INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY
ENVIRONMENTAL HEALTH CRITERIA 70
PRINCIPLES FOR THE SAFETY ASSESSMENT OF
FOOD ADDITIVES AND CONTAMINANTS IN FOOD
This report contains the collective views of an international group of
experts and does not necessarily represent the decisions or the stated
policy of the United Nations Environment Programme, the International
Labour Organisation, the World Health Organization, or the Food
Agriculture Organization of the United Nations
Published under the joint sponsorship of
the United Nations Environment Programme,
the International Labour Organisation,
and the World Health Organization in
collaboration with the Food and Agriculture
Organization of the United Nations
World Health Orgnization
Geneva, 1987
The International Programme on Chemical Safety (IPCS) is a
joint venture of the United Nations Environment Programme, the
International Labour Organisation, and the World Health
Organization. The main objective of the IPCS is to carry out and
disseminate evaluations of the effects of chemicals on human health
and the quality of the environment. Supporting activities include
the development of epidemiological, experimental laboratory, and
risk-assessment methods that could produce internationally
comparable results, and the development of manpower in the field of
toxicology. Other activities carried out by the IPCS include the
development of know-how for coping with chemical accidents,
coordination of laboratory testing and epidemiological studies, and
promotion of research on the mechanisms of the biological action of
chemicals.
ISBN 92 4 154270 5
The World Health Organization welcomes requests for permission
to reproduce or translate its publications, in part or in full.
Applications and enquiries should be addressed to the Office of
Publications, World Health Organization, Geneva, Switzerland, which
will be glad to provide the latest information on any changes made
to the text, plans for new editions, and reprints and translations
already available.
(c) World Health Organization 1987
Publications of the World Health Organization enjoy copyright
protection in accordance with the provisions of Protocol 2 of the
Universal Copyright Convention. All rights reserved.
The designations employed and the presentation of the material
in this publication do not imply the expression of any opinion
whatsoever on the part of the Secretariat of the World Health
Organization concerning the legal status of any country, territory,
city or area or of its authorities, or concerning the delimitation
of its frontiers or boundaries.
The mention of specific companies or of certain manufacturers'
products does not imply that they are endorsed or recommended by the
World Health Organization in preference to others of a similar
nature that are not mentioned. Errors and omissions excepted, the
names of proprietary products are distinguished by initial capital
letters.
CONTENTS
PRINCIPLES FOR THE SAFETY ASSESSMENT OF FOOD ADDITIVES AND CONTAMINANTS
IN FOOD
FOREWORD
PREFACE
1. INTRODUCTION
2. HISTORICAL BACKGROUND
2.1. Introduction
2.2. Periodic review
2.2.1. Concept of periodic review
2.2.2. Mechanism of periodic review
3. CRITERIA FOR TESTING AND EVALUATION
3.1. Criteria for testing requirements
3.1.1. Estimating exposure
3.1.2. Predicting toxicity from chemical structure
3.1.3. Other factors to consider when developing criteria
3.2. Priorities for testing and evaluation
3.3. Quality of data
4. CHEMICAL COMPOSITION AND THE DEVELOPMENT OF SPECIFICATIONS
4.1. Identity and purity
4.2. Reactions and fate of food additives and contaminants in food
4.3. Specifications
5. TEST PROCEDURES AND EVALUATION
5.1. End-points in experimental toxicity studies
5.1.1. Effects with functional manifestations0
5.1.2. Non-neoplastic lesions with morphological manifestations
5.1.3. Neoplasms
5.1.4. Reproduction/developmental toxicity
5.1.5. In vitro studies
5.2. The use of metabolic and pharmacokinetic studies in safety
assessment
5.2.1. Identifying relevant animal species.
5.2.2. Determining the mechanisms of toxicity
5.2.3. Metabolism into normal body constituents
5.2.4. Influence of the gut microflora in safety assessment
5.2.4.1 Effects of the gut microflora on the chemical
5.2.4.2 Effects of the chemical on the gut microflora
5.3. Influence of age, nutritional status, and health status on the
design and interpretation of studies
5.3.1. Age
5.3.1.1 History
5.3.1.2 Usefulness of studies involving in utero exposure
5.3.1.3 Complications of aging
5.3.2. Nutritional status
5.3.3. Health status
5.3.4. Study design
5.4. Use of human studies in safety evaluation
5.4.1. Epidemiological studies
5.4.2. Food intolerance
5.5. Setting the ADI
5.5.1. Determination of the no-observed-effect level
5.5.2. Use of the safety factor
5.5.3. Toxicological versus physiological responses
5.5.4. Group ADIs
5.5.5. Special situations
5.5.6. Comparing the ADI with potential exposure
6. PRINCIPLES RELATED TO SPECIFIC GROUPS OF SUBSTANCES
6.1. Substances consumed in small amounts
6.1.1. Food contaminants
6.1.2. Food flavouring agents
6.2. Substances consumed in large amounts
6.2.1. Chemical composition, specifications, and impurities
6.2.2. Nutritional studies
6.2.3. Toxicity studies
6.2.4. Foods from novel sources
REFERENCES
ANNEX I. GLOSSARY
ANNEX II. STATISTICAL ASPECTS OF TOXICITY STUDIES
REFERENCES TO ANNEX II
ANNEX III. GUIDELINES FOR THE EVALUATION OF VARIOUS GROUPS OF FOOD
ADDITIVES AND CONTAMINANTS
REFERENCES TO ANNEX III
ANNEX IV. EXAMPLES OF THE USE OF METABOLIC STUDIES IN THE SAFETY
ASSESSMENT OF FOOD ADDITIVES
REFERENCES TO ANNEX IV
ANNEX V. APPROXIMATE RELATION OF PARTS PER MILLION IN DIET TO MG/KG PER
DAY
ANNEX VI. REPORTS AND OTHER DOCUMENTS RESULTING FROM MEETINGS OF
THE JOINT FAO/WHO EXPERT COMMITTEE ON FOOD ADDITIVES
WHO TASK GROUP ON UPDATING THE PRINCIPLES FOR THE SAFETY ASSESSMENT OF
FOOD ADDITIVES AND CONTAMINANTS IN FOOD
1985 and/or 1986 JECFA Members
a,b,c,d,e Dr H. Blumenthal, Division of Toxicology, Center for
Food Safety and Applied Nutrition, US Food and
Drug Administration, Washington DC, USA
c,d Dr I. Chakravarty, Department of Biochemistry and
Nutrition, All India Institute of Hygiene and
Public Health, Calcutta, India
c Dr W.H.B. Denner, Food Composition and Information
Unit, Food Sciences Division, Ministry of Agri-
culture, Fisheries and Food, London, United
Kingdom
c Dr A.H. El-Sebae, Pesticides Division, Faculty of
Agriculture, Alexandria University, Alexandria,
Egypt
c Professor P.E. Fournier, Hôpital Fernand-Widal,
Paris, France
a,c,d,f Dr S. Gunner, Food Directorate, Health Protection
Branch, Health and Welfare Canada, Ottawa,
Ontario, Canada
d Mr J. Howlett, Food Science Division, Ministry of
Agriculture, Fisheries and Food, London, United
Kingdom
c,d,f Professor K. Kojima, College of Environmental Health,
Azabu University, Sagamihara-Shi, Kanagawa-Ken,
Japan
c Dr W. Kroenert, Food Chemistry Division, Max von
Pettenkofer Institute, Federal Office of Public
Health, Berlin (West)
a,b,c,d,f Dr B. MacGibbon, Division of Toxicology, Environ-
mental Pollution and Prevention, Department of
Health and Social Security, London, United
Kingdom
c,d Dr R. Mathews, Food Chemicals Codex, National
Academy of Sciences, Washington DC, USA
d Mrs I. Meyland, National Food Institute, Ministry of
the Environment, Soborg, Denmark
c,d,e Dr J. Modderman, Food Additives Chemistry Evaluation
Branch, Center for Food Safety and Applied
Nutrition, US Food and Drug Administration,
Washington DC, USA
d Dr G. Nazario, Ministry of Health, National Health
Council, Rio de Janeiro, Brazil
c,d Professor K.A. Odusote, College of Medicine, Univ-
ersity of Lagos, Lagos, Nigeria
c,d Professor F. Pellerin, Faculté de Pharmacie de
l'Université Paris-Sud, Hôpital Corentin Celton,
Issy-les-Moulineaux, France
c,f Dr P. Pothisiri, Food Control Division, Food and
Drug Administration, Ministry of Public Health,
Bangkok, Thailand
b,c,d Professor M.J. Rand, Department of Pharmacology,
University of Melbourne, Parkville, Victoria,
Australia
a,b,d,e,g Dr P. Shubik, Green College, Oxford, United Kingdom
d Dr A. Slorach, Food Research Department, The National
Food Administration, Uppsala, Sweden
d Dr V.A. Tutelyn, Institute of Nutrition, Academy of
Medical Sciences of the USSR, Moscow, USSR
Secretariat
d Dr Y.K. Al-Mutawa, Division of Public Health Labor-
atory, Ministry of Public Health, Safat, Kuwait
a Dr E.A. Bababunmi, University of Ibadan, Ibadan,
Nigeria
e Dr A. Bär, Department of Vitamin and Nutrition
Research, F. Hoffmann-LaRoche and Company, Ltd.,
Basel, Switzerland
a,b,d,f Dr J. Cabral, Unit of Mechanisms of Carcinogenesis,
International Agency for Research on Cancer,
Lyons, France
e Dr J. Caldwell, St. Mary's Hospital Medical School,
London, United Kingdom
a,e Dr D.M. Conning, British Nutrition Council, London,
United Kingdom
f Dr J.L. Emerson, External Technical Affairs, The
Coca-Cola Company, Atlanta, Georgia, USA
d Mr A. Feberwee, Committee on Food Additives, Nutri-
tion and Quality Affairs, Ministry of Agriculture
and Fisheries, The Hague, The Netherlands
d Professor C.L. Galli, Toxicology Laboratory,
Institute of Pharmacology and Pharmacognosy,
University of Milan, Milan, Italy
e Dr M.J. Goldblatt, Consumer Nutrition Affairs,
General Foods Corporation, White Plains,New
York, USA
f Dr W. Grunow, Divison of Food Toxicology, Max von
Pettenkofer Institute, Federal Office of Public
Health, Berlin (West)
d Mr R. Haigh, Commission of the European Communities,
Brussels, Belgium
a,d,f Dr Y. Hayashi, Division of Pathology, Biological
Safety Research Center, National Institute of
Hygienic Sciences, Tokyo, Japan
b,d,e,g Dr J. Herrman, Division of Food and Color Additives,
Center for Food Safety and Applied Nutrition,
US Food and Drug Administration, Washington DC,
USA
f Dr D. Krewski, Environmental Health Directorate,
Health Protection Branch, Health and Welfare
Canada, Ottawa, Ontario, Canada
e Mr P.N. Lee, Consultant in Statistics and Adviser in
Epidemiology and Toxicology, Surrey, United
Kingdom
b Dr M. Mercier, International Programme on Chemical
Safety, World Health Organization, Geneva,
Switzerland
d Dr R.W. Moch, Center for Food Safety and Applied
Nutrition, US Food and Drug Administration,
Washington DC, USA
a Dr V.H. Morgenroth, Chemicals Division, Environment
Directorate, Organization for Economic Cooper-
ation and Development, Paris, France
a Dr I. Nir, Department of Pharmacology and Experi-
mental Therapeutics, The Hebrew University
Hadassah Medical School, Jerusalem, Israel
d Dr E. Poulsen, National Food Institute, Institute of
Toxicology, Soborg, Denmark
b,d,f Dr A.W. Randell, Food Policy and Nutrition Division,
Food and Agricultural Organization of the United
Nations, Rome, Italy
d Dr N. Rao Maturu, Joint FAO/WHO Food Standards
Programme, Food and Agricultural Organization of
the United Nations, Rome, Italy
e,f Dr A.G. Renwick, Clinical Pharmacology Group, Univ-
ersity of Southampton, Southampton, United
Kingdom
e,f Dr F.J.C. Roe, Consultant in Toxicology and Adviser
in Experimental Pathology and Cancer Research,
London, United Kingdom
d,e Dr S.I. Shibko, Division of Toxicology, Center for
Food Safety and Applied Nutrition, US Food and
Drug Administration, Washington DC, USA
a,b,e,f Dr V. Silano, Department of Comparative Toxicology,
High Institute of Health, Rome, Italy
d Professor A. Somogyi, Department of Drugs, Animal
Nutrition and Residue Research, Institute for
Vetinary Medicine, Berlin (West)
b,d Professor R. Truhaut, Faculté des Sciences Pharma-
ceutiques et Biologiques de Paris Luxembourg,
Laboratoire de Toxicologie et Hygiene Indus-
rielle, Université Rene Descartes, Paris, France
f Dr G.J. Van Esch, National Institute for Public
Health and Environmental Hygiene, Bilthoven,
The Netherlands
a,b,d Dr G. Vettorazzi, International Programme on Chemical
Safety, World Health Organization, Geneva,
Switzerland
e Dr M.J. Wade, Division of Toxicology, Center for
Food Safety and Applied Nutrition, Washington DC,
USA
a,b,d,e,g Dr R. Walker, Department of Biochemistry, University
of Surrey, Guildford, Surrey, United Kingdom
-----------------------------------------------------------------
a Present at strategy meeting, Oxford, United Kingdom, 19-23
September, 1983.
