series Number 17, Geneva, World Health Organization, 1982. FESTA M D, ANDERSON H L, DOWDY R P and ELLERSICK M R, ‘Effect of zinc on copper excretion and retention in men’, American Journal of Clinical Nutrition, 1985 41 285–92. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Survey of aluminium, antimony, chromium, cobalt, indium, nickel, thallium and tin in food’, Food Surveillance Paper No. 15, London, The Stationery Office, 1985. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘1994 Total Diet Study (Part 2) – Dietary intakes of metals and other elements’, Food Surveillance Information Sheet No. 149, 1998. NATIONAL ACADEMY OF SCIENCES, ‘Recommended Dietary Allowances’, USA, 1980. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Nitrate, nitrite and Nnitroso compounds in food’, Food Surveillance Paper No. 20, London, The Stationery Office, 1987. The Preservatives in Food Regulations (S.I. 1989 No. 533), as amended. London, The Stationery Office, 1989.

45.

COMMISSION OF THE EUROPEAN COMMUNITIES SCIENTIFIC COMMITTEE FOR FOOD, ‘Opinion on nitrate and nitrite, expressed on 19 October 1992’,

46.

COMMISSION OF THE EUROPEAN COMMUNITIES SCIENTIFIC COMMITTEE FOR

1992. FOOD,

47. 48.

‘Opinion on nitrate and nitrite, expressed on 22 September 1995 (Annex to Document III/5611/95)’, 1995. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Nitrate, nitrite and Nnitroso compounds in food: Second report’, Food Surveillance Paper No. 32, London, The Stationery Office, 1992. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘1997 Total Diet Study – nitrate and nitrite’, Food Surveillance Information Sheet No. 163, 1998.

8 Environmental organic contaminants in food N. Harrison, Food Standards Agency, London

8.1

Introduction

There is a continuing interest in contamination of the food chain with organic chemicals that are either produced by industry or are by-products of industrial activity. The need to identify and quantify any risks posed by such chemicals has led to the development of sensitive analytical techniques, utilising novel approaches such as immumochemical analysis. These enable the determination of these compounds at concentrations ranging from micrograms (10 6 grams) to femtograms (10 15). Where analytical methods are available it is largely because of a crossfertilisation of effort from well-established areas of food contaminants work. For example, the steady development since the 1960s of methods of analysis for chlorinated pesticides led to the analysis of food for polychlorinated biphenyls (PCBs) since PCBs were readily detectable by general methods used to analyse food for organochlorine pesticides. The analysis of food for chlorinated dioxins and furans (PCDDs and PCDFs) at the very low levels at which they are found in food is a more recent development, and one that is an important precedent since it arose from interest in environmental contamination rather than because of cross-fertilisation of scientific methodology from an established area of food chemistry. Although dioxins were detectable some years ago at much less sensitivity in some pesticides, it was environmental interest that led to their study at very low levels in the food chain. In addition to the well-characterised groups of contaminants such as PCBs and polycyclic aromatic hydrocarbons (PAHs), there are hundreds of thousands of other organic chemicals that are emitted by industrial processes and which have the potential to enter the food chain by a variety of routes. It would be

170

Food chemical safety

impossible to monitor all these organic chemicals, so a systematic process is needed for selecting chemicals to study. A simple and pragmatic scheme was developed in a paper by Wearne et al.1 The following criteria are probably the key ones in the selection process: • Production volume. This criterion is based on the assumption that the amount of chemical released to the environment is related to the quantity produced. Small volumes of production of a given chemical should not necessarily eliminate it from study, but since there are as many as 50,000 different chemicals produced by industry worldwide, it is important to exclude lowvolume chemicals where there is evidence that they are not toxic or environmentally persistent. • Pattern of usage. This is important because it takes account of the different potentials for release into the food chain arising from the various uses of an industrial chemical. A higher priority would be given to highly diffuse uses, such as would arise from the use of organic chemicals as non-active components of pesticide formulations. • Possible fate in the environment. An industrial chemical that has been released into the environment will exist in differing concentrations in the various environmental compartments. The concentrations of a substance in air, water, soil and other media following release can be modelled using the concept of fugacity.2 At its simplest, this involves only the use of standard physico-chemical data to estimate the partitioning between the various media. • Likelihood of chemical entering the food chain. The likelihood of an organic environmental contaminant entering the food chain is determined by the availability of the chemical in the environment. This in turn is largely controlled by the rate of degradation of the chemical in the phase in which it is present. This factor must be assessed on the basis of observed data rather than by a theoretical approach. • Mechanism of entry into the food chain. This is necessarily empirical and is based on the number of reports of food contamination. It will tend to introduce a bias towards chemicals known to contaminate food and which are frequently the subject of targeted surveys. • Persistence in the food chain. This is clearly a key factor as this reflects the ease with which chemicals are taken up from air and soil by plants and animals and thus the potential for bioaccumulation. However, there is very little known about it for most industrial chemicals and by-products, although the bioaccumulation factor can be measured experimentally or calculated using mathematical modelling. • Toxicity. This factor takes account of available data on animal and human studies. Studies on worker and environmental safety have created a lot of information about the toxicity of industrial chemicals. In general, enough is known about the toxicology of some industrial chemicals for the researcher to discriminate between those chemicals that are worth studying in food and those that are not.

Environmental organic contaminants in food

171

The scheme described provides a simple and pragmatic tool to select chemicals that should be looked for in the food chain by considering sound evidence in each of the above areas. Further study of the factors influencing the contamination of the food chain by organic chemicals and improved modelling techniques should enable a greater range of chemicals to be prioritised for detailed investigation. Substances that are currently being studied in food or related materials for both academic interest and by regulators are covered in the next sections of this chapter.

8.2

Aromatic hydrocarbons

Benzene is a naturally occurring chemical found in crude petroleum. Very large quantities are also produced industrially worldwide. It has been widely used as a solvent, mainly in paints and adhesives, but the recognition of its carcinogenic potential has led to a reduction in these uses. It is still used in the manufacture of other chemicals and end products, including ethyl benzene (to produce styrene), cumene (to produce phenol) and cyclohexane (used in the production of nylon). Benzene and other aromatic hydrocarbons are constituents of petrol and diesel fuels – petrol typically contains 3 percent benzene, 1 percent toluene, 4 percent xylenes, 2 percent ethylbenzene, and less than 1 percent of cumene and naphthalene. Most of the benzene, cumene and naphthalene released to the atmosphere in the UK arises from motor vehicle exhaust gases. Road transport also makes a large contribution to releases of toluene, but is a less important source of ethylbenzene and xylenes for which industrial solvent usage is the major source of releases to the atmosphere. As a result of their wide range of uses and their potential for loss to the environment, aromatic hydrocarbons have been reported in air and surface water throughout industrialised nations.3–8 Once released to the atmosphere, aromatic hydrocarbons may enter foods via a number of routes, including direct absorption from the atmosphere by fatty foods at different points of the food chain.9 Benzene is considered to be a genotoxic carcinogen and has been shown to induce leukaemia in some humans exposed to relatively high occupational levels.10 There are less toxicological data for toluene, ethylbenzene, xylenes, cumene and naphthalene. Estimates of daily exposure to benzene from urban or suburban air range from 180 to 1300 g/person/day.11, 12 Urban air concentrations of the other aromatic hydrocarbons are similar to those of benzene and the vast majority of exposure of the general population to these other aromatic hydrocarbons will be due to road transport or solvent-containing products rather than food. A 1995 survey of these compounds in samples from the UK Total Diet Study showed that average dietary exposures to benzene and related compounds from food in the UK are low, and very much lower than estimated exposure from active smoking of tobacco or intakes from air by urban dwellers.13 The mean dietary exposure to benzene was estimated to be in the range 0.9–2.4 g/person/day.

172

Food chemical safety

This estimate is similar to the 1.2 g/person/day for the exposure to benzene by Canadian adults from food.14 These exposures to benzene from food are up to three orders of magnitude smaller than the estimated daily exposure to benzene from active smoking of tobacco, or exposures from air by urban dwellers. Owing to their greater exposure to motor vehicle exhaust emissions, it is possible that fatty foods on sale at shops attached to petrol stations or at stalls and shops in busy roads could contain higher concentrations of aromatic hydrocarbons than similar foods on sale at other shops. A study in Germany found that concentrations of benzene and toluene were higher in retail packs from petrol stations on busy roads than from petrol stations in rural areas.15 It also found that retail packs from shops in busy roads contained higher concentrations of benzene, toluene, xylenes and ethylbenzene than retail packs from shops in residential areas. A similar survey in the UK, carried out to investigate this hypothesis, analysed fatty foods purchased across the UK from urban areas and locations believed to be remote from sources of aromatic hydrocarbons.16 No discernible differences were found in the concentrations of aromatic hydrocarbons in the fatty foods on sale in the various locations.

8.3

Polycyclic aromatic hydrocarbons

Polycyclic aromatic hydrocarbons (PAHs) are a large group of substances with the common structural feature of two or more fused benzene rings. Examples of their structures are given in Fig. 8.1. PAHs can be formed as products of the incomplete pyrolysis of organic materials and are present in considerable quantities in fossil fuels from which they are released by a variety of combustion processes. High temperatures and open flames are not necessarily required, however, for the formation of PAHs. For example, the aromatic hydrocarbons of crude oil were formed over millions of years in sediments that were at temperatures between 100 and 150 degrees Celsius. Sources of environmental PAHs include power plants, domestic heating systems that burn oil, coal or wood for example, gasoline and diesel engines, waste incineration, various industrial activities, and tobacco smoke.17 More specifically, petroleum refining processes contribute to localised loadings of PAHs into the environment through industrial effluents from coal gasification and liquefaction processes and accidental spillage of raw and refined petroleum. PAHs also partition from the atmosphere to vegetation, soil, water, and sediment. PAHs are hydrophobic compounds and their persistence in the environment is mainly due to their low water solubilities and electrochemical stability. Higher molecular weight PAHs exhibit greater environmental persistence than lower molecular weight PAHs due to an increase in hydrophobicity and an increase in stability with increasing mass. Evidence suggests that the lipophilicity, environmental persistence, and toxicity of PAHs increase as the molecular size of the PAHs increases up to four or five fused benzene rings.18 This relationship

Environmental organic contaminants in food

Fig. 8.1

173

Structures of typical polycyclic aromatic hydrocarbons.

of environmental persistence and increasing number of benzene rings is consistent with the results of studies correlating biodegradation rates and PAH molecule size.19, 20 The many natural and anthropogenic sources of PAHs in combination with global transport phenomena result in the worldwide distribution of these compounds. The concentration of PAHs in any given area can vary widely, depending on the level of industrial development and transport processes. In a study of different soils collected from rural and urban areas (not from grossly contaminated locations such as near gas works or refineries) at 49 locations in Wales, UK, Jones et al.21 found a range of PAH contamination from between 0.1 to 55 g/g soil. PAHs have been described as the largest group of known environmental carcinogens22 and are probably the most widely distributed environmental contaminants. Human exposure to them is unavoidable. There is a considerable amount of data on levels of PAHs in food, covering a large number of different members of the group as there are differing views on the priority to be given to the individual PAHs. Table 8.1 lists the 16 PAHs identified as the priority environmental pollutants, with the 12 compounds identified by the UK Committee on Carcinogenicity (COC) starred. The PAHs analysed tend to be those with 4- to 5- rings as some members with these structures have been found to be carcinogens, teratogens and mutagens. However, benzo[a]pyrene (BaP) is

174

Food chemical safety

Table 8.1

EPA priority environmental pollutants

Naphthalene Acenaphthene Acenaphthylene Fluorene Anthracene* Phenanthrene* Fluoranthene* Pyrene*

Benzo[a]anthracene* Chrysene* Benzo[b]fluoranthene* Benzo[k]fluoranthene* Benzo[a]pyrene* Dibenz[a,h]anthracene* Benzo[g,h,i]perylene* Indeno[1,2,3-cd]pyrene*

Note: COC list also includes anthanthrene, benzo[e]pyrene, benzo[b]naph[2,1-d]thiophene and cyclopenta[cd]pyrene * ˆ priority PAH by COC.

usually highlighted, although it comprises less than 5 percent of the total amount of PAHs present in the atmosphere, since it is thought to be the most toxic PAH. The International Agency for Research on Cancer (IARC) considers BaP a known animal carcinogen and a probable human carcinogen (Group 2A).23 Published surveys include a UK Total Diet Study in 1983.24 This estimated that the total daily dietary exposure to PAHs was around 3700 ng/person, with BaP contributing 250 ng/person. Similar figures have been reported for the Netherlands, where a total diet study found that the average daily dietary exposure to PAHs ranged from 5000 to 17,000 ng/person.25 A market basket study of components of the Italian diet found that the average daily dietary exposure was 3000 ng/person.26 In a follow-up study of those food groups making major contributions to the overall dietary exposure in the UK, the cereals and fats and oils groups were each shown to contribute around one-third of the total dietary exposure.27 Much of this was due to the edible oils in the cereal products. The PAHs were traced back through margarine production to oilseed growth. The source appeared to be on the farm, leaving an unidentified environmental source as the most likely explanation. Recent studies on PAHs in food include those on vegetables in China and Greece.28, 29 The study in Linxian, China was to investigate whether environmental exposure to PAHs via ingestion of food cooked with soft coal was the reason for a high rate of oesophageal cancer.30 High levels of BaP were found in most food samples. The median values for BaP were: 4.6 ng/kg for cooked wheat products and 4.9 ng/kg for cooked corn products. These levels were found to be higher that the mean BaP concentration of 0.12 ng/kg in comparable wheat and corn products obtained in the US and similar to levels in very well-done charcoal-grilled red meat (2.6–25.2 ng/kg).31, 32 The dietary exposure to PAHs via vegetables in an industrialised area of Greece was low, at between 1.6 and 4.5 g/person/day.29 The vegetables in this study were prepared as for consumption, i.e. after washing with tap water and peeling of carrot roots. This reflects the ease by which PAHs can generally be removed from fruit and vegetables after deposition.

Environmental organic contaminants in food

8.4

175

Dioxins and PCBs

Polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (generally known as PCDDs and PCDFs) are ubiquitous environmental contaminants. The known sources of PCDDs and PCDFs are many, including the following: • Chemical manufacture. Chlorination of organic chemicals such as pesticides may lead to the formation of PCDDs and PCDFs as by-products. Action has generally been taken to avoid formation of these compounds where this is known to occur. New environmental contamination from this source is now unlikely, but most chlorinated chemicals are persistent and will be found in environmental media for many years. • Bleaching processes. Historically, the use of chlorine to bleach wood pulp was an important source of PCDDs and PCDFs, both to the environment and to paper, some of which was used in sensitive applications such as food packaging and disposable nappies. However, this type of bleaching is now usually done using non-chlorine based processes, for example by using hydrogen peroxide. • Combustion processes. Combustion is still a major source of these substances. PCDDs and PCDFs have been found in exhaust gases from a wide variety of combustion processes – from cigarette smoke to emissions by fossil fuel power plants. Stringent controls have been introduced on waste incinerators in the UK and several other countries. It is likely now that most formerly significant sources have been controlled, leaving diffuse combustion sources, e.g. bonfires and car exhausts, as major contributors to the environmental background.

The wide range of potential sources and their environmental persistence may well explain why PCDDs and PCDFs are ubiquitous. It is believed that there are two main routes by which PCDDs and PCDFs are released into the environment: via the atmosphere leading to deposition on soil, water and plants and via solid or liquid waste with subsequent contamination of land (e.g. sewage sludge). These findings should help to reduce the already extensive effort needed to trace back the contamination to its sources, although the persistence of PCDDs and PCDFs in the environment for many years means that historical as well as current sources need to be taken into account. PCBs have been used since the 1930s in a wide variety of industrial applications, since they are chemically inert and have good dielectric properties. They were used as dielectric fluids in transformers and capacitors in electrical machinery, and in materials such as carbonless copy paper as commercial formulations known as Arochlors. However, as evidence grew that they are very persistent in the environment and that they might be toxic, their usage decreased. The manufacture and general use of PCBs ceased in the mid-1970s and was banned in the UK under the Control of Pollution (Supply and Use of Injurious Substances) Regulations 1986 (S.I. 1986 No. 902).33 The main potential sources are now persistent residues in the environment and PCBs in old industrial

176

Food chemical safety

equipment, although controls over the disposal of equipment mean that the latter source is decreasingly likely to contaminate the food chain. PCBs sealed inside older electrical equipment were required to be destroyed by the end of 2000 under UK PCB Regulations.34 PCBs must be disposed of in an environmentally sound manner and are generally destroyed by high-temperature incineration or dechlorination processes. Concern over dioxins originally arose over one particular dioxin, 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD). This was found to produce clinical effects (chloracne) in workers exposed to it through industrial accidents.35 It is also toxic to some species of laboratory animals. A few PCBs have been shown in experimental systems to exert a number of toxic responses similar to those observed for TCDD. In consequence, the total concentrations of both dioxins and PCBs in environmental media, including food, are now commonly determined as part of the same chemical analysis, and each cannot be discussed now without reference to the other. 8.4.1 Toxic equivalents There are 75 possible isomers of polychlorinated-p-dibenzodioxin (PCDD) and 135 such isomers of polychlorinated dibenzofuran (PCDF). These related compounds are known as congeners. In addition, PCBs form a similar group of 209 related congeners that differ only in the number and pattern of chlorine atoms attached to the parent biphenyl molecule. The complex structures of these substances (Fig. 8.2) are apparently highly resistant to biological degradation which results in a high degree of persistence in the environment. However, the degree of chlorination, which is from one to eight chlorine atoms per molecule, seems to determine the toxicity of the molecules. For PCDDs and PCDFs, it is four chlorine atoms at positions 2, 3, 7 and 8 that leads to greatest chronic toxicity. The sixteen other dioxins which contain chlorine at these positions are thought to be less toxic, while the other PCDDs and PCDFs are assumed to have relatively little biological activity. A much wider range of PCBs is of interest. There is some evidence suggesting that TCDD and some PCBs have a common mechanism of action in biological systems, based on the binding of these compounds to a specific cellular receptor, the Ah-receptor. As PCDDs and related compounds are normally present in environmental and food samples as complex mixtures of congeners, the concept of Toxic Equivalents (TEQs) has been developed to aid risk assessment. This concept uses the available toxicological and biological data to generate a series of weighting factors, called Toxic Equivalency Factors (TEFs), each of which expresses the toxicity of a ‘dioxin-like’ compound in terms of the equivalent amount of TCDD. Multiplication of the concentration of a compound by its TEF gives a TEQ. The initial system that was widely adopted was that devised by the North Atlantic Treaty Organisation, Committee on the Challenges of Modern Society (NATO-CCMS) of International Toxic Equivalency Factors (I-TEF).36 Similar

Environmental organic contaminants in food

Fig. 8.2

177

Structures of polychlorinated dibenzo-p-dioxin, polychlorinated dibenzofuran and polychlorinated biphenyl.

systems to that of CCMS have been developed by a Nordic Expert Group, the German Federal Health Office and the US Environmental Protection Agency. These systems tend to give lower total figures for toxic equivalents as they use lower toxicity equivalencies for one or more of the pentachloro-PCDDs or PCDFs. More recently, the World Health Organization (WHO) has agreed a further set of TEFs (WHO-TEFs), which also apply to PCBs.37 These effectively supersede those proposed by Ahlborg et al.38 The TEFs for PCDDs and PCDFs have been slightly revised from the I-TEF scheme and Table 8.2 shows the differences between the two systems. The WHO system of TEFs is becoming increasingly widely recognised and used for risk assessment of these chlorinated compounds. In this way the amounts of the different PCDDs, PCDFs and PCBs can now be converted to one figure which is a measure of the toxicological impact of the detected mixture. This approach simplifies the reporting of surveys for PCDDs, PCDFs and PCBs in food, allowing the correlation of different surveys from country to country and a greater understanding of dietary exposure. 8.4.2 Toxicity of PCDDs, PCDFs and PCBs An expert group convened by the WHO Regional Office for Europe in 1990 recommended a Tolerable Daily Intake (TDI) of 10 pg/kg bodyweight for

Table 8.2

Toxic Equivalency Factors using I-TEF and WHO-TEF systems Toxic Equivalency Factor 36

Toxic Equivalency Factor 37

PCDDs and PCDFs

CCMS 1988

WHO 1997

PCBs

Ahlborg et al. 199438

WHO 199737

2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD 2,3,7,8-TCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF

1 0.5 0.1 0.1 0.1 0.01 0.001 0.1 0.05 0.5 0.1 0.1 0.1 0.1 0.01 0.01 0.001

1 1 0.1 0.1 0.1 0.01 0.0001 0.1 0.05 0.5 0.1 0.1 0.1 0.1 0.01 0.01 0.0001

3,3’,4,4’-CB(77) 3,4,4’,5-CB(81) 2,3,3’,4,4’-CB(105) 2,3,4,4’,5-CB(114) 2,3’,4,4’,5-CB(118) 2,3,4,4’,5-CB(123) 3,3’,4,4’,5-CB(126) 2,3,3’,4,4’,5-CB(156) 2,3,3’,4,4’,5’-CB(157) 2,3’,4,4’,5,5’-CB(167) 3,3’,4,4’,5,5’-CB(169) 2,2’,3,3’,4,4’,5-CB(170) 2,2’,3,4,4’,5,5’-CB(180) 2,3,3’,4,4’,5,5’-CB(189)

0.0005 – 0.0001 0.0005 0.0001 0.0001 0.1 0.0005 0.0005 0.00001 0.01 0.0001 0.00001 0.0001

0.0001 0.0001 0.0001 0.0005 0.0001 0.0001 0.1 0.0005 0.0005 0.00001 0.01 – – 0.0001

Environmental organic contaminants in food

179

2,3,7,8-TCDD.39 Taking into account toxic equivalency of the many PCDFs and PCDDs, this TDI can be regarded as 10 pg TCDD equivalents/kg bodyweight per day or seven times this for a time-weighted average tolerable intake per week. These are very useful parameters with which to compare toxic equivalents of PCDDs and PCDFs found in food surveillance. The UK Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment (COT) endorsed this TDI in 1992.40 More recently, the COT has tentatively accepted41 that this TDI can be applied to mixtures of dioxins and those PCBs for which TEFs have been set by the WHO,37 as a pragmatic approach to the evaluation of complex mixtures of these compounds in food. A re-evaluation of the TDI for these compounds has led to the publication of the conclusions of a group of WHO experts in 199942 that the TDI should be in the range 1–4 pg/kg bodyweight for 2,3,7,8-TCDD, the lower end of the range being seen as a target. 8.4.3 PCDDs, PCDFs and PCBs in food It is only since the mid-1980s that surveys for these substances in food have been carried out to any great extent. This is mainly because reliable, sensitive methods for analysing food for them have been developed only recently, and are still being refined. The methods need to be capable of analysing at levels in the order of 0.01 ng/kg (1 part in 1014). At this degree of sensitivity and with the complex mixtures of PCDDs, PCDFs and PCBs that can be present, very strict criteria must be applied in carrying out and assessing the analytical work involved. The key criteria were defined to include use of an adequately documented method; correct use of internal standards; specified limits of detection; quality control and criteria for data acceptance.40 Applying these criteria should ensure that analytical results are properly representative of the sample that has been analysed. Survey results need to be very carefully assessed before they can be accepted and this must take account of how the analytical methods were applied, as well as the representativeness of the samples of food of the general supply. Since analysis for PCDDs, PCDFs and PCBs in food is a particularly complex and hence expensive and time-consuming process, particular care needs to be taken that statistically sufficient samples are analysed in each survey. Given all these difficulties it is not surprising that only recently have major surveys of PCDDs and PCDFs in food become a viable prospect. This now extends to congener-specific surveys for PCBs, although it is not usual to quantify all congeners, principally those with TEFs and those selected by the International Committee for the Exploration of the Sea43 for analysis in fish. The principal surveys on food have been market-basket or Total Diet Studies (see Chapter 7) to determine the average dietary exposures of national populations. There are values for average exposure from the UK diet for 1982, 1992 and 1997; the 1982 samples being measured at the same time as the 1992 samples to ensure comparability. These studies have shown that average

180

Food chemical safety

dietary exposures of UK consumers fell substantially between 1982 and 1992, from an upper bound figure of 6.8 pg I-TEQ/kg (7.2 pg WHO-TEQ/kg bodyweight) to 2.4 pg I-TEQ/kg bodyweight (2.5 pg WHO-TEQ/kg bodyweight). More recently, analysis of composite food groups from 1997 found that concentrations of PCDDs and PCDFs were generally lower in most food groups than those in the 1982 and 1992 surveys.44 These decreases were reflected in the dietary exposures estimated from the results of the surveys, with the average dietary exposure now at an upper bound figure of 1.8 pg WHOTEQ/kg bodyweight. This can be compared with the UK TDI of 10 pg WHOTEQ/kg bodyweight/day and also now with the WHO TDI of 1–4 pg WHOTEQ/kg bodyweight/day. This latest dietary exposure is similar to those found in other countries, although most figures are quoted in terms of I-TEQs. Converted to I-TEQ from WHO-TEQs, the UK figure of 0.8 pg I-TEQ/kg for PCDDs and PCDFs bodyweight alone in 1997 may be compared with 0.7 pg I-TEQ/kg bodyweight for 1997 in Italy,45 1.4 pg I-TEQ/kg bodyweight in Norway,46 2.4–3.5 pg ITEQ/kg bodyweight for 1995 in Spain47 and the very low figure of 0.2 pg ITEQ/kg bodyweight for New Zealand.48 The converted figure of 0.9 pg I-TEQ/kg bodyweight for PCBs alone may be compared with 1.8 pg I-TEQ/kg bodyweight in Norway,46 2.2 I-TEQ/kg bodyweight for 1995 in Spain47 and another low value of 0.2 pg I-TEQ/kg bodyweight for New Zealand.48 In addition to the above studies of population average exposure, there have been surveys for dioxins and PCBs in individual foods. One example is discussed below. A survey of fish for PCDDs, PCDFs and PCBs in samples of various UKlanded and imported marine fish species, salmon and fish fingers collected in 1995/96 found concentrations of dioxins and PCBs in the range 0.9–140 ng WHO-TEQ/kg fat.49 Further details are given in Table 8.3. Exposure to dioxins and PCBs from the consumption of fish in combination with the rest of the diet was estimated to be 2.6 pg WHO-TEQ/kg bodyweight/day for an average UK adult consumer and 5.6 pg WHO-TEQ/kg bodyweight/day for a high-level adult consumer. PCDDs, PCDFs and PCBs are highly fat-soluble and accumulate in adipose tissue. They can also pass through the placenta and are excreted in human breast milk, resulting in exposure of the nursing infant (Table 8.4). The results from UK participants in a WHO inter-laboratory trial showed that the concentrations of PCDDs and PCDFs fell from 29–37 ng I-TEQ/kg milk fat in 1987–1988 to 21–24 ng I-TEQ/kg milk fat in 1993–94.44 Although PCBs were not analysed in 1987–1988, the concentrations in the 1993–94 milk samples were 10–12 ng ITEQ milk fat. The concentrations were similar to those reported by other European countries and other participants in the WHO trial. As PCDDs and PCDFs are lipophilic, the major sources of PCDDs and PCDFs in the diet tend to be fat-containing foods of animal origin. Cows’ milk is a useful indicator of environmental pollution since cows graze relatively large areas and any PCDDs and PCDFs present on or in grass eaten

Table 8.3 Fish type

Concentrations of PCDDs, PCDFs and PCBs in edible tissue samples from marine fish (ng WHO-TEQ/kg fat) PCDDs and PCDFs Mean Range

Mean

UK landed: Cod Haddock Plaice Whiting Herring Mackerel Salmon Trout

9.0 6.9 25 8.3 24 3.8 6.5 5.7

2.1–24 1.1–14 3.6–43 2.0–20 13–38 1.0–9.0 4.6–11 2.4–14

Fish fingers

0.7

Imported: Cod Haddock Plaice Salmon Red fish

6.1 4.6 20 3.4 14

Note: Data from reference 49.

PCBs

Range

Dioxins and PCBs Mean Range

17 7.4 42 23 59 14 19 18

3.3–76 2.2–22 9.5–55 2.4–91 12–110 2.5–31 12–30 8.7–50

26 14 67 32 83 17 25 24

7.2–98 5.5–24 13–90 4.4–110 26–140 3.4–40 16–38 12–61

0.3–2.4

1.6

0.3–6.2

2.3

0.9–6.6

1.4–18 1.9–8.5 16–27 3.4 12,16

9.7 5.4 33 12 43

2.0–32 1.9–12 21–57 12 42,44

16 10 54 16 57

6.3–50 4.2–19 37–84 16 57,57

182

Food chemical safety

Table 8.4

PCDDs, PCDFs and PCBs in human milk (ng I-TEQ/kg milk fat)

Country

Year

PCDDs + PCDFs

PCBs

Reference

Denmark Russia Germany

1993/94 1993 1995

16.7 16 17

17.2 13 32

Hilbert et al. 199650 Polder et al. 199651 Malisch 199652

by the cows would concentrate in the milk fat. Milk and milk products are also an important source of nutrients for much of the general population. These targeted surveys cannot provide information about general dietary exposure. Indeed their results can easily be misinterpreted if they are taken as a measure of general exposure. UK surveys of cows’ milk from individual farms near to potential industrial point sources of PCDDs and PCDFs, such as steelworks and waste incinerators, have shown few indications of concentrations elevated above the expected range for the UK, apart from a few sites where there was known pollution. These sites included Bolsover in Derbyshire where it is likely that localised pollution of agricultural land and associated streams occurred from a nearby incinerator that was used to dispose of chlorinated chemical waste.40 Some farms with elevated levels of PCDDs and PCDFs were also found close to the site of a former municipal waste incinerator.53 Both sites have been the subject of continued monitoring to ensure that no risk was posed to local consumers. It is unlikely that future work to quantify dietary exposure and identify potential environmental sources of PCDDs, PCDFs and PCBs will be on a much greater scale than recent research unless quicker and less expensive methods of food analysis are developed for these substances. Recent developments in immunochemical techniques offer this potential.

8.5

Chlorinated hydrocarbons

Much of the effort on environmental chemicals that contaminate food has concentrated on a small range of chlorinated chemicals. In addition to the chlorinated PCDDs, PCDFs and PCBs already mentioned, other chlorinated compounds can be separated into two groups: chlorinated aromatic compounds and chlorinated aliphatic compounds. Although there is a number of organochlorine pesticides that are persistent in the environment, these will not be considered here, as they comprise an extensive field of study in their own right. The approved uses for chlorinated chemicals have now become so restricted that for some of them at least they are now more likely to contaminate food as a result of their persistence in the environment rather than from their direct use on the food chain. A former pesticide in this category is hexachlorobenzene. There

Environmental organic contaminants in food

183

has also been some work on residues of other chlorinated benzenes in food, and other industrial organochlorines such as 1,1,1-trichloroethene which is used as a dry cleaning fluid. 8.5.1 Aromatic compounds There are 12 chlorobenzene compounds: monochlorobenzene (MCB), three dichlorobenzenes (1,2-, 1,3,- and 1,4-DCB), three trichlorobenzenes (1,2,3-, 1,2,4- and 1,3,5-TCB), three tetrachlorobenzenes (1,2,3,4-, 1,2,3,5- and 1,2,4,5TeCB), pentachlorobenzene (PeCB) and hexachlorobenzene (HCB). HCB has been used in the UK as a pesticide, while the other chlorobenzenes all have industrial uses, mostly as intermediates in the production of pesticides and dyes, with 1,4-DCB used as a disinfectant. In addition, MCB is used as a solvent and the higher chlorobenzenes are used in dielectric fluids.54 Chlorobenzenes are widely dispersed in the environment as a result of their industrial usages and emissions from incinerators. Chlorobenzenes have an increasing tendency to bioaccumulate and to become more persistent as the number of chlorine atoms increases. They are also present in sewage sludge and have been shown to be taken up by carrots via this route.55 There have been reports of some chlorobenzenes being found in foodstuffs: in meat, fish, vegetables and crude seed oils in the UK,54, 56 in cod from the North Sea,57 in fish and shellfish in Canada,58 in fish from contaminated waters in Europe59–61 and in human milk in Canada.62 A survey carried out in 1995 for 11 chlorobenzenes in UK TDS samples and retail samples did not detect six of the chlorobenzenes in any of the samples analysed at an LOD of 0.002 mg/kg fresh weight for samples with a fat content of less than 10 per cent and 0.01 mg/kg fat for samples with a fat content greater than 10 per cent.63 The remaining five chlorobenzenes (1,2-DCB, 1,4-DCB, 1,2,4-TCB, 1,2,4,5-TeCB, 1,2,3,4-TeCB) were found only in 8 percent of the retail samples (meat and meat products, milk and milk products, fish and fish products, cereal products, potato products and olive oil). None of the chlorobenzenes was detected in any of the TDS samples. Estimated upper bound average and high level UK dietary exposures to individual chlorobenzenes were in the ranges 0.10–0.85 and 0.29–1.8 micrograms/person/day respectively, which intakes are below the TDIs published by the World Health Organization.64 Although hexachlorobenzene (HCB) was used as an active component of insecticides and fungicides in the UK until the 1970s, its major source to the environment now is now as a by-product of industrial processes such as aluminium smelting and production of perchloroethylene and vinyl chloride monomer.65 The MAFF survey mentioned earlier did not include HCB, but a later survey on milk samples from farms around potential point sources did analyse for HCB. No HCB was detected at or above the reporting limit of 1 g/ kg in whole milk in samples from either around potential point sources or control farms.66

184

Food chemical safety

The findings of the survey mentioned above point to the need for continuing surveillance and control of the food supply for these organochlorines even though their uses as pesticides are likely to decline to negligible levels in coming years. Fortunately methods of analysis for them in food are well developed, there are extensive databases built up over many years with which to compare new surveillance data, and there are ADIs for some if not all of the substances. 8.5.2 Aliphatic compounds Chlorinated aliphatic compounds used industrially include tetrachloroethene, 1,1,1-trichloroethane and trichloroethene. Tetrachloroethene (also known as perchloroethene, hence PCE) and trichloroethene (TCE) are both products of the petrochemical industry. Uses of PCE and TCE include the following: as solvent in dry-cleaning operations; a metal degreaser; as a solvent for fats, greases, waxes, rubber, gums, dyeing, and caffeine from coffee; to remove soot from industrial boilers; in the manufacturing of paint removers and printing inks; in the manufacture of trichloroacetic acid; as a heat transfer medium; as a fumigant for rodents; in the manufacture of fluorocarbons; and as an anaesthetic.65 Such organochlorines are not normally very persistent in either the environment or the food chain, as they are volatile. Since tetrachloroethene is widely used as a dry-cleaning agent, foods stored close to a source of TCE, such as a dry-cleaning operation, may be exposed to atmospheric concentrations of TCE that could then be absorbed by the food. TCE was found in butter and lard bought from shops close to dry cleaners, at concentrations up to 763 g/kg.67 Residues of 1,1,1-trichloroethane, chloroform and trichlorofluoromethane were also found in some samples of butter and lard, although at much lower concentrations than the TCE.

8.6

Phthalic acid esters

The diesters of phthalic acid, commonly known as phthalates, are a group of organic chemicals that have a variety of industrial uses, including use as plasticisers in a wide range of household and consumer goods,68 uses in lubricating oils and uses as carriers for perfumes in cosmetics. Their use in plastic food packaging is now limited to some adhesives and some printing inks. They are no longer used in the manufacture of ‘cling film’ or ‘cling wrap’. The release of phthalates to the environment may occur during the production and distribution of phthalates, during the manufacture or use of products in which they are used, or after the dispersal of these products. Although some phthalates occur naturally in coal, crude oil and shale, the contribution of such sources to general environmental levels is likely to be insignificant.68 As a result of their extensive use and their moderate resistance to degradation, phthalates are widely distributed in the environment and are often found at low levels in food. Tolerable Daily Intakes (TDIs) have been established by the EC

Environmental organic contaminants in food Table 8.5

185

Tolerable Daily Intakes for phthalate diesters

Abbreviation

Chemical name

Tolerable Daily Intake (mg/kg bodyweight/day)

DEHP BBP DBP DCHP DEP

Di-2-ethylhexyl phthalate Benzyl butyl phthalate Dibutyl phthalate Dicyclohexyl phthalate Diethyl phthalate

0.05 0.1 0.05 0.1 0.2

Scientific Committee on Food (SCF) for a number of compounds in the group permitted for use in food contact materials and these are given in Table 8.5. The SCF has also recommended a ‘group restriction’ for the sum of other phthalates. This is simply a precautionary limit set as a guide pending sufficient toxicological data to set TDIs. A Total Diet Study carried out in the UK in 1993 detected low levels of a range of individual phthalates in composite samples of carcass meat, poultry, eggs and milk.69 Only low levels (generally < 0.07 mg/kg) were found. Total phthalate was determined (by conversion of any phthalate in the samples to dimethyl phthalate) in carcass meat, eggs, meat products, offals, poultry, fish, fats and oils, milk and milk products. Each food group was found to contain total phthalate in a range between 0.5 and 8.8 mg/kg. This difference may be due to the presence of phthalate monoester metabolites, which would be included in the analytes. The dietary exposures of even 97.5th percentile consumers were estimated to be within the TDIs for DEHP, BBP and DBP. A study of infant formulae showed that phthalates were present in the majority of samples analysed.70 However, 31 percent of the samples did not contain any of the seven individual phthalates analysed. These were di-isopropyl phthalate (DIPP), dipropyl phthalate (DPP), di-iso-butyl phthalate (DIBP), dibutyl phthalate (DBP), benzyl butyl phthalate (BBP), di-2-ethylhexyl phthalate (DEHP) and di-isodecyl phthalate. Only DBP, BBP and DEHP were present at measurable concentrations. DEHP was the most abundant individual phthalate and was found at concentrations ranging from 0.05 to 0.44 mg/kg. The dietary exposures to all phthalates were below the relevant TDIs.

8.7

Endocrine disrupters

There has been considerable concern in recent years over the possible increasing trends for adverse effects on the reproductive capabilities of animals and man. In humans, effects such as decreased sperm count and increased testicular cancer in men and breast cancer in women are of particular concern.71 It has been proposed that a common cause for these diverse observations may be the disruption by certain environmental contaminants of the endocrine system which

186

Food chemical safety

regulates reproductive health. Such chemicals may mimic or interfere with the action of endogenous oestrogenic hormones to produce oestrogenic and/or oestrogenic activities.72, 73 Much of the concern focuses on man-made – and mainly organic – chemicals, as highlighted in the book Our Stolen Future.74 Possible adverse effects of endocrine disrupters include cancers, behavioural changes and reproductive abnormalities. There is conclusive evidence for effects on wildlife, but the evidence for effects on humans is still not conclusive. It is still unclear whether the presence of environmental chemicals could lead to actual exposure. 17 -oestradiol and other endogenous hormones occur naturally within the body. Endocrine disrupters are those chemicals entering the body from the external environment that may mimic or interfere with the human endocrine system. Such endocrine disrupters may derive from many sources. They may be deliberately administered as pharmaceuticals or as oral contraceptives. They also occur naturally in foodstuffs, particularly vegetables, and may be found in food moulds such as mycotoxins on grain and dried fruit. Several synthetic organic compounds have been shown to have oestrogenic properties. Some PCB congeners possess oestrogenic properties, but with a potency only about one millionth that of oestradiol.75 In contrast, the major characteristic of PCDDs and PCDFs is their anti-oestrogenic activity. For example, TCDD was shown to decrease liver and uterus weight in an in vivo bioassay for oestrogenic activity, while oestradiol caused an increase in both these parameters.76 Other organic compounds associated with endocrine disrupting activity include alkylphenol polyethoxylates, alkyl phenols such as p-octyl phenol and p-nonyl phenol, and phthalates with DEHP in particular. This is a known reproductive toxicant in animals and effects have been demonstrated on the reproductive systems of rats, including reduced serum testosterone levels and atrophied testes.68 A firm assessment of the risk to humans is not possible at present because of a lack of relevant data about the effects of exposure to endocrine disrupters. While high levels of exposure to some chemicals thought to have endocrine disrupting properties could theoretically increase the risk of reproductive and developmental disorders, no direct evidence is available at present. Further investigation of the relationship between potential endocrine disrupters and human health is needed. In particular, the levels of exposure of humans to such chemicals must be reliably established.

8.8 1.

References WEARNE S J, GEM G DE M, HARRISON N, COLLIER P P, FAIRWEATHER F,

and WALTON H ‘Prioritisation scheme to identify manufactured organic chemicals as potential contaminants of food’, Environmental Science and Pollution Research, 1996 3(2) 83–88. FIELDING M, FRANKLIN A, STARTIN J R, TREGUNNO, R J

Environmental organic contaminants in food 2. 3. 4. 5. 6. 7. 8. 9. 10.

