Document e74gQ1yNYO6oaovkvJdbNE2gq

FILE NAME: No Safe Threshold (NST) DATE: 1970 Apr 22 DOC#: NST009 DOCUMENT DESCRIPTION: Report to the Surgeon General - Evaluation of Environmental Carcinogens EVALUATION OF ENVIRONMENTAL CARCINOGENS Report to th e Surgeon General, USPHS Ap r il 22, 1970 Ad Hoc Com mittee on t h e Ev a l u a tio n of Law Le v e l s o f Environm ental Ch e m ic a l Carcinogens T h is r e p o r t was in tro d u c e d a s a n e x h ib i t and p u b lis h e d tn f u ll in both th e follow ing Senate hearings; "CHEMICALS AND THE FUTURE OF MAN" Hearings before the Subcommittee on Executive Reorganization ana Government Research o f th e Committee on Government O perations: U nited S ta te s Senate, A p ril S and 7, 1971j pages 180-198. and th e "FEDERAL ENVIRONMENTAL PESTICIDE CONTROL ACT" Hearings before the Subcommittee on A g ricu ltu ra l Research and General L e g is la tio n o f th e Committee on A g ric u ltu re and F o restry, U nited S ta te s Senate, March 23, 24, 25, and 28, 1971; pages 877-697. R eprinted by the U.S. Department o f H ealth, E ducation, and W elfare N ational In s titu te s of Health m B S OF THE AD HOC COMITTEE ON THE EVALUATION OF LOW LEVELS OF ENVIRONMENTAL CHEMICAL CARCINOGENS Na t io n a l Cancer I n s titl ite Beth esd a- Maryland Umberto S a f f i o t t i , Chairm an, A sso c ia te S c ie n tif ic D ire c to r f o r C arcinogenesis, E tiology, N ational Cancer In s titu te , B uilding 37, Room 3A21, B eth esd a, M aryland. Hans L. F a lk , A sso c ia te D ire c to r f o r L aboratory R esearch, N a tio n a l I n s titu te o f Environm ental H ealth S ciences, Research T rian g le Park, North C arolina. Paul K otin, D ire cto r, N ational I n s titu te o f Environmental H ealth S ciences, Research T rian g le Park, North C arolina. W illiam L ijin sk y , P ro fe sso r o f B iochem istry, The Eppley I n s titu te f o r R esearch on C ancer, U n iv e rsity o f Nebraska C ollege o f M e d icin e, Omaha, N ebraska. Marvin Schneiderman, A sso ciate C hief, Biometry Branch, N ational C ancer I n s t i t u t e , 'Viscon B u ild in g , Room 5C10, 8 e th e s d a , M aryland. P h ilip p e S h u b ik , D i r e c t o r , The E ppley I n s t i t u t e f o r R esearch on C an cer, U n iv e r s ity o f N eb rask a, C o lle g e o f M edicine, Omaha Nebraska Sidney W einhouse, D ire c to r, F e ls R esearch I n s t i t u t e , Temple U n iv ersity School o f M edicine, P h ilad elp h ia, Pennsylvania G e rald Wogan, P ro fe s s o r o f Food T o x ic o lo g y , M a ssac h u se tts I n s t i t u t e o f T echnology, 77 M a ssac h u se tts Avenue, Cambridge, M a ss a c h u se tts. S t a f f Members: John A. C ooper, E x ec u tiv e S e c r e ta r y , R ich a rd R. B ates Jam es A. P e t e r s , Howard R. R osenberg, E liz a b e th K. W eisburger, John H. W eisburger INTRODUCTION Establishment of this Ad Hoc Committee was requested on October 24, 1969, by the Deputy Assistant Secretary for Health and Scientific Affairs. The task of the Committee is to review the problems relating to the evalua- tlon of low levels of environmental chemical carcinogens, to consider the scien tific bases on which such evaluations can be made, and to advise the Department of HEW on the implications of such evaluations. The Committee, in addressing itself to the problems of environmental exposures to chemical agents from ail sources, has considered the scientific criteria for evaluation of carcinogenic hazards. Many previous recommendations on the criteria to be used for evaluating environmental chemical carcinogenic hazards have been made for specific sources of exposure or for specific groups of substances (e.g. food additives, pesticides, certain occupational carcinogens). In some eases this approach has led to. an uneven assessment of risks from different sources and to an uneven approach to preventive measures. The task of this Committee covers a broader area and includes an appraisal or the scientific criteria for evaluation of chemical carcinogenesis hazards In the total environment. i. recommendations In full consideration of the past and present states of carcinogenesis investi gation this Committee offers the following recommendations: 1. a. Any substance which is shown conclusively to. cause tumors in animals 6hould be considered carcinogenic and therefore a potential cancer hazard for man. Exceptions should be considered only where the carcinogenic effect is clearly shown to result from physical, rather than chemical, induction, or where the route of administration is shown to be grossly inappropriate in terms of con ceivable human exposure. b. Data on carcinogenic effects in man are only acceptable when they repre sent critically evaluated results of adequately conducted epidemiologic studies. 2. No level of exposure to a chemical carcinogen should be considered toxi cologically insignificant for man. For carcinogenic agents a "safe level for man" cannot be established by application of our present knowledge. The concept of "socially acceptable risk" represents a more realistic notion. 3. The statement made in 1969 by the Food Protection Committee, National Kesearch Council (see Appendix II) that natural or synthetic substances can be considered safe without undergoing biological assay should be recognized as scientifically unacceptable. 4. No chemical substance should lie assumed safe for human consumption without proper negative lifetime biological assays of adequate size. The mini mum requirements for carcinogenesis bioassays should provide fo r: adequate numbers of animals of at least two species and both sexes with adequate con trols, subjected for their lifetime to the administration of a suitable dose range, Including the highest tolerated dose, of the test material by routes of adminis tration that include those' by which man is exposed. Adequate documentation of the test conditions and pathologic standards employed are essential. 5. Evidence of negative results, under the conditions of the test used, should be considered superseded by positive findings iti other tests. Evidence of posi tive results should remain definitive, unless and until new evidence conclusively proves that the prior results were not causally related to the exposure. 6. The implication of potential carcinogenicity should be drawn both from tests resulting In the induction of benign tumors and those resulting in tumors which are more obviously malignant. 1 7. The principle of a zero tolerance for carcinogenic exposures should be re tained in all areas of legislation presently covered by it and should be extended to cover other exiweures as well. Only in the cases where contamination of an environmental source by a carcinogen has been proven to be unavoidable should exception be made to the principle of zero tolerance. Exception should be made only after the most extraordinary justification, including extensive documenta tion of chemical and biological analyses and a specific statement of the estimated risk for man, are presented. All eltorts should he made to reduce the level of contamination to the minimum. Periodic review of the degree of contamination and the estimated risk should be made mandatory. K. A basic distinction should be made between intentional and unintentional a. No substance developed primarily for uses Involving exposure to man should be allowed for wide-spread human intake without having been properly tested for carcinogenicity and found negative. lx Any substance d e v e lo p for use not primarily involving exposure in man but nevertheless resulting in such exposure, if found to be carcinogenic, should be either prevented from entering the environment or, if it already exists in the environment, progressively eliminated. . 9. A system should be established for ensuring that bioassay operations pro viding data upon which regulatory decisions are made be monitored so that their results are obtained in accordance with scientifically acceptable standards. 