b Present at pre-consultation of contributors, Geneva,
Switzerland, 29-31 May, 1985.
c Member of JECFA-85, Geneva, Switzerland, 3-12 June, 1985.
d Participant in JECFA-86, Rome, Italy, 2-11 June, 1986.
e Consultant who contributed written material.
f Submitter of written comments.
g Member of Editorial Committee.
NOTE TO READERS OF THE CRITERIA DOCUMENTS
Every effort has been made to present information in the
criteria documents as accurately as possible without unduly
delaying their publication. In the interest of all users of the
environmental health criteria documents, readers are kindly
requested to communicate any errors that may have occurred to
the Manager of the International Programme on Chemical Safety,
World Health Organization, Geneva, Switzerland, in order that
they may be included in corrigenda, which will appear in
subsequent volumes.
* * *
FOREWORD
The WHO activities concerned with the safety assessment of
food chemicals were incorporated into the International
Programme on Chemical Safety (IPCS) in 1980. Since this time, a
keen interest has developed in all aspects pertaining to the
toxicological evaluation of food additives and contaminants,
including the methodological aspects. These activities are part
of the responsibilities of the Programme insofar that its
objectives include the formulation of "guiding principles for
exposure limits, such as acceptable daily intakes for food
additives and pesticide residues, and tolerances for toxic
substances in food, air, water, soil, and the working
environment".
The present publication on "Principles for the Safety
Assessment of Food Additives and Contaminants in Food" has been
developed in response to repeated recommendations by the Joint
FAO/WHO Expert Committee on Food Additives (JECFA). Its
inclusion in the methodology section of the Environmental Health
Criteria series will make it readily available to both Member
States and the food industry.
The IPCS gratefully acknowledges the financial support of
the United Kingdom Department of Health and Social Security
(DHSS), and the US Food and Drug Administration (FDA), which was
indispensable for the completion of the project.
Dr M. Mercier
Manager
International Programme on
Chemical Safety
PREFACE
For the last thirty years, the internationally sponsored
committee known as the Joint FAO/WHO Expert Committee on Food
Additives (JECFA) has played a major role in providing a unique
international mechanism for the identification and safety
assessment of food chemicals, including food additives, food
contaminants, and residues of veterinary drugs. With no
regulatory aspirations, this Committee has probably contributed
more to the elaboration of sound national food regulation in
this area than any other international body aimed at harmonizing
or normalizing often divergent national approaches to the
problem of food safety, food technology, and food control.
JECFA achieved this by providing recommendations based on
scientific evidence and by establishing a rational model of
safety assessment that is widely reputed and accepted.
Hundreds of highly skilled international specialists have
given, and continue to give, freely of their time and talents to
foster advances in toxicological methodologies and analytical
procedures, to consolidate accessible presentations of data, and
to keep abreast with scientific developments, which often
requires readjustment of previous conclusions. Through reports,
toxicological monographs, and profiles of chemical
specifications by JECFA, national food regulatory authorities
and the Codex Alimentarius Commission are provided with all the
necessary elements for making the best decisions on the rational
use of chemicals in food.
The present undertaking has several precedents in the
history of JECFA. For example, in 1957, the second report of
the Committee elaborated on procedures for the testing of
intentional food additives to establish their safety for use
and, in 1960, the fifth report contained a series of guidelines
for the evaluation of the carcinogenic hazards of food
additives. It should also be mentioned that, in 1966, the
Committee commissioned a special scientific group to develop
procedures for investigating intentional and unintentional food
additives. Finally, in 1981, after realizing that a significant
interval of time had elapsed since previous methodological
updatings, the Committee called for a state-of-the-art review of
methodology. A favourable answer was received from the newly
established International Programme on Chemical Safety (IPCS), a
cooperative programme sponsored by the International Labour
Organisation (ILO), the United Nations Environment Programme
(UNEP), and the World Health Organization (WHO). It should be
noted that the implementation of the recommendation by the IPCS
was significantly facilitated by the fact that the toxicological
component of the JECFA came within the scope of the programme.
The contents of this publication are the result of sustained
efforts of an IPCS Task Group during a number of meetings
including: a strategy meeting in 1983, a consultation of
contributors in 1985, and the JECFA Working Groups at their
annual meetings in 1985 and 1986. The members of the Task Group
contributed either written material or comments, or both. An
Editorial Board was responsible for preparing the final draft
for publication. The Task Group benefited widely from the large
number of recommendations and observations on the methodology of
testing and assessing chemicals in food, found in the previous
reports of JECFA and related scientific groups.
Thus, this publication reflects faithfully the
recommendations of the Committee regarding the safety assessment
of food additives and contaminants by reaffirming the validity
of recommendations that are still appropriate, while pointing
out the problems associated with those that, in the light of
modern advances in methodology, are no longer valid. New
recommendations are also made, as might be expected, with the
advancing state of the toxicological sciences. Particularly
enlightening are the section dealing with the principles related
to the safety assessment of substances consumed in large
amounts, and Annexes II and IV dealing with the statistical
aspects of toxicity studies and examples of the use of metabolic
studies in the safety assessment of food additives,
respectively. An Index has also been included at the end of the
book.
It is the most earnest wish of all concerned with the
production of this publication that it should make an important
contribution to the field of food toxicology and that it will be
found useful by the members of JECFA, national food regulatory
authorities, and industry, who are involved in the development
of safety data and in making the consumer aware of the problems
of the safe use of food additives.
Dr G. Vettorazzi
Dr A. Randell
1. INTRODUCTION
This publication is concerned with reviewing the basis for
decision-making by the Joint Expert Committee on Food Additives
(JECFA) of the Food and Agriculture Organization of the United
Nations (FAO) and the World Health Organization (WHO). Because
the toxicological and chemical characteristics of food additives
are the primary concern of the Committee, both aspects are dealt
with in this publication, which has been prepared by WHO- and
FAO-appointed consultants, assisted by the WHO and FAO
secretariats. The twenty-eighth JECFA report (1) includes a
summary of the areas that the Committee considered to be most in
need of evaluation.
The major concerns of this monograph are with the testing of
chemicals in food and with the evaluation of the test results.
In keeping with the approach developed during the past 30 years
by JECFA, the recommendations for test procedures and safety
assessment are discussed in broad terms taking into account the
latest scientific advances in the relevant fields. No effort
has been made to provide instructions for test procedures.
Differences in national approaches to toxicological evaluations
exist, and some variations in the data submitted must be
considered by JECFA in making their evaluations. However,
certain basic data requirements are necessary in order to enable
an expert committee to make sound judgements. The elements of
this base line are included in various sections of this report.
In essence, the problems under consideration fall into three
general categories: first, the determination of the chemical and
toxicological test requirements for individual chemicals that
are added to or occur in food; second, the assessment methods
that are to be applied; and third, the updating of the test
procedures and methods of assessment as the science progresses.
Methods for testing and assessment have changed considerably
during the life of JECFA. However, the Committee has, by no
means, been a static organization. Not only have various
approaches been continuously updated at the individual meetings
of the Committee, but intervening meetings of scientific groups
have been held to consider the impact of new scientific
developments on the procedures used (2, 3). Thus, every effort
has been made in this publication to record the views of JECFA
and to detail the changes that have come about with the course
of time.
It is recognized that, with advances in science, it is
possible to obtain more complete toxicological profiles for
individual chemicals. For example, it is becoming increasingly
easy to learn more about the disposition and metabolic fate of
xenobiotics. Until relatively recently, the toxicologist in
this field had to rely almost entirely on a set of routine
tests, the results of which were then assessed and used to
establish arbitrarily-determined safe levels.
JECFA has long recognized that a number of factors can be
used to determine test requirements; these include the structure
of the chemical, its natural occurrence in foodstuffs, its
metabolic characteristics, and knowledge of its effects in man.
However, systematic guidelines incorporating these factors have
not been developed by JECFA. In recent years, some of the
problems posed to JECFA have concerned the testing and
evaluation of additives and food ingredients that are consumed
in large amounts. Other key problems in the determination of
the appropriate level of testing involve the largest group of
food additives, the flavouring agents that are used, generally,
at very low levels and that are often "nature identical" or
derived from natural sources. Therefore, both "high
consumption" substances and flavouring agents are discussed in
detail in sections 6.1 and 6.2
There have been considerable changes in laboratory studies
used in other areas of toxicology, which are yet to have a major
impact on the evaluation of food additives. Of particular
interest are the series of mutagenicity/clastogenicity tests,
often, but not invariably, using sub-mammalian test organisms in
vitro. The number, diversity, and uses of these tests have
increased rapidly in the past decade. In general, such tests
are effective for measuring an intended genetic end-point. How
effectively these tests identify chemical carcinogens is much
less clear. In the absence of a clear correlation with
carcinogenicity, it is difficult to know how such tests should
be interpreted and used in safety evaluations. Even though
these in vitro tests may not be required for the evaluation of
the safety of food additives, it is becoming more and more
frequent for chemicals to be tested in this way for other
reasons, e.g., to detect potential environmental or occupational
hazards. Thus, JECFA may have to decide on the relevance of
such information (section 5.1.5).