11. 12. 13. 14.

15. 16. 17.

187

and SHIU W Y ‘Generic models for evaluating the regional fate of chemicals’, Chemosphere, 1992 24(6) 695–717. NIELSEN I R and HOWE P D (1991) Environmental Hazard Assessment: Toluene. TSD/1. Building Research Establishment, Watford, UK. NIELSEN I R, REA J D and HOWE P D (1991) Environmental Hazard Assessment: Benzene. TSD/4. Building Research Establishment, Watford, UK. CROOKES M J and HOWE P D (1992) Environmental Hazard Assessment: Ethylbenzene. TSD/7. Building Research Establishment, Watford, UK. CROOKES M J, DOBSON S and HOWE P D (1993) Environmental Hazard Assessment: Xylenes. TSD/12. Building Research Establishment, Watford, UK. NIELSEN I R, DIMENT J and DOBSON S (1994) Environmental Hazard Assessment: Cumene. TDS/20. Building Research Establishment, Watford, UK. GAVIN N, BROOKE D N, HOWE P D and DOBSON S (1996) Environmental Hazard Assessment: Naphthalene. TSD/27. Building Research Establishment, Watford, UK. GROB K, FRAUENFELDER C and ARTHO A ‘Uptake by foods of tetrachloroethylene, trichloroethylene, toluene, and benzene from air’, Z. Lebensm. Unters. Forsch. 1990 191, 435–441. DEPARTMENT OF THE ENVIRONMENT, ‘Benzene’, 1991 Annual Report of the Committees on Toxicity, Mutagenicity and Carcinogenicity of Chemicals in Food, Consumer Products and the Environment, pp. 45– 46, HMSO, London, 1992. WALLACE L A ‘Major sources of benzene exposure’, Environmental Health Perspectives, 1989 82 165–169. HOWARD P H in Handbook of Fate and Exposure Data for Organic Chemicals. Volume II – Solvents. Lewis Publishers, Chelsea, Michigan, 1990. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, ‘Benzene and other aromatic hydrocarbons in food – average UK dietary intakes’, Food Surveillance Information Sheet No. 58, 1995. HOLLIDAY M G and PARK J M ‘Exposure of Canadians to benzene’, unpublished document by Michael Holliday and Associates prepared for the Environmental Health Directorate, Department of National Health and Welfare, Ottawa, 1989. ¨ WE S, TASCHAN H and BRUNN H ‘Benzene, toluene and other alkyl STU benzenes in foodstuffs from central Hesse’, Lebensmittelchemie, 1991 45 116. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Hydrocarbons in foods from shops in petrol stations and stalls or shops in busy roads’, Food Surveillance Information Sheet No. 98, 1996. HALL, M, and GROVER P L, (1990). ‘Polycyclic aromatic hydrocarbons: metabolism, activation, and tumor initiation’, pp. 327–372. In C S COOPER MACKAY D, PATERSON S

188

18. 19. 20.

21. 22. 23. 24. 25. 26. 27.

28.

29. 30. 31.

Food chemical safety and P L GROVER (eds), Chemical Carcinogenesis and Mutagenesis I. Handbook of Experimental Pharmacology, vol. 94/I. Springer Verlag, New York. CERNIGLIA C ‘Biodegradation of polycyclic aromatic hydrocarbons’, Biodegradation, 1992 3 351–368. BOSSERT I D and BARTHA R ‘Structure biodegradability relationships of polycyclic aromatic hydrocarbons in soil’, Bulletin of Environmental Contamination and Toxicology, 1986 37 490–495. HEITKAMP M A and CERNIGLIA C E ‘The effects of chemical structure and exposure on the microbial degradation of polycyclic aromatic hydrocarbons in freshwater and estuarine ecosystems’, Environmental Toxicology and Chemistry, 1987 6 535–546. JONES K C, STRATFORD J A, WATERHOUSE K S and VOGT N B ‘Organic contaminants in Welsh soils: polynuclear aromatic hydrocarbons’, Environmental Science and Technology, 1989 23 540–550. GRIMMER G (eds) Environmental Carcinogens: Polycyclic Aromatic Hydrocarbons, CRC Press, Boca Raton, 1983. IARC 1972–1990 Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 1–49, International Agency for Research in Cancer, Lyons, France, 1992. DENNIS M J, MASSEY R C, MCWEENY D J, KNOWLES M E and WATSON D ‘Analysis of polycyclic aromatic hydrocarbons in UK total diets’, Food and Chemical Toxicology, 1983 21 569–574. DE VOS R H, VAN DOKKUM W, SCHOUTEN A and DE JONG-BERKHOUT P ‘Polycyclic aromatic hydrocarbons in Dutch total diet samples (1984– 1986)’, Food and Chemical Toxicology, 1990 28 263–268. LODOVICI M, DOLARA P, CASALINI C, CIAPELLANO S and TESTOLIN G ‘Polycyclic aromatic hydrocarbons contamination in the Italian diet’, Food Additives and Contaminants 1995 12(5) 703–713. DENNIS M J, MASSEY R C, CRIPPS G, VENN I, HOWARTH N and LEE G ‘Factors affecting the polycyclic aromatic hydrocarbon content of cereals, fats and other food products’, Food Additives and Contaminants, 1991 8(4) 517– 530. ROTH M J, STRICKLAND K L, WANG G Q, ROTHMAN N, GREENBERG A and DAWSEY S M ‘High levels of carcinogenic polycyclic aromatic hydrocarbons from Linxian, China may contribute to that region’s high incidence of oesophageal cancer’, European Journal of Cancer, 1998 34(5) 757–758. VOUTSA D, SAMARA C ‘Dietary intake of trace elements and polycyclic aromatic hydrocarbons via vegetables grown in an industrial Greek area’, The Science of the Total Environment, 1998 218 203–216. LI L D, LU F Z and ZHANG S W ‘Analysis of cancer mortality rates and distribution in China: 1990–92’, Chinese Journal of Oncology 1982 18(6) 403–407. KANG D H, ROTHMAN N and POIRIER M C ‘Interindividual differences in the

Environmental organic contaminants in food

32.

33. 34. 35. 36.

189

concentration of 1–hydroxypyrene-glucuronide in urine and polycyclic aromatic hydrocarbon-DNA adducts in peripheral white blood cells after charbroiled beef consumption’, Carcinogenesis, 1995 16(5) 1079–1085. ROTHMAN N, POIRIER M C and BASER M E ‘Formation of polycyclic aromatic hydrocarbon-DNA adducts in peripheral white blood cells during consumption of charcoal-broiled beef’, Carcinogenesis, 1990 11(7) 1242– 1243. The Control of Pollution (Supply and Use of Injurious Substances) Regulations 1986. (S.I. 1986 No. 902) London, The Stationery Office, 1986. The Environmental Protection (Disposal of Polychlorinated Biphenyls and other Dangerous Substances) (England and Wales) Regulations (S.I. 2000 No. 1043) London, The Stationery Office, 2000. SKENE S A, DEWHURST I C and GREENBERG M ‘Polychlorinated dibenzo-pdioxins and polychlorinated dibenzofurans. The risks to human health – A review’, Human Toxicology, 1989 8(3) 173–204. NORTH ATLANTIC TREATY ORGANISATION, COMMITTEE ON THE

‘International Toxicity Equivalency Factor (I-TEF) method of risk assessment for complex mixtures of dioxins and related compounds. Pilot study on international information exchange on dioxins and related compounds’. CCMS Report Number 176, Environmental Protection Agency, Washington D.C., 1988. VAN DEN BERG M ‘Toxic Equivalency Factors (TEFs) for PCBs, PCDDs and PCDFs for humans and wildlife’, Environmental Health Perspectives, 1998 106 775–792. CHALLENGES OF MODERN SOCIETY.

37. 38.

AHLBORG U G, BECKING G C, BIRNBAUM L S, BROUWER A, DERKS H G M, FEELEY M, GOLOR G, HANBERG A, LARSEN J C, LIEM A K D, SAFE S H, ¨ NHEIKKI E ‘Toxic equivalency SCHLATTER C, WÆRN F, YOUNES M and YRJA

39.

40. 41. 42. 43.

factors for dioxin-like PCBs: Report on a WHO-ECEH and IPCS consultation, December 1993’, Chemosphere, 1994 28 1049–1067. WORLD HEALTH ORGANISATION, REGIONAL OFFICE FOR EUROPE. ‘Summary Report. Consultation on Tolerable Daily Intake from food of PCDDs and PCDFs, Bilthoven, Netherlands, 4–7 December 1990’. EUR/ICP/PCS 030(S) 0369n. Regional Office for Europe, Copenhagen 1991. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Dioxins in food’, Food Surveillance Paper No. 31, The Stationery Office, London, 1992. Statement by the Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment on the health hazards of polychlorinated biphenyls (1997). VAN LEEUWEN F X R ‘Dioxins: WHO’s Tolerable Daily Intake (TDI) revisited’, Chemosphere, 2000 40 1095–1101. INTERNATIONAL COUNCIL FOR EXPLORATION OF THE SEAS ‘PCBs in the marine environment – an overview. Report of the Advisory Committee on Marine Pollution’, ICES Co-operative Research Report, 1982 112 (Annex 4) 43–50.

190

Food chemical safety

44.

‘Dioxins and PCBs in the UK diet: 1997 Total Diet samples’, Food Surveillance Information Sheet No. 4/00, 2000. ZANOTTO E ‘PCDD/Fs in Venetian foods and a quantitative assessment of dietary intake’, Organohalogen Compounds, 1999 44 13–17. BECHER G ‘Dietary exposure and human body burden of dioxins and dioxin-like PCBs in Norway’, Organohalogen Compounds, 1998 38 79– 82. JIMENEZ B ‘Estimated intake of PCDDs, PCDFs and co-planar PCBs in individuals from Madrid (Spain) eating an average diet’, Chemosphere, 1996 33 1465–1474. BUCKLAND S J ‘Concentrations of PCDDs, PCDFs and PCBs in New Zealand retail foods and assessment of dietary exposure’, Organohalogen Compounds, 1998 38 71–74. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Dioxins and PCBs in UK and imported marine fish’, Food Surveillance Information Sheet No. 184, 1999. ¨ CHER A and ANDERSEN L S ‘Time trend studies HILBERT G, CEDERBERG T, BU of chlorinated pesticides, PCBs and dioxins in Danish human milk’, Paper presented at 16th Symposium on Chlorinated Dioxins and Related Compounds, Amsterdam, 12–16 August 1996. POLDER A, BECHER G, SAVINOVA T N and SKAARE J U ‘Dioxins, PCBs and some chlorinated pesticides in human milk from the Kela Peninsula, Russia’, Paper presented at 16th Symposium on Chlorinated Dioxins and Related Compounds, Amsterdam, 12–16 August 1996. MALISCH R ‘Dioxin-like PCBs in food and breast milk samples’, Paper presented at 16th Symposium on Chlorinated Dioxins and Related Compounds, Amsterdam, 12–16 August 1996. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Dioxins and PCBs in cows’ milk from farms close to industrial sites: – Huddersfield 1997’, Food Surveillance Information Sheet No. 135, 1997. PEATTIE M E, LINDSAY D G and HOODLESS R A ‘Dietary exposure of man to chlorinated benzenes in the United Kingdom’, The Science of the Total Environment 1984 34 73–86. WANG M J and JONES K C ‘Uptake of chlorobenzenes by carrots from spiked and sewage sludge-amended soil’, Environmental Science & Technology, 1994 28 1260–1267. WANG M J and JONES K C ‘Occurrence of chlorobenzenes in nine United Kingdom retail vegetables’, Journal of Agricultural Food Chemistry, 1994 42 2322–2328. DE BOER J ‘Organochlorine compounds and bromodiphenylethers in livers of Atlantic cod (Gadus morhua) from the North Sea 1977–1987’, Chemosphere, 1989 18 2131–2140. HAFFNER G D, TOMEZAK M and LAZAR R ‘Organic contaminant exposure in the Lake St. Clair food web’, Hydrobiologia, 1994 281 19–27. OFSTAD E B, LUNDE G, MARTINSEN K and RYGG B ‘Chlorinated aromatic

45. 46. 47. 48. 49. 50.

51.

52. 53. 54. 55. 56. 57. 58. 59.

FOOD STANDARDS AGENCY

Environmental organic contaminants in food

60.

61. 62.

63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74.

191

hydrocarbons in fish from an area polluted by industrial effluent’, The Science of the Total Environment, 1978 10 219–230. KYPKE-HUTTER K, VOGELGESANG J, MALISCH R, BINNEMAN P and WETZLAR H ‘The origin of contamination of fish from the River Neckar with hexachlorobenzene octachlorostyrene and pentachlorobenzene: formation in an industrial process. 1. The course of the contamination in the upper section of the River Neckar’, Z. Lebensm. Unters. Forsch., 1986 182 464–470. JAN J, ZUPANCIC-KRALJ L and ZIGON D ‘Residue profile of PCB, PCN, and CBz in fish, algae, moss and sediment from the River Krupa (Slovenia)’, Toxicological and Environmental Chemistry, 1994 43 235–243. DAVIES D and MES J ‘Comparison of residue levels of some organochlorine compounds in breast milk of the general and indigenous Canadian populations’, Bulletin of Environmental Contamination and Toxicology, 1990 39 743–749. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Chlorobenzenes in food’, Food Surveillance Information Sheet No. 141, 1998. INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY ‘Chlorobenzenes other than hexachlorobenzene’, Environmental Health Criteria 128, World Health Organization, Geneva. 1991. VERSCHUEREN K Handbook of Environmental Data on Organic Chemicals, 2nd Edition, Van Nostrand Reinhold, New York, 1983. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Hexachlorobenzene in cows’ milk from farms close to industrial sites’, Food Surveillance Information Sheet No. 114, 1997. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Tetrachloroethylene in butter and lard’, Food Surveillance Information Sheet No. 5, 1993. INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY ‘Diethylhexyl phthalate’, Environmental Health Criteria 131, World Health Organization, Geneva, 1992. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Phthalates in food’, Food Surveillance Information Sheet No. 82, 1996. MINISTRY OF AGRICULTURE, FISHERIES AND FOOD ‘Phthalates in infant formulae – follow-up survey’, Food Surveillance Information Sheet 168, 1998. SHARPE R M and SKAKKEBAEK N E ‘Are oestrogens involved in falling sperm counts and disorders of the male reproductive tract?’, Lancet, 1993 341 1392–1395. NATIONAL RESEARCH COUNCIL Hormonally Active Agents in the Environment, National Academy Press, Washington, 1999. IUPAC ‘Natural and anthropogenic environmental oestrogens – the scientific basis for risk assessment’, Pure and Applied Chemistry, 1998 70(9) 1617–1865. COLBORN T, MYERS J P and DUMANOSKI D Our Stolen Future: How Manmade Chemicals are Threatening our Fertility, Intelligence and Survival, Little Brown, London, 1996.

192

Food chemical safety

75.

LIONE A

76.

‘Polychlorinated biphenyls and reproduction’, Reproductive Toxicology, 1988 2 83–89. ROMKES M, PISKORSKA-PLISZCZYNSKA J and SAFE S ‘Effects of 2,3,7,8– tetrachloro-p-dioxin on hepatic and uterine estrogen receptor levels in rats’, Toxicology and Applied Pharmacology, 1987 87 306–314.

9 Chemical migration from food packaging L. Castle, Ministry of Agriculture, Fisheries and Food, York

9.1

Introduction

9.1.1 Food packaging The modern packaging industry can be traced back as far as 1810 when the French pharmacist Nicolas Appert invented the canning process. Modern civilisation requires a continuous and reliable supply of safe and high-quality food. Since unprotected food is liable to deteriorate rapidly, it is necessary to provide appropriate protection and, although many other protective methodologies are currently in use, a prime feature of modern food protection is packaging. 9.1.2 Migration from packaging Packaging is the most important, and most obvious, example of materials and articles intended to come into contact with food. Although food packaging is the most common example, there are many other situations where materials are deliberately used in contact with food during manufacture, transport, storage, preparation and consumption. These include the materials used to construct storage vessels, conveyor belts, tubing, food preparation surfaces, and cooking and eating utensils. Food and beverages can be very chemically aggressive milieux and may interact strongly with materials that they touch. Collectively, they are as good as many of the solvents used in a chemistry laboratory. For example, food acids can corrode metals, fats and oils can swell and leach plastics, and beverages can disintegrate unprotected paper and cartonboard. In fact, no food contact material is completely inert and so it is possible for their chemical constituents to

194

Food chemical safety

‘migrate’ into the packaged food. Metals, glass, ceramics, plastics, rubber and paper can all release minute amounts of their chemical constituents when they touch certain types of foods. This release of chemicals to the food is known technically as migration. Migration can be defined scientifically as ‘the mass transfer from an external source into food by sub-microscopic processes’. More colloquial terms used are ‘leaching’, ‘bleeding’ or ‘leaking’ of substances from the packaging into the food. Note, the term food is used throughout this chapter to include beverages also. Food contact materials in general can be classified into ten main categories. As even a cursory glance around any supermarket will confirm the first four categories in this list dominate the packaging of our food in the developed world. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

plastics, including varnishes and coatings paper and board metals and alloys glass regenerated cellulose ceramics elastomers and rubbers paraffin waxes and micro-crystalline waxes wood, including cork textile products.

9.1.3

Why is understanding and controlling chemical migration important? Any chemical migration into food is important because it can have two impacts on the food: 1. 2.

Food safety – some substances used to manufacture packaging materials could be harmful if they migrated to the food and were ingested in a large enough quantity. Food quality – migrating substances may impart taint or odour to the food and so reduce consumer appeal.

Chemical migration from packaging is not an inconsequential process. For some food-package combinations the concentration of chemical migrants in the food can approach that of substances used as direct food additives, at levels of tens of parts-per-million (mg substance per kg of food, mg/kg). For obvious reasons, legislation has concerned itself mainly with the safety implications of any chemical migration and we shall do likewise here for this volume on food chemical safety. Specifically not discussed here are food quality issues such as organoleptic properties and prolongation of shelf-life. The quality of food is clearly of importance, but on this the consumer is empowered and needs less help and protection from the rulemakers and legislators.

Chemical migration from food packaging

195

All parties involved in the production, transport, selling and consumption of foods need to be aware of the potential for chemical migration. Each needs to ensure that packaging materials are correctly specified for the intended application so that there is no excessive chemical migration. The chain of care involves: • primary manufacturers of the raw materials; e.g. polymer and paper manufacturers • converters who turn the raw material into packaging for food use • vendors of the materials, e.g. retailers of articles, supply companies • users of the material – the food packer • food retailers • enforcement authorities • consumers – with respect to the proper use both of pre-packaged foods and of materials and articles used in the home.

9.2

Chemical migration and the main factors that control it

9.2.1 The physico-chemical basis of migration Migration was defined above as ‘the mass transfer from an external source into food by sub-microscopic processes’. Migration of chemical substances is a diffusion process subject to both kinetic and thermodynamic control and is described by diffusion mathematics as derived from Fick’s Law. The mathematics describe the diffusion process as a function of time, temperature, thickness of the material, amount of migrant in the material, partition coefficient and distribution coefficient. The kinetic dimension of migration dictates how fast the process of migration occurs. The thermodynamic dimension dictates how extensively the transfer of substances will be when migration is finished – or more properly when the system is at equilibrium. The kinetic and thermodynamic aspects should not be confused. Just because a migration process is fast does not necessarily mean that it will also be extensive with high concentrations resulting in the food. Similarly, migration may proceed at a slow rate but, if the chemical migrant has a higher affinity for the food than for the packaging material, then given enough time (e.g. a long shelf-life) it may still migrate extensively into the food. Being a molecular diffusion process, chemical migration is subject to the normal laws of physics. There are several determinants of chemical migration and exactly what migration occurs depends first on the identity and concentration of any chemicals present in the packaging material. Other important parameters are the nature of the food along with the conditions of contact. Lastly, the intrinsic properties of the packaging material itself are important considerations. If it interacts strongly with the food it could give high migration by leaching. Conversely, an inert material with low diffusivity is likely to give low migration values. It is important to understand the factors that

196

Food chemical safety

control chemical migration, because from this understanding springs the ability to prevent or limit any undesirable migration into foods. 9.2.2 Composition of the packaging material The packaging material is the source of any chemical migration. Both the rate and the extent of any migration depends first on the concentration of the putative migrant in the packaging. As a general rule, migration is proportional to the starting concentration in the packaging. It has been suggested that for some combinations of substances and materials, a so-called threshold effect operates such that below a minimum concentration in the packaging then no migration occurs. This has been suggested for vinyl chloride monomer (VCM) migrating out of polyvinylchloride (PVC), for example. If such a threshold effect really does operate, it would require that a certain fraction of the VCM is locked into the PVC matrix and is immobilised so that it cannot migrate. Any VCM above the critical concentration in the PVC would, in contrast, exist in more accessible locations in the polymer matrix and so be available for migration. It is difficult to explain such a postulate in terms of polymer structure and molecule structure interactions. It may be that for some substance-packaging combinations there is not a strictly linear relationship between packaging composition and propensity to migrate. It may also be the case that some purported threshold effects have more to do with the difficulty of measuring migration near to the limit of detection of analytical methods. The general case is that if a substance is not present in a packaging material then it cannot migrate (self-evident but easily forgotten!) and if a substance is present in the packaging then migration levels increase as the concentration in the packaging increases. 9.2.3 The nature and extent of contact The nature and extent of any contact between the packaging and the food is the next important parameter to consider. Most chemicals need direct contact between the food and the packaging if migration is to occur. The exceptions are chemicals that have a significant vapour pressure at the temperature at which the packaging is used. These chemicals may then migrate through an air gap even if the food does not touch the package directly. The extent of contact is also important and it can be expressed as the ratio of food weight to contact area. An example is the situation where a film is used to pack a solid food like potato crisps which make only point-contact with a limited area of the film. Clearly, in this case the potential for migration is less than if the same film was used in more intimate contact with a food – for example if it were used to vacuum-pack a food item. Even if the food contact is intimate with a liquid or semi-solid food, there can still be large differences in the food mass to contact area ratio. Common examples would be an individual portion pack of margarine or butter (say, 7 g in contact with 28 cm2, or 4000 cm2/kg) compared with a catering pack of the margarine or butter (say, 2 kg in contact with

Chemical migration from food packaging

197

1050 cm2 or 525 cm2/kg). If the same plastic was used to make the two pack sizes (polystyrene or polypropylene would be candidates) then the same migration on a unit area basis would give rise to an eight-fold higher concentration of any migrating substance in the individual portion pack compared with the catering pack. The most extreme examples of this surface area : food mass ratio are to be found outside the area of general packaging materials; for example, the relatively limited contact made by small gaskets used in a large food processing plant, gloves or a conveyor belt used to handle tonnes of food in a packing plant, or tubing used to handle tens or hundreds of thousands of litres of liquid during its service life. A third factor that determines the nature and extent of any contact with the food is the presence of a barrier layer. If the chemical that may migrate is located in one layer of the packaging material but this is separated from the food by an intervening layer, then this barrier layer may retard or prevent migration from occurring. This is quite a common situation with modern multi-laminate packaging materials where inks, adhesives, or one or more of the laminate plys do not touch the food directly. The packaging industry is very aware of the benefits provided by barrier layers in protecting the food product from air, light and moisture, in controlling the inner atmosphere of the pack (MAP and CAP, modified and controlled atmosphere packs), and in retaining desirable aromas whilst protecting the packed food from undesirable odour pick-up. The same physico-chemical laws that endow these quality benefits can also be used to prevent or limit chemical migration that could otherwise contaminate the food. For example, a thin barrier layer of virgin polyethylene terephthalate (PET) can prevent any detectable migration from recycled PET incorporated into multilayer bottles used for soft drinks. 9.2.4 The nature of the food The nature of the food that touches the packaging is important for two reasons. First, if the packaging is not compatible with that type of food then there can be a strong interaction leading to an accelerated release of chemical substances. Examples are the interaction of fats and oils with certain plastics leading to swelling of the plastic and leaching of substances from that plastic. Leaching, formally known as Class III migration, occurs because the diffusivity of the plastic increases with any swelling. An even more extreme example of an undesirable interaction between packaging and food is the corrosion of uncoated metal surfaces leading to high metal release into certain acidic foods. It is important to avoid such obvious mis-matches and ensure that packaging materials are compatible with the food that it is intended to pack. These effects notwithstanding, even if the packaging does not suffer from such an obvious strong interaction with the food, the nature of the food still has a pronounced influence on chemical migration because it determines the solubility of any packaging chemical in the food and so influences the amount of migration that may occur. Foods can be conveniently classified into five

198

Food chemical safety

Table 9.1 Nature of the food in contact

Nature of chemicals most likely to migrate

Acidic foods, aqueous foods and low alcohol beverages Fatty foods, distilled spirits

Polar organic chemicals, salts, metals

Dry foods

Non-polar, lipophilic (‘fat-loving’) organic substances Low molecular weight, volatile substances

categories; aqueous, acidic, alcoholic, fatty, and dry. It is generally accepted that acidic foods will cause migration similar to that caused by aqueous and lowalcohol food and beverages, and that acidic foods are generally more aggressive. It is also generally accepted that fatty foods cause migration similar to highalcohol beverages. Consequently, the three main drivers of migration with respect to food type, can be characterised by the type of substances that have a high affinity for them and so tend to migrate more readily. The three main categories are summarised in Table 9.1. 9.2.5 The temperature of contact The migration of chemicals is like virtually all chemical and physical processes in that it is accelerated by heat. So more migration will occur if the temperature is raised. Packaging materials are increasingly used under a very wide range of temperature conditions, ranging from deep frozen storage through refrigerated and ambient temperature storage, to boiling, sterilisation, microwaving and even baking in the pack. A material suitable for one particular application may not necessarily be suitable for another. 9.2.6 The duration of contact Materials suitable for short-duration contact may not be suitable for longer service times. The kinetics of migration are, to a first approximation, first-order in that the extent of migration increases according to the square-root of the time of contact; M/t1/2. The time (duration) of contact for common packaging can vary enormously and the performance requirements of the material must be specified accordingly: • • • • • •

seconds (food handling) minutes (take-away foods) hours (fresh bakery, sandwiches) days (fresh milk, meat, fruit and vegetables) weeks (butter, cheese) months and years (frozen foods, dry goods, canned foods, drinks).

Chemical migration from food packaging

199

9.2.7 Mobility of the chemicals in the packaging The value of the diffusion coefficient, DP, for migration of a substance, is a function of the material type, the molecular size of the substance, and the temperature. There are many possibilities for modelling this function. Currently, there exists only a limited number of reliable DP-values due to the enormous requirements needed for the experimental determination. However, the molecular size of the migrant can, to a first approximation, be equated to its molecular mass if the influence of molecular shape is ignored. In turn, the diffusion coefficient DP can be calculated from the molecular weight and from an empirical parameter that describes the intrinsic permeability (diffusivity) of the packaging material. The development of numerical models using finite element analysis to calculate and predict migration levels is advancing at a rapid rate; most especially for plastics. Modelling holds great promise as a means to reduce the burden and expense of measuring chemical migration by experimental means. Further, when reliable migration models come available for different packaging materials, it will be possible to put specifications on the composition of packaging materials such that any migration limits in regulations are not exceeded during the actual and foreseeable conditions of that material. 9.2.8 Practical examples – plastic usage As examples of the general considerations detailed above, with the interplay between conditions of use and the intrinsic properties of materials, Table 9.2 gives a general guide to the suitability of some common packaging plastics. The applications are listed in order of severity – as you move down the table the applications place greater demands on the plastic and so more inert plastics are needed. Thus, for example, foods high in sugar or fats (e.g. jams, meat) can be more demanding than aqueous foods (e.g. coffee, vegetables) since they tend to interact more strongly with plastics and also they can get hotter if heated in a microwave oven. This table is a guide only and each case much be considered individually. In every case the plastic must be of ‘food grade’ and should be manufactured using only permitted ingredients.

9.3

The range and sources of chemicals in food packaging that pose a potential risk

Substances present in food packaging materials can originate from a number of sources: • known ingredients used to make the basic packaging materials of plastics, paper, coated and uncoated metals, and glass; for example monomers and additives in plastics, or chemicals used in paper-making • known or unknown isomers, impurities and transformation products of these known ingredients

200

Food chemical safety

Table 9.2 Application

p-PVC PE

PS

PP

PVDC

PVC

PET

General use (e.g. frozen, refrigerated and room temperature storage) Limited contact with hot foods (e.g. take-away) – aqueous foods Limited contact with hot foods (e.g. take-away) – fatty foods Microwave oven use – aqueous foods not high in sugar or fat Microwave oven use – high sugar and fatty foods Use in conventional ovens





























?

?































?

?

?

















Key:  = likely to be suitable. ? = needs careful consideration.  = likely to be unsuitable. PE = Polyethylene (‘polythene’) including high-density and low-density PE (HDPE and LDPE) PP = Polypropylene PS = Polystyrene including expanded polystyrene (EPS) and acrylonitrile-butadiene-styrene (ABS) copolymers PET = Polyethylene terephthalate PVC = Polyvinyl chloride – both unplasticised (PVC) and plasticised (p-PVC) PVDC = Polyvinylidene chloride

• chemicals used to convert or fabricate the basic packaging material into its finished form; for example, inks and adhesives • unknown contaminants in the raw materials used, and especially those in the feedstock if materials are recycled.

9.3.1 Ingredients used to make the basic packaging materials A tremendous range and variety of chemicals are used to make modern packaging materials. The European inventory list of chemicals used to make plastics intended for food contact numbered more than 1500 listed substances and inventory lists of a similar length exist for chemicals used to make paper, can coatings, inks and adhesives. The chemicals on the plastics inventory list run literally from A (acetic acid) to X (xylene) with more than 1500 chemicals in between. Substances needed to make effective plastics packaging materials include: • • • •

monomers and other starting substances catalysts solvents and suspension media additives – antioxidants – antistatic agents

Chemical migration from food packaging – – – – – –

201

antifogging agents slip additives plasticisers heat stabilisers nucleating agents dyes and pigments.

The inventory list of chemicals used to make plastics is comprehensive because plastics are of synthetic origin and their composition is known more or less completely (with the caveat about transformation products, see later). In contrast, paper as a natural material is less well defined and characterised and so any inventory list of chemicals used in paper-making will be only an incomplete list of chemicals that may be present in the paper or cartonboard packaging and that therefore may migrate. 9.3.2

Impurities and transformation products formed during manufacturing processes Transformation products formed during manufacturing processes present quite a different problem from chemicals that are used deliberately and that will appear on an inventory list. Transformation products could derive from parent substances that may not themselves be present as residues in the finished packaging, and the formation and fate of the transformation products may be difficult to predict. For example, plastics intended for food contact use may contain residues derived from aids to polymerisation. These polymerisation aids are defined as substances necessary for the manufacture of polymers, and include the catalysts and initiators. They are typified by the organic azo and peroxy initiators which are unlikely to be present in the polymer in their initial state since they are purposely decomposed to radicals at elevated temperatures so that they can participate in the polymerisation reaction. The high temperatures subsequently used for processing the polymer and moulding the plastic into packaging materials and articles can only further decompose these initiators. However, substances such as their decomposition and recombination products, may be present in the finished plastic and may migrate to foods. As a second example, there is a wide variety of breakdown products and oligomeric products that may be formed from the reactive monomers that are the building blocks of plastics. For plastics, the general assumption has been that any side-reaction products and breakdown products are likely to be significantly less toxic than the monomers, and so restricting the migration of the monomer was accepted as an indirect way to limit any hazard from the oligomers also. Whilst this approach is probably acceptable for addition polymers, such as those made from the unsaturated monomers vinyl chloride, butadiene and acrylonitrile where the unsaturated monomer is far more noxious than their products, the validity of this means of indirect control is questionable for condensation polymers such as polyesters and for polyethers formed from epoxide monomers.

202

Food chemical safety

Whilst it is self-evident that any controls should apply to those species that may migrate, rather than simply those chemicals that are used at the outset to make the packaging material, the problem of reaction products is particularly difficult and has not yet been solved. Control can be either directly on the substances per se, or indirectly through restrictions on their precursors if the likelihood of formation is well understood. One approach is to control migration through limits based on the total content of certain functional groups and this approach is taken in a Council of Europe (CoE) Resolution on aids to polymerisation which specifies limits on metal residues and other chemical moieties, as considered appropriate. 9.3.3 Inks and adhesives used by converters to make packages Inks and adhesives are two examples of materials used to make packaging, which are not intended to touch the food directly, but from which migration can nevertheless occur in some circumstances. A printing ink typically consists of the following: colourant (pigment or dye), resin, solvent and various additives including plasticisers for flexibility, slip agents to prevent scuffing, and antioxidants to protect the ink during processing and curing. Printing inks, overlacquers and varnishes are normally applied to the outside of food packaging materials and there is a generally held view that inks applied to the outside then stay on the outside. Consequently, most inks are formulated using substances that do not have specific approval for direct food contact applications under the existing approval schemes such as Directive 90/128 in the EU, the USFDA code, or the German BgVV recommendations. There is evidence, however, that paper, cartonboard, and even many plastic films, are in fact rather permeable and volatile low molecular weight substances can diffuse through them over time. It is likely therefore that residues of these substances must be considered to make direct contact with the packaged food. A large fraction of the substances used in inks will not permeate through the printed substrate because their molecular size is large and so diffusion is slow. These substances will not come into contact with (and possibly contaminate) the packaged food unless set-off occurs when the printed surface touches the foodcontact surface, such as when films are stored on reels or when printed containers are stored stacked or nested. Of course, printers constantly inspect the quality of their work and are alert to the possibility of set-off. But visual examination will detect only the coloured pigments and dyes. The uncoloured, invisible, components of inks go undetected and there is a need to minimise any set-off of these invisible components onto the food contact surface of the packaging material. There is also an increasing use of printing inks applied with or without a coating to the inner surfaces of packaging such that the ink touches the food directly. Some manufacturers do supply inks specially formulated for direct contact printing. However, given the technical requirements of printing inks it is difficult to see how all of the components used in an ink can be replaced with direct food additives.

Chemical migration from food packaging

203

Adhesives are used to fabricate rigid packs from cartonboard. They are also used to seal flexible packaging to form wrappers and pouches and for gluing plastic-plastic, plastic-paper and plastic-paper-foil laminates. Commercial adhesives are used because they are effective, but their detailed chemical composition is poorly catalogued. Within the broad categories of contact and reactive adhesives, including coldseals, hot-melt glues, epoxy resins and isocyanates (polyurethanes), manufacturers produce an array of formulations each tailored to suit a particular end-use. Unlike bulk plastics and paper where unwanted substances can be removed by vacuum stripping (e.g. vinylchloride monomer from polyvinylchloride, styrene from polystyrene) or by washing (e.g. organic and metallic residues in mass-polymerised plastics), adhesives by their ‘gummy’ nature are difficult to clean-up. Residues of incomplete polymerisation and reaction by-products could be effectively retained and may subsequently migrate. On the other hand, adhesives are generally not used in direct contact with the packaged foods. Rather, they are applied at seams and pack ends and any contact with the food is likely to be incidental and limited in area. 9.3.4

Contaminants including those potentially present in recycled materials The recycling of plastics or paper into food packaging materials introduces the potential inclusion of adventitious contaminants. Since by definition the recycling of plastics or paper involves destruction of the original material or article with co-mingling on a large scale, it is the potential build-up of chronic toxic contaminants that is of greatest concern rather than acutely toxic contaminants. It will never be possible, however, to test each and every batch of recycled packaging for all potential chemical contaminants. If one accepts this premise but still requires food contact use of recovered plastics, then safety assurance must take another approach. Source control is a key element for recycling. Additional factors are how effectively any input of contaminated feedstock may be cleaned or diluted in a large-scale recycling process A further option is to use a functional barrier – either of virgin material or of another material entirely – as an integral layer which under normal and foreseeable conditions of use reduces all possible material transfers to foodstuffs to a quantity that should not endanger human health. Guideline notes exist in the USA and in Europe. The essence of these guidelines is that, due to the impossibility of defining the exact composition of recycled plastics and paper, they should not generally be used in direct contact with food but should be used only as a buried layer protected by virgin polymer or by some other functional barrier. Exceptions to this general rule could be where a well-controlled closed-loop cycle is employed, such as the re-cycling of plastic beverage bottles back into beverage bottles, where contact is with only one food type and there is practically no cross-contamination with other polymers. Other exceptions could include the packaging of foods that have a low

204

Food chemical safety

tendency to elicit migration, including dry foods, frozen foods, and foods with a protective skin or shell. 9.3.5 Recycled fibres – a difficult problem Whereas there has been a lot of research and discussion about the potential use of recycled plastics for food contact, in fact the use of recycled paper for food packaging far exceeds the use of recycled plastics. For example, a large proportion of the feedstock for paper and board production in Europe, is recycled fibres. Figures available for the individual EU member states indicate between 25 and 75% use of recycled fibre. The figures for direct and indirect contact with food are not known and they will of course be lower than these figures for all applications. Nevertheless, the volume of recycled fibres going into food packaging is very significant – as any examination of the things we buy as consumers will confirm. The easiest way to recognise that packaging has been made using recovered fibres is when the food producer and packer declare this on the packaging. If the packaging is not labelled in this way then one telltale sign can be the presence of specks of inks that have not been removed completely during the recycling process. This is not always a reliable indication, however, since the recovered source material may not have been heavily inked or because any ink residues may not be detectable. The most reliable indication of recycled fibres is a microscopic examination which usually reveals that recycled paper contains distinguishable fibres from a variety of different wood sources – because of co-mingling and mixing of collected papers on a large scale – whereas virgin paper will contain just one or relatively few different wood types. Unless the source material is carefully controlled and the clean-up stages are efficient, then as discussed above recycling carries the risk of adventitious contamination of packaging materials. For example, it is well known that polychlorinated biphenyls (PCBs) can be contaminants from the historical use of carbonless copy paper. Following discontinuation of this practice some 30 years ago, levels of PCBs in recycled paper have now declined such that there is no detectable migration into packaged foods. More recently, there has been a world-wide problem with the presence of diisopropylnaphthalenes (DIPNs) in food packages. DIPNs are widely used for ink-jet printers and as solvents used in the preparation of special papers such as carbonless and thermal copy paper. Not all of the DIPNs may be removed by the treatment of the recycled fibres. Some may be present in the finished board and thus can under some circumstances migrate into food. DIPNs have been detected in a range of food packaged in recycled cartonboard including even dry foods like rice and pasta. Additional toxicological information is needed on DIPNs and, in the meantime, industry has been instructed to ensure that levels of DIPNs in food packaging made from recycled paper and board are kept as low as reasonably practicable to minimise migration into food. These examples serve to illustrate a general point, that understanding the clean-up requirements of paper recycling and linking these

Chemical migration from food packaging

205

with the severity of the intended food packaging application, is an urgent research need.

9.4

Research on health issues

The risk assessment process involves describing the toxicological hazard profile of a chemical substance, using qualitative and quantitative data, and coupling this to an estimate of exposure to assess any risk. Consequently, the information that is required on packaging chemicals comprises: (a) toxicity data; and, (b) exposure data. 9.4.1 The biological properties of the substance With some caveats, the toxicological information required is the same as for other chemical contaminants in our food, irrespective of their source. Chronic effects are the main concern – i.e. low-level long-term exposure to migrating substances. One exception is the migration of tin from tinplate steel into canned tomato products where high tin concentrations in food may cause short-term stomach upsets in some people but without any lasting harm. As a general principle, the greater the extent of migration into food, the more toxicological information will be required. For example, to establish the toxicological profile of packaging chemicals the Scientific Committee for Food, which is an independent group of experts that gives advice to the Commission of the EU, looks for information on the following essential core set of tests: • a 90-day oral study • three mutagenicity studies: – a test for gene mutations in bacteria – a test for chromosomal aberrations in cultured mammalian cells – a test for gene mutations in cultured mammalian cells • ADME studies (adsorption, distribution, metabolism and excretion) • data on reproduction • data on teratogenicity • data on long-term toxicity/carcinogenicity • data on the potential for accumulation.