10. A unified approach to the assessment and prevention of carcinogenesis risks should be developed in the federal legislation; It should deal with all sources of human exposure to carcinogenic hazards. U. Clear channels should be identified for the regulatory function of different Government departments and agencies in the field of cancer prevention. Estab lishment of a surveillance and information program would alert all concerned Government agencies to the extent and development of information on formation on carcinogenic hazards. 12. An ad hoc committee of exiierts should be charged with the task of recom mending methodos for extrapolating dose-response bioassay data to tbe low response region (1-10000% to 1-10000000%). The low doses corresponding to the responses in this range are the ones which have direct relevance to the human situation. II. BACKQBOUND Knowledge of cancer causation by chemicals originates from clinical observa tions, going back us far as 1775 with I'ott's discovery of soot as the causative gent in chimney sweeps' cancer. Several major classes of carcinogenic agents were first discovered by their effects on man. Experimental animal models for the determination of the potential carcinogenic activity of chemicals were only devel oped tn the last 50 years, and most of them have been studied only in the last 20 years. The effects of carclnagens on tissues appear irreversible. Exposure to small doses of a carcinogen over a period of time results in a summation or potentiation of effects. The fundamental characteristic which distinguishes the carcinogenic effect from other toxic effwts is that the tissues affected do not seem to return to their normal condition. This summation of effects in time and the long interval (latent period) which pusses after tumor induction before the tumor becomes clinically manifest demonstrate that cancer can develop in man and in animals long after the causative agent has been in contact uml disappeared. It is, therefore, important to realize that incidences of caucer iu man today reflect exposure of 15 or more years ago: similarly, any increase of carcinogenic contaminants iu man's environment today will reveal Its carcinogenic effect some 15 or more years from now. For this reason it is urgent that every effort be made to detect and control sources of carcinogenic contamination of the environment well before damaging effects become evident in man. Similar concepts may apply to the needs for evaluation of other chronic toxicity hazards. Environmental cancer remains one of the major disease problems of modern man. An agent which is causally related to the occurrence of cancer in man or animals is defined as a carcinogen or oncogen. The number of known carcinogenic agents includes several groups of viruses, various physical factors, and hundreds of chemicals. Viruses of different types are known to induce cancer in animals; none has yet been proven to evoke cancer In man. If specific viruses are proven to be causally 2 related to cancer Induction In man, tbe frequency of certain human tumora might be reduced In the future by Immunization procedures. . Physical factors are known to cuuse cancers in man and animals. For example, ultraviolet radiation causes skin cancer, and ionizing radiation cancer of various organs (e.g. leukemias, lung cancer, bone sarcomas, skin cancer). Exposure to a '`background level" has been widely considered as unavoidable and, in the case of ultraviolet light, even necessary as an Integral part of our natural environ ment. Strong epidemiologic and experimental evidence indicates the existence of a direct dose-response relationship between exposure to radiation and carcino genic effects. Tolerance levels have been suggested for various forms of radiation and health benefits have been realized from their application. Evaluation of radiation hazards has been approached through measurement of the total cumu lative dose of radiation exposure. Some carcinogenic radiation hazards, such as certain occupational exposures (e.g. radiation in uranium mines), are still not effectively controlled. Chemicals of many classes produce cancer in a large number of organ sites In animals. Cancers in man ure known to be caused by several individual chemicals and by materials composed of mixtures of chemicals. Chemical carcinogens have been shown to act by surface contact with skin or mucosae, by inhalation, by in gestion, and occasionally by injection or implantation (medical or accidental). Chemicals may induce cancer at the site of initial contact (e.g. skin cancer from 'radionuclides), the site of metabolism and detoxification (e.g. liver or kid ney cancer from ufiatoxin or nitrosamines). or the site of excretion (e.g. urinary [polynuclear hydrocarbons), the site of selective localization (e.g. bone cancer Diadder cancer from aromatic amines). A complex and often uneven approach to the problem of preventing exposure to chemical carcinogens has developed over the years. It has become increasingly obvious that the hazard from a single chem ical carcinogen cannot be evaluated out of coutext of tbe total environmental ex posure (J). Estimation of the ``cumulative carcinogenic dose'' resulting from all possible chemical carcinogens or even from all sources of a single type or class of chemical carcinogens is presently impossible. Prevention of exposure to known carcinogenic chemicals depends largely on man's ability to control their entry into the environment. Certain chemical car cinogens are natural products (e.g. metabolites of the amino acid tryptophan) or naturally occurring contaminants (e.g. mycotoxins). Others are formed in the processing of natural products. Many, such as polynuclear hydrocarbons (e.g. benzito[a]pyrene), occur almost ubiquitously in our modern industrialized en vironment. They derive from most sources of organic combustion. A class of very potent carcinogens discovered only in recent years, tbe N-nitrosamines, include compounds that may be formed in the environment from nitrites and secondary amines. Many other known chemical carcinogens have been introduced as syn thetic materials or by-products into man's present environment through a wide range of newly-developed industrial processes. Some of these, such as food addi tives, medicinal products, cosmetics, and certain household products or pesticides, were developed for human use. Several carcinogens derive from products such as tobacco smoke, developed exclusively for human use. In other cases chemical carcinogens not intended primarily for human exiiosure are introduced Into the general environment and eventually come in contact with its Inhabitants ; many substances (certain polynuclear hydrocarbons, pesticides, metals, dusts, and fumes, etc.) gain widespread environmental distribution, thereby becoming pol lutants of the air, soil, water, and food. Prevention of exposure to this broad spectrum of chemical carcinogens must take a variety of forms. The production of chemicals recognized as carcinogens for uses involving intentional human exposure can be identified and effectively eliminated. Ex ceptions to this approach should be made for substances that involve a well- defined health benefit (e.g. certain chemotherapeutic drugs). Use of such substances should be accepted on the basis of extraordinary evidence that their health benefit outweighs their risk. The production of specific carcinogenic chemicals for uses thnt do not pri marily involve an. intentional exposure of man, but which result in such en vironmental contamination that extensive human exposure becomes inevitable, must also be controlled. The most effective prevention of exposure in man is the elimination of carcinogen production, or control of entry into tlie environment. A large group of chemical carcinogens (e.g. combustion products, myco toxins, and other natural products) is widely disseminated in the environment from sources that can only be partly controlled. For these contaminants, as 3 well as for products which have been widely spread in the environment