During the past decade, there has been a major increase in
the number of chemicals tested routinely for chronic toxicity in
standardized in vivo tests. Although these tests may not be
designed to evaluate food additives, the results have to be
carefully considered at JECFA meetings. Among the major
problems that occur are those arising from results obtained when
chemicals are administered to animals by routes other than the
diet or drinking-water. For example, some years ago, JECFA
faced the problem of assessing the significance of the induction
of subcutaneous sarcomas at the site of injection of certain
food chemicals into rodents. It was found that many substances,
including inert plastics, could give rise to similar sarcomas on
implantation. As a result, the Committee concluded that such
findings could not be used in a definitive manner for assessing
the safety of food additives (4, pp. 16-17). However, these
findings cannot be totally ignored and may be an indicator of
the need for further carcinogenicity studies using the oral
route. Another problem with interpretation arises with many of
the more recent "routine" in vivo studies that record the
enhancement of a variety of common "spontaneous" tumours in
rodents, including lymphomas, hepatomas, and pheochromocytomas
(section 5.1.3).
The scope of long-term toxicity tests has been discussed
extensively by JECFA. For example, several food chemicals have
been tested in 2-generation studies rather than the commonly-
used single-generation study. While the use of this more
extensive test is advisable under certain conditions, it should
not necessarily be a routine procedure (section 5.3).
Many of the chemicals of concern to those responsible for
food safety evaluation are present in food at very low levels
and may be present as environmental contaminants or may result
from the migration of substances from food packaging or residues
from the use of solvents, pesticides, or veterinary drugs.
These situations often require very different approaches to test
requirements than those used for intentional food additives
(Annex III). One case, the use of anabolic agents in livestock,
has posed various problems for JECFA that cannot be answered
within the scope of current procedures (Annex III). JECFA will
soon be developing methodology for the testing and evaluation of
veterinary drug residues in support of a new committee on
residues of veterinary drugs in foods that has been established
by the Codex Alimentarius Commission.
In assessing the significance of data, a major issue to be
resolved concerns the distinctions that should be made among
different toxicological manifestations. The carcinogenic
potential of chemicals has been emphasized in the past few
decades to the exclusion of most other toxic end-points. There
was a general consensus that chemicals found to be carcinogenic
were not appropriate as food additives at any level whatsoever.
More recently, however, it has become widely accepted that the
term "carcinogen" has become harder and harder to define
(section 5.1). It is apparent that cancer can be induced by a
variety of chemicals acting by very different mechanisms and
that the mechanism should be an important consideration when
determining whether a safe level can be established. Other
questions concern whether high-dose animal data are relevant to
human exposure to low levels, and how teratogenicity data in the
absence of maternal toxicity are to be interpreted (section
5.1.4).
2. HISTORICAL BACKGROUND
2.1. Introduction
The Joint FAO/WHO Expert Committee on Food Additives (JECFA)
was established following recommendations made to the Directors-
General of FAO and WHO by the Joint FAO/WHO Expert Committee on
Nutrition at it fourth session (5), and the subsequent first
Joint FAO/WHO Conference on Food Additives was held in Septem-
ber, 1955 (6). The terms of reference of the earlier meetings
of JECFA related to the formulation of general principles gov-
erning the use of food additives and consideration of suitable
uniform methods for evaluating the safety of food additives.
For these purposes, food additives were defined by the Joint
Conference as "non-nutritive substances added intentionally to
food, generally in small quantities, to improve its appearance,
flavour, texture, or storage properties."a Following recom-
mendations of the third Joint FAO/WHO Conference on Food Addi-
tives (8), these terms of reference were broadened to include
substances unintentionally introduced into human food and JECFA
has subsequently considered and evaluated such materials,
including growth promoters, components of packaging materials,
solvents used in food processing, aerosol propellants, enzymes
used in food processing, and metals in foods. Novel foods and
ingredients that may be incorporated into foods at levels higher
than those previously envisaged for food additives have also
been referred to JECFA and pose special problems in safety
evaluation, which will be discussed later in this report
(section 6.2).
The first (9), second (10), and fifth Reports (4) of JECFA
established principles for the use of food additives and made
recommendations on methods for establishing the safety-in-use of
food additives and for the evaluation of the carcinogenic
hazards of food additives. From the outset, JECFA recognized
that:
"no single pattern of tests could cover adequately, but
not wastefully, the testing of substances so diverse in
structure and function as food additives" and that
"the establishment of a uniform set of experimental
procedures that would be standardized and obligatory is
therefore undesirable" (10).
----------------------------------
a From a practical standpoint, the "food additive" definition
has been expanded since the time it was drafted, as a
variety of compounds, including nutritive substances
consumed in high amounts, have been brought under the
umbrella of food additives. Indeed, the second Joint
FAO/WHO Conference on Food Additives (7) recommended that
the scope of the JECFA programme be expanded beyond the
substances included in the original definition.
Accordingly, this Committee concluded "that it was only possible
to formulate general recommendations with regard to testing
procedures." The Committee also recognized that advances in the
basic sciences might suggest new approaches to toxicological
investigations and that these might be used immediately by the
scientist but would take longer to become incorporated into any
officially recommended testing procedures. Subsequent meetings
of JECFA have consistently adopted this approach and have
avoided the adoption of fixed protocols for the testing and
evaluation of all classes of intentional and unintentional food
additives. This has had the advantage of allowing the Committee
to respond to new problems as they have arisen, with minimal
inertia, and to encompass non-routine and ad hoc studies in the
safety evaluation process. Within this framework, the Committee
has found it possible to formulate guidelines for the evaluation
of several groups of intentional and unintentional food
additives that posed their own peculiar problems; several of
these guidelines, which serve as specific examples to support
general principles, are contained in Annex III.
The requirement to keep abreast of scientific developments
in toxicology and related scientific disciplines implies the
need for a periodic review of testing methodology. Following
recommendations to this effect made by the eighth (11) and ninth
(12) meetings of JECFA, a WHO Scientific Group on Procedures for
Investigating Intentional and Unintentional Food Additives was
convened in 1966
"to review, in the light of new scientific knowledge,
the criteria used in establishing acceptable daily
intakes. . . ." and "to suggest further studies on
toxicological procedures used for the evaluation of
intentional and unintentional food additives in order
to establish their safety to the consumer" (2).
Subsequent meetings of JECFA have taken cognisance of the report
of this Scientific Group and of the report of a more recent WHO
Scientific Group on the Assessment of Carcinogenicity and Muta-
genicity of Chemicals (3). Some aspects of the reports of other
WHO Scientific Groups on the Principles for the Testing and
Evaluation of Drugs for Carcinogenicity (13), Mutagenicity (14),
and Teratogenicity (15) are also pertinent to the methodology
of testing food additives. However, significant developments in
the science of toxicology and related disciplines led the
seventeenth meeting of JECFA to recommend that "the methods and
procedures for the toxicity testing of food additives should be
comprehensively reviewed and brought into line with advances in
toxicology and cognate disciplines" (16). This recommendation
was reiterated in the reports of the eighteenth (17) and
nineteenth (18) meetings, the latter of which called for the
convening of an appropriate meeting for the purpose of the
review, and reaffirmed at the twentieth meeting (19).
Safety evaluation of food additives is a 2-stage process.
The first stage involves the collection of relevant data
including the results of studies on experimental animals and,
where possible, observations in man. The second stage involves
the assessment of data to determine the acceptability of the
substance as a food additive. While the recommendations
referred to in the preceding paragraph emphasize the impact of
scientific advances on the first testing stage, the impact of
such advances extends also to the assessment stage. This was
made explicit in the twenty-first report of JECFA (20, p. 31),
which stated that:
"in view of the rapid progress of the science of toxi-
cology and the increasing refinement of evaluation pro-
cedures, the Committee felt strongly that the tradi-
tional concepts of setting ADIs, the application of
safety factors, and the relationship of these safety
factors to the observed toxicological manifestations in
animal experiments should be reconsidered".
This recommendation was endorsed by the twenty-fourth meeting of
JECFA (21).
Many features of toxicity testing and evaluation of advent-
itious food additives and contaminants that fall within the
terms of reference of JECFA, are common to pesticides that are
within the scope of the Joint FAO/WHO Meeting on Pesticide
Residues. In recognition of this, the twenty-fifth meeting of
JECFA (22) recommended that a group of experts should be
convened, as soon as possible, to study the application of
advances in methodology to the toxicological evaluation of food
additives and contaminants, and also of pesticide residues. The
urgency of the need to implement this recommendation was
stressed by the twenty-sixth (23) and twenty-seventh (24) meet-
ings of JECFA.
In response to the Committee's repeated recommendations, a
meeting of a group of experts to study the application of
advances in methodology to the toxicological evaluation of food
additives and contaminants was convened in September 1983. The
objectives of the meeting were to formulate specific recommend-
ations in order to bring up to date:
(a) the principles set out in earlier reports of JECFA
concerning safety evaluation in relation to specific
toxicological problems or specific chemical entities or
groups;
(b) the test methods used in the toxicological evaluation
of chemicals in food; and
(c) the assessment procedures adopted by JECFA in
determining quantitative end-points.
The report of the Working Group (Updating Principles of
Methodology for Testing and Assessing Chemicals in Food: Report
of a Strategy Meeting) (unpublished WHO document ICS(Food)/83.3)
and working papers on specific issues were considered by the
twenty-eighth meeting of JECFA (1). Several questions were
identified as remaining to be considered, including special
problems associated with:
(a) bulking agents and novel foods;
(b) food contaminants;
(c) animal feed additives and veterinary drug residues;
(d) test methods and principles (including alternative
methods of testing);
(e) testing for allergenicity;
(f) lesions observed in bioassays that are difficult to
interpret (a number of examples are cited in the
report); and
(g) assessment procedures; extrapolation and quantitative
assessment.
The Committee recommended that a unified document on these
issues should be prepared for consideration by JECFA at a future
meeting. The present publication has been prepared in response
to that recommendation.
In carrying out the review of methodology for the testing
and evaluation of intentional and unintentional food additives,
the working group has taken notice of recommendations, guide-
lines, and procedures adopted by national regulatory authorities
and international/supra-national organizations including the
Organization for Economic Cooperation and Development (OECD),
the International Agency for Research on Cancer (IARC), and the
European Economic Community (EEC) Scientific Committee for
Food. It is recognized that, where possible, a unified approach
should be adopted. However, the purposes for which these other
bodies have formulated guidelines differ in detail from those of
JECFA, and it is inappropriate to adopt these without modifi-
cation to meet the needs of JECFA.
2.2. Periodic Review
2.2.1. Concept of periodic review
JECFA has indicated that, in discharging its duty to eval-
uate the safety-in-use of intentional and unintentional food
additives, it may be necessary to carry out a periodic re-
evaluation of substances previously assessed by the Committee.
The first JECFA meeting, in looking ahead, envisaged, in
addition to the continuing evaluation of food additives, that
there would be a re-evaluation process associated with the
programme on food additive safety assessment (9). It stated:
"Permitted additives should be subjected to continuing
observation for possible deleterious effects under
changing conditions of use. They should be reappraised
whenever indicated by advances in knowledge. Special
recognition in such reappraisals should be given to
improvements in toxicological methodology."
This principle was endorsed in the third (25), seventh (26),
eighth (11), and ninth reports (12) of JECFA.
The "need for review of past recommendations" was high-
lighted in the thirteenth JECFA report as follows (27, p. 22):
"There is a widespread but fallacious belief that
clearance of an additive for use in food constitutes an
irrevocable decision. Such a view renders a grave dis-
service to the cause of consumer protection for it
fails to recognize the need for regular review of all
safety evaluations."
Periodic review of past decisions on safety is made
necessary by one or more of the following developments (27):
(a) A new manufacturing process for the food additive.
(b) A new specification.
(c) New data on the biological properties of the compound.
