Threshold of Toxicological Concern (TTC) - International Life ...

ILSI EUROPE CONCISE MONOGRAPH SERIESTHRESHOLD OFTOXICOLOGICALCONCERN (TTC)A TOOL FOR ASSESSINGSUBSTANCES OF UNKNOWNTOXICITY PRESENT ATLOW LEVELS IN THE DIET

ABOUT ILSI / ILSI EUROPEThe International Life Sciences Institute (ILSI) is a nonprofit, worldwide foundation established in 1978 to advance the understanding ofscientific issues relating to nutrition, food safety, toxicology, risk assessment, and the environment. By bringing together scientists fromacademia, government, industry, and the public sector, ILSI seeks a balanced approach to solving problems of common concern for the wellbeingof the general public. ILSI is headquartered in Washington, DC, USA. Branches include Argentina, Brazil, Europe, India, Japan, Korea,Mexico, North Africa and Gulf Region, North America, North Andean, South Africa, South Andean, Southeast Asia Region, the focal point inChina, and the ILSI Health and Environmental Sciences Institute (HESI). ILSI is affiliated with the World Health Organization as a nongovernmentalorganisation (NGO) and has specialised consultative status with the Food and Agriculture Organization of the United Nations.ILSI Europe was established in 1986 to identify and evaluate scientific issues related to the above topics through symposia, workshops, expertgroups, and resulting publications. The aim is to advance the understanding and resolution of scientific issues in these areas. ILSI Europe isfunded primarily by its industry members.This publication is made possible by support from the ILSI Europe Threshold of Toxicological Concern Task Force, which is under the umbrellaof the Board of Directors of ILSI Europe. ILSI policy mandates that the ILSI and ILSI branch Boards of Directors must be composed of at least50% public sector scientists; the remaining directors represent ILSI’s member companies. Listed below are the ILSI Europe Board of Directorsand the ILSI Europe Threshold of Toxicological Concern Task Force members.ILSI Europe Board of Directors membersMrs. K. Duffin-Maxwell, Kraft Foods (DE)Prof. G. Eisenbrand, University of Kaiserslautern (DE)Prof. A. Flynn, University College Cork (IE)Prof. A. Grynberg, National Institute for Agricultural Research (FR)Dr. M.E. Knowles, Coca-Cola Europe, Eurasia and Middle East (BE)Dr. I. Knudsen, Danish Institute for Food and Veterinary Research (DK)Dr. M. Kovac, Food Research Institute (SK)Dr. G. Kozianowski, Südzucker (DE)Dr. D.J.G. Müller, Procter & Gamble European Service GmbH (DE)Prof. G. Pascal, INRA (FR)Dr. J. Schlatter, Swiss Federal Office of Public Health (CH)Prof. L. Serra Majem, University of Las Palmas de Gran Canaria (ES)Prof. C. Shortt, Yakult (UK)Dr. G. Thompson, Groupe Danone (FR)Prof. V. Tutelyan, National Nutrition Institute (RU)Prof. P. van Bladeren, Nestlé Research Center (CH)Prof. W.M.J. van Gelder, Royal Numico (NL)Mr. P.M. Verschuren, Unilever Health Institute (NL)Prof. em. P. Walter, University of Basel (CH)Dr. J. Wills, Masterfoods (UK)ILSI Europe Threshold of Toxicological Concern Task Force industry membersCoca-ColaDow EuropeDSM Nutritional ProductsGroupe DanoneNestléNutrinovaSüdzucker

THRESHOLD OF TOXICOLOGICALCONCERN (TTC)A TOOL FOR ASSESSING SUBSTANCESOF UNKNOWN TOXICITY PRESENTAT LOW LEVELS IN THE DIETby Susan BarlowILSI Europe

© 2005 International Life Sciences InstituteAll rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in anyform or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior writtenpermission of the copyright holder. The International Life Sciences Institute (ILSI) does not claim copyright on U.S.government information.Authorisation to photocopy items for internal or personal use is granted by ILSI for libraries and other usersregistered with the Copyright Clearance Center (CCC) Transactional Reporting Services, provided that $0.50 perpage per copy is paid directly to CCC, 222 Rosewood Drive, Danvers, MA 01923. Phone: (+1) 978 750 8400, Fax: (+1)978 750 4470.ILSI®, “A Global Partnership for a Safer, Healthier World.®”, and the International Life Sciences Institute (ILSI) logoimage of the microscope over the globe are registered trademarks of the International Life Sciences Institute. The useof trade names and commercial sources in this document is for purposes of identification only and does not implyendorsement by ILSI. In addition, the views expressed herein are those of the individual authors and/or theirorganisations, and do not necessarily reflect those of ILSI and ILSI Europe.ILSI PressOne Thomas Circle, NWNinth FloorWashington DC 20005-5802USATel.: (+1) 202 659 0074Fax: (+1) 202 659 3859E-mail: ilsipress@ilsi.orgWebsite: http://www.ilsi.orgILSI Europe a.i.s.b.l.Avenue E. Mounier 83, Box 6B-1200 BrusselsBelgiumTel.: (+32) 2 771 00 14Fax: (+32) 2 762 00 44E-mail: info@ilsieurope.beWebsite: http://europe.ilsi.orgPrinted in BelgiumISBN 1-57881-188-0

CONTENTSForeword ........................................................................................................................... 1Introduction ....................................................................................................................... 3Current approaches to toxicity testing and safety evaluation of chemicals ........................................... 5Deciding on the likely level of concern .................................................................................. 5Deciding on whether there are enough toxicity data ................................................................. 5Assessment of exposure .................................................................................................... 6Use of toxicity data to assess risks and safe levels of intake ........................................................ 6The Threshold of Toxicological Concern (TTC) concept: a generic approach ........................................ 9History and evolution of the TTC concept ............................................................................. 9Proposals for generic TTCs according to chemical structure ........................................................ 12Further validation and refinement of the TTC concept .............................................................. 14The ILSI decision tree ...................................................................................................... 16Issues and limitations ............................................................................................................ 21Allergenicity ................................................................................................................... 21Accumulation ................................................................................................................. 21Endocrine disruption ........................................................................................................ 21Uncertainties, limitations and strengths of the databases ............................................................ 22Dealing with mixtures ....................................................................................................... 24Application of the TTC approach to subpopulations ................................................................. 24Current applications of the TTC concept .................................................................................... 25FDAexperience .............................................................................................................. 25JECFAexperience ............................................................................................................ 25Use by other organisations ................................................................................................. 26Summary and conclusions ..................................................................................................... 27Glossary ............................................................................................................................ 28Further reading ................................................................................................................... 31Author: Susan Barlow (UK)Scientific Editor: Erik Dybing, Norwegian Institute of Public Health (N)Scientific Referees: Wolfgang Dekant, University of Würzburg (D),Philippe Verger, National Institute for Agricultural Research – INRA (F)Concise Monograph Series Editor: Ron Walker, University of Surrey (UK)

Threshold of toxicological concern (TTC) 1FOREWORDMan is exposed to thousands of chemicals whethernaturally occurring or man-made. The human diet, forexample, contains innumerable low molecular weight,organic compounds that could, at some level of intake,represent a risk to human health. Extensive toxicitystudies, utilising many animals, are necessary toevaluate the safety of chemicals applied in food or toestablish if contaminants to which humans are exposedmay cause harm.The Threshold of Toxicological Concern (TTC) asdescribed in this Monograph is a principle that refers tothe establishment of a generic human exposurethreshold value for (groups of) chemicals below whichthere would be no appreciable risk to human health.The concept proposes that such a value can be identifiedfor many chemicals, including those of unknowntoxicity when considering their chemical structures.Evidently the establishment of a more widely acceptedTTC would benefit consumers, industry and regulators.For example, there is an ongoing concern that humansare exposed to a diverse array of chemicals and there isa demand to evaluate large numbers of chemicals. At thesame time there exists a strong pressure to reduce ourreliance on animal experimentation and to relyincreasingly on in vitro and in silico data. Use of the TTCprinciple would eliminate the necessity of extensivetoxicity testing and safety evaluations when humanintakes of a chemical are below a certain level of concern,would focus limited resources of time, funding, animaluse and expertise on the testing and evaluation ofsubstances with greater potential to pose risks to humanhealth and would considerably contribute to a reductionin the use of animals.In addition, the principle may be applied to theassessment of chemicals in sectors of health riskassessment other than food and could moreover befurther developed for environmental risk assessment.For example, application of the TTC principle could alsobe extended to other categories of chemical use such ascosmetics and consumer products. In this case, of course,appropriate methodologies should be developed toallow for route to route extrapolation and to assesscombined multi-route exposure. In addition, the TTCprinciple can be used to indicate analytical data needs(as, for example, it is used in the USA for indirect foodadditives), or for setting priorities among chemicals forlevels of “inherent concern”.In addition, since the principle is based on safetyevaluations relating to daily intake throughout life, theapproach may further be used in the assessment ofimpurities present in compounds, for contaminants atlarge, and as a science-based approach to indicatepotentially acceptable concentrations of chemicalspresent in nature, which could be utilised in theapplication of the precautionary principle.An International Life Sciences Institute (ILSI) – Europeexpert group has examined this TTC principle for itsapplicability to food safety evaluation. This Monographdescribes the history and development of the principleand its application to chemicals in food that humans areexposed to at low levels.Robert KroesUtrecht University

Threshold of toxicological concern (TTC) 3INTRODUCTIONWhat is a threshold of toxicologicalconcern (TTC)?The threshold of toxicological concern (TTC) is aconcept that refers to the establishment of a humanexposure threshold value for all chemicals, below whichthere would be no appreciable risk to health. The storywhich follows describes how and why this concept wasdeveloped, the scientific basis for the human exposurethreshold values that have been derived, where the TTCprinciple is now being applied and its value to thescientific community and to society.The world of chemicalsAs humans, we are exposed to thousands of chemicalsubstances in our daily lives. Over 70,000 chemicals areused commercially and more than 100,000 naturallyoccurring chemicals have been identified. Exposure canoccur at work, from the air we breathe, from consumerproducts used in the home and garden, from the waterwe drink or use for bathing, showering or swimmingand from the food we eat.Exposure to chemicals in foodsSome of the chemical substances to which we areexposed come from our diet. The main components offoods themselves, such as fats, carbohydrates, proteins,vitamins and minerals, are all chemicals. These areusually not of concern, unless particular ones are takenin excess or in nutritionally inadequate amounts.In addition, processed foods may contain chemicaladditives to preserve, colour, emulsify, sweeten andflavour foods, or to perform some other functional rolein the food. Foods may also contain residues ofpesticides that are used on crops and traces of veterinarydrugs used in food-producing animals. Chemicals usedas processing aids, such as machinery lubricants orantibacterial substances in salad washing water, can alsoleave residues on foods. Chemicals present in foodpackaging materials and kitchen utensils can migrateinto foods during manufacture, transport, storage,heating or cooking of food. Foods may containcontaminants of natural origin, such as toxins fromfungi, or metals from natural minerals and soils, andman-made contaminants that find their way into thegeneral environment, such as persistent polychlorinatedbiphenyls (PCB) and dioxins. Lastly, undesirablechemicals may be generated during cooking or smokingof food, such as acrylamide in fried potatoes and coffee,and polycyclic aromatic hydrocarbons (PAH) fromsmoking or barbecuing of meat and fish.What do we know about thesechemicals?For some food chemicals, such as additives, pesticidesand veterinary drugs, we have a wealth of informationon their chemical and toxicological properties and onwhat levels of exposure are likely to be safe for humans.Similarly, for vitamins and minerals in food there isinformation and experience from human consumptionabout what levels are safe. The situation is different,however, for many other chemicals found in food, suchas chemicals migrating from food packaging, flavouringsubstances, processing aids, unexpected contaminantsand substances formed as reaction products orbreakdown products during processing, heating and