before their carcinogenicity was recognized, the only possible approach to exposure reduction is to monitor their environmental distribution and subsequently mini mize their contact with humans. Modifying factors are known to condition the development of neoplasia in man and animals. They can act intrinsically or extrinsically (e.g. hormonal im balances, metabolic characteristics or abnormalities, caloric intake, dietary factors). Understanding of their specific effects in man, however, is still not adequate to serve as a reliable basis for preventative action. Interactions among multiple factors have received limited attention to date. There are well-documented instances in animal studies of strong synergistic effects produced by chemicals in combination with radiation, viruses, or other chemicals. The epidemiological patterns of certain human cancers implicate combined effects of multiple agents (e.g. inhnlatlon of radon and radon daugh ters in uranium mines and cigarette smoking). The types of cancer in man that are due, directly or indirectly, to extrinsicfactors are thought to account for a large i>ercentage of the total cancer incidence (2). These include tumors of the skin, the respiratory, gastrointestinal and urinary tracts, hormone-dependent organs (such as the breast, thyroid, and uterus), and the hemopoietic system. During the past decade considerable progress has been made In the detection of carcinogenic agents and the analysis of their biological effects. New approaches to the interpretation of quantitative relationships between eximsures and carcinogenic effects in man and animals are being developed. It Is estimated, therefore, that the majority of human cancers are potentially preventable (2). IIL ANIM AL BIOASSAY RESULTS AND EVALUATION OF B IS K S IN MAN In order to evaluate the hazard of a chemical for man, one must extrapolate from the animal evidence. It is essential to recognize that no level of exposure to a carcinogenic substance, however low it may lie, can be established to be a "safe level" for man. This concept, put forward in the lDoO's, remains true in lhTO. The current legislation in the field of food additives, with its "anticancer clause", Is based on this principle (Federal Food, Drug and Cosmetic Act, as amended, Sect. 40b (e) (3) (A )). The reusons for retaining this "nntleaneer clause" were effectively summarized in 1080 by Secretary of Health, Education and Welfare Arthur S. Flemming in testimony to Congress (d) on the subject of extending the clause to cover the use of food colors, with the following statement. "The rallying point against the anticancer provision is the catch phrase that it takes away the scientist's right to exercise judgment. The issue thus made is a false one. because the clause allows the exercise of all the judgment that can safely be exercised on the basis of our present knowledge. The clause is grounded on the scientific fact of life that no one, at this time, can tell us how to establish for man a safe tolerance for a cancer-producing agent. "Until cancer research makes a breakthrough at this point, there simply is no specific basis on which judgment or discretion could be exercised in tolerating a small amount of n known carcinogenic color on food additive. "As 1 pointed out in my original testimony, the opposition to Inclusion of an anticancer clause arises largely out of a misunderstanding of how this provision works. It allows the IkiKUttneirt and its scientific jieople fuil discretion and Judgment in deciding whether a substance has been shown to produce cancer when added to the diet of test animals. But once this decision is made, the limits of Judgment have been reached and there is no reliable basis on which discretion could be exercised in determining a safe threshold dose for the established carcinogen. "So long as the outstanding experts in the National Cancer Institute and the Food and Drug Administration tell us that they do not know how to establish with any assurance at all u safe dose iu man's food for a cancer-producing substance, the principle in the anticancer clause is sound. "I want to emphasize the statement I made on January 28 that the Food, Drug, ami Cosmetic Act, as it now stands, will be enforced to prohibit the addi tion of caticer-pro'ducing substances to food unless a law should be passed direct ing us to follow another course of action. ''Eren-Dmugh.we.have-tliiu authority -bv-tbedaw, me urge the Ci-nyr-fm tr "Even though we have this authority in the law, we urge the Congress to 4 Join with the executive branch to. give added assurance to the consuming public by directing the anticancer clause in the proposed color additives amendment "Again, we say, however, that we believe the issue is so important that the elected Representatives of the people should have the opportunity of examining the evidence and determining whether or not the authority should be granted." The scientific basis on which the Government's position was established in 1960 remains valid. The progress of knowledge in carcinogenesis in the last decades has only strengthened the points made in Secretary I'lemming's testi mony. IV. DETECTION OF LOW LEVELS OF CARCINOGENS IN TH E ENVIRONMENT To establish the presence of "low levels of carcinogen in the environment" requires that 1) the presence of the material in question be recognized in the en vironment and 2) the material be recognized as carcinogenic. To evaluate the impact of a chemical in the human environment, it is useful to prepare an "en vironmental profile" to reflect the distribution of this material in time and space. Failure to detect the presence of a compound implies only that the compound is present, if at all, in concentrations below the detectable limit of the analytical method used. These "sub-detection levels" cannot he differentiated from "zero". From the distribution profile and additional information on tbe conditions of uptake in man the approximate level and extent of exposure for population seg ments can be estimated. In recognizing a chemical as a carcinogen, the limiting factor is the sensi tivity and specificity of the bioassay system used. A bioassay system designed to detect tumor induction only at or above a given level under the conditions of the test (e.g. a 257c incidence of a specific tumor type) will fail to reveal car cinogenicity below that level. Compounds whose carcinogenic effects fall below specific bioassay detection limits must not be considered innocuous. Such mate rials must be characterized as presenting a carcinogenic risk no greater than that defined by this lower limit. Methodology for the determination of chemical contamination in the environ ment and of biological activity of carcinogens are discussed in the following sections. A. Chemical Detection Methods Methods for detection of low levels of carcinogens in the environment have in creased in accuracy and reliability over the past several years. The lower limits of detection for different types of known carcinogenic substances are extremely variable, extending over several orders of magnitude from very sensitive methods (e.g. 1 part per billion of benzo() pyrene or alia toxin; to rather insensitive ones (e.g. for aromatic amines). In principle, analytical methods should be capable of detecting carcinogenic materials at any level nr in any condition which has rele vance to'human exposure. For tills reason, increasingly sensitive analytical tech niques are needed, and indeed many have been developed over the last 10 years. Appendix I summarizes the present state of the art in the analytical determina tion of low levels of environmental carcinogens ( I ) and indicates the lower limits of detection available today, in contrast with those of 1950. Much of the improve ment in methodology is attributable to the application of gas-liquid chromato graphic techniques. Within the next few years sizable additional improvements in the sensitivity of analytical methods are likely to be achieved. It is