(d) New data concerning the nature, or the biological
properties, or both, of the impurities present in a
food additive.
(e) Advances in scientific knowledge germane to the nature
or mode of action of food additives.
(f) Changes in consumption patterns or level of use of a
food additive.
(g) Improved standards of safety evaluation. This is made
possible by new scientific knowledge and the quality
and quantity of safety data considered necessary in the
case of new additives. Since JECFA began the eval-
uation of food additives in 1956, the paucity of
information available on many food additives has been
such that assessments have often been difficult to
make. Tests of too short duration, conducted with a
very small number of animals at inappropriate dose
levels, and without adequate clinical, haematological,
chemical, or histopathological examinations have
frequently been encountered among the data submitted
for evaluation. Tests of this sort cannot be regarded
as having permanent validity; with the passing of time,
they need to be supplemented by studies carried out in
full accordance with the recommendations set out in the
report of the WHO Scientific Group on Procedures for
Investigating Intentional and Unintentional Food
Additives (2).
It should also be noted that the second Joint FAO/WHO
Conference on Food Additives (7) recommended that it (JECFA)
"should revise, if needed, the toxicological evaluation of all
additives considered in previous meetings of the Expert
Committee."
The seventeenth report of JECFA (16) reads, in part:
"The objective in assessing the toxicological data on
food additives is to ensure their safety for the con-
sumer on the basis of all the evidence available to the
Committee at the time. Future results with present
methods or with techniques yet to be developed will
necessitate reassessments that may lead to changes in
earlier decisions."
Other meetings reaffirmed the need to take advantage of
recent developments in toxicological techniques for research and
safety assessment.
These Committee recommendations and observations, the
rationale set forth in the thirteenth report (27) and other
reports for the need to review past decisions, and the ensuing
years of progress in the science of toxicology and refinement of
research, evaluation procedures, and changing consumption
patterns all point towards the advisability of periodic review
of this large class of substances.
2.2.2. Mechanism of periodic review
That a considerable amount of re-evaluation of substances is
already carried out within the system is evident when the year-
to-year agenda of JECFA is examined. Food additives are reass-
essed when new biological and chemical data are made available
to FAO and WHO. In fact, new data are mandated on timetables
established by JECFA when temporary ADIs are established (sec-
tion 5.5.5). In addition, re-evaluations are made at the request
of Member States and by the Codex Alimentarius Commission.
However, for many additives, the assessment has not been
conducted using the more recently adopted procedures for
investigating intentional and unintentional food additives. A
review of past decisions also reveals that some additives have
only had a cursory examination. The evaluation of these addi-
tives may have been based on limited data.
A periodic review programme on substances previously
reviewed by JECFA should be constituted to reflect the changing
state-of-the-art and to provide the best possible assurance to
consumers of food additive safety. However, a mechanism has not
yet been developed for the continuous systematic updating of
safety information on food additives. Of course, even in the
absence of a periodic review programme, if new data on a food
additive raises suspicion of significant hazard, then immediate
re-evaluation is conducted.
The use of an international forum to devise and implement a
system for the periodic review of chemicals used in or on food
and contaminants of food could also be of great economic and
practical value to Member States. It would ensure a uniform
approach to a complex toxicological problem, duplication of
effort would be minimized, and emphasis on such a programme
would give added reassurance to consumers throughout the world
that the food supply continues to be safe. Perhaps such a
programme could be developed in cooperation with the Codex
Alimentarius Commission.
3. CRITERIA FOR TESTING AND EVALUATION
JECFA has always operated on the principle that testing
requirements for all food additives should not be the same.
Such factors as expected toxicity, exposure levels, natural
occurrence in food (section 6.1), occurrence as normal body
constituents (section 5.2), use in traditional foods, and
knowledge of effects in man (section 5.4) should be taken into
account. In relation to carcinogenic hazards, the Committee has
stated that "the scope of the test required should depend on a
number of factors, such as the nature of the substance, the
extent to which it might be present in food, and the population
consuming it" (4). More generally, the Committee has requested
data on, inter alia, method(s) of manufacture, impurities, fate
in food, levels of use of additives in food, and estimates of
actual daily intake, and concluded that such information "was
important and relevant both for the toxicological evaluation and
for the preparation of specifications" (22). However, difficul-
ties arise when an attempt is made to determine testing require-
ments because of problems in predicting toxicity, estimating
levels of food additive use and natural occurrence, and
obtaining human data. As discussed below, criteria for testing
requirements can also be used to allocate priorities for the
testing and evaluation of food chemicals.
3.1. Criteria for Testing Requirements
The establishment of principles for determining the
appropriate amount of data that will be required to adequately
evaluate the safety of additives at their estimated consumption
levels is urgently needed to ensure consistency in decision-
making by the Committee and to provide guidance to sponsors of
food chemicals. Both exposure data and potential toxicity
should be important considerations in the establishment of these
principles. Consideration of only one of these elements to the
exclusion of the other leaves serious deficiencies. If only
exposure data are used, then no consideration is given to the
wide range of toxicities observed among chemicals and no
advantage is taken of the vast amount of bioassay data already
in existence. If, on the other hand, only toxicity information,
predicted or known, is used, then chemicals with known toxic
properties or those related to chemicals of known high toxicity,
particularly carcinogenicity, would automatically require the
most data, with no consideration given to relatively low
exposure levels.
3.1.1. Estimating exposure
For the purpose of this publication, exposure is defined as
the total intake of a chemical substance by human beings. For
the majority of substances evaluated by JECFA, the primary mode
of exposure is through ingestion of the substance in the food
supply.
Estimates of exposure used by Committees in previous years
are of three general types: per capita estimates, estimates from
dietary food intake surveys, and analytical values from market-
basket/total-diet surveys. For a detailed discussion of the
advantages and the use of these different types of estimates see
reference no. 28.
The per capita approach is an estimated value that repre-
sents the exposure level if a food additive or contaminant were
equally distributed across a population. For example, a per
capita intake for a nation may be calculated by dividing the
total yearly production volume, corrected for imports and
exports, of a chemical used in food, within a nation, by the
national population. Another form of per capita intake may be
computed from a nation's per capita disappearance of a certain
food commodity multiplied by the usual level of an additive or
contaminant in the food commodity. These per capita intakes can
be converted to daily intake per kilogram body weight.
In some countries, dietary surveys are performed on food-
stuffs consumed by a representative group of individuals, within
a national population, over a short period of time, e.g., 1 - 14
days. The intake of an additive or contaminant, per food type,
can be calculated by multiplying the usual additive or contam-
inant level in each type of food by the dietary intake of the
food. The intakes per food type can then be summed to derive a
total additive or contaminant intake. An advantage of the
dietary survey approach is that additive or contaminant intakes
for selected subpopulations, such as different age groups or
high-frequency consumers of certain foodstuffs, may be computed,
depending on the specificity of the dietary survey.
When considering intakes computed by the dietary survey
approach, the tenth meeting of JECFA (29, pp. 23-24) reaffirmed
the validity of calculating the average daily intake of a food
additive based on: (a) levels arising from good technological
practice; (b) average consumption of foods containing the
additive; and (c) average body weight. This Committee also
noted at the time that, while data on average food consumption
were available from many countries, high consumption data were
available from only two countries. The Committee recognized a
special need for determining how much of a food additive is
likely to be consumed by groups that have a high level of
consumption and strongly recommended that every effort should be
made to obtain such information on food consumption.
The fourteenth meeting of JECFA (30) considered methodol-
ogies for computing additive intakes from dietary food surveys,
and recognized the importance of experimental design so that
collective data can be used for calculating reliable intakes on
an individual basis. The Committee noted difficulties in common
descriptors for food items, when information is gathered in
surveys performed by different organizations, and in obtaining
confidential information about food additive use from industry.
Market-basket surveys (also called total-diet surveys)
involve analyses of representative diets for the usual level of
additive or contaminant in the diet. The analyses may be per-
formed on food mixtures or on individual foodstuffs. The
selection of foods represents a normal diet for a certain popu-
lation, such as the typical daily diet for a certain nation's
average consumer. Market-basket surveys can be used for
estimating the actual level of additive or contaminant in a
selected total diet, which is of value for substances present in
food at levels that are less than the amounts added. However,
the difficulties of analyses usually restrict this approach to
estimations of average intakes of contaminants in samples repre-
sentative of the average dietary habits of a nation's general
population rather than estimations of intakes for selected
subpopulations. In this regard, data on certain contaminants in
food are available from the Global Environmental Monitoring
System (GEMS).
These procedures are useful for estimating exposure to food
additives in individual countries. However, accurate estimation
is much more difficult when attempted on a global scale.
Clearly, consumption of a food additive will not be the same in
two countries in which it is regulated with differing restric-
tions or with very different food consumption patterns. To use
exposure estimates on such a scale as a criterion for testing
requirements or for setting priorities for the testing of food
additives is an extremely ambitious exercise that would require
extensive resources.
3.1.2. Predicting toxicity from chemical structure
Chemical structure determines to a great extent the attitude
of the toxicologist towards a compound. As a result, there have
been many efforts to systematize the use of chemical structure
as a predictor of toxicity. The use of such relationships has
been suggested by JECFA with certain classes of flavouring
agents (section 6.1.2), and chemical structure is an important
consideration in the selection of compounds for carcinogenicity
testing. Structure/activity relationships also form the basis
for establishing group ADIs (section 5.5.4).
Structure/activity relationships appear to provide a
reasonably good basis for predicting toxicity for some
categories of compounds, primarily carcinogens, which are char-
acterized by specific functional groups (e.g., nitrosamines,
carbamates, epoxides, and aromatic amines) or by structural
features and specific atomic arrangements (e.g., polycyclic
aromatic hydrocarbons and aflatoxins). However, all these chem-
ical groups have some members that do not seem to be carcino-
genic or are only weakly so. Since structure/activity relation-
ships are better established for carcinogens than for other
toxic end-points, dependence on such predictions emphasizes
suspect carcinogens at the expense of other forms of potential
toxicity. However, as more chemicals are tested for toxicity
and other end-points are identified in the future, the data base
will become larger, which should permit more valid comparisons
between structure and toxicity among more classes of compounds.
In terms of carcinogenic substances, another system that has
sometimes been used for predictive purposes is a battery of
tests for genotoxicity (possible applications of such tests are
discussed in section 5.1.5).
3.1.3. Other factors to consider when developing criteria
The value of structure/activity relationships and exposure
data in determining the extent of testing required may be
considerably enhanced by collateral information on metabolism
and pharmacokinetics. It has been previously accepted that:
"if a series of chemical analogues can be shown to give
rise to the same metabolic product. . . it may be suf-
ficient to carry out toxicological studies on a suit-
able representative of the series" and "where adequate
biochemical and toxicological data on closely related
chemicals are available, the objective (of toxicity
tests) becomes the detection of any deviation from the
established pattern. This can usually be determined by
intensive studies of a few months duration when these
are adequately designed and evaluated" (2).
More recently (31), JECFA has concluded that:
"if the chemical structure of a compound under consid-
eration did not closely resemble that of any known
toxic or carcinogenic compound, and, if the toxicolo-
gical data on it, its metabolites, and its homologues
did not give any cause for concern, these less exten-
sive toxicological data might be used for the evalua-
tion of the compound. . . . In the evaluation of a
series of structurally-related compounds, complete
toxicological data should be available for at least one
member of the series. Other compounds in the series
should be evaluated on the basis of these data, plus
data on their natural occurrence and metabolism, and on
the toxicology of their homologous compounds."