4 Concise Monograph Seriescooking. For many of these and for many of the nonfoodchemicals to which humans may be exposed, weoften have little or, in some cases, no information ontheir potential for toxicity. In addition, analyticalcapabilities for the detection and quantification ofchemicals in foods are continuously improving, suchthat minute traces of a huge array of chemicals can nowbe identified. Scientists, governments and industry aremaking concerted efforts to test chemicals to whichhumans are known to be exposed, according to agreedpriorities, but this takes time and considerableresources. It is clearly not feasible to test all knownchemicals and probably unnecessary to subject everychemical to extensive testing for toxic effects.How much is toxic?Exposure is often used as one of the aspects to be takeninto account when setting priorities for testing. This isbecause the likelihood of adverse or harmful effects isrelated to the magnitude, frequency and duration ofexposure to a chemical. In the laboratory, scientistsobserve that for most toxic effects, there is an exposuredose, or threshold, below which no adverse effects areseen. If a general threshold, or several thresholds, couldbe determined for the world of chemicals, below whichexposure did not raise safety concerns for humans, thenthis could be a useful tool, among others, in deciding onthe need for toxicity testing. This concept has becomeknown as the Threshold of Toxicological Concern (TTC).How might a TTC be used?The TTC concept could be particularly useful, forexample, when there is a new discovery of the presenceof a contaminant in food, for which there is notoxicological information. It could also be useful insetting priorities for testing among large functionallysimilar groups of chemicals to which exposure isgenerally very low, such as flavourings and substancesused in food contact materials.The use of such a tool would have benefits, not only forindustry and regulatory authorities, but also for consumers,because it would enable the world’s limited resources fortoxicity testing and safety evaluation to be focused onchemicals that may pose a real threat to human health. Byeliminating the need for unnecessary toxicity tests, it wouldalso reduce the number of animals used in laboratory testingwhich would be welcomed both by the scientists involvedand the general public.

Threshold of toxicological concern (TTC) 5CURRENT APPROACHES TOTOXICITY TESTING ANDSAFETY EVALUATION OFCHEMICALSDeciding on the likely level of concernThe present system for safety evaluation of chemicals islargely based on a case-by-case approach. Scientists firstassemble all the information they have on a chemical andmake a judgement on the likely level of concern. In theinitial stages, the information available may be limited toknowledge of the chemical’s structure, where it occursand what degree of human exposure can be anticipated.For some chemicals there may also be limited toxicityinformation but this will often be far from complete. Atthis stage, a decision has to be taken on whether furthertoxicity or exposure data need to be generated. It is at thispoint that the TTC concept could be useful (see later).Deciding on whether there are enoughtoxicity dataIdeally, for a full assessment of any human safety risks ofa chemical in food, results from a range of laboratorytoxicity tests are needed (see Box 1). These tests shouldreveal any adverse effects on the structure and functionof the cells, organs, tissues and fluids in the bodyresulting from short-term exposure or from long-term,daily exposure. Life stages covered in the tests shouldinclude not only adulthood including pregnancy, butalso infancy and the juvenile period. Testing wouldnormally include not only investigation of any effects onthe various organs and systems of the body, but alsomale and female fertility, reproduction, development ofthe embryo and foetus and postnatal growth anddevelopment. If during these tests effects are revealed onparticular systems in the body, such as the immunesystem or the nervous system, then additional testsfocusing on these aspects may be needed. Anyknowledge of the effects of the chemical on humans isBOX 1Type of laboratory toxicity testSub-chronic toxicityChronic toxicityCarcinogenicityGenotoxicityReproductive toxicityDevelopmental toxicityImmunotoxicityNeurotoxicityWhat it can revealAdverse effects on structure and function in any part of the body following repeated dailyexposure for up to 10 per cent of lifetimeAdverse effects on structure and function in any part of the body following repeated dailyexposure over a substantial part of the lifetimeCancerDamage to the inherited genetic material inside cells (DNA)Adverse effects on fertility and reproductionAdverse effects on the embryo and foetusAdverse effects on the structure and function of the immune system, or in reaction toimmune challengeAdverse effects on the structure and function of the nervous system and behaviour

6 Concise Monograph Seriesparticularly valuable, but for many chemicals, no suchinformation is available.(For more details about toxicity testing methods, see the ILSIEurope Concise Monograph on The Acceptable Daily Intake)If the data available cover all or most of the above typesof tests, then a comprehensive safety evaluation can beconducted. If a non-critical piece of information is notavailable, those conducting the safety evaluation canuse their scientific judgement to make allowances forthe missing data. If the missing data are considered tobe critical to the safety evaluation, then more tests mustbe conducted.Assessment of exposureAt an early stage in safety evaluation, considerationmust be given to whether there is enough informationon how much of the chemical concerned is present infood, which foods may contain it, how much of therelevant foods are consumed in the daily diet, whichsections of the population may be most exposed and atwhat level. Other, non-dietary routes of exposure to thechemical may also need to be taken into consideration.If the chemical concerned is not detectable in foods,further consideration must be given to whether theanalytical limit of detection is sufficiently sensitive topick up amounts that may still be of toxicologicalrelevance. If the available data are insufficient to enablea good estimate of average and high exposures to bemade, it may be necessary to undertake more chemicalanalyses of foods or to generate more information onfood consumption. The generation of such data can becostly. For that reason, exposure assessment should be astepwise procedure, in which each step contributes tothe reduction of uncertainty. The process can be stoppedat the point where the estimated exposure is below thelevel of toxicological concern.Use of toxicity data to assess risks andsafe levels of intakeEffect levels and no effect levelsToxicological studies in animals are usually conductedusing several doses covering a wide range of exposure.For assessment of food chemicals the preferred route ofadministration is oral. The results of each study willgenerally, but not always, show some adverse orharmful effects at higher doses and no effects at lowerdoses. If the substance is toxic, the study will identifythe dose (or doses) at which adverse effects areobserved, known as an Effect Level (EL). The nature andseverity of the effects observed will vary, depending onthe type of test, the species of animal and the durationof exposure. The study will also normally identify themaximum dose at which there are no observed effects,and this is called the No Observed Effect Level (NOEL).Thus, from a range of toxicity studies there may beseveral NOELs and the risk assessment will as a rulefocus on the most sensitive relevant study giving thelowest NOEL. Sometimes the term No ObservedAdverse Effect Level (NOAEL) is used instead of NOELto distinguish between an observed effect that isadverse and an effect that is not necessarily adverse. Inthis Monograph the term NOEL is used and should beinterpreted as synonymous with NOAEL.The results from toxicity studies can be used in twodifferent ways:1. To predict safe levels of exposure for humans.2. To predict potentially harmful levels of exposure andthe likely nature of the harmful effects.Setting an Acceptable Daily Intake (ADI)In the first case, the results from toxicity studies can beused to predict the highest amount of a chemical

Threshold of toxicological concern (TTC) 7ingested on a daily basis by humans that is likely to besafe. For food chemicals this is often expressed as theAcceptable Daily Intake (ADI) or Tolerable Daily Intake(TDI). The term ADI is generally used for substancesintentionally added to food, while TDI is generally usedfor substances appearing in food but not intentionallyadded. The ADI or TDI is defined as the amount of achemical, expressed on a body weight basis that can beingested daily over a lifetime without appreciable riskto health.An ADI or TDI for a chemical is generally calculated bydividing the lowest NOEL revealed by the toxicity testsby a factor, usually 100, known as a safety factor oruncertainty factor.ADI/TDI =NOELSafety/Uncertainty FactorThe incorporation of a safety or uncertainty factor givesan additional margin of reassurance to take account ofthe possibility that humans may be more sensitive thananimals and that among humans some may be moresensitive than others. Thus, although it is consideredthat the toxicity tests conducted in laboratory animalsare predictive of likely effects in humans, it is knownthat there can be variations between species and withinspecies, including humans, in the way a chemical isabsorbed, metabolised and excreted in the body(toxicokinetics) and variations in the way a chemicalacts on the cells, organs and tissues of the body(toxicodynamics). The NOEL is therefore divided by asafety or uncertainty factor to allow for thesepossibilities. Thus, an ADI or TDI errs on the safe side,producing a conservative estimate of the intake of afood chemical likely to be without risk for humans.Predicting adverse effectsIn the second case, the results of the toxicity studies canbe used to predict the nature of the adverse effects thatmay occur in humans (the hazard) and at what level(s)of exposure these adverse effects may occur. Most typesof adverse effect for any particular chemical only occurabove a particular dose, but the magnitude of that dosemay vary depending on species and duration ofexposure. As in the previous case of predicting intakesthat are likely to be safe, when using laboratory animaldata to predict a potentially harmful level of intake for afood chemical in humans, variability between andwithin species needs to be taken into account.(For more details of how risks from intakes exceeding the ADIcan be assessed, see the ILSI Europe Report on Significance ofExcursions of Intake above the Acceptable Daily Intake (ADI))Do all toxic effects have a threshold?For most toxic effects caused by a particular chemicalthere is an exposure threshold below which adverseeffects do not occur. At low exposures, the body canusually tolerate some disturbance to its normalbiochemical and physiological functions without anyovert signs or symptoms of illness. The body also hasinbuilt mechanisms for rapidly getting rid of chemicalsvia metabolism and excretion and for repairing damagedcells and tissues. However, there are some particulartypes of toxic effect which can be triggered by exposureto very low amounts of a chemical and which can resultin long-term illness or permanent (irreversible) damage.This relates to damage to the genetic inherited materialin cells (DNA and chromosomes) and cancers caused bydamage to the DNA. These are known as genotoxic andcarcinogenic effects.(For more details of derivation of the ADI and dealing withuncertainty, see the ILSI Europe Concise Monograph on TheAcceptable Daily Intake)