Where migration is above 5 mg/kg of food, all the core studies are required. Where migration is lower, in the range of 0.05 to 5 mg/kg of food, then a limited data set is required comprising information on the accumulation potential, mutagenic potential and the 90-day oral toxicity study. Where migration is lower than 0.05 mg/kg of food, then a reduced data set is called for which is limited to demonstrating the absence of mutagenic potential in the three mutagenicity tests. This last requirement, simply to demonstrate the absence of mutagenic potential if the migration of a substance is below 0.05 mg/kg into foods, is similar in principle to the US-FDA threshold of regulation approach. If a substance is not a

206

Food chemical safety

carcinogen and if calculations can show that it does not migrate to more than 0.0005 mg/kg in a model US diet, then no further information is required on that substance in packaging materials. 9.4.2 Consumer exposure The second part of any risk assessment requires an estimate of consumer exposure to the migrating substance. This depends on a number of factors: • the types of packaging materials that contain the substance • the fraction of each packaging material that contains the substance and quantities of the substance incorporated • the types of food packaged • the proportion of each type of food that is packaged in each type of packaging material • the length of contact of the foods with the materials, the unit weight of food in relation to the surface area of packaging and temperatures encountered while food is in contact with the material • the extent of migration of the substance or of its breakdown products into each type of food and its possible reactions with food components • the quantities of foods consumed that have been in contact with each of the packaging materials containing the substance • the frequency with which food containing the substance or its breakdown products or its reaction products with food is consumed • the period over which food containing the substance is consumed.

Different countries have this information to a greater or lesser extent. Some have almost everything known and other have almost nothing known. In the absence of detailed statistics then conservative assumptions of intake are made in order to protect consumers.

9.5

Regulatory context

9.5.1 Historical context In one of the first comprehensive schemes to regulate the use of food contact materials, the US Food and Drug Administration (FDA) issued legal provisions during the 1950s in the sector of plastics. This was closely followed by German and Italian regulations in the field of migration. These regulations were designed to avoid excessive release into food of substances contained in the materials, especially in plastics, and rule out the possibility of a health hazard to the consumer. French, Dutch and Belgian authorities later issued similar laws. During 1972 the Commission of the European Communities drew up a broad programme of action designed to harmonise existing laws. The Commission drew up a Framework Directive setting out the underlying principles, listing the materials to be regulated, and defining the procedures and criteria to be used in

Chemical migration from food packaging

207

adopting specific regulations for each type of material. As community-wide rules are enacted in the EU then the rules of the individual members states, e.g. the BgVV Recommendations in Germany and the Warenvet in the Netherlands, are gradually superseded. 9.5.2 General principles All regulations on migration from food packaging materials are part of the risk management process which follows the risk assessment process given above; describing the toxicological hazard profile of a chemical substance using qualitative and quantitative data, and coupling this to an estimate of exposure to assess any risk. Control measures are then introduced if needed, as part of the risk management process to ensure that exposure remains (or is lowered to) within acceptable limits. Because foodstuffs themselves are difficult to analyse and have a variable composition, testing packaging materials for migration almost invariably uses model foods, known as ‘food simulants’, which are designed to mimic the main classes of foodstuffs. There are some minor differences in detail between EU and US-FDA requirements but the essential elements are listed in Table 9.3. Neither European nor USA regulations specifically require the testing of packages used for dry foods that have no free fat or oil phase on the surface. The general assumption is that with no intimate contact with a liquid food phase, the potential for migration is negligible. The consequence of this rulemaking, applied to plastics in the EU and to packaging materials in general in the USA, is that packers and producers take the cue that dry foods are less demanding applications. There is, however, mounting evidence that certain substances, and especially mobile, low molecular weight ingredients and impurities, can migrate rather freely to dry foods. This topic needs some consideration by the rulemakers. The aim of migration tests using simulants is to anticipate levels of migration expected to foods. Conventional test conditions are specified in the different regulation schemes to correspond to typical food packaging applications. The exposure conditions of time and temperature are the most important parameters along with considerations of the effect of symmetry if, for instance, just one side of an assymmetrical laminate is the food contact surface. If a food simulant is more aggressive than the food it is intended to model, with respect to eliciting Table 9.3 Class of food

Typical simulants used

Aqueous foods Acidic foods pH < 4.5 Alcoholic foods Fatty foods

Distilled water Dilute (3%) acetic acid solution in water Ethanol solutions in water Olive oil, heptane, 95% ethanol, Miglyol

208

Food chemical safety

chemical migration, then either the time and temperature conditions of exposure can be adjusted accordingly or a so-called reduction factor can be applied to the test result using simulant to bring the test result down to that expected for the food. Testing with food simulants also allows accelerated tests to be performed over a short time which is especially useful when packing foods that have a long shelf life and waiting for several months to learn the test result is not often acceptable. 9.5.3 EU directives on chemical migration In the European Union (EU), the legal foundations governing materials and articles intended to come into contact with food are contained in the Framework Directive 89/109/EEC, which states in Article 2: Materials and articles must be manufactured in compliance with good manufacturing practice so that, under their normal or foreseeable conditions of use, they do not transfer their constituents into foodstuffs in quantities which could: • endanger human health • bring about an unacceptable change in the composition of the foodstuffs or a deterioration in the organoleptic characteristics thereof.

The EU has taken a sectorial approach to meet the general aims of the Framework Directive by introducing individual directives for different types of materials. This approach has been adopted because rules and test procedures can be tailored to suit the different materials best. Thus, individual directives have already been drawn up to implement the Framework Directive for specific types of food contact materials. These include detailed regulations for plastics, ceramics and RCF (regenerated cellulose film). For paper and board intended for food packaging, there is as yet no specific EU Directive although the Council of Europe Committee of Experts on Materials Coming into Contact with Food (CoE) has been working for many years now on a draft resolution. There is an interesting contrast between the directives for the three types of material listed in Table 9.4. For RCF the directive gives a list of permitted substances – a so-called positive list – that can be used to make RCF and control is largely in the form of compositional limits. There are just two migration limits: for monoethylene and diethylene glycol. The plastics directive also has a positive list but here there are relatively few compositional rules and control is exercised rather through numerous migration limits. In the ceramics directive, there is no positive list but rather a negative list with limits on the release of lead and cadmium. As depicted in Table 9.4, it has been suggested that any future directives on paper and board, on rubber and on metals could follow these three paradigms. Up-to-date details can be found on the excellent EU Food Contact Materials Resource Centre web-page listed in section 9.8.

Chemical migration from food packaging

209

Table 9.4

#

#

#

Framework directive 89/109/EEC Regenerate cellulose directive 93/10/EEC

Plastics directive 90/128/EEC

Ceramics directive 84/500/EEC

Paper and board (planned)

Rubber (planned)

Metals (planned)

9.5.4 US-FDA rules The US-FDA rules on chemicals that may migrate from food packaging are very different in character from the European rules. Migrating chemicals that may reasonably be expected to become a component of foods are classed as indirect food additives. The FDA system is very prescriptive and requires a lot of information to be supplied at the petition stage. As in the EU, this naturally includes toxicological information on the substance(s) used along with indicative migration data. A major difference though is that a petition is very specific to a particular application – for example, a petition would request to use substance A in polymer B, used to pack food C under storage conditions of D days at E degrees Fahrenheit. If the petition was accepted, this permission would be given – no more and no less. If in future it was desired to use the same substance but in an application not covered by the original, then a new petition would be needed. This contrasts strongly with the EU system in which, for plastics at least, once a migration limit is issued then it applies to all uses of that substance. A particular strength of the US system is that very specific information can be provided at the petition stage and this allows detailed estimates of possible consumer exposure to be made and also allows some fairly pragmatic end-tests to be proposed to ensure that future production falls within the specifications of the materials tested for migration in the original petition. One draw-back of the US system is that it cannot be enforced experimentally by any third party – although compliance can be checked via documentation. The US authorities have been very progressive in introducing a threshold approach to regulation (designed to prevent spending time and resources on the trivial) and also have introduced recently a system of pre-market notification to help speed up the approvals process. Up-to-date details can be found on the excellent FDA web-page listed in section 9.8.

9.6

Migration testing

9.6.1 What sets it apart from any other field of contaminant analysis? The term ‘migration testing’ is almost universally associated with tests using food simulants (model foods) rather than testing for migration into foods.

210

Food chemical safety

Testing for chemical migration from packaging intended for food is conducted in two distinct steps. First, the food contact material is exposed to the food simulant under conditions that are carefully selected and controlled. Second, the simulant is analysed to determine the level of any migration. It is the first step that sets migration testing apart and establishes it as a distinct discipline. Whilst the analysis of chemical contaminants at parts per million (ppm, mg/kg) or parts per billion (ppb, g/kg) is a demanding task, the problems faced and the analytical methods available as solutions, are not unique to food contact materials but are common to the analysis of all food, biological or environmental matrices for trace constituents or contaminants. In the somewhat esoteric discipline of testing food contact materials, a major challenge lies in the initial choice and execution of the migration exposure rather than in the subsequent chemical analysis. 9.6.2 The purpose of migration testing The purpose of testing food packaging materials is to ensure that food packed therein remains wholesome and safe. As described above, packaging materials may contain several types of substances that can migrate to the food. These include known ingredients; known or unknown isomers, impurities, reaction products and breakdown products of these ingredients; and unknown contaminants. These are the targets for the analytical chemist given the task of assessing the suitability of the material from a migration viewpoint. 9.6.3 The users of migration test data There are a number of stakeholders that need information on migration. Each group has different requirements and starts with a different level of knowledge of the material under scrutiny. This dictates the approach that the analytical chemist must take in order to fulfil their requirements. These users include manufacturers of the basic packaging material who know the full list of ingredients used and the manufacturing history of the material. Provided they use only substances on approved lists, in principle they need only ensure that these substances are of the required purity, that the material is made according to good manufacturing practice, and that any compositional limits, extraction limits or migration limits imposed on any of the ingredients, are not exceeded. As one moves down the chain of use, to the converter, the packer and finally to the retailer, full disclosure of this compositional information and manufacturing history can be difficult to obtain and verify. These users will require a compliance statement from the primary manufacturer but will need also to conduct migration tests themselves to ensure that the conversion process and conditions of actual use, do not cause unacceptable migration. At the end of the chain, acting for the consumer at the point of sale, an enforcement laboratory can be presented with a sample (perhaps an import) already in contact with food and for which a priori none of this compositional

Chemical migration from food packaging

211

and manufacturing information is available, and which must all be gathered by analysis. Here, the use of food simulants may not be possible or even desirable – the real interest is in the migration levels in the foods themselves as eaten. 9.6.4 The migration data required There are two general principles regarding chemical migration, the purity of the foodstuff and the inertness of the material. The first principle, the purity of the foodstuff, is the raison d’eˆtre for the control of specific substances on the basis of known or suspected toxicity. Control can take the form of limitations on the quantity of substance permitted to migrate to foods or food simulants. Alternatively, control can be indirectly via a limit on the quantity of the substance permitted to be present in the finished material or article. These are termed migration limits and compositional limits respectively. The second principle underlying migration testing can be considered the subject of tests for overall migration. This has also been known as ‘global migration’ in the EU. Where the test with a food simulant is replaced with an extraction solvent – not intended to mimic a food but to give a more exhaustive extraction – overall migration equates to ‘total extractable matter’. The test for overall migration has its roots in limiting adulteration of the food. The rationale is that above a certain level even the migration of innocuous substances should not be tolerated since it affects the quality of the food. This upper limit – a ceiling on the total quantity of all substances permitted to migrate – is based on a gravimetric measurement with no requirement to identify the substances migrating. The analytical chemist is therefore required to measure at times: overall migration, total extractables, the composition of the material with regard to both approved ingredients and unauthorised contaminants, and the migration of these ingredients and contaminants to foods or food simulants. The limits placed on each measured parameter are defined by regulations and recommendations and the methodology used will, of course, be dictated by these requirements. The following gives an indication of the limits of interest in each case. These are illustrative only and the exact control limits will vary from one regulatory system to another. 9.6.5 Approved ingredients in the material or article A compositional analysis is very straightforward. Some methods for direct in situ analysis are known but, more generally, a complete extraction of the target substance(s) from the packaging matrix using solvent is required followed by chemical analysis of the extract. 9.6.6 Migration of approved ingredients Limits on the migration of specific substances of toxicological concern span the concentration range from 0.01 mg/kg to about 60 mg/kg. The lower limits are

212

Food chemical safety

applied to the most toxic ingredients, e.g. residues of the reactive monomers vinyl chloride and acrylonitrile, whereas the higher limits are more common for additives used in the packaging material to confer desirable technological properties, e.g. plasticisers, antioxidants and stabilisers. The upper limit in the EU is the overall migration limit of 60 mg/kg – above 60 mg/kg specific migration measurements are not required (for non-volatile substances at least) since the material would fail an overall migration test. The methodology of testing packaging articles with food simulants is quite prescriptive and no useful purpose is served in reproducing it here. The key stages are as follows: • • • • • •

Step 1. Take a representative sample. Step 2. Choose appropriate simulant (model food). Step 3. Choose appropriate test conditions (exposure time and temperature). Step 4. Choose appropriate exposure configuration. Step 5. Conduct and monitor the exposure. Step 6. Analyse the simulant for specific substances or for the total (‘overall’, ‘global’) migrate.

9.6.7 Contaminants in the material or article The explicit control of named contaminants requires that the possibility of their presence has been recognised and recorded in, for example, a negative list. Almost by definition, the majority of contaminants are not of this nature. Analysis for contaminants in packaging materials and for any migration of same, is difficult and almost open-ended. For this reason, much emphasis is needed on Good Manufacturing Practice and in record-keeping. 9.6.8 Overall migration and extractables The gravimetric determination of the residue from overall migration tests or extraction tests is at the milligram scale. As mentioned above, the EU limit for overall migration is 60 mg/kg. With some caveats, measuring at these levels is relatively simple in even the most basic laboratory and official methods are well described.

9.7

Case studies

Three case studies have been selected in order to draw out some of the principles and considerations described above. They are: (i) the migration of a residual monomer from a can coating; (ii) the migration of an additive from a plasticised plastic film; (iii) the migration of a photoinitiator used in a printing ink.

Chemical migration from food packaging 9.7.1

213

Case study 1: migration of residual TPA and IPA monomer from can coatings Terephthalic acid (benzene 1,4-dicarboxylate, TPA) and isophthalic acid (benzene 1,3-dicarboxylate, IPA) are reactive bi-functional acids used as monomers to make plastics and coatings. TPA is a starting material for polyethylene terephthalate (PET). TPA and IPA are used to make polyester lacquers and coatings including the internal coatings of food cans. Industry has carried out work on the effective use of these substances in can coatings. TPA and IPA may be present in can coatings either because they are in resins that are used to make coatings, or because of cross-contamination of coating materials. Liquid coatings that may contain TPA and IPA are applied to the internal surface of some metal cans, usually by a roller. The solvent in the coating is evaporated at a moderately low temperature, before the coating is cured at a temperature of approximately 200ºC. The coated sheet or coil of metal is then used to make the can. Stripes of PET are used to cover the welded joint in the side of some cans. These side stripes are made by applying powdered PET, by a roller or an electrostatic process, to the welded joint on the inside of the can. The can is then heated to fuse the PET and form the side stripe. After the coating or coating and side stripe have been applied the can is then filled with food or beverage, sealed and the contents sterilised, typically at 121ºC for up to an hour with the product either static, or rotated to assist heat transfer. The UK Food Standards Agency reported a study in 2000 that was conducted to investigate if TPA and IPA could migrate in these applications. The first phase of sampling was to identify which products were sold in cans with coatings made from polyester resin. This illustrates the first rule of migration from packaging that is self-evident but bears repeating – if a substance is not present in the original packaging material then it is pointless testing that packaging for the migration of that substance. Considerable time and money can be saved on analysing materials before food or food simulants themselves are tested. In the second phase of the investigation, samples of products packaged in cans with such coatings were purchased and the can contents analysed for the presence of TPA and IPA. Twenty-eight of fifty-four samples purchased were found to have coatings made from polyester resins and so their contents were analysed for TPA and IPA. Analysis was by GC-MS after extraction and derivatisation to form the dimethyl esters of TPA and IPA. Included in the analytical quality assurance steps taken was participation in a check-sample exercise whereby four samples were supplied by a second laboratory and analysed blind. The performance of the testing laboratory was assessed from the accuracy and precision of their measurements and was demonstrated to be acceptable. TPA was found in three of the twenty-eight samples, at or just above the limit of quantification and in seven samples at levels between the limit of detection and limit of quantification (0.2 and 0.7 mg/kg respectively). The highest concentration of TPA observed was 0.8 mg/kg. Seven of the ten food samples

214

Food chemical safety

containing TPA were from cans coated with PET. The other three samples were from cans that contained a PET side stripe. IPA was detected in four of the twenty-eight samples at levels between the limit of detection and limit of quantification. The highest concentration of IPA observed was less than 0.7 mg/ kg. All the samples containing IPA were from cans with internal surfaces coated with a polyester resin other than PET. The investigation provided evidence of limited migration of both TPA and IPA. The UK Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment (COT) advised that the levels of TPA and IPA found in foods in this survey were not of concern for public health based on available information. Nevertheless, they recommended that further studies be carried out for TPA and IPA to see if they have endocrine disrupter activity. This case study is interesting for a number of reasons. The migration concentrations of TPA and IPA in the foods tested were slightly higher than might have been anticipated from the published literature on migration from can coatings. Testing coatings that are stoved to achieve a full cure can be problematic and the migration results obtained are very dependent on the exact stoving conditions of high temperature and short duration. It is also fairly common practice to prepare test specimens of coatings applied to an inert substrate and test in sheet form rather than finished cans, for convenience of migration testing. Surveys of this type, in addition to protecting the consumer, also provide invaluable data on migration into foods themselves and this can be used to ensure that simulants are modelling foodstuffs reliably. 9.7.2 Case study 2: DEHA migration from plasticised PVC films DEHA (di-2-ethylhexyl adipate) is widely used as a plasticiser in the production of plastic materials and especially PVC cling films. The plasticiser imparts flexibility and stretchability to the films along with the property of ‘cling’ (tackiness; the sticking both to itself and to overwrapped containers that users find convenient). In the period 1986–89 there were several investigations made in the UK of plasticiser migration into foods from packaging materials, which included DEHA along with phthalates, citrates, epoxidised soybean oil, sebacates and phosphates. The dietary intake of DEHA was estimated then at around 16 mg/day for a high consumer. Levels were not a cause for concern but the studies showed markedly higher migration when the plasticised film was used in cooking in microwave or conventional ovens. Manufacturers consequently agreed to provide instructions for appropriate use on the packaging of the cling films. During this time they also conducted a programme of work to reduce the amount of DEHA used to plasticise films by making thinner films (‘down-gauging’) and by using alternative, so-called ‘polymeric’, plasticisers as partial or total replacements for DEHA. The alternatives were also organic esters like DEHA but they had a higher molecular weight and migrated less readily. By 1990 and following the introduction of labelling guidelines and reformulation of films, maximum intake of DEHA had fallen to about half of

Chemical migration from food packaging

215

the earlier estimate, to around 8.2 mg/person/day. Information had also been obtained about the migration levels of the polymeric plasticisers used as alternatives and a mathematical model had been established along with empirical rules for estimating migration of plasticisers other than DEHA. At that time, the maximum intake figure of 8.2 mg/day was not considered to be a hazard to health but it was considered prudent to reduce intakes further. In particular, advice was given to consumers that the plasticised films should not be used in direct contact with foods having a high fat content, such as cheese. In 1991 the UK Committee on Toxicology (COT) set a tolerable daily intake value (TDI) of 0.3 mg/kg body weight/day for DEHA. One year later, in 1992, a urinary biomarker study was reported for DEHA in a limited population exercise in the UK. A skewed distribution was determined with a median value of 2.7 mg/day and this confirmed by an independent route, the earlier estimates of DEHA intake made using dietary survey data. The COT opinion of 1991 on DEHA was followed in 1994 when both the COT and the SCF set a TDI for epoxidised soybean oil (ESBO, a heat-stabiliser and secondary plasticiser used in PVC clingfilms) of 1 mg/kg bw/day weight and the SCF confirmed the COT TDI for DEHA of 0.3 mg/kg bw/day. For the first time this gave concrete figures against which to assess dietary exposure. This safety clearance on two substances which were major ingredients in the manufacture of cling film, DEHA and ESBO, meant that there were now cling films available that could be used to wrap high-fat foods. The advice on cling film, originally issued in 1990, was revised in 1995 to: • Do not use cling films where they may melt into the food, such as in conventional ovens or with pots and pans on cooker hobs. • When reheating or cooking food in a microwave oven ensure that the cling film does not touch the food. • Only use in contact with high-fat foods when the manufacturer’s advice states it is suitable for this. Examples of high-fat foods include some types of cheese, raw meats with a layer of fat; fried meats, pastry products, and cakes with butter icing or chocolate coatings.

At the time of writing, DEHA has not yet been included in the EU positive list of additives used to make food contact plastics. With an established TDI of 0.3 mg/kg bw/day and using the normal conservative assumptions of 1 kg food consumed daily and a 60 kg bw, this TDI would give a specific migration limit of 18 mg DEHA per kg food or food simulant. It must be recognised that noncompliance with a migration limit of 18 mg/kg based on this convention does not necessarily mean that the TDI could be exceeded. It seems unlikely that DEHA will be listed until the issue of food consumption factors is resolved. This case study is interesting because it exemplifies many of the key factors that determine migration and that were described in sections 9.2 and 9.4 – the effect of food type and of contact temperature in driving migration, along with the need for concrete numerical limits based on toxicological information coupled with realistic estimates of exposure.

216

Food chemical safety

9.7.3 Case study 3: migration of benzophenone from printed cartonboard Benzophenone is used as a photoinitiator in the curing of inks with ultra-violet (UV) light. UV-cure inks contain typically 5–10 per cent photoinitiator. Only a small portion of the initiator is used up during the curing process. Benzophenone can therefore remain in the printed material. The use of UV-cure inks for printing cartonboard has become widespread because the fast cure permits online cutting and folding, enabling rapid production of finished packaging. Research has shown that benzophenone can be present in printed cartonboard packages at up to 0.7 mg/100 cm2 and can migrate to foods packaged in this material even during frozen storage, up to 0.4 mg/kg. For frozen foods in microwavable packaging, the extent of migration into the food during storage and subsequent heating was shown to increase further, with levels of benzophenone after heating ranging up to 1.0 mg/kg. In a UK survey reported in 2000, 350 samples of food packaged in printed cartonboard were purchased from retail outlets. The packaging was analysed for residues of benzophenone, using solvent extraction and GC-MS analysis. One hundred and forty-three packaging samples contained measurable benzophenone. Seven of the 143 samples were in the concentration range 0.05 to 0.2 mg/100 cm2, 60 of the 143 were in the range from 0.2 to 0.8 mg/100 cm2 and the remaining 76 packaging samples contained benzophenone in the range from 0.8 to 3.3 mg/100 cm2. Seventy-one samples were then selected at random from the 143 packaging samples that contained benzophenone, and the food contained in these 71 packages was analysed for benzophenone. Benzophenone was detected in 51 of the 71 food samples. Twenty-nine samples contained benzophenone at 0.05 to 0.5 mg/kg, 17 samples were in the range of 0.5 to 5 mg/ kg and three exceeded 5 mg/kg. The highest level was 7.3 mg/kg. Two samples contained levels that were detectable but not quantifiable. When there was no benzophenone in the packaging there was none detected in the food. However, there was not a clear numerical correlation between levels of benzophenone in cartonboard and food. Differences in the composition of the foods tested may have been important in this respect, along with other factors such as the storage time, the storage temperature, the extent of any direct contact made between the food and the package, and the presence or absence of any barrier material between the printed cartonboard and the food. The results confirmed that benzophenone can migrate from printed cartonboard to food. The estimated intake of this substance was far less than the TDI of 0.01 mg/kg bw/day set by the SCF. Using a 97.5 percentile value for daily consumption for the foods that were tested (102 g), a mean value of bodyweight of 60 kg and a mean contaminant level, the intake was estimated at 0.001 mg/kg bw/day (one-tenth of the TDI) and no health effects would be expected in an individual’s lifetime from the levels found in the survey. In additional work, benzophenone was found to migrate even from polyethylene-coated board and this was attributed in part to the fact that polyethylene is rather permeable to low molecular weight substances and is not a good barrier to migration. Additional studies were conducted with two other

Chemical migration from food packaging

217

frozen foods by incorporating model ink components into their cartonboard packaging. The substances were benzophenone, benzylbutyl phthalate, butyl benzoate, chlorodecane and dimethyl phthalate. Migration was readily detected after storing the food at 20ºC for one week in the impregnated cartonboard. Migration levels in the food increased only slowly thereafter, up to one year, except for the most volatile substances used. Subsequent heating of the frozen food using a microwave oven decreased the concentration of some migrants, by volatilisation, but increased the concentration of other, less volatile substances, by heat-accelerated migration. It was concluded that for inks used to print food contact materials, if the content of low molecular weight volatiles is controlled and if transfer of higher molecular weight components via set-off is controlled also, then migration levels could be kept low in these conditions. This case study is interesting for a number of reasons. It emphasises that the potential for migration to dry and frozen foods cannot be ignored. It also illustrates that barrier layers may be ineffective in preventing migration. Although the inks were applied to the outside of the cartonboard, the board itself presented little if any barrier to migration and even some thin plastic films were inadequate barriers if made from high permeability plastics like polyethylene. Lastly, it shows the benefits of having extensive food consumption statistics that enable reliable, upper-bound estimates of consumer exposure to be made, thus placing migration data firmly in the context of what matters – consumer protection.

9.8

Suggested further reading

Migration from Food Contact Materials. L. L. Katan (ed.), Blackie Academic and Professional, Glasgow, UK, 1996. ‘Migration of packaging components to foods: Regulatory considerations’. C. V. Breder, in Food Packaging Interactions, J. H. Hotchkiss (ed.), ACS Symposium Series 365, 1988, pp. 159–169. Food Packaging Migration and Legislation. R. Ashby, I. Cooper, S. Harvey and P. Tice. Revised edition, Pira International, Leatherhead, UK, 1997. Plastic Packaging Materials for Food: Barrier Function, Mass Transport, Quality Assurance and Legislation. O. Piringer and A. L. Baner (eds), Wiley-VCH, Weinheim, FRG, 2000. FDA webpage on indirect food additives at http://vm.cfsan.fda.gov/~lrd/ foodadd.html EU Food Contact Materials Resource Centre, at http://cpf.jrc.it/webpack/ Analysis of Food Contaminants. J. Gilbert (ed.), Elsevier Applied Science Publishers, Barking, UK, 1984. Food and Drink Laboratory Accreditation: A Practical Approach. S. Wilson and G. Weir, Chapman & Hall, London, UK, 1995.

10 Pesticides I. Shaw and R. Vannoort, Institute of Environmental Science and Research, Christchurch

10.1

Introduction

Pesticide residues in food in the developed world are extremely unlikely to ever have caused acute illness. Their levels are so low and relatively infrequent that it would be impossible to consume sufficient food to attain an intake of pesticides that even approached toxic levels. The chronic effects of pesticide residues, in particular mixed residues, is very much more difficult to assess in terms of biological effects on the consumer. We know very little about the synergistic or additive toxicological effects of pesticide residues, but we do know that in every meal that we consume it is very likely that more than one pesticide will be present, albeit at a vanishingly low concentration. These ideas are confirmed by the recently completed New Zealand Total Diet Survey (Cressey et al., 2000) which shows that the most commonly encountered pesticide residue is chlopyriphos-methyl and that in the average New Zealand diet this reaches approximately 1.2% of the Acceptable Daily Intake (ADI – see later for definition). If, however, the intakes of the five most prevalent organophosphorus (OP) pesticides (Table 10.1) are calculated in terms of their percentage ADIs and added together, the total is still only approximately 3.4% of the combined ADI. It is questionable whether it is appropriate to sum ADIs. However, since the example given here comprises only OPs and they all have the same mechanism of toxicity (i.e. inhibition of synaptosomal acetylcholine esterase), this method does have a genuine toxicological basis. The point remains that the total intake is unlikely to have any toxicological effect on the consumer. Despite good evidence that suggests that pesticide residues in food present a very low risk to the consumer, when we recently asked shoppers in a supermarket in the UK whether they were concerned about pesticides residues in

Pesticides

219

Table 10.1 Organophosphorus pesticide intakes calculated as % ADI in the New Zealand diet. Data from the 1997/98 New Zealand Total Diet Survey (Cressey et al., 2000) Pesticide

% ADI

Chlorpyriphos Chloryriphos-methyl Dimethoate Fenitrothion Pirimiphos-methyl TOTAL

0.2 1.2 0.8 0.2 1.0 3.4

their food they all said, Yes! When asked about natural toxins, no one knew what we meant. This is an interesting paradox because it is likely that the natural toxins in food present a greater risk to the consumer than pesticide residues. Psoralens (furocoumarins) (Fig. 10.1), a group of naturally occurring pesticides in parsnips, celery, parsley and related vegetables are photoactivated carcinogens often present at 10s or 100s mg.kg 1 (Table 10.2). They are present in every parsnip, stick of celery and leaf of parsley that we eat. On the other hand pesticide residues are not omnipresent and their residues are generally at exceptionally low concentrations (i.e. <0.1 mg.kg 1). Perhaps the press should make an issue of this hidden risk rather than focusing on other food related issues. Continuing with the psoralen example, psoralens must present a greater risk to the consumer of parsnips than any of the pesticide residues that are likely to be present in parsnips simply because psoralens are carcinogens. The testing necessary before a pesticide is approved for use in countries of the developed world makes it very unlikely indeed that a carcinogenic pesticide will be approved. It is important that specific risks (e.g. pesticides in food) are considered in the context of related unavoidable or accepted risks (e.g. natural toxins in food). Indeed, pesticides are likely to be very low in the hierarchy of life’s risks (Fig. 10.2).

Fig. 10.1

Structures of the linear furocoumarins: psoralen; xanthotoxin (8-MOP); and bergapten (5-MOP).

220

Food chemical safety

Table 10.2 Furocoumarins in fruit and vegetables. The levels in celery and parsnip are high (data from MAFF, 1996b) Family of species of plant Umbelliferae Celery

Parsnip Parsley

Coriander Rutacea Grapefruit Lemon Orange Lime

Moraceae Fig

Types of furocoumarins found

Levels of total furocoumarins (mg/kg) and main sources

psoralen bergapten xanthotoxin isopimpinellin 9-methoxypsoralen 4-methoxypsoralen 9,4-methoxypsoralen bergapten xanthotoxin isoimperatorin oxypeucedanin graveolone bergapten xanthotoxin coriandrin dihydrocoriandrin

1.3–46.7 25–100 0.65 – – – – 40–1740 2500 11.4–112 300 – – – 0.45 –

psoralen bergaptol bergapten bergamottin phellopterin 8-geranoxypsoralen

– – – 3000 – –

bergapten

480 620

stalk root seed

root diseased root fresh leaf dried leaf

fresh

lime oil

leaves sap

Toxic chemical intake must be quantified in order that comparisons between different chemicals (e.g. pesticides), diets and countries can be made. Most countries in the developed world conduct surveys which involve analysing food for pesticide residues. These values are then put in context by comparing them to benchmarks of toxicity (Acceptable Daily Intake – ADI) or with trading standards (Maximum Residue Level – MRL) which are set to ensure that countries exporting food do not export excessive pesticide residues with that food. 10.1.1 Acceptable Daily Intake (ADI) The ADI is a measure of toxicity. It is calculated from a known toxicological effect of a chemical (e.g. pesticide), for example increased serum glutamate pyruvate aminotransferase (SGPT) is often measured as an indicator of liver damage. The dose of the pesticide just below the point at which the defined

Pesticides

221

RISKS GETTING RUN OVER ON THE WAY TO THE SHOP Choking on a Brussels Sprout Natural toxins Pesticide residues Fig. 10.2

A light-hearted look at the potential risks of eating fruit and vegetables (adapted from Shaw, 1999).

toxicological effect occurs in animal studies is determined (No Observable Adverse Effect Level – NOAEL). This is the maximum dose that can be consumed without any adverse effect resulting. Assuming that an extrapolation can be made from the animal used in the NOAEL study (this is by no means certain and is the subject of many a heated discussion between toxicologists!), this is also the amount that a person can eat without any ill effects. The ADI is calculated by dividing the NOAEL by a safety factor (usually 1000) to account (in fact over-account) for differences in response to the test chemical between the animal used in the toxicity test and humans. The ADI thus represents a very low level of risk (Winter and Francis, 1997). Since the NOAEL studies are usually lifetime studies (e.g. two-year rat study), the ADI relates to consumption over an entire lifetime. If a pesticide residue level exceeds the ADI, this is the point at which toxicologists start to get worried and would seriously consider removing the offending food from the market. It is, however, very important to remember that the ADI exceedance over an entire lifetime is necessary to result in ill effect. This, of course, is very unlikely to happen with pesticide residues in food in any developed country. Nevertheless, an ADI exceedance is taken seriously. ADI exceedences are very rare indeed. In fact annual surveillance reports or total diet surveys from the USA, EU, New Zealand and Norway over the past two years have not reported a single ADI exceedance. The UK, however, had problems with lindane in milk (MAFF, 1996a) and chlormequat in pears (MAFF, 1999); in both cases the ADI was very closely approached and ministers took both cases very seriously indeed. In fact, the UK officials coined an interesting euphemism to describe this situation: ‘erosion of safety margins’.

222

Food chemical safety

10.1.2 Maximum Residue Level (MRL) The MRL is not a toxicological parameter, but rather a trading standard set by national and international authorities (e.g. Codex Alimentarius) to ensure that residues are controlled in world food trade. The MRL is the pesticide residue level in a particular food following its production according to Good Agricultural Practice (GAP). For example, if a carrot is grown properly and all of the pesticide applications are in compliance with GAP (i.e. applied at the right time and according to the label directions) the pesticide residues will not exceed the MRL. The MRL is determined by a smallscale farm study in which the pesticide under test is applied to the particular crop, the appropriate withdrawal period is allowed (i.e. the time necessary between application of pesticide and harvest), the crop is harvested and residues determined. The residue level is the MRL. If a residue level exceeds the MRL the crop has not been grown according to GAP and the product is not permitted to be sold, imported or exported. If the MRL is exceeded there is no health implication; however, the farmer producing the crop has breached national or international regulations and is liable to prosecution. MRL exceedances are far more common than ADI exceedances. In the most recent UK survey 1.2% of fruit and vegetables analysed as part of the survey exceeded MRLs (MAFF, 2000). Of these, three farmers were traced and subjected to court action. In the USA, MRL violations were 4.8% of samples analysed in 1996 (USDA, 1998). 10.1.3 Pesticide residues and health There are no examples of illness caused by pesticide residues in food in the developed world. The definition of illness in this context is generally accepted to be acute. This is hardly surprising because pesticide residues levels in food are exceptionally low – far below the levels that could result in an intake that would have an acute effect. This gives pesticides a very clean bill of health. However, very little is known about the long-term effect of pesticides and mixtures of pesticides in food. It is incredibly difficult to study these potential effects. In theory, the most commonly associated long-term toxic effect, namely cancer, is very unlikely to be caused by pesticide residues in food because pesticides must pass stringent toxicity tests before being approved for use. These tests include carcinogenicity studies in animals. No carcinogenic pesticide would ever be approved for use. Despite this there are pesticides that have been on the market for many years. Their ‘approval’ for marketing came long before the need for toxicity testing. Perhaps the best example of a pesticide in this class is lindane ( -hexachlorocyclohexane, -HCH). It has been suggested that lindane is associated with breast cancer, although the evidence for this is scant and highly subjective. In addition to this, pesticides used in the past, even if they are no longer in use, might remain in the environment and therefore contaminate food. The best example of this class is DDT. DDT’s heavy use in New Zealand in the 1950s and early 1960s has led to a significant environmental contamination

Pesticides

223

which, due to its incredibly long environmental half-life, will remain a problem for many years. New Zealand butter imported into the UK has been found to contain relatively high DDT breakdown products (e.g. ppDDE) (although not exceeding the MRL) (MAFF, 1997) and eggs surveyed in New Zealand have been found to contain traces of DDT and breakdown products (Cressey et al., 2000). Despite this, DDT in the New Zealand diet is declining (Cressey et al., 2000) (Fig. 10.3). It is difficult to explain this steady decline in terms of the biodegradation of DDT. However, it could be due to the steady decline in fatty food intake (DDT is lipid soluble and therefore will be taken in with dietary food) due to health education programmes in New Zealand. In addition, countries such as Indonesia and the People’s Republic of China still use DDT despite its worldwide ban and therefore foods from these countries are likely to contain DDT residues (e.g. rabbit imported into the UK from China (MAFF, 1997)). DDT is carcinogenic and therefore its residues in food add to the total dose of carcinogens in our diet. Further to the carcinogenic risk of long-term exposure to pesticide residues, the potential for endocrine disruption by pesticides that mimic hormones is becoming an emotive issue (Muller et al., 1995). There are a number of pesticides (e.g. DDT) that have structure activity relationships with steroid hormones. These molecules fit, albeit poorly, the cellular steroid hormone receptors (e.g. estrogen receptors (ER)) (Figs 10.4 and 10.5) and either activate (receptor agonists) or inhibit (receptor antagonists) the receptor. The estrogen receptor agonists are termed xenoestrogens and represent a broad class of environmental pollutants ranging from surfactive detergents (e.g. nonylphenol) to pesticides (e.g. DDT and its metabolite DDE) (Shaw and Chadwick, 1998). The pesticides have only a fraction of the activity of 17 -estradiol – for example DDT is only one millionth the activity of 17 -estradiol. Exposure to

Fig. 10.3 Estimated daily dietary exposure to total DDT for a young male (YM) in the 1997/98 New Zealand Total Diet Survey (NZTDS) compared to previous NZTDSs (adapted from Cressey et al., 2000).

224

Food chemical safety

Fig. 10.4

Estrogen receptor.

xenoestrogens (generally, not just pesticides) has been blamed for a number of human and animal effects, amongst which is the steadily declining human sperm count and the decrease in length of alligators’ penises in Laka Apopca in the USA. Calculation of the dietary intakes of estrogenic pesticides compared to the intakes of natural phytoestrogens present in soya and other legumes clearly shows that pesticides represent a drop in the ocean of dietary xenoestrogen intake (Table 10.3) (Shaw and McCully, 2000). This casts doubt upon the relevance of chronic exposure to pesticides as a contribution to declining sperm count and other hormone agonist-mediated effects in humans.

Fig. 10.5

Estrogen receptor with DDT in place.

Pesticides Table 10.3

Theoretical estrogen levels (data from Shaw and McCully, 2000)

Estrogen class/food Phytoestrogens Beans, peas, spinach Soya Plasticisers Pesticides

225

Exogenous estrogen

Theoretical plasma concentration (ng.dm 3)

Coumesterol Coumesterol Genistein/genistin Bisphenol-A Phthalates Total estrogenic pesticides*

11.5 3 143 0.1 0.003 0.005

*includes DDT, dicofol, endosulphan, dieldrin and b-HCH

10.2

Monitoring pesticides in food

Governments are keen to assure their national consumers and potential international importers of their food, that it is safe. From the point of view of pesticide residues this generally involves the operation of surveillance schemes. Such schemes comply with national legislation and involve the analysis of prescribed numbers of samples of food for selected pesticides on an annual basis. The number of samples analysed varies greatly from country to country. For example, in Europe (Table 10.4) the greatest number of samples analysed on a per capita population basis is in Sweden (1001  10 6 per capita in 1996 – total number of samples analysed = 8908) and the smallest is in the UK (15  10 6 per capita – total number of samples analysed = 878) (Shaw, 1999). Whether this reflects the individual country’s concern for consumer well-being is uncertain! Countries that rely on fruit and vegetable exports often appear towards the top of the samples analysed per capitum league table. For example, Holland analysed 706  10 6 samples per capita in 1996 (total number of samples analysed = 11015 (Shaw, 1999) and was second in the European league. Some countries do not carry out pesticide monitoring schemes (e.g. New Zealand), but instead opt for the total diet survey approach. 10.2.1 The New Zealand Total Diet Survey A distinguishing characteristic of Total Diet Surveys like the NZTDS (Cressey et al., 2000) is that foods are ‘analysed as normally consumed’ (e.g. bananas peeled, meat cooked etc.). They therefore frequently provide the most relevant means of assessing the consumer’s exposure. A Total Diet Survey is essentially a public health risk assessment tool, and not a compliance monitoring tool. In New Zealand, the 1997/98 NZTDS food list consisted of 114 different commodities, representing about 70% of those most commonly consumed by New Zealanders. Sixty-six of the foods were considered to be distributed nationally, and these were sampled from one city over two seasons; whereas the 48 regional foods

226

Food chemical safety

Table 10.4 Fruit and vegetable sampled in Europe in 1996 as part of member states’ surveillance programme (adapted from Shaw, 1999) Country

Population (106)

No. samples analysed

No. samples analysed per capita population (106)

Belgium Denmark Germany Greece Spain Ireland Italy Luxemburg Netherlands Portugal Finland Sweden Norway UK

10.2 5.3 83.5 10.5 39.2 3.6 57.5 0.42 15.6 9.9 5.1 8.9 4.4 58.5

932 1273 4257 1132 3022 505 7194 212 11015 600 3368 8908 2936 878

91 240 51 108 77 140 125 506 706 61 660 1001 667 15

were sampled from four different regions, also over two seasons. Each different foodstuff was analysed separately for pesticide residues, as well as other contaminants (e.g. Cadmium) and selected nutrient elements (e.g. Iodine). Fortnightly simulated typical diets, using the 114 foods of the 1997/98 NZTDS, were derived mainly from food frequency and 24-hour diet recall data, and developed for six different age–sex groups in NZ. From these fortnightly diets, the weight of each individual food item consumed was determined for each age–sex group. Exposures to pesticide residues from a particular food in the 1997/98 NZTDS were estimated by multiplying the mean concentration of the particular pesticide residue in the food by the amount of that food consumed. By adding together the contributions of all foods in the simulated diet, the total dietary exposure to the pesticide residue can be estimated for each age–sex group. This can then be compared to the ADI in order to assess the potential risk to the health of the consumer. The NZTDS thus provides a reliable snapshot of the overall quality and safety of the NZ food supply, and is a means (albeit not ideal) of checking the effectiveness of regulatory systems established to control pesticide residues in food. Total Diet Studies are also valuable in determining whether particular pesticide residues occur across the diet as a whole, or are restricted to certain food groups or even individual foods. The NZTDS provides readily understandable information on the dietary exposures of pesticide residues for the use of regulatory agencies, lawmakers and the public. In the 1997/98 NZTDS, some 2,440 food samples were collected. After compositing some samples of the same food type, 460 samples were actually screened, with 272 samples (59%) being found to contain detectable residues.