These principles can form the basis for determining the lim-
ited amount of testing that may be required for compounds that
are closely related structurally. If the toxicological data
base is adequate for the homologous compounds and suggests a low
intrinsic toxicity, metabolic and pharmacokinetic data alone may
be sufficient to make an evaluation of a related compound.
The results of studies on absorption, distribution, and
metabolism may either increase or decrease the health concern
from the use of the additive. For example, a relatively non-
toxic additive may be transformed by liver enzymes into a sub-
stance with a much greater toxic potential, or vice versa.
Correlations between structure and activity will often auto-
matically include these considerations, because substances of a
particular class will often be absorbed, distributed, and meta-
bolized in similar ways. However, this will not always be the
case, and these parameters should be specifically considered
when making such correlations.
Other factors influence the extent and type of testing
required for safety assessment. For example, the need for
extensive testing may be mitigated when the substance occurs
naturally in food and has a history of human use or when it is
metabolized into normal body constituents (section 5.2.3). More
extensive testing in animals may be necessary when the additive
will be used in special populations at risk, such as pregnant
women and very young infants (section 5.3). Human testing may
be needed if problems of intolerance arise (section 5.4.2). The
types of end-points, as discussed in section 5.1, must be
considered in any criteria system that is established.
The development of criteria for determining the extent of
required testing is worthy of extensive future study. Its value
would be considerably enhanced by including it in the context of
a priority-setting scheme, as discussed below, because additives
should not be considered in isolation from one another.
3.2. Priorities for Testing and Evaluation
The primary basis for establishing the list of substances to
be considered by JECFA is the recommendations of the Codex
Committee on Food Additives (CCFA) and Member Governments.
However, Committees have recognized the need for the estab-
lishment of a "priority list as a means of selecting the most
relevant compounds for future evaluation. In order to establish
priorities for the toxicological testing and evaluation of
intentional and unintentional food additives", JECFA recommended
at the twenty-second (32) and twenty-third (31) meetings that:
"FAO and WHO convene an inter-disciplinary group of
experts to establish an inventory of compounds that
have not yet been fully evaluated and to classify them
in terms of their potential hazard to health on the
basis of toxicological knowledge and extent of use."
The Committee has recognized that the most obvious need for
allocating priorities is for the testing and safety evaluation
of food flavouring agents (19). Committees continue to stress
the need to establish priorities for testing and evaluating food
additives (24,33).
One basis for establishing priorities for testing food
additives is by using an index based on exposure levels and
predicted toxicity. For examples of approaches using these
parameters, see references 34-38.
As discussed in section 3.1.1, valid exposure data are
extremely difficult to develop. However, comparative levels of
consumption may be sufficient for the purpose of setting prior-
ities for the testing of food additives. Therefore, even though
accurate consumption estimates of wide geographical relevance
will probably never be achieved, the lesser requirement of com-
parative estimates may be achievable to the extent necessary for
JECFA's use, if the Committee decides to develop the informa-
tion.
Because of the semiquantitative nature of much of the
biological data available for predicting toxicity, rigorous
analytical or statistical interpretation is not always possible.
Therefore, expert interpretation and evaluation of the data, a
time-consuming and expensive procedure, must be integrated into
any automated decision-making mechanism that is developed. To
ensure maximum usefulness, the priority-setting system should
take account of all available toxicity and other biological
information, including metabolic and human data. A properly-
devised system will be capable of considering new data and can
be modified using modern data processing methods and equipment.
3.3. Quality of Data
In recent years, various national regulatory agencies and
international bodies have instituted codes of Good Laboratory
Practice (GLP), the aim of which is to help underwrite the
validity of studies by ensuring that they can be verified and
reproduced. GLP codes require the maintenance of certain
records regarding the performance of studies, including data
from chemical and toxicological tests, which help ensure full
documentation of the conduct and results of studies. However,
GLP codes are not a substitute for scientific quality; an
inappropriate study may be conducted according to GLP standards.
On the other hand, a study that does not meet GLP criteria may
still be scientifically sound.
JECFA has always judged studies on their merits, the main
criteria being that the study was: (a) carried out with
scientific rigor, and (b) reported in sufficient detail to
enable comprehensive evaluation of the validity of the results.
Usually, studies that are published in the scientific
literature are subjected to peer review prior to publication,
and after publication, the results are open to refutation or
confirmation. Unpublished reports, on the other hand, are not
necessarily subjected to this scrutiny. For this reason, JECFA
has repeatedly recommended that data brought before it be
published (10, p. 6; 12, p. 7; 39, p. 7; 40). However, in point
of fact, JECFA does review many high-quality studies that remain
unpublished for proprietary and other reasons. Also, the
Committee often requests unpublished raw data when published
reports do not include sufficiently detailed data for an
adequate safety review. Studies performed in compliance with
GLP codes provide added assurance that the quality of
unpublished data is acceptable. For these reasons, it is
appropriate that JECFA experts continue to consider all the data
brought before them, published, and unpublished, and make
decisions about the validity of these data on an ad hoc basis.
This means that the studies reviewed by the Committee should be
fully documented.
4. CHEMICAL COMPOSITION AND THE DEVELOPMENT OF SPECIFICATIONS
The proper safety evaluation and use of food additives
requires that they be chemically characterized. Therefore, the
Committees review data relating to the identity of additives,
impurities that may be present, and possible reaction products
that may arise during storage or processing. "JECFA
specifications" are then elaborated, taking these and other
factors into account.
4.1. Identity and Purity
To establish the chemical identities of food additives, it
is necessary to know the nature of the raw materials, methods of
manufacture, and impurities (22). This information is used to
assess the completeness of analytical data on the composition of
additives and to assess the similarity of materials used in
biological testing with those commercially produced. From
information on raw materials and methods of manufacture,
potential impurities in commercially manufactured chemical
materials due to carry-over of contaminants in raw materials and
by-products of the manufacturing process can be predicted.
To evaluate biological testing data from multiple studies,
JECFA must have information on the chemical composition of the
tested materials, which necessitates manufacturing information.
Analytical data on the chemical composition of materials used in
biological testing should be more detailed than a standard
presentation of chemical specifications. Furthermore, materials
used in biological testing should be representative of
substances manufactured by actual commercial processes so that
the materials administered to experimental animals will
represent those ingested by consumers.
A food additive may be a single chemical substance, a
manufactured complex chemical mixture, or a natural product.
The need for complete information on chemical composition,
including description, raw materials, methods of manufacture,
and analyses for impurities, is equally valid for each type of
additive. However, implementation of the requirement for
chemical composition data may vary depending on the type of
substance. For additives that are single chemical substances,
it is virtually impossible to remove all impurities in their
commercial production; therefore, analyses are generally
performed on the major components and predicted impurities, with
the highest significance placed on potentially toxic impurities.
For commercially manufactured complex mixtures, such as mono-
and diglycerides, information is needed on the range of
substances commercially produced, with emphasis on descriptions
of manufacturing processes, supported by analytical data on the
components of the different commercial products. Natural pro-
ducts present particularly difficult problems because of their
biological variability and because the chemical constituents are
too numerous for regular analytical determinations; thus, the
analyst is starting with an "unknown". For additives derived
from natural products, it is vital that the sources and methods
of manufacture are precisely defined. Chemical composition data
should include analyses for general chemical characteristics,
such as proximate analyses for protein, fat, moisture, carbo-
hydrate, and mineral content, and analyses for specific toxic
impurities carried over from raw materials or chemicals used in
the manufacture of the substance. Further information necessary
for the evaluation of "novel foods", which are usually sub-
stances derived from natural products, is provided in sections
6.2.1 and 6.2.4.
4.2. Reactions and Fate of Food Additives and Contaminants in Food
Biological testing of food chemicals must relate to their
presence in food as consumed. This is an important consider-
ation when added substances undergo chemical change in food.
Therefore, data are necessary on the reactions and fate of addi-
tives or contaminants in food and their effects on nutrients
(22).
Certain food additives perform their functional effect by
reaction with undesirable food constituents (e.g., antioxidants
react with oxygen in food and EDTA reacts with trace metals) or
by reactions that modify food constituents (e.g., potassium
bromate reacts with dough constituents). Food additives may
also degrade under certain conditions of food processing, though
such degradation is detrimental to their functional effect. For
example, the sweetener aspartame is transformed to a diketo-
piperazine derivative at rates varying with the acidity and the
temperature of the food. In previous evaluations of "reactive"
additives, Committees have evaluated analyses for additive
reaction products in food, as consumed, and biological testing
data on either specific reaction products or samples of food
containing the reaction products as consumers would ingest
them.
For all intentional food additives proposed for evaluation,
Committees request submission of four types of data related to
reactivity:
(a) the general chemical reactivity of the additive;
(b) its stability during storage and reactions in model
systems;
(c) reactions of the additive in actual food systems; and
(d) the additive's fate in living systems. These data are
important for relating biological test data to the
actual use of the additive in food.
Processing aids are substances that come into contact with
food during processing and may unintentionally become part of
food because of their incomplete removal. Committees have eval-
uated a number of processing aids, such as extraction solvents
and enzyme preparations, for their safety in use. When evalua-
ting a processing aid, information should be provided on its use
and either analytical data on the amount of the processing aid
carried over into food or a computed estimate of the amount to
be expected in food. In some cases, a component of the proces-
sing aid may have the greatest potential for biological effects,
such as ethylenimine leaching from polyethylenimine, an immobil-
izing agent used in the preparation of immobilized enzyme
preparations.
Contaminants in food evaluated by previous Committees
include environmental contaminants and substances migrating from
food packaging. Of environmental contaminants, metals have been
considered the most often. Committees request information on
the chemical forms of metals in the food supply (e.g., ionic
form and/or covalently bonded chemical form) and their concen-
tration distribution in the food supply, as determined by
analyses of food or experimental models for carry-over from
environmental sources. For contaminants derived from food
packaging, data are required on the identification of chemicals
migrating from the packaging material and concentrations in food
(analysed or estimated from migration modelling studies).
4.3. Specifications
Specifications are a necessary product of Committee eval-
uations, the purposes of which are 3-fold:
(a) to identify the substance that has been biologically
tested;
(b) to ensure that the substance is of the quality required
for safe use in food; and
(c) to reflect and encourage good manufacturing practice.
The first Joint FAO/WHO Conference on Food Additives (6)
established a programme for the collection and dissemination of
information on the chemical, physical, pharmacological, toxico-
logical, and other properties of individual food additives. The
first two meetings of the Joint Expert Committee, in preparing
reports on "General Principles Governing the Use of Food
Additives" (9) and "Procedures for the Testing of Intentional
Food Additives to Establish Their Safety for Use" (10),
recommended that specifications should be prepared, citing the
need for:
(a) limiting impurities in food;
(b) identifying materials used in toxicity testing; and
(c) ensuring that the additive tested is the additive used
in practice.
The third meeting of JECFA was devoted in its entirety to dev-
eloping principles governing the elaboration of specifications
and developing provisional specifications for the first group of
additives evaluated by the Committee (25).
