PAGENO="0001" CONTJ~OL OF CA~C~NOGI~NS lilk! TIF~IE ENVIf~WNR'AENT HEARING BEFORE THE STIBOOMIMIETTEE ON OOMIMIEIROE, TRANSPORTATION, Al~fD TOURISM OF THE OOMIMTTTEE ON ENERGY AND COMIMIERCE HOUSE OF REIIRESENTATIVES NINETY-EIGHTH CONGRESS FIRST SESSION -MARCH 17, 1983 Serial No. 98=31 Printed for the use of the Committee on Energy and Commerce U.S. GOVERNMENT PRINTING OFFICE 22-1430 WASHINGTON : 1983 PAGENO="0002" COMMITI'EE ON ENERGY AND COMMERCE JOHN D. DINGELL, Michigan, Chairman JAMES H. SCHEUER, New York RICHARD L. OTHNGER, New York HENRY A. WAXMAN, California TIMOTHY E. WIRTH, Colorado PHILIP R. SHARP, Indiana JAMES J. FLORIO, New Jersey EDWARD J. MARKEY, Massachusetts THOMAS A. LUKEN, Ohio DOUG WALGREN, Pennsylvania ALBERT GORE, JR., Tennessee BARBARA A. MIKULSKI, Maryland AL SWW]~, Washington MICKEY LELAND, Texas RICHARD C. SHELBY, Alabama CARDISS COLLINS, Illinois MIKE SYNAR, Oklahoma W. J. "BILLY" TAUZIN, Louisiana RON WYDEN, Oregon RALPH M. HALL, Texas DENNIS E. ECKART, Ohio WAYNE DOWDY, Mississippi BILL RICHARDSON, New Mexico JIM SLA~I'ERY, Kansas GERRY SIKORSKI, Minnesota JOHN BRYANT, Texas JIM BATES, California FrL&Ies~ M. Po~rmR, Jr., Chief Counsel and Staff Director SHARON E. DAVIS, Chief Clerk/Administrative Assistant DONALD A. WA'rr, Printing Editor ARNOLD I. HAvaNs, Minority Counsel SUBCOMMITTEE ON COMMERCE, TRANSPORTATION, AND TOURISM JAMES J. FLORIO, New Jersey, Chairman BARBARA A. MIKULSKI, Maryland NORMAN F. LENT, New York W. J. "BILLY" TAUZIN, Louisiana DON RITI'ER, Pennsylvania DENNIS E. ECKART, Ohio JAMES T. BROYHILL, North Carolina WAYNE DOWDY, Mississippi (Ex Officio) BILL RICHARDSON, New Mexico JOHN D. DINGELL, Michigan (Ex Officio) GREGORY E. LAWLER, Staff Director RICHARD C. FORTUNA, Toxicologist CI..krnE Wssrrres~y, Associate Minority Counsel JAMES T. BROYHILL, North Carolina NORMAN F. LENT, New York EDWARD R. MADIGAN, Illinois CARLOS J. MOORHEAD, California MA11~HEW J. RINALDO, New Jersey TOM CORCORAN, Illinois WILLIAM E. DANNEMEYER, California BOB WHI'ITAKER, Kansas THOMAS J. TAUKE, Iowa DON RITI'ER, Pennsylvania DAN COATS, Indiana THOMAS J. BLILEY, JR., Virginia JACK FIELDS, Texas MICHAEL G. OXLEY, Ohio HOWARD C. NIELSON, Utah (II) PAGENO="0003" CONTENTS Statement of: Anderson, Betty, Office of Health Effects, Office of Research and Develop- Page ment, Environmental Protection Agency 372 Browning, Jackson B., member, Board of Directors, American Industrial Health Council 281 DeMuth, Christopher C., Administrator for Information and Regulatory Affairs, Office of Management and Budget and Executive Director, Presidential Task Force on Regulatory Relief 356 Hernandez, John W., Deputy Administrator, Environmental Protection Agency 372 Keyworth, Dr. G. A., II, Science Adviser to the President, Director, Office of Science and Technology, Executive Office of the President 348 Neal, Robert A., Ph.D., president, Chemical Industry Institution of Toxi- cology 265 Nelson, Norton, Ph. D., professor Of environmental medicine and former director, Institute of Environmental Medicine, New York University Medical Center 69 Perera, Frederica P., Dr. P. H., senior staff scientist, Natural Resources Defense Council 96 Pitot, Henry C., M.D., director, McArdle Laboratory for Cancer Research and former chairman, National Cancer Advisory Board, University of Wisconsin-Madison 22 Silbergeld, Ellen K., Ph. D., chief toxics scientist, Environmental Defense Fund 157 Weinstein, I. Bernard, M.D., professor of medicine and public health, director of the division of environmental sciences, School of Public Health, Columbia University 3 Weir, Russ, Acting Director, Hazardous Sites, Environmental Protection Agency 372 Material submitted for the record by: National Cancer Advisory Board, report of the Subcommittee on Environ- mental Carcinogenesis 542 Commerce, Transportation, and Tourism Subcommittee, letter, dated Oc- tober 5, 1982 from Rita M. Lavelle to John .Hernandez re NTP study 415 (III) PAGENO="0004" PAGENO="0005" CONTROL OF CARCINOGENS IN THE ENVIRONMENT THURSDAY, MARCH 17, 1983 HOUSE OF REPRESENTATIVES, COMMITTEE ON ENERGY AND COMMERCE, SUBcOMMITTEE ON COMMERCE, TRANSPORTATION, AND TOURISM, Washington, D.C. The subcommittee met, pursuant to notice, at 10 a.m., in room 2123 Rayburn House Office Building, Hon. James J. Florio (chair- man) presiding. Mr. FL0RI0. The subcommittee will come to order. A number of the members are on their way. In the interest of conserving our time and going forward so as to accommodate all our witnesses today, we will convene at this time. I am pleased to welcome all in attendance and particularly our witnesses to our hearings today, which are designed to examine proposed changes to the Nation's policy on the control of carcino- gens in the environment. The subcommittee has legislative and oversight responsibilities for all of this Nation's hazardous waste and toxic substances pro- grams, including the superfund, the Resource Conservation and Re- covery Act and the Toxic Substances Control Act. While these stat- utes deal with a wide range of public health threats, they are also instructed to pay attention to the evaluation and control of carcino- gens in the general environment, particularly when those carcino- gens are involved in hazardous waste management and cleanup. Cancer occupies center stage in American concern about disease because of the toll that it takes in lives, suffering, and medical ex- pense. Cancer strikes one out of every four Americans, killing one out of five. It is the second leading cause of death, claiming over 400,000 lives in the United States each year. Studies over the last two decades have estimated that 60 to 90 percent of cancer is associated with the environment and is, there- fore, theoretically preventable. While many of these estimates in- corporate factors such as smoking and diet, under the heading of environment in its broadest sense, involuntary exposures to car- cinogens in the air, water, land, workplace, and consumer products have profound effects on the incidence of cancer. It is the prevention of involuntary exposures to those agents that remains our most productive avenue of preventive intervention. For example, the OTA-Office of Technology Assessment-has esti- mated that anywhere between 4 and 18 percent of total cancer oc- (1) PAGENO="0006" 2 currence in the United States is attributable to occupational expo- sure to asbestos alone. In recent months we have seen the emergence of several new proposals that substantially and, in certain cases, radically change the methods and protocols by which human risk from environmen- tal carcinogens are identified, evaluated, and controlled. These pro- posals and, in certain cases de facto changes in cancer policy, raise very serious concerns and questions regarding their impact on public health and the environment. I am deeply concerned that efforts to reevaluate cancer policy not be used as an excuse to define problems out of existence or to reduce our comment to protect the environment or to pervert sci- ence to achieve some political or economic goal. This morning we will examine these policies and the underlying scientific assumptions in great detail. In addition, the subcommit- tee will specifically examine the application and impact of these recent endeavors on the control of dioxin and formaldehyde in the general environment and the effect of these proposals on the con- trol of carcinogenic groundwater contaminants at hazardous waste sites. This is the first in a series of hearings on this subject. We plan to take a long and considered look at the subject through the course of this entire year. Congressman Henry Waxman, the Chairman of the Health and Environment Subcommittee of this full Committee, has expressed his interest in this area as well, and I am sure that he and I will be working together in a collaborative way on this very important area. I am pleased that the subcommittee has been able to assemble such a distinguished gathering of experts from the academic world, the environmental and the business communities, as well as the participation of three very high-ranking representatives from the administration. We will at this time call our first panel. It is made up of Dr. I. Bernard Weinstein, Professor of Medicine and Environmental Sci- ences at Columbia University, the Institute for Cancer Research; Dr. Henry Pitot, Director of the McArdle Laboratory for Cancer Research and former Chairman of the National Cancer Advisory Board; and Dr. Norton Nelson, Professor of Environmental Medi- cine of NYU. Gentlemen, if you would come forward, we welcome you to the committee. As with all of our witnesses, your statements will be made a part of the record in their entirety. You may feel free to proceed as you see fit. Dr. Weinstein. PAGENO="0007" :3 STATEMENTS OF I. BERNARD WEINSTEIN, M.D., PROFESSOR OF MEDICINE AND PUBLIC HEALTH, DIRECTOR OF THE DIVISION OF ENVIRONMENTAL SCIENCES, SCHOOL OF PUBLIC HEALTH, COLUMBIA UNIVERSITY; HENRY C. PITOT, M.D., DIRECTOR, McARDLE LABORATORY FOR CANCER RESEARCH AND FORMER CHAIRMAN, NATIONAL CANCER ADVISORY BOARD, UNIVERSITY OF WISCONSIN-MADISON; AND NORTON NELSON, PH.D., PROFESSOR OF ENVIRONMENTAL MEDICINE AND FORMER DIRECTOR, INSTITUTE OF ENVIRONMENTAL MEDI- CINE, NEW YORK UNIVERSITY MEDICAL CENTER Dr. WEINSTEIN. Thank you, Chairman Florio, members of the committee and guests. Public policies related to the regulation of potential carcinogens in our environment must be firmly rooted in scientific principles. I am pleased, therefore, to have this opportuni- ty to begin this session by briefly reviewing with you certain basic scientific principles of environmental carcinogenesis as I and my colleagues see them. As our Chairman, Mr. Florio, has emphasized, cancer is a disease of major magnitude. Advances are being made in the treatment of this disease, but many of us feel that there should be an intensifi- cation of efforts at the level of prevention. Indeed, in the general area of health, advances in prevention have often made the major strides in contrast to improved methods of treatment. We are encouraged by the evidence that a major fraction of cancer is preventable because it is due to exposure to external fac- tors rather than inherited or genetic factors. Advances in our un- derstanding of cancer causation are intimately tied to advances in basic cancer research, and indeed the major rate-limiting steps in rational approaches in this area remain our understanding of the fundamental basis of the process. At the same time, as in other areas of public health, we must apply the knowledge as it now exists and, as in other disease areas, prevention has often been fea- sible in the advance of complete understanding of the disease. As we all know, within the past few decades there has been a remarkable expansion in the Ŕhemical industry in terms of the number of new compounds synthesized and introduced into our en- vironment, many of which have highly beneficial effects. They have raised the standard of living. We must, however, at the same time have a comparable technology to monitor the potential health hazards of these substances. We know that several of them have been responsible for the cau- sation of human cancer, and we know that a number of them are carcinogenic in experimental animals. It is imperative, therefore, that as we move forward in this important area of technology we have the appropriate surveillance and regulatory system to limit human exposure to those chemicals which are potentially carcino- genic in humans. Now this task is a complicated one for several reasons. The first is the very long lag or latent period between exposure to a carcino- gen and the occurrence of cancer. In the