Pesticides

227

This is very similar to the frequency (56%) found in the 1990/91 NZTDS, but quite different from the average of about 30% found in other countries’ pesticide residue monitoring programmes (e.g. the UK). This difference highlights an important point. When looking at statistics from different surveys, it is crucial that they are of similar design, have robust sampling regimens, and use similar methods, limits of reporting and analytical quality control. The most relevant comparisons are thus between the same survey over time, and much less so between different surveys or countries. The 59% of samples with detectable pesticide residues in the 1997/98 NZTDS is higher than found in the compliance programmes of the UK and USA for two reasons. First, with constraints on resources always an issue, there is usually a need to balance effectively the range of pesticide residues screened for and the number of samples analysed. The sampling plan for the NZTDS was devised to look closely at foods that were more likely to contain pesticide residues. For example, previous NZTDSs had indicated that carbonated cola beverage was unlikely to contain pesticide residues and so only two composite samples were analysed. By contrast, bread often contains residues of organophosphorus pesticides and so eight samples of each of three different types of bread were analysed. Second, the limits of reporting needed for TDS studies are generally one to two orders of magnitude lower than for compliance monitoring programmes, so TDSs pick up more samples with residues. To illustrate this, it is worth noting that of the 272 food samples with detectable residues in the 1997/98 NZTDS, 160 would not have been detected with the higher limits of reporting employed in 1990/91. This means that while the percentage of samples with detectable pesticide residues had risen slightly in the 1997/98 NZTDS, the majority of the residue levels were significantly lower than those in the 1990/91 survey. Only 20 different pesticide residues were detected in the 1997/98 NZTDS out of a screen of 90. Of these, none of the pesticide residue levels detected exceeded the New Zealand Food Regulations 1984 MRL, where one was listed for the specific food item. Of approximately 29,000 individual analytical pesticide residue results in the 1997/98 NZTDS, only 397 (1.4%) were detectable residues. Many of these detectable residues might not have been detected in a compliance monitoring programme, with the often higher limits of reporting. 10.2.2 Pesticide monitoring vs. Total Diet Surveys Pesticide residue compliance monitoring programmes analyse individual foods ‘as purchased’ (i.e. bananas, including skin; meat, raw, etc.). Because a smaller range of foods are generally targeted, sample numbers per commodity are generally more robust than in a TDS. Any pesticide residues found are checked for compliance with the MRL for that commodity, which has been set to reflect GAP. Total Diet Surveys are often cheaper than full surveillance programmes. Countries with small populations and a consequent low tax revenue cannot

228

Food chemical safety

afford to set aside a considerable proportion of their tax revenue to monitor pesticides in food (this is hardly warranted when the low risk of pesticides to the consumer is considered). For example, the UK has a population of approximately 60,000,000 and spends about £2,000,000 on pesticide residue analysis each year. New Zealand, on the other hand, has a population of 3,500,000 and spends about NZ$1,000,000 (£220,000) spread over three years on its Total Diet Survey which covers both pesticide and elemental compounds. Countries clearly must get their food safety issues in perspective, and pesticide residues present a very low risk to the consumer. In this context, it is appropriate that New Zealand spends money on reducing its high incidence of Campylobacter food poisoning (Orchard et al., 2000). Monitoring pesticide residues is very unlikely to save lives; reducing Campylobacter contamination of food and water will. But the public continue to demand information on pesticide residue levels in the food supply. Not monitoring pesticide residues in food therefore represents a significant political risk.

10.3

High risk groups

It is important to investigate high risk consumers specifically. They might be in the high risk category because they consume huge amounts of a particular food or they are particularly susceptible to the toxicity of pesticides or a pesticide. One example is infants whose metabolism might not be fully developed and therefore they might be unable to detoxify and excrete toxic chemicals as well as adults. In addition, they consume more food per body weight than adults to provide the massive energy intake that they require to grow and develop. A second example is people who include a large amount of a particular food in their diets. Some people, for reasons that are difficult to understand, follow erratic diets based on a particular food or vegetable. If ‘their’ fruit or vegetable has pesticide residues they will be exposed to greater amounts of the pesticide than a ‘normal’ consumer. The question is whether this high risk group should be specifically taken account of. Indeed, in the UK such extremes are no longer considered as part of pesticide intake risk assessments. 10.3.1 Infants The tolerance of the newborn infant to environmental toxicants is not well understood, therefore it is prudent to assume that that the very young are more susceptible to toxic chemicals than adults. Indeed, exposure of the immature organs to pesticide residues can permanently alter the structure or function of an organ system (NRC, 1993) and therefore it is important to protect infants from the potential harm of pesticide residues in their food. Equally important is the fact that infants are dependent on a very narrow range of foods to provide their complete nutrition. The diversity of food intake in adults tends to protect against imbalances; this protective diversity is not

Pesticides

229

present during the early period of life, when the infant receives, as its sole source of nutrition, human milk or infant formula. Infants also consume more kilojoules per unit bodyweight than adults (NRC, 1993). With infant diets based on comparatively high consumption of a limited range of foods it is particularly important that these foods do not contain ‘excessive’ levels of pesticide residues. For these reasons, it has been stated that infant formulae products are the most highly regulated and controlled of all commercially available foods (Tanner and Barnett, 1986). The EU has recently introduced legislation limiting all pesticide residues to a maximum of 0.01 mg.kg 1 in infant formulae and weaning foods. In a survey of 25 domestic and imported infant formulae (IF) products (20 milk based, five soy based) and 30 weaning foods commercially available in New Zealand (Vannoort and Cressey personal communication; data presented at New Zealand Institute of Food Science and Technology meeting, July 1997), some pesticide residues were detected, but all extremely low and well below regulatory limits and health safety standards. Residues of pp-DDE were found in seven out of 20 (35%) milk-based infant formulae at concentrations ranging from 0.00003 to 0.00048 mg.kg 1, on a ready-to-feed basis. The parent compound, pp-DDT was found in one milk-based IF at 0.00065 mg.kg 1, and dieldrin residues detected in four out of five (80%) soy-based IF at concentrations ranging from 0.00005 to 0.00008 mg.kg 1. Mean intakes for fully formula-fed infants less than one month of age (the age at which the energy requirements per kilogramme bodyweight are the greatest) were estimated to be less than 0.1% of the ADI for total DDT, while for dieldrin mean intake was 9.5% of the ADI. Two OPs were detected in the survey: azinphos-methyl in one soy-based formula at 0.022 mg.kg 1 on RTF basis, and pirimphos-methyl in two cerealbased weaning foods at 0.005 and 0.014 mg.kg 1. All were well below MRLs. No other pesticides of the 140 screened for were detected.

10.4

The UK’s approach to pesticide surveillance

The UK runs a pesticide residues food surveillance programme via the independent Pesticide Residues Committee (PRC; formerly the Working Party on Pesticide Residues (WPPR)). The independence of the Committee is important because it gives credibility to published data – there can be no massaging of data by government officials! The budget for the programme in 1999/2000 is £1.67 million, most of which will be spent on analysis. Data from the PRC show that approximately 30% of food consumed in the UK contains measurable residues of pesticides (Table 10.5) and that approximately 1% contains residues above the MRL based on the Good Agricultural Practice (GAP)-compliant use of pesticides. The frequency of food contamination by pesticides is remarkably stable. This suggests that if we accept the use of pesticides in food production in the way in

230

Food chemical safety

Table 10.5 Frequency of residues of pesticides in food determined as part of the UK’s monitoring programme (data from MAFF 1995–1999) Year 1994 1995 1996 1997 1998

% with residues 30 31 34 29 26

% above MRL <1 <1 <1 <1 1.3

which they are currently used, we must also accept these residue levels. Indeed, from the point of view of human health there is perhaps good reason to accept the situation because ADI exceedances are very rare. In the UK there have been no exceedances in the past four years. However, ADIs are based on single pesticides, but our diet contains complex cocktails of pesticide residues and there are no data to allow us to decide what effect these might have on the consumer in the long term. Despite this, it is likely that such effects will be minimal because the total intake of pesticides with food is very low when compared with other toxins (e.g. natural plant toxins). In addition, the ADIs have large safety factors (100–1000) built into them. Perhaps the best argument for reducing the use of pesticides is the proven deleterious effect that pesticides have on the environment. The rapidly increasing anti-pesticide lobby is having an effect upon pesticide use. For example, public concern about pesticide residues in food has led to supermarkets controlling the use of particular pesticides by their growers in order to present a market advantage over their rivals. For example, Alar (Daminozide) is a plant growth regulator which is a suspect carcinogen. Its potential toxic effects became apparent to consumers in the UK in the 1990s. One supermarket assured their customers that no Alar has been used by its apple growers. Soon others followed and Alar was effectively removed from the apple growers’ repertoire of pesticides without the need for government legislation. Supermarkets are very powerful in this respect and can instigate significant changes in pesticide usage both nationally and internationally. This, combined with the organic movement and environmental lobby, will almost certainly reduce pesticide use over the coming years. It will be interesting to see if this is reflected in a reduced residues frequency. The UK has taken advantage of the power of retailers (in particular supermarkets) in dictating to their suppliers (and therefore ultimately the producers) by releasing the names of retail outlets from where samples were sourced for its pesticide residues surveillance scheme (MAFF, 1999). A media furore met the first report of the UK’s Working Party on Pesticide Residues to ‘name and shame’ retail outlets. Indeed the Guardian, a major UK newspaper, featured a leaked version of the report on its front page the day before the planned release. As a result of this there was significant interest in a problem with chlormequat in pears. The supermarkets took a particular interest in the

Pesticides

231

issue because of the adverse publicity surrounding it and might have been instrumental in its solution. This is an interesting use of the power of the supermarkets by government to aid in their control of a residues programme. And it does not require legislation and is free!

10.5

Findings from the UK pesticide monitoring scheme

10.5.1 Yams and carbendazim In 1998 there was an apparent rise in the frequency of residues above MRLs (Table 10.5). This was because 75% of yams contained residues of the fungicide carbendazim above its MRL (none exceeded the ADI). On the face of it this is a significant problem, particularly for West Indian families who might include yams as a major carbohydrate source in their diet. However, the reason for these MRL exceedances was because a default MRL (based on the analytical limit of determination for carbendazim) had been set for carbendazim in yams because GAP trials data were not available. These were therefore technical exceedances. Discussions between the UK government and the yam-exporting countries will hopefully result in an Import Tolerance being set which means that even though the yams imported into the UK in 1999 might contain residues of carbendazim similar to those for 1998, they will not exceed MRLs. If a correction is made to the MRL exceedance frequency for 1998 (i.e. yam data are removed) the value is <1% and therefore does not represent an increase over previous years. This is given as an example because it illustrates how misleading statistics in MRLs can be. This is particularly important when MRLs are erroneously used as an indicator of toxicity. 10.5.2 Illegal chlormequat in UK pears Residues data cannot all be explained away. In the 1998 UK survey (MAFF, 1999), 80% of UK-grown pears was found to contain the growth regulatory pesticide chlormequat. Chlormequat is not approved for use in the UK, therefore this represented an illegal use of the pesticide. There are two issues here, the fact that growers were using a pesticide illegally and the potential harm that this might cause the consumer of pears. The latter is particularly important because pears and pear juice are commonly used in proprietary and home-prepared infant foods. ADI calculations showed that one of the pears was only marginally below the ADI (residues in pear = 11 mg.kg 1; ADI = 15 mg.kg 1). This is a situation that regulators and government take very seriously. On this occasion a follow-up study was initiated and legal action taken against the offending growers. Discussions between the UK, Dutch and Belgian governments resulted in the use of chlormequat in pear growing in Belgium being banned and import/export surveillance for chlormequat in pears being stepped up. This is an excellent example of surveillance working for the good of the consumer.

232

Food chemical safety

10.5.3 Lindane in continental-style chocolates A rather unexpected residue problem arose in the UK in 1998 (MAFF, 1999); 73% of high-quality continental-style chocolate on sale in the UK was found to contain lindane ( -hexachlorocyclohexane; -HCH), an organochlorine (OC) pesticide banned in many countries. There were no ADI exceedances, and from the point of view of human health there was little or no concern because most people would only eat small quantities of this expensive commodity on a relatively irregular basis. The lindane originated from the cocoa butter used in the manufacture of the chocolate. Cocoa is grown in parts of the world where pesticide regulation and use are poor and therefore residues are difficult to control. This is an example where the power of the supermarkets might eliminate the problem; press interest in the issue led to public concern which in turn meant that supermarkets were forced to assure their customers that their chocolates were safe. We have no doubt that the supermarkets will put considerable pressure on their suppliers to ensure that their source of chocolates does not contain lindane. In turn, the suppliers will pressurise the manufacturers and the manufacturers only buy cocoa butter from growers who do not use lindane. This is a useful chain reaction to minimise pesticide use which even affects countries where pesticide use is poorly controlled. 10.5.4 Lindane in milk in 1995 Lindane is a hydrophobic OC which has an affinity for lipids and therefore is commonly found in high lipid content foods. For many years UK milk has contained just detectable residues of lindane, partly due to its continued use in the UK (particularly in sugar beet growing) and partly because its residues are relatively long-lived in the environment. In 1995 (MAFF, 1996a) unexpectedly high lindane residues were found in milk sampled in June. Lindane residues in subsequent months’ samples continued to rise, reaching a peak in September (Fig. 10.6). The milk levels in September were only marginally below the ADI. Residues near to the ADI in a staple dietary commodity (i.e. milk, potatoes, bread) are worrying. Fortunately October’s residues were significantly reduced and therefore a potential crisis did not come to fruition. In such cases it is important to explain the effect in order to attempt to prevent its re-occurrence. In this case there were several important contributory factors: 1995 was a hot summer and it is possible that milking cows were marginally malnourished and so in order to maintain their milk output it is likely that they mobilised fat reserves. It is well known that animal fat reserves contain long-lived lipidsoluble residues (e.g. lindane, DDT), therefore lindane from fat reserves might have been mobilised and incorporated into milk. In addition, since the summer’s drought had resulted in a poor cereal, grain and forage crop, these commodities were imported. It is possible that such imports contained lindane and were incorporated into compound feed which was fed to dairy herds. These explanations could not be proved, or even investigated. However, in 1996 (Fig. 10.6) a slight increase in milk lindane levels occurred at approximately the

Pesticides

Fig. 10.6

233

Lindane residues in UK milk in 1995 and 1996. (Graph reproduced by kind permission of MAFF Publications (MAFF 1997), London.)

same time as the previous year’s enormous increase. This was possibly due to farmers feeding their cattle left-over compound feed from the previous year. Surprisingly, the 1998 survey showed, for the first time, that no milk samples contained detectable lindane residues (MAFF, 1999). 10.5.5 Vinclozolin in lettuce Vinclozolin is a fungicide used in lettuce growing. Its approval for use in the UK was revoked because of its potential toxicity to workers applying it to lettuces in greenhouses. It is an androgen receptor blocker and therefore might interfere with sexual development in boys, and have effects on sperm production and secondary sex characteristics in men. Despite this toxicity it is still approved for use in several European countries, including France. Warmer countries probably do not grow their lettuces in greenhouses and therefore the exposure risk to workers applying the vinclozolin in the open air is lower than within the confined space of a greenhouse; this may explain why vinclozolin is approved in some EU member states. Despite its ban in the UK, residues of vinclozolin have been detected in UKgrown winter lettuce for at least six years (Table 10.6). These residues present a negligible risk to the consumer because their concentration is far too low to result in an anti-androgen effect. Nevertheless residues indicate that the pesticide has been used illegally by UK growers and therefore action must be taken. Much to the chagrin of the lettuce growers, lettuce imported from countries where vinclozolin is approved for use is legal in the UK providing its residues do

234

Food chemical safety

Table 10.6 Residues of vinclozolin in UK winter lettuce. Data from MAFF, 1998 and 1999 Year

% samples with vinclozolin residues

1997 1998

3.2 (n = 94) 5.7 (n = 70)

not exceed the MRL. This illustrates well that the MRL is a trading standard rather than a safety factor. At no time have vinclozolin residues exceeded the ADI and so have not presented a risk to the consumer. The only injured party in this incident is MAFF whom some UK growers seem reluctant to obey.

10.6

Human exposure monitoring

It is difficult to assess human exposure to pesticides. Several studies have measured pesticide residues in human fat (MAFF, 1995) or human milk (MAFF, 1997), as both matrices are good indicators of long-term exposure, but they tell little of the subject’s exposure to short-lived pesticides such as the OPs or pyrethroids. 10.6.1 Pesticides in human fat In a MAFF study (MAFF, 1995) of pesticides in human fat in the UK, 99% of samples analysed (n = 203) had detectable residues of DDT (as p,p’-DDT, o,pDDT, p,p’-TDE and/or p,p’-DDE). Since the fat samples were taken at routine autopsy, it is likely that most of the subjects were at least 70 years old and therefore had lived through times when DDT was permitted in the UK. Their residues reflect their lifetime exposure. Twenty-three per cent of the subjects had DDT fat residues between 1 and 9.3 mg.kg 1 which suggests higher exposures. These people might have been exposed directly during DDT’s heyday, or might have been fond of oily fish which contains higher residues of OCs than most other foods. In the same MAFF study other long half-life OCs were also found in human fat (Table 10.7). Again these residues are indicative of the subjects’ lifetime exposures to these pesticides rather than indicating recent exposure. It is interesting that the shortest environmental half-life OC, namely lindane, has the lowest human fat residue frequency. 10.6.2 Pesticides in human milk Levels of pesticide residues in human milk from women from different countries give an indication of exposure to pesticides in their respective countries (Table 10.8). For example, in the USA, p,p’-DDT was found at 0.039 mg.kg-1 (Mattison et al., 1992) in human milk, whereas in milk from women in Faridkot in India a

Pesticides Table 10.7

235

Frequency of OC residues in human fat (n = 203) (MAFF, 1995)

Pesticide

% samples with detectable residues

Chlordane DDT Dieldrin -HCH Lindane ( -HCH) Heptachlor Hexachlorobenzene

53 99 59 98.5 3 30 94

residue level of 13.81 mg.kg 1 was reported (Kalra, 1994). This, perhaps, illustrates the difference in DDT use policies between the two countries. This point is illustrated very well indeed if women from the former East Germany are compared with women from West Germany (Table 10.8). Milk p,p’-DDT residues in the former have been reported (Anon., 1991) at 2.28 mg.kg 1, whereas in the latter a value of 0.81 mg.kg 1 is reported (Anon., 1991). In this example, even though the women originated from bordering countries, their national policies on DDT use is likely to account for the vast difference in pesticide residues found in human milk. Very recent studies (Shaw et al., 2000) in Indonesia have shown human milk DDT residues as high as 17.7 mg.kg 1 (Table 10.9), even though the country insists that DDT is no longer used. Professor Shaw spent time talking to farmers in the Puncak region 50 miles outside Jakarta, Indonesia. He talked with four farmers, three of whom regularly used DDT on their crops. Clearly government policy has not filtered through to the farmers!

Table 10.8

DDT in human milk from around the world (from Shaw et al., 2000)

Country/city USA Arkansas New York Germany E. Berlin W. Berlin Thailand Bangkok S. Vietnam Ho Chi Minh Papua New Guinea India (Punjab) Ludhiana Faridkot

p,p-DDT (mg.kg-1) 0.039 0.023 2.28 0.81 0.734 4.22–7.3 0.023 0.42 7.18 13.81

236

Food chemical safety

Table 10.9

DDT in human milk from four Indonesian women volunteers

Volunteer 1 2 3 4

10.7

p,p-DDT residues in milk fat (mg.kg-1; mean  SD) 4.8 17.7 0.4 4.8

   

0.6 9.1 0.1 0.8

Should we ban pesticides?

The consumers’ perception of the risk of pesticide residues in food is far greater than the actual risk. Indeed, it is likely that natural toxins in food present a far greater risk. Driving to the shop to buy food certainly poses an order of magnitude greater risk than the toxic effects of pesticides in the food. Therefore it is folly for the lobby groups to use residues in food as part of their argument to reduce the use of pesticides. They would be much wiser to concentrate their campaigning efforts on the effects of pesticide use on the environment. The environmental impact of pesticides is a far sharper nail for the pesticides coffin.

10.8

References

(1991) Legacy in milk New Scientist, 130 (1772); 13. and THOMSON, B (2000), 1997/98 New Zealand Total Diet Survey, Part 1 – Pesticides. Ministry of Health, Wellington, New Zealand. KALRA, R L (1994) Organochlorine pesticide residues in human milk in Punjab, India Environ. Pollut, 85; 147. MAFF (1995) Annual Report of the Working Party on Pesticide Residues – 1994. MAFF Publications, London. MAFF (1996a) Annual Report of the Working Party on Pesticide Residues – 1995. MAFF Publications, London. MAFF (1996b) Annual Report of the Steering Group on Chemical Aspects of Food Surveillance, Food Surveillance Paper No. 51, The Stationery Office, London. MAFF (1997) Annual Report of the Working Party on Pesticide Residues – 1996. MAFF Publications, London. MAFF (1998) Annual Report of the Working Party on Pesticide Residues – 1997. MAFF Publications, London. MAFF (1999) Annual Report of the Working Party on Pesticide Residues – 1998. MAFF Publications, London. MAFF (2000) Annual Report of the Working Party on Pesticide Residues – 1999. MAFF Publications, London. ANON.

CRESSEY, P, VANNOORT, R, SILVERS, K

Pesticides

237

and SELEVAN, (1992) Pesticide concentrations in Arkansas breast milk J. Ark. Med. Soc., 88; 553. MULLER, A M F, MAKROPAULOS V and BOLT H M (1995) Toxicological aspects of oestrogen-mimetic xenobiotics present in the environment TEN, 3; 68–73. NRC (NATIONAL RESEARCH COUNCIL) (1993) Pesticides in the Diets of Infants and Children. Washington: National Academy Press. ORCHARD, V A, BAKER, M and MARTIN D (2000) Prioritising health services: the communicable disease picture in New Zealand today. Healthcare Review Online http:///www.enigma.co.nz/hcro articles/0005/vol14no5 001.htm SHAW I C (1999) Pesticides in food, pp. 421–428 in Pesticide Chemistry and Bioscience: The Food Environment Challenge (eds G T Brooks and T R Roberts). Royal Society of Chemistry, London. 1 SHAW I C, BURKE E, SUHARYANTO F and SIHOMBING G (2000) Residues of pp – DDT and hexachlorobenzene in human milk from Indonesian women Environ. Sci. Pollut. Res. 7; 75–77. SHAW I C and CHADWICK J (1998) Principles of Environmental Toxicology. Taylor & Francis, London. SHAW I C and MCCULLY S (2000) The impact of endocrine disrupters on people, in Endocrine Disrupters: The Analytical Challenge. Royal Society of Chemistry, London. In press. TANNER J T and BARNETT S A (1986) Methods of analysis for infant formula: Food and Drug Administration and Infant Formula Council Collaborative Study, Phase III JAOAC, 69(5); 777–785. USDA (1998) Pesticide Data Program – Annual Summary Calendar Year 1996. United States Department of Agriculture, Washington. WINTER C K and FRANCIS F J (1997) Assessing, managing and communicating chemical food risks Food Technology, 51(5); 85–92. MATTISON, D R, WOHLEB, T, TO, Y, LAMB, S, FAITAK, M A, WALLS, R D SG

11 Mycotoxins J. E. Smith, University of Strathclyde, Glasgow

11.1

Introduction

Mycotoxins are extremely toxic chemical substances produced by certain filamentous fungi growing naturally in many agricultural crops but especially in cereals including maize, wheat, barley, rye and most oilseeds both in the field, after harvest and during storage, and later when processed into food and animal feed concentrates (Table 11.1). The consumption of such mycotoxin contaminated foodstuffs can produce toxic symptoms in animals and humans which are known as mycotoxicoses. There is an anomaly in the use of the term mycotoxin. Zootoxins and phytotoxins are compounds toxic to animals and plants respectively, and by analogy the term mycotoxin could be expected to refer to compounds toxic to fungi but instead it is used to describe compounds produced by fungi toxic to animals in a restricted sense. Because of the relatively high intake of cereals and oilseeds in the diet of intensively farmed animals such as poultry, pigs and cattle, there has been extensive documentation of the adverse effects on animal health and productively when mycotoxincontaminated feeds have been consumed (Berry, 1988; Smith and Moss, 1985; Smith et al., 1994). In contrast, human mycotoxicoses are less well understood but are increasingly been diagnosed and studied (Smith et al., 1995). Mycotoxins are undoubtedly the most unfamiliar and least investigated of the diseases that affect man. Human intake of mycotoxins mainly occurs from plantbased foods and from animal-derived foods such as milk, cheese and certain meat products. Mycotoxins are, in general, low molecular weight, non-antigenic fungal secondary metabolites formed by way of several metabolic pathways, e.g. the polyketide route (aflatoxins), the terpene route (trichothecenes), the amino acid

Mycotoxins 239 Table 11.1 mycotoxins

Some important toxigenic species of filamentous fungi and related

Species

Toxin

Aspergillus flavus A. parasiticus A. ochraceus A. versicolor Penicillium verrucosum P. purpurogenum P. expansum Fusarium sporotrichioides F. moniliforme F. graminearum Alternaria alternata Stachybotrys atra

Aflatoxins B1, B2; cyclopiazonic acid Aflatoxins B1, B2, G1, G2 Ochratoxin A; penicillic acid Sterigmatocystin, cyclopiazonic acid Ochratoxin A, citrinin Rubratoxins Patulin, citrinin T-2 toxin Fumonisin B1 Deoxynivalenol, nivalenol, zearalenone Tenuazonic acid Satratoxins

route (aflatoxin) and the tricarboxylic acid route (rubratoxin). Some mycotoxins such as cyclopiazonic acid are formed from a combination of two or more of the principal pathways. While some mycotoxins are formed by only a limited number of fungal species others can be produced by a range of species from several genera. At least 300 potential mycotoxins have been isolated under laboratory conditions but only a relatively small number (about 20) have been shown to occur in foods and feeds at significant levels and frequency. All of these mycotoxins have been shown to have significant animal toxicity. Mycotoxins normally enter the human and animal dietary system by indirect or direct contamination (Table 11.2). Direct contamination occurs when the food or feed becomes infected with a toxigenic fungus with subsequent toxin formation. In contrast, indirect contamination can take place when an ingredient of a process has previously been contaminated with toxin-producing fungi and while the fungus itself may be killed or removed during processing, the mycotoxin will mostly remain in the final product. Most of the common mycotoxins are, in general, quite resistant to most forms of food and feed processing. It is now widely agreed that approximately 25% of the world’s food crops are affected each year by variable levels of mycotoxins which can have considerable economic consequences for the crop, livestock producers, grain handlers, processors, consumers and indeed national economies (Miller, 1998). It has been conservatively estimated that annual losses in the US and Canada, as a result of mycotoxin occurrence, to the feed and livestock industries alone can be in the region of US$5 billion (Charmley et al., 1995). In developing countries, regular mycotoxin presence in the diet will also affect the human populations, causing morbidity and premature deaths. Direct economic losses due to the presence of mycotoxin- (especially aflatoxin-)producing fungi in agricultural crops can be detected in reduced crop yields and lower quality, reduced animal performance and reproductivity

240

Food chemical safety

Table 11.2

Probable routes for mycotoxin contamination of foods and feeds

Mould-damaged foodstuffs Agricultural products, e.g. cereals major source oilseeds

} fruits minor source vegetables }

Consumer foods (secondary infections) Compounded animal feeds (secondary infections) Residues in animal tissues and animal products Milk (animal and human) Dairy produce Meats (liver, kidney) Mould-ripened foods Cheeses Fermented meat products Oriental fermentations Fermenter-derived products Microbial proteins Food additives

capabilities and increased disease incidence. In the past, indirect losses have been greatly underestimated. For instance, crop producers with aflatoxincontaminated products will incur downgrading of crops, reduced markets, increased handling and processing, and increased costs of detoxification or dilution. Feed and food processors will also identify increased costs for further processing needs, in particular, analyses and monitoring for presence of mycotoxins. Similarly, animal producers will identify increased production costs related to veterinary requirements, reduced output and possibly seeking new supplies. Predicting the impact of mycotoxins on net revenues for grain producers has been the subject of intense studies especially in Canada (Charmley et al., 1995) where scientists have devised a prediction formula: NCR = [(100 where

X0/100) AP PP] + [(X/100) AP PC]

NCR = net cash receipts X = percentage of crop contaminated AP = annual grain prediction in metric tonnes per year PP = average prices of premium grain in dollars per metric tonne PC = average price of contaminated grain in dollars per metric tonne

The net loss of cash income is estimated by calculating NCR when X is 0 and subtracting the actual NCR. When such economic analyses are carried out, it is

Mycotoxins 241 clear that when, for instance, aflatoxin contamination does occur at levels above the legal limit, it can lead to multimillion dollar losses. Recognising the economic and health problems created by mycotoxins has been the first step towards implementing appropriate programs for their prevention and control. These must include not only the prevention of mycotoxin formation in agricultural products but also, when possible, their removal through detoxification or destruction. The economies of many developing countries rely greatly on the export of specific agricultural raw materials and due to insufficient drying equipment, coupled with greatly humid atmospheric conditions, can lead to unacceptable levels of aflatoxins in harvested maize, ground nuts and palm kernels. Since the import regulations of most developed countries have become increasingly stringent, this has resulted in restriction of export potential and concentrated economic damage to producer countries.

11.2

Health implications of mycotoxins

11.2.1 Animal toxicity There is an extensive literature documenting the wide range of adverse effects on animal health and productivity, when mycotoxins, especially aflatoxins, are present in the diet (Smith et al., 1994). Animals can demonstrate variable susceptibilities to mycotoxins, depending on genetic factors (species, breed and strain), physiological factors (age, nutrition etc.) and environmental factors (climatic conditions, husbandry, and management). It is also now documented that compound animal feeds, due to their range of sourcing, are exposing animals to multiple mycotoxin assault. In acute forms of mycotoxicosis, individual animals will show marked signs of disease and, in extreme cases, animals may die. Specific observable acute disease syndromes include hepatitis, haemorrhage, nephritis and necrosis of oral and enteric epithelia (Fig. 11.1). However, in most developed countries it is now recognised that natural contamination levels of most mycotoxins in animal feeds are not normally occurring at levels likely to cause overt toxic symptoms but, instead, to induce symptoms of chronic primary mycotoxicoses and immune suppression. It is extremely difficult to diagnose such manifestations of disease but undoubtedly they represent the most common forms of mycotoxicoses in farmed animals, e.g. reduced productivity in the form of slower growth rates, reduced reproductive efficiency, inferior market quality, reduced feed conversion efficiency, reduced milk yields or reduced egg production. Does chronic primary mycotoxicoses exist in humans? It would be surprising if there were no physiological upsets in humans with chronic levels of mycotoxins but, at present, there is no acceptable form of measurement. Because of their diversity of chemical structures and differing physical properties, mycotoxins exhibit a wide array of biological effects on mammalian systems and individual mycotoxins can be genotoxic, mutagenic, carcinogenic, embryotoxic, teratogenic or oestrogenic (Smith and Henderson, 1991). Some

242

Food chemical safety

Fig. 11.1

Toxic phenomena associated with mycotoxins

mycotoxins can cause cancers in a variety of animal species and man. The relatively short life span of most intensively reared animals reduces the significance of cancer development. In contrast, the opposite is true for humans. 11.2.2 Human toxicity The effects of mycotoxins on human health are complex and mostly little understood. Historically, the mycotoxicosis known as ergotism caused massive sufferings and deaths throughout most of Northern Europe in the Middle Ages (Table 11.3). The causal fungal organism Claviceps purpurea growing on rye produced poisonous alkaloids (mycotoxins) which were not inactivated by the baking process and when consumed in the rye bread, could cause nervous disorders, necrosis, gangrene and limb loss, and in many cases there were deleterious effects on human reproducibility (Matossian, 1989). Other historical episodes of mycotoxicosis causing huge loss of life include Alimentary Toxic Aleukia (ATA), Stachybotryotoxicosis in Russia and Yellow Rice Disease in the

Mycotoxins 243 Table 11.3 Extreme diseases of humans commonly recognised as having been caused by mouldy food, the moulds involved, and the mycotoxins that have been implicated Disease

Moulds involved

Toxins implicated

Aflatoxicosis

Aspergillus flavus A. parasiticus Claviceps purpurea, C. paspali, C. fusiformis Fusarium sporotrichioides, F. poae Fusarium spp. Fusarium spp.

Aflatoxins

Ergotism Alimentary toxic aleukia Urov or Kaschin-Beck ‘Drunken bread’ toxicosis (Scabby grain toxicosis) Stachybotryotoxicosis Yellow rice (cardiac beri beri)

Stachybotrys atra Penicillium citreoviride, P. citrinum, P. islandicum

Ergot alkaloids (T-2 toxin) – – (Various macrocyclic trichothecenes) (Luteoskyrin, islanditoxin, cyclochloritine, citrinin, citreoviridin)

Far East. In the above examples of human mycotoxicoses, there has been strong correlation with the consumption of heavily toxin-contaminated foods. In most developed economies the consumption of foods heavily contaminated with mycotoxins is not likely to occur at any substantial level because there exist strict food regulations, although isolated high levels will occasionally occur. Improved agricultural practices and better storage and transportation facilities have curtailed toxigenic mould growth in raw materials destined for the human food chain. However, there is genuine concern on the possible adverse effects resulting from long-term exposure to low levels of certain mycotoxins in the food chain (Smith et al., 1994). There continues to be insufficient understanding of the effects of varying levels of ingestion of a single mycotoxin or a mixture of mycotoxins, the susceptibility of individuals and the effects of other types of toxins in the environment on synergy or even antagonism. Such effects have been well demonstrated in animal systems through careful experimentation yet the food industry continues to underestimate their undoubted presence (with the exception of the aflatoxins) within the human food chain. Little is also known about the consequences of mycotoxin intake on what would be considered as the most vulnerable individuals such as children, pregnant women, the aged, the immuno-compromised and individuals who are receiving a nutritionally inadequate diet. Most of the knowledge concerning modes of action, pathological conditions and degrees of toxicity of mycotoxins has been developed from experimental studies on animals other than man. Genotoxicity resulting in carcinogenicity, mutagenicity and/or teratogenicity has long been associated with the aflatoxins and demonstrated in a wide range of animal species (IARC, 1993). It is now accepted that the aflatoxins are undoubted human carcinogens. The potential carcinogenicity of sterigmatocystin, ochratoxin A, T-2 toxin, patulin,

244

Food chemical safety

griseofulvin and zearalenone is now accepted for animals and highly suspected for humans. Establishing a causal relationship between mycotoxin exposure and human disease is extremely complicated because of the many uncertainties associated with human epidemiological studies. Of special concern is establishing an accurate assessment of human mycotoxin exposure. Recently developed methods for biomonitoring individual markers of a population should contribute significantly to the confirmation of the perceived linkage between mycotoxin exposure and human disease. The development of biomarkers for aflatoxin and other mycotoxins, relies on knowledge of the metabolism and macromolecular adduct formation of these mycotoxins. Methods are now available to use such markers present in urine, blood and milk to estimate the exposure of individuals to these toxins and then predict the risk of developing cancer and other diseases (Coker, 1997; Wild et al., 1998). Where aflatoxin is regularly present in the human diet, such as warm, humid developing countries, it is incriminated in the aetiology of kwashiorkor, present in human breast milk, the blood of pregnant women and in babies’ cord blood at birth. It has also been found in heroin samples in the UK (Smith et al., 1995). The immune system of an organism plays an essential role in defence against infections and tumour formation. Any damage or suppression of the immune system resulting from exposure to toxic substances can make the host more susceptible to the challenges of microbes, tumour cells and other toxins. There is now strong scientific evidence that several mycotoxins can damage parts of the human immune system involving cellular immune phenomena and non-specific humoral factors associated with immunity. Many toxicologists are now suggesting that the single most important toxic role of many mycotoxins on humans could be the insidious effects on the human immune system (Pestka and Bondy, 1994). Chronic exposure to mycotoxins in the human diet could have substantial effects on human health. In many developing countries immunedamaging mycotoxins, especially aflatoxin, are endemic and must cause increased levels of immune-compromised individuals, especially children. Infectious diseases such as respiratory infections, diarrhoeal diseases, tuberculosis, measles, HIV and several others, are amongst the main killers of children in developing countries (Miller, 1998). A World Bank report strongly incriminates mycotoxins as a significant modulating factor in such childhood diseases and deaths (World Bank, 1993). In general, mycotoxins demonstrate their toxicity through association with cellular components which have important roles in cellular regulation and function. Thus, knowledge of the interaction between mycotoxins and cellular components is critical for a full understanding of the mode of action of mycotoxins. While some mycotoxins may act directly on cellular systems, others require to be converted into active forms. While the toxicity of any mycotoxin will be dependent on the dose, it is not the actual administered dose of toxin but the actual concentration of the mycotoxin at the site(s) of cellular action that will determine cytotoxicity. The

Mycotoxins 245 concentration, and thus, the toxicity of a mycotoxin at a target site or organ will be influenced by many factors. If the resultant concentration of the mycotoxin is very small it may never achieve a sufficient toxic concentration at the potential site of action, while biotransformation of the mycotoxin could result in the production of a less or more toxic metabolite. The rate of excretion of a mycotoxin can be closely interrelated and dependent on its distribution and/or biotransformation. Absorption of a mycotoxin will occur when it crosses body membranes and enters the blood stream. The primary sites of mycotoxin absorption are the gastrointestinal tract (ingestion of contaminated food), lungs (inhalation of contaminated particles or toxin-containing fungal spores) and the skin (direct contact with contaminated materials or pure mycotoxins). When the mycotoxin enters the blood it is then available for distribution. Livers and kidneys have a high capacity to bind many mycotoxins while other mycotoxins are highly lipophilic and can concentrate in body fat. In the final outcome a toxic response by a mycotoxin will be critically influenced by the rate of absorption, distribution, biotransformation and excretion (Smith et al., 1994). 11.2.3 Mycotoxins and risk assessment Over the last few years, risk assessments of mycotoxins have been carried out and serve as the basis for their risk management (Kuiper-Goodman, 1998). Mycotoxin levels in food raw materials will depend on many factors including adverse conditions favouring fungal invasion and growth either in the field or during storage (especially cereals and oilseeds). A high level of monitoring of mycotoxins is required in areas where problems with high levels, which could cause adverse health effects, are known to occur. This is especially pertinent where two or more mycotoxins may be present, such as aflatoxins and fumonisin, resulting in synergistic effects. Epidemiological studies and animal toxicity studies are important to provide the greatest predictive information and more recently there has been the use of in vitro assays with human cell types (Lewis et al., 1998). Several mycotoxins are now recognised to be carcinogenic. Aflatoxin B1 is believed to be able to induce the process of carcinogenisis whereas the fumonisin and zearalenone affect the promotion of carcinogenesis. While several mycotoxins may be considered as acute hazards to animals (i.e. occurring at levels able to cause death) this will seldom be the case for humans. Rather mycotoxins will represent a chronic hazard to humans because of lower exposure levels in general. The widespread exposure to many populations of the potent liver carcinogen, aflatoxin B1, and the potent renal carcinogen ochratoxin A, and other carcinogenic mycotoxins, confirms that mycotoxins should be considered to pose the highest risk of any diet-related risk to humans when compared with other microbial toxins, phycotoxins, food additives and pesticide residues. Yet, this insidious aspect of mycotoxins still goes surprisingly unappreciated and unrecognised by most food microbiologists!