JECFA specifications are minimum requirements for the compo-
sition and quality of food-grade additives, allowing for accep-
table variation in their production (18). These specifications
are meant to be used internationally and to the extent that data
are available, specifications are elaborated to cover suitable
products manufactured in various parts of the world. The third
meeting considered the value of specifications with regard to
protection for the consumer, advice to regulatory organizations,
standards for the food industry, and establishment of safety for
use (relative to identification of materials used in biological
testing in comparison with materials produced for commercial
use). The format for specifications established by this meeting
continues to be used in current JECFA specifications, that is,
the additive is identified by synonym, definition (chemical
name, formula, relative molecular mass, etc.), and description,
its functional uses are listed, tests of identity and impurities
are provided, and an assay for the major component(s) is pro-
vided. The third meeting of JECFA, recognizing that practical
specifications could not specify every impurity, limited the
scope of impurity tests to constituents of commercially produced
substances that: (a) were related to the safe use of the
additive; (b) might affect the usefulness of the additive; or
(c) would serve as an indicator of good manufacturing practice.
Finally, the meeting concluded that, for specifications to be
acceptable on a global basis, they must be subject to continuing
review and evaluation to take into account the presentation of
new information, particularly with respect to different
manufacturing processes and improved analytical methods.
In detailing the purposes of JECFA specifications, Commi-
ttees have, through the years, refined the scope of their spec-
ifications and provided advice on how they should be used.
Specifications developed by each Committee should be read in
conjunction with the report of that Committee. JECFA specifi-
cations apply to the material(s) that was toxicologically
reviewed and take into account the uses of the additive (17,
41). Periodic review of specifications is required, because of
changes in patterns of additive use, in raw materials, and in
methods of manufacture. Comments on JECFA specifications by
national and international organizations are valuable sources of
information for periodic review (18, 27, 29).
JECFA specifications in their entirety describe substances
of food-grade quality, and as such, they are directly related to
toxicological evaluations and to good manufacturing practice.
However, though specifications may include criteria that are
important for commercial users of additives, they do not include
requirements that are of interest only to commercial users
(42).
Differences may exist between specifications prepared by
national and international organizations; however, the Committee
is not aware of any information indicating that these differ-
ences incur health risks for consumers (23). JECFA specifica-
tions are meant to be minimum requirements for the safe use of
additives, and not every component of commercially manufactured
chemical substances is subject to an impurity test (11). Test
requirements in JECFA specifications are sufficient to ensure
the safe use of commercially manufactured food additives. Sub-
stances of higher chemical purity (e.g., analytical grade
reagents) are not excluded from use in food, even though such
substances may deviate somewhat from the identification tests in
the specifications, provided that they meet the stated require-
ments for specified purity tests and are otherwise suitable for
use as food additives (18).
Since 1956, the meetings of JECFA have designated specifi-
cations as either full or tentative. Specifications were given
the tentative designation from the third to twenty-second
meetings because either the chemistry data needed to prepare
specifications were not adequate or a temporary ADI was assigned
to the additive. At, and since, the twenty-third meeting of
JECFA, the tentative designation has been assigned only when the
data necessary for preparing specifications were insufficient.
JECFA policy has been to prepare specifications for sub-
stances added to food, whenever constituents of the substance
had the potential to be present in food. Initially, specifica-
tions were prepared only for intentional food additives that
were added directly, to accomplish a functional effect in food.
The fourteenth JECFA (30) evaluated extraction solvents, the
first group of "processing aids" that had been reviewed by
JECFA. This Committee concluded that, although extraction sol-
vents are substantially removed from food, evaluation of the
conditions of safe use of these solvents depends on the identity
and purity of the solvents. Therefore, JECFA specifications
were prepared. Since then, specifications have been reported
for other processing aids such as anti-foaming/defoaming agents,
clarifying agents, decolourizing agents, enzyme preparations,
filtering aids, packing gases, propellants, lubricants/release
agents, odour/taste-removing agents, and yeast "food" (yeast
nutrients). The twenty-seventh JECFA (24) decided that chemical
reagents used in the preparation of food additives or processing
aids (such as glutaraldehyde in the preparation of immobilized
enzyme preparations or acetic anhydride in the manufacture of
modified starches) do not usually need specifications. Carry-
over of these reagents or their contaminants into food may be
controlled in the specifications for purity of the additive or
processing aid.
Food additives may be marketed as formulated preparations,
such as a mixture of a main ingredient with a solvent vehicle
and emulsifier. Specifications refer only to each ingredient in
the formulated preparation as individual commercially-manufac-
tured food additive substances. Mixtures should not be formu-
lated in such a way that the absorption or metabolism of any
ingredient is altered so that the biological data are invali-
dated (12, 42). Added substances such as anticaking agents,
antioxidants, and stabilizers may also influence the results of
tests given in specifications. Therefore, in its nineteenth
report, JECFA recommended that manufacturers of food additives
should indicate the presence of such added substances (18).
In considering whether specifications apply to food additive
quality as manufactured or as received, JECFA has decided to
prepare specifications to cover the normal shelf-life of the
product. Limits are set for decomposition products that may
form during normal storage. Manufacturers and users of food
additives should ensure good packaging and storage conditions
and use good handling practices to minimize deleterious changes
in quality and purity (18). Information on changes in the
composition of food additives during storage should be submitted
for evaluation by the Committee.
In addition to periodic reviews to examine the consistency
of specifications within classes of similar additives, JECFA
periodically reviews specification test methods to update the
analytical methodology of specifications. Two summaries of
specification test methods have been published (43, 44), which
provide guidelines for the application and interpretation of
specification requirements and test methods. JECFA has made
considerable progress in adopting modern analytical methodology
for specification tests, whenever equipment and other supplies
needed to perform the tests are accessible on a world-wide
basis. However, because JECFA specifications are elaborated for
world-wide use, certain analytical methods involving recently-
developed techniques or equipment cannot be included until such
techniques are available on an international scale. Alternative
methods of analysis can be used to test products for conformity
with specifications, provided that the methods and procedures
used produce results of equivalent accuracy and specificity.
In order to foster international agreement on specifications
for food-grade substances, JECFA seeks comments from member
countries and international organizations. The Codex Aliment-
arius Commission systematically provides comments on JECFA
specifications through the Codex Committee on Food Additives
(CCFA) and endorses certain JECFA specifications as "Codex
Advisory Specifications". The systematic review of JECFA
specifications by the CCFA has provided JECFA with valuable data
on novel manufacturing processes, previously unknown impurities,
updated methodology, and advice on the format of JECFA speci-
fications.
Although JECFA specifications and those of the Codex
Alimentarius Commission are elaborated for many of the same
purposes, the interpretation of these purposes may result in
differences in specific requirements or test methods for the
same food-grade substance. In replying to suggested changes in
JECFA specifications from the CCFA or other interested parties,
it may be decided to amend existing specifications, providing
that the requested changes do not significantly lessen the
assurance of food-grade quality embodied within the JECFA
specifications and that the requested change conforms with the
principles for elaboration of specifications established at
previous meetings. A requested change in an existing full JECFA
specification must be supported by scientific data.
5. TEST PROCEDURES AND EVALUATION
5.1. End-Points in Experimental Toxicity Studies
There are virtually no findings in experimental toxicology
that can be simply extrapolated to man without careful thought.
During the past two decades, there has been an increasing emph-
asis on carcinogenesis as a manifestation of chemical toxicity.
Most of the other manifestations of chemical intoxication, for
example, immunosuppression, have, by comparison, received rela-
tively little attention. This has resulted in an unbalanced
approach by the toxicologist in which the emphasis on end-points
bears a less and less obvious relationship with disease patterns
in man. For example, a survey of the recommendations for the
evaluation of food additives has revealed that little or no
attention is paid to the detection of cardiovascular lesions,
even though these lesions are the most common cause of fatal-
ities in the human population in developed countries. In
addition, certain lesions commonly found in rodents that are the
primary targets of the toxicologist do not have any counterpart
in man. It seems reasonable that an effort should be made to
relate toxicological findings more carefully to the human situ-
ation, recognizing that this will be a long-term project. In
the meantime, when conducting experimental animal tests, special
attention should be paid to alterations that indicate a poten-
tial for the test compound to adversely affect the cardio-
vascular, immunological, reproductive, or central nervous
systems. If such alterations are detected, they should be
investigated further using special studies aimed at clarifying
their significance.
The end-points discussed in this section have been grouped
for convenience into effects with:
(a) functional manifestations only;
(b) non-neoplastic morphological characteristics;
(c) neoplastic manifestations; and
(d) reproduction/developmental manifestations.
In view of the large number of effects encountered, it is
possible to summarize only some of the specific observations.
However, situations that have become controversial are dealt
with in more detail.
Finally, this section concludes with a short discussion on
the role of short-term in vitro tests in the safety assessment
of food additives.
5.1.1. Effects with functional manifestations
Generalized weight loss, although having causes that are not
solely physiological (section 5.5.3), does not necessarily
involve any particular pathological lesions (section 5.5.3).
Reduced weight gain has played a major role as an end-point in
toxicological determinations in various ways. In a sense, it
has often been used for determining various empirical indices in
the absence of other manifestations. The procedures followed by
JECFA for determining an ADI demand that a no-observed-effect
level should be established. For this level to be established,
it is necessary to establish an effect level and, when all else
has failed, a generalized decrement in weight gain has been used
for this purpose, provided reduced food intake is not the
obvious cause. The other major use of a decrease in weight gain
has been in establishing a maximum tolerated dose (MTD) (for a
definition of the MTD, see Annex I).
Among the commonest effects observed in studies on food
additives is a laxative effect; the physiological reasons for
this are usually quite apparent and can be taken into account
when considering the appropriate levels of use of additives
causing this effect. In most instances, additives have been
permitted that cause laxation at high levels in man when they
have been otherwise non-toxic and can be used effectively at
levels at which laxation does not occur.
Although a great deal has been said about the need to
evaluate food additives and contaminants for the induction of
possible behavioural changes, JECFA has hitherto devoted little
time to evaluating such changes. Since it has been suggested
that certain food constituents can produce behavioural changes
in man, JECFA will, in the future, undoubtedly have to consider
such effects. Unfortunately, good animal models have not been
developed, and objective human data are difficult to obtain. It
is not possible to recommend a simple series of tests at this
time, primarily because there is no clear consensus on the kinds
of studies that should be performed nor on the interpretation of
the results.
Intolerance to food additives should always be considered a
possibility, even though tests for reactions to food additives
are not part of the normal data package that JECFA considers
when assessing new food additives. Even when evidence of
widespread intolerance to a food additive appears, it may still
be very difficult to determine a cause-and-effect relationship.
Problems include the often-anecdotal nature of much of the
evidence, psychological factors, and problems with developing an
adequate central data collection system. These points are
discussed in more detail in section 5.4.
5.1.2. Non-neoplastic lesions with morphological manifestations
A number of lesions are relatively frequently observed in in
vivo studies, particularly in rodents, that often give rise to
controversy.
Non-specific liver enlargement or hypertrophy has been
discussed by a WHO scientific group (2, pp. 18-19). In the
past, this occurrence has been considered to be a manifestation
of toxicity. More recently, it has been realized that this can
often be a physiological response involving the induction of
microsomal enzymes in the detoxification process that is
reversible on removal of the test compound.
The formation of calculi in the urinary bladder is a
frequent finding in rodent studies. Often, the formation of
calculi may be followed by the formation of bladder tumours. It
is not uncommon to find calculus formation in one rodent species
and not in another. Under these circumstances, the nature of
the calculi can sometimes be associated with specific metabolic
changes that have led to their formation. This, in turn, may
allow for a scientifically-based extrapolation to man, providing
that human clinical studies are possible.
Caecal enlargement, a common finding in rodent studies, is a
normal finding in rodents maintained on standard laboratory
diets under germ-free conditions. It is also a common response
of rodents, especially rats, to diets that include non-nutrient
substances (e.g., certain permitted food colours and saccharin)
or certain nutrients in excessive concentrations (e.g., modified
starches, plant gums, lactose, and various polyols).