8 Concise Monograph SeriesGenotoxic and carcinogenic effectsGenotoxic effects may be detected by in vitro tests, suchas exposing bacteria to the chemical (e.g. the Ames test)or exposing isolated animal cells or human cells to thechemical. If genotoxic effects are detected in vitro, furthertests in live animals (in vivo) can then be conducted tosee if the harmful effects on chromosomes and DNAobserved in vitro could actually cause damage in thebody. Damage to DNA is an everyday event (e.g. fromcell division, exposure to the sun’s ultraviolet rays orinternal exposure to reactive oxygen species) so the bodyhas repair mechanisms to deal with it and every daymillions of repairs are successfully carried out. However,studies on genotoxic chemicals offer strong evidencethat damage to DNA can occur at very low doses,without an apparent threshold, and that the damageincreases steadily with increasing dose. Thus chemicalsare described either as “positive” (cause damage) or“negative” (do not cause damage) for genotoxicity. It isat present not possible to define no effect levels forpositive genotoxic chemicals. Unrepaired damage to acell’s chromosomes or DNA can have two detrimentalconsequences; it can cause its growth and division to goout of control (cancer) or, in the case of germ cells (ovaand sperm), it can cause mutations that can be passed onto the offspring. However, it is important to note thatbecause of repair mechanisms, damage to DNA does notnecessarily result in a mutation or cancer and thatongoing research may eventually allow thresholds forgenotoxic substances to be established.Carcinogenic effects are investigated by exposinganimals, usually rats and mice, from a young agethroughout and until the end of adult life, to daily dosesof a chemical and examining the number and type oftumours that develop. Even though a laboratory animalstudy on cancer may appear to show a dose at whichthere is no increase in tumours, if the cause of the cancercan be linked to a genotoxic mechanism of action, onpresent evidence it is prudent to assume that there is nothreshold for the toxic effect. Exposure to any amount,however small or transient, might have a harmful effectin the long-term. This assumption is made becauseanimal experiments cannot, for logistical reasons, utilisesufficiently large numbers of animals to detect smallincreases in cancers at very low doses and therebypreclude the possibility that they occur.Thus for chemicals that are shown to be genotoxic, orgenotoxic and carcinogenic, when given to animals, it isnot possible to set an ADI or TDI using the NOEL/safety factor approach. However, it should be noted thatcancer can also be caused by non-genotoxic mechanismsof action for which thresholds can be established. Forchemicals acting in this way it is possible to set an ADIor TDI.Predicting the risk of cancerFor carcinogenic chemicals with genotoxic mechanismsof action, different approaches can be used to assess therisk of cancer at exposures likely to be encountered byhumans. Ordinarily, this involves making estimates ofrisks at low or very low exposures. The approachestaken usually involve use of the dose-response curveobtained in an animal carcinogenicity test. This curverelates incidence of cancer to the various daily doses ofthe chemical given to the animals over their lifetime. Asdoses used in experiments are normally high, relative tolikely human exposures, an estimate of risk at lowexposures is made by extrapolating the dose-responsecurve downwards to a point below the range of thedoses used in the experiment. A variety of mathematicalmodels can be applied to the dose-response curve tomake such a low-dose risk estimate. The mathematicalmodels are generally considered to be highlyconservative, and so give estimates of risk which not

Threshold of toxicological concern (TTC) 9only err on the side of safety, but may considerablyoverestimate the likely risk to humans. They can be usedto produce an estimate either of the exposure associatedwith a particular level of risk, or the risk associated witha particular level of exposure. Risk managers can then begiven choices about what they would consider to be anacceptable or a “virtually safe dose” (e.g. a dose whichresults in a predicted incidence of cancer of 1 in a millionpersons exposed for a lifetime to a particular dose).Because of inherent limitations in animal carcinogenicityexperiments and in the mathematical models used, somerisk assessors and risk managers do not view the aboveapproach as an appropriate way to estimate risks forhumans. If that view is taken and a chemical is shown tobe genotoxic and to cause cancer in animals, riskmanagers may decide that human exposure to thatchemical should be as low as reasonably practicable(ALARP) or as low as reasonable achievable (ALARA).Risk management measures then have to be taken toreduce or eliminate human exposure. It is evident thatlimits set on this basis may imply different risks forsubstances of different potencies.Are genotoxic substances permitted in foods?In the case of food chemicals, substances are notauthorised for deliberate addition to foods (additives), orfor use on crops (pesticides) or in food-producinganimals (veterinary drugs), if they are shown to begenotoxic, or genotoxic and carcinogenic, when tested invivo. However, many other natural and man-madechemicals can appear in foods as contaminants, andsome of these are known to be genotoxic. The TTC mayalso be useful for assessing these types of substances (seelater, step 4 of the decision tree).THE THRESHOLD OFTOXICOLOGICAL CONCERN(TTC) CONCEPT:A GENERIC APPROACHHistory and evolution of the TTCconceptThe TTC concept has evolved from a lengthy history ofattempts by scientists over the years, in regulatoryauthorities and elsewhere, to develop generic approachesto the safety assessment of large groups of chemicals orof individual chemicals of unknown toxicity.The driving forces behind these efforts have been:• the continuing improvements in analytical capabilitieswhich allow more and more chemicals to beidentified in food at lower and lower concentrations,• the widely accepted premise that exposure to verylow amounts of chemicals is usually without harm,• the view that the time and attention devoted to aparticular chemical should be in proportion to therisk to health,• the limited toxicological resources worldwide, bothin capacity for toxicity testing and for evaluation,• the desire to minimise the use of animals,• and the ability to analyse large sets of existingtoxicity data to make predictions about the behaviourof other structurally-related chemicals.Frawley’s approachOne of the first efforts was in relation to food packagingmaterials and was published by Frawley in 1967.Starting from the premise that there must be some usesof food packaging materials that do not involve anyhazard to health of the consumer of food, he set aboutdefining a dose which he considered would be without

10 Concise Monograph SeriesBOX 2Frawley’s classification of 220 chemicalsDistribution of Number of Heavy metalsNOELs (mg/kg chemicals (220)* andin the diet) pesticides (88)

Threshold of toxicological concern (TTC) 11FIGURE 1Distribution of TD 50 s for chemical carcinogens and extrapolation to a 1 in a million risk0.5 ppb5 ppb500 ppmLog VSDs for709 carcinogensRelative Probability10 ppmLog TD 50 for709 carcinogensOverlap-12 -9 -6 -3 0 +3 +6 +9 +12VSD: Virtually safe doseLog (mg/kg/day)Reprinted from Food and Chemical Toxicology Vol 37. Cheeseman MA, Machuga EJ and Bailey AB; A tiered approach to thresholdof regulation, pp387-412, Copyright 1999, with permission from Elsevier.distribution of exposures calculated to represent anestimated lifetime risk of one in a million of developingcancer or “virtually safe dose” (VSD) (see Figure 1).Thus, the distribution of carcinogenic potencies could beused to derive an estimate of the dietary concentration ofmost carcinogens which would give rise to less than aone in a million lifetime risk of cancer, assuming that therisks in animals were representative of those in humans.That concentration was estimated to be 0.5 microgrammes/kgof diet. It is this figure which is used as thebasis of the Threshold of Regulation policy. From this, ahuman daily exposure level of 1.5 microgrammes/person was derived, by assuming that a personconsumes 1500 g of food and 1500 g of fluids daily andthat the chemical is distributed evenly throughout thetotal diet.Later the carcinogenic potency database was enlarged toover 700 chemicals (Gold and colleagues, 1995), but thisdid not alter the distribution of the calculated risks.Based on this analysis, should any untested chemical towhich the Threshold of Regulation policy is applied turnout to be a carcinogen, the consumer should still beprotected. Since toxic effects other than cancer usuallyoccur at much higher exposures, consumers wouldautomatically be protected from those effects too.It can be seen that the policy contains elements of bothscientific and risk management judgements. The

12 Concise Monograph SeriesThreshold of Regulation policy means that producers canapply for an exemption from regulation of any chemicaloriginating from food contact materials estimated to bepresent in the diet at levels not exceeding 0.5 microgrammes/kg.If the FDA is satisfied that the conditionsfor exemption are met, the chemical does not have toundergo toxicological testing nor the normal pre-marketsafety evaluation by the agency.Proposal for generic TTCs according tochemical structureAnalysis of chemical structuresMunro and colleagues in 1996 went on to develop theconcept of generic thresholds by analysing toxic, butnon-carcinogenic, effects of chemicals, according totheir chemical structure. The chemicals were dividedinto three structural classes, based on a “decision tree”developed earlier by Cramer and colleagues. The threeclasses are shown in Box 3.The toxicity databaseA reference database was built up using results fromoral toxicity tests in rats and rabbits on 613 chemicalswith a wide range of structures and uses. The testsincluded sub-chronic, chronic, reproductive anddevelopmental toxicity studies. From these, the mostconservative NOEL for each chemical was selected,based on the most sensitive species, sex and toxic effect.The 613 NOELs were then plotted in three groups,according to structural class (see Figure 2).Human exposure thresholdsFor each of the three distributions of NOELs, a valuecoinciding with the point on the distribution where 5%of the chemicals had lower NOELs and 95% had higherBOX 3Structural classes for chemicals within the TTCconceptClass IClass IIClass IIISubstances with simple chemical structures and forwhich efficient modes of metabolism exist,suggesting a low order of oral toxicity.Substances which possess structures that are lessinnocuous than class I substances, but do not containstructural features suggestive of toxicity like thosesubstances in class III.Substances with chemical structures that permit nostrong initial presumption of safety or may evensuggest significant toxicity or have reactive functionalgroups.NOELs was selected (i.e. the fifth percentile NOEL). Thelower fifth percentile NOELs were then divided by afactor of 100 to ensure substantial margins of safety. Thisyielded three values termed “human exposurethresholds”, one for each structural class of chemical,shown in Box 4. These human exposure thresholds arealso referred to as TTCs.According to this scheme, a threshold can be selected fora chemical of known structure but unknown toxicity: ifhuman exposure is below the relevant threshold ofconcern for that structural class, it can be assumed withreasonable confidence that the likelihood of any risk tohuman health is low. Later work increased the number ofchemicals in the database from 613 to 900 but this did notalter the cumulative distributions of NOELs, addingfurther reassurance about the validity of using the databaseto derive thresholds of toxicological concern (TTC).

Threshold of toxicological concern (TTC) 13FIGURE 2Cumulative percentage10090807060504030201000.01 0.1 1.0 10 100 1000 10000NOEL (mg/kg body weight/day)FitteddistributionClass IClass IIClass IIIReprinted from Food and Chemical Toxicology Vol 34. Munro IC, Ford RA, Kennepohl E and Sprenger JG; Correlation of astructural class with no-observed-effect levels: a proposal for establishing a threshold of concern, pp 829-867, Copyright 1996,with permission from Elsevier.BOX 4Generic TTCs: Derivation of human exposure thresholds from toxicity dataStructural class Fifth percentile NOEL Human exposure threshold(mg/kg bw/day)(mg/person/day)*I 3.0 1.8II 0.91 0.54III 0.15 0.09* The human exposure threshold was calculated by multiplying the fifth percentile NOEL by 60 (assuming an individual weighs60 kg) and dividing by a safety factor of 100.