246

11.3

Food chemical safety

Analytical methods

11.3.1 General principles Efficient sampling, sample preparation and methods of analysis are the continuing aims for identifying the extent of mycotoxins presence in foods, feed ingredients, human and animal tissues, blood, urine and milk in order to form the basis of quality control procedures (Smith et al., 1994). Since mycotoxins display a wide diversity of chemical structures, there are no uniform methods of analysis either collectively or for specific mycotoxins in various food or feeds (Steyn et al., 1991). The analytical procedures for isolating mycotoxins from complex biological matrices and subsequent separation and purification follows well documented flow patterns, viz. sampling, extraction, clean-up, separation, detection, quantification and, finally, confirmation (Table 11.4). Most mycotoxins can now be readily identified qualitatively and quantitatively, and current methodologies concentrate on increasing sensitivity, accuracy and reproducibility and, above all, to decrease the time of analysis. Mycotoxins are rarely uniformly distributed throughout natural products and their presence will be uneven and spasmodic. Thus, critical to any analytical process for mycotoxin analysis is the design and efficacy of the sampling plan. All sampling steps must be carried out with a high degree of accuracy so that the final samples to be chemically analysed are truly representative of the batch in question. Difficulties arise due to variation in raw material particle size. Thus, a single heavily contaminated peanut when ground up with a large number of (e.g. 10,000) uncontaminated peanuts could still lead to an overall level of contamination that is unacceptable. Absolute recommendations on sample size are not feasible but should be related to size of the specific particles and level of homogeneity. Thus, larger samples should be taken with increasing particle size, e.g. peanuts > corn > wheat > rice > milled products. Fluids and well-mixed process products such as milk, milk products, beers etc., do not normally present such sampling problems. The complexities associated with the design of acceptable sampling protocols for mycotoxins have been comprehensively reviewed by Coker et al. (1995). Mycotoxins usually occur in extremely low concentrations (ppb-ppm) in the complex chemical mixture of natural organic material. Before the mycotoxins can be accurately measured, they must be extracted from the contaminated commodity and other co-extracted organic molecules substantially removed by clean-up producers. Mycotoxin purification from commodities has been improved immensely in recent years by the use of solid phase extraction columns which are quick, solvent efficient and extremely economical. The analyte (mycotoxin) can be eluted into a small solvent volume to be then injected into a liquid or gas chromatograph with possible automation for largescale sample handling. Multifunctional clean-up columns provide an extremely rapid removal of interference from food extracts by lipophilic and charged sites present in the proprietary packing material. Yet another improved clean-up methodology is the immunoaffinity cartridge which can be used only when

Mycotoxins 247 Table 11.4

Basic protocol in the analyses of mycotoxins from natural products

Step

Description

Sampling Sample

Probe or automatic sampler Representative sample Grinding, mixing, subsampling Representative preparation sample Shaker or blender Separate the toxin from compounds insoluble in the extraction solution Liquid-liquid partitioning Separate the toxin from (separatory funnel) groups of compounds Column chromatography in the sample extract Divalent metal clean-up (Pb2+, Fe2+, Cu2+) Thin layer Separate the toxin from chromatography (TLC) remaining compounds in the Gas-liquid chromatography sample extract that might (GLC) interfere with the toxin Liquid-liquid chromatography (LC) Minicolumn chromatography

Extraction Clean-up

Final

Purpose

Derivatisation (where appropriate) Detection and quantitation Confirmation

Fluorescence on TLC plate Fluorescence in solution UV absorption in solution GLC-flame detector TLC separation and detection of derivative of mycotoxin Biological test Mass spectrometry

Detection measurement of response Identification of chemical compound

suitable monoclonal antibodies to specific mycotoxins are available (Dietrich et al., 1995). The crude extract is passed through a cartridge containing the specific monoclonal mycotoxin antibodies absorbed onto an inert support, and the retained mycotoxins are eluted from the affinity column with solvent prior to quantification using the most suitable and efficient method. Affinity cartridges are now available for the analysis of the aflatoxins B1, B2, G1, G2 and M1 and also for fumonisin B1, ochratoxin A, zearalenone and deoxynivalenol. 11.3.2 Chromatographic separation Analytical methods for mycotoxins have historically utilised chemical detection methods which are easily quantifiable, highly sensitive, less subject to interference by co-extractives but can be quite time demanding. Most quantitative methods use a final chromatographic separator followed by a suitable detection

248

Food chemical safety

step. Most mycotoxins can be detected visually by their inherent fluorescence or by colour development after spraying with a chromogenic reagent. Thin layer chromatography (TLC) is the most widely used quantification step although high performance liquid chromatography (HPLC) is increasingly used since it offers increased sensitivity, improved accuracy and the ability to automate. Reverse phase liquid chromatography (LC) is now being increasingly used for many mycotoxins. As most mycotoxins are non-volatile, gas-liquid chromatography (GLC) has offered limited assistance but has been used especially for the non-fluorescing trichothecene mycotoxins. While HPLC does not always produce superior results to those with TLC it allows greater versatility and is more suitable for the analysis of complex organic matrices such as cereals. HPLC coupled to sensitive detection and sophisticated data retrieval has improved the identification of selected mycotoxins at levels much less than achieved by TLC. Additional chromatographic modes such as normal-phase, reverse phase and ion-exchange chromatography have been employed. There are no truly universal detectors available for HPLC. Detectors presently in use include Fourier transform infrared detections (FT-IRD), diode array ultraviolet detection (DAD) and mass selection detectors (MSD) (Coker, 1997). While GLC has been much less used for mycotoxin analyses than TLC and HPLC it has been extensively used for trichothecenes as the method of choice. While HPLC and GLC techniques have excellent sensitivity and reproducibility and can be adapted to automation they do have certain disadvantages, viz: • samples require to be highly processed in relation to clean-up to achieve sensitive detection limits; • only one sample can be analysed at any given time; • capital output is high; • running costs are high; • highly trained personnel are normally required to operate the equipment and interpret the results; • they are time-consuming.

In many instances of mycotoxin analysis there is a great need for screening methods that can analyse large volumes of food and feed samples, most of which will be mycotoxin-free. It would also be extremely helpful if such methods could be used by relatively inexperienced operators and in situations where good laboratory facilities are not available (Table 11.5). 11.3.3 Immunoassays Immunoassay methodologies are now a major method for rapid analysis of many mycotoxins, especially aflatoxins. These immunochemical techniques are based upon quite different principles to chromatographic procedures. In essence, immunochemical procedures involve reversible binding between antigens (the

Mycotoxins 249 Table 11.5

Properties required in rapid mycotoxin analytical methodology

1. Minimum sample preparation prior to analysis. 2 Good sensitivity and specificity for the appropriate mycotoxin(s) under test. 3. Applicability to a wide range of natural substrates. 4. Easy adaptability to in-field situations using untrained personnel or at least using the minimum of laboratory equipment and technical assistance. 5. Laboratory-oriented prototype to assess the full capabilities of the system for toxin determination. 6. Availability of final results (e.g. a distinct and measurable end-point) within a short period of time from sample preparation (5–30 min).

analyte or mycotoxin) and selective antibodies, resulting in a specific antigenantibody complex (Candlish, 1991, Dietrich et al., 1995). For mycotoxin analyses radioimmunoassay (RIA), enzyme-linked immunosorbent assay (ELISAs) and affinity chromatography are the principal immunochemical methods in commercial application. Immunoaffinity columns or cartridges for specific mycotoxins are now being increasingly used in preliminary clean-up of extracts prior to final analysis by HPLC or GLC methods. A considerable number of immunoassay kits for several mycotoxins are now being marketed under various trade names and while most of these kits give satisfactory results when compared with other accepted methods their evaluation can be a source of controversy. Consequently, AOAC Intervention in the USA has established a new Research Institute to test the performance of such proprietary test kit methods (Smith et al., 1994) (Table 11.5). 11.3.4 Novel approaches In recent years there has been increasing interest in the use of primary and established animal and human cell lines in which mycotoxin toxicity is determined on morphological criteria and end-points like protein or DNA synthesis. The development of immortalised human cell lines is now being actively pursued to serve as standardised bioassays for the detection of food toxins including mycotoxins (Lewis et al., 1998). The development and improvement of analytical methodologies for mycotoxins has been greatly improved by the increased availability of matrix matched certified reference materials (CRMs) (Boenke, 1995) (Table 11.6). The type of matrix CRMs and concentration of the specified mycotoxin are based on the natural occurrence pattern of the toxin in specific foods and feeds. The recent availability of suitable CRMs, while being a prerequisite for the implementation of regulations and standards, will also be invaluable in many ways for the validation of new methods, solving trade disputes and for harmonising proficiency schemes. Throughout the world, there are statutory limits or advisory guidelines for the maximum levels of, especially, the aflatoxins permitted in foods and feeds (Tables

250

Food chemical safety

Table 11.6 Overview on the different CRMs for aflatoxin analysis and corresponding certified values (adapted from Boenke, 1995) CRM No. or RM No.

Matrix

CRM385

Peanut butter

CRM401

CRM282 CRM283 CRM284 CRM285 CRM262 CRM263 CRM264 CRM375

Mycotoxin

Aflatoxin B1 Aflatoxin B2 Aflatoxin G1 Aflatoxin G2 Total aflatoxins Peanut butter Aflatoxin B1 Aflatoxin B2 Aflatoxin G1 Aflatoxin G2 Total aflatoxins Full cream Aflatoxin M1 Full cream Aflatoxin M1 Milk powder Aflatoxin M1 Milk powder Aflatoxin M1 Defatted peanut Aflatoxin M1 meal Compound feed

Certified value mass fraction or mass concentration

Uncertainty mass fraction or mass concentration

7.0 ug/kg (d) 1.1 ug/kg (d) 1.7 ug/kg (d) 0.3 ug/kg (d) 10.1 ug/kg (b) < 0.2 ug/kg(d) < 0.2 ug/kg(d) < 0.3 ug/kg (d) < 0.2 ug/kg (d) < 0.9 ug/kg (d) < 0.05 ug/kg (a) 0.09 ug/kg (d) 0.31 ug/kg (d) 0.76 ug/kg (d) < 3 ug/kg (d)

±0.8 ug/kg ±0.2 ug/kg ±0.3 ug/kg ±0.2 ug/kg ±1.5 ug/kg (c) – – – – – – + 0.04 ug/kg ±0.06 ug/kg ±0.05 ug/kg –

43.3 ug/kg (d) 206 ug/kg (d) < 1 ug/kg (d) 9.3 ug/kg (d)

±2.8 ug/kg ±13 ug/kg – ±0.5 ug/kg

11.7 and 11.8). Such levels have been set surprisingly low when there is considerable evidence of the difficulties encountered in aflatoxin sampling and analysis. While consistency can be more easily achieved within an individual laboratory, there still exists considerable inter-laboratory variability when analysing at threshold limits. In most cases, current limits are set down to the limits of performance of available analytical methods (van Egmond and Dekker, 1995).

11.4

Application of HACCP systems to reduce mycotoxin presence

Hazard Analysis Critical Control Points (HACCP) systems with respect to foods are a systematic and practical approach to enhance the safety of foods from primary production to final consumption through the identification, evaluation and control of hazards that are significant for food safety (WHO, 1995; Brera et al., 1998). Inherent in this approach is that it is the producer who will be responsible for identifying and implementing preventative actions in all aspects of the food production chain to ensure maximum safety from specific food safety hazards. Thus, the producer must set in motion a plan to identify, monitor

Mycotoxins 251 Table 11.7

FDA guidance levels for total aflatoxins in livestock feeds and human foods

Class of animal

Feed

Total AF level mg/kg

Finishing beef cattle Beef cattle, swine, poultry Finishing swine over 100 lbs Breeding cattle, breeding swine and mature poultry Immature animals

Corn and peanut products Cotton seed meal Corn or peanut products Corn and peanut products

0.3 0.3 0.2 0.1

Animal feeds and ingredients excluding cottonseed meal Animal feeds and ingredients Milk Any food except milk

0.02

Dairy animals and others not listed above Humans

0.02 0.0005 (AFM1) 0.02

and rectify any harmful condition(s) that might occur during the whole production cycle. The need for hazard identification and characterisation for mycotoxins has been required because of the now well-recognised animal and human toxicological effects of many commonly occurring mycotoxins, especially aflatoxin B1. For a HACCP system to function adequately it is essential that reliable analytical and sampling methods are available to assess the level of mycotoxins in the commodity. Also, processing technologies may affect levels of mycotoxins, in some cases reducing the level while in others concentrating. The frequency and duration of consumption of the specific commodity should also be considered. Unlike toxic bacterial contamination where toxicity will be apparent relatively quickly, with ingestion of mycotoxin-contaminated foods there will generally be no evident symptoms in humans. Long-term exposure to low levels of certain mycotoxins can lead to chronic diseases, i.e. cancer (IARC, 1993). It is largely because of the lack of observable symptoms from mycotoxin intake that there has been insufficient attention given to this insidious problem. At the farm level due attention must be given to pre-harvest conditions including Good Agricultural Practice involving correct irrigation, crop turnover and suitable use of pesticides. At post-harvest stages the correct drying of the crop is possibly the most critical of all the control points and retaining the low level of moisture within the commodity throughout storage and transportation is critical. All stages of storage and transportation to the factory must be examined to avoid entry points of moisture and mould contamination. During processing new hazards may arise due to inadequate factory hygiene, prolonged periods in poor environmental conditions, lack of refrigeration, addition of contaminated ingredients, i.e. spices, contaminated machinery, poor packaging etc. When mould fungi are used in the processing of some foods (cheese and fermented sausages etc.), it is imperative to use mycotoxin-free strains (Smith and Moss, 1985).

252

Food chemical safety

Table 11.8

EC guidance levels for aflatoxins in feedingstuffs and human foods

Substances, products

Feedingstuffs

Maximum contents in mg/kg of unadulterated matter at moisture content of 12%

Aflatoxin B1

Straight feedingstuffs with the exception of: Groundnut, copra, palm kernel, cotton seed, babassu, maize, and products derived from the processing thereof Complete feeding stuffs for cattle, sheep and goats (with the exception of complete feedingstuffs for calves, lambs and kids) Complete feedingstuffs for pigs and poultry (with the exception of young animals) Other complete feedingstuffs Complementary feedingstuffs for cattle, sheep and goats (with the exception of complementary feedingstuffs for dairy animals, calves and lambs) Complementary feedingstuffs for pigs and poultry (with the exception of young animals) Other complementary feedingstuffs

0.05

Aflatoxin B1 Aflatoxin M1

Human foods Cereals, nuts, nut products and dried fruits (mostly imported) Baby food and food for infants Milk/milk products Baby food and food for infants

0.02

0.05

0.02 0.01 0.05

0.03 0.005

0.002–4 0.001–2 0.0005 0.0005

Furthermore, it is essential that the analytical laboratory can ensure continued, accurate and reproducible mycotoxin analyses especially to ensure that negative results mean just that. Mycotoxins present a high severity of hazard level because of the serious long-term effects on humans, especially immunological disturbances and cancer developments. While the public are increasingly aware of the immediate health-disturbing effects of viral and bacterial-contaminated foods

Mycotoxins 253 sadly this is not the case yet for mycotoxins. They remain the ‘Cinderellas’ of food poisoning toxins.

11.5

Prevention and control of mycotoxins

The principal path of entry of mycotoxins into the human and animal food chains is through agricultural products, mainly cereals and oilseeds or products later derived from such sources (Smith and Moss, 1985). Toxigenic mould spores are almost universally present in the atmosphere and especially around agricultural crops and products, and successful contaminant control programs will include the development of methods to inhibit the germination and proliferation of such spores. While mycotoxin contamination mostly arises from inadequate storage conditions of agricultural produce there are many examples where seeds may be contaminated at the pre-harvest stage. In particular, the Fusarium mycotoxins, zearalenone, the trichothecenes and the fumonisins, the ergot alkaloids, the tremorgens and the aflatoxins can be formed in seeds in the field. 11.5.1 Pre-harvest control Preventative treatment to avoid pre-harvest fungal penetration of seeds and subsequent toxin formation include: • breeding fungal resistant crop plants; • good agronomic practices (to reduce crop stress by good irrigation, crop rotation etc.); • harvesting the crop at the optimum stage of development; • rapid reduction in moisture level by correct drying; • application of pesticides.

Many environmental parameters have been shown to promote mycotoxin formation in growing crops and include varietal susceptibility or resistance, insect infestation, drought conditions, mechanical damage, nutritional deficiencies and unseasonable temperatures and rainfall. The ultimate aim in pre-harvest control must be to generate a high level of resistance into the main agricultural food crops (Lisker and Lillehoj, 1991). 11.5.2 Post-harvest control Post-harvest contamination by toxigenic fungi usually occurs during storage and transportation and is caused by improper drying or re-wetting of the crop from rain or condensation. Thus, effective post-harvest management technologies involve correct drying and storage. It should be noted that post-harvest deterioration of grain annually amounts to about 10–20% of world production, much of which is by fungi (Williams, 1991).

254

Food chemical safety

By far the most critical environmental factors determining whether a substrate will support mould growth are temperature, moisture content and time, and each of these parameters can be used correctly for the prevention of mould growth. While low temperature storage can be valuable for controlling mould growth in some conditions it is of little use for large-scale storage of agricultural crops. Agricultural products, especially cereals and oil seeds, will usually be held under storage for considerable periods of time. The prevention of mould growth in storage conditions is normally accomplished by modification of the inter-seed environment, e.g. controlling moisture, temperature, and the gaseous environment. In most cases, the control of moisture content is the main commercial approach used to inhibit fungal growth and subsequent mycotoxin formation. Long-term storage of grain must be considered as both a microbiological and an engineering problem. Correct design of sustainable storage is where environmental control by engineering practices will endeavour to hold the moisture level at the minimal value. In practice, it has been noted that the greatest risks will occur with the storage of inadequately dried agricultural products and re-wetting of dried and stored products. Such problems are compounded in countries with high ambient humidities. In most developing countries crop drying is normally achieved by sun drying with repeated turning of the grain masses (Aidoo, 1991). While this can be successfully achieved in dry countries inclement weather in tropical environments can hinder successful drying. In temperate countries successful drying is achieved by various hot air methods. Grain and other agricultural produce can be stored under hermetic or anaerobic conditions in silos where the aerobic toxigenic mould spores cannot germinate. The duration of storage can be an important factor when considering mycotoxin formation since the longer the retention in storage systems, the greater will be the possibility of changes in environmental conditions conducive to mould proliferation. 11.5.3 Chemical methods The use of selective chemicals (mostly organic acids) to prevent mould spore germination and proliferation in moist cereal grain in storage is a widely used practice. Such methods can be effective with a relatively small capital outlay since the need to achieve low water content decreases. The disadvantage of these acids include corrosiveness, they are unpleasant to handle and the need to seek special areas for preparation. Furthermore, the germinating ability of the seed is lost after treatment and consequently chemical preservation of grain crops is carried out mostly to store grain in good condition for animal feed. The ideal chemical preservative should demonstrate broad spectrum activity, low mammalian toxicity, low volatility, and be fungicidal at low concentrations (Lacey et al., 1982). Various organic acids such as sorbic, benzoic, propionic, acetic and formic acids, have been widely used as preservatives and are normally used as the corresponding sodium, potassium or calcium salts. Such methods are seldom used where the stored produce is destined for human consumption.

Mycotoxins 255 11.5.4 Elimination of mycotoxins from contaminated materials Whenever a product has been deemed contaminated with mycotoxins there are only two approaches that can be used for ultimate safe human or animal consumption: • the toxin contaminated product (i.e. seed) can be removed; • the toxin can be degraded into less toxic or non-toxic products.

Since most mycotoxins in agricultural materials are usually contained in a very small proportion of individual seeds or kernels the most practical and effective method of reducing the mycotoxin content of the whole commodity is to remove the contaminated seeds or kernels mechanically (West and Bullerman, 1991). Various techniques have been devised, based on colour and visual appearance of decay or damage to separate out contaminated seed etc. This may be manual or by more advanced electronic instrumental selection. There has been limited use of high temperature exposure for destruction of the aflatoxins in various nuts and coffee beans but this has been offset by the deteriorative changes in the raw material. There has been considerable interest in the development and application of phyllosilicate clays (HSCAs) to selectively chemisorb aflatoxins in aqueous suspensions, including milk, reduce the uptake of aflatoxin by blood and its distribution to body organs such as the liver, reduce the transmission of aflatoxin M1 to milk in lactating animals and to decrease the toxic effect of aflatoxin to many animal species (Phillips et al., 1995). It is believed that the HSCAs when added to feeds act by the selective chemisorption of the aflatoxin in the gastrointestinal tract of the animal resulting in a marked reduction in the bioavailability of the aflatoxins. Chemical detoxification processes or decontamination will include degradation, destruction and/or inactivation of the mycotoxin. In any such process the reduction of the mycotoxin to safe levels should not result in toxic degradation products or reduce the palatability or nutritional properties of the commodities. Aflatoxin has been the subject of most studies and only a relatively small number of these offers any hope of success. There is as yet no FDA or EC fully approved method for aflatoxin detoxification in human foods. Current methods in advanced stages of approval use ammonia in the gaseous form or as an ammonium hydroxide solution at various temperatures, pressure, moisture contents and reaction time to degrade aflatoxins in various animal feedstuffs. There have been extensive studies using two processes, viz: • high pressure and high temperature • atmospheric pressure and ambient temperature

and it has been possible to reduce aflatoxin levels in contaminated products by up to 99% (Coker, 1997) (Table 11.9). Finally, biological detoxification or biotransformation, or degradation of mycotoxins, especially aflatoxin by microbial systems to a metabolite(s) that is

256

Food chemical safety

Table 11.9 procedures

Parameters and application of ammonia/aflatoxin decontamination Process

Temperature

Ammonia level (%) Pressure (PSI) Temperature (ºC) Duration Moisture (%) Commodities

Application

High temperature High pressure

Ambient temperature Atmospheric pressure

1–4 45 (3 bars) 80–120 0.5–3.0h 14–20 Whole cottonseed, cottonseed meal, peanut meal, peanut cakes Feed mill

1.8–2 Atmospheric Ambient 14–21 days 17 Whole cottonseed, corn Farm

either non-toxic when consumed by animals or less toxic than the original toxin and readily excreted from the body is being extensively studied in many laboratories but, as yet, does not constitute realistic application (Smith and Bol, 1989; Bhatnagar et al., 1991).

11.6

Conclusion and future trends

While there is a wealth of knowledge concerning the toxic effects of mycotoxins on animals, there is still insufficient knowledge concerning the health effects of mycotoxins in humans. There is urgent need to gain a better understanding of mycotoxins in the epidemiology of human disease and to further confirm the growing evidence of a relationship between mycotoxin exposure and immunotoxicity. The insidious long-term effects of low concentrations of certain mycotoxins on the human immune system identifies mycotoxins as serious human environmentally-derived poisons. Firm establishment of these relationships must be a central feature of future research. There should be an extension of research into the analysis of human body fluids, e.g. blood, urine, milk by way of adducts specific to the perceived most harmful mycotoxins. There should also be continued examination of novel ways of determining toxic effects on the immune system. The use of human cell lines for toxicity studies should be further developed especially using cells that have been transfected with human cytochromes. The continued development of rapid, cost-effective methods of analysis of mycotoxins is essential, especially methods that can be used in unsophisticated laboratories. While low levels of tolerance are now accepted for the aflatoxins in most developed countries, it is doubtful if most laboratories can meet such requirements.

Mycotoxins 257 It is ultimately required to reduce/eliminate mycotoxins in the main food crops and research should continue on improved farm management, storage and transport facilities and general awareness of toxic fungal contamination not only in industry but also in the domestic environment. Of particular significance would be the development of mould- and/or mycotoxin-resistance crop varieties which should be more possible using modern genetic engineering techniques. There should also be continued research to optimise methods to detoxify contaminated agricultural crops to produce safe products. It has been stated that the single most important influence on the human diet worldwide would be the successful exclusion of mycotoxins (Miller, 1998). The awareness of the toxic role of mycotoxins in the human diet would be greatly improved by the wider teaching of mycotoxicology in courses on food science and microbiology.

11.7

References

AIDOO, K.E. ‘Postharvest

storage and preservation of tropical crops’, in Smith J.E. and Henderson R.S., Mycotoxins and Animal Foods, Boca Raton, CRC Press, 747–764, 1991. BERRY, C.L. ‘The pathology of mycotoxins’, J. Pathology, 154, 301–311, 1988. BHATNAGAR, D., LILLEJOHN, E.B. and BENNETT, J.W. ‘Biological detoxification of mycotoxins’, in Smith, J.E. and Henderson, R.S. Mycotoxins and Animal Foods, Boca Raton, CRC Press, 815–826, 1991. BOENKE, A. ‘BCR- and M&T-activities in the area of mycotoxin analysis in food and feedstuffs’, Nat. Toxins 3, 243–247, 1995. BRERA, C., MIRAGLIA, M. and ONORI, R. ‘HACCP and mycotoxins: what is the point of interaction’ in Miraglia, M., van Egmond, H., Brera, C. and Gilbert, J. Toxicology and Food Safety, Alaken Inc., Fort Collins, Colorado, 145–149, 1998. CANDLISH, A.A.G. ‘The determination of mycotoxins in animal feeds by biological methods’ in Smith, J.E. and Henderson, R.S., Mycotoxins and Animal Foods, Boca Raton, CRC Press, 223–246. 1991. CHARMLEY, L.L., TRENHOLM, H.L., PRELUSKY, D.B. and ROSENBERG, A. ‘Economic losses and decontamination’, Nat. Toxins, 3, 199–203, 1995. COKER, R.D. Mycotoxins and Their Control: Constraints and Opportunities. Bulletin 73, Natural Resources Institute, The University of Greenwich, 1997. COKER, R.D., NAGLER, M.J.H., BLUNDEN, G., SHASKEY, A.J., DEFIZE, P.R., DERKSEN,

and WHITAKER, T.B. ‘Design of sampling plans for mycotoxins in foods and feeds’, Nat. Toxins, 3, 257–262, 1995. DIETRICH, R., SCHNEIDER, E., USLEBER, E. and MARTLBAUER, E. ‘Use of monoclonal antibodies for the analysis of mycotoxins’, Nat. Toxins, 3, 288–293, 1995. IARC, ‘Some naturally occurring substances: food items and constituents, heterocyclic aromatic amines and mycotoxins’, IARC Monographs on the G.B.

258

Food chemical safety

Evaluation of Carcinogenic Risks to Humans, Vol. 56, IARC, Lyon, France, 1993. KUIPER-GOODMAN, T. ‘Food safety: mycotoxins and phycotoxins in perspective’ in Miraglia M., van Egmond H.P., Brera, C. and Gilbert, J. Mycotoxins and Phycotoxins: Developments in Chemistry, Toxicology and Food Safety, Alaken Inc., Fort Collins, Colorado, 25–48, 1998. LACEY, J., LORD, K.A. and CAYLEY, G.R. ‘Problems in preserving crops with chemicals’, in Proceedings 4th Meeting, Mycotoxins in Animal Disease, Ministry of Agriculture, Fisheries and Foods, Alnwick, U.K., Ref. Book360, 104–120, 1982. LEWIS, C.W., SMITH, J.E., ANDERSON, J.G. and FRESHNEY, R.I. ‘Comparative cytotoxicity of mycotoxins to transformed and transfected normal human bronchial epithelial cells’, in Miraglia, M., van Egmond, H.P., Brera, C. and Gilbert, J. Mycotoxins and Phycotoxins: Developments in Chemistry, Toxicology and Food Safety, Alaken Inc., Fort Collins, Colorado, 159– 164, 1998. LISKER, N. and LILLEHOJ, E.B. ‘Prevention of mycotoxin contamination (principally aflatoxins and Fusarium toxins) at the preharvest stage’, in Smith, J.E. and Henderson, R.S. Mycotoxins and Animal Foods, Boca Raton, CRC Press, 689–720, 1991. MATOSSIAN, M.K. Poisons of the Past: Moulds, Epidemics and History, New Haven, Yale University Press, 1989. MILLER, J.D. ‘Global significance of mycotoxins’, in Miraglia, M., van Egmond, H., Brera, C. and Gilbert, J. Mycotoxins and Phycotoxins: Developments in Chemistry, Toxicology and Food Safety, Alaken Inc., Fort Collins, Colarado, 3–15, 1998. PESTKA, J.J. and BONDY, G.S. ‘Mycotoxin-induced immune modulation’, in Dean J.H., Luster, M.I., Munson, A.E. and Kimber, I., Immunotoxicity and Immunopharmacology, 2nd Edition, Raven Press, New York, 163–182, 1994. PHILLIPS, T.D., SARR, A.B. and GRANT, P.G. ‘Selective chemisorption and detoxification of aflatoxins in phyllosilicate clay’, Nat. Toxins, 3, 204– 213, 1995. SMITH, J.E. and BOL, J. ‘Biological detoxification of aflatoxin’, Food Biotechnol. 3, 127–138, 1989. SMITH, J.E. and HENDERSON, R.S. Mycotoxins and Animal Foods, Boca Raton, CRC Press, 1991. SMITH, J.E., LEWIS, C.W., ANDERSON, J.G. and SOLOMONS, G.L. Mycotoxins in Human Nutrition and Health, Directorate-General XII Science, Research and Development, EUR 16048 EN, 1994. SMITH, J.E. and MOSS M.O. Mycotoxins: Formation, Analysis and Significance, John Wiley and Sons Ltd., Chichester, 1985. SMITH, J.E., SOLOMONS, G., LEWIS, C. and ANDERSON, J.G. ‘The role of mycotoxins in human and animal nutrition and health’, Nat. Toxins, 187–192, 1995. STEYN, P.S., THIEL, P.G. and TRINDDER, D.W. ‘Detection and quantification of

Mycotoxins 259 mycotoxins by chemical analysis’, in Smith, J.E. and Henderson, R.S. Mycotoxins and Animal Foods, Boca Raton, CRC Press, 165–222, 1991. VAN EGMOND, H.P. and DEKKER, W.H. ‘Worldwide regulations for mycotoxins in 1994’, Nat. Toxins, 3, 332–336, 1995. WEST, D.I. and BULLERMAN, L.B. ‘Physical and chemical separation of mycotoxins from agricultural products’, in Smith, J.E. and Henderson, R.S., Mycotoxins and Animal Foods, Boca Raton, CRC Press, 777–784, 1991. WHO ‘Applications of risk analysis to food standard issues’. Report of the Joint FAO/WHO Expert Consultation, Geneva, Switzerland, 13–17 March 1995. WILD, C.P., DAUDT, A.W. and CASTEGNARO, M. ‘The molecular epidemiology of mycotoxin-related diseases’ in Miraglia, M., van Egmond, H.P., Brera, C. and Gilbert, J. Mycotoxins and Phycotoxins: Developments in Chemistry, Toxicology and Food Safety, Alaken Inc., Fort Collins, Colorado, 213– 232, 1998. WILLIAMS, P.C. ‘Storage of grains and seeds’, in Smith, J.E. and Henderson, R.S. Mycotoxins and Animal Foods, Boca Raton, CRC Press, 721–764, 1991. WORLD BANK, World Development Report: Investing in Health, Oxford University Press, 1993.

This page intentionally left blank

Part III Regulation

This page intentionally left blank

12 The international regulation of chemical contaminants in food T. Berg, Danish Veterinary and Food Administration, Soborg

12.1

Introduction

Dioxins and polychlorinated biphenyls (PCBs) in fatty foods, methylmercury in fish, tetrabutyltin in molluscs or fumonisins in corn products are some examples of chemical contaminants. Frequently problems concerning chemical contaminants in food reach the headlines of major newspapers and the reader may be left with a scare, sometimes with good reason, often not. In any event, those responsible for the sale of foods must be alert. They bear the responsibility for the safety of the foods they offer for sale, and they also know that any such alert may instantly influence the sale, whether the problem raised is more or less serious. An adequate supply of safe food is considered to be essential for everyone – more or less a right for all people. The United Nations Conference on Environment and Development at its meeting in Rio de Janeiro in 1992 adopted the Agenda 21 Strategy Document, which in Chapter 6 stresses specifically that particular attention should be directed towards food safety, with the priority placed on the elimination of food contamination.1 Both the Bible and the holy books of other religions contain restrictions that are essentially regulations on contaminants, chemical or microbiological, in food. Somewhere in its legislation every country has regulations to the effect that foods offered for sale for human consumption may not be contaminated to an extent that can cause disease or poisoning. The national legislation may limit itself to such general rules, or it may be much more detailed, giving specific regulations as to whether and on what conditions food in which contaminants are found may be sold. Such legislation often has the form of maximum limits or guideline values for the tolerable concentration of the contaminant in individual

264

Food chemical safety

foods or food groups. Whereas food in general, even in the recent past, was produced not far from where it was offered for sale and later consumed, the trend is now clearly towards foods being produced where it is economically most attractive. Moreover, the trend is towards sale of processed foods to the consumer, rather than raw foods for home cooking. Hence, trade in both commodities for food production and in finished food is ever increasing. Consequently, food legislation, including regulations for chemical contaminants in food, becomes an international issue, and efficient food regulation will in future be internationally based. Chemical contamination does not respect international borders. The contaminants are spread worldwide by air and water. Environmental organic contaminants and inorganic contaminants such as metals and metal compounds, nitrate and nitrite will be present in all foods, though sometimes in quantities below the limit of detection of the analytical methods of today. Moreover, foods as well as raw materials and ingredients for food production are to an increasing extent traded across borders. In 1961/62 the Food and Agriculture Organization (FAO) and the World Health Organization (WHO) established the Joint FAO/WHO Codex Alimentarius Commission (CAC) to elaborate international food standards and codes of practice for questions related to food. Questions concerning contaminants were dealt with, partly by the Commodity Committees and partly in the Codex Committee for Food Additives and Contaminants (CCFAC). The Codex Alimentarius system concerning contaminants is described in section 12.2. Section 12.3 deals with decision-making and enforcement. The Codex Alimentarius procedures for establishing Codex Standards and other texts are described. Once adopted by the CAC, Codex standards should be accepted by Member States, in accordance with the Codex statutes, in order to be binding. This happened, though only to a limited extent. The World Trade Organisation (WTO) agreements on Sanitary and Phytosanitary Measures (SPS) and Technical Barriers to Trade Agreement (TBT) establish basic rules for international trade.2 The Codex Alimentarius standards, guidelines and recommendations, being internationally agreed legislative standards for measures applied to protect human health from risks arising from contaminants and toxins, etc. in foods, are mentioned in the SPS Agreement as the basis for regulations concerning contaminants in food in international trade. The Codex standards, guidelines and recommendations are to be used as the basis for settling disputes in international trade in food. The introduction of the SPS and TBT Agreements, which came into force in 1995, gave impetus to the Codex Alimentarius system, and the Codex regulations became more important through this decision. The Codex General Standard for Contaminants and Toxins in Food (GSCTF) was accepted in 1997 by the CAC, the superior body of the Codex system, in the form of a Preamble with five Annexes.3 The GSCTF, however, does not yet contain figures pertaining to the maximum limits (MLs) for contaminants and

The international regulation of chemical contaminants in food

265

toxins in the various food groups. The MLs are presently under development by the CCFAC for the contaminants included in the GSCTF. Section 12.4 covers the GSCTF. Some recent developments and future trends in international regulation of chemical contaminants in food are discussed in section 12.5. It is attempted to foresee some issues that could be of interest and concern for food scientists and legislators dealing with contaminants in food in the early part of the new millennium. Certainly, persistent organic pollutants, many of which are chlorinated hydrocarbons, and some of which have been used in the past and in fact may still be used in parts of the world as pesticides, will be among those issues of justified concern. It is always important to locate sources where more detailed and in depth information may be sought. Section 12.6 discusses a number of such sources of information, and also offers some advice concerning the problems that most often will be met by food legislation administrators, producers and consumers. Similar problems are of concern both for those who are to administrate legislation concerning contaminants in food, and those producers and consumers who are legitimately concerned about potential chemical contaminants in the food they offer or find offered for sale.

12.2

The nature of international regulation: Codex Alimentarius

Environmental organic and inorganic contaminants will be present in all foods, even if in quantities below the limit of detection. Foods, raw materials and ingredients for food production are traded across borders to an increasing extent. It is only natural that there is a need for international food regulation, and this regulation takes place on a worldwide basis in the Codex Alimentarius. Regionally, too, international food regulation is harmonised in the European Union (EU), for Australia and New Zealand (ANZFA) and in other international fora. The Joint FAO/WHO Food Standard Programme and the Codex Alimentarius Commission elaborate international food standards and codes of practice for questions related to food. Codex Alimentarius is Latin and means food code. The purpose of the Joint FAO/WHO Food Standards Programme, as laid down in the statutes of the Codex Alimentarius Commission includes • protecting the health of the consumers and ensuring fair practices in the food trade • promoting co-ordination of all food standards work undertaken by international governmental and non-governmental organisations • determining priorities and initiating and guiding the preparation of draft standards, etc.

The CAC is the superior body of the Codex system, meeting regularly every second year, alternating between the FAO headquarters in Rome and the WHO

266

Food chemical safety

in Geneva. There are by 2001 165 Member Nations of the CAC, and the CAC sessions are attended by approximately 100 delegations. An Executive Committee acts on behalf of the Commission between sessions. The CAC sessions as well as the sessions of the subsidiary bodies such as the CCFAC are attended not only by government representatives, but also international professional organisations and consumer fora are also present, sometimes as members of the governmental delegations and sometimes as observers in their own right. The subsidiary bodies of the CAC are • the – – – – – – –

world-wide General Subject Codex Committees Codex Committee on Food Additives and Contaminants (CCFAC) Codex Committee on General Principles (CCGP) Codex Committee on Pesticide Residues (CCPR) Codex Committee on Food Hygiene (CCFH) Codex Committee on Residues of Veterinary Drugs in Food (CCRVDF) Codex Committee on Methods of Analysis and Sampling (CCMAS) Codex Committee on Nutrition and Food for Special Dietary Uses (CCNFSDU) – Codex Committee on Food Import and Export Inspection and Certification Systems (CCFICS) – Codex Committee on Food Labelling (CCFL) • the world-wide Commodity Codex Committees, such as the – Codex Committee on Natural Mineral Waters – Codex Committee on Fish and Fishery Products – Codex Committee on Fats and Oils • the six regional Co-ordinating Committees for Africa, Asia, Europe, Near East, Latin America and the Caribbean, and North America and the South West Pacific, respectively. In the early years, most work was done in the Commodity Codex Committees, and they succeeded in elaborating numerous Codex Standards, mostly for individual foods, for example the Codex Standard 13-1981 for Canned Tomatoes (CODEX STAN 13-1981). By 1994, 237 such commodity food standards containing ‘requirements for food aimed at ensuring the consumer a sound, wholesome food product, free from adulteration, correctly labelled and presented’ were agreed.4 However, it became clear that it would never become possible to agree standards for all foods. As international trade and consumer interest grew in the 1970s, emphasis shifted towards more flexible commodity standards, and many Commodity Committees completed their agenda and were adjourned sine die. In the 1990s General Standards treating contaminants and other issues on a horizontal basis were developed. This development accelerated in the 1980s and was established as a priority for future Codex work with the FAO/WHO Conference on Food Standards, Chemicals in Food and Food Trade in 1991.5 Horizontal standards apply both to food additives and to contaminants and toxins in food; establishing the General Standard for Contaminants and Toxins in Food (GSCTF) is the most important development concerning contaminants (see section 12.4).

The international regulation of chemical contaminants in food

267

For the purpose of Codex Alimentarius, a contaminant is defined as: ‘A substance not intentionally added to food, which is present in such food as a result of the production (including operations carried out in crop husbandry, animal husbandry and veterinary medicine), manufacture processing, preparation, treatment, packing, packaging, transport or holding of such food or as a result of environmental contamination. The term does not include insect fragments, rodent hairs and other extraneous matter.’6 For the work of Codex Alimentarius, it is important that pesticide residues and residues of veterinary drugs are treated in the general subject Codex Committees Codex Committee on Pesticide Residues (CCPR) and Codex Committee on Residues of Veterinary Drugs in Food (CCRVDF), respectively. In the WTO system, the term contaminant covers also pesticide residues and residues of veterinary medicine. Questions concerning contaminants were originally dealt with, partly by the many Commodity Committees, and partly in the Codex Committee for Food Additives, which became in the 1980s – when discussions on contaminants gained more prominence in the deliberations of the Committee – the Codex Committee on Food Additives and Contaminants (CCFAC). Now, the CCFAC, which meets every year, normally in The Hague, The Netherlands, is the appropriate forum in Codex for discussions concerning contaminants. The CCFAC agenda includes • endorsement or revision of maximum levels for contaminants in Codex Standards • completion of the GSCTF, by establishing maximum limits for all contaminants of importance for food safety and for trade in all relevant foods • establishment of priorities for evaluation of contaminants by the FAO/WHO Joint Expert Committee on Food Additives and Contaminants (JECFA).