Most of the enlargement is attributable to increased luminal
contents; in addition, the weight of the caecal wall after
washing out the luminal contents is usually marginally more than
normal. In haematoxylin- and eosin-stained sections, the caecal
wall shows no remarkable features, and there is no evidence that
caecal enlargement predisposes to any form of neoplasm of the
caecum. Caecal enlargement may be due to osmotic effects, but
its mechanism is not well understood. In some cases, caecal
enlargement is an incidental finding, with the primary effect
being nephrocalcinosis (23, pp. 11-12). Various forms of
mineral deposition occur in the kidneys of laboratory rodents,
more commonly in rats than in mice or hamsters. Unless
appropriate diagnostic staining or chemical analysis is carried
out, it is not strictly justifiable to refer to these changes as
renal "calcinosis", though most of the mineral deposits do, in
fact, contain calcium in one form or another. Mineral
deposition can take place in almost any part of the nephron and
deposition may predominate in one or more areas of the kidney.
The main forms of renal mineralization are basement membrane
mineralization, corticomedullary nephrocalcinosis, pelvic neph-
rocalcinosis, and nephrocalcinosis associated with acute tubular
nephrosis. All these forms of nephrocalcinosis may co-exist,
and one and the same agent may cause more than one form of neph-
rocalcinosis.
Magnesium deficiency in standard laboratory diets undoubt-
edly contributes to the high incidence of corticomedullary neph-
rocalcinosis in rats. Excessive dietary phosphate and possibly
excessive dietary calcium may predispose to pelvic nephrocal-
cinosis. Such observations lead to the conclusion that more
attention needs to be paid in the future to the formulation of
diets for rodents with respect to physiologically-relevant
levels of calcium, magnesium, and phosphorus.
Testicular atrophy, which is sometimes observed in rodents,
may occur as a result of reduced caloric intake, frequently as a
result of the addition of an unpalatable chemical to the
animal's food or drinking-water. This should be distinguished
from testicular atrophy resulting from the direct action of the
chemical on the testicular cells. This distinction can be
achieved by undertaking paired feeding studies. It is important
that the function of the testes be investigated in reproduction
studies when atrophy is detected.
Manifestations of vitamin deficiencies, notably of the fat-
soluble vitamins, are sometimes observed in studies on agents
that may be fat solvents and are only partly absorbed in the
gastrointestinal tract; an example of such a substance is
mineral oil. Another effect that is sometimes observed is dis-
colouration of mesenteric lymph nodes after feeding a coloured
substance. This is a normal physiological response and should
not be considered a toxic end-point, as long as it is not asso-
ciated with proliferative reactions.
Hormonally-associated effects occur with certain additives
and require special endocrinological evaluations. Recently,
JECFA has been faced with the task of evaluating the use of
certain anabolic agents used in the raising of meat-producing
animals (22, 23). These agents result in the presence of low-
level residues of hormonally-active compounds in meat. The
evaluation of the potential toxic effects of such compounds
requires knowledge of the levels of naturally-occurring com-
pounds with similar effects (Annex III).
5.1.3. Neoplasms
The most difficult decisions facing the toxicologist arise
from the varied end-points found in long-term in vivo carcino-
genesis studies. These problems have existed for a long time,
but they have been greatly exacerbated by the large number of
effects observed in the many rote tests now performed on
chemicals including many that may be present in the food supply.
These chemicals include certain direct food additives, some
processing and carrier solvents, components of packaging
materials, and a variety of contaminants.
Oncologists now generally recognize that different mech-
anisms of carcinogenesis exist with different chemicals acting
on different tissues in the body (45, 46). Many believe that it
may be possible to determine tolerance levels for some carcino-
gens, though this is still not possible with any degree of cer-
tainty with the majority of them. The view that a tolerance may
be set for carcinogens giving rise to tumours through either a
hormonal mechanism or by the formation of bladder calculi has
been expressed by a WHO scientific working group (3, p. 11).
The perception that a chemical for which there is evidence
of "carcinogenicity" in any system should not be permitted for
use as a food additive at any dose whatsoever has become wide-
spread. Although this philosophy has never been promulgated or
officially adopted by JECFA, it has, in practice, influenced the
Committee's approach to decision making. Probably more experi-
mental work has been undertaken in cancer research over the past
two decades, since this view was first established, than in all
the preceding years, and clearly there is a great need to
clarify the issue in terms of practical interpretation.
The assessment of the evidence for the carcinogenicity of
chemicals is a major issue to be resolved by JECFA. Not only
have many chemicals been tested by a variety of routes of
administration that may not be relevant to food additive use
(such as the repeated injection of food colours and other addi-
tives in rats and mice with consequent development of subcut-
aneous sarcomas at the site of injection (27)), but, in addi-
tion, new end-points are continually revealed and interpretation
becomes more confusing as studies become more and more detailed.
Positive results may be obtained due, for example, to a carcino-
genic impurity. The extrapolation of such data has become very
complex. One possibility, to make the term "carcinogen" more
generalized, clearly would not solve the problem of how best to
interpret these data.
Much of this issue centres around the meaning of the various
types of enhancement of the tumours that occur in the rodents
used for in vivo bioassays, since the rodents in common use, the
mouse and the rat, develop extremely high incidences of a
variety of tumours in the untreated state. Many reports indi-
cate that one or more of these tumours has an increased inci-
dence or has appeared earlier (or both) in treated, compared
with untreated animals. One problem in interpreting the signi-
ficance of these tumours is the difficulty in deciding whether
these naturally-occurring tumours are spontaneously induced or
whether the agent is able to induce them. The problem is fur-
ther confounded by the fact that the incidence of tumours in the
untreated control animals varies considerably with time. As a
result, there is now a debate as to the importance of historical
as well as concurrent control animals. It appears without doubt
that both such controls are of importance (especially when the
historical control data come from the same laboratory, using the
same standardized diet, and do not go back in history beyond 5
years of the study under consideration) and that, if the chem-
ical in question has only enhanced the incidence of a commonly-
seen tumour to a level seen in historical controls, then the
level of concern will be much less than would otherwise be the
case.
The evaluation of studies in which these commonly-occurring
tumours are a complicating factor need careful individual
assessment. The tumours that have given rise to the most
controversy in recent years are hepatomas (particularly in the
mouse), pheochromocytomas in the rat (see below), lymphomas and
lung adenomas in the mouse, pancreatic adenomas and gastric
papillomas in the rat, and certain endocrine-associated tumours,
including pituitary, mammary, and thyroid tumours, in both rats
and mice. Some of these tumours, such as hepatomas and lung
adenomas, may occur in the majority of untreated animals.
With the exception of lymphomas, some of which are virally
associated, the endocrine-associated tumours, and possibly
hepatomas in high-incidence strains of mice, which may involve
oncogenes (47), there is no clue as to the origin of tumours
that occur commonly in experimentally-used rodents. Indeed,
there is not even any cogent speculation as to the mechanisms by
which these tumour incidences are increased.
Adrenal medullary lesions in rats provide a good example of
the problems encountered in interpreting the significance of
high tumour incidences. An overview of the literature indicates
that untreated rats of various strains may exhibit widely dif-
fering incidences of lesions described as "pheochromocytomas"
(24, 48, 49). There are no clear criteria for distinguishing
between prominent foci of hyperplasia and benign neoplasms, and
pathologists differ in the criteria that they use for disting-
uishing between benign and malignant adrenal medullary tumours.
Rats fed ad libitum on highly nutritious diets tend to
develop a wide variety of neoplasms, particularly of the
endocrine glands, in much higher incidences than animals
provided with enough food to meet their nutritional needs but
not enough to render them obese. The adrenal medulla is just
one of the sites affected by overfeeding. Controlled feeding,
especially early in life, reduces the life-time expectation of
developing either hyperplasia or neoplasia of the adrenal
medulla in rats.
A complicating factor in assessing carcinogenicity studies
is the question of how to consider benign tumours. If benign
and malignant tumours are observed in an animal tissue and there
is evidence of progression from the benign to the malignant
state, then it is appropriate to combine the tumour types before
performing statistical analysis. Assessment of the relative
numbers of benign and malignant tumours at the various dose
levels in the study can help determine the dose response of the
animal to the compound under test. On the other hand, if only
benign tumours are observed and there is no indication that they
progress to malignancy, then, in most cases, it is not appro-
priate to consider the compound to be a frank carcinogen, under
the conditions of the test (this finding may suggest further
study). Often, how benign tumours should be considered is much
less clear. Some clarification can be achieved by classifying
and analysing tumours on the basis of their histogenic origin.
This is helpful, not only for determining the significance of
benign tumours, but also for preventing different malignant
tumours occurring in the same organ from being grouped together
for statistical analysis. For further discussion of these
points, see (50), pp. 226-230.
The results of statistical analyses are often misunderstood
and misused. An effect may be statistically significant but not
be of any biological significance, because the animal's well-
being is not affected by its occurrence. On the other hand, an
event that is of biological significance, such as the occurrence
of one or two tumours of a very rare type in treated animals,
may not be significant by the usual battery of statistical
tests. This difference between biological and statistical
significance underscores the need for critical analysis of
statistical results rather than the blind acceptance of the
numbers obtained. The statistical aspects of the design and
interpretation of toxicity studies are discussed in Annex II.
Generally, it is becoming increasingly difficult to identify
a substance as a carcinogen with confidence. In particular,
when animals with a high background incidence of tumours are
involved, it is extremely difficult to know when to draw the
line between a result that indicates that a compound that is
potentially hazardous for man has been discovered compared with
a compound that is merely an experimental curiosity. The
International Agency for Research on Cancer (IARC) reviews
evidence for the carcinogenicity of chemicals on a continuous
basis and drafts monographs on many groups of substances.
However, faced with the difficulty of separating different
levels of "carcinogenicity", IARC working groups on the
Evaluation of the Carcinogenic Risk of Chemicals to Humans
designate compounds as being possessed of either "limited" or
"sufficient" evidence for carcinogenicity. Given the natural
desire of the food additive toxicologist to be as cautious as
possible, this terminology is not very practical. After a
compound has been designated as possessing "limited carcino-
genicity", it is very difficult from a regulatory point of view
to approve it as a food additive, even if extensive further work
on the compound shows it to be safe at expected levels of con-
sumption with no other evidence of carcinogenicity.
The decisions that are being made on the basis of the
present state of knowledge of carcinogenesis, may, in the
future, prove to have been excessively conservative. However,
it is now possible to make certain reasoned decisions, provided
that each instance where carcinogenicity is the problem is
examined individually and all relevant factors are taken into
account.
5.1.4. Reproduction/developmental toxicity
Most food additives are consumed by men and women during the
reproductive stages of their lives and by pregnant and lactating
women. Some food additives are also consumed by infants. Thus,
a thorough safety evaluation requires that the effects of the
substance on reproductive performance and development from
fertilization through weaning be studied. JECFA recognizes,
however, that it is unrealistic to expect that such studies be
performed in all cases (sections 5.3.1 and 5.3.4).
Adverse effects on reproduction may be expressed through
reduced fertility or sterility in either the parents or off-
spring due to morphological, biochemical, or physiological dis-
turbances. Adverse effects on development may be expressed
through structural or functional abnormalities due to either
mutations or to biochemical or physiological disturbances.