14 Concise Monograph SeriesComparison with the Threshold of RegulationMunro and colleagues emphasised that the humanexposure thresholds are intended to apply only tostructurally defined chemicals for which there is noevidence of genotoxic carcinogenicity and no structuralalerts for genotoxicity. A structural alert is a feature of achemical structure, such as an epoxide group, which isknown to have a predisposition for damaging DNA.Comparing these human exposure thresholds, rangingfrom 90-1800 microgrammes/day, derived from data onnon-carcinogenic effects, with the figure of 1.5 microgrammes/dayfor the FDA’s Threshold of Regulation,based on carcinogenic effects, it can be seen that thethresholds for non-carcinogenic effects are higher by atleast an order of magnitude. This is in accordance withwhat would be expected from our knowledge of themechanisms of various toxic effects and the doses thatinduce them, i.e. it is biologically plausible that somecarcinogens induce tumours at lower exposures than theexposures needed to induce other toxic effects.Further validation and refinement ofthe TTC conceptA tiered approach to Threshold of RegulationFurther work by the FDA has provided support for theuse of thresholds higher than 1.5 microgrammes/day forless potent carcinogens. Cheeseman and colleagues usedthe expanded carcinogenic potency database of over 700chemicals, together with short-term toxicity data, resultsof genotoxicity testing and structural alerts, to identifypotent and non-potent subsets. This work confirmed thevalidity of 1.5 microgrammes/day as an appropriatethreshold for most carcinogens, but went on to proposethat a tiered threshold of regulation could be justified.Examination of the expanded database led them toconclude that a dietary threshold of 4-5 microgrammes/kg could be appropriate for substances without structuralalerts and even for substances with structural alerts if theywere negative in tests for genotoxicity. The two exceptionsto this were N-nitroso and benzidine-like compoundswhich are more potent carcinogens. If substances had nostructural alerts, were negative in tests for genotoxicityand had acute toxicity (LD 50 ) above 1000 mg/kg bw, adietary threshold of regulation of 10-15 microgrammes/kg could be possible. The tiered approach has not yet beenadopted by the FDA.Cheeseman and colleagues also re-examined theunderlying premise of the Threshold of Regulationpolicy that carcinogenic effects generally occur at lowerdietary concentrations than other toxic effects. Theyanalysed information from a database (the Registry ofToxic Effects of Chemical Substances – RTECS) on 3306substances for which there were oral reproductivetoxicity data and on 2542 substances for which therewere data from other repeat-dose toxicity tests. For eachchemical, they searched for the lowest dose at which atoxic effect was seen. They then divided the lowesteffect level for each substance by an uncertainty factorof 1000 to derive a range of “pseudo-acceptable dailyintakes” (PADIs). The most likely (median) value for thePADI was 8300-fold above the threshold value derivedfrom the carcinogenic potency database. These resultssupported the contention that a “virtually safe dose”based on carcinogenicity data would also protectagainst other toxic effects.Do human exposure thresholds cover all possibleeffects?One issue raised in scientific discussions of the TTCconcept proposed by Munro and colleagues waswhether potentially sensitive toxicological effects thatmight occur at low dose levels would be covered by thederived human exposure thresholds (see Box 4). Inparticular, concerns were raised with regard to whethereffects on the nervous system, immune system,endocrine system and development would be absent at

Threshold of toxicological concern (TTC) 15the human exposure threshold values. Although theoriginal database published by Munro and colleagues in1996 did include some studies measuring thesepotentially sensitive endpoints, they were insufficient innumber to provide a robust answer to the question ofpotential low-dose effects. An Expert Group wastherefore set up by ILSI Europe to examine this questionin more detail (Kroes and colleagues, 2000).Expanded databases were developed for the toxicologicalendpoints of neurotoxicity (82 substances,comprising 45 with subchronic and chronic neurotoxicitydata and 37 with acute neurotoxicity data), immunotoxicity(37 substances), developmental neurotoxicity(52 substances) and developmental toxicity (81 substances).They were analysed to see if these endpointswere more sensitive than those for structural Class IIIcompounds in the original database compiled by Munroand colleagues and to see whether the TTC of 1.5 microgrammes/person/dayderived from the carcinogenicpotency database adequately covered such endpoints.Once again the distributions for the NOELs were plotted.There was no difference in the cumulative distribution ofNOELs for any of the selected endpoints other thanneurotoxicity. The cumulative distribution of NOELs forneurotoxicity was not only lower than those of the otherselected endpoints, but it was also lower than that forstructural Class III compounds. None of the selected noncancerendpoints were more sensitive than cancer.Moreover, the TTC of 1.5 microgrammes/person/day,based on cancer endpoints, comfortably covered all theseeffects, including neurotoxicity, being 2-3 orders ofmagnitude lower than the neurotoxicity NOELs dividedby a safety factor of 100.The ILSI Europe Expert Group concluded that a TTC of1.5 microgrammes/person/day is conservative and thatchemicals present in the diet that are consumed at levelsbelow this threshold pose no appreciable risk. It furtherconcluded that for chemicals which do not possessstructural alerts for genotoxicity, further analysis mayindicate that a higher TTC may be appropriate.Exclusion of high potency carcinogensThe TTC of 1.5 microgrammes/person/day used in theThreshold of Regulation policy is designed to protectagainst the toxicity of most chemicals, including those ofunknown toxicity should they turn out to be carcinogens.Nevertheless, the FDA acknowledges that there may besome chemicals with a very high carcinogenic potencythat may be unsuitable for the Threshold of Regulationapproach. The ILSI Europe Expert Group set out toexplore the issue of exceptionally potent chemicals (Kroesand colleagues, 2004).The carcinogenic potency database used by Cheesemanand colleagues (see earlier) comprising 709 compoundswas further expanded to 730 compounds and analysed inorder to identify structural alerts that would give thehighest calculated risks if present at very lowconcentrations in the diet. This analysis identified 5groups of compounds having a significant fraction oftheir members that may still be of concern at an intake of0.15 microgrammes/person/day. This is 10-fold belowthe Threshold of Regulation figure. These 5 structuralgroups, shown in Box 5, were termed the “Cohort ofConcern”. Three of the groups are genotoxic (aflatoxinlike-,azoxy- and nitroso-compounds), while two are nongenotoxic(TCDD and steroids). The ILSI Europe ExpertGroup concluded that compounds with these structuralalerts for high potency require compound-specifictoxicity data and should be excluded from any TTCapproach. The peer review Workshop (see below)recommended using a TTC of 0.15 microgrammes/dayfor all other substances with structural alerts forgenotoxicity which were not part of the cohort of concern.

16 Concise Monograph SeriesBOX 5Cohort of ConcernHigh potency carcinogens identified by structuralalerts and not suitable for the TTC approachAflatoxin-like compoundsAzoxy-compoundsNitroso-compounds2,3,7,8-dibenzo-p-dioxin and its analogues (TCDD)SteroidsExclusion for reasons other than carcinogenicpotencyIn addition to excluding compounds with structuralalerts for high potency carcinogenicity, the ILSI EuropeExpert Group also made a number of other recommendationsfor exclusion of particular groups from theTTC approach. It recommended that polyhalogenated-dibenzodioxins, -dibenzofurans and -biphenyls, alongwith heavy metals, should be excluded on the groundsthat they are known to accumulate in the body (seelater). Other non-essential metals in elemental, ionic ororganic forms should also be excluded because theywere not included in the original database of Munroand colleagues. In addition, proteins were not includedin the original database and should also be excludedbecause of their potential for allergenicity (see later) andbecause some peptides have potent biological activities.NeurotoxicantsThe ILSI Europe Expert Group further explored whetherparticular neurotoxicants should be considered as aseparate class. Using the expanded database from theearlier ILSI Europe work (see above) and locating themost sensitive indicators of effects that they could find,they plotted the NOELs for the most potent neurotoxicants,the organophosphorus compounds (OPs),separately from the other neurotoxicants. They noted thatthe 5 th percentile NOEL for OPs was lower, by around anorder of magnitude, than the corresponding NOEL forother neurotoxicants. The other neurotoxicants resulted ina plot comparable to the Class III chemicals, as publishedby Munro and colleagues. By applying a safety factor of100 to the 5 th percentile NOEL for OPs they derived ahuman exposure threshold of 18 microgrammes/person/day. The ILSI Europe Expert Group therefore recommendedthat this figure be used for OPs rather than thevalue of 90 microgrammes/person/day used for othercompounds in structural Class III (see Box 4).The ILSI decision treeDevelopment of the decision treeFollowing the development of the TTC concept and itssubsequent refinements described above, the work ofthe ILSI Europe Expert Group culminated in theconstruction of a decision tree, based on a tieredapproach, to act as guidance on how and when the TTCprinciple could be applied as a preliminary step in foodsafety evaluation. The decision tree was finalisedfollowing a peer review Workshop held in March 2003,at which the science behind the various steps in thetiered approach was presented and critically discussed.The decision tree is shown in Box 6.