important to consider how widely the new analytical methods can he applied for the detection of a given carcinogenic contaminant in different mate rials. While highly sensitive analytical methods can lie devised to detect a chemi cal in specific materials, these same methods might lie powerless in the analysis of the same chemical from other source materials le.g. dimethylnitrosamine can be detected in the alcoholic beverages at 1 ppb. hut in foods only at l(V-ltf) ppbh All uneven evaluation of the sources of environmental contamination may result. De velopment of widely applicable procedures will provide a more balanced evalua tion of environmental contamination. B. Biological Detection Methods The carcinogenic activity of materials can only be detected by long-term bio logical tests. At the present time the chemical structure or physico-chemical properties of a compound do not provide a reliable basis for prediction of free dom from carcinogenic activity. Several structure-activity correlations are valu able indicators of the possible carcinogenicity of a compound, but none can be 5 Q9ed to classify the compound as non-carclnogenic. Short-term bloassays that de termine the effect of certain chemicals on selected biologic targets have not been reliable for prediction of carcinogenic activity. The present state of the art requires long-term bioassays in mammalian species tor the experimental identification of carcinogenic activity. United States law requires that food additives and various other materials be tested in animals by the intended route of human exposure. Similar tests have not been required for some materials to which humans are exposed by other than the orai route. The expanding production and use of chemicals in household products results In extensive human exposure (via the skin and respiratory tract) to dusts and aerosols; littie information is available on the chronic toxicity of these materials by these routes of administration. It would not be wise to wait for the results of these "experiments In man" before instituting animal experimentation. Bioassays are always performed on a number of animals which is extremely small when compared with the millions of humans exposed to most environmental carcinogens. Such studies can only detect carcinogenic effects resulting in fairly high incidences. For example, an observed outcome of no tumors in a test group of 100 animals, as well as in 100 negative controls, only provides assur ance, at the 99 percent probability level, that the true tumor risk is under 4.5 percent. The maximum probable risk is 0.46 percent if groups of 1000 animals are used. It would require tumor-free results in 450 animals to establish with like probability that the risk is under 1 percent U ). The assessment of the carcinogenic activity of a chemical depends on a variety of parameters. These inciude not only the total number of tumors In duced but also their multiplicity, latent period, morphologic type, and degree of malignancy. The induction of tumors diagnosed as benign as a result of treatments has been interpreted by certain groups in the past as not sufficient to demonstrate a ''carcinogenic" effect. This is a dangerous position since lew, if any, substances are known to have produced only benign tumors and no malignant ones when properly and repeatedly tested. This has been pointed out in the Report of the Subcommittee on Carcinogenesis of the FDA Com mittee on Protocols for Safety Evaluation (5). The Important scientific problem of defining the sensitivity of a bioassay system used for testing materials of unknown activity has received insufficient attention. The interpretation of both positive and negative findings is strictly dependent on such definition as well as on the results obtained In negative, vehicle, positive and colony control animals. A bioassay result is meaningful only when accompanied by a statement of the sensitivity and specificity of the bioassay design used. An observed incidence of a given tumor type in a test group has no meaning without adequate information on the appropriate con trols. Far too littie work lias been done using adequate positive controls. Daek of tumor response in a given experimental system cannot be interpreted as negative evidence if positive controls also yield negative results or if uo posi tive controls have been included to show that the experimental system used Is appropriate. , . A body of knowledge has developed over the years on the response ot ex perimental animals to chemical carcinogens. Several committees of experts in the field of carcinogenesis convened by national and international bodies over the past 15 years have formulated general principles for performance and evaluation of carcinogenesis studies in auimals. The recommendations put forth by these committees have shown remarkable unanimity ffi, 5-10) and are widely accepted in principle by the scientific community (11-15). General requirements for testing procedures, which have been outlined by these groups, include speci fication of criteria for the following: 1. selection of materials to be tested; if. chemical and physical characterization of the test m aterials: 3. selection of appropriate animal species and group size; and 4. choice of appropriate routes and levels of administration. In addition, recommendations concerning the lifetime maintenance and path ological examination of experimental animals have been outlined. Two principles are recognized as fundamental to the evaluation of carcino genesis bloassays. 1. The minimum requirements for carcinogenesis bioassay should include adequate numbers of animals of at. least two species and both sexes with adequate positive and negative controls, subjected for their lifetime to the 09-044 0 -- 71------ 14 6 administration by appropriate routes of a suitable dose range of the test ma terial, including doses considerably higher than those anticipated for human exposure. 2. Any substance which is shown conclusively to produce tumors in animals, when tested under these conditions, should be considered potentially carcinogenic for man. V. QUANTITATIVE RELATIO N SH IPS The major new argument presented today against the "anticancer clause" is that the marked increase in sensitivity of many analytical methods makes it possible to detect low levels of carcinogens in a broader segment of the environ ment and that, therefore, the immediate enforcement of regulations requiring a zero tolerance becomes more difficult, in some instances impossible. New and very potent classes .flt.chemical carcinogens, such as aflatoxin and nltrosamines, have been detected in the environment. Striking examples of potentiation in cancer induction have been reported in experimental animal tests and in epidemiologic observations. Bioassays have revealed the carcino genicity of such widespread environmental chemicals as DDT and cyclamate, to which a large majority of the American population has been exposed. In contrast to the analytical methods, bloassay methods have 'remained tools of low senitivlty, capable only of detecting the highest peaks of carcinogenic activity. The factor which limits bioassay sensitivity is usually the small number of test animals used. If the bioassay design has a low probability of detecting carcinogenic effects produced by hazards at levels comparable to those present in environmental samples, then tests at such levels are wastes of time, effort and money. The need to test dim levels higher than those found in the environ ment is thus founded. Some substances, on the other hand, are potent carcino gens in animal test systems at levels not currently detectable In the environment. An example is provided by the recent evidence on a (Intoxin, its lowest analytical ly detectable level is 1 ppb. One hundred percent tumor Incidence was produced in rats by a dose as low ns 15 In the diet. Experiments now under way suggest that aflatoxin, when fed to rats at the lowest detectable level (1 ppb), is still carcinogenic (10). It