The elaboration of Codex standards follows a stepwise procedure to ensure that all relevant parties, including governments, professional and consumer organisations and other Codex Committees that may have an interest in the content of a standard have several opportunities to express their opinion. The procedures are described in section 12.3. Once a standard is adopted by the CAC, it is published, and it is up to the Member States to accept the Standard and notify their acceptance to the Secretariat. There are also detailed guidelines laid down for the acceptance procedure for Codex standards.6 Though many Commodity Codex Standards were adopted in the 1970s and 1980s, there was considerable reluctance among Member States formally to accept the standards adopted. However, when the WTO in 1994 decided to use Codex Alimentarius regulations as the instrument to solve problems in international trade in food, the whole scenario changed. The importance of the uncertainty that may have been caused by this general backlog of formal acceptance of the standards decreased. The WTO Agreement on the Application of Sanitary and Phytosanitary Measures (WTO-SPS, 1994) and the Agreement on Technical Barriers to Trade (WTO-TBT, 1994) recognise as the international standards, guidelines and

268

Food chemical safety

recommendation for food safety those standards, guidelines and recommendations that are established by the Codex Alimentarius Commission. This decision meant that the work of Codex Alimentarius immediately gained momentum, as the Codex documents will be those that a WTO panel will use as the basis for settling disputes in international trade. If and when Member States wish to use their right to take sanitary and phytosanitary measures that are different, such measures shall be based on scientific principles and not be a disguised restriction on international trade (see section 12.3). There certainly is a need from time to time to revise provisions in established Codex standards. In order to take into account the development in general knowledge or changes in evaluation of problems related to food safety, to food production technology and to methods of analysis and sampling, there are procedures for how such revisions are done. When a standard established by a Commodity Committee, which has been adjourned sine die, is to be revised, the Codex Secretariat in co-operation with the national secretariat of the adjourned Committee will look after the revision.

12.3

Decision making and enforcement mechanisms

The Codex Alimentarius statutes and rules of procedure provide specific information about what form the Codex decisions can take (such as Standards, Guidelines and Codes of Practice etc.), and on how decisions are made. The elaboration of Codex standards follows a stepwise procedure to ensure that all relevant parties, including governments, professional and consumer organisations and other Codex Committees that may have an interest in the content of a standard have several opportunities to express their opinion (Fig. 12.1). Typically, the need for a draft standard for a contaminant in foods will be recognised by the CCFAC as a result of a discussion in the Committee. Such a discussion will often be based on a Position Paper produced by one or more Member States that can provide particular experience or expertise on the problem. The CCFAC Chairman and the Codex Secretariat will ensure that this Member State or states, sometimes supported by a drafting group, will provide a draft standard as basis for the further, more formal work. The draft standard will be sent for comments to the parties concerned and subsequently discussed first time in the CCFAC on step 3. When general agreement has been reached in CCFAC, the draft standard goes for adoption to the CAC or the Executive Committee at step 5. Member states may again comment at this stage. Following adoption by the CAC at step 5, a further round of discussions in CCFAC takes place at step 7. This may again lead to general agreement, perhaps a few Member States may express reservations, and the CCFAC decides to send the draft again to the Commission, now for final adoption by the CAC at step 8. The procedure appears to be designed to give ample time for consideration of any draft standard, given the frequency of the meetings of CCFAC and CAC. If there are problems in finding

The international regulation of chemical contaminants in food

Fig. 12.1

269

A brief overview of the stepwise procedure for the elaboration of Codex Standards and related texts.

agreement, a draft standard may well be kept at a certain step of the procedure for years. There are at present draft standards for lead, cadmium and tin as well as for several mycotoxins under development in the CCFAC, in view of inclusion into the GSCTF. The Codex Alimentarius Statutes contain provisions for voting, and each Member has one vote. Decisions of the Commission may be taken by a majority of the votes cast. However, traditionally much time and effort is spent to reach decisions by consensus, and in the large majority of cases these efforts are successful, leading to sustainable and durable standards. This principle has been endorsed several times recently, the latest in the FAO Conference on International Food Trade beyond 2000: Science-based Decisions, Harmonisation, Equivalence and Mutual Recognition in Melbourne in October 1999.7 The World Trade Organisation Sanitary and Phytosanitary Agreement (WTO-SPS) and Technical Barriers to Trade Agreement (WTO-TBT) regulate international trade. The Codex Alimentarius standards, guidelines and recommendations, being internationally agreed legislative standards for measures applied to protect human health from risks arising from contaminants and toxins etc., in foods, are mentioned in the SPS Agreement as the fundament for regulations concerning contaminants in food in international trade. Hence, the Codex standards, guidelines and recommendations are to be used as the basis

270

Food chemical safety

for settling disputes in international trade in food. The SPS and TBT Agreements accepted in Marrakech in 1994 came into force in 1995. They gave impetus to the Codex Alimentarius system, and the Codex regulations achieved more importance through this decision. The SPS Agreement applies to sanitary and phytosanitary measures which may affect international trade. Such measures include laws, regulations, requirements and procedures, including inter alia end product criteria, sampling and testing procedures, as well as packing and labelling requirements directly related to food safety. Measures applied to protect life and health from risks from contaminant and toxins in food and beverages are specifically mentioned. In WTO-SPS connection, ‘contaminants’ include residues of pesticides and veterinary drugs and extraneous matter, whereas these are not included in the Codex definition of a contaminant. In the Codex system residues of pesticides and veterinary drugs are dealt with specifically in the general purpose Codex Committees on Pesticide Residues (CCPR) and Codex Committee on Residues of Veterinary Drugs in Food (CCRVDF), respectively. The SPS Agreement does not differentiate between Codex standards, guidelines and recommendations. All such texts have similar status, and how a text would be applied depends on its substantive content rather than on the category of the text. The WTO Members have the right to establish their own level of protection, which means that they may choose their own sanitary measures necessary for the protection of life or health, provided they are not inconsistent with the provisions of the SPS Agreement. Such measures applied shall be based on scientific principles and may not be maintained without sufficient scientific evidence, except in the case where relevant scientific evidence is insufficient. Then measures may be taken on a provisional basis, on the basis of available pertinent information, and they shall reviewed within a reasonable period of time. These provisions in the SPS Agreement are often referred to as ‘the precautionary principle’. WTO Member States shall notify changes in their food regulations to WTO, as they are sanitary measures covered by the SPS Agreement, and they shall provide information about them in accordance with the agreement. They are also obliged to accept the regulations, etc., of other Member States as equivalent, if the other Member State can show that their measures achieve the appropriate level of health protection. If so requested, Member States shall enter into consultations in order to achieve agreement on recognition of the equivalence of specified sanitary measures. Disputes shall be settled in accordance with the General Agreement on Trade and Tariffs (GATT) provisions. A WTO panel shall seek advice on scientific or technical issues from experts chosen by the panel and the parties involved in the dispute. It is important to note that the Codex Alimentarius texts are the basis for decisions taken in such connections. When a country introduces regulations that are more detailed, or stricter, than similar Codex provision, that country should be prepared to defend the measures taken on scientific grounds in front of a WTO panel.

The international regulation of chemical contaminants in food

12.4

271

The Codex General Standard on Contaminants and Toxins in Food

In the Codex Alimentarius system, questions concerning contaminants were originally dealt with partly by the many Commodity Committees and partly in the Codex Committee for Food Additives. In the early 1980s, when discussions on contaminants became more prominent on the agenda of the Committee, the name was changed to the Codex Committee on Food Additives and Contaminants (CCFAC). Nowadays the CCFAC is the forum in Codex for all discussions concerning contaminants. Every year, the CCFAC agenda includes as a major item the completion of the General Standard for Contaminants and Toxins in Food (GSCTF). The intention is to establish maximum limits for all contaminants of importance for food safety and for trade in all relevant foods. As a result of an analysis provided by an independent consultant, Dr W. H. B. Denner, the CCFAC decided in 1989 to begin work, which will eventually lead to a General Standard for Food Additives (GSFA).8 The draft GSFA is a horizontal standard for food additives, providing maximum limits for food additives in all foods. The same work provided inspiration to a discussion concerning contaminants in the 23rd CCFAC. It was decided to accept an offer from the Dutch and the Danish delegations to develop a paper on philosophy and procedure concerning the setting of Codex maximum and guideline levels for contaminants for discussion in the next session of the CCFAC.9 Having decided in 1991 to develop the GSFA, it was not surprising that the debate about the Dutch–Danish philosophy paper in the 24th CCFAC the following year emerged into a decision to develop a General Standard on Contaminants – and later also Toxins – in Food (GSCTF). The GSCTF would lay down provisions for all contaminants important for the health of the consumer as well as for the international trade in food.10, 11 The Dutch and the Danish delegations to the CCFAC (Dr David Kloet and the author) have since 1992 continued to provide the draft papers for the GSCTF. Six years later, in 1997, the superior body of the Codex Alimentarius, the CAC, accepted the CCFAC proposal for a GSCTF at the final step 8, in the form of a Preamble with five Annexes. The five annexes cover, respectively: 1. 2. 3. 4. 5.

Criteria for the Establishment of Maximum Limits in Food Procedure for Risk-management Decisions Format of the Standard Annotated List of Contaminants and Toxins Food Categorization System to be used in the GSCTF.

The GSCTF, however, does not yet contain figures pertaining to the maximum limits (MLs) for contaminants and toxins in the various food groups. The MLs are presently under development by the CCFAC for the contaminants, which are to be included in the GSCTF. This is done in the form of draft standards for each individual contaminant in foods.

272

Food chemical safety

The Dutch–Danish 1991 philosophy paper highlighted that the Codex definition of a contaminant does not include inherent natural toxins (e.g. the glucosinolates and phycotoxins), since these substances are present in food as a result of the metabolic processes in the organism. It was then decided to have the GSCTF also to include such toxins, as they may in many ways be similar to contaminants. Many toxins are at least as toxic to humans as most contaminants, and they may also cause problems in international trade.10 It was also recommended and agreed by the CCFAC that the GSCTF should be based upon a horizontal approach, i.e. covering the important contaminants in all relevant foods, and that the MLs should be set as low as reasonably achievable – the ALARA principle. The paper recommended that a decision on whether international action should be taken on a contaminant in food should be based upon the following criteria: • The substance is demonstrated to be present in the food at a certain level, which is determined by reliable analysis. • The substance is of toxicological concern at this level. • The foodstuff for which action is to be taken plays a sufficiently important role in the intake of the substance concerned. • The foodstuff appears in international trade.

In parallel to the development of the GSCTF, the CCFAC discussed and decided on which contaminants are considered immediately to comply with the criteria. This is being done through discussions based upon Discussion or Position Papers developed by one, or a few, delegations, which have a particular expertise or experience with the contaminant concerned. The GSCTF contains provisions concerning the content of such papers. If the CCFAC deems it appropriate, a draft standard is to be developed. In order to have a sufficient toxicological basis, the JECFA will be asked to assess the contaminant, if there is not already a recent JECFA evaluation available. At present there are draft standards for the contaminant metals lead, cadmium and tin in the pipeline. The Draft Standard for Lead in Food provides MLs for the many foods in which lead may be found in amounts that are significant for the total exposure of the consumer, as lead is a ubiquitous contaminant. The draft standard was agreed by the CAC in 1997 on step 5. It has been further developed by the CCFAC since then. For tin, the draft standard agreed by the CAC on step 5 in 1999 covers canned foods only. The draft standard for cadmium is discussed on step 3, and step 6 for cereals, pulses and legumes. A JECFA re-evaluation for tin and cadmium took place in June 2000, as a sine qua non for further progress. There are certain provisions concerning mercury in the form of Guideline Levels for methylmercury in fish (CAC/GL 7, 1991), whereas for arsenic it has been agreed that MLs shall be laid down only for certain toxic arsenic species or arsenic compounds. Most arsenic is present in food in the form of arsenic species of no toxicological concern.12 A draft standard for arsenic in food shall be developed only when appropriate routine methods of analysis become available.

The international regulation of chemical contaminants in food

273

For zinc, copper and iron, the 31st CCFAC in 1999 recommended that the relevant Commodity Committees take appropriate steps to remove the provisions concerning zinc, copper and iron from the contaminant section of all commodity standards. These provisions should be considered as quality parameters and can be included in the standards as such.13 Moreover, there are draft standards for the mycotoxins aflatoxins, ochratoxin A and patulin, and Position Papers are becoming available for fumonisins and zearalenone. For aflatoxins, an ML of 15 g/kg for total aflatoxins in peanuts intended for further processing was agreed by the 23rd CAC in 1999, as well as a draft sampling plan, on an interim basis, to be further developed by the CCFAC and the CCMAS. A proposal for an ML for aflatoxin M1 in milk is also in the pipeline, whereas a Codex Code of Practice for the reduction of aflatoxins in raw materials and supplementary feedstuffs for milk-producing animals was adopted by the CAC in 1997. For ochratoxin A, a Position Paper from Sweden was agreed in the 31st CCFAC in 1999,14 as well as a proposal to develop a Code of Practice for the Prevention of Contamination by Ochratoxin A in Cereals. Moreover, a draft ML of 5 g/kg for ochratoxin A in cereals and cereal products was proposed. This proposal was circulated to governments and international organisations for comments at step 3, including suggestions for a sampling plan. The CCFAC at the same time asked JECFA for a risk assessment on the levels of 5 and 20 g/kg ochratoxin A in cereals and cereal products. There may well be included foods other than cereals, such as coffee, wine and grape juice in a future standard for ochratoxin A, as cereals are the most important source of this contaminant in food, but far from the only one. France has provided the Position Paper for patulin.15 A draft ML for patulin in apple juice and apple juice ingredients in other beverages of 50 g/ kg was agreed by the CAC in 1999 on step 5, while further comments were sought at step 6 for possibly justifying a lower ML of 25 g/kg. In 2000, the CCFAC agreed to forward this ML to the CAC for adoption at step 8, with the understanding that the level will be incorporated in the GSCTF. An offer from the UK to develop a draft Code of Practice for the Prevention of Patulin Contamination in Apple Juice was accepted by the 32nd CCFAC in March 2000. For zearalenone, a Position Paper produced by Norway was finalised in the 2000 CCFAC, and this delegation is also elaborating a draft Code of Practice for the Prevention of Contamination of Cereals by Zearalenone.16 Whereas the draft Code of Practice will be circulated for comments at step 3, there is no indication that an ML will be proposed in the near future. For fumonisins, the USA presented a Position Paper to the 32nd CCFAC in 2000.17 This paper was finalised in the 33rd CCFAC, and fumonisins as well as ochratoxin A and other mycotoxins were evaluated by JECFA in February 2001. Hence, for some mycotoxins, such as aflatoxins, patulin and zearalenone, individual draft Codes of Practice are being developed, in order to reduce the

274

Food chemical safety

problems at the source. Moreover, the 32nd CCFAC agreed in March 2000 to create also a General Code of Practice for the Prevention of Mycotoxin Contamination in Cereals, and the USA accepted to direct this work. In general, application of source-related measures as an efficient tool to reduce food contamination is another of the ideas from the philosophy paper that gained support. Sweden is preparing a specific paper on source-directed measures to reduce contamination of food with chemicals.18 The problems related to dioxins and dioxin-like PCBs in foods of animal origin, caused by contamination of feedstuffs, have served to highlight the need for accelerated progress in this field. A revised Discussion Paper on Dioxins from the Netherlands was a major item for the 32nd CCFAC, and a Position Paper will be produced for the 33rd CCFAC in 2001, exploring inter alia the arguments for and against setting MLs for dioxins and dioxin-like PCBs in foods. The GSCTF is at present only partially complete. The framework is accepted by the CAC, and there is considerable consensus on which contaminants shall be included first. For some of the contaminants, a consensus that covers the MLs in various foods appears to be within reach. For others there remain a JECFA assessment, or re-assessment, as well as another thorough discussion in CCFAC. The house of the GSCTF is built, but the apartments reserved for some individual contaminants still remain empty. Some potential tenants are more or less ready to move in, whereas others still are not quite ready.

12.5

Future trends

The increase in international trade in food is one factor that drives international regulation of contaminants in food. Enhanced food safety problems, real or perceived, is another factor. The occurrence of chemical contaminants in food, no matter whether it is caused by industrial or environmental contamination, or a result of production processes, are among such problems. Authorities respond to consumer concerns by more stringent food safety assurance systems, including both lower MLs and more efficient control, but also self-control and certification control systems. In a world where so much trade goes across national borders, the response will, to an increasing degree, become international regulations, framed in the Codex Alimentarius system, or by a group of trade partner countries like the EU or the Association of South East Asian Nations (ASEAN). The trend is towards international regulation in the form of Codex Alimentarius MLs for important contaminants in the relevant foods. The development of cleaner production methods and better methods of analysis support the trend towards lower MLs for contaminants in foods. A logical consequence will be the subsequent need for further development of common international standards for validated methods of sampling and chemical analysis. The SPS Agreement of the WTO highlights the need for scientific justification for measures like MLs. However, there are cases of concern where

The international regulation of chemical contaminants in food

275

the relevant scientific evidence is insufficient, and possibly will remain so for some time, so it is recognised as legitimate to base provisional measures on the pertinent information available. In such cases a government or an international group of countries may apply a precautionary principle when framing regulations in order to protect the health of the consumer. Application of the precautionary principle in international regulation on contaminants in food remains under discussion, in particular in the CCGP, where proposals for inclusion of the principle and guidelines for its use have been tabled. The precautionary principle has been defined as: ‘A general customary rule of international law or at least a general principle of law, the essence of which is that it applies not only in the management of risk, but also in the assessment thereof.’19 There is, however, not yet general agreement about this or any other definition. The move in favour of accepting concerns other than those that can be strictly scientifically argued appears to be gaining momentum, and the trend is towards governments acknowledging the need to protect consumer confidence and hence taking action and legislating beyond what is fully scientifically based. The first full discussion of these ‘other legitimate factors’ took place in the CAC in 1993 and in the last few years the discussion has gained momentum.20 Taking into account such factors is advocated both by governments and consumers, and the discussion about approval of Maximum Residue Levels for the hormone bovine somatotropin (BST) highlighted the need for further clarification of these other legitimate factors on an international level.21 Among such factors in the 32nd CCFAC were mentioned:22 • economic costs associated with the establishment of MLs and methods of analysis • general technical and technological need and feasibility • availability of resources to undertake analysis and enforce standards • prevention of environmental contamination through the use of Source Directed Measures • the proper use of labelling to inform consumers • consumer concerns related to the safety of food additives and contaminants.

It is also significant that in connection with the development of the Codex GSCTF it was decided to use MLs and discontinue with the use of other terms like ‘guideline levels’. The same idea is expressed in the SPS Agreement where there is no distinction between the application of measures whether they are called international standards, guidelines or recommendations. In summary, the most important trends in international regulation of chemical contaminants in food are the trends towards: • • • •

taking into account also ‘other legitimate factors’ international MLs, rather than national fixed MLs, rather than guideline levels lower MLs, rather than higher.

276

Food chemical safety

When considering for which contaminants it is relevant to lay down maximum limits in relevant foods, some contaminants, which have so far been unnoticed or may have been disregarded, will become pertinent. Improved chemical methods of analysis help to identify and quantify contaminants that may be or become of concern, such as brominated flame-retardants. Improved methods of analysis will also serve to separate, e.g. PCBs and dioxin congeners, and it will be possible to analyse separately for organometal species of different toxic potentials, such as the different organic arsenic or organotin compounds that may be found in fish and seafood. When the different chemical species are sufficiently stable in food to have different toxicological properties, and when it becomes possible to perform control analysis reliably, there will be every good reason to legislate individually for the different chemical species. This should be done taking into account their toxicity, etc. In future, regulation of some chemical contaminants will be based on speciation rather than on total PCBs or the sum of metal species of different properties.23 Allergy is an important and increasing problem world-wide. Whereas most food allergies are related to natural ingredients in foods, there are also some allergies that may be provoked by food additives. There may also be cases where an allergy is either provoked by a contaminant in food, or may otherwise be related to a contaminant. So far, problems related to allergy have generally not been taken much into account when food legislation was laid down. In summary, trends with respect to the chemical contaminants that may be considered for regulation in future will comprise: • contaminants so far unnoticed • speciation of contaminants so far treated only as a sum • contaminants so far disregarded, such as allergy-provoking contaminants.

12.6

Sources of further information and advice

A review of the international legislation concerning food was recently presented in International Standards for Food Safety.24 This review contains a chapter on the Codex Alimentarius General Standard for Contaminants and Toxins in Food where more information about the development and the content of the GSCTF can be found.25 There are also other chapters of significance for readers who wish to seek further information on specific subjects concerning international food legislation. To follow the future development of the GSCTF, and the preparatory documents produced in order to promote the standard, the best way would be to attend the CCFAC regularly, but much could also be achieved by following the paperwork on the web-site. Progress in scientific research concerning contaminants in food will often be reported in international journals such as Food Additives and Contaminants,

The international regulation of chemical contaminants in food

277

while questions related both to science and legislation may be discussed in Regulatory Toxicology and Pharmacology. Quick access to day-to-day news is found, e.g., in World Food Chemical News.

12.7 1. 2. 3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14.

References Agenda 21 Strategy Document, United Nations Conference on Environment and Development, Rio de Janeiro, 1992. Agreement on the Application of Sanitary and Phytosanitary Measures, Legal Texts, Uruguay Round of Multilateral Trade Negotiations, World Trade Organization, Geneva, Switzerland, 1995. Preamble and Annexes, Codex General Standard for Contaminants and Toxins in Food, CODEX-STAN 193-9995 (Rev. 1-1997), Joint FAO/ WHO Food Standards Programme, FAO, Rome, 2000. MILLER, R. W., This is Codex Alimentarius, Joint FAO/WHO Food Standards Programme, FAO, Rome, 2nd Edition, 1994. Report of the FAO/WHO Conference on Food Standards, Chemicals in Food and Food Trade, FAO/WHO Conference on Food Standards, Chemicals in Food and Food Trade, FAO, Rome, 1991. Codex Alimentarius Commission, Procedural Manual, 11th Edition, Joint FAO/WHO Food Standards Programme, FAO, Rome, 2000. Science-based Decisions, Harmonization, Equivalence and Mutual Recognition, Report of the Conference on International Trade beyond 2000, Melbourne, Australia. FAO, Rome, 1999. DENNER, W. H. B., Future Activities of the Committee in regard to the Establishment and Regular Review of Provisions Related to Food Additives in Codex Standards and Possible Mechanisms for the Establishment of General Provisions for the Use of Food Additives in Non-Standardised Foods, CX/FAC 89/16. Rome, Codex Alimentarius Commission, 1989. Report of the 23rd Session of the Codex Alimentarius Committee on Food Additives and Contaminants, 1999. BAL, A. and BERG, T., Contaminants in Food. Towards a Codex Approach, CX/FAC 92/10, Codex Alimentarius Commission, Rome, 1991. Report of the 24th Session of the Codex Alimentarius Committee on Food Additives and Contaminants. ALINORM 93/12, paragraphs 2–4 and 64– 78, Codex Alimentarius Commission, Rome, 1992. Position Paper on Arsenic. CX/FAC 99/22, Codex Alimentarius Commission, Rome, 1998. Report of the 31st Session of the Codex Alimentarius Committee on Food Additives and Contaminants. ALINORM 99/12A, paragraph 96, Codex Alimentarius Commission, Rome, 1999. Position Paper on Ochratoxin A. CX/FAC 99/14, Codex Alimentarius Commission, Rome, 1998.

278

Food chemical safety

15.

Position Paper on Patulin. CX/FAC 98/17, Codex Alimentarius Commission, Rome, 1997. Position Paper on Zearalenone. CX/FAC 00/19, Codex Alimentarius Commission, Rome, 2000. Position Paper on Fumonisins. CX/FAC 00/22, Codex Alimentarius Commission, Rome, 1999. Source-Directed Measures to Reduce Contamination of Food with Chemicals, CL 1999/23-FAC Codex Alimentarius Commission, Rome 1999. Report of the Appellate Body: EC Measures Concerning Meat and Meat Products (Hormones). World Trade Organization, Geneva, 1998. Report of the 20th Session of the Codex Alimentarius Commission. Codex Alimentarius Commission, FAO, Rome, 1993. MCCREA, D., ‘A View from Consumers’ in REES, N. and WATSON, D., International Standards for Food Safety, Gaithersburg, Maryland, Aspen Publishers, 2000. Report of the 32nd Session of the Codex Alimentarius Committee on Food Additives and Contaminants. ALINORM 01/12, paragraphs 143–147, Codex Alimentarius Commission, Rome, 2000. BERG, T. and LARSEN, E. H., ‘Speciation and legislation – where are we today and what do we need for tomorrow?’ Fresenius J Anal Chem, 1999, 363 431–434. REES, N. and WATSON, D., International Standards for Food Safety, Gaithersburg, Maryland, Aspen Publishers, 2000. BERG, T., ‘Development of the Codex Standard for Contaminants and Toxins in Foods’ in REES, N. and WATSON, D. International Standards for Food Safety, Gaithersburg, Maryland, Aspen Publishers, 2000.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

13 The regulation of chemical contaminants in foodstuffs in the European Union S. A. Slorach, National Food Administration, Uppsala

13.1

Introduction

This chapter briefly describes how European Community (EC) legislation on chemical contaminants in foodstuffs is developed, adopted and enforced. In this context the term ‘chemical contaminants’ covers residues of pesticides and veterinary drugs, heavy metals, mycotoxins and nitrate. Community procedures for preparing and adopting legislation on contaminants in foodstuffs involve, among others, the European Commission (‘the Commission’) and its scientific advisory bodies, the Council of the European Union (the ‘Council’), the 15 Member States of the European Union (EU) and the European Parliament. Proposals for new legislation are made by the Commission. Decisions on new legislation on contaminants in foodstuffs are generally made by the Council, unless authority to do so has been delegated to the Commission. In addition to the complications arising from the above situation, different procedures are used for different classes of substance (for example, the procedure for elaborating legislation on mycotoxins is different from that for veterinary drug residues) and the procedures have changed over the years and are still changing. Responsibility for initiating work on legislation on pesticide residues, mycotoxins, heavy metals and nitrate rests with the Commission’s Directorate-General for Health and Consumer Protection (usually referred to as DG SANCO), whereas veterinary drug residues are the responsibility of the Commission’s Directorate-General for Enterprise. A brief outline of the procedures used in developing and enforcing regulations on chemical contaminants is given here and readers wishing to immerse themselves in the detail, of which there is plenty, should consult the references given at the end of the chapter.

280

Food chemical safety

All Community legislation and proposals for legislation are published in the Official Journal of the European Communities (the ‘Official Journal’). Legislative instruments are mainly of two types – Regulations and Directives. Regulations are binding in their entirety and apply directly (verbatim) in all Member States. Directives are binding, as to the result to be achieved, but allow the national authorities the choice of form and methods: they are transposed into the national legislation of the Member States and may therefore differ slightly from country to country. In addition, there are Decisions, which are binding in their entirety upon those to whom they are addressed. Recommendations and Opinions, on the other hand, have no binding force.

13.2

Scientific advisory committees

Most of the expert advice that forms the scientific basis of EC legislation on chemical contaminants in food is provided by the Commission’s scientific advisory committees working in the food safety area. The experts on these committees are expected to provide independent advice and not represent their countries or the organisations that employ them. The results of the deliberations of the scientific advisory committees are published and are also available on the Internet at DG SANCO’s web site (http://europa.eu.int/comm/dgs/ health_consumer/index_en.htm). In 1997, the Commission transferred responsibility for these committees to DG SANCO and restructured the system. It established a Scientific Steering Committee and defined the mandates for eight scientific advisory committees (Commission Decision 97/579) (Fig. 13.1). The aim of the reform of the system was to increase the independence of the expert committees from national government and other interests and to increase transparency. Members of the expert committees are required to declare any interests they may have that could affect their impartiality in dealing with a particular subject or substance on a committee’s agenda. The committees most involved in setting limits for contaminants in food are the Scientific Committee on Food, the Scientific Committee on Veterinary Measures Relating to Public Health and the Scientific Committee on Plants. Prior to autumn 1997, there was a separate Scientific Committee for Pesticides, but its responsibilities are now included in the mandate of the Scientific Committee on Plants. In order to facilitate their work, many of the scientific committees have subsidiary working groups which carry out a lot of the preparatory work that needs to be done (preparation of working documents, etc.) before opinions can be given by the committee. The Scientific Committee on Food has a Working Group on Contaminants (see Fig. 13.2). In November 2000, the Commission proposed the creation of a European Food Authority (EFA), which would take over responsibility for much of the food safety risk assessment work at present carried out by the Commission and

The regulation of chemical contaminants in foodstuffs in the EU

Fig. 13.1

281

European Commission Scientific Advisory Committees.

its advisory bodies. Thus if and when the EFA is established and comes into operation (perhaps in 2002), it will assume responsibility for providing risk assessments as a basis for EC legislation on contaminants in food.

Fig. 13.2

SCF Working Group Structure.

282

13.3

Food chemical safety

Pesticide residues

The procedures used to set permanent maximum residue levels (MRLs) for pesticide residues is described briefly below. In parallel with this, proposals for provisional MRLs are prepared in connection with the work on regulating the placing of plant protection products on the market according to Council Directive 91/414. 13.3.1 Toxicological evaluations The DG SANCO’s Working Group on Pesticide Residues proposes MRLs for pesticide residues in foodstuffs. For each pesticide, MRLs are proposed for residues in the relevant individual commodities/crops. Since a large number of pesticides have to be dealt with, the workload is spread by appointing each Member State in the Working Group as Rapporteur for a number of specified pesticides. When proposing an MRL, the Rapporteur Member State (RMS) identifies the Acceptable Daily Intake (ADI) and Acute Reference Dose (ARfD) for man that is valid for the pesticide in question. The ADI thus identified is often the same as that recommended by the Joint FAO/WHO Meetings on Pesticide Residues (JMPR), whose recommendations on ADIs and MRLs are used within the Codex Alimentarius system. If the ADI proposed by the RMS is not that recommended by JMPR, the RMS has to provide an explanation for the difference. The other Member States comment on the RMS proposal at meetings of the Working Group on Pesticide Residues. If the Member States cannot reach agreement on the evaluation, the matter is referred to one or more of the Commission’s scientific advisory committees. Prior to autumn 1997, such questions were referred to the Scientific Committee for Pesticides, but since then they have been referred to the Scientific Committee on Plants. 13.3.2 Residue data Data on the levels of residues found in supervised trials, in which the pesticide is used according to Good Agricultural Practice (GAP), are the main basis for proposing MRLs. Such trials must include studies that reflect the use that results in the highest residue levels or critical GAP. MRLs are established for individual crops or groups of crops. In certain cases, data from supervised trials on one crop may be extrapolated to another crop. The results of the supervised trials are evaluated by the RMS and then discussed by the Working Group on Pesticide Residues. 13.3.3 Pesticide intake calculations Although the main basis for proposing MRLs is the data from supervised trials, when considering the proposals Member States calculate the theoretical maximum daily intake that could occur if the proposed MRLs were adopted,

The regulation of chemical contaminants in foodstuffs in the EU

283

using the World Health Organization (WHO) European Diet or national dietary information as the basis for their calculations. If the theoretical maximum daily intake exceeds the proposed ADI, Member States may not be prepared to accept the proposed MRL without further refinement of the intake calculations. The EC takes into consideration not only chronic exposure to pesticide residues via foodstuffs, but also acute exposure. The acute exposure is assessed in accordance with the procedures and practices used in the EC, taking account of guidelines published by WHO. 13.3.4 Preparation and adoption of legislation Nowadays, EC MRLs are laid down according to the procedures in Directive 97/41, which amended Directives 76/895, 86/362, 86/363 and 90/642. Earlier decisions on MRLs were made by a Council procedure, i.e. the Council made the final decision. However, they are now usually decided by the Commission according to a Regulatory Committee decision procedure (Procedure IIIb) involving the Commission’s Standing Committee on Plant Health. This Standing Committee consists of representatives of the Member States, but is chaired by the Commission. If the members of this Standing Committee cannot agree, i.e. there is not a qualified majority for a proposal, the question is referred to the Council for a decision. The Council shall act by a qualified majority. If the Council fails to reply within the stipulated time limit, the proposal shall be adopted by the Commission, except where the Council has decided against the measure by a simple majority. Before the proposals for MRLs are notified to the World Trade Organization, they are sent to the Scientific Committee on Plants. 13.3.5 EC legislation on maximum residue levels Council Directive 76/895 (and Directives 81/36, 82/528, 88/298, 2000/24, 2000/ 57 and 2000/82, which contain amendments to that Directive) contains recommendations for MRLs for pesticide residues in or on fruit and vegetables. However, these MRLs are not mandatory and Member States may set higher MRLs in their national legislation, but not lower levels. Council Directives 86/362, 86/363 and 90/642 contain MRLs for pesticide residues in or on cereals, foods of animal origin and fruit and vegetables, respectively. These MRLs are mandatory and must be incorporated into the national legislation of the Member States. These Directives have been amended several times (Directives 88/298, 93/57, 94/29, 95/39, 96/33, 98/82, 1999/71 2000/24, 2000/42, 2000/48, 2000/58, 2000/81 and 2000/82 concerning MRLs for cereals and foods of animal origin and Directives 93/58, 94/30, 95/38, 95/61, 96/32, 98/82, 1999/71, 2000/24, 2000/42, 2000/48, 2000/57, 2000/58, 2000/81 and 2000/82 concerning MRLs for fruit and vegetables). Unfortunately, official consolidated versions of the amended Directives are not produced, which makes it difficult to get a complete picture of all the MRLs that have been adopted to

284

Food chemical safety

date. However, an unofficial list of EC MRLs can be found on the web site of DG SANCO. 13.3.6 Interaction with Codex Many EU Member States participate actively in the work of the Codex Committee on Pesticide Residues (CCPR), which is hosted by The Netherlands. Before and during each CCPR meeting, EC positions are co-ordinated as far as possible. In view of the special status attached to Codex MRLs since the signing of the Agreement on Sanitary and Phytosanitary Measures (the SPS Agreement), the EC now attaches great importance to Codex work. 13.3.7 Enforcement Enforcement of EC regulations on pesticide residues in foodstuffs is the responsibility of the competent authorities in the Member States. The authorities in each Member State should ensure that the products produced in that country and those imported directly from countries outside the EU (‘Third Countries’) comply with EC legislation. In principle, when such a system is fully developed and operational in all Member States, there should be no real need for countries to examine products coming from other Member States. Each year the Commission issues recommendations concerning co-ordinated Community monitoring programmes to ensure compliance with the MRLs in and on cereals, fruit and vegetables, etc. The latest recommendation was published as Commission Recommendation 2001/42 in January 2001. It contained detailed information on the pesticide residue/ product combinations to be monitored, the number of samples to be taken by each country and quality control procedures for the analysis, etc. In addition to the recommended (minimum) monitoring programme, the different Member States carry out their own individual programmes, which can vary in both content and scope. Member States are encouraged to publish the results of their pesticide control work and the results of the recommended Community programme are collected and published by the Commission. The Commission’s Food and Veterinary Office (FVO) carries out inspections in Member States to check that they have incorporated EC legislation into their national regulations and that they are enforcing it properly. The reports of these inspections are placed on the Internet at DG SANCO’s web site.

13.4

Veterinary drug residues

13.4.1 Introduction The procedures used within the Community for establishing MRLs for veterinary medicinal products in foodstuffs of animal origin are laid down in Council Regulation 2377/90, which came into force on 1 January 1992. The

The regulation of chemical contaminants in foodstuffs in the EU

285

most important principle laid down in these procedures is that foodstuffs obtained from treated animals must not contain residues that might constitute a health hazard for the consumer. When calculating MRLs, the aim is to ensure that the total intake of residues of the substance via foodstuffs of animal origin does not exceed the ADI. For the purposes of these calculations, the bodyweight of the consumer has been assumed to be 60 kg and the daily intakes of various foods has been assigned certain values, e.g. milk 1500 g, muscle 300 g, liver 100 g. In cases where a substance is also used as a pesticide, the intake from such use is also taken into account. The procedures described below apply to veterinary drugs used as medicines, but not to the medical substances used as feed additives. MRLs are determined by the Committee for Veterinary Medicinal Products (CVMP), and its Safety Working Party, attached to the European Medicine Evaluation Agency (EMEA) in London. Different procedures are used, depending on when the pharmacologically active substance was first authorised for use. For medicines containing substances authorised after 1 January 1992 (‘New substances’), MRLs must be set at the European level for all pharmacologically active substances, including excipients, before approval procedures can be started in the Member States. MRLs for medicines authorised before 1 January 1992 (‘Old substances’) must have been evaluated by 31 December 1999 if their use after that date is to continue. Regardless of the procedure to be followed in setting the MRL, the manufacturer of a veterinary medicinal product must provide safety and residue dossiers containing the information required to set an ADI and MRLs. The safety dossier contains the pharmacodynamic, kinetic, metabolic and toxicity data and the residue dossier the data on kinetics, metabolism and residues, as well as the analytical method(s) for the substance. Regulation 2377/90 contains the following four annexes in which the substances are listed after evaluation: I. Substances for which final MRLs have been fixed; II. Substances for which MRLs are not deemed necessary in order to protect public health; III. Substances with provisional MRLs – if a dossier is incomplete, the manufacturer may be given a set time (up to five years) in which to provide the necessary information; IV. Substances for which it is not possible, due to safety concerns, to set an MRL – the administration of substances listed in this annex is prohibited throughout the EU and the marketing authorisation for the medicines concerned has been withdrawn. 13.4.2 MRLs for substances authorised after 1 January 1992 Applications from industry are submitted to the EMEA. Since a large number of substances have to be dealt with, the workload is spread by appointing each Member State represented in the CVMP as rapporteur or co-rapporteur for a number of specified veterinary drugs. Using the information in the dossier provided by the manufacturer, the Rapporteur Member State (RMS) proposes an ADI for the drug in question and MRLs for relevant foodstuffs of animal

286

Food chemical safety

origin (e.g. muscle, liver, milk), using the guidelines in Regulation 2377/90 and in the Rules Governing Medicinal Products in the European Community, Volume VI (soon Volume VIII). The ADI and MRLs are often, but not always, the same as those proposed by JECFA or JMPR for the same drug. In the CVMP other Member States then comment on the ADI and MRLs proposed by the RMS. When the CVMP has reached agreement on MRLs for a drug, they are submitted to the Commission, for adoption by the Committee for the Adaptation to Technical Progress of the Directives on the Removal of Technical Barriers to Trade in the Veterinary Medical Products Sector. If a qualified majority of the Member States in that Committee supports adoption, the MRLs are then incorporated into the relevant annex to Council Regulation 2377/90. If a qualified majority is not obtained, the Commission proposes to the Council the measures to be adopted: the Council acts by a qualified majority. If the Council has not acted within three months, the proposed measures are adopted by the Commission, unless the Council has voted against them by a simple majority. All amendments to the annexes of Regulation 2377/90 are published in the Official Journal. 13.4.3 MRLs for substances authorised before 1 January 1992 After examining dossiers supplied by industry, Rapporteur Member States in the CVMP’s Safety Working Party (SWP-V) propose ADIs and MRLs for substances authorised before 1 January 1992. These proposals are then discussed by the CVMP. When the CVMP has reached agreement on MRLs, they are then adopted by the procedure described above for ‘New substances’. 13.4.4 Withdrawal periods The ‘withdrawal period’ is the time between the last dose given to the animal and the time when the level of residues in the tissues (muscle, liver, kidney, skin/fat) or products (milk, eggs, honey) is lower than or equal to the MRL. For veterinary medicinal products intended to be marketed in only one Member State, withdrawal periods are set at the national (Member State) level. This is also the case for all the old substances authorised before 1 January 1992. For products intended to be used in more than one Member State, the mutual recognition procedure has been obligatory since 1 January 1998. In this procedure evaluation is carried out in one country and the proposed withdrawal period is then accepted (or rejected) by other Member States. If the proposal is not accepted, the matter can go to arbitration at EMEA. For products intended to be marketed throughout the whole of the European Union, withdrawal periods are determined by the CVMP by a central procedure analogous to that used for MRLs. This procedure must always be used for certain special groups of substances, e.g. those produced by biotechnology and innovative products.