human, this lag can be in the range of 5 to 30 or more years. This complicates considerably epidemiologic studies since one has to wait a long period of time, often decades, to know whether or not a newly introduced com- PAGENO="0008" 4 pound is a carcinogenic risk to humans, if we are going to rely only on the human experience. In addition, given the large number of compounds that are poten- tially introduced into the environment each year, prudence would dictate that we have predictive tests in advance of human expo- sure. A related principal is that carcinogenesis is dose-dependent. This is intuitively correct and most of the data supports this. The larger the dose, the more tumors that will occur, and often, the shorter the latent period between exposure and the time that the tumors occur. My reading of the evidence is that there is no clear evidence of a threshold for any carcinogen that we are concerned with with re- spect to human exposure. In risk extrapolation, therefore, it is rea- sonable to assume a linear dose-dependent, nonthreshold model for all types of carginogens unless there is direct evidence to the con- trary. To my knowledge, at the present time there is not direct evi- dence to the contrary for the sorts of substances that we are cur- rently concerned with respect to regulation. Another basic aspect of the carcinogenic process is the fact that it is a multi-step process occurring, as I have emphasized, over a long period of time, and that the evolution of this multi-step proc- ess is a function of exposure to multiple factors which can, in some cases, act synergistically. We define initiating agents as substances which act at the early stages of the process, and often these substances act by damaging cellular DNA, the genetic material. We define promoters as agents that act at later steps in the process. Many of the tumor promoters that have been studied thus far in the laboratory do not appear to act by causing direct damage to cellular DNA. Some of them bind to receptors and alter membrane structure and function, growth and differentiation. The initiating agents, as I have emphasized, do damage DNA, and this provides the basis for so-called short-term tests, simple laboratory tests in which one asks whether or not the compound is capable of damaging DNA or producing mutations or damaging chromosomes. These tests are highly valuable, although they do not in themselves establish whether or not a substance is a carcino- genic risk. Unfortunately, at the present time we do not have a battery of similar short-term tests to assay for tumor promoters. There are advances being made in this area and I would guess that within the next few years such tests will become available and will be validated. At the present time this means that our battery of pre- dictive short-term tests is not complete since it does not assay for tumor promoters. The precise mechanism of carcinogenesis at the DNA level is also not fully understood. There are now very exciting advances in cancer research related to oncogenes-genes which may be critical to the carcinogenic process-and should this prove to be the case, then this could revolutionize our approaches to genetic toxicology, to risk assessment, and to cancer prevention. I must emphasize, however, that this remains a speculative area of research and these advances are not directly applicable at the present time to assessment. Therefore, we must fall back upon PAGENO="0009" 5 more conventional and established approaches in the process of risk assessment. Now having reviewed briefly these principles, I would like to relate them to certain issues that have surfaced recently with re- spect to alterations in guidelines for risk extrapolation. These issues have surfaced in certain recent documents, including a memo to the Administrator of EPA on formaldehyde from Dr. John Todhunter, dated February 10, 1982, a paper on water-borne car- cinogens circulated by the EPA, and a paper from the Office of Sci- ence and Technology on cancer policy. The first issue has to do with the validity of animal bioassays and short-term tests in predicting potential human carcinogens. Al- though epidemiologic studies have played a key role in identifying a number of human carcinogens, they have certain inherent limita- tions. These include the following. They are largely retrospective, rather than predictive. That is, they depend upon the occurrence of cancer and so we obtain the evidence after the fact. They have lim- ited sensitivity. Generally the risk in the expo~cd versus the con- trol group must be 1.5 or greater to be significant. A 1.5-fold in- crease in a certain cancer would present a major public health hazard. In addition unless the epidemiologic studies are coupled with laboratory approaches, they often provide evidence for an as- sociation rather than the identification of specific causative factors. Because of these limitations, particularly the fact that they are retrospective rather than predictive, it is necessary to employ rodent bioassays and short-term laboratory tests in the assessment of potential human carcinogens~ Rodent bioassays have a long tra- dition. There is extensive experience with these procedures and there is reason to believe that the results obtained from such stud- ies are predictive of the human experience and must, therefore, be taken seriously~ Thus, almost all of the known human carcinogens-there are more than two dozen-are also positive as carcinogens in the rodent bioassays, thus establishing a correlation between the re- sponse of rodents and humans. Rodent carcinogens indeed have predicted or anticipated the carcinogenicity of several compounds which were subsequently proven to be human carcinogens. We should not assume, therefore, that any compound tested in rodents will be carcinogenic and, therefore, we can ignore the re- sults of such assays. Indeed, of the thousands of compounds tested, only a fraction are carcinogenic, in rodent bioassays, even though the list of compounds selected for testing was probably skewed in terms of the likelihood of compounds that were suspected of being toxic. We, of course, recognize that rodent assays have major limita- tions because of inter-species variations interns of metabolism of compounds, DNA repair, differentiation and other effects. Much is being learned about the metabolism of carcinogens and there is considerable evidence that the metabolism of several classes of car- cinogens in humans is qualitatively similar to that in rodents, al- though there may be major quantitative differences. I must stress, therefore, that when positive results of carcinogen- icity are obtained and judged to be valid within the experimental design of the rodent bioassay, and when confounding factors are PAGENO="0010" 6 eliminated, such results must be taken as strong evidence that the substance in question is likely to pose a carcinogenic risk to humans. Negative results in rodent bioassays obtained in one species or strain of animals do not invalidate positive results obtained in other species or strain because of known species and strain differ- ences. Unless there is evidence to the contrary, `it seems prudent to assume that humans may be as sensitive as the most sensitive animal strain or species tested. Again, I emphasize, unless there is convincing evidence to the contrary. I will not go into detail on the short-term tests. They are ex- tremely valuable, but I believe that at the present time they do not in themselves establish the carcinogenicity or lack of carcinogen- icity of a substance. They can be used for understanding mecha- nism and for establishing priorities in terms of which compounds deserve further testing or more extensive epidemiologic studies. Another issue which has surfaced in these recent documents is the question of whether different methods of risk extrapolation should be used for agents which are genotoxic-that is, agents which damage cellular DNA-versus agents which act by other mechanisms, for example tumor promoters. In a recent letter in Science (219, 794, 1983) I have listed six or eight reasons why I think this distinction is faulty. I will not go into these in detail but I think that it most prudent until we know more about the mechanism of action of initiators and promoters that we assume similar dose response relationships and not assume the existence of a threshold dose for either class of compounds. There has been the impression that tumor promoters are some- how safer and less potent. My colleague, Dr. Henry Pitot, will dis- cuss TCDD (dioxin) in greater detail. I simply want to stress that it appears to act in the carcinogenic process as a promoter and yet it is several orders of magnitude more potent than most of the con- ventional initiating carcinogens. Furthermore, a number of the substances which may act as pro- moters accumulate in the environment or in body fat stores and are not biodegradable. Therefore, they can cause repeated and pro- longed exposure to human tissues. I have emphasized the lack of evidence for the existence of a threshold for both initiators and promoters and I would stress that even if a simple experimental system, a threshold, had been established at a given level, it would be difficult to extrapolate with confidence that specific dose level, to a heterogeneous human population, a population which may also be exposed to other agents that could synergize with the effects of the compound undef consideration. In summary, therefore, although there has been exciting prog- ress in our understanding of the mechanism of action of environ- mental carcinogens, the field is in a sufficient state of flux that at the present time I believe it would be premature to alter certain well-established guidelines for assessing potential human carcino- gens. I think it is premature to alter the general guidelines for risk extrapolation to the human population. Thank you. [Testimony resumes on p. 22.] [Dr. Weinstein's prepared statement follows:] PAGENO="0011" I. Bernard Weinstein, M. D. Professor of Medicine and Public Health Director of the Division of Environmental Sciences of the School of Public Health Columbia University College of Physicians and Surgeons Basic Principles of Environmental Carcinogenesis Since public policies related to the regulation of potential carcinogens in our environment must be firmly rooted in scientific principles, I am pleased to have this opportunity to review with you certain basic principles of environmental carcinogenesis. I have recently reviewed this subject in considerable detail elsewhere and, therefore, I will only briefly list the major principles as I see them. 1. Cancer is the second leading cause of death in the United States accounting for over 1400,000 deat~hs per year. It is, therefore, a disease of major magnitude. Although progress is being made in cancer treatment, I believe that we must also intensify our efforts in cancer prevention. Many of the major advances in the control of other diseases have been made at the level of prevention rather than treatment. 