has already been demonstrated to be carcinogenic at 1 ppb In the trout. These data indicate that aflatoxin may be present ln'food at undetectable levels and still be capable of producing cancer Incidences so high as to be detectable in tests involving relatively small numbers of experimental animals. ^ It is impossible to establish any absolutely safe level of exposure to a carcino gen for' man. The concept of "toxicologieally insignificant" levels (as advanced by the Food Protection Committee of the XAS/XRC in 1060; see Appendix II). of dubious merit In any life science, has absolutely no validity In the field of carcinogenesis. Society must be willing to accept some finite risk as the price of using any carcinogenic material In whatever quantity. The best that science can do Is to estimate the upper probable limit of that risk. For this reason, the concept of "safe level for man", as applied to carcinogenic agents, should be replaced by that of a "socially acceptable level of risk". While science can provide quantitative information regarding maximum risk levels, the task 'ultimately selecting socially acceptable levels of human risk rests with society and its political leaders. The evaluation of the balance of bene fits and risks, required for such a decision by society, should not be the result of uninformed guesswork but should be reached on the basis of complete and perti nent data, social as well as scientific. It is necessary, therefore, to define the extent in the processes of interpreting animal response data (Appendix III) and subsequently extrapoluting them to man (Appendix IV), The principle of zero tolerance should be applied in all but the most extraordinary of cases. VI. CONCLUSION Modern society has been extremely fortunate--given the technical limits on detection of carcinogenic effects--that at Least some environmental carcinogens have been identified. So-called negative data, obtained in bioassays often In capable of detecting effects below the 10 percent level, are grossly Inadequate to give assurance of safety for man. Information on about 2,500 compounds tested for carcinogenic activity up through 1900 has been compiled and pub lished (17). Most of these materials, however, are of no environmental signifi cance. Data on tests reported since I960 will be published shortly. It is estimated that data on 3,500 previously unevaluated chemicals will be included In the 7 forthcoming volume. It is seen, then, that ahout 6,000 chemicals are documented as having undergone carcinogenesis bioassay to date. Many of the referenced tests, however, were inadequate according to presently recommended standards. If this nation wishes to identify a large segment of existing and potential carcinogenic hazards, it must institute a comprehensive program involving a coacert of actvities. Scientific and technical plans for the development of method ological standards should be provided by experienced agencies in collaboration with qualified advisors. It is essential that the objectivity of these advisors not be damaged by any conflicting interests. Resources needed for the extensive bioassay screening of environmental chemicals will be considerable. In addition to the myriad of substances pres ently in the environment, several thousand new compounds are introduced each year. Up to 20,000 materials should be tested for carcinogenicity as a first screen ing of the environment. Testing 20.000 compounds by bloassay would cost about $1 billion. This estimate would increase accordingly for the more extensive testing required in less superficial evaluation. Yet even were such funds avail able today, they could not nearly be spent effectively. Bioassay laboratory and professional resources are just not available in quantities capable of supporting a huge testing program. A great deal of "tooling up" is prerequisite to any such expanded level of effort. Because the latent period In human carcinogenesis Is so long, epidemiologic evidence develops only over periods of 15 to 20 years. Timely decisions to exclude materials from uses involving exposure to man, therefore, must be based solely on adequately conducted animal bioassays. Retrospective human, evidence of risk moat.not be allowed to show itself 'oefoie controlling action is taken. Chemicals should be subjected to scientific scrutiny rather than given individual "rights" ; they must be considered potentially guilty unless and until proven innocent. Valid evidence must come from biological assays; every bioassay report should include a statement of its limits of sensitivity. Experimental design should provide for reproducibility of test results. Since the bioassay plays such a key role in a total carcinogen control scheme, more effort must he devoted to setting standards for both the performance of tests and the interpretation of results. Only given good bioassay data can science possibly provide sound information to those who are charged with making social decisions regarding the accept ability of carcinogenesis risk levels. An effective program to protect man from the mass of environmental cancer hatari's Is within reach. A'o more time should be allowed to pass before the recommendations set forth in this report are applied to reality. Appendix I--Chemical D etection Methods : Analytical P rocedures and T heir Sensitivity A review of these methods has recently been published by the International Union Against Cancer (/). A brief summary follows. I. POLYNUCLEAR COMPOUNDS Area or concern.--Carcinogenic materials failing into this general class are, at least potentially, produced under all circumstances where pyrolysis of organic substances occurs. Many meinliers. of this chemical group are carcinogenic. Analytical methods provide separation and determination of benzaia )pyrene, dibenz(o,ft) anthracene and dila*nz( a.ii)acridine and other compounds commonly found together in combustion products (tars, soots, petroleum derivatives, etc.). Analytical methods.--Acceptable analytical methods must be capable of sepa rating and measuring at least the specific mateilals mentioned above. At present the methods which appear most effective dejiend on Absorption, fluorescence spectrometry, mass spectrometry, or gas-liquid chromatography for quantifica tion and one of several chromatographic techniques for separation. Sensitivity.--The minimum concentration of polynuclear compounds which may be determined by present methods is approximately 1 part of an individual compound per billion (1959 sensitivity approximately 190 ppb.). 8 n . CHLORINATED HYDROCARBONS Area of concern.--The chlorinated hydrocarbons have been and still are ex tensively used as industrial solvents and reactive intermediates. Carbon tetra chloride is a well-known liver toxin and is suspected of being a carcinogen. Dur ing the past decade, other chlorinated hydrocarbons have been widely used as persistent pesticides and general exposure of man has occurred. DDT, DDD, aldrin, dieldrin and heptaehlor have been reported as carcinogenic in animals. Analytical method*.--Current methods are based on extraction and condensa tion followed by separation and quantification by gas chromatography. Sensitivity.--The advent of electron capture detectors has made methods of analysis for this class of compounds extremely sensitive. Cutrent minimum levels of detection are about 0.1 ppb (1959 sensitivity approximately 10 ppb). H I . AROMATIC A M IN E S AND CHEM ICALLY RELATED COMPOUNDS Area of concern.--For purposes of this discussion the "aromatic amine" are considered to include not only the substituted aromatic amines but also those compounds (nltro-. azo-, or hydroxylamine compounds) capable of being con verted into aromatic amines by metabolic processes. Occupational exposure aud exposure through the air, foodstuffs, plastics, and drinking water pose threats of carcinogenic hazards for humans. One must also consider the contribution to the carcinogenic load of endogenous aromatic amines such as certain tryptophan metabolites. Analytical methods.