Mutations may occur in either somatic or germ cells. Mutations
in male or female germ cells represent potentially the most
long-lasting and severe effects on the human population that a
chemical could cause.
Adverse effects on reproduction or development induced by
chemicals may be expressed immediately or they may be delayed,
sometimes for many years, as exemplified by transplacental
carcinogens (section 5.3.1.2).
Structural or functional abnormalities are most likely to
develop during embryogenesis, the period of development during
which cells differentiate into the various organ systems.
Typical teratogenicity studies investigate the effects of expo-
sure to test substances during this period. Effects due to
exposure during fetogenesis, the developmental period after the
organ systems have formed, generally involve growth retardation
and functional disorders, though the external genitalia and the
central nervous system are also susceptible to injury during
this period (51, 52). Such structural or functional abnorm-
alities often do not become obvious until some time after birth
and, in some cases, not until adulthood.
Neonatal development may be influenced by the consumption of
milk containing chemicals (or their metabolites) that were
consumed by the mother. Agents may also affect neonatal
development by influencing maternal behaviour, hormonal balance,
or nutrition. Direct neonatal exposure to xenobiotic compounds
also occurs, but is less common, since JECFA considers it pru-
dent that food intended for infants younger than 12 weeks of age
should not contain any additives (42).
Guidelines for reproductive toxicity investigations have
been developed by various legislative and international organi-
zations, including the US Food and Drug Administration (US FDA),
US Environmental Protection Agency (US EPA), the United Kingdom
Committee on Safety of Medicines (CSM), Committee on Toxicity
(COT), and Pesticides Safety Precautions Scheme (PSPS), the
Japan Ministry of Agriculture, Forestry and Fisheries (MAFF)
and Ministry of Health and Welfare (MHW), the World Health
Organization (WHO), the Organization for Economic Cooperation
and Development (OECD) (all listed in (51)) and the Inter-
national Programme on Chemical Safety (IPCS) (53). A review of
the methodology for assessing the effects of chemicals on repro-
ductive function has been published under the auspices of the
IPCS and the Scientific Committee on Problems of the Environment
(SCOPE) of the International Council of Scientific Unions (54).
The procedures described in these publications are designed to
assess the reproductive and developmental toxicity potential of
test compounds using lower mammals as model systems. These pro-
cedures generally involve combining various stages of the life
cycle in one test, as it is usually not practicable to examine
the effects of a chemical in each separate stage of the repro-
ductive cycle. An exception is the so-called "teratogenicity"
study, where exposure is limited to the period of organogenesis
(see below).
The goal of reproduction/developmental toxicity studies is
to assess whether the organism is more sensitive to the agent
under test during its reproductive and developmental stages than
during its adult phase. Therefore, the highest dose of food
chemical that is administered is generally the amount that would
be expected to cause slight maternal toxicity, and the lowest
dose is the amount that is not expected to cause an effect in
either the mother or the conceptus. If profound toxicity is
observed in the offspring at the high dose (the dose that causes
only slight maternal toxicity), then the conclusion would be
that the substance is more toxic for offspring than for adults.
This conclusion would be reinforced by the appearance of adverse
effects in the conceptus at the mid- and/or low-dose levels. On
the other hand, if the test substance injures reproduction or
development at levels comparable with levels that cause toxicity
in adults, then no special concern should be attached to the
results of the reproduction/developmental toxicity studies.
Single-generation and multi-generation reproduction studies
are useful for assessing potentially deleterious effects on
reproduction and development through parturition and lactation.
However, because of the long-term exposure inherent in these
studies, detoxifying enzymes may be induced in mothers before
embryogenesis takes place. Under these circumstances, the
observed toxicity would be understated. Studies in which
mothers are exposed to the test substance only during organo-
genesis, as in teratogenicity studies, reduce the possibility of
the mother adapting to the test compound.
The range of effects arising from maternal exposure to
chemicals during organogenesis includes:
(a) death and resorption of the embryo;
(b) teratogenic defects (malformations of a structural
nature);
(c) growth retardation or specific developmental delays;
and
(d) decreased postnatal functional capabilities (55).
Which of these effects will be expressed depends on the level
and gestational timing of the dosage of the food chemical, and
the duration of the period of treatment (56). Thus, a substance
given at one dose level may result in growth retardation, while,
at a higher level, it may result in death and resorption of the
embryo. Sometimes, the slope of the dose-response curve of, and
between, these effects is very steep, making the interpretation
of the studies very difficult. Because all of these outcomes
are unacceptable, the most important consideration when evalu-
ating these studies should not be which effect is observed, but
rather, at what dose level the adverse effect became evident.
This dosage information can then be used to set exposure limits.
Because teratogenic effects are only one part of the total
spectrum of the embryotoxic effects that should be investigated
in such studies, a better term for "teratogenic studies" might
be "embryotoxicity studies". In those rare situations when the
studies are performed during the period of fetal development,
the term "fetotoxicity studies" should be used.
The appropriate role of studies involving in utero and
neonatal exposure in the evaluation of food additives is
discussed in section 5.3.
5.1.5. In vitro studies
In recent years, a great deal of effort has gone into the
development of in vitro test systems. Generally, these systems
are segregated according to two kinds of functions: (a) to
reveal whether a particular kind of toxicity is produced by the
agent under study; or (b) to help elucidate the mechanism of
toxicity displayed by a chemical. The former tests are being
developed to serve both as predictors of toxicity (section
3.1.2) and as substitutes for complex, lengthy in vivo pro-
cedures. The latter are more directly focused than the former,
and their value has been clearly demonstrated as a means of
establishing the metabolic mechanisms at the organ, tissue, or
cellular level (section 5.2).
Much effort has been devoted to the development of in vitro
test systems based on isolated cells, tissues, and organs. Some
of these systems are reported to be related to specific toxic
end-points such as mutagenicity and carcinogenicity (e.g., DNA
damage and repair in mammalian cells, covalent binding to DNA,
cell transformation, mitotic recombination, and gene conversion
in yeast (57)) and to embryotoxicity (e.g., whole embryo
cultures, cultures of embryonic tissues, teratocarcinoma cells,
and embyronated eggs (58)).
The number, diversity, and use of these tests have increased
rapidly in the past decade and are likely to continue to
increase in the future. However, correlations among results of
various in vitro tests and reported correlations between the
results of batteries of short-term in vitro tests and in vivo
carcinogenesis bioassays (which have been the primary thrust of
these assays) are not high. Such short-term in vitro tests are
generally effective at measuring their intended genetic end-
point, i.e., mutagenicity in the particular system under study.
However, the relevance of mutagenic effects to food additive
toxicity has not been established, and the results of many
current in vitro test procedures do not relate to genetic
effects in mammalian reproductive tissues. Neither is it clear
how well these tests identify chemical carcinogens or how they
should be used in the absence of corroborative data on carcino-
genicity.
In a similar fashion, culture techniques designed to measure
prenatal toxicity are extremely useful for research purposes,
but, at their present stage of development, they are not very
suitable for screening (58). By excluding the maternal-
placental-fetal relationship, such dissected systems permit the
compound to reach the target directly (membrane systems that
provide biological barriers are missing) without permitting the
potential moderating or activating influences of the maternal
tissues.
Attention should be paid to scientific developments in in
vitro test systems. However, because of the many experimental
uncertainties and controversial issues surrounding the efficacy
of these tests as predictors of specific toxic end-points, it
would be inappropriate for JECFA to request that all food
additives brought before it should be subjected to such tests on
a systematic basis. On the other hand, data obtained with in
vitro systems sometimes help to clarify the mechanism of action
of chemicals observed in in vivo systems. Therefore, JECFA
should continue to determine the relevance of available in
vitro data on an ad hoc basis when assessing the safety of
specific compounds.
5.2. The Use of Metabolic and Pharmacokinetic Studies in Safety Assessment
Chemical toxicity results from reactions between the
ingested toxic chemical, or its metabolites, with constituents
of the body. Therefore, the complete safety evaluation of a
substance such as a food additive must consider its metabolism
and pharmacokinetics. Unfortunately, a great deal more has been
said and recommended in this area than has been done in
practice. The importance of metabolic and pharmacokinetic data
in the proper planning and interpretation of in vivo toxicity
testing of chemicals is obvious, but the fact is that such data
are either inadequate or unavailable to aid in the interpreta-
tion of the majority of long-term studies undertaken with chem-
icals, including food additives.
Detailed metabolic studies have gained added importance in
determining the extent of appropriate toxicological testing
since the advent of novel and modified foods. This is con-
sidered in detail in section 6.2. However, it is necessary to
repeat some general aspects of the subject here.
Biochemical studies play two separate roles in the safety
evaluation of chemicals. These are:
(a) to design animal studies by identifying the appropriate
species for, and helping to determine the appropriate
level of, testing; and
(b) to extrapolate experimental animal toxicity data to
human beings, by elucidating the mechanism of toxicity
of the chemical, thus facilitating the establishment of
a no-observed-effect level; a comparison of biochemical
data between experimental animals and man helps
determine the relevance of any toxicity observed in
animals.
The ingested chemical itself may exert a toxic effect, or a
metabolite(s) may be the toxic agent. Many polar, non-
lipophilic chemicals are rapidly metabolized and/or excreted,
while lipophilic compounds may be stored, excreted into the
bile, or metabolized into more polar, water-soluble compounds,
which are eliminated from the body, in the urine, more rapidly
than the ingested additive.
Absorbed substances, except those that enter the lymphatic
system, are transported directly to the liver via the portal
vein. Many substances that are metabolized in the liver are
transported via the hepatic vein to the kidneys to be excreted
in the urine. Through enterohepatic circulation, some sub-
stances that are conjugated in the liver are excreted with the
bile, reabsorbed, and then excreted once again in either the
bile or the urine.
Metabolism, primarily involving enzymatic reactions, may:
(a) convert the additive into a body constituent or a
source of energy;
(b) lead to the detoxification of the ingested chemical and
the excretion of its metabolites; or
(c) result in activation of the chemical into reactive
intermediates that then react most importantly with
glutathione, tissue proteins, RNA, or DNA.
Biotransformation reactions are catalysed by intra- and extra-
cellular enzymes and by enzymes of the microflora of the gastro-
intestinal tract. Knowledge of the rates of formation, reaction
with tissue components, and excretion of various metabolites is
essential for full understanding of the disposition and elimin-
ation of the chemical from the body and of the mechanism and
extent of its toxicity.
This section contains a general discussion of the role of
metabolism and pharmacokinetic data in the safety assessment of
food additives. Simple guidelines have not been generated, as
it is doubtful that such guidelines are feasible or desirable.
Several food additives that have been extensively studied bio-
chemically are discussed in Annex IV. These examples are
designed to provide an appreciation of the value and problems
involved with investigating the metabolic bases for the bio-
logical responses to food additives.
5.2.1. Identifying relevant animal species
The occurrence of interspecies differences in response to
foreign compounds complicates the extrapolation of animal
toxicity data to human beings. The resolution of this problem
depends on an understanding of such interspecies variations in
the disposition of ingested compounds. In this context,
disposition is meant to encompass metabolism and pharmaco-
kinetics.
The rates of absorption, rates and sites of distribution,
and rates and routes of excretion determine the concentration-
time profiles of the parent molecule and metabolites in the
various tissues and organs of the body. The overall biological
response is thus the product of the fluxes of the unchanged
molecule and its metabolites occurring in the animal under
examination. Definition of the pharmacokinetic properties of a
food additive may require various routes of administration. The
influence of any vehicle to be us