2. The majority (50-80%) of human cancers are due to environmental (i.e., exogenous) rather than hereditary factors. They are, therefore, in principal preventable by identifying the causative agents and reducing human exposure to them. In certain cases it may also be possible to enhance host defense mechanisms, but at the present time this approach is largely theoretical. Advances in our understanding of cancer causation are intimately tied to advances in basic cancer research. PAGENO="0012" 8 3. Within the past few decades there has been a marked increased in the exposure of humans to ~y~~etic chemi and their various products, both in the workplace and the general environment. Although there is, thus far, no evidence that this has led to a marked increase in total cancer incidence, a number of these synthetic compounds are carcinogenic in experimental animals, and several have clearly been responsible for the causation of cancers in humans (i.e., aromatic amines, diethylstilbestrol, vinyl chloride, etc.). It is imperative, therefore, that we prevent unnecessary pollution of our environment by synthetic chemicals, and that we maintain a thorough surveillance and regulatory system to limit human exposure to chemicals which are potentially carcinogenic to humans. L~. There is a ~ lag, or, latent period between exposure to carcinogens and the clinical appearance of tumors. In humans this lag is in the range of 5-30 or more years. Carcinogens can even act transplacentally, with tumors appearing only later in the adult progeny. These aspects considerably complicate epidemiologic studies and emphasize the need for predictive laboratory assays. - 5. Carcinogenesis is dose ~pendent. Within a given range of doses, the larger the dose the greater the incidence of tumors. Increasing doses also often shorten the latent period. There is little or no evidence of a threshoi4 level, i.e., a level below which carcinogens produce no biologic effects. In risk extrapolations it is, therefore, reasonable to utilize a linear non-threshold model for all types of carcinogens. PAGENO="0013" 9 6. The conversion of a normal tissue to a ~ malignant cancer is a multist~p process. It seems likely that the cellular and biochemical events that underlie the early steps are different from those related to the later steps. Initiating agents specifically enhance the early steps, whereas tumor promoters enhance the later steps. When applied alone tumor promoters produce few or no tumors, but when applied following application of an initiating carcinogen they markedly enhance tumor development. Most, of the initiators are also complete carcinogens, i.e., when applied repeatedly, or in certain cases as a single large dose, they produce tumors by themselves. 7. Current evidence indicates that tumor initiators act by genera~4~g h~ghly reactive chemical intermediates, sometimes spontaneously but usually as a result of metabolism by the host. These intermediates, called ultimate carcinogens, form a chemical bond with cellular DNA. It is assumed that this modification of the genetic material leads to a mutation in the affected cell, i.e., a heritable change in DNA sequence. More complex heritable changes may also be involved, i.e., specific DNA or chromosomal rearrangements, activation of cellular proto-oncogenes, changes in DNA methylation, alterations in chromatin structure, etc.. On the other hand, the known tumor promoters do not act by binding directly to DNA. Their cellular targets appear to be cellular membranes or cytoplasmic receptors. Tumor promoters often stimulate cell proliferation and modify gene expression and cellular differentiation. For these reasons their effects are sometimes described as "epigenetic." Indirect effects of promoters on DNA and chromosomes have not, however, been excluded as important components PAGENO="0014" 10 in the process of tumor promotion. 8. In addition to specific tumor promoters, hormones and various cofactors can also enhance the carcinogenic process. The effects of these agents also do not appear to involve direct damage to cellular DNA. There are also important examples of synergistic interactions between chemicals and specific viruses in the causation of cancer in experimental animals and there is reason to believe that similar interactions might be responsible for certain human cancers (i.e., liver and nasopharyngeal cancer). Thus, in considering the potential hazards to humans of a given substance we must be concerned with possible synergistic interactions with other agents to which humans might already be exposed. Current Public Policy Issues Related to Carcinogens and Risk Assessment In the light of the above basic principles of carcinogenesis, I would now like to consider four current issues regarding public policies related to the regulation of potential carcinogens in our environment. These issues have surfaced in certain recent documents including: a memo to the Administrator of EPA on formaldehyde from Dr. John Todhunter, dated February 10, 1982; a paper on water-borne carcinogens, circulated by the EPA; and a paper from the Office of Science.and Technology, on cancer policy. The first issue has to do with the validity of animal bioassays and various short term tests in predicting potential human carcinogens. Although epidemiologic studies have played a key role in PAGENO="0015" 11 identifying certain human carcinogens they have certain inherent limitations. These include the following: 1) They are largely retrospective, rather than predictive, i.e., they depend on the occurrence of human cancer. 2) They have limited sensitivity, i.e., generally, the risk in the exposed versus control group must be 1.5 or greater to be significant. 3) Unless coupled with laboratory approaches, they often provide evidence for an association rather than identification of the specific causative factors. There are exciting opportunities to couple newly developed laboratory methods with epidemiologic studies, an approach we term molecular cancer epidemiology. This approach mayin certain cases circumvent extrapolation from data obtained in lower organisms but it requires development and validation. Because of the limitations of epidemiologic studies particularly the fact that they are retrospective rather than predictive, it~ is necessary to employ rodent bioassays and short term laboratory tests in the assessment of potential human carcinogens. Rodent bioassays. At present, the long-term (usually lifetime) - administration of the compound in~ question to laboratory animals, usually mice or rats, to see if it induces cancer is the most reliable laboratory method for determining whether a substance is carcinogenic. There is extensive experience with such assays, and their validity has been demonstrated in several ways (Table 1). Thus, almost all of the known human carcinogens are also positive as carcinogens in the rodent bioassays. Rodent carcinogen bioassays detected several compounds which were later proven to be carcinogenic in humans. PAGENO="0016" 12 We should not assume that almost any compound that is tested will be carcinogenic in rodents and that, therefore, we can ignore the results of rodent assays. Of the thousands of compounds tested, only a fraction are carcinogenic in rodent bioassays, even though the list of compounds chosen for testing is probably skewed in terms of its high probability of toxicity. Likewise, of 1140 pesticides tested, only 10 percent were positive. At the same time, animal bioassays present certain major problems. They are quite expensive and time consuming. In addition, the standard rodent bioassays provide data at only two dose levels,.' often require a high dose of the agent to give statistically significant results, and the data are obtained in only specific strains of two species, usually rats and mice. We also know that there can be marked variations between strains and species in terms -of their responses to specific carcinogens. In some cases this may be due to species differences in carcinogen metabolism. Nevertheless, it has been demonstrated that with certain classes of carcinogens cellular metabolism and DNA binding are qualitatively similar in rodents and in humans. It would be prudent, therefore, to assume that the metabolism by humans of a newly discovered compound is qualitatively similar to that in rodents, unless there is evidence to the contrary. - Despite the limitations of rodent bioassays, I must stress that when positive results for carcinogenicity are obtained, when, the results are reproducible, and when possible confounding factors are eliminated, such results must be taken as strong evidence that the substance in question is likely to pose a carcinogenic hazard to PAGENO="0017" 13 humans. Negative results obtained in one species or strain of animals do not invalidate positive results obtained in another species or strain because of known species and strain differences in response to various toxic chemicals. Short terre tests. Within the past decade, there has bee& a major advance in methodology for detecting potential carcinogens through the use of short-terre in vitro tests, i.e., assays that use simple biologic or biochemical markers as endpoints, rather than the induction of tumors in whole animals. The current major short-terra' tests are listed in Table 2. No single short-term test will detect all potential carcinogens. Therefore, a battery of short-term tests must be used. Such tests can play a valuable role in identifying which of the many compounds that enter our environment might be potential carcinogens, thus highlighting chemicals that deserve a high priority for rodent bioassays or epidemiologic studies. Negative results in these assays do not, in themselves, indicate that the compound is not carcinogenic. It should be stressed that almost all of the available short-term tests are designed to detect initiating carcinogens or complete carcinogens. There is a major need to develop a similar battery of tests to detect tumor promoters, various cofactors, and synergistic interactions. - 22-143 O-83--_..2 PAGENO="0018" 14 The second issue is related to whether or not separate methods of risk extrapolation should be used for agents that exert genotoxic effects (initiating agents) and those that appear to act by epigenetic mechanisms (promoters). Implicit in this proposed approach is the assumption that agents that might act via epigenetic effects are less hazardous. There are several reasons why I think that at the present time this is not a rational policy. These reasons include the following (for a more detailed discussion and specific references see Weinstein, I. B., Science 219: 7914, 1983). 1. We do not know with certainty that certain carcinogenic -. agents act exclusively through genotoxic mechanisms and others through epigenetic mechanisms. Leading cancer researchers differ on this question. Therefore, much more basic research is required before this distinction becomes the basis of public policy. 2. Current laboratory methods for predicting whether or not a substance will be genotoxic in humans need refinement and validation. Furthermore, as stressed above, we do not at the present time have well validated short-term tests for tumor promoters. Therefore, identification of these agents would often be done on the basis of negative evidence. Furthermore, a given agent may produce multiple effects in the intact organism and also interact synergistically with other environmental factors. 3. Certain tumor promoters produce a low but significant incidence of tumors in the absence of application of an initiating agent. It is also possible that tumor promoters produce genotoxic effects through indirect mechanisms, some of which may not be revealed in the standard short terms tests. PAGENO="0019" 15. ~. It is often assumed that tumor promoters are more likely to show a threshold effect than initiating agents. There is, however, little or no data to support this~ assumption. 5. Although the known promoters require repeated application to exert their tumor-promoting effect, this does not necessarily imply a comfortable margin of safety. With many of the substances that are of concern (such as water pollutants, industrial chemicals, hazardous wastes, and food additives) there is likely to be repeated and prolonged human exposure. Moreover, some of these substances are only slowly degraded and, therefore, will persist or even accumulate in -the general environment or in body tissues. 