--General methods for the determination of "total aro matic amines" are of little value because of the wide variation in carcinogenicity between members of the class. Estimation of specific substances is generally effected by their conversion to primary aromatic amines and then to azo dyes by diazotization and coupling; the azo dyes are separated chromaiographically and estimated coloriiuetrlcall.r. . Sensitivity.--Methods for detection of members of this general class vary widely in their ensitivity due to the diverse properties of the compounds included. Azo compounds are easily estimated at sensitivities on the order of 10-100 ppb, while aromatic amines and hydroxylamine can be estimated with precision only at levels of about 1 ppm (1950 sensitivity about the same). IV. N-NITROSO COMPOUNDS Area of concern.--During recent years many members of this class of chemi cals (nitrosamines and nitrosamldes) have been recognized as highly potent carcinogens capable of inducing cancel' in a variety of organ sites. There Is a growing concern that these materials may represent significant cancer hazards to man. These compounds first came to our attention as a result of their utiliza tion as industrial solvents and chemical intermediate, but they may be widely distributed at. low levels in a variety of natural products. It is suspected that they may even be formed by the interaction *oi natural products in the organism or during food preparation. Analytical methods.--Methods in current use depend on the extraction of these substances from a variety of starting materials and cleanup of the extract by various chromatographic techniques followed by quantification of the individual compounds by gas-liquid chromatography. Specific methods have only been de veloped for a few individual members of the class. No satisfactory methods for the measurement of nonvolatile nitroso compounds have yet been developed. Sensitivity.--Minimum detectable levels for the individual compounds are at present about 100 ppb (1959 sensitivity= relatively insensitive colorimetric or polargraphic procedures). V. AFLAT0XIW8 Area of concern.--Human exposure to the aflatoxins occurs principally by way of contaminated foods. It is now clear that such contamination can occur in essentially any food commodity, when harvest, storage, or transport conditions permit the growth of specific spoilage molds. Analytical methods.--Currently used methods involve extraction and cleanup followed by identification of the compounds on thin layer chromatograms viewed under long wave ultraviolet light. 9 Sensitivity.--Although problems are associated with all of the presently avail able techniques, detection is possible at the level of 1 ppb (1959 sensitivity=prob- lem was unrecognized). VI. INORGANIC SUBSTANCES Area of concern.--A number of inorganic substances of common occurrence in the environment have been reported in the past to be carcinogenic in man or certain animal species. They include compounds containing the following ele ments : arsenic, beryllium, cadmium, chromium, cobalt, lead, nickel, silver, titanium. Analytical methods.--Analysis of the metals is performed today by atomic absorption spectroscopy after apropriate extraction and concentration procedures. Sensitivity.--The lower limits of detection for the metals listed are on the order of 1 ppb (1959 sensitivity=about 10 ppb). vn. asbestos Area of concern.--This substance is a fibrous partlculnte material produced in vast amounts for many industrial uses and recently found to be a widespread air pollutant. Major monitoring needs concern its presence in the general atmosphere and as an industrial dust. Analytical methods.--Although sensitive methods are available for the detec tion of the elemental components of this substance, no specific chemical methods for identification of asbestos itself are available. Its estimation is still based on microscopic identification and count of asbestos particles in measured samples. Sensitivity.--Extremely low and little improved since 1959. VIII. OTHERS A number of other carcinogens, belonging to a wide variety of chemical classes, need to be analyzed with specific methods. Their enumeration exceeds the limits of the present discussion. It is also necessary to improve the detection of materials which are recognized as potential sources of carcinogens and for which no ade quate analytical methods are presently available (e.g. secondary amines and nitrites, which may interact to form nitrosamines). Appendix II--Comments on 1969 R eport of the F ood P rotection Committee This Committee has examined a report entitled "Guidelines for Estimating Toxicologically Insignificant Levels of Chemicals in Food" published in 1969 by the Food Protection Committee---Food and Nutrition Board of the National Academy of Sciences---National Research Council. It records its strong objections to the principles expressed in that report, which states that natural or synthetic substances can be considered safe without experimental support under certain vaguely-stated conditions. The Food Protection Committee Report assumes that ". . . a level of insignifi cance may be determined if: (11 There are available adequate scientific studies that establish safe levels of similar magnitude for at least two analogous sub stances. (2) The acute or subacute toxicity of the new substance and two analo gous substances is of the same nature and degree." For "Chemicals in Commer cial Production" it recommends that: "If a chemical has been in commercial pro duction for a substantial jariod, e.g. 5 years or more, without evidence of toxi cological hazard incident to its production or use. if it is not a heavy metal or a compound of a heavy metal, and if it is not intended for use because of its bio logical activity, it is consistent with sound toxicological judgment to conclude that a level of 0.1 ppm of the chemical in the diet of man is toxicologically insigInficant." To assume (a) that a 5-year period of use has any meaning for the evaluation of chronic toxicity in man, (b) that any chemical may he considered safe simply because two "analogous substances" are "safe", and (c) that acute or subacute toxicity are reliuble. guidelines for evaluating long-term toxicity is to display a lack of understanding and appreciation of factors involved in chronic toxicity, particularly of the irreversible and delayed toxic effects which occur in carcino genesis. 10 Since the purpose of the report ts to recommend guidelines and priorities for electing chemicals for human use without direct experimental toxicological evaluation, the lack of consideration of irreversible long-term toxic effects (which would not be ruled out by the suggested criteria) makes the suggested approach practically Inapplicable and potentially dangerous. Appendix I I I --A Method fob D etebmininq the Dose Compatible With Some "Acceptable" Level o r Risk (Contributed by Dr. M. A. Schneiderman) iN TKODucnor.- In establishing the concept of an "acceptable" risk dose (ARD) we must bear in mind that the dose level arrived at gives a tolerable risk only 1or the speciea for which the extrapolation has been made. Given this caveat, to make a conservative* estimate of an ARD. iwo assumptions need to be made: A. The dose-responsive curve at low response levels is concave npward (e.g. the left-hand tail of the common S-sbaped curve). The upper (dose) limit of this S-shaped curve is a straight line going through the 0.0 point (no dose= no response). The true dose-response curve is shallowest at this point B. A given set of data wiil provide an estimate for the upper limit of the other point through which the straight line must pass (which then gives the upper limit of the slope of the dose-response curve). As an example of the u-e of the ARD concept, let us suppose that we have observed 100 control animals and 100 treated animals at a given dose rate, d, and have seen no tumors in either group. The upper Si.