Threshold of toxicological concern (TTC) 17BOX 6Decision tree proposed by ILSI Europe to decide whether substances can be assessed by the TTC approach(From Kroes et al., Food and Chemical Toxicology 42, p76, 2004)1. Is the substance a non-essential metal or metal containing compound, or is it a polyhalogenated-dibenzodioxin,-dibenzofuran, or -biphenyl?NO2. Are there structural alerts that raiseconcern for potential genotoxicity?NO5. Does estimated intake exceedTTC of 1.5µg/day?YESSubstance would not beexpected to be a safety concern6. Is the compound an organophosphate?NONOYESYESRisk assessment requires compound-specifictoxicity data3. Is the chemical an aflatoxin-like-,azoxy-, or N-nitroso- compound?NO4. Does estimated intake exceed TTCof 0.15µg/day?NOYESYESYESNegligible risk (low probability of a life-timecancer risk greater than 1 in 10 6 – see text)8. Is the compound in Cramerstructural class III?NO YES10. Is the compound in Cramerstructural class II?NO12. Does estimated intakeexceed 1800µg/day?YES NO7. Does estimated intake exceedTTC of 18µg/day?9. Does estimated intake exceed90µg/day?YESYESNOSubstance would not be expected tobe a safety concernNOYESSubstance would not be expectedto be a safety concern11. Does estimated intake exceed540µg/day?NOYESRisk assessment requirescompound-specific toxicity dataRisk assessment requirescompound-specific toxicity data

18 Concise Monograph SeriesUse of the decision treeThe decision tree comprises a series of steps, each oneframed as a question, to which the answer, either ‘Yes’or ‘No’, will carry the user through to the next step. Thequestions relate to whether the chemical is suitable forassessing via the TTC concept (see exclusions describedearlier), the presence or absence of structural alerts forgenotoxicity, and, depending on the chemical’sstructure, how the level of exposure relates to therelevant human exposure threshold. For any chemicaltaken through the decision tree process, one of tworecommendations will be reached:either,the substance would not be expected to be a safetyconcern,or,risk assessment requires compound-specific toxicitydata.The decision tree is only applicable to chemicals ofknown structure and with low molecular mass asrepresented in the database. Accordingly, it is notapplicable, for example, to polymers. A good estimate ofintake or exposure (see later) is critical to the use of thetree, since this determines whether or not the TTC isexceeded. The steps in the tree are described below.Steps of the decision treeStep 1. This removes from consideration types ofsubstances and chemical structures that are notadequately represented in the carcinogenicity andtoxicity databases used to develop the TTC values.Step 2. If the substance is not removed at step 1, it canproceed to step 2. This identifies compounds that havethe potential for genotoxicity and could be possiblegenotoxic carcinogens.Step 3. If the answer at step 2 is YES – it does havestructural alerts for genotoxicity – then step 3 identifiesthose structures that are likely to be the most potentgenotoxic carcinogens, i.e. aflatoxin-like, azoxy- and N-nitroso-compounds. These require compound-specifictoxicity data and cannot be further assessed by the TTCapproach.Step 4. Substances evaluated at step 4 would all bepotential genotoxic carcinogens, but with the most potentstructures removed at steps 2 and 3. Step 4 asks if theestimated intake exceeds the TTC of 0.15 microgrammes/day (or 0.0025 microgrammes/kg bw/day). The rationalefor this TTC was described earlier. For any substancereaching step 4, with an intake at or below this TTC, theprobability that any risk of cancer exceeds 1 in a millionis considered to be very low. The inclusion of this step isnot designed to allow genotoxic substances to be addeddeliberately to food, but rather to determine whetherthere is a safety concern, should they be detected in food,say, as a contaminant.Step 5. If the answer at step 2 is NO – it does not havestructural alerts for genotoxicity – then step 5 asks if theestimated intake exceeds 1.5 microgrammes/day (or0.025 microgrammes/kg bw/day). This TTC is the oneused in the Threshold of Regulation, based on an analysisof carcinogenic compounds, including both genotoxicand non-genotoxic compounds. For any substancereaching step 5, with an intake at or below this TTC, theprobability that any risk of cancer exceeds 1 in a millionis considered to be very low. As the TTCs for other formsof toxicity are all higher than this value, other forms oftoxicity would not be of concern either at intakes at orbelow 1.5 microgrammes/day.Step 6. This step identifies organophosphates which havea lower TTC (see earlier) than that for structural Class IIIcompounds in general. This step is not intended toreplace the normal regulatory assessments and controlsfor organophosphates used as pesticides, but can be usedto determine if there is any safety concern should a nonapprovedor unregulated OP be detected in food, forexample, as a contaminant.

Threshold of toxicological concern (TTC) 19Step 7. If the substance is identified as an OP at step 6,step 7 asks if the estimated intake exceeds the TTC forOPs of 18 microgrammes/day (or 0.3 microgrammes/kgbw/day). If the answer is NO, the substance would notbe expected to be a safety concern. If the answer is YES,the substance requires compound-specific toxicity dataand cannot be further assessed by the TTC approach.Step 8. Having by this stage eliminated potentialgenotoxic carcinogens and organophosphates, step 8 asksif the chemical falls into Cramer structural class III (seeBox 3).Step 9. If the answer to step 8 is YES – the chemical is inCramer structural class III – step 9 asks if the estimatedintake exceeds the TTC for that class of 90 microgrammes/day(or 1.5 microgrammes/kg bw/day). If theanswer is NO, the substance would not be expected to bea safety concern. If the answer is YES, the substancerequires compound-specific toxicity data and cannot befurther assessed by the TTC approach.Step 10. If the substance is not in Cramer structural classIII, step 10 asks if the chemical falls into Cramerstructural class II (see Box 3).Step 11. If the answer to step 10 is YES – the chemical is inCramer structural class II – step 11 asks if the estimatedintake exceeds the TTC for that class of 540 microgrammes/day(or 9 microgrammes/kg bw/day). If theanswer is NO, the substance would not be expected to bea safety concern. If the answer is YES, the substancerequires compound-specific toxicity data and cannot befurther assessed by the TTC approach.Step 12. If the substance is not in Cramer structural classII, step 12 assumes that the chemical falls into Cramerstructural class I (see Box 3) and asks if the estimatedintake exceeds the TTC for that class of 1800 microgrammes/day(or 30 microgrammes/kg bw/day). If theanswer is NO, the substance would not be expected to bea safety concern. If the answer is YES, the substancerequires compound-specific toxicity data and cannot befurther assessed by the TTC approach.Potential applications of the TTC principleThe ILSI Europe Expert Group has recommended that theTTC principle can be used for substances that are presentin food in low concentrations, which lack toxicity data, butfor which exposure assessment can provide reliable intakeestimates. The decision tree provides a structuredapproach that allows the consistent application of the TTCprinciple in a risk assessment context.Its main applications are anticipated to be in thefollowing situations:• As a preliminary step in the safety assessment ofchemicals present at low concentrations in food.Substances expected to have generally lowconcentrations in food and for which toxicity dataare often lacking are flavourings, substancesmigrating from food contact materials, some naturalcontaminants, contaminants of environmental originand substances used at low concentrations in a verylimited number of food items which are consumedin very low quantities.• In the setting of priorities, depending on the level ofconcern, for more in depth risk assessment.Use of the decision tree will help identify thosesubstances for which exposure estimates exceed therelevant TTC and which may therefore requirefurther information for risk assessment.• In the setting of priorities, depending on the level ofconcern, for further toxicological testing.Substances for which exposure estimates show thatthey do not exceed the relevant TTC can beconsidered as low priority for further testing, whilesubstances for which exposure estimates exceed theTTC may require prioritising for further testing,

20 Concise Monograph Seriesdepending on their structure and the degree towhich they exceed the relevant TTC.• In setting priorities for analytical method development.Substances for which present analytical methods donot allow accurate measurement at concentrationsthat are relevant to their particular structural classTTC, may point to the need for more sensitiveanalytical methods.• In setting priorities for more refined intake data.Substances for which intake estimates are close tothe relevant TTC but contain some uncertainties,may require more refined estimates of intake.Exposure data needed for application of the TTCprincipleA critical aspect of the appropriate application of the TTCprinciple is the necessity for reliable exposure data. As theTTCs are expressed in terms of microgrammes per personper day, exposure estimates need to be similarlyexpressed or related to body weight. As use of the TTCapproach could mean that consumers are exposed via thediet to substances on which there is little or no toxicityinformation as long as exposures are below the relevantthreshold value, it is important to ensure that exposureestimates are as complete and as accurate as possible, orbuild in adequate conservatism to account for possibleunderestimates.It is necessary to consider not only exposure from food,but also other possible sources of exposure (air, water,consumer products, workplace). In foods, the substancemay be widely distributed across many items in the dietor present only in a restricted number and type of fooditems. Food intake data and analytical data on levels infoods, or information on uses or occurrence in foods,need to be sufficiently robust and comprehensive toenable reliable estimates of intake to be made. Analyticalmethods used to determine levels in foods need to besufficiently sensitive to detect low concentrations,relative to the human exposure thresholds, otherwise alarge number of ‘non-detect’ values might give amisleading picture of total exposure. Since particulargroups in the population may consume differentamounts of specific foods, food intake data may need tobe sufficiently detailed to enable these groups to beexamined separately, for example, by age, gender orethnicity. For infants and children in particular, becauseof their smaller size, food intake expressed on a bodyweight basis is generally higher than that for adults.Infants and children may also consume greater absoluteamounts of some types of foods (e.g. fruits) than adultsbecause of dietary preferences. They may also have a lessvaried diet than adults (e.g. a high consumption of infantformula or processed baby foods) which has considerableimplications for intake estimates.It is assumed that an adult person may consume 1.5 kg offood and 1.5 kg of beverages per day (for children see“Adjusting TTCs for body weight”, page 25). As anexample, for a substance that occurs uniformly in thewhole diet and is in Cramer structural class I, for whichthe TTC is 1800 microgrammes/day, the TTC would bereached if there was a concentration of 600 microgrammes/kgin the whole diet. If the substance was onlypresent in beverages, then a concentration of 1200 microgrammes/kgwould reach the TTC. If the substance waspresent in only one food item, consumed in daily amountsof no more than 100 g, then the TTC would be reached bya concentration of 18000 microgrammes/kg. The situationis more complex when a substance is present only in a fewfood items consumed by a limited number of consumers(e.g. candies consumed by children). In such cases, the percapita assessment should consider the number ofconsumers exposed as a proportion of the wholepopulation, in order to avoid underestimation ofindividual intake. Obtaining such data can be resourceintensiveand new methodologies, such as post-marketingestimation of the number of consumers could be helpful.