7. There is the impression that tumor promoters are much less potent than initiating carcinogens. This is not necessarily the case. On a molar basis certain phorbol ester and indole alkaloid tumor promoters are about two orders of magnitude more potent in exerting - biologic effects than the initiating carcinogen benzo[a)pyrene. TCDD (dioxin) is about four orders of magnitude more potent than benzo[a]- pyrene and yet it does not appear to act by binding to DNA. 8. We know that nature has evolved specific defense mechanisms against some of the genotoxic agents, including conjugation and detoxifying mechanisms and DNA excision repair. We do not know to what extent humans have evolved protective mechanisms against tumor promoters. 9. A final reason for being concerned about the potential health hazards of tumor promoters and various carcinogenic cofactors that do not appear to act by directly damaging cellular DNA is the evidence that a major fraction of human cancer is due to "lifestyle factors" PAGENO="0020" 16 and that many of these may not act as simple genotoxic agents. The third issue is the general question of the existence of a threshold level for certain carcinogens in humans. As stressed above, although the response to various types of carcinogens is likely to be dose dependent, I know of no evidence that clearly establishes a threshold level for any carcinogen. Furthermore, even if this were established in a given experimental system, it would be difficult to predict with confidence the threshold level in a heterogeneous human population. It is prudent, therefore, not to assume the existence of a threshold dose for any carcinogen. Obviously, it will not be possible to completely remove all carcinogens from our environment. Nevertheless, we should attempt to reduce human exposure to the lowest feasible levels. The fourth issue is the attempt to utilize structure-~vi~y relationships (the SAP approach) to predict the carcinogefl~p~ or lack of ~ ciuneriicit of ~ synthesized compounds, by simply examining their chemical structures, physical properties and chemical reactivities. Although remarkable advances have been made in understanding the chemistry of certain carcinogens, our knowledge of carcinogen metabolism, and of factors that influence DNA binding and repair is too fragmentary to make such predictions reliable. Moreover, at the present time the diversity of potential cellular targets for tumor promoters and various cofactors, and the lack of a unifying principle for understanding the mechanism of multistage carcinogenesis, makes it impossible to predict, from chemical data alone, tumor promoting activity across a wide range of compounds and species. Although there has been exciting progress in our understanding of the mechanism of action of environmental carcinogens, the field is in a sufficient state of flux that at the present time it would be premature to alter the existing, well_established guidelines for detecting potential carcinogens and for risk extrapolations to the human population. PAGENO="0021" 17 TABLE 1 How Reliable are Animal Bioassays? 1. Almost all of the known "human" carcinogens are also carcinogenic in rodents. 2. Hone of the approximately 1~43 "rodent" carcinogens have been proven to be noncarcinogenic in humans. 3. Animal assays have predicted several human carcinogens-- aflatoxin, 14-amino-biphenyl ,bis(chloromethyl)ether, diethyistilbestrol, melphalan, and vinyl chloride. 4. For several carcinogens, the tissue speci~ficity in rodents is the same as in humans, but this is not always the case. TABLE 2 Methods for Detecting Carcinogens IN VIVO 1. Clinical observations--by astute physicians and patients. 2. Epidemiologic studies. 3. Experimental animal bioassays. SHORT-TERM TESTS 1. Mutagenesis assays: A. Bacteria -- Ames test. B. Mammalian cell cultures. C. Other eukaryotes -- yeast, Drosophila, mice. 2. Assays for cell transformation. 3. Assays for DMA binding, damage and repair -- and binding to other macromolecules. ~4. Assays for chromosomal abnormalities and sister chromatid exchange. 5. Inclusion of systems for metabolic activation. PAGENO="0022" 18 For a more detailed review of subjects discussed in this statement see: 1. Weinstein, I. B. Current Concepts and Controversies in Chemical CarcinogenesiS. J. of Supramol. Structure and Cellular Biochem. 17:99-120, 1981. 2. Weinstein, I. B. Molecular and cellular mechanisms of chemical carcinogenesiS. In: Cancer and Chemotherapy, Vol. 1. Introduction to Neoplasia and Anti-Neoplastic Chemotherapy, Crooke, S. and Prestayko, A. (Eds), Academic Press, NY, pp 169-196, 1980. 3. Weinstein, I. B. The scientific basis for carcinogen detection and primary cancer prevention. Ca-A Cancer Journal for Clinicians, American Cancer Society, Vol. 32, No. 6, pp 3148-362, 1982. 14. Weinstein, I. B. Carcinogen policy at EPA. Science 219: 7914-796 (Letters), 1983. PAGENO="0023" 19 The following is a list of specific references related to these subjects: 1. Ames, B. Identifying environmental chemicals causing mutations and cancer. Science 204: 587-593, 1979. 2. Doll, R. and Peto, R. The Causes of Cancer, Oxford University Press, Oxford, 1981. 3. Cancer Testing Technology and Saccharin, Office of Technology Assessment, Library of Congress #77-600051, U.S. Government Printing Office, 1977. 4. Clayson, D.B. Principles underlying testing or carcinogenicity. The Cancer Bulletin 29: 161-166, 197~7. 5. Davis, D.K. and Magee, B.H. Cancer and industrial production. Science 206: 1356-1358, 1978. 6. Fishbein, L. Potential Industrial Carcinogens and Mutagens. U.S, Environmental Protection Agency, (Publ 560/5-77 005) Washington, D.C., 1977. 7. Fisher, P.8. and Weinstein, 1.8. Chemical-viral interactions and multi- step aspects of cell transformation: In: Molecular and Cellular Aspects of Carcinogen Screening Tests, Montesano, R., Bartsch H., and Tomatis L. (Eds.) IARC Scientific Publications 27: 113-131, 1979. 8. Fisher, P.8. and Weinstein, 1.8. In Vitrb Screening Tests for Potential Carcinogens. In: Carcinogens in~Industry and Environment, Sontag, J.M. - (Ed.), Marcell Dekker, pp. 113-166, 1981. 9. Foulds, L. Neoplastic Development, Academic Press, Volume I, London and New York, 1969. 10. Gehring, P.J., Watanabe, P.G. and Blau, G.E. Risk Assessment of Environmental Carcinogens Utilizing Pharmacokinetic Parameters, Ann. N.Y. Acad. Sci. 329: 137-152, 1979. 11. General Criteria for Assessing the Evidence for Carcinogenicity of Chemical Substances, Report of the Subcommittee on Environmental Carcinogenesis, National Cancer Advisory Board, June 2, 1976. 12. Harris, C.C., Trump, B.F., Grafstrom, R. and Antrup, H. Differences in metabolism of chemical carcinogens~ in cultured human epithelial tissues and cells, In: Mechanisms of Chemical Carcinogenesis, Harris, C.C. and Cerutti, P.A., (Eds.), Alan R.~ Liss, Inc., New York, pp. 239-243, 1981. 13. Hiatt, H.H., Watson, J.D. and Winsten, J.A. (Eds.) Origins of Human Cancer, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1977. PAGENO="0024" 20 14. Hoel D.C., Gaylor, D.W., Kirschtein, R.L., Saffiotti, U. and Schneiderman, M.A. Estimation or risks of irreversible delayed toxicity. JI. of Tox. and Env. Health I, (1), 133-151, 1975. 15. Hollstein, H., McCann, J., Angelosanto, F.A. and Nichols, W.W. Short-term tests for carcinogens and mutagens, Mutation Res., 65: 133-226, 1979. 16. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Supplement 4; Chemicals, Industrial Processes and Industries Associated with Cancer in Humans. IARC, Lyon France, 1982. 17. IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Man. Volumes 1-29, [ARC, Lyon France. 18. Marx, J.L. Do tumor promoters affect DNA after all? Science 219, 158-159, 1983. 19. Maugh, T.H., Chemical Carcinogens: How Dangerous are Low Doses?, Science 202: 37-41, 1978. 20. Miller, E. Some current perspectives on chemical carcinogenesis in human and experimental animals: Presidential Address, Cancer Res., 38: 1479-1496, 1978. 21. Nago, H., Sugimura, T. and Matsushima, T. Environmental Mutagens and Carcinogens, Ann. Rev. Genet. 12: 117-159, 1978. 22. National Cancer Institute, Carcinogenesis Technical Report Series, No. 1, Guidelines for Carcinogen Bioassay in Small Rodents, DREW Publ. No. (NIH) 76-801, 1976. 23. Perera, F. and Weinstein, I.B. Molecular Epidemiology and Carcinogen-DNA Adduct Detection: New Approaches to Studies of Human Cancer Causation, J. Chron. Dis, 35: 581-600, 1982. 24. Rall, D.P., Validity of extrapolation of resultsof animal studies to man, Annals of N.Y. Acad. Sci. 329: 85-91, 1979. - 25. Schneiderman, M.A., Decoufle, P. and Brown, C.C. Thresholds for environmental cancer: biologic and statistical considerations. Ann. N.Y. Acad. Sci. 329: 92-130, 1979. 26. Slag, T.J., Sivak, A. and Boutwell, R.K., (Eds.) Carcinogenesis, Vol. 11, Mechaniams pf Tumor Promotion and Cocarcinogenesim, Raven Press, N.Y. 1978. 27. Starr, C. and Whipple, C. Risks of Risk Decisions, Science 208: 1114-1119, 1980. 28. Tomatis, L. The predictive value of rodent carcinogenicity tests in the evaluation of human risks, Ann. Rev. Pharmacol. Toxicol. 19: 511-530, 1979. 29. Tomatis, L., Breslow, N.E. and Bartsch, H. Experimental studies in the assessment of human risk. In: Cancer Epidemiology and Prevention, D. Schottenfeld and J. Fraumeni, (Eds.), in press, 1980. 30. Van Ryzin, J. Quantitative Risk Assessment, Jl. of Occup. Medicine 22: 321-326, 1980. PAGENO="0025" SCIENCE. VOL 2I~ Lettei~1 Marshall ~ very disturbing trend in government reg- ulatory policies, namely the attempt to establish separate guidelines for evaluat- ing the health eftects of "genotoxic" and "epigenctic" carcinogens, with empha- sis on sofLening the restrictions for the latter class of agents. The distinction between these two classes of agents is largely theoretical and has no factual. basis in terms of our current knowledge of mechanisms of action of carcinogens, for several reasons. I) We do nol know with certainty that certain carcinogens act through geno- toxic mechanisms and others through epigenetic mechanisms. Indeed, recent studies in molecular genetics, develop- mental biology, and immunology tend to blur the classical distinctions hetsseen genetic and epigenetic mechanisms, even in normal biologic processes (I). 2) Eves if this distinction were true, our current methods for assessing whether or not a given agent is likely to be genotoxic in humans have very seri- ous limitations (2); and what is worse, at the present time we do not have well- validated short-term tests for assessing agents that might act through nongeno- toxic mechanisms, that is, tumor pro- moters, hormones, and so forth. Identifi- cation of the "epigenetic" agents must, therefore, often be done by exclusion, a risky approach. 3) Most of the known* carcinogens I produce multiple effects. In fact, when given at sufficient dosage the genotoxic chemicals are usually complete carcino- gens and, therefore, probably produce both tumor-initiating and tumor-promot- ing effects (3, 4). Simple tests for geno- toxicity may fail to assess the promoting capacity of these compounds. This, and other factors, severely limit attempts to predict the mechanism(s) of action and relative potencies of carcinogens, `shen findings based simply on genotoxic ac- tisity are used. The paradigm of random point mutation as a basis for understand- ing the carcinogenic action of agents that display genotoxic effects nsa) itself be antiquated, in view of the multistage aspects of the carcinogenic process, probable 5) nergistic (and sometimes in- hibitory) nsultifactor interactiOns, and the possibility that carcinogcncsis in' Berecemen 1. 