The regulation of chemical contaminants in foodstuffs in the EU

287

13.4.5 Interaction with Codex Many EU Member States and the Commission take an active part in the Codex work on MRLs for veterinary drugs, especially in the Codex Committee on Residues of Veterinary Drugs in Food (CCRVDF). Before and during each CCRVDF meeting, EC positions are co-ordinated as far as possible. In view of the special status attached to Codex MRLs since the signing of the SPS Agreement, the EC now attaches great importance to Codex work. The MRLs proposed by JECFA and JMPR and discussed in CCRVDF are in many, but not all, cases accepted by the Member States of the EU. The most notable exceptions to this in recent years are the MRLs for hormones used for growth promotion, which have been the subject of much acrimonious debate in Codex and also the subject of an SPS Dispute Panel. 13.4.6 Enforcement Enforcement of EC regulations on veterinary drug residues in foodstuffs is the responsibility of the competent authorities in the Member States. There are also regulations on checks to be carried out on live animals to check the absence of growth-promoters whose use is prohibited in the EU. The authorities in each Member State are required to ensure that the relevant products comply with EC legislation. Council Directive 96/23 and Commission Decision 97/747 lay down the measures to be taken to monitor residues of veterinary drugs in foods of animal origin and specify the minimum level and frequency of sampling for such control. Each year Member States are required to submit their monitoring programmes to the Commission for approval and also to report the results of their monitoring work. The Commission’s Food and Veterinary Office, based in Ireland, also carries out inspections in Member States to ensure that EC legislation on veterinary drug residues is enforced.

13.5

Mercury and histamine in fishery products

Council Directive 91/493 lays down the health conditions for the production and placing on the market of fishery products. Chapter V of that Directive contains, among other things, maximum limits for histamine in certain fish species and instructions for checking that these limits are not exceeded. That chapter also makes provision for the establishment of limits for the presence in fish of contaminants from the aquatic environment. Commission Decision 93/351 lays down maximum limits for mercury in fishery products and methods of sampling and analysis to check compliance with these limits. This Decision was made after consulting the Standing Veterinary Committee. The mean total mercury content of the edible parts of fishery products must not exceed 0.5 mg/kg of fresh weight. However, this average limit is increased to 1 mg/kg fresh weight for the edible parts of certain

288

Food chemical safety

species listed in the annex to the Decision (a revision of this list of fish species is under way). The higher limit applies to inter alia sharks, tuna, swordfish, halibut and pike. In future, the maximum levels for mercury in fish will be regulated in a similar way to that described below for other heavy metals.

13.6

Other chemical contaminants

13.6.1 General procedure Council Regulation 315/93 lays down Community procedures for establishing maximum limits for contaminants (other than pesticide and veterinary drug residues) in food. The Scientific Committee for Food must be consulted on all questions which may have an effect on public health and this committee carries out the toxicological evaluations which underpin the limits set for contaminants. The scientific data which form the basis of the evaluations are obtained mainly from the scientific literature and from the Member States. Data on human exposure to contaminants, such as nitrates, cadmium, aflatoxins and ochratoxin A, have been collected and collated in projects in the programme on scientific co-operation between the Member States (known as SCOOP). Proposals for new limits prepared by Commission Working Parties are submitted to the Standing Committee for Foodstuffs, which consists of representatives of the Member States, but is chaired by the Commission. Decisions on new limits are usually made by the Commission according to a Regulatory Committee procedure (Procedure IIIb) – for details see section 13.3.4. The Commission publishes the limits as a Regulation in the Official Journal. Methods of sampling and analysis to check compliance with the maximum levels laid down are also published.

13.6.2 Mycotoxins Aflatoxins, ochratoxin A, patulin, nivalenol, deoxynivalenol, fumonisins and zearalenone have been evaluated by the Scientific Committee for Food. The question of maximum levels for some of these mycotoxins in foodstuffs has been discussed for several years in the Committee of Experts – Working Party on Agricultural Contaminants under DG VI (now under DG SANCO). Proposals from this committee are then considered by the Standing Committee on Foodstuffs, prior to adoption by the Commission as Commission Regulations. Maximum levels for aflatoxin MI in milk and for aflatoxin B1 and the sum of aflatoxins B1, B2, G1 and G2 in groundnuts and certain other foods were laid down in Commission Regulation 1525/98 (which amended Regulation 194/97) and came into force on 1 January 1999. The Commission is expected to adopt maximum levels for aflatoxins in spices in the near future. The question of maximum levels for ochratoxin A and patulin in certain foods has been under discussion for some time and a decision is expected soon.

The regulation of chemical contaminants in foodstuffs in the EU

289

A recommendation has recently been made regarding the maximum level of deoxynivalenol in cereal products. 13.6.3 Heavy metals other than mercury Discussions on limits for lead and cadmium in a wide range of foodstuffs have been going on for several years in a Working Party under DG III ( now under DG SANCO). The Scientific Committee on Food has carried out toxicological evaluations on these metals. As yet, limits for these metals have been adopted by the Commission, but a decision is expected soon. 13.6.4 Nitrate in lettuce and spinach Proposals for limits for nitrate in certain vegetables were prepared by a Committee of Experts in the Working Party on Agricultural Contaminants under the former Directorate-General VI (now under DG SANCO). The proposals were then considered under the above-mentioned procedure and the Commission has issued Regulation 194/97 setting maximum levels for nitrates in lettuce and spinach. 13.6.5 Interaction with Codex Within the Codex system, the contaminants considered in this section are mainly dealt with by the Codex Committee on Food Additives and Contaminants (CCFAC), which is hosted by the Netherlands. Many of the Member States of the European Union are very active in CCFAC. For example, Denmark and the Netherlands have been instrumental in developing the Codex General Standard on Contaminants and Toxins and draft limits for lead in various foods. Sweden has developed a proposal for a limit for ochratoxin A in cereals and cereal products and France has proposed a maximum level for patulin in apple juice.

13.7

Future trends

The EC has established procedures for preparing, adopting and enforcing legislation on limits for various chemical contaminants in foodstuffs. In recent years, the work of the scientific advisory committees, which provide the scientific basis for most of the limits, has become more independent and transparent and this trend is likely to continue. Much work still remains to be done on limits for pesticide residues. In addition to new substances, there is an urgent need to re-evaluate many of the older pesticides in the light of new toxicological data. Council Directive 91/414 concerning the placing of plant protection products on the market provides for the Commission to assess the safety aspects of pesticides. The programme of work for setting MRLs for

290

Food chemical safety

pesticide residues is gradually being aligned with that on the evaluation of pesticides according to Directive 91/414. A timetable for the work planned for the next few years has been agreed. Continued co-operation with countries outside the EU should expedite matters. The setting of MRLs for residues of veterinary drugs has been simplified somewhat since 1999, when the evaluation of ‘old substances’ was completed and all MRLs are now developed and adopted by a unified central procedure. Much work has still to be done on the preparation and adoption of maximum levels for mycotoxins, heavy metals and other contaminants, such as PCBs and dioxins. Here one of the main factors delaying progress is the lack of data for toxicological evaluations and setting tolerable daily or weekly intakes. Furthermore, there is a lack of reliable data on levels of contaminants in individual foodstuffs and on dietary intakes of such substances. Many EU Member States already play an active role in the development of Codex limits for contaminants. It is foreseen that this will continue and that coordination between the Member States and the European Commission on Codex matters will further improve. Limits for many substances mentioned above, e.g. ochratoxin A, lead, cadmium and some pesticides, are being discussed in parallel in the EC and in Codex and often by the same people. This facilitates co-ordination of the work in these different fora and should hopefully expedite the establishment of limits which can be widely accepted.

13.8

References

All the references to legislation published in the Official Journal of the European Communities are listed here in chronological order. Council Directive 76/895/EEC of 23 November 1976 relating to the fixing of maximum levels for pesticide residues in or on fruit and vegetables. Official Journal of the European Communities. No. L 340, pp. 26–31. Council Directive 81/36/EEC of 9 February 1981 amending Annex II to Directive 76/895/EEC relating to the fixing of maximum levels for pesticide residues in and on fruit and vegetables. Official Journal of the European Communities. No. L 46, pp. 33–34. Council Directive 82/528/EEC of 19 July 1982 amending Annex II to Directive 76/895/EEC relating to the fixing of maximum levels for pesticide residues in and on fruit and vegetables. Official Journal of the European Communities. No. L 234, pp. 1–4. Council Directive 86/362/EEC of 24 July 1986 on the fixing of maximum levels for pesticide residues in and on cereals. Official Journal of the European Communities. No. L 221, pp. 37–42. Council Directive 86/363/EEC of 24 July 1986 on the fixing of maximum levels for pesticide residues in and on foods of animal origin. Official Journal of the European Communities. No. L 221, pp. 43–47.

The regulation of chemical contaminants in foodstuffs in the EU

291

Council Directive 88/298/EEC of 16 May 1988 amending Annex II to Directives 76/895/EEC and 86/362 relating to the fixing of maximum levels for pesticide residues in and on fruit and vegetables and cereals, respectively. Official Journal of the European Communities. No. L 126, pp. 53–54. Council Regulation 90/2377 EEC of 26 June 1990 laying down a Community procedure for the establishment of maximum residue limits of veterinary medicinal products in foodstuffs of animal origin. Official Journal of the European Communities. No. L 224, pp. 7–14. Council Directive 90/642/EEC of 27 November 1990 on the fixing of maximum levels for pesticide residues in and on certain products of plant origin, including fruit and vegetables. Official Journal of the European Communities. No. L 350, pp. 71–79. Council Directive 91/493/EEC of 22 July 1991 laying down the health conditions for the production and placing on the market of fishery products. Official Journal of the European Communities No. L 268, pp. 15–34. Council Directive 91/414/EEC of 15 July 1991 concerning the placing of plant protection products on the market. Official Journal of the European Communities. No. L 230, pp. 1–32. Rules governing medicinal products in the European Community. Volume VI. Establishment by the European Community of maximum residue limits (MRLs) for residues of veterinary medicinal products in foodstuffs of animal origin. Commission of the European Communities. DirectorateGeneral for Internal Market and Industrial Affairs. October 1991. Council Regulation 93/315/EEC of 8 February 1993 laying down Community procedures for contaminants in food. Official Journal of the European Communities. No. L 37, pp. 1–3. Commission Decision 93/351/EEC of 19 May 1993 determining analysis methods, sampling plans and maximum limits for mercury in fishery products. Official Journal of the European Communities. No. L 144, pp. 23–24. Council Directive 93/57/EEC of 29 June 1993 amending the Annexes to Directives 86/362/EEC and 86/363/EEC on the fixing of maximum levels for pesticide residues in and on cereals and foodstuffs of animal origin, respectively. Official Journal of the European Communities. No. L 211, pp. 1–5. Council Directive 93/58/EEC of 29 June 1993 amending Annex II to Directive 76/895/EEC relating to the fixing of maximum levels for pesticide residues in or on fruit and vegetables and the Annex to Directive 90/462/ EEC relating to the fixing of maximum levels for pesticide residues in and on certain products of plant origin, including fruit and vegetables, and providing for the establishment of a first list of maximum levels. Official Journal of the European Communities. No. L 211, pp. 6–39. Council Directive 94/29/EC of 23 June 1994 amending the Annexes to Directives 86/362/EEC and 86/363/EEC on the fixing of maximum levels

292

Food chemical safety

for pesticide residues in and on cereals and foodstuffs of animal origin, respectively. Official Journal of the European Communities. No. L 189, pp. 67–69. Council Directive 94/30/EC of 23 June 1994 amending Annex II to Directive 90/ 642/EEC relating to the fixing of maximum levels for pesticide residues in and on certain products of plant origin, including fruit and vegetables, and providing for the establishment of a list of maximum levels. Official Journal of the European Communities. No. L 189, pp. 70–83. Council Directive 95/38/EC of 17 July 1995 amending Annexes I and II to Directive 90/642/EEC on the fixing of maximum levels for pesticide residues in and on certain products of plant origin, including fruit and vegetables, and providing for the establishment of a list of maximum levels. Official Journal of the European Communities. No. L 197, pp. 14–28. Council Directive 95/39/EC of 17 July 1995 amending the Annexes to Directives 86/362/EEC and 86/363/EEC on the fixing of maximum levels for pesticide residues in and on cereals and foodstuffs of animal origin. Official Journal of the European Communities. No. L 197, pp. 29–31. Council Directive 95/61/EC of 29 November 1995 amending Annex II to Directive 90/642/EEC on the fixing of maximum levels for pesticide residues in and on certain products of plant origin, including fruit and vegetables. Official Journal of the European Communities. No. L 292, pp. 27–30. Council Directive 96/23/EC of 29 April 1996 on measures to monitor certain substances and residues thereof in live animals and animal products and repealing Directives 85/358/EEC and 86/469/EEC and Decisions 89/187/ EEC and 91/664/EEC. Official Journal of the European Communities. No. L 125, pp. 10–32. Council Directive 96/32/EC of 21 May 1996 amending the Annex II to Directive 76/895/EEC relating to fixing of maximum limits for pesticide residues in and on fruit and vegetables and Annex II to Directive 90/642/EEC relating to the fixing of maximum levels for pesticide residues in and on certain products of plant origin, including fruit and vegetables, and providing for the establishment of a list of maximum levels. Official Journal of the European Communities. No. L 144, pp. 12–34. Council Directive 96/33/EC of 21 May 1996 amending Annexes to Directives 86/362/EEC and 86/363/EEC on the fixing of maximum levels for pesticide residues in and on cereals and foodstuffs of animal origin. Official Journal of the European Communities. No. L 144, pp. 35–38. Commission Regulation 194/97/EC of 31 January 1997 setting maximum levels for certain contaminants in foodstuffs. Official Journal of the European Communities. No. L 31, pp. 48–50. Council Directive 97/41/EC of 25 June 1997 amending Directives 76/895/EEC, 86/362/EEC, 86/363 EEC and 90/462/EEC relating to the fixing of maximum levels for pesticide residues in and on, respectively, fruit and vegetables, cereals, foodstuffs of animal origin, and certain products of

The regulation of chemical contaminants in foodstuffs in the EU

293

plant origin, including fruit and vegetables. Official Journal of the European Communities. No. L 184, pp. 33–49. Commission Decision 97/579/EC of 23 July 1997 setting up Scientific Committees in the field of consumer health and food safety. Official Journal of the European Communities. No. L 237, pp. 18–23. Commission Decision 97/747/EC of 27 October 1997 fixing the levels and frequencies of sampling provided for by Council Directive 96/23/EC for the monitoring of certain substances and residues thereof in certain animal products. Official Journal of the European Communities. No. L 303, pp. 12–15. Commission Regulation 1525/98/EC of 16 July 1998 amending Regulation (EC) No.194/97 of 31 January 1997 setting maximum levels for certain contaminants in foodstuffs. Official Journal of the European Communities. No. L 201, pp. 43–46. Commission Directive 98/82/EC of 27 October 1998 amending the Annexes to Council Directives 86/362/EEC, 86/363/EEC and 90/642/EEC on the fixing of maximum levels for pesticide residues in and on cereals, foodstuffs of animal origin and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 290, pp. 25–54. Commission Directive 1999/71/EC of 14 July 1999 amending the Annexes to Council Directives 86/362/EEC, 86/363/EEC and 90/642/EEC on the fixing of maximum levels for pesticide residues in and on cereals, foodstuffs of animal origin and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 194, pp. 36–44. Commission Directive 2000/24/EC of 28 April 2000 amending the Annexes to Council Directives 76/895/EEC, 86/362/EEC, 86/363/EEC and 90/642/ EEC on the fixing of maximum levels for pesticide residues in and on cereals, foodstuffs of animal origin and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 107, pp. 28–37. Commission Directive 2000/42/EC of 22 June 2000 amending the Annexes to Council Directives 86/362/EEC, 86/363/EEC and 90/642/EEC on the fixing of maximum levels for pesticide residues in and on cereals, foodstuffs of animal origin and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 158, pp. 51–75. Commission Directive 2000/48/EC of 25 July 2000 amending the Annexes to Council Directives 86/362/EEC and 90/642/EEC on the fixing of maximum levels for pesticide residues in and on cereals and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 197, pp. 26–31. Commission Directive 2000/57/EC of 22 September 2000 amending the Annexes to Council Directives 76/895/EEC and 90/642/EEC on the

294

Food chemical safety

fixing of maximum levels for pesticide residues in and on fruit and vegetables and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 244, pp. 76–77. Commission Directive 2000/58/EC of 22 September 2000 amending the Annexes to Council Directives 86/362/EEC, 86/363/EEC and 90/642/ EEC on the fixing of maximum levels for pesticide residues in and on cereals, foodstuffs of animal origin and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 244, pp. 78–83. Commission Directive 2000/81/EC of 18 December 2000 amending the Annexes to Council Directives 86/362/EEC, 86/363/EEC and 90/642/ EEC on the fixing of maximum levels for pesticide residues in and on cereals, foodstuffs of animal origin and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 236, pp. 56–62. Commission Directive 2000/82/EC of 20 December 2000 amending the Annexes to Council Directives 76/895/EEC, 86/362/EEC, 86/363/EEC and 90/642/EEC on the fixing of maximum levels for pesticide residues in and on cereals, foodstuffs of animal origin and certain products of plant origin, including fruit and vegetables, respectively. Official Journal of the European Communities. No. L 3, pp. 18–26. Commission Recommendation 2001/42/EC of 22 December 2000 concerning a co-ordinated Community monitoring programme for 2001 to ensure compliance with maximum levels of pesticide residues in and on cereals and certain products of plant origin, including fruit and vegetables. Official Journal of the European Communities. No. L 11, pp. 40–45.

14 Contaminant regulation and management in the United States: the case of pesticides C. K. Winter, University of California, Davis

14.1

Introduction

The agricultural use of pesticides such as insecticides, herbicides, and fungicides has clearly reduced crop losses due to insects, weeds, and plant diseases in the US and throughout the world. The benefits from agricultural pesticides include improved crop yields, greater availability of fruits, vegetables, and grains, and lower consumer costs (Ecobichon, 1996). Since pesticide chemicals rely upon their toxicity to provide control of pests, they also present unique potential risks to humans and to the environment. Occupational exposure to workers involved in the mixing, loading, or application of pesticides and to workers entering fields treated with pesticides is of significant concern in the US as hundreds of cases of worker illnesses and injuries resulting from occupational pesticide exposure are reported annually (California Department of Pesticide Regulation, 1999). In addition, epidemiology studies have linked increases in the incidence of certain types of cancers to occupational exposure to some pesticides (Hoar et al., 1986, American Medical Association, 1988). Pesticides also pose risks through their potential to destroy natural vegetation, reduce natural pest populations, poison non-target organisms such as fish, wildlife, and livestock, reduce honeybee populations, or create secondary pest problems (Pimentel et al., 1992). The environmental and occupational impacts from pesticide use may well represent the major types of risks posed by pesticides. In terms of public opinion in the US, however, the pesticide risks most commonly identified are those from consumer exposure to pesticide residues in foods. In a national consumer attitude survey performed annually in which respondents were asked to indicate the perceived magnitude of risk from pesticide residues in food, 72 to 82% of

296

Food chemical safety

US consumers considered pesticide residues to pose a serious health hazard (Bruhn et al., 1998). This level of concern contrasts greatly with the food safety priorities of the US Food and Drug Administration (FDA) which considers pesticide residues as its fifth food safety priority and of less concern than • • • •

microbiological contamination nutritional imbalance environmental contaminants naturally occurring toxins (Winter, 1996).

The National Research Council (NRC) of the US National Academy of Sciences recently concluded that the cancer risks posed by naturally occurring chemicals in the food supply were considerably greater than the risks from synthetic chemicals such as pesticides (NRC, 1996). Consumer awareness concerning pesticide residues in foods in the US has risen greatly in the past two decades and much of this awareness may be traced to several events that have captured significant media, legislative, and regulatory attention. In 1985, a California watermelon grower illegally used the insecticide aldicarb and more than one thousand cases of probable or possible pesticide poisoning were identified (Goldman et al., 1990). In 1987, a report of the NRC was issued which estimated alarmingly high cancer risks from pesticides in foods (NRC, 1987). While the body of the NRC report indicated that the methodology relied upon worst-case assumptions that dramatically overestimated exposure levels (subsequent research indicated risks were exaggerated by factors of 10,000 to 100,000 times), this report nevertheless captured media headlines, consumer interest, and legislative concern (Archibald and Winter, 1989). Two years later, an environmental advocacy organization issued a report alleging ‘intolerable’ risks to children from exposure to residues of cancer-causing or neurotoxic pesticides in foods which received widespread media attention (Natural Resources Defense Council, 1989). The NRC issued another report in 1993 recommending that the US Environmental Protection Agency (EPA) improve its regulatory and risk assessment practices to take into account the potential greater risks of infants and children to pesticides (NRC, 1993). Several of the NRC recommendations became law following the passage of the Food Quality Protection Act (FQPA) of 1996. The implementation of FQPA by the EPA has already resulted in restrictions in pesticide use and future decisions are likely to involve more restrictions. Additionally, some controversial human epidemiological studies have linked exposure to some insecticides to increases in cancer incidence while others have suggested that pesticides used in agriculture may have the ability to cause endocrine system disruption in the general population (Wolff et al., 1996, Colburn et al., 1996). While it is clear that much public and legislative concern exists regarding pesticide residues in foods, it is also clear that the practices for regulating residues and estimating levels of potential risks are complicated and often confusing. This chapter addresses the regulation and management of pesticide residues in the US. It focuses upon the regulations that pertain to pesticide

Contaminant regulation and management in the United States

297

management and on the various US agencies responsible for pesticide regulation. The scientific basis for regulating and managing pesticide residues is discussed and methods to improve the regulatory and management processes are suggested.

14.2

Pesticide regulation in the US

Three US federal agencies – the EPA, the FDA, and the US Department of Agriculture (USDA) – all play roles in the regulation of pesticides. The EPA, under provisions of the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) of 1947 and several subsequent FIFRA amendments, makes the determination as to whether the benefits of the use of a specific pesticide outweigh its competing risks. Benefits might include considerations such as increased agricultural productivity, lower food costs, or public health protection, while risks might include consumer exposure, occupational exposure, or environmental disruption. The EPA reviews toxicological and environmental fate data submitted by the pesticide manufacturer prior to granting a pesticide registration. The time for completion of required studies and EPA review frequently exceeds 10 years at a cost of approximately $30 million (Ecobichon, 1996). In cases where EPA has deemed the benefits to outweigh the risks, the EPA determines the conditions for appropriate pesticide use practices and may establish tolerances representing the maximum allowable residue level on food and feed crops. Individual states in the US also have the authority to register pesticides for use and to specify appropriate use conditions that may be more (but never less) stringent than those established by EPA. Of the 50 US states, California has the most comprehensive program for pesticide regulation. The FDA has the responsibility to enforce tolerances in domestic and imported foods (FDA, 2000). Samples of domestic foods are often taken near the source of production or at the wholesale level. Imported food samples are typically taken at the point of entry into the US. FDA monitoring efforts primarily focus upon the raw agricultural commodities that are analyzed as the unwashed, whole (unpeeled) raw product. The types and numbers of samples taken in the FDA monitoring programs are determined by several factors including regional intelligence on pesticide use, the dietary importance of the food, information on the amount of domestic and imported food that enters interstate commerce, chemical characteristics and toxicity of the pesticide, and pesticide usage patterns (FDA, 2000). Once samples are collected, they are analyzed using multiresidue methods capable of determining approximately one-half of the approximately 400 pesticides for which EPA has established tolerances as well as other pesticides for which no tolerances exist. Single residue methods are also frequently used to supplement results obtained from the multiresidue methods although they are much less cost effective than the multiresidue methods.

298

Food chemical safety

The FDA has also performed its Total Diet Study annually since 1961. This study involves a ‘market basket’ collection of foods from four chosen geographical regions and three cities in each region per year. Each market basket is comprised of 261 different food samples and the foods are prepared for table-ready consumption prior to analysis. Estimates of daily exposures of population subgroups to pesticide residues in the diet are attained by considering the results of the Total Diet Study samples and food consumption estimates obtained from national food consumption surveys (FDA, 2000). While the FDA is the primary federal agency responsible for monitoring food for pesticide residues, the USDA also has a role as it monitors meat, poultry, and egg products (Garcia and Winter, 2000). In addition, the USDA operates the Pesticide Data Program (PDP) which takes additional samples of fruits, vegetables, and processed foods. Results from the PDP are more representative of the residue profiles of pesticides in the food supply than those obtained from FDA’s regulatory monitoring programs and are frequently used by the EPA to aid in the risk assessment process (USDA, 2000).

14.3

Regulatory monitoring of pesticides in the US

In 1999, the FDA analyzed 9,438 food samples for pesticide residues in its regulatory monitoring program. The majority of samples involved imported foods (6,012 samples or 63.7 percent) rather than domestic foods (3,426 samples or 36.3 percent) (FDA, 2000). The results of FDA’s 1999 monitoring of imported foods are shown in Fig. 14.1. Overall, 65.0 percent of the samples had no residues detected while 3.1 percent had violative residues. Residues and violations were more commonly detected in samples of fruits and vegetables than from grains and grain products, milk/dairy products/eggs, or fish/shellfish. Figure 14.2 shows the results of FDA’s 1999 monitoring of domestic foods. No residues were detected in 60.2 percent of the samples and the violation rate for domestic foods was 0.8 percent. As was the case with the imported samples, fruits and vegetables were responsible for the highest percentages of residue detections and the greatest number of violations. Two distinct types of violations are commonly seen for pesticide residues in foods. One type involves detection of a residue of a pesticide at a level that exceeds the established tolerance for that commodity. The other type involves the case where a residue of a pesticide, at any quantifiable level, is detected on a commodity for which a tolerance has not been established. This type of violative residue may result from application of the pesticide to the wrong commodity, drift of a pesticide from an adjacent field, or uptake from soil contaminated from a prior use of the pesticide on a different commodity. For domestic samples in 1999, 8 of 26 (31 percent) violations involved residues detected in excess of tolerances while the rest (18 of 26, 69 percent) involved residues detected on commodities for which no tolerances were established. Violative imported

Contaminant regulation and management in the United States

Fig. 14.1

299

FDA 1999 monitoring of imported foods.

samples included 19 of 188 (10 percent) where tolerances were exceeded and 169 of 188 (90 percent) where residues were detected on commodities for which tolerances were not established (FDA, 2000). Monitoring of pesticide residues in the US also occurs at the state level. California is the largest state in the US in terms of both agricultural production and population and also possesses, by far, the most comprehensive state regulatory program for pesticides. In 1997, California’s Marketplace Surveillance Program analyzed 5660 samples (62.2 percent from California, 6.7 percent from other states, 31.1 percent from other countries) for pesticide residues. The results from this program are shown in Fig. 14.3 and demonstrate that the majority of samples (62.1 percent) contained no detectable residues

300

Food chemical safety

Fig. 14.2

FDA 1999 monitoring of domestic foods.

while 36.7 percent of the samples contained legal residues and 1.2 percent contained violative residues (California Department of Pesticide Regulation, 2000). As was also seen from the FDA monitoring, most of the violations (82 percent) involved detection of residues on commodities for which no tolerances were established while 18 percent of the violations involved detection of residues exceeding the tolerance levels. A comparison of domestic and imported samples from California’s 1997 program indicated that a higher percentage (67.0 vs. 51.3) of domestic foods contained no detectable residues than imported foods and that violation rates from imported foods were more than three times greater (2.4 percent vs. 0.7 percent) than those of domestic foods (California Department of Pesticide Regulation, 2000).

Contaminant regulation and management in the United States

301

Fig. 14.3 California 1997 pesticide residue monitoring.

USDA’s PDP has been in existence since 1991. In contrast to the FDA’s monitoring efforts which focus primarily upon enforcing tolerances in the raw commodities, the PDP was designed to determine more accurately and representatively the incidence of pesticide residues in foods and the levels of such foods following normal procedures such as washing or processing (USDA, 2000). Laboratories participating in PDP were located in ten different states and analytical procedures were refined to allow more sensitive detection than methods commonly used in regulatory monitoring programs that enforce tolerances. The selection of commodities chosen for sampling in PDP varies annually and is based upon both EPA risk assessment data needs and USDA’s food consumption surveys. In 1998, PDP collected samples of apple juice, cantaloupe, grape juice, green beans, orange juice, pears, spinach, strawberries, sweet potatoes, tomatoes, winter squash, corn syrup, milk, and soybeans. A total of 8500 samples were collected. Most of these were from fruits and vegetables (7017) with lower numbers of samples collected for whole milk (595), soybeans (590) and corn syrup (298). The majority of samples (84 percent) was of domestic origin. Overall, 45 percent of the samples contained no detectable residue while 26 percent contained one residue and 29 percent contained more than one residue. Residues exceeding the tolerance level were detected on 0.15 percent of the samples. In another 3.7 percent of the samples, residues of pesticides were detected on commodities for which no tolerances of the pesticides were established (USDA, 2000).

14.4

Managing pesticides in foods in the US

Federal and state regulatory monitoring programs that enforce pesticide tolerances provide a direct mechanism to manage pesticides in foods. In cases where violative residues are detected, the foods on which the residues are

302

Food chemical safety

present are subject to seizure or injunction; such actions would not allow the foods to be distributed in commerce. While these regulatory monitoring programs do provide economic disincentives to food producers that may not use pesticides correctly they sample only a small percentage of the total food supply and therefore are incapable of serving as tools for the comprehensive management of pesticides in foods. The primary tool for managing pesticide residues in foods is the establishment of tolerances by the EPA. The practices used by EPA to establish tolerances are often confusing. Such practices are described in detail by Winter (1992a) and are summarized in the following paragraphs. Tolerances are normally required in cases where the legal use of a pesticide might result in residues on a food or feed crop. Tolerances are specific to commodity/pesticide combinations; the same pesticide may have different tolerance levels established for different commodities while the same commodity may have several different tolerance levels established for different pesticides. Tolerances are primarily established to represent the maximum residues anticipated from the legal use of the pesticide on the commodity. The maximum residue levels are determined from the results of controlled field studies performed by the pesticide manufacturer in a variety of geographical regions. The manufacturer performs the studies under conditions that would likely yield the maximum residue levels such as applying the pesticide at the maximum recommended rate for the maximum number of applications anticipated and harvesting the commodity at the minimum expected preharvest interval. The highest residue levels detected under these ‘worst-case’ application scenarios are identified and the manufacturer petitions the EPA to establish the tolerance levels at or slightly above the maximum residues encountered. It is correct to note that the specific tolerance levels requested by the manufacturer are determined solely on the basis of agricultural practices and not upon potential human health considerations. As such, tolerances represent enforcement tools to determine whether pesticide applications were made in accordance with the law but should not be considered as ‘safety standards.’ In the case where a pesticide is used properly, the resulting residue level should be below the tolerance level. Residues detected in excess of the established tolerance are likely encountered only in cases where applications are not made in accordance with the legal directions. Results obtained from federal and state monitoring programs demonstrate that the incidence of residues detected in excess of tolerances is very low and suggest that most pesticide applications are made legally. Although confusing, it is also correct to note that potential human health risks are considered before a tolerance is established. The EPA will perform a risk assessment of the potential dietary risk to consumers from exposure to the pesticide from all registered (and proposed) uses of the pesticide. If such a risk is determined to be excessive, the EPA will deny the manufacturer’s petition to establish a tolerance. If the level of risk is not considered to be of concern, the

Contaminant regulation and management in the United States

303

EPA will establish the tolerance at or slightly above the maximum residue level detected from the manufacturer’s controlled field trials. The processes EPA uses to assess the risks from pesticides are complicated and constantly changing to meet the needs of new regulations and evolving toxicological and computational methods. The provisions of FQPA, passed in 1996, are likely to have significant impacts upon the establishment of tolerances for new pesticides as well as the reevaluation of existing tolerances. Many of the scientific issues raised by FQPA are discussed in section 14.5. At the present time, most US pesticide tolerances were established prior to the passage of FQPA. In assessing consumer risk from exposure to pesticides, the EPA first estimates consumer exposure. The maximum legal exposure to the pesticide is usually first calculated by assuming that • the pesticide is always used on all food items for which it is registered (or proposed for registration) • all residues on the commodity will be present at the established (or proposed) tolerance levels • no reduction in residue level will occur through the effects of factors such as washing, peeling, cooking, transportation, and processing.

This maximum legal exposure, often referred to as the Theoretical Maximum Residue Contribution, or TMRC, is compared with established toxicological criteria such as the reference dose (RfD) or Acceptable Daily Intake (ADI) which represent, after analysis of animal toxicology data and extrapolations to humans, the daily exposure that is not considered to present any appreciable level of risk. When it is determined that the TMRC exposure is below the RfD or ADI, the EPA usually considers the risks from the pesticide in question to be negligible and approves the manufacturer’s petition to establish a tolerance at or slightly greater than the maximum levels identified from the manufacturer’s controlled field trials (Winter, 1992a). Present risk assessment practices normally differentiate between non-cancer risks that are assumed to possess a toxicological threshold dose below which no effects are likely to occur and cancer (carcinogenic) risks for which it is frequently assumed that any level of exposure possesses at least a finite mathematical chance of leading to cancer. For potentially carcinogenic pesticides, the EPA will calculate an estimate of the pesticide’s oncogenic (tumor producing) risk using mathematical models for tumor production and substituting the TMRC value as the estimate of exposure. If the oncogenic risk is below the ‘negligible risk’ level of one excess case of cancer per million persons exposed, the EPA will generally approve a tolerance. It should be noted that the calculation of oncogenic risk involves conservative assumptions of tumor development that extrapolate the results of moderate and high dose carcinogenicity studies to predict oncogenic risks at low levels of exposure; this practice may dramatically overestimate the ‘actual’ incidence of cancer development (Winter, 1992b, Winter and Francis, 1997).

304

Food chemical safety

There are many cases in which the TMRC exceeds the RfD or ADI or the oncogenic risk at which the TMRC exceeds one excess cancer per million. When these cases arise, the EPA may refine its exposure assessment practices to consider factors such as more realistic estimates of pesticide use, residue levels, and/or postharvest effects upon residue levels. Studies have indicated that TMRC values often exaggerate pesticide exposure estimates by factors of 10,000 to 100,000 times. The EPA’s refinements may yield a value known as the Anticipated Residue Contribution (ARC) which may be substituted for the TMRC to determine the potential risks from the pesticide. In cases where the ARC is below the RfD or ADI and the oncogenic risk at the ARC is below one excess cancer per million, the EPA will generally approve a tolerance (Winter, 1992a). The procedures used by the US to establish tolerances are similar to those used by the Codex Alimentarius Commission to determine their analogous Maximum Residue Levels (MRLs). One comparison of US tolerances and Codex MRLs demonstrated that the two sets of standards were equivalent 47 percent of the time while US tolerances were lower 19 percent of the time and Codex MRLs were lower (and therefore more stringent) 34 percent of the time. Some of these differences were explained to result from different agricultural production and pest control practices, the use of different data sets, and differences in how the breakdown products of some pesticides are regulated (General Accounting Office, 1991). It should be recognized that US pesticide tolerances established by the aforementioned process, in combination with regulatory monitoring programs, serve important roles as enforcement tools that provide economic disincentives for pesticide users to misuse pesticides. At the same time, the tolerances should not be considered as safety standards since violative residues rarely represent residues of toxicological concern (Winter, 1992a). As was mentioned previously, the management of pesticides involves regulating environmental and occupational risks from pesticides in addition to potential dietary risks faced by consumers. Regulatory actions taken to eliminate or reduce environmental or occupational risks may frequently impact consumer risks since the potential for consumer exposure to the pesticides may be altered. As an example, the EPA may establish worker reentry intervals that represent the minimum legal amount of time following a pesticide application that must pass before workers may enter a treated field. If the EPA were to lengthen the worker re-entry interval for a particular pesticide on a specific commodity, it is possible that pesticide applications would need to be made earlier in the growing season to allow workers to harvest the commodity while obeying the new re-entry interval. Under this scenario, the pesticide will have more time for its residues to dissipate and consumer exposure to the pesticide could be reduced.

Contaminant regulation and management in the United States

14.5

305

Improving the management of pesticides in foods

The procedures used by the EPA to estimate the potential risks associated with pesticide residues in the food supply and to determine subsequently the appropriateness of pesticide tolerances were reviewed in a comprehensive report of the US National Research Council (NRC, 1993). It was concluded in this report that the EPA needed to improve its risk assessment methods to account specifically for the potential increased risks faced by infants and children exposed to pesticides. Several of the recommendations of the NRC report were incorporated into FQPA when it was signed into law in August 1996. An important provision of FQPA is the ‘10 factor’ which requires the EPA to consider if an additional ten-fold uncertainty factor in the determination of acceptable levels of exposure is needed to provide greater protection for infants and children. An ‘aggregate exposure’ provision was included which directed the EPA to determine consumer pesticide exposure from residues in the water supply and/or the residential setting in addition to dietary sources. In cases where specific pesticides were members of pesticide ‘families’ which all possessed a common mechanism of toxicological action, the EPA was directed to consider the ‘cumulative exposure’ to all members of the pesticide family rather than consider the risks posed by the individual pesticides of the family. Additionally, FQPA provided for development of a ‘consumer right to know’ brochure, established a timetable for which all existing pesticide tolerances needed to be reassessed, and required the EPA to develop guidelines for evaluating the potential risks of pesticides that could affect endocrine system function. Ironically, the NRC report did not conclude that the risks of infants and children exposed to pesticides were excessive but simply recommended improvements to the EPA’s risk assessment methods. The impetus for FQPA’s adoption was court rulings that set up timetables for the elimination of several existing pesticide uses because they fell under the provisions of the 1958 Delaney Amendment of the US Federal Food, Drug, and Cosmetic Act. The Delaney Amendment did not allow the use of food additives shown to ‘induce cancer in man or in animals’ in the US food supply regardless of the amounts and associated risks posed by the chemicals. While most pesticide residues were not legally considered to be food additives and were thus not subject to the Delaney Amendment, pesticides that were either added directly to processed foods or whose residues concentrated during processing were regulated as food additives (Winter, 1993). FQPA provided a legislative solution to the anachronistic Delaney Amendment by excluding all residues of pesticides as food additives. Contrary to many beliefs, the Delaney Amendment still applies to food additives in the US. Improvements in pesticide residue risk assessment practices should improve the scientific basis for managing pesticide residues in foods and the FQPA provides a blueprint for making such improvements. While most of the FQPA provisions are considered in theory to represent improvements in the risk assessment process, the practical adoption of methods to comply with such

306

Food chemical safety

provisions presents scientific challenges. Methodologies to perform aggregate and cumulative assessments of risk, for example, are still in their early stages of development. The EPA, however, is also confronted with statutory deadlines and political pressures that may compromise its ability to use the best scientific methods, many of which are still being developed, to assess pesticide risks and make regulatory decisions. One of the major improvements in dietary pesticide risk assessment is the use of probabilistic methods to assess exposure. The NRC report strongly recommended probabilistic techniques to predict acute (single day) exposure to pesticides and improvements in our computational capabilities are making such techniques easier to use. Briefly, previous methods to assess pesticide exposure, such as those already described for determining TMRCs and ARCs, relied upon deterministic methods that assign finite values for both food consumption and pesticide residue levels. Using a deterministic approach, an exposure estimate is attained by multiplying the food consumption and pesticide residue levels together. In reality, neither food consumption nor pesticide residue levels exist as single points but are best considered to represent distributions. Food consumption databases, for example, often indicate that a majority of the population may not have eaten any of a particular food item on a given day. For those that did consume the item, several frequently consume small to moderate amounts and only a few consume very large amounts. An analogous situation exists for pesticide residues; most foods contain no detectable residues. When residues are detected, they are frequently present at very low levels and it is rare that levels approach tolerance levels. In California’s 1997 Priority Pesticide Program, for example, 81.1 percent of 1,695 food samples analyzed contained no detectable residues. Another 10.4 percent of the samples contained residues at levels less than 10 percent of the tolerance level, 8.1 percent had residues between 10 and 50 percent of the tolerance level, and only 0.35 percent had residues between 50 and 100 percent of the tolerance (California Department of Pesticide Regulation, 2000). Probabilistic models are capable of using all of the data from the food consumption and the pesticide residue databases rather than focusing upon single point estimates. Through the use of computational techniques frequently known as Monte Carlo models, single food consumption and residue values are continuously and randomly selected from the full databases and multiplied to yield exposure estimates (Petersen, 2000). This process is repeated for a determined number of events (frequently in the thousands or tens of thousands) and the corresponding exposure estimates are combined to yield a distribution of daily dietary exposure to the pesticide. Such exposure distributions provide much more information than that resulting from a deterministic approach and allow for determination of estimates for the mean daily exposure as well as the upper percentiles such as the 95th, 99th, and 99.9th percentiles. The additional data provided from the probabilistic exposure assessment also require greater interpretation. Under provisions of FQPA, the EPA is not allowed to grant tolerances for pesticides unless the aggregate and cumulative

Contaminant regulation and management in the United States

307

exposures provide a ‘reasonable certainty of no harm.’ In the past, using deterministic approaches, such a determination was not hard to make since the point exposure estimate could simply be compared with the appropriate toxicological criterion such as the RfD or ADI to determine if such a criterion was exceeded. The definition of a ‘reasonable certainty of no harm’ becomes more difficult using probabilistic methods as a determination is required to be made for the appropriate level of protection. For example, exposures at the 99.9th percentile may be far greater than those of the 99th percentile or the 95th percentile so the choice of the desired level of protection could dramatically influence the acceptability of the exposure. Under current EPA guidelines, the ‘reasonable certainty of no harm’ determination applies if exposure to a pesticide at the 99.9th percentile for a population subgroup is less than the RfD (EPA, 2000). Such an approach is criticized by Chaisson et al. (1999) who contend that significant bias, error, and uncertainty exist in the upper percentiles which may dramatically overestimate exposure. Sources of error and bias include the inaccuracy of dietary food consumption surveys, insufficient sample sizes, improper weighting of sample data, and reliance upon subpopulation data to characterize the entire population. They argue that regulatory decisions should use a ‘point of regulation’ approach which represents the highest percentile of exposure that is not dominated by overestimation bias and error. The quality and availability of residue data may also influence the accuracy of exposure estimates using probabilistic methods. Pesticide residues are often taken as composite samples containing several individual servings of the commodities rather than as single servings. Research has shown that single serving size subsamples of composite samples are quite variable and the possibility exists for an individual consuming a single serving of a commodity to experience an exposure far different to that of the composite sample or from another single serving (Andersson, 2000; Harris, 2000). Composite samples also may contain residues of a greater number of pesticides than single serving samples. In the USDA’s 1998 PDP summary, as many as eight different pesticides were detected from composite pear samples and residues of two, three, or four different pesticides were detected on 36.4, 18.1, and 10.1 percent of the samples, respectively. This contrasts to results of single serving size pear samples where the maximum number of pesticides detected was three (0.9 percent) and residues of two different pesticides were detected only 19.5 percent of the time (USDA, 2000). Some of the different pesticides detected might be members of the same chemical family and may effectively serve as alternatives or substitutes for each other; as such the probability of their co-occurrence in a single serving sample would be very slight although composite samples might indicate a much greater probability of cooccurrence. In cases where the cumulative exposure to the entire family of chemicals is determined, reliance upon composite samples rather than single serving samples could dramatically overestimate exposure at the upper percentiles.