Threshold of toxicological concern (TTC) 21ISSUES AND LIMITATIONSAllergenicityAllergic reactions to food are common, sometimes lifethreateningand of public concern. Once an individual issensitised to a particular food or chemical, allergicreactions can occur from exposure to very low amounts.The ILSI Europe Expert Group therefore gaveconsideration to whether any threshold could beestablished for allergic reactions. It was concluded thatwhilst thresholds undoubtedly exist, they have not beenestablished so far, even for common allergens, and areknown to vary with each individual and within anindividual over time. ILSI Europe’s examination ofpotentially sensitive endpoints, described earlier,included immunotoxicity, but excluded allergicresponses, which are a special sub-category of immunereactions. Thus, although the TTC approach does takeaccount of substances causing immunotoxicity otherthan allergenicity, it cannot be used to assess theconcern for allergenicity.AccumulationAccumulation describes the process by which theamount of a substance in the body (the ‘body burden’)increases with repeated exposure. This occurs when theamount ingested exceeds the body’s capacity toeliminate it via metabolism and excretion in urine,faeces and expired air. If a substance is not readilymetabolised and is also very soluble in fat it willaccumulate if exposure is frequent. For such substancesthere may be considerable differences between speciesin rates of elimination from the body and the differencesmay be greater than the safety or uncertainty factoremployed in risk assessment to take account of speciesdifferences in metabolism and elimination. The TTCprinciple should not be applied to such substances.An example is, TCDD (2,3,7,8-tetrachloro-dibenzo-pdioxin,the chemical released in the Seveso disaster),which is eliminated much more rapidly by rodents thanby humans. TCDD belongs to a group of chemicalsknown as the polyhalogenated dibenzo-para-dioxins.These are closely related structurally to polyhalogenateddibenzofurans and polyhalogenated biphenyls. Even thelow TTC for Cramer structural class III compounds is notappropriate for substances like these which accumulatein the body. Furthermore, such chemicals were notincluded in the database of Munro and colleagues onwhich the TTC approach is based. Accordingly, thesechemicals are not appropriate for assessment using theTTC approach.Heavy metals, such as cadmium, can also accumulate inthe body, and they were not included in the database ofMunro and colleagues. Thus, the TTC approach shouldnot be used for the assessment of metals in elemental,ionic or organic forms. Moreover, for a number of heavymetals it would be unnecessary as there is a vasttoxicological literature on the effects of exposure tometals such as lead, cadmium and mercury.Other compounds present in the diet may also showmarked differences between species in their potential toaccumulate in the body (e.g. the naturally occurringfungal toxin, ochratoxin A). If this is known thenapplication of the TTC approach is not appropriate.Endocrine disruptionAn important current issue in toxicology is theidentification and risk assessment of substances that actto perturb the endocrine system which produces thenumerous hormones in the body. Chemicals thatdirectly or indirectly affect either the structure and/orthe function of the hormone producing glands or theparts of the brain that control them are known as

22 Concise Monograph Series‘endocrine disrupters’. Exposure during development,either before birth or after, is a particularly vulnerableperiod for endocrine disruption. The issue of whetherendocrine disrupters may be active at very lowexposures is an unresolved, ongoing debate amongscientists. In view of the uncertainties, it would bepremature to include low-dose, endocrine-mediatedeffects in the TTC approach. Moreover, it is likely thatfor any chemical already identified as a potentialendocrine disrupter, toxicological data will be availablewhich can be used to perform a more comprehensiverisk assessment.Uncertainties, limitations and strengthsof the databasesUncertaintiesIn any method of risk assessment, there are inherentuncertainties in toxicity, exposure and extrapolationaspects, which risk assessors need to identify and, ifpossible, quantify. The TTC approach is little different,having its own particular uncertainties, but in this case,any significant uncertainty in exposure estimates wouldpreclude use of the TTC approach. The uncertainties ofthe TTC approach relate mainly to:• the validity of assuming the likely toxicity of aknown chemical structure, based on toxicity informationfrom similar chemicals falling into one of threebroad structural groups;• the validity of the factor of 100 applied to the 5 thpercentile NOELs in the database to derive thenumerical values for the TTCs;• whether the database on chemical toxicity used toderive the various TTCs is sufficiently comprehensiveto be representative, both of chemical structuresand of toxic effects;• the validity of extrapolating to a ‘virtually safe dose’for genotoxic carcinogens, by applying a mathematicalmodel to laboratory animal data.How the TTC approach addresses the uncertainties• the validity of assuming the likely toxicity of a knownchemical structure, based on toxicity information fromsimilar chemicals falling into one of three broad structuralgroupsThe concept that toxic activity and potency bear arelationship to chemical structure has evolved over theyears and has been widely studied and broadlyconfirmed. Some of this work was mentioned earlier indescribing the origins of the TTC approach. Threeaspects of chemical structure are important - the easewith which particular structures are metabolised (andhence eliminated from the body), whether the structureoccurs naturally in the body or is a normal product ofintermediary metabolism, and the presence or absenceof particular chemical groupings within a structure thatare known to cause toxicity.These three elements were used by Cramer andcolleagues to devise their original decision tree in 1978,in which they proposed three main structural classes.The examination of a large number of these threestructural classes of chemicals in relation to their NOELsby Munro and colleagues confirmed the expectedrelative ranking of low, moderate and higher toxicity tostructural classes I, II and III, respectively. It is of courserecognised that due to the biological complexities ofliving organisms, including humans, such structureactivitypredictions may occasionally turn out to bewrong. It is for this reason that the NOELs used toderive the TTCs are divided by a factor of 100 to providean extra margin of safety, in case a particular chemical ofunknown toxicity does not behave as predicted.

Threshold of toxicological concern (TTC) 23In the case of structural alerts for those substances likelyto represent some of the most toxic at low exposures, i.e.those that are potentially genotoxic, this has been themost intensively studied aspect of structure-activityrelationships in toxicology and it is widely acceptedamong scientists that predictions based on these typesof alert are robust. Examination of structures for thesealerts is incorporated at an early stage in the TTCdecision tree.• the validity of the factor of 100 applied to the 5 th percentileNOELs in the database to derive the numerical values forthe TTCsAs explained above, the factor of 100 was selected toprovide an extra margin of safety over and above the 5 thpercentile NOEL for each structural class. This factorwas chosen because historically a factor of 100 has alsobeen used to derive Acceptable Daily Intakes forindividual compounds from their compound-specificNOELs (see earlier). Although the original selection of afactor of 100 some fifty years ago was based on scientificjudgement rather than detailed evidence, considerablesupport has emerged in recent years for the use of thisfigure from studies on human and animal metabolicdifferences and species differences in adverse responsesto chemicals including drugs (toxicokinetics andtoxicodynamics). Thus, it is now widely accepted thatuse of a factor of 100 when extrapolating from NOELsderived from animal studies to predicted safe intakesfor humans should generally provide a reasonablemargin of safety.• whether the database on chemical toxicity used to derivethe various TTCs is sufficiently comprehensive to be representative,both of chemical structures and of toxic effectsThis issue has been addressed in several publications ofMunro, Cheeseman and their colleagues, and by theILSI Europe Expert Group, in response to comments byother regulatory and research scientists that someendpoints of toxicity, that might be particularlysensitive, were insufficiently represented within theoriginal database. Accordingly, particular attention hasbeen paid to the endpoints of carcinogenicity,neurotoxicity including developmental neurotoxicity,other developmental toxicity (teratogenicity) andimmunotoxicity and evidence provided to show thatthey are adequately represented within the updateddatabase. For substances which are not represented atall or not well represented within the database, the ILSIEurope Expert Group has recommended that they arenot evaluated using the TTC approach. These includehigh potency carcinogens, metals and polyhalogenatedring-structured compounds. Similarly, substances withparticular toxicological features such as endocrinedisruption at low doses or potential allergenicity arealso excluded.• the validity of extrapolating to a ‘virtually safe dose’ forgenotoxic carcinogens, by applying a mathematical modelto laboratory animal data.This perhaps represents the most contentious area ofuncertainty in the TTC approach. As explained earlier,under “Predicting the risk of cancer”, not all scientistsagree that the application of mathematical modelling tothe results from laboratory animal studies on cancer, inorder to derive a virtually safe dose, gives an accurateprediction of likely risks for humans. But whatever theirviewpoint, they generally agree that the mathematicalmodels used are highly conservative and so are unlikelyto underestimate risks to humans. Thus, in utilising thisapproach to derive a TTC for carcinogens, together withadditional decision-tree steps to exclude high potencycarcinogens, the TTC approach is very conservative.

24 Concise Monograph SeriesDealing with mixturesIn principle, the TTC approach could be used fordealing with mixtures of substances which have similartoxic mechanisms of action at the biochemical level. Ifconsumers simultaneously ingest a food or foodscontaining potentially toxic substances that act in thesame way (e.g. carrots containing residues of more thanone organophosphate pesticide), it would be possible tosum their exposures/intakes and compare thecombined exposure/intake with the relevant TTC,provided they were of similar potency or were correctedto a similar potency. If the combined intake was belowthe TTC, this would indicate that the substances wouldnot be expected to be a safety concern.If mechanisms of action of the substances in the mixtureare known to be dissimilar, then the TTC approach canbe followed to assess each individual substance, one byone. Similarly, with a mixture of impurities, some ofknown structure and some unknown, if the level of theimpurity present in the highest concentration is belowthe human exposure threshold value for structural classIII (the class most suspect for toxicity), then it can beassumed that all other impurities, present at lowerconcentrations, would also be below that threshold.Application of the TTC approach tosubpopulationsPotentially vulnerable subpopulationsAn important issue to consider in deciding whether it isappropriate to apply the TTC approach is the nature ofthe subpopulation(s) predicted to be most at riskbecause of their exposure. Some subpopulations may beconsidered vulnerable, not only because of higherexposure, but because of potentially greater sensitivityto toxicity. Such groups might include:• the elderly due to a reduced capacity for metabolismand excretion of chemicals;• the very young with immature metabolising capacityfor some, but not all, chemicals;• pregnant women because of vulnerability of theembryo and foetus;• persons of any age who have a particular geneticmakeup (called a ‘genetic polymorphism’) thatimpairs or alters the way they handle and respond topotentially toxic substances.The database used to identify the NOELs for derivationof the numerical TTC values includes toxicity studies onageing animals, pregnant animals, newborn, very youngand juvenile animals. Thus, most of the scenariosidentified above are represented in the database and thederived TTCs should cover these subpopulations. Inaddition, the use of a factor of 100 to derive a TTC from aNOEL takes into account potential metabolic differencesbetween laboratory animals and humans and differencesbetween individual humans.Exceptional subpopulations which might not be coveredare those with certain genetic polymorphisms that haveprofound effects on metabolic capacity and metabolicpathways. Present knowledge on the nature andprevalence of these polymorphisms in different ethnicgroups is far from complete, but already a few are knownwhich would result in certain substances being handledin the body in ways which would considerably erode the100-fold margin of safety built into the TTC values. Atpresent, it is not possible in most risk assessmentsituations to identify these potentially vulnerable people.However, this uncertainty applies equally to other,conventional risk assessment approaches.