1. B. Weinstein. I. Snpvont't. S:i:e:. Cell. Rio- chen. 17.99(1981). 2. P. B. Fisherund 1. B. \`.einstein. in Coeci'oges in Iethsuey :edE,e,'imn,nvot. 3. M. Soeug. Ed. (Dnkker. Non Yock. t98t). p. 113: M. Holstcn. ~J. McCnnn. F. A. Angelounto. w. w. Nichols. Mum. Rm. 65. 3 (1979). -. - 3. T. 3. Slog:. A. Si:nk. R. K. Bvstnefl Edt.. Co:dnocenesi.e. sot. 2. Mevh.:ei:e,o of lot-nor ?sue'th;e oedCo:u'eie-gv~vei:(Rooee. Neon Ye,).. 978). 4. V. to:eoo'ii ond t. B. SVc)n:tcin. (ude:geor. Si: 3. 705 (1991). 3. A. R. Kinoell.s and H. R.odo:.,v. Peas. No). Avon!. Sn!. U.S.A. 75. 6)49(1575). H. Na1o'.:o'a end I. B. Lie):. ibid. 76. 9)5 1)979); H. C. Bino9oine. Seine 215. 1247 1952). 6. B. Doll and R. Fo:o. The Ca-toes of Cone,, (Osfoed Unis. rams. Oxf'vd. 1991). 21 [ a more coniplex genomic changes ~8)W ow that nature has cvolved~ (gene rearrangements, chromosomal specific defense mechanisms against transloeation, oneogene activation, al- some of the genotoxic agents, including tered DNA methylation, and so forth) (3). conjugation and detoxifying mechanisms 4) Certain tumor promoters (such as I *I4 DNA excisios rep.ir. We do not the phorbol esters and, TCDD) can in- , L!~'~░ what extent humans have duce a significant number oftumors in evolved protective mechanisms' aT~riei animals, even without prior application tumor Fomotera. I do aot doubt that of an initiating carcinogen (3). In addi. `such me~hanisms essist, but at the pies- tion, thereare a few studies suggesting ent time we do not know their properties that, although the primary target of the or relative efficiencies. phorbol ester tumor promoters is cellular 9) A final reason for being concerned membranes rather than DNA (I), these about the potential health hazards of compounds `may indirectly inflict chro- tumor promoters and v9rious eareino,- mosomal damage, perhaps via the gener- `genie cofactors that do not appear to act ation of activated forms of oxygen (5), If by directly damaging cellular DNA is the this is the case,~then these compounds evidencethat a major fraction of human also have genotoxic activity, albeit cancer is due to "lifestyle factors" and through an indirect effect, that many of these may not act as simple 5) It is often assumed that'tumor pro genotoxic agents (6). It is essential,- moters and other agents that might act therefore, that we not overemphasize through epigenetic mechanisms-will, in- our concern with genotoxic agents, contrast to initialing and genotoxic car- downplay the potential health hazards of cinogens, display a threshold in their other types of agents, and thus distort dose response. The data on dose-re- priorities in our efforts at primary cancer sponse relationships with tumor promot- prevention. ers are skimpy, and I know of no cvi- In summary, although there has been dence that clearly establishes a threshold `-exciting progress in our understanding of for tumor promoters in humans or in the mechanism of action of environmen- experimental systems. Even if this were tal carcinogens (1), the field is in a suffi- the case, how would we know how to cient state of flux that at the present time extrapolate from a specific set of data the it would be premature to alter the exist- actual threshold level in a heterogeneous ing, well-established guidelines for risk human population? - . ..~. extrapolations of potential hazards to the 6)Ifis true that the known tumor human population. Specifically, I see no promoters require repeated application justification for assuming a nonlinear to exert their tumor-promoting effect, dose response and threshold model for whereas the single application of certain . certain carcinogens simply because they initiating carcinogens is sufficient (3). do no) give a positive response in certain This does not necessarily imply a com- currently used assays for gcnotoxicity. fort.able margin of safety for tumor pro - I. BERNARD V/EINSTEIN- moters, because for many substances Division of Ensironrnenial Sriences, that are of concern (such as waterpollut- School of Publir Health, and ants, industrial chemicals, and food addi- Cancer Cenzer/Instiruze of canrer - lives) there is likely to be repeated and Research, ~olu,nhia University, prolonged human exposure. Moreover, New York 10032 some of these substances are only slowly degraded and, therefore, will persist or esen accumulate in body tissues or the general environment. 7) There is the impression that tumor promoters are much less potent than initiating carcinogens and, therefore, are less hazardc~us. This is not necessarily the case. On a molar basis TPA is about two orders of magnitude more potent in exerting biologic effects than bÚnzo[a)- pyrcne, and TCDD is about four orders of magnitude more potent than bcnzo(a]- pyrene (3). - . I PAGENO="0026" 22 Mr. FI~oRIo. Thank you very much. Dr. Pitot. STATEMENT OF HENRY C. PITOT, M.D., PH.D. Dr. PrroT. Thank you, Mr. Florio, members of the subcommittee, ladies and gentlemen. I appreciate the privilege of testifying on areas of our own and other's research which we feel may be of po- tential importance to the present deliberations of the subcommit- tee. Cancer is the most fearful disease of our present society. Unfor- tunately, in some sectors of our society the fear of cancer has truly become a cancer of fear. Although better educational efforts con- cerning the nature of the disease may ameliorate some of these concerns, only the ultimate control of the disease through preven- tion and cure can overcome these fears. I believe today's hearing is concerned with the prevention of cancer and wish to speak to that point. Although cancer is a dis- ease, is many things to many people, we have begun to learn there are certain common denominators in the development of what we call the natural history of the disease. My colleague, Dr. Weinstein, has discussed a number of aspects of the natural history of neoplasia with you, including the fact that the first beginnings of cancer appear to start in a small number of cells within the body. Such cells we have termed initiated cells. For the most part, these few and rare initiated cells are extremely diffi- cult to recognize and identify by available techniques. That is, we can only know they exist in that at some future time one or more of these cells will develop into cancer. While there is evidence that many vertebrate organisms possess a number of such initiated cells, it is likely that only rarely does one of these cells progress to cancer. A variety of agents in our en- vironment, as Dr. Weinstein has mentioned, are capable of initiat- ing cells. These include such chemicals as 2-naphthylamine, which caused so much bladder cancer in workers in the last and early part of this century, gamma radiation, which induced a number of different types of cancers in survivors of the atomic bomb blasts in Hiroshima and Nagasaki, and some viruses which as yet have been best studied in lower animals. At very low doses, many of these agents only appear to initiate cells, but at higher doses, especially when given repeatedly, such as with a number of chemicals, the initiated cells are continually stimulated by repeated administration of the agent until a few of their descendents become malignant. This stimulation, as Dr. Weinstein has mentioned, has been termed tumor promotion. We now know that there are a number of chemicals that appear to exert their effects in the natural history of cancer development during the promotion phase. Examples of chemicals that are in this category, as far as the human is con- cerned, include many of the hormones which naturally occur within our bodies, constituents of the diet, especially dietary fat, certain drugs and probably asbestos and alcoholic beverages. Examples of chemicals that are in cigarette smoke itself contains many promoting agents. But it also has many agents that initiate cells and, for this reason, we call cigarette smoke a complete car- PAGENO="0027" 23 cinogen. Most cancer-causing chemicals that we presently know of and which cause cancer after prolonged administration are com- plete carcinogens and possess the ability both to initiate and to pro- mote cells to cancer. Theoretically, to some extent there might also exist another group of chemicals which can initiate cells but possess no promot- ing action and such agents have been designated as incomplete car- cinogens. Therefore, in keeping with this classification and the points that Dr. Weinstein has raised, it appears that both complete and incom- plete carcinogens exert their effects by altering the structure or chemically reacting with the genetic material DNA. Promoting agents, on the other hand, do not appear to have the capacity to initiate cells, and do not appear to interact chemically directly with DNA, although recent studies suggest that at some stage of the interaction of some of them with the cell, they may induce re- actions which indirectly alter DNA structure. A major characteristic of all known promoting agents is that they effect one or more changes in the expression of the DNA itself. In very simplistic terms, then, initiators damage DNA and promoting agents may force the expression of the altered DNA. And as Dr. Weinstein has pointed out, can promoting agents cause cancer? Yes. Since they can promote cells initiated by a vari- ety of means, if not direct exposure to low levels of complete car- cinogens, then common environmental factors such as background radiation, dietary contaminants, and a number of others. The agent under question today, or one of the agents, 2378, te- trachlorodibenzodioxin, or more commonly known as dioxin, has been shown in our laboratory, and that of my colleague, Dr. Poland, to be a highly effective promoting agent, both for cancer production in rodent liver and in rodent skin, promoting carcino- mas of each tissue after appropriate initiation. Dr. Poland has also demonstrated that within the sensitivity of our methods, dioxin does not chemically react with DNA nor, as other workers have shown, is it significantly mutogenic, at least as evidenced by present methodologies. Therefore, to date, it has not been possible to demonstrate that dioxin is capable of initiating cells. On the other hand, in rodent liver initiated with a potent carcinogen, dioxin is effective as a pro- moting agent at levels between 1 and 10 micrograms, and to put this in perhaps more realistic terms, a millionth of a gram or one thirty-millionth of an ounce. This is the total dose which is given over a 7- to 8-month period, and which effectively promotes cancer in this tissue. Dr. Poland and his associates have demonstrated that similar doses of dioxin will promote skin cancer in hairless mice. He has pointed out that promotion by dioxin in this model may be relevant to humans because the reaction produced by the agent in the skin of the hairless mouse is similar to that evoked by this and related agents in human skin. Although, as Dr. Weinstein has pointed out, there is some discus- sion as to whether promoting agents exhibit a threshold or no effect level, we have as yet been unable to demonstrate any such PAGENO="0028" 24 threshold for dioxin as a promoting agent for liver cancer in the rodent. Furthermore, even if such a threshold or no effect level exists for dioxin, the relatively high degree of retention of this agent within the body, such as is also seen with asbestos, probably tends to in- crease its effectiveness as a promoting agent. In at least one study by Dr. Kociba and his associates, long-term, that is, 2 years chronic administration of extremely low levels of dioxin, similar to what I mentioned earlier, as would be expected, does result in