%% confidence limit on such a result Is 3/100 (i.e. the data are consistent with the statement that with a true response rate of .03, we might expect to see zero tumor incidence in 05% of similar experiments). If the socially "acceptable" risk selected is to be one additional cancer case for each hundred thousand persons, It is now desired to establish an ARD which will produce a maximum lifetime tumor incidence of 1/100,000. We can then construct the graph : The procedure outlined here is essentially the "one hit" procedure and gives a some what more conservative answer than the procedure of Mantel and Bryan (1). 11 The line connecting the (0,0) point with the (d, .08) point ha Slope-03/d and the equation of the line la : Response= (Dose) (.03/d) If we wish to determine some dose (e.g. the ARD) which would be predicted to yield a response of not more than DIO- *we would write : D o se= (l'10rJ) (d/.03)=3.3-10-`d) ARD "Working towards an ARD this way has several consequences: a. An increase in the experimental size (at the same dose, d, and with the same experimental result, i.e. no tumors) will reduce the upper confidence limit. Thus, if there were 200 animals in each contrast group, the upper 95% confidence limit would be about halved and the estimated ARD would be doubled. b. An Increase in the dose, d, at which the result (no tumors) occurred would Increase the estimated ARD in direct proportion to the tested dose. Thus, if the Initial experiment had been conducted at twice as high a dose with the name result, the estimated ARD would be twice as high as that determined from this experiment. The procedure outlined has the virtue of lending greater credence to con clusions reached on the basis of larger experiments. Higher estimated ARD's will be found under such circumstances and the confusion between "hot statisti cally significantly different'', and "not different" can be avoided. It must be stressed that these calculations do not establish an "acceptable risk dose" for man. The procedures do not tell us for example what factors are required to extrapolate from animal to man, although the following factors must certainly be considered. 1. Dose-response curves In man (a grossly non-homogeneous animal) are likely to be much shallower than dose-response curves in experimental animals. This implies higher levels of response at iow doses for man, other things being eqnal. 2. It is possible that even the use of the straight line between the 0,0 point a n d some experimentally determined dose-response point (plus upper confidence limit) may yield too high an estimated ARD for man. 3. Whether the proper blow-up factor for doses in man should be on a mg/kg or mg/MJ basis, or any other basis, must be established independently. 4. Nothing in this procedure would enable one to know where the doseresponse curve for man belongs along the dose scale in comparison with the dose-response curve for the experimental animal. If man's curve lies much to the left of the experimental animal's curve, then nn ARD for the experimental animal may be a gross overdose for man. Where several species are studied, it would seem safest in the absence of better Information, to accept for man an ARD no greater than the lowest ARD dose derived from the results for the several species. I t has been suggested that an appropriate "safety factor" for man should involve a reduction in the ARI) for animals by a factor of 100 to allow for si<ecies' differences, another factor of 100 to allow for Interactions with other carcinogens, and another factor of 100 to hedge against the incorrect choice of "blow up" (weight, or surface areal from animals to man. This would Imply an ARD In man of about 1 X 10--*the ARD in animals. INSTRUCTIONS r o a CALCULATION OF T H E ARD USING 9 5 % AND CONFIDENCE L IM IT S (Experiment in which no animals in either treated or control group show a response, e.g. development of a tumor) : We need the following values: C=Upper confidence limit on the result (computed as shown below). S=Arbitrary "acceptable" response level. d=Acceptable risk dose (ARD), that which causes no more than the arbi trarily allowed acceptable risk. t d=Test dose, some multiple, t, of the tolerated dose. These values are related as shown : 12 From the above figure, we have the following relationships: C /S= (t-d )/d = t which can also be stated as C=St. Since C is a function of the size of the experiment, N, we have a relationship between S, t, and N where N=samplesize. For zero positive responses, the upper a7b confidence limit is S t= C = l--e [ ln ( l--a)1/N , where In is the natural logarithm. From this rela tionship, given any two of the values X, t. and S, we can solve for the third. The accompanying graphs have this relationship shown for difference values of N. Figure A1 s for a 95% confidence limit: Figure A2 is for a 99% confidence limit. , In any specific situation St is constant <=C), so that we can increase the range of the chart by dividing S (the tolerable risk level) by k while multiplying t (the dose multiplier) by k. Two sets of values for S and t are shown on each graph. . EXAM PLES OE MOW TO U S E GRAPHS 1. Given: a. That we agree to a socially "acceptable1' risk level, say .00001 (1 in 100,000). b. The expected average dose in man is I), and we wish to determine whether it is compatible with the acceptable risk. c. The maximum pharmacologically tolerated dose in our test, species is tu, where, say, t=600, (i.e. this species can survive a dose 600 times larger than the average dose in man). Procedure ( Using "95%" figure) a. Find t=600 along the bottom scale of the graph. Notice that It is in Paren thesis. This means we must look for our S value in parenthesis, too. (bmau arrow at bottom of graph.) b. Find S--10f X10*). (Small arrow at side of graph. ) ^ ^ ., _ . c. Find where the t=000 vertical line crosses the b=.00001 horizontal line. This Is at the diagonal line labeled 500. . ,, . This means we must conduct an experiment with a00 animals in each group, yielding 0 positives, at a dose level 600 times the "average" d e in man, in ovdn to have some assurance that the average dose in man is within the newpraMersk level, where acceptable risk is .00001. [This dose may be far too high, for it has been assumed that man's dose-response curve is the same (except for slope) as 13 the animal specie In which we did the experiment. No "safety''factors have been applied.] 2. Given To repeat the above experiment using a dose 300 X the expected ABD. S=.00001, and t=300. Procedure Find t=300 along the bottom scale (In parenthesis). Find S=10(X 10~*) Note the crossing point at diagonal N=1000. This means If we test at a lower dose, we most have more test animals to achieve the same ARD. 8. Given To repeat experiment 1 using only 100 anlmalg. X=100; S=10(X10'*) Procedure Follow along the horizontal line until you come to diagonal N=100 Look down to the t scale and note that we are at t=3000. This means if we test with fewer animals, we can have assurance only that a dose 1/t times the test dose will yield less than the acceptable level. Fewer animals In a test imply a lower ARD. In this case the ARD Is 1/3000 the test does as com pared with 1/600 in Example 1. TEST DOSE (t) AS A MULTIPLE CF THE '"ACCEPTABLE" RISK DOSE 14 S "ACCEPTABLE" RISK X 10' "ACCEPTABLE" RISK DOSE Appendix IV--R elationship Between Chemical Analysis, Biological Assays and Carcinogenic Risks to Man The following set of Figures presents a conceptual scheme for dealing with the problems stated. In practice any or all of the steps may l>e rendered Im possible by limitations in quantifying pertinent variables. Operational problems In Implementing the plan result from uncertainties in the correspondence of human and animal dose response curves, extrapolation of animal data to ex tremely low response levels, differences in response from species to species, synergistic effects, etc. The scheme itself may be useful, however, In defining the specific areas in which further efforts should be made. P roblem Xo. 1 : Given the P resence of a C hemical in the E nvironment. E stimate I ts Carcinogenic H azard for Max and Determine I ts Compati bility With a Socially "Acceptable" R isk The inputs and the estimations required for this problem are represented In Figure 1 and discussed in the following paragraphs. 