Threshold of toxicological concern (TTC) 25Adjusting TTCs for body weightThe numerical values for the various TTCs calculated bythe ILSI Europe Expert Group are based on a 60kg adult(see Box 4). Where a substance occurs in foods consumedby infants and/or children, users of the TTC approachmay wish to consider whether intake estimates should becalculated separately for these groups and compared tothe relevant TTC, adjusted for bodyweight. For example,for a substance in Cramer class I, the TTC for a 10 kg,12-month-old infant would be 300 microgrammes/dayinstead of 1800 microgrammes/day (i.e. 1800 x 10/60),after adjustment for body weight.CURRENT APPLICATIONS OFTHE TTC CONCEPTFDA experienceSince the implementation of the Threshold ofRegulation in the USA in 1995, applying to foodpackaging migrants present in the diet at levels below0.5 ppb, the FDA has dealt with 183 applications underthis regulation and issued 78 exemptions using thisconcept. Applications are considered under anabbreviated review process in order to ascertainwhether the dietary concentration that would resultfrom the intended use is at or below the threshold leveland that there is no reason to suspect, based on test dataor chemical structure, that the substance may be acarcinogen. Although the Threshold of Regulation isdesigned to protect against all types of toxicityincluding carcinogenicity, under US law the FDA is notallowed to regulate known carcinogens as foodadditives (in the USA food contact materials areregulated as indirect food additives). The main reasonfor rejection of an application has been the submissionof inadequate exposure data. The FDA has commentedthat the Threshold of Regulation has been extremelyuseful because it is based on sound science and can beapplied rationally, consistently and effectively, case bycase. It is estimated to have reduced the workload of theagency by around 15%.JECFA experienceThe Joint FAO/WHO Expert Committee on FoodAdditives, known as JECFA, first considered using a newprocedure for the safety evaluation of flavouring agentsin 1995. JECFA was tasked with the evaluation of over2500 flavouring substances in current use. For many

26 Concise Monograph Seriesindividual flavouring substances, no toxicity ormetabolic data existed. The Committee agreed that inview of the large number of substances requiring safetyevaluation and the fact that, for the majority of flavouringagents, human intakes are relatively low and selflimiting,a different approach from that normally usedfor food additives should be followed.The proposed procedure put forward by Munro andcolleagues (1999) was based on the TTC concept, i.e. thethree Cramer structural classes and their respectiveTTCs. It involved integration of per capita intake data inrelation to the human exposure thresholds, withinformation on structure-activity relationships, metabolismand toxicity. As many flavouring substances areclosely related structurally, this procedure wasconsidered promising, since it would allow flavouringsto be evaluated in chemical groups, not only applying theTTC principle, but also incorporating, where available,any metabolic and toxicity information on any of theflavourings in a group. After applying the proposedprocedure to the evaluation of 3 groups of flavouringsubstances in 1996, JECFA adopted the new procedurefor safety evaluation (WHO, 1997). Since then JECFA hasevaluated in excess of 1400 flavourings using thisscheme. The JECFA recognised the limitation of percapita intake estimates, especially for estimatingexposure of individual groups of consumers consumingparticular foods, and further methodological developmentsof this aspect are under discussion in JECFA(WHO, 2001).Use by other organisationsThe TTC principle is also used by the EuropeanMedicines Agency (EMEA) to assess genotoxic impuritiesin pharmaceutical preparations ( 1 ). It was moreover usedby the former EC Scientific Committee on Food and isnow used by the European Food Safety Authority toevaluate flavouring substances ( 2 ). The TTC principle hasfurthermore been endorsed by the WHO InternationalProgram on Chemical Safety for the risk assessment ofchemicals ( 3 ) and also by the EU Scientific Committee onToxicology, Ecotoxicology and the Environment ( 4 ).(1) The European Medicines Agency. Committee for Medicinal Productsfor Human Use (CHMP). Guideline on the limits of genotoxicimpurities. CPMP/SWP/5199/02. London, 23 June 2004. Available at:www.emea.eu.int/pdfs/human/swp/519902en.pdf(2) European Food Safety Authority. Scientific Panel on Food Additives,Flavourings, Processing Aids and Materials in Contact with Food.Opinion on Flavouring Group FGE.03 Acetals of branched- andstraight-chain aliphatic saturated primary alcohols and branched- andstraight-chain saturated aldehydes, and an orthoester of formic acid,from chemical groups 1 and 2. Opinion expressed on 7 October 2004.Available at: www.efsa.eu.int/science/afc/catindex_en.html(3) International Program on Chemical Safety, World HealthOrganization. WHO Food Additive Series 35, WHO, Geneva,Switzerland.(4) Bridges, J. Strategy for a future chemicals policy. The view of theScientific Committee on Toxicology, Ecotoxicology and theEnvironment (CSTEE). Available at:www.eutop.de/chp/Download/BridgesRe.doc

Threshold of toxicological concern (TTC) 27SUMMARY ANDCONCLUSIONSThe world of chemicals to which humans may beexposed is very large. Some of these chemicals are closelyregulated on the basis of safety evaluations performed byscientists, using extensive information on toxicity. This isthe case for chemicals added deliberately to food orknown to be present in food, such as food additives,pesticides and veterinary drugs. However, for manyother substances also present in the diet, such ascontaminants (natural or man-made), flavouringsubstances and chemicals arising from food processing,including cooking, there may be little or no toxicityinformation. The threshold of toxicological concernconcept has been developed as a tool to enable apreliminary assessment of the likely risk from exposureto a known amount of a substance of known chemicalstructure, but of unknown toxicity.The threshold of toxicological concern (TTC) refers to theestablishment of human exposure threshold values forchemicals, below which there would be no appreciablerisk to health. Extensive exploration over several years ofexisting data on the relationship between chemicalstructure and toxic effects has enabled three broad classesof chemical structure to be defined and a numerical valuefor a TTC for each class to be established.The ILSI Europe Expert Group on the TTC principle hasdeveloped a decision tree which provides a structuredapproach that allows the consistent application of theTTC principle for risk assessment of substances ofunknown toxicity present in the diet for which there arereliable exposure estimates. In the application of the TTCconcept to safety evaluation of a chemical present in thediet (and possibly elsewhere), the intake or exposure tothe chemical is compared to the relevant TTC for itsstructural class. If the intake or exposure is below therelevant TTC, this indicates that there is unlikely to beany safety concern. If the intake or exposure exceeds therelevant TTC, this indicates that further information,including chemical-specific toxicity data, may be neededto perform a risk assessment. Thus, the TTC approachoffers a tool for risk assessors and risk managers toprioritise chemicals in need of further evaluation oradditional safety data.The ILSI Europe Expert Group has also identified certaintypes of substances for which the TTC approach shouldnot be used. These include heavy metals, substances thataccumulate, such as dioxins, allergens and endocrinedisrupters with low-dose effects.Useful applications of the TTC approach are envisaged toinclude situations when there is a new discovery of thepresence of a contaminant in food, for which there is notoxicological information, and in setting priorities fortesting among large functional groups of chemicals towhich exposure is generally very low, such as flavouringsubstances and substances used in food contact materials.The TTC concept is already being applied by organisationssuch as the US Food and Drug Administration inthe regulation of food contact materials and by the JointFAO/WHO Expert Committee on Food Additives(JECFA) in evaluations of flavouring substances.The wider use of such a tool would have benefits forindustry, regulatory authorities and consumers because itenables the world’s limited resources for toxicity testingand safety evaluation to be focused on exposures tochemicals which may pose a real threat to human health.By eliminating the need for unnecessary toxicity tests, itwould also reduce the number of animals used inlaboratory testing, and this would be welcomed by boththe scientists involved and the general public.

28 Concise Monograph SeriesGLOSSARYAcceptable Daily Intake (ADI): Estimate of the amountof a substance in food or drinking water, expressedon a body mass basis (usually mg/kg body weight),which can be ingested daily over a lifetime byhumans without appreciable health risk.Acute toxicity: Adverse effects occurring within a shorttime (usually up to 14 days) after administration of asingle dose of test substance, or after multiple dosesadministered within 24 hours.Adverse effect: Change in morphology, physiology,growth, development or lifespan of an organismwhich results in impairment of functional capacity orimpairment of capacity to compensate for additionalstress or increased susceptibility to the harmfuleffects of other environmental influences.Allergen: A substance which provokes an allergicresponse.Allergy: An inappropriate and exaggerated immuneresponse.Carcinogen: A substance capable of inducing cancer.Carcinogenesis: The complex, multistep process ofcancer causation.Chromosome: In the cell, DNA is tightly packagedtogether with particular proteins into structurescalled chromosomes. Packaging into chromosomesenables the organised assortment of genes intodaughter cells upon cell division, as well as playing arole in controlling gene expression.Chronic toxicity: Adverse effects following continuedexposures over an extended period of time (morethan 10 per cent of the lifespan).Decision tree: A structured approach for making stepby-stepdecisions about individual chemicals.Developmental toxicity: Adverse effects on the embryoand/or foetus following exposure during the prenatalperiod.Dioxins: A group of environmentally persistentsubstances with structures containing threeconnected rings made up of carbon, oxygen andeither chlorine, or chlorine and hydrogen. Dioxinsmay interact with the Ah receptor in the body toproduce cancer, reproductive toxicity and immunesystem effects.DNA (Deoxyribonucleic acid): A long molecule madeup of repeating units (each unit contains deoxyribose,a sugar, a phosphoric acid and one of four bases)joined together in a particular order. Each DNAmolecule consists of two strands in the shape of adouble helix. Genes are made of DNA, and areresponsible for the transfer of genetic informationfrom one cell/generation to the next.Effect level: The concentration or amount of an agent,found by study or observation, that causes alterationof morphology, functional capacity, growth,development or life-span of the target.Endocrine disrupter: A substance or mixture that altersthe function(s) of the endocrine system andconsequently causes adverse health effects in anintact organism, or its progeny, or (sub)populations.Endocrine system: Organs and tissues in the bodywhich produce hormones.Exposure: Concentration or amount of a particularchemical agent that reaches the target population,organism, organ, tissues or cell, usually expressed innumerical terms of substance concentration, durationand frequency.

Threshold of toxicological concern (TTC) 29Genotoxicity: Ability to cause damage to geneticmaterial. Such damage may lead to mutations andcancer.Hormone: A chemical substance produced in one partor organ of the body that initiates or regulates theactivity of an organ or group of cells in another partof the body.Human exposure threshold (of toxicological concern):A generic value for human exposure to a chemicalfalling within a particular structural class, belowwhich there would be no appreciable risk to health.Immunotoxicity: Adverse effects on the structure andfunction of the immune system, or in reaction toimmune challenge.In silico: Data generated and analysed using modellingand information technology approaches.In vitro: Literally “in glass”, referring to a study in thelaboratory usually involving isolated organ, tissue,cells or cellular fractions.In vivo: In the living body, referring to a studyperformed on a living organism.Long-term toxicity study: A study in which animals areobserved during the whole life span (or the majorpart of the life-span) and in which exposure to thetest material takes place over the whole observationtime or a substantial part thereof. The term chronictoxicity study is used sometimes as a synonym for“long-term toxicity study”.Margin of safety: The ratio of the no-observed-adverseeffectlevel (NOAEL) for the critical effect to thetheoretical, predicted or estimated exposure dose orconcentration.Mutation: The change in the DNA sequence caused bydamage by a mutagen, or by errors in cellularprocesses that may occur during cell division. Somemutations have no effect on the function of the genesin which they occur, while others inactivate or changethe activity of the genes. Some mutations aredetrimental to the organism, a few are beneficial.Mutations are a source of variation betweenindividuals, and are a driving force of evolution.Neurotoxicity: Adverse effects on the structure andfunction of the nervous system and behaviour.No observed adverse effect level (NOAEL): Thegreatest concentration or amount of an agent, foundby study or observation, that causes no detectableadverse alteration of morphology, functionalcapacity, growth, development or life-span of thetarget.No observed effect level (NOEL): The greatestconcentration or amount of an agent, found by studyor observation, that causes no detectable alteration ofmorphology, functional capacity, growth, developmentor life-span of the target.Polymorphism: A single gene trait that is caused by thepresence in the population of pairs of differing butrelated genes, resulting in more than one phenotypewithin the population, the less common geneoccurring in more than 1% of individuals.Potency: The extent, relative to dose, to which achemical is active with respect to a specific particulartoxic endpoint.Pseudo-acceptable daily intake (PADI): An intake for asubstance derived by applying a 1000-folduncertainty factor to the lowest low-effect level fornon-carcinogenic endpoints.Reproductive toxicity: Adverse effects on fertility andreproduction.