the appearance of an increased number of malignant cancers in the liver, nose, and lung of the animals. Therefore, while significant experimental evidence suggests that promoting agents act quite differently from complete carcinogens in the natural history of cancer development, such arguments may not have much reference to dioxins whose action in the second stage of cancer development is so extremely effective in at least two mammalian species. Although different species react to the toxic effects of dioxin quite differently, with the primate appearing to be much less sensi- tive than the rodent, one cannot equate toxicity with the promot- ing aspect of carcinogens, and until we know a great deal more about the carcinogenic potential of dioxin, or its actual role in the causation of human disease, including cancer, prudence indicates that we should take great care in handling this material. Thank you, sir. [Testimony resumes on p. 69.] [Attachments to Mr. Pitot's prepared statement follow:] PAGENO="0029" 25 191 Biochimica er Biophysica Acre, 605 (1980) 191 -215 ę Elsever/North.Ho~and Biomedical Press BBA 87079 THE STAGES OF INITIATION AND PROMOTION IN HEPATOCARCINOGENESIS * HENRY C. P1TOT and ALPHONSE E SIRICA McArdle Laboratory for Cancer Resea,-ch, The Medical School, Departments of Oncology and Pathology, University of Wisconsin, Madison, WI 53 706 (USA.) (ReceivedJujy 24th, 1979) Contents I. Introduction 192 A. Stages in tumor development - historical aspects 192 IL Definitions and concepts of the two-stage process of caicinogenesis 193 A. Characteristics of initiation and promotion during skin carcinogenesis -193 B. Identification end characterization of stages during carcinogenesis in tissues other than skin 194 C. Definitions of initiation and promotion 196 III. The natural history of hepatocarcinogenesis - stages in the formation of hepatic carcino- mas 197 A. `Preneoplastic lesions' identified in experimental hepatocarcinogenesis and their rela- tionship to stages in this proceaa 197 B. The experimental definitionof stages in hepatocarcinogenesis 201 1. ModifIers of hepatocarcinogenesis 201 2. Biologic demonstration of the stages occurring during hepatocarcinogenesjs . . . . 202 a. Promoting agents for hepatocarcinˇgenesjs 203 b. Systems demonstrating multiple stages during hepatocarcinogenesis 205 c. Evidence for the sequential relationship of enzyme-altered foci to hepatocellu. lar carcinomas 206 3~ Analogies between the stages of carcinogenesis in liver and in skin 206 IV. Mechanisms of tumor promotion during hepatocarcinogenesis 208 V. Applications of the two-stage process of hepatocarcinogenesjs to problems in human can- cer 209 A. Stages in human liver cancer 209 B. Implications of the two-stage process. of hepatocarcinogenesis in the monitoring an~ regulation of potential human carcinogens, - 210 Vt. Conclusions - --- 211 References 212 * This paper is dedicated to Dr. Harold P. Rusch on the occasion of his retirement from the faculty of the University of Wisconsin-Madison. Dr. Rusch was the founder and long-time Director of the McArdle Laboratory for Cancer Research and was responsible for creating its productive research atmosphere. Abbreviation: TCDD, 2,3,7,8-tetrachlorodibenzo-p-dfoxin. PAGENO="0030" 26 192 I. Introduction It has been known for many years that a latent period exists between the administra-(~ tion of a carcinogen to an organism and theactual appearance of pathologically or cm- ically manifested neoplastic disease. This latency, which can be readily demonstrated fol- lowing treatment with chemical carcinogens, has been shown to occur even when the car- cinogen is continuously administered to an experimental animal [1]. Furthermore, the latency period prior to the appearance of neoplasia was found to follow the administra- tion of chemical carcinogens in utero as well as to the neonatal animal [2]. Thus, the latency period of the carcinogenic process may be considered to be a general feature of the natural history of neoplasia, although the time until tumor development from first contact with the carcinogenic agent varies with the type of agent itself, with its dosage, and with certain characteristics of the target cells. This review will be concerned with the mechanisms for the latency period in the pro- cess of hepatocarcinogenesis. Earlier reviews [3-6] have already dealt with the biology and biochemistry of stages in the process of carcinogenesis. as well as with co-carcino- genesis, tumor promotion, tumor progression and other aspects of the natural history of neoplastic development. However, since an understanding of the relationship between the natural history of the development of hepatic carcinomas has been influenced by studies of other organ systems, a brief recapitulation of the more general aspects of the stages in tumor development will be presented first. IA. Stages in tumor development - historical aspects Our understanding of the latency phenomenon in the natural history of carcinogenesis was pioneered by Mottram [7] and greatly increased by the studies of Berenblum and Shubik [8,9] and later by those of Boutwell [4], van Duuren [10], Hecker [11], and others [12,13]. Rous and his associates [14,15] were among the first to provide experi- mental evidence suggesting a two-stage mechanism for carcinogenesis in skin. These inves- tigators demonstrated that local application of coal tar to the ears of rabbits for a period of time, followed by wounding with a cork borer, resulted in the appearance of tumors growing along the edge of the wound. In this respect, Rous used the terms `initiate' for the process resulting from the tar application and `promote' for the function of the wounding. In subsequent studies, Berenblum and Shubik [8,9] clearly demonstrated that skin carcinogenesis could be divided into a stage of initiation as produced by the direct action of a carcinogen on the target cell followed by a second and longer stage of promo- tion which was brought about by the repeated application of a promoting agent. This latter agent was by itself considered to be noncarcinogenic, but possessed the property of enhancing or stimulating neoplastic expression in the initiated celL More recently, Bout- well [161 has proposed that skin tumorigenesis may be divided into three different stages. According to this proposal, the carcinogenic process was operationally defined to include initiation, which involves a specific and irreversible intracellular event, followed by pro- motion, which in turn could be subdivided into two separate phases: those of conversion and of propagation. Conversion defined the phase of promotion whereby the initiated cell becomes converted to a dormant neoplastic cell and propagation referred to the phase whereby these dormant cells are stimulated to proliferate. A more general concept which extends the two-stage mechanism is that of tumor pro- gression as described by Foulds [17]. While this author has proposed that the entire PAGENO="0031" 27 193 natural history of carcinogenesis encompasses the phase of progression, an alternate inter- pretation is that tumor progression is a' phenomenon which occurs after the formation of a distinct neoplasm. Pitot [18] has suggested that tumor progression is dependent on karyotypic changes in the neoplasm, implying that tumors with normal karyotypes have not yet entered this latter stage of their natural history. For a more extensive discussion of the historical aspects of the stages in tumor devel- opment, the reader is referred to the papers by Scribner and SŘss [19] and by Boutwell [4]. II. Definitions and concepts of the twostage process of carcinogenesis At the present time much of our knowledge of the stages of carcinogenesis is based on experiments concerning the genesis of epidermoid carcinoma in the mouse. However, as we learn more of the characteristics of specific stages in carcinogenesis of other tissues, it becomes apparent that the definition of ea˘h stage as first formulated in the skin has a broader application to the neoplastic process in many other tissues and organ systems. liii. Characteristics of initiation and promotion during skin carcinogen esis Based largely on studies in mouse skin, it is possible to delineate specific qualitative differences between the two processes of initiation and promotion. These differences, in turn, can then serve to provide a framework for a more general and complete definition of each of the distinct stages of carcinogenesis. Some qualitative differences in the stages of initiation and promotion in epidermal carcinogenesis are summarized in Table I. One of the most important characteristics of initiating agents is that a cell, once initi- ated, does not lose this induced property with time. Thus, the initiating agent evokes an irreversible change in one or more of the cells of the target tissue. Some agents. such as urethane, have been found to be capable of inducing neoplasms in the liver and lung, but not in the skin [20]. However, after urethane administration, subsequent treatment of the skin of mice with croton oil, a well-known tumor promoter, resulted in the appear- ance of papillomas and carcinomas of the skin [20]. These data suggested that urethane was capable only of initiating cells in the skin and not inducing the remaining process of TABLE! CHARACTERISTICS OF STAGES OF INITIATION AND PROMOTION IN SKIN CARCINOGEN- ESIS Initiation Promotion Ixrcvcrsiblc, with `memory' Reversible, at least in early stages Initiated celLs and immediate progeny not usually identi- fiable Promoted neoplasm seen grossly `Pure' initiation (incomplete carcinogen) causes lire- versible change but not neoplasm unless promoter applied Promoting agents not carcinogenic but may promote fortuitously initiated cells (i.e., background) Dependent on cell cycle and, for many chemicals, on the metabolism of the cell Modulated by diet, hormonal, environ- mental, and related factors PAGENO="0032" 28 194 epidermal carcinogenesis. Urethane was termed an `incomplete' carcinogen or `pure' initi- ating agent for the skin while it was a `complete' carcinogen for the lung and liver [20]. In contrast, a number of polycycic hydrocarbons (e.g., benzo(a)pyrene, methyichol- anthrene) have been shown to be complete carcinogens for the skin, having the capability both of initiation and promotion since repeated applications of such chemicals result in the production of neoplasms [4,19]. One of the experimental difficulties with the two-stage system in mouse skin is that initiated cells can only be identified at present by the production of neoplasms. If one assumes that the neoplasms are clonal in origin, then the process of initiation in the skin would seem to be extremely inefficient when based solely on the production of neo- plasma since only a few tumors have been shown to develop relative to the total cell population of the epidermal area treated by carcinogenic agents. Although initiating agents were found to exert their effects in a manner such as to produce an irreversible change, it was clear that the efficiency of initiation was dependent on the cell cycle [21 ,22] and that many initiators depended on intracellular metabolic pathways for their conversion to active forms which then are responsible for the actual initiating event in the cell [23]. In contrast, agents which inhibited the process of carcino- genesis have been termed anti-carcinogens [51. Some of these latter agents have been shown to stimulate drug metabolic pathways in the cell which destroy the carcinogenic potency of many chemical agents [24]. Thus, if such compounds are given slightly before or simultaneously with the administration of the initiating agent, no tumor production