1. Analytical Detectability in the Environment--Limited by the aensitlvtty and specificity of the available analytical method. ,, ___ 2 Quantitative Distribution in Time and Space in the Environment-Prepara tion of "environmental profiles" limited chiefly by extent of sampling. Segments of the environment should not be considered non-contributory, because tdej lau below the minimum levels for analytical detection. 15 PROBLEM NO. 1 - GIVEN THE PRESENCE O F A CHEMICAL IN THE ENVIRONMENT, ESTIMATE I I S CARCINO GENIC HAZARD J A R MAN AND DETERMINE I T S C O M P A T IB IL IT Y WITH A SO C IA L ],? "ACCEPTABLE" r a S K . IN P U T S : ANALYTICAL DATA E T O L O G IC A ! DATA ARBITRARY SELECTIO N S BOX NUMBERS REFER TO TEXT FIGURE I 8. Calculated Response In Animals Corresponding to a Dose Equal to Ex posure in Man--Obtained by calculation from Items 3 ancl 7. This provides the basis for the extrapolation from the response in animals to that in man in Steps 0 and 10. 9. Selected "Safety Factors" for the Extrapolation of Effects from Animals to Man--This selection is of necessity an arbitrary one, but it must be based on informed judgment. It is severely limited by our lack of specific information on effects in man, and should be made so as to avoid underestimating the risk. The choice of a safety factor of 100. frequently applied in cases of re versible toxicity, is inadequate to cover the complex assessment of carcinogenic risks in man. A factor should be applied for at least each of the following parameters: species susceptibility, interaction and potentiation effects, correc tion for animal size or body weight. 10. Calculated Projected Incidence in Man--Obtained by calculation from Items 3, 8, and 9, i.e., by taking the level of exposure In man. calculating the tumor incidence that this level of exposure would produce in the animal, and multiplying it by the total selected "safety factor". U . Social Selection of '`Acceptable" Risk for Man--This arbitrary selection is a determining element in the process of decision. It must be bused on the most extensive information available on the consequences of accepting a given risk and on the projected ''benefits" that would result. 12. (a) Determination of the Maximum Levels in the Environment and Extent of Exposure in Man Compatible with Socially "Acceptable" Risks--Compare Items 19 and 11: if projected incidence in man is greater than acceptable risk, determine what level of exposure in man (Item 3) would give an incidence in man (Item 10) equal to or lower than the acceptable risk (Item 11). Exposure can be reduced by changes in the environmental levels and distribution, or by reducing the number of people exposed. _ 13. Implementation of Corrective Measures, or Acceptance of Projected Risks-- Implementation of corrective measures necessary to reduce the projected risk to a socially acceptable level may be extremely difficult. In such eases, the extent of the risk should be clearly stated and a continuing effort developed towards control of the problem. PROBLEM NO. 2 : G O T X A PROPOSAL TO INTRODUCE A CHEM ICAL INTO TH E ENVIRON M ENT, E6T IM A T E T H E M AX IM U M EXPOSURE LEVEL KOR MAN COMPATIBLE W IT H A SOCIALLY " ACCEPTABLE" RISK The inputs and estimates required for this problem are represented In Figure 2. The same definitions used for Problem 1 apply. Two different calculations are involved, those in steps 14,10, and 12(h ). 14. Calculated Dose in Animals Corresponding to a Response Equal to the "Acceptable" Risk for Man--Obtained by calculation from items 7 and 11. 17 PROBLEM NO. 2 - GIVEN A PROPOSAL TO INTRODUCE A CiiEM ICAL INTO IT S ENVIRONMENT, ESTIM A TE THE MAXIMUM EXPO SU RE L E V E L FOR MAN CO M PATIBLE WITH A SOCIALLY "ACCEPTABLE" R ISK IN P U T S : ANALYTICAL BATA B IOLOGICAL DATA ARBITRARY SELECTIONS BOX NUMBERS REFER TO TEXT F I GORE I I 1. Factual information on the sensitivity of analytical methods and on the stent of environmental distribution and of exposure in man (Items 1, 2, and 3). 2. Factual information on the limits of the biological determination of car cinogenicity, represented by the sensitivity of bioassay methods, the significance levels of the results obtained, and the dose-response data (Item 4, 5). 8. Documentation of the criteria and procedures used for the selection of the arbitrary factors used in the evaluation process. These factors include the selection of a socially "acceptable" risk, of a procedure for dose-response extraie olations in the animal model and of -`safety factors" for the extrapolation from animal toman (Items (5,!), 11). Only in the case of substances for which we have extensive analytical en vironmental data and extensive bioassay results (probably including a doseresponse study of acceptable quality) can we attempt the exercise of estimating some dose corresponding to an "acceptable" risk for man. This can only be done by clearly stating Che justification for each of the arbitrary extrapolation factors' chosen. The selections will be justified only if they represent an upper limit of risk acceptance, that is, if they assure us that the risk in man is not underestimated. REFERENCES (!) Shubik I*. Clayson DB. Terracini B: The Quantification of Environ mental Carcinogens. tIC C Technical Report Series, Volume 4. Geneve, Switzer land, Braillard E, 1970. 33 p. . (2) Report of a WHO Expert Committee: Prevention of cancer. WId Hlth Org Techii Rep Ser 276 :l-53. 1964 (J) Fleming AS: Statement of Arthur S Fleming. Secretary, D H M \, In U) Mantel X. Urvan WR: "S afety testing of carcinogenic agents. J Nat Cancer Inst 27:455-47, 1K31 , , (5) FDA Advisory Committee on Protocol for Safety Evaluation: I anel on Carcinogenesis: Re[X>rt on cancer testing and the safety evaluation of food addi tives and pesticides. Toxicol Appl Pharmacol: Iin press), 1910 (6) Committee on Causative Factors of Cancer and Committee on Cancer Prevention: Report of symposium on potential cancer hazards from chemical additives and contaminants to foodstuffs. Acta Fn Int Cancer 13:179-193. 1957 (7) Fifth Report of the Joint FAO/WHO Exiiert Committee on Food Addi tives: Evaluation of carcinogenic hazards of food additives. Mid Hlth Org Techn Rep Ser 220 :l-32. 1961 (S) Food Protection Committee. Food and Nutrition Board: Problems in the Evaluation of Carcinogenic Hazard from Fse of Food Additive. Washington DC, Xational Academy of Sciences--National Research Council Publ No <49. 1960, 44 p: Cancer Res 21 :42D--*56. 1961 _ , . ,a (9) Berenblmn I: Carcinogenicity Testing. tIC(- Technical Report Series, Volume 2. Geneve, Switzerland, de Bursen G, 1969 56 p ( 10) Report U'" Euv (, . A Review. Cancer Res 16:728-742. 1956 ( 12) Clnysnn D B : Chemical Carcinogenesis. Boston, Massachusetts, Little Brown and Company. 1962. 467 p , ,, ____ (id) Hneper \VC. Conway \VD: Chemical Carcinogenesis and Cancers, bprmg- tleld, Illinois, Thomas OC, 1964. 744 p o t l i \ Weislmrger, JH, Weisburger F.K: Tests for chemical carcinogens. In Methods in Cancer Research. Volume I (Busch H. ed.b New \ork. New lorh, Academic Press Inc. 1967. pp 307-387 Tn<lne (15) Arcos JG. Argus MG. Wolf G: Testing procedures.^ In Chemical Induc tion of Cancer, Volume I (Arcos JG. Argus MG, Wolf G, eds.). New lorh, New Vork. Academic Press Inc.. 1968. pp 310--163 (16) Wogan GX: Personal communication .,,.,, ,, (17) Hartwell JL : Survey of Compounds Which Have Been fested for Car cinogenic Activity. 2nd ed., Public Health Serv Publ Xo 149. Washington DC. 19 US GoTt Print Office, 1051, 583 p; Shubik P, Hartwell, JL : Supplement 1 to the b o r e . Washington DC', l\S Govt Print Office, 1037, 388 p ; Peters J : Suppl a n t 2 to the a lw v e . Washington DC, ITS Govt Print Office, 1009, 055 p. NIH-77762