30 Concise Monograph SeriesRisk: The probability of an adverse effect in anorganism, system or (sub)population caused underspecified circumstances by exposure to an agent.Risk assessment: A process intended to calculate orestimate the risk to a given target organism, systemor (sub)population, including the identification ofattendant uncertainties, following exposure to aparticular agent, taking into account the inherentcharacteristics of the agent of concern as well as thecharacteristics of the specific target system.Safety: Practical certainty that adverse effects will notresult from exposure to an agent under definedcircumstances. It is the reciprocal of risk.Safety factor: A factor applied to the no-observedadverse-effect-levelto derive an ADI. The value ofthe safety factor depends on the size and type ofpopulation to be protected and the quality of thetoxicological information available.Short-term toxicity study: An animal study (sometimescalled a subacute or subchronic study) in which theeffects produced by the test material, whenadministered in repeated doses (or continuously infood or drinking water) over a period of about 90days (less than 10 per cent of the lifespan), arestudied.Structural alert: A particular chemical grouping withina chemical structure which is known to be associatedwith a particular type of toxic effect, e.g. genotoxicity.Threshold: Dose or exposure concentration of an agentbelow which a stated effect is not observed orexpected to occur.Threshold of Regulation: A policy of the USGovernment allowing regulation of food contactmaterials present only at very low levels in the dietby an abbreviated procedure.Threshold of Toxicological Concern (TTC) concept: Aconcept that proposes human exposure thresholdvalues for groups of chemicals, below which therewould be no appreciable risk to health.Tolerable Daily Intake (TDI): Regulatory valueequivalent to the Acceptable Daily Intake, used forfood contaminants, i.e. an estimate of the amount of asubstance in food or drinking water, expressed on abody mass basis (usually mg/kg body weight),which can be ingested daily over a lifetime byhumans without appreciable health risk.Toxicity: Inherent property of an agent to cause anadverse biological effect.Toxicodynamics: Description of the interaction betweena toxic agent and the target tissue on which it has anadverse effect.Toxicokinetics: Description of the absorption,distribution, metabolism and excretion of a chemicalin the body.Uncertainty factor: An alternative description of safetyfactor, which is being used increasingly because itindicates that the factor is to allow for uncertainties inthe risk assessment process.Virtually safe dose (VSD): A human exposure over alifetime to a carcinogen which has been estimated,using mathematical modelling, to result in a very lowincidence of cancer, somewhere between zero and thespecified incidence, e.g. 1 cancer in a million people.

Threshold of toxicological concern (TTC) 31FURTHER READINGMore details on the principles and methodologydescribed in this monograph may be found in:Cheeseman, M.A., Machuga, E.J. and Bailey, A.B. (1999).A tiered approach to threshold of regulation. Food andChemical Toxicology, 37, 387-412.Cramer, G.M., Ford, R.A. and Hall, R.L. (1978).Estimation of toxic hazard - a decision tree approach.Food and Cosmetic Toxicology, 16, 255-276.Frawley, J.P. (1967). Scientific evidence and commonsense as a basis for food packaging regulations. Food andCosmetics Toxicology, 5, 293-308.Gold, L.S., Sawyer, C.B., Magaw, R., Backman, G.M., deVeciana, M., Levinson, R., Hooper, N.K., Havender,W.R., Bernstein, L., Peto, R., Pike, M. and Ames, B.N.(1984). A carcinogenesis potency database of thestandardized results of animal bioassays. EnvironmentalHealth Perspectives, 58, 9-319.Gold, L.S., Manley, N.B., Slone, T.H., Garfinkel, G.B.,Ames, B.N., Rohrbach, L., Stern, B.R. and Chow, K.(1995). Sixth plot of the carcinogenic potency database:Results of animal bioassays published in the generalliterature 1989-1990 and by the National ToxicologyProgram through 1990-1993. Environmental HealthPerspectives, 103, (Suppl.8), 3-122.ILSI Europe (2000). The Acceptable Daily Intake: A Toolfor Ensuring Food Safety. By Diane Benford. ILSIEurope Concise Monograph Series. Brussels, Belgium.ISBN 1-57881-091-4.ILSI Europe (1999). Significance of Excursions of Intakeabove the Acceptable Daily Intake (ADI). ILSI EuropeReport Series. Brussels, Belgium. ISBN 1-57881-053-1.Kroes, R., Galli, C., Munro, I., Schilter, B., Tran, L.-A.,Walker, R. and Wurtzen, G. (2000). Threshold oftoxicological concern for chemical substances present inthe diet: a practical tool for assessing the need fortoxicity testing. Food and Chemical Toxicology, 38, 255-312.Kroes, R., Renwick, A.G., Cheeseman, M., Kleiner, J.,Mangelsdorf, I., Piersma, A., Schilter, B., Schlatter, J., vanSchothorst, F., Vos, J.G. and Wurtzen, G. (2004).Structure-based Thresholds of Toxicological Concern(TTC): Guidance for application to substances present atlow levels in the diet. Food and Chemical Toxicology, 42,65-83.Munro, I.C., Ford, R.A., Kennepohl, E. and Sprenger,J.G. (1996). Correlation of a structural class with noobserved-effectlevels: a proposal for establishing athreshold of concern. Food and Chemical Toxicology, 34,829-867.Munro, I.C., Kennepohl, E. and Kroes, R. (1999). Aprocedure for the safety evaluation of flavouringsubstances. Food and Chemical Toxicology, 37, 207-232.US Food and Drug Administration (1995). FoodAdditives: Threshold of Regulation for Substances Usedin Food-Contact Articles; Final Rule. Federal Register60, 36582-36596, Monday July 17, 1995.World Health Organization (1995). Evaluation ofCertain Food Additives and Contaminants. WHOTechnical Report Series 859. WHO, Geneva, pp2-3.World Health Organization (1997). Evaluation ofCertain Food Additives and Contaminants. WHOTechnical Report Series 868. WHO, Geneva, pp2-6.World Health Organization (2001). Evaluation ofCertain Food Additives and Contaminants. WHOTechnical Report Series 901. WHO, Geneva, p3.

Other ILSI Europe PublicationsConcise Monographs• Alcohol – Health Issues Related to AlcoholConsumption• A Simple Guide to Understanding andApplying the Hazard Analysis CriticalControl Point Concept• Calcium in Nutrition• Carbohydrates: Nutritional and healthAspects• Caries Preventive Strategies• Concepts of Functional Foods• Dietary Fat – Some Aspects of Nutritionand Health and Product Development• Dietary Fibre• Food Allergy• Food Biotechnology – An Introduction• Genetic Modification Technology and Food– Consumer Health and Safety• Health Issues Related to AlcoholConsumption• Healthy Lifestyles – Nutrition and PhysicalActivity• Microwave Ovens• Nutrition and Immunity in Man• Nutritional and Health Aspects of Sugars –Evaluation of New Findings• Nutritional Epidemiology, Possibilities andLimitations• Oxidants, Antioxidants, and DiseasePrevention• Principles of Risk Assessment of Food andDrinking Water Related to Human Health• The Acceptable Daily Intake – A Tool forEnsuring Food SafetyReports• Addition of Nutrients to Food: Nutritionaland Safety Considerations• An Evaluation of the Budget Method forScreening Food Additive Intake• Antioxidants: Scientific Basis, RegulatoryAspects and Industry Perspectives• Applicability of the ADI to Infants andChildren• Approach to the Control of EnterohaemorrhagicEscherichia coli (EHEC)• Assessing and Controlling IndustrialImpacts on the Aquatic Environment withReference to Food processing• Assessing Health Risks from EnvironmentalExposure to Chemicals: The Example ofDrinking Water• Detection Methods for Novel FoodsDerived from Genetically ModifiedOrganisms• Exposure from Food Contact Materials• Foodborne Protozoan Parasites• Foodborne Viruses: An Emerging Problem• Food Consumption and Packaging UsageFactors• Food Safety Management Tools• Food Safety Objectives – Role inMicrobiological Food Safety Management• Functional Foods – Scientific and GlobalPerspectives• Guidance on the Safety Assessment ofBotanicals and Botanical Preparations forUse in Food and Food Supplements• Markers of Oxidative Damage andAntioxidant Protection: Current status andrelevance to disease• Method Development in Relation toRegulatory Requirements for the Detectionof GMOs in the Food Chain• Mycobacterium avium subsp. paratuberculosis(MAP) and the Food Chain• Nutrition in Children and Adolescents inEurope: What is the Scientific Basis?• Overview of Health Issues Related toAlcohol Consumption• Overweight and Obesity in EuropeanChildren and Adolescents: Causes andConsequences – Prevention and Treatment• Packaging Materials: 1. PolyethyleneTerephthalate (PET) for Food PackagingApplications• Packaging Materials: 2. Polystyrene forFood Packaging Applications• Packaging Materials: 3. Polypropylene as aPackaging Material for Foods andBeverages• Packaging Materials: 4. Polyethylene forFood Packaging Applications• Packaging Materials: 5. Polyvinyl Chloride(PVC) for Food Packaging Applications• Packaging Materials: 6. Paper and Board forFood Packaging Applications• Recycling of Plastics for Food Contact Use• Safety Assessment of Viable GeneticallyModified Micro-organisms Used in Food• Safety Considerations of DNA in Foods• Salmonella Typhimurium definitive type(DT) 104: A multi-resistant Salmonella• Significance of Excursions of Intake abovethe Acceptable Daily Intake (ADI)• The Safety Assessment of Novel Foods• The Safety Assessment of Novel Foods andConcepts to Determine their Safety in use• Threshold of Toxicological Concern forChemical Substances Present in the Diet• Transmissible Spongiform Encephalopathyas a Zoonotic Disease• Trichothecenes with a Special Focus onDON• Validation and Verification of HACCPTo orderILSI Europe a.i.s.b.l.83 Avenue E. Mounier, Box 6B-1200 Brussels, BelgiumPhone (+32) 2 771 00 14, Fax (+32) 2 762 00 44E-mail: publications@ilsieurope.beILSI Europe’s Concise Monographs and ReportSeries can be downloaded fromhttp://europe.ilsi.org/publications

ISBN 1-57881-188-0,!7IB5H8-ibbiif!