results. Promotion also can be markedly affected by environmental factors. Boutwell, Rusch, Tannenbaum and others [25,261 demonstrated the effect of nutrition on tumor promotion, with respect to both the quality and the quantity of nutritional components given to animals during the process of carcinogenesis. Furthermore, Boutwell [161 demonstrated by suitable dosage regimens that promotion in skin carcinogenesis could be reversed under appropriate circumstances. Retinoids have been shown to inhibit the effects of promoting agents in the skin [27] and have also been found to prevent the appearance of several types of epithelial neoplasms when administered at appropriate times during carcinogenesis [28]. Since tumor promotion depends on continued application of the promoting agent, it is not surprising to find that the repeated application of promoting agents sometimes gives rise to a very low incidence of neoplasms. Based on this finding, Roe and Clack [291 have suggested that tumor promoters, specifically croton oil and related compounds, are sim- ply very weak carcinogens. On the other hand, efficient promoting agents would be expected to promote the formation of neoplasms from a small minority of initiated cells produced either by ambient conditions of dietary mutagens, background radiation and the like or by contaminating components of the promoting agent itself. However, the reader should realize thatat this point in time it is not possible to distinguish on an abso- lute basis between an efficient promoting agent and a weak carcinogenic agent by any known experimental methodology. IIB. Identification and characterization of stages during carcinogenesis in tissues other than skin Within the last two decades it has become apparent that the two-stage phenomenon in the development of neoplasms is not unique to the skin. Furthermore, with our increased knowledge of the natural history of neoplasia in both the experimental animal and in the PAGENO="0033" Dog bladder Rat bladder Rat bladder Rat colon Rat bone marrow (leukemia) Mouse embryo fibroblasts (in culture) Mouse epidermis Mouse forestomach Rat liver 2-Naphthylamine Metlsylnitrosourea N-(4 J-(5-nitro-2-furyl)-2-tliiazolylfornotmide N-methyl-N'-nitro-N-nitrosoguanidine N,N'-2,7-fluorenylbisacetarnjcie 3-Methyicholanthrene Ultraviolet radiation 3-Methylcholanthrene, ~-propioIactone, urethane, etc. 3-Methyicholanthrene, bcnzo~a)pyrcne, dimethylbenzo(a)anthracene 2-Acetylaminofluorene, dietliylnitrosamine, azo dyes Urethane 7,1 2-Dimethylbenzo(a)~ntliracene Azascrine N-Methyl-N-nitrosourea 2-Acetylaminofluorene `Preneoplastic' lesions Promoting agent Alkaline phosphatase-deficient D,L-tryptophan foci Saccharin Allopurinol Proliferative foci Litbocholic acid Blood loss `letradecanoylphorbol acetate `letradecanoylphorbol acetate Croton oil or tetradecanoyl phorbol acetate Croton oil or lime oil Phenobarbital, DDT, PCB's, butylated hydroxytoluene, estrogens Butylated hydroxytoluene L)uctular hyperplasia Phorbol, prolactin llypcrplassic alveolar nodules Acinar nodule Adenonias Ref. [38,39J [401 [41J [42,431 [44J (45J [46J [4,19j [47,48 [49,50,51,52,94j [531 [54,551 [56,57j [58J "a 0 00 00 03 TABLE II INITIATORS, PROMOTERS AND "PRENEOPLASTIC" LESIONS IN VARIOUS ORGAN SYSTEMS Tissue Initiating agent Mouse lung Rat mammary gland Rat pancreas Rat thyroid Ilyperplastic nodules Enzyme-altered foci Methyithiouracil ~0 PAGENO="0034" 30 196 human, it has been possible to identify specific `preneoplastic' lesions, many of which appear to be the forerunners of frank neoplasms [301. It is very likely that many of these preneoplastic lesions are composed of altered cells that represent the progeny of a single or a group of initiated cells. As we shall see later in this review, and as is recounted in other papers in this volume, experimental liver cancer may represent one of the best experimental models in which the earliest progeny of the initiated cell can be identified. Table II lists a number of model systems, both experimental and human, in which the two.stage mechanism appears to be evident. This table also gives examples of the `preneo~ plastic' lesions which are characteristically found during the natural history of a specific neoplasm of that organ system. In addition to the specific examples noted in the fable, both Sivak [61 and Scribner and SŘss [19] have pointed out other examples of promoting agents in skin as well as in a variety of other organ systems. The early studies of Rous and his co-workers [14,151 demonstrated a cooperative effect following the treatment of rabbit skin with both the Shope papilloma virus and coal tar. Subsequently, numerous studies have indicated that chemical carcinogens may `activate' latent or endogenous oncogenic viruses [6] and have also suggested that hor- mones may act as promoters in at least one virus-induced type of cancer, the mouse mam- mary tumor [31,32]. Furthermore, in at least two instances, it has been suggested that viruses may exhibit a promoting action on chemical carcinogenesis [33 ,34}. T'ne effect of nutrition and caloric restriction on the incidence and rate of formation of a variety of neoplasms in experimental animals and even in the human has now been well documented [35]. Thus, it is apparent that the two~stage phenomenon in the natural history of neoplastic development is not limited to epidermal carcinogenesis, but is a much more general phenomenon. In this respect, it is likely that promoting agents effec- tive in the genesis of a wide variety of neoplasms. if not all, can be expected to be found. The implications of this statement can have considerable effect with respect to the evalua- tion of risk to the human of agents that are found to be `carcinogenic' in experimental animals. IIC. Definitions of initiation and promotion The definitions of initiating and promoting agents presented herein are obviously sub- ject to criticism and final experimental confirmation, but at the present time. in these reviewers' opinion, such definitions represent a view of the most probable mechanisms involved. Initiating agent: a chemical, physical, or biological agent which is capable of directly altering in an irreversible manner the native molecular structure of the genetic component (DNA) of the cell. Such alteration(s) may be the result of a covalent reaction of DNA with the initiating agent itself or with one of its metabolites, but this alteration may also include a distortion of the structure of DNA without covalent binding to its components. Finally, the agent may cause one or more complete scissions of the DNA chain, an elimi- nation of on& of its component parts (e.g., bases or sugars), or errors in DNA repair. Such capabilities of an initiating agent, however, do not in themselves prove that alteration of DNA is the only or the absolute requirement for the neoplastic transformation. Promoting agent: an agent that alters the expression of genetic information of the cell. Examples of such agents include hormones, drugs, plant products, etc., which in them- selves do not directly react with the genetic material but rather affect its expression by a variety of mechanisms including their interaction with cell surface receptors or with PAGENO="0035" 31 197 cytoplasmic and nuclear protein receptors, or by an alteration of other cellular compo- nents and functions. These definitions are restrictive and characterize initiating and promoting agents as being distinct from co-carcinogens, anti-carcinogens and other similarly used terms. Hecker [36] has sought to distinguish between these latter agents and promoting agents as well as initiating agents. However, it is not the object of this review to discuss the semantics of nomenclature, but rather to concentrate on the two-stage phenomenon seen in hepatocarcinogenesis. Thus, the development of this subject will utilize the definition of initiating and promoting agents as stated above, and also attempt torelate the experi- mental findings to these definitions as well as to the characteristics of initiation and pro. motion described in Table I. These proposed definitions of initiating and promoting agents do not cover all exam- ples of carcinogenic agents. Clearly oncogenic viruses and radiation can be considered as initiating agents since they alter the basic structure of the DNA of the cell. But certain hormones and the enigmatic `plastic film' as carcinogenic agents do not appear to con- form completely to these definitions. Several of these problems have been pointed out in a review by Berenbium [37] in which he has proposed that hormones may actually be considered as a class of promoting agents (sometimes as co-carcinogens) which may effi- ciently promote the expression of cells initiated by ambient environmental factors as noted earlier. On the other hand, the mechanism of the `plastic film' carcinogenesis clearly does not conform to the concepts stated herein and thus will have to remain an unexplained mechanism for tumor induction. III. The natural history of hepatocarcinogenesis - stages in the formation of hepatic car- cinomas It is obvious from the data presented in Table II that carcinogenesis in a variety of dif- ferent tissues is characterized by distinctive stages, which in certain instances may be identified with specific morphologic components, but in most cases can be separated by distinctive responses to specific agents, the principal ones being initiators and promoters. While each system has its distinct advantages and disadvantages, a major advantage of experimental hepatocarcinogenesis as a niodel system is the large amount of background information available on the morphogenesis of hepatocellular carcinoma [59,60] and the even larger body of data available on the biochemistry and physiology of the liver [61]. Thus, in studies aimed at dissecting and characterizing the distinct stages of hepatocar- cinogenesis, the investigator has as a base a wealth of information which can be applied to the understanding of each distinctive process as it is uncovered. Initially, however, one must clearly define the stages of the process of hepatoca.rcinogenesis in relation to our general knowledge of the natural history of neoplasia and especially its analogies to the well-defined stages seen in epidermal carcinogenesis. lilA. `Preneoplastic lesions' identified ui experimental hepatocarcinogenesis and their relationship to stages in this process It is well established that during the administration of hepatocarcinogens to the rodent, several distinct focal and nodular hyperplastic lesions develop in the liver before the appearance of hepatocarcinorna. Some of the specific biochemical `markers' used to distinguish these `preneoplastic' lesions from normal liver parenchyma are shown in Table III and are discussed in detail below. PAGENO="0036" 32 198 TABLE III SOME PROPERTIES OF ENZYME-ALTERED FOCI AND HYPERPLASTIC NODULES INDUCED IN THE RAT LIVER BY CHEMICAL CARCINOGENS Refs. 1. Show altered metabolism of glycogen [67,741 2. Deficient in the enzyme activities of: [50,63,70,89,901 glucose-6-phosphatase [50,71,90,1201 adenosina triphosphatase(cailaliCUlaI) [72,901 ~-glucuronidare [90,172] serinc dehydratase [90,1681 acid phosphatase [67] glycogen phosphorylaso trans- [50,83,891 3. Exhibit increased activity of the fetal hepatocyte enzyme -y.glutamyl peptidase [65,661 4. Express preneoplastic antigen [75,138] 5. Show a deficiency to store iron [1731 6. Resistant to the cytotoxic action of hepatotoxins and carcinogens [67,71,72,891 7. Exhibit an elevated DNA synthesis and mitosis characteristics [50,89,90,172] 8. Possess `phenotypic heterogeneity' with respect to above Hyperplastic nodules. Although considerable descriptive literature