308

Food chemical safety

The ability to use probabilistic approaches to assess dietary pesticide exposure has also changed much of the emphasis of pesticide risk assessment practices from assessing long-term (chronic) exposure to short-term (acute) exposure. Deterministic approaches worked well with chronic assessments since the day-to-day variability in food consumption patterns and the variability of pesticide residue levels tended to average out over the course of a 70-year exposure period. Deterministic approaches have also often been used in the assessment of acute dietary risk by assuming an upper percentile level of food consumption and the maximum detected or allowable level of residue. The point estimate determined in this manner is then compared with the RfD to determine the acceptability of exposure under the specified conditions. It is now possible to apply acute probabilistic and deterministic methods of exposure assessment for the same chemical using the same food consumption and residue databases and derive conflicting results concerning the acceptability of exposure. A deterministic method, for example, could conclude that a residue present at the maximum detected level consumed by someone eating at the upper 95th percentile of consumption poses an exposure below the RfD. A probabilistic assessment using the same consumption and residue databases may indicate that exposure at the 99.9th percentile exceeds the RfD. In this case, regulatory actions may ensue even though a prior deterministic exposure estimate did not suggest an excessive risk. The relatively new focus upon acute risks from pesticides also requires improvements in the toxicological databases. Toxicology tests have commonly focused upon identifying the maximum amount of exposure (No Observed Adverse Effect Level, or NOAEL) that does not cause effects in test animals. The RfD, in turn, is derived from the NOAEL by dividing the NOAEL by uncertainty factors to account for differences in intra- and inter-species variability in toxicological response. Most NOAELs and RfDs are derived from the results of chronic toxicology studies that involve repeated daily exposures of test animals for periods of a year or more. When shorter-term studies have been conducted to establish shorter-term RfDs, they have typically involved 28- to 90-day repeated exposures rather than single-day exposures. While it is possible to similarly derive acute RfDs using single-day exposure studies in test animals, such single-day studies have traditionally been used to determine animal lethality and not threshold effects such as the NOAEL. It is most appropriate for the exposure estimates determined from acute probabilistic assessment techniques to be compared with acute single-day RfDs to determine the acceptability of such exposures. Unfortunately, since accurate acute RfDs for most pesticides have not been determined, the exposure estimates are often compared with RfDs derived from longer-term (28 to 90 days) or chronic toxicology studies. In most cases, the acute RfDs may be much higher than those obtained from longer-term studies. This is particularly important in cases where pharmacokinetic factors such as absorption, distribution, biotransformation, and excretion of a pesticide have been established and demonstrate that repeated exposure to the pesticide could cause an increase in

Contaminant regulation and management in the United States

309

the concentration of the pesticide at the biochemical site of toxicity. In effect, this improper comparison of single-day exposure with longer-term toxicological effects may lead to an exaggeration of the actual level of risk. In other cases where acute RfDs do exist, they may be derived from studies that do not include sufficient numbers and ranges of dose levels to allow an accurate determination of the acute NOAEL. A recent EPA decision concerning the dietary risks of the insecticide methyl parathion illustrates the need for accurate acute RfDs. In preliminary assessments of the risk, the EPA relied upon a single day toxicology study that showed a NOAEL of 0.025 mg/kg/day and a Lowest Observed Adverse Effect Level (LOAEL) of 7.5 mg/kg/day. The range between doses was 300-fold and suggested that the ‘true’ NOAEL would be anywhere from 0.025 mg/kg/day to 7.5 mg/kg/day; comparisons of the toxicity of methyl parathion to other members of its chemical family of organophosphate insecticides would suggest that the ‘true’ NOAEL would be much closer to 7.5 mg/kg/day than to 0.025 mg/ kg/day. In EPA’s revised risk assessment for methyl parathion, it rejected this abnormally low NOAEL and replaced it with a NOAEL of 0.11 mg/kg/day derived from a one-year feeding study in rats (EPA, 1999). After relying upon this exaggeratedly low NOAEL to determine an acceptable level of exposure, the EPA concluded that the daily dietary risks for infants and children from exposure to methyl parathion were excessive and the registrations for methyl parathion on several food crops were revoked. The EPA also acknowledged that a recently submitted single-day toxicology study of methyl parathion in rats indicated a NOAEL of 1.0 mg/kg/day. Had this more accurate NOAEL been used to determine the acute RfD, all exposures to all population subgroups would have been below the acute RfD at the upper 99.9th percentile of exposure and no regulatory action may have been taken. This example clearly indicates the need for development of appropriate tests to determine acute NOAELS and subsequent RfDs. In addition to the need for scientific improvements to allow probabilistic risk assessments to be properly performed and interpreted, there also exists a need to educate stakeholders about what the US system for tolerance establishment and monitoring does and does not do. In simplest terms, the US system can be described as a ‘food quality’ system but not necessarily a ‘food safety’ system. This results from the fact that the pesticide tolerances are not safety standards but rather exist as enforcement tools that allow an assessment of how well pesticide application regulations are adhered to. Violative residues demonstrate the likelihood of pesticide misuse but should not be considered, in the vast majority of cases, to represent ‘unsafe’ residues. Safety considerations govern whether or not the use of pesticides on specified commodities will be permitted; tolerances, when granted, serve as indicators of good agricultural practices rather than as toxicological benchmarks.

310

Food chemical safety

14.6

Future trends

Clearly, future US risk assessment and regulatory trends will focus upon improving risk assessment capabilities for pesticide residues in foods. This manuscript has provided a general overview concerning many of the challenges faced by risk assessors and risk managers but is by no means comprehensive. The provisions of FQPA call for the EPA to perform both aggregate and cumulative risk assessments for pesticides. To perform these tasks adequately, much more data are needed and methodologies to interpret the data appropriately must also be developed. In the case of aggregate exposures, for example, the data available to evaluate exposure of consumers from pesticides in drinking water are minimal and variable while even less information exists concerning residential exposure to pesticides. Cumulative assessments have yet to be put into place by the EPA although FQPA became law more than four years ago; risk assessments are still considering pesticides on an individual basis rather than as members of a family possessing a common mechanism of toxicity. The management of risk resulting from cumulative risk assessments also presents challenges in cases where the risks are deemed to be excessive as decisions need to be made concerning which uses of pesticides should be eliminated to reduce the magnitude of the risk to an acceptable level. The development of processes to determine aggregate and cumulative exposures adequately to pesticides will require the use of sophisticated modeling practices to generate sufficient data to guide such assessments. An accurate knowledge of pesticide use and pest management practices is essential to modeling efforts. To date, pesticide use information has been incorporated into risk assessment practices but only in a limited manner. Deterministic approaches have frequently considered the percentage of acres of a commodity treated with a pesticide to refine exposure assessments. For example, if the TMRC value for a pesticide used on a single commodity were determined and use data indicated that the pesticide was only used on 50 percent of the acres of the commodity, the ARC would be calculated to be 50 percent of the TMRC (Winter, 1992b). In other cases, the percentage of acres treated may be used to assist in the quantitation of possible residues that may have been present on the food item but were below the limit of analytical detection. Residues below the detection limits are frequently and conservatively assigned a value of one-half of the limit of detection; the use of percentage of acres treated provides an assignment of zero for the fraction of samples assumed to have not been treated with the pesticide. Pesticide use data are also important in risk management decisions that consider the availability of pesticide alternatives and the subsequent additional risks posed by the alternatives resulting from elimination of specific pesticide uses. It is clear that improved knowledge of pesticide use practices could significantly improve risk assessments by developing models that predict food residues based upon modeling actual pesticide use practices. Data concerning the frequency of pesticide application, weather and climate conditions,

Contaminant regulation and management in the United States

311

environmental fate, the intervals between application and harvest, and the application rate could allow predictions of residues on a variety of commodities produced throughout the world that could also be tested through analysis of realworld samples. Such a process would require a comprehensive and accurate reporting of pesticide use. The state of California in the US is presently the only state that has laws requiring the reporting of all agricultural applications of pesticides although similar approaches are being considered in other states. To predict residues based upon pesticide use data more appropriately, it is necessary to have uniform reporting throughout the US as well as in other countries that frequently export their agricultural products to the US.

14.7

Further information and advice

Pesticide risk assessment and management practices in the US are dynamic and evolving. The FQPA has presented great scientific challenges that require advances in scientific methodologies as well as development in science policies. The EPA has identified several science policy issues including many discussed in this manuscript and provides up-to-date policy statements, guidelines, and public comments concerning such issues. This information is available on the internet at http://www.epa.gov/pesticides/trac/science/. Specific risk assessment documents developed for many organophosphate insecticides are also available from the EPA at http://www.epa.gov/oppsrrd1/op/status.htm.

14.8

References

(1988), ‘Cancer risk of pesticides in agricultural workers’, JAMA, 260(7), 959– 966. ANDERSSON A (2000), ‘Comparison of pesticide residues in composite samples and in individual units: the Swedish approach to sampling’, Food Addit. Contam., 17, 547–550. ARCHIBALD S O and WINTER C K (1989), ‘Pesticide residues and cancer risks’, Calif. Agric. 43, 6–9. AMERICAN MEDICAL ASSOCIATION, COUNCIL ON SCIENTIFIC AFFAIRS

BRUHN, C M, WINTER C K, BEALL G A, BROWN S, HARWOOD J O, LAMP C L, STANFORD

and TURNER B (1998), ‘Consumer response to pesticide/ food safety risk statements: implications for consumer education’, Dairy Food Environ. Sanit. 18(5), 278–287. CALIFORNIA DEPARTMENT OF PESTICIDE REGULATION (1999), Overview of the California Pesticide Illness Surveillance Program, 1996, California Environmental Protection Agency, Sacramento, CA. CALIFORNIA DEPARTMENT OF PESTICIDE REGULATION (2000), Residues in Fresh Produce 1997, California Environmental Protection Agency, Sacramento, CA. G, STEINBRING Y J

312

Food chemical safety

and WAYLETT D K (1999), ‘Overestimation bias and other pitfalls associated with the estimated 99.9th percentile in acute dietary exposure assessments’, Reg. Toxicol. Pharmacol., 29, 102–127. COLBURN T, DUMANOSKI D and MYERS J P (1996), Our Stolen Future: Are We Threatening Our Fertility, Intelligence, and Survival? A Scientific Detective Story, Dutton Publishing, New York. ECOBICHON D J (1996), ‘Toxic effects of pesticides’, in Klaassen C D, Casarett & Doull’s Toxicology: The Basic Science of Poisons, 5th Edn, McGrawHill, New York, 643–685. EPA (1999), Human Health Risk Assessment: Methyl Parathion, EPA Office of Pesticide Programs, Health Effects Division, Washington, DC. EPA (2000), Choosing a Percentile of Acute Dietary Exposure as a Threshold of Regulatory Concern, EPA Office of Pesticide Programs, Washington DC. FDA (2000), Food and Drug Administration Pesticide Program: Residue Monitoring 1999, Food and Drug Administration, Washington DC. GARCIA E L and WINTER C K (2000), ‘Pesticide residues in food’, in Francis F J, Wiley Encyclopedia of Food Science and Technology, 2nd Edn, John Wiley & Sons, New York, 1868–1871. GENERAL ACCOUNTING OFFICE (1991), International Food Safety: Comparison of US and Codex Pesticide Standards, GAO/PEMD-91–22, Washington DC. GOLDMAN L R, BELLER M and JACKSON R (1990), ‘Aldicarb food poisonings in California, 1985–1988: toxicity estimates for humans’, Arch. Environ. Health 45, 141–147. HARRIS C A (2000), ‘How the variability issue was uncovered: the history of the UK residue variability findings’, Food Addit. Contam. 17, 491–495. HOAR, S K, BLAIR A, HOLMES F F, BOYSEN C D and ROBEL R J (1986), ‘Agricultural herbicide use and risk of lymphoma and soft tissue sarcoma’, JAMA 256(9), 1141–1147. NATURAL RESOURCES DEFENSE COUNCIL (1989), Intolerable Risk: Pesticides in Our Children’s Food, Washington DC. NRC (1987), Regulating Pesticides in Food: The Delaney Paradox, National Academy Press, Washington DC. NRC (1993), Pesticides in the Diets of Infants and Children, National Academy Press, Washington DC. NRC (1996), Carcinogens and Anticarcinogens in the Human Diet, National Academy Press, Washington DC. PETERSEN B J (2000), ‘Probabilistic modelling: theory and practice’, Food Addit. Contam. 17, 591–599. CHAISSON C F, SIELKEN R L

PIMENTEL D, ACQUAY H, BILTONEN M, RICE P, SILVA M, NELSON J, LIPNER V,

and D’AMORE M (1992), ‘Environmental and economic costs of pesticide use’, BioScience 42, 750–760. USDA (2000), Pesticide Data Program Annual Summary Calendar Year 1998, Agriculture and Marketing Service, Science and Technology, Washington DC. WINTER C K (1992a), ‘Pesticide tolerances and their relevance as safety standards’, Reg. Toxicol. Pharmacol. 15, 137–150. GIORDANO S, HOROWITZ A

Contaminant regulation and management in the United States

313

(1992b), ‘Dietary pesticide risk assessment’, Rev. Environ. Contam. Toxicol. 127, 23–67. WINTER C K (1993), ‘Pesticide residues and the Delaney Clause’, Food Technol. 47(7), 81–86. WINTER C K (1996), ‘Pesticide residues in foods: recent events and emerging issues’, Weed Technol. 10, 969–973. WINTER C K and FRANCIS F J (1997), ‘Assessing, managing, and communicating chemical food risks’, Food Technol. 51(5), 85–92. WOLFF M S, COLLMAN G W, BARRETT J C and HUFF J (1996), ‘Breast cancer and environmental risk factors epidemiological and experimental findings’, Ann. Rev. Pharmacol. Toxicol. 36, 573–596. WINTER C K

Index

acceptable daily intakes (ADIs) 150 pesticides 220–1, 230, 282, 303, 304, 307 veterinary drug residues 112, 116 accreditation agencies and proficiency testing 48 laboratory accreditation 42–3 accuracy 56 acetyl-choline esterase inhibitors 100–1 acidic foods 197–8 acute risk assessments 32, 308–9 ad hoc analysis 46–7 Additional Measures Concerning the Food Control of Foodstuffs (AMFC) Directive 38–9 adhesives 200, 202–3 Advisory Group on Veterinary Residues in Animal Products (AGVR) (UK) 132–3, 133–4 aflatoxins 243–4, 245, 273–4, 288 elimination from contaminated materials 255–6 guidance levels 249–50, 251, 252 Agenda 21 Strategy Document 263 aggregate exposure assessment 33–4, 310 Ah-receptor 92, 93 see also DR-CALUX bioassay Alar Daminozide 230 ALARA (as low as reasonably possible) principle 272 albendazole 126, 127

aldicarb 296 algal toxins 9 alimentary toxic aleukia 242–3 aliphatic chlorinated hydrocarbons 184 alkyl phenols 186 alkylmercurials 156 alkylphenol polyethoxylates 186 aluminium 159–60 aminoglycosides 121, 122 Amnesic Shellfish Poisoning (ASP) 101–2 anabolic agents 4, 5, 111, 123–5, 136, 145 analytical methods 37–70, 71–2 future direction 61–2 FSA surveillance requirements 41 laboratory accreditation and quality control 41–7 legislative requirements for laboratories 38–41 mycotoxins 246–50 general principles 246–7 novel approaches 249–50 proficiency testing schemes for laboratories 47–52 standardised 57–61 validation 53–7 veterinary drug residues surveillance 134–8 see also bioassays; molecular imprinted polymers

Index androgens 125 anilides 128, 129 animal feeds 17–18 Belgian dioxin incident 97–8 animals 142–3 toxicity of mycotoxins 241–2 anthelmintic agents 4, 5, 126–8 antibiotics 4, 117–23 Anticipated Residue Contribution (ARC) 304 antimicrobial agents 4–5, 110–11, 117–23, 144–5 results of surveillance 138–9, 140 antimony 161–2 anti-pesticide lobby 230 Appert, N. 193 apples 26–8 aromatic hydrocarbons 171–2 chlorinated 183–4 polycyclic (PAHs) 8, 98, 99, 172–4 arsenic 155–6, 272 atrazine 77 avermectins 128 azarperone 129, 131 bacteria 9, 59–60 barrier layers 197, 203 beer 152 Belgian dioxin incident 97–8 benzimidazoles 98, 126, 127, 139 benzene 171–2 benzo[a]pyrene (BaP) 99, 173–4 benzophenone 216–17 benzyl butyl phthalate (BBP) 185 beta-adrenergic agonists (beta-agonists) 4, 5, 129, 131, 145 beta-lactam antibiotics 118, 119–20 MIP-based extraction 80–4, 85 beta-lactamase inhibitors 118, 119 bias 307 bile 136 bioassays 91–106 acetyl-choline esterase inhibitors 100–1 dioxins and the DR-CALUX bioassay 92–9 future developments 102 oestrogen assays 100 shellfish poisons 101–2 biological detoxification (biotransformation) 255–6 biomarkers 244 biotoxins 8–9, 58–9, 219, 220 see also mycotoxins

315

bleaching 175 Brazilian citrus pulp incident 96–7 brevitoxins 101–2 cadmium 152–5, 272, 289 California 299–300, 301, 306, 311 cambendazole 126, 127 can coatings 213–14 cancer pesticide risk 222–3, 303–4 risk from mycotoxins 245 carbadox 130–2, 136 carbendazim 231 catalytic MIPs 86 cattle 142 CEN (European Committee for Standardization) 54–5, 57–61 cephalosporins 118, 120 ceramics 208–9 cereals/grain 157–8, 174, 240–1, 253–4 certified reference materials (CRMs) 249, 250 chemical detoxification 255, 256 chemical migration see migration from contact materials chemical preservation 254 chemotherapeutic antimicrobial agents 4–5 chlorinated hydrocarbons 182–4 chlorination of organic chemicals 175 chlormequat 231 chlorobenzenes 183 chlortetracycline 121 chocolate 232 chromatography 137–8, 247–8 chromium 162 Ciguatera toxins 101–2 ciprofloxacin 123 citrus pulp 96–7 clean-up methodology 72, 246–7 clenbuterol 129, 131 MIP-based extraction 79–80, 81, 82 cling films 214–15 cobalt 162 coccidiostats 4, 5, 128, 130, 143 Codex Alimentarius 265–8 committees 10–11 decision making and enforcement mechanisms 268–70 General Standard for Contaminants and Toxins in Food 11, 264–5, 266, 271–4 MRLs compared with US tolerances 304 standards 264, 266, 267, 268–9

316

Index

Codex Alimentarius Commission (CAC) 264, 265–6, 268 analytical methods 53 criteria for laboratories 40–1 Codex Committee on Food Additives and Contaminants (CCFAC) 10, 11, 267, 268, 271, 289 Codex Committee on Pesticide Residues (CCPR) 10, 267, 284 Codex Committee on Residues of Veterinary Drugs in Food (CCRVDF) 10, 115–17, 267, 287 collaborative trials 56–7 collusion 51 combustion 175 Committee for Veterinary Medicinal Products (CVMP) (EU) 115, 285, 286 working group on the safety of residues of veterinary medicines 115, 285 Commodity Codex Committees 266 composite samples 307 concentration 196 confirmatory analyses 135–6 consumer exposure estimates 149–50 migration 206 pesticides 303 consumer perceptions of risk 30–1 consumption data 21, 23–4, 148–50 contact materials see migration from contact materials continental-style chocolates 232 continuous monitoring 45 control materials 44 control measures 9–11 see also regulation; surveillance/ monitoring cooking 18 copper 160–1, 273 corticosteroids 98 Council Directive on the Official Control of Foodstuffs (OCF) (EU) 37, 38 cows’ milk see milk crops impact of mycotoxins on 239–41 inherent toxins 9, 219, 220 prevention and control of mycotoxins 253–6 cumene 171 cumulative risk assessments 310 Danofloxacin 123 DDT 2, 222–3, 234–6

decision analysis 31 decision making 268–70 Delaney Amendment 20–1, 305 deoxynivalenol 288, 289 deterministic exposure estimates 306, 307, 308 detoxification 255–6 di-2–ethylhexyl adipate (DEHA) 214–15 di-2–ethylhexyl phthalate (DEHP) 185, 186 diamphenethide 128, 129 Diarethic Shellfish Poisoning (DSP) 101–2 diary studies 150 dibutyl phthalate (DBP) 185 dietary supplements 156 diffusion 195–6 diffusion coefficient 199 diisopropylnaphthalenes (DIPNs) 204 dimetridazole 139, 140, 141 dioxins 6–7, 175–82, 274 and DR-CALUX bioassay 92–9 domoic acid 101–2 dose-response characterisation 15–16, 19–21, 33 DR-CALUX bioassay 92–9 Belgian dioxin incident 97–8 citrus pulp incident 96–7 development of 92–3 future developments 99 specificity 98–9 use for other types of samples 98 validation for milk fat 93–6 drug withdrawal periods 113, 286 drying 253–4 duplicate diet studies 150 duration of contact 198 E-screen 100 economic losses 239–41 ectoparasiticides 4 see also pesticides eggs 141, 143, 158 electrochemical MIP-based sensors 77 endocrine disrupters 185–6, 223–5 enforcement Codex Alimentarius 268–70 EU 284, 287 Enrofloxacin 123 environmental organic contaminants see organic contaminants Environmental Protection Agency (EPA) (US) 34, 297, 302–4, 305, 306 enzyme linked immunoassays (ELISAs) 138, 249

Index enzyme mimics 86 ER-CALUX assay 100 ergotism 242–3 EROD-assay 92 error 307 ethyl carbamate 8 ethylbenzene 171–2 European Committee for Standardization (CEN) 54–5, 57–61 European Food Authority (EFA) 280–1 European Medicines Evaluation Agency (EMEA) 146, 285 European Union (EU) 150 CVMP 115, 285, 286 directives on chemical migration 208–9 laboratory data 37 analytical methods 53–4 legislative requirements 38–9 regulation 279–94 future trends 289–90 mercury and histamine in fishery products 287–8 other chemical contaminants 288–9 pesticide residues 114–15, 282–4, 289–90 scientific advisory committees 280–1 veterinary drug residues 284–7, 290 SCF 150, 280, 281, 288 trade dispute with USA 31 exposure analysis 16, 21–8 aggregate 33–4, 310 estimating intakes 24–6 food consumption data 23–4 occurrence data 22–3 pesticides 234–6 USA 306–11 probabilistic intake modelling 26–8 falsification of results 51 farmed fish 140–1 fat human 234, 235 milk 93–6 fatty foods 57–8, 197–8 fenbendazole 126, 127 fibres, recycled 204–5 fish dioxins and PCBs in 180, 181 EU regulation of mercury and histamine 287–8 metals and metalloids in 153, 154, 155, 157, 158 veterinary drug residues 140–1, 143

317

flock management 111 fluoroquinolones 123 Food Advisory Committee (UK) 31 Food Balance Sheets (FBSs) 23 food chain environmental organic contaminants entering 170 hazard identification in 16–19 food consumption data 21, 23–4 Food and Drug Administration (FDA) (US) 209, 296, 297–8 food mass:contact area ratio 196–7 food processing 18 food quality 194, 309 Food Quality Protection Act (FQPA) 296, 303, 305 food safety 194, 309 food simulants 207–8 migration testing 209–12 Food Surveillance Programme 22 Food and Veterinary Office (FVO) (EU) 284 Food Standards Agency 37, 41 Four-Plate Test (FPT) 137 fugacity 170 fumonisins 273 functional barrier 197, 203 furocoumarins 219, 220 gas-liquid chromatography (GLC) 248 General Standard for Food Additives (GSFA) 271 General Standard for Contaminants and Toxins in Food (GSCTF) 11, 264–5, 266, 271–4 General Subject Codex Committees 266 genetically modified organisms 61 Good Agricultural Practice (GAP) 222, 251, 282 good practice in the use of veterinary drugs (GPVD) 117 grain/cereals 157–8, 174, 240–1, 253–4 growth-promoters 110–11 non-hormonal 130–2 see also anabolic agents; antimicrobial agents; hormones ham 142 Harmonisation Protocol on Collaborative Studies 56–7 Hazard Analysis Critical Control Points (HACCP) 32 reducing mycotoxins 250–3 hazard identification 15, 16–19

318

Index

health implications of mycotoxins 241–5 animal toxicity 241–2 human toxicity 242–5 pesticide residues and 222–5 potential effects of veterinary drug residues 143–4 research and chemical migration 205–6 herd management 111 hexachlorobenzene (HCB) 183 high performance liquid chromatography (HPLC) 248 high risk groups 228–9 histamine 287–8 hormones 5, 123–5, 136, 145 residues in meat 31 see also anabolic agents human fat 234, 235 human health see health human milk dioxins and PCBs 180, 182 pesticides in 234–6 immortalised human cell lines 249 immune system 244 immunoaffinity clean-up columns 72, 138, 246–7, 249 immunoassays 248–9 imported animal products 145–6 impurities 199, 201–2 in-house method validation 62 in-line analysis 45 infants 228–9 ingredients, packaging 199, 200–1 inks 200, 202–3, 216–17 inorganic contaminants 7, 148–68 consumer exposure estimates 149–50 diary studies 150 metals and metalloids 150–63 nitrate and nitrite 163–5 population exposure estimates 149 risk assessment 150 intake estimates 21, 24–6 pesticides 282–3 probabilistic intake modelling 26–8 internal quality control (IQC) 43–7 basic concepts 43–5 recommendations 45–7 scope of the guidelines 45 International Harmonised Protocol for Proficiency Testing of (Chemical) Analytical Laboratories 48–51 international regulation 10–11, 263–78 Codex General Standard 11, 264–5,

266, 271–4 decision making and enforcement mechanisms 268–70 future trends 274–6 nature of 265–8 ionophores 128, 139, 140 iron 273 isophthalic acid (IPA) 213–14 ivermectin 126, 127, 128 Joint Expert Committee on Food Additives and Contaminants (JECFA) 10, 11, 116, 150 Joint Meetings on Pesticide Residues (JMPR) 10, 282 kidneys, animal 22, 23–4 laboratory accreditation 42–3 laboratory analysis see analytical methods lasalocid 141 leaching 197 lead 150–2, 272, 289 in animal kidneys 22 intake estimates 24–5 legislation see regulation lettuce 233–4, 289 levamisole 126, 127–8 LFRA Ltd 87 LIFELINE model 34 lindane 222 in continental-style chocolates 232 in milk 232–3 Liverpool Bay 158 macrolides 121–3 magnetic MIPs 86 Marbofloxacin 123 maximum limits (MLs) 264–5, 271–4, 274–5 maximum residue levels/limits (MRLs) compared with US tolerances 304 pesticides 222, 282, 283–4, 289–90 veterinary drug residues CCRVDF 116 EU 114, 115, 284–6, 290 UK 113, 114 matrix matched CRMs 249, 250 measurement uncertainty 62, 307 meat curing 163–4 veterinary drug residues 138–40, 142 mebendazole 126, 127 mercury 156–8, 287–8

Index metalloids 150–63 metals 7, 59–60, 150–63, 272–3, 289 see also under individual names methyl parathion 309 microbial inhibition tests 137 microbiology 59–60 micro-exposure event modelling 34 migration from contact materials 8, 193–217 case studies 212–17 factors controlling 196–9 composition of packaging 196 duration of contact 198 mobility of chemicals 199 nature and extent of contact 196–7 nature of the food 197–8 temperature 198 importance 194–5 migration testing 209–12 physico-chemical basis 195–6 range and sources of chemicals posing risks 199–205 impurities and transformation products 199, 201–2 inks and adhesives 200, 202–3 known ingredients 199, 200–1 unknown contaminants 200, 203–5 regulatory context 206–9 research on health issues 205–6 migration testing 209–12 approved ingredients 211 migration of 211–12 contaminants 212 data required 211 overall migration and extractables 212 purpose 210 users of test data 210–11 milk dioxins and PCBs in 180–2 DR-CALUX bioassay validation for milk fat 93–6 human 180, 182, 234–6 lindane in 232–3 MIP-based extraction of beta lactam antibiotics 80–4, 85 veterinary drug residues 141, 143 Minimata disease 156–7 Ministry of Agriculture, Fisheries and Food (MAFF) VMD 112, 132–3 molecular imprinted polymers (MIPs) 71–90 case studies of contaminant analysis 79–84

319

development and application of MIPbased sensors 76–8 future trends 84–6 principles of MIP-based techniques 73–6 Molecularly Imprinted Materials for Integrated Chemical Sensors (MIMICS) 87 molecularly imprinted sorbent assays 86 monitoring see surveillance/monitoring monochlorobenzene (MCB) 183 monomer migration 213–14 Monte Carlo modelling 26–8, 306–9 Morantel 127, 128 multivariate IQC 45 mutual recognition procedure 286 mycotoxins 2, 9, 238–59 analytical methods 246–50 application of HACCP 250–3 contamination routes 239, 240 future trends 256–7 health implications 241–5 prevention and control 253–6 regulation 273–4, 288–9 N-nitrosamines 7 naphthalene 171 naphtoflavone 98 National Adult Dietary Survey (NADS) (UK) 149–50 National Research Council (NRC) (US) 296, 305, 306 natural toxins 8–9, 58–9, 219, 220 see also mycotoxins Neurotoxic Shellfish Poisoning 101–2 New Zealand Total Diet Survey (NZTDS) 218, 219, 225–7 nicarbazin 140, 141 nickel 162–3 nitrate 163–5, 289 nitrite 163–5 nitroxynil 128, 129 No Observed Adverse Effect Level (NOAEL) 19, 220–1, 308, 309 non-hormonal growth promoters 130–2 Non-Statutory Surveillance programmes 133, 141–2, 142–3 non-thresholded contaminants 20, 30 Northern Ireland 142–3 occurrence data 21, 22–3 ochratoxin A 273, 288, 289 oestradiol 186 oestrogen assays 100

320

Index

oestrogens 125, 223–5 okadaic acid 101–2 olaquindox 130–2 organic acids 254 organic contaminants, environmental 6–7, 169–92 aromatic hydrocarbons 171–2 chlorinated hydrocarbons 182–4 dioxins and PCBs 175–82 endocrine disrupters 185–6 phthalic acid esters 184–5 polycyclic aromatic hydrocarbons 172–4 organochlorine chemicals 2, 234, 235 see also DDT organophosphorus (OP) pesticides 2, 218, 219 overall migration 211, 212 oxacillin 80–4, 85 oxfendazole 126, 127 oxyclozanide 128, 129 p-aminobenzoic acid 117, 118 packaging 193 composition of material 196 ingredients used to make 200–1, 211 migration from see migration from contact materials mobility of chemicals in 199 paper and board 8, 216–17 recycled 204–5 Paralytic Shellfish Poisoning (PSP) 101–2 patulin 273–4, 288, 289 pears 231 penicillins 118, 119 persistent environmental chemicals see inorganic contaminants; organic contaminants Pesticide Data Program (PDP) (US) 298, 301 Pesticide Residues Committee (PRC) (UK) 229 pesticides 2–4, 10, 218–37 ADIs 220–1, 230, 282, 303, 304, 307 analytical methods 57–8 EU regulation 114–15, 282–4, 289–90 high risk groups 228–9 human exposure monitoring 234–6 monitoring in food 225–8 MRLs 222, 282, 283–4, 289–90 naturally occurring 219, 220 probabilistic intake modelling 26–8 regulation in USA 295–313 residues and health 222–5 risk from 218–20, 236

UK surveillance 229–31 findings 231–4 use data 310–11 photoinitiator 216–17 phthalic acid esters (phthalates) 2, 8, 184–5, 186 physiologically-based pharmacokinetic (PB-PK) modelling 21, 33 piezoelectric transducer-based MIP sensors 77 pigs 143, 144–5 plant breeding 9 plasticised PVC films 214–15 plastics 8, 199, 200, 208–9 chemicals used to make 200–1 recycled 203–4 plastizyme 86 polychlorinated biphenyls (PCBs) 2, 6, 175–82, 186, 204, 274 polychlorinated dibenzofurans (PCDFs) 175–82, 186 see also dioxins polychlorinated dibenzo-p-dioxins (PCDDs) 175–82, 186 see also dioxins polycyclic aromatic hydrocarbons (PAHs) 8, 98, 99, 172–4 polyethylene terephthalate (PET) 213–14 polymerisation aids 201–2 polyvinylchloride (PVC) 196 DEHA migration from plasticised PVC films 214–15 population exposure estimates 149 postharvest control 253–4 poultry 139–40, 143 precautionary principle 270, 275 precision 56 preharvest control 253 primary food production 16–18 printed cartonboard 216–17 probabilistic exposure assessment 26–8, 306–9 processing, food 18 processing contaminants 7–8 proficiency testing 47–52 international harmonised protocol 48–51 statistical procedure for analysis of results 51–2 prophylactic agents 110 propriopromazine 129, 131 proteins 84–5 Provisional Tolerable Daily Intakes (PTDIs) 150

Index Provisional Tolerable Weekly Intakes (PTWIs) 19–20, 28, 29, 150 psoralens (furocoumarins) 219, 220 pyrethroids 2 quality assurance measures 45 quality of lab data 37–70 internal quality control 43–7 quantitative risk assessment (QRA) 20–1, 29 quinolones 123 rafoxanide 128, 129 reaction products 201–2 ‘reasonable certainty of no harm’ 307 recovery 62 recycled materials 203–5 red meat 138–9 reference doses (RfDs) 303, 307, 308–9 regenerated cellulose film (RCF) 208–9 regulation analytical methods 38–41 chemical migration 206–9 general principles 207–8 historical context 206–7 EU 38–9, 208–9, 279–94 international 10–11, 263–78 USA 209, 295–313 results reporting 51 statistical procedure for analysis of 51–2 retailing 18 reverse phase liquid chromatography (LC) 248 risk assessment 15–36 dose-response characterisation 15–16, 19–21 exposure analysis 16, 21–8 future trends 32–4 hazard identification 15, 16–19 inorganic contaminants 150 mycotoxins and 245 risk communication 16, 32 risk evaluation 16, 28–9 risk management 16, 29–31 USA 301–11 runs, analytical 44–5, 46–7 salmon 140 sample distribution frequency 50 sample extraction and preparation 71–2, 138, 246–8 sampling 45, 49–50

321

Sanitary and Phytosanitary (SPS) Agreement 264, 267–8, 269–70 Sarafloxacin 123 saxitoxins 101–2 scientific advisory committees 280–1 Scientific Committee for Food (SCF) (EU) 150, 280, 281, 288 screening techniques 135 sensors, MIP-based 76–86 shellfish metals in 153, 154, 155, 158, 162–3 poisons 101–2 Society for Molecular Imprinting 87 soil 16–17 solid phase extraction (SPE) 72, 246 specificity 98–9 spinach 289 stachybotryotoxicosis 242–3 standard reference materials 135–6 standardised methods of analysis 57–61 Standing Committee for Foodstuffs (EU) 288 statistical analysis 57 proficiency testing 51–2 Statutory Surveillance programmes 133, 138–41, 142–3 storage 253–4 Streptomycin 121, 122 substituted phenols 128, 129 sulphonamides 117–18, 139, 144–5 supermarkets 230–1, 232 surface acoustic wave (SAW) MIP sensor 77 surface area:food mass ratio 196–7 surveillance/monitoring 10 pesticides 225–8 monitoring vs total diet surveys 227–8 UK 229–34 USA 298–301 veterinary drug residues 132–43 TCDD 176, 186 Technical Barriers to Trade (TBT) Agreement 264, 267–8, 269–70 temperature 198 terephthalic acid (TPA) 213–14 tetrachloroethane (PCE) 184 tetracyclines 118–21 tetrahydro-imidathiazoles 127–8 tetrahydro-pyrimidines 127–8 Theoretical Maximum Residue Contribution (TMRC) 303–4 therapeutic agents 110 thiabendazole 126, 127

322

Index

thin layer chromatography (TLC) 247, 248 3-monochloropropane-1,2-diol (3-MCPD) 8 threshold effect 196 thresholded contaminants 19–20, 29–30 tiger prawns 142 tin 159, 272 tissue selection for analysis 136 Tolerable Daily Intakes (TDIs) 150 tolerances 302–4 toluene 171–2 total diet surveys New Zealand 218, 219, 225–7 pesticide monitoring vs 227–8 UK 149 US-FDA 298 Toxic Equivalency Factors (TEFs) 176–7, 178 toxic equivalents (TEQs) 176–7 toxicity 170 dioxins and PCBs 177–9 mycotoxins 241–5 animal toxicity 241–2 human toxicity 242–5 toxicological hazard profile 205–6 toxicological thresholds contaminants with 19–20, 29–30 contaminants without 20, 30 trace elements 60–1 see also metals trade dispute between EU and USA 31 international 269–70 tranquillisers 4, 5, 129, 131 transformation products 199, 201–2 trichloroethane (TCE) 184 trout 141 ‘true’ result 50 2,3,7,8–tetrachlorodibenzo-p-dioxin (TCDD) 176, 186 2,4–D MIP sensor 77 Tylosin 121–3 uncertainty, measurement 62, 307 United Kingdom pesticide surveillance 229–31 findings 231–4 surveillance programmes for veterinary drug residues 132–4 analytical methods 134–8 results 138–43 United Kingdom Accreditation Service (UKAS) 39, 42 United States (USA)

Department of Agriculture 298 EPA 34, 297, 302–4, 305, 306 FDA 209, 296, 297–8 pesticide regulation 295–313 future trends 310–11 improving pesticide management 305–9 managing pesticides in foods 301–4 monitoring 298–301 regulatory agencies 297–8 trade dispute with EU 31 validation of analytical methods 53–7 in-house 62 requirements for valid methods 56–7 vegetables 164–5, 174, 289 veterinary drug residues 2, 4–6, 10, 109–47 chemical substances used 117–32 control in UK 112–17 current issues 144–6 EU regulation 284–7, 290 growth-promoting agents 110–11 herd and flock management 111 and human health 143–4 prophylactic agents 110 surveillance 132–4 analytical methods 134–8 results for UK 138–43 therapeutic agents 110 Veterinary Medicines Directorate (VMD) (UK) 112, 132–3 Veterinary Products Committee (VPC) (UK) 112–13 vinclozolin 233–4 vinyl chloride monomer (VCM) 196 violative residues 298–301, 309 water 152 wine 152 withdrawal periods 113, 286 worker re-entry intervals 304 World Trade Organization (WTO) 31 SPS and TBT 264, 267–8, 269–70 xenoestrogens 223–5 xylenes 171–2 yams 231 Yellow Rice Disease 242–3 z-score 51–2 zearalenone 273–4 zinc 161, 273