on the morphology and morphogenesis of hepatic neoplasms induced in rodents by chemicals had been reported prior to 1955 [59], Farber and his associates [62] were among the first to emphasize the importance of the `hyperplastic' nodule as a possible progenitor or preneo- plastic' precursor to hepatocellular carcinoma. Goldfarb and Zak [631 further emphasized the association between hyperplastic nodules and hepatocellular carcinomas. Their studies were based not only on morphology but also on the deficiency in glucose-6-phosphatase activity of these nodules, a characteristic found in a number of hepatocellular carcinomas. However, Teebor and Becker [64] reported that many, but not all hyperplastic nodules regressed or disappeared after the animal was removed from the carcinogenic stimulus. Such a fmding called into question the possible `preneoplastic' characteristic of this lesion since initiated cells and their progeny are stable and irreversible according to the charac- teristics outlined in Table 1. In later studies, Farber and his associates [65,66] demon- strated with imrnunochemical procedures that a number of cells within hyperplastic nodules expressed a new antigen, the preneoplastic antigen. Using this antigenic marker, Farber further demonstrated that upon regression, many cells of the hyperplastic nodules became~ `remodeled' into the normal anatomical structure of liver, but without showing a concomitant loss of their expression of the preneoplastic antigen [65]. Thus, despite the findings that many hyperplastic nodules were not stable as specific structures following withdrawal of the carcinogenic stimulus, cells within such lesions may well represent the progeny of initiated cells. The morphology of hyperplastic liver nodules has been reviewed [67] and the nomen- clature re-evaluated [68]. While most of the studies concerning the morphology and natural history of these lesions have been described in rats, Butler and Hempsall have recently emphasized the structure of comparable lesions in the mouse [69]. However, there is no evidence to these reviewers' knowledge that liver `nodules' in the mouse regress upon removal of the inducing stimulus. Thus, it does appear that in the rat at least, there PAGENO="0037" 33 199 are a number of `preneoplastic' nodules which can be identified grossly and microscopically which are transient or unstable upon removal of the carcinogenic stimulus, but that other such lesions persist even when the carcinogenic stimulus is removed [64,67]. tjnfortu- nately, it is not possible at present to distinguish between the stable and transient nodule lesions in the rat exclusively through the use of light microscopic and standard histo- pathological staining procedures. Enzyme-altered foci. A second type of preneoplastic lesion, the enzyme-altered focus or liver island, was first described by Friedrich-Freksa and his associates [70] as occurring soon after injections of the hepatocarcinogen, diethylnitrosamine, to rats. These lesions, which developed long before carcinomas appeared, were characterized by small foci of cells that were devoid of histochemically demonstrable glucose-6-phosphatase activity. In addition, the islands could only be distinguished from normal hepatocytes through the use of histochemistry and were found to occur at any of the administered doses of the carcinogen used in this study. In an extension of these studies, Rabes et at. [71] demon- strated that such foci could also be identified by their deficiency of canalicular adenosine triphosphatase activity. Furthermore, these workers showed that the thymidine-labeling index of cells from such foci was greater than that of the surrounding normal paren- chyma, and they also described lesions which appeared to represent intermediate transi- tional stages between the development of the early foci and the appearance of hepatocel- lular carcinomas. Cells of these enzyme-altered foci were designated by Farber as a new hepatocyte population, type III [65]. These findings were confirmed by Kitagawa and Sugano [72] who also demonstrated that such foci weredeficient in ~3-glucuronidase. In addition, Scherer and Hoffmann [73] showed that such enzyme-altered foci were proba. bly clonal in origin. Bannasch [74] has carefully studied the morphological development of hepatocellular carcinoma from microscopic `preneoplastic' lesions in the liver of the rat. In this respect, he was able to distinguish with light and electron microscopy at least four different types of altered hepatocytes; ~clear' glycogen storage cells, acidophiic glycogen storage cells, fat storage cells, and basophiiic cells which were poor in glycogen and rich in ribosomes. In addition, there were diverse intermediate cell types. Based on these findings, Bannasch proposed that the cytogenesis of neoplastic lesions and of hepatocellular carcinoma was characterized by a progressive loss of glycogen in the clear and acidophilic cells with an increase in ribosomes (basophilia) and often, a transitory accumulation of fat. Many of these cellular lesions were also shown to be deficient in glucose.6-phosphatase, but in some, this histochemicafly demonstrable enzyme was found to be normal or even increased in its staining intensity. Studies from our laboratory, which employed other markers as well (Pitot, H.C., unpublished data), have confirmed Bannasch's findings and have further demonstrated that all of the lesions described by Bannasch exhibit one or more histochemicaj abnormalities when several such biochemical markers were investi- gated simultaneously (see subsection tuB).. Other histochemical and functional characteristics have been used to localize altered foci of hepatocytes during carcinogenesis. Williams and Watanabe [75] have demon- strated that `carcinogen-induced foci' of altered hepatocytes are refractory to the accu- mulation of stainable iron following the feeding of ferrous gluconate together with 8-hydroxyquinoline for several months or after the injection of saccharated iron oxide at intervals over a two-week period. However, with discontinuation of the administration of the carcinogen, 2-acetylaminofluorene, many of the iron-deficient foci were shown to dis- appear [75], thus indicating a transient nature to these foci which is reminiscent of that PAGENO="0038" 34 200 seen with many of the hyperplastic nodules in the rat following cessation of the carcino- genic stimulus. One of the most interesting biochemical markers for `preneoplastic' and neoplasti liver lesions in the rat is the enzyme y-glutamyl transpeptidase. This enzyme has been shown by histochemical and biochemical methods to be either nondetectable or present only in very low amounts in normal adult rat hepatocytes [76,77] as well as in the hepa- tocytes of regenerating rat liver [78]. In. contrast, the activity of this enzyme, which was found to be very high in the hepatocytes of fetal and early neonatal rodents [79,801, was first shown by Fiala and his associates [78,80-821 to be markedly elevated in many rat hepatocellular carcinomas. Subsequent studies by many other investigators have extended Fiala's findings and have shown this enzyme to be increased in `preneoplastic' liver lesions and in hepatocellular carcinomas induced in rats by a variety of chemical carcinogens [79. 83-88]. Pugh and Goldfarb [891, using serial cryostat liver sections of rats fed 2-acetylamino- fluorene. showed a significant variation in enzyme staining among the various preneo- plastic and neoplastic lesions identified. As markers, they used glucose-6-phosphatase and adenosine triphosphatase deficiency, as well as the positive staining reaction of -y-glutamyl transpeptidase. In this way, they demonstrated foci which exhibited every possible com- bination of `phenotype', although the predominance of foci did exhibit a positive stain for 7-glutamyl transpeptidase as predicted by the data of others [83,84]. Kitagawa [90] also reported variations in the enzyme phenotypes of enzyme-altered foci and of hyper- plastic nodules. Miscellaneous `preneoplasric' lesions. In addition to hyperplastic nodules and enzyme- altered foci, other `neoplastic' lesions in the liver have been reported. Daoust and Molnar [91] have suggested that thyperbasophiic' foci of altered hepatocytes seen in the rat liver following the feeding of carcinogenic azo dyes may represent a transition between hyper- plasia and neoplasia as well as contain thLactual sites of the neoplastic transformation. Karasaki [92] has further shown that the cells of these hyperbasophilic foci possess an altered histochemistry for membrane ATPase and alkaline phosphatase which is distinct from that exhibited by hyperplastic nodules and enzyme-altered foci, but which persists during the development of azo dye-induced hepatocellular carcinoma. However, like hyperplastic nodules, such hyperbasophilic areas when induced by another carcinogen, N-nitrosomorpholine, have been reported to be transient [93], although such a finding does not rule out the presence of some initiated cells in hyperbasophilic areas. In addition to the morphologic changes described above, other cellular alterations have been reported to occur during hepatocarcinogenesis. One of the best studied is the prolif- eration of a cell population which was first termed by Farber as `oval cells' [94]. There is some indication that such oval cells may have characteristics of the hepatocyte [95] and that they may actually transform into hepatic parenchymal cells [96]. The oval cells appear to originate from the canal of Hering and to disappear upon removal of the car- cinogenic stimulus [97] While not shown to arise directly from oval cells, another pos- sible `preneoplastic' lesion is seen as areas of cholangiofibrosis which arise during the feed ing of various chemicals to rodents [98]. These areas of fibrosis containing atypical chol- angioles have been claimed to give rise to cholangiocarcinomas and may contribute to the formation of poorly differentiated hepatocellular carcinoma. In the human, Takahashi et al. [99] have proposed that hepatic lobules having a dis- tinctly greater mean radius than lobules in normal liver represent a `precancerous' growth of hepatocytes. PAGENO="0039" 35 201 IIIB. The experimental definition of stages in heparocarcinogenesis As can be seen from the variety of preneoplastic lesions which have been shown to occur during hepatocarcinogenesis, the morphological delineation of specific stages during hepatocarcinogenesis requires functional studies in order to define clearly this process as a summation of initiation, promotion and progression as seen in other systems, especially that of the skin. In order to delineate such stages it is necessary to establish experiments patterned after those of Berenbium, Shubik, Boutwell and others who have defined these stages in the natural history of epidermal carcinogenesis. IJJB-1. Modifiers of heparocarcinogenesis Long before any experimental evidence indicated that hepatocarcinogenesis could be divided into distinct stages, there was evidence to show that this process could be modi- fied by both exogenous (dietary) and endogenous (hormonal) factors. Some of the earlier studies were reported by Bielschowsky [100] who demonstrated that goitrogens inhibited hepatocarcinogenesis by 2-acetylaminofluorene. Early studies by the Millers et al. [101] further demonstrated that the riboflavin content of the diet in rats undergoing azo dye. induced hepatocarcinogenesis markedly~ influenced the incidence of hepatomas in these animals. Bielschowsky extended his earlier studies by investigating tlie effects of endocrine ablation and replacement on hepatocarcinogenesis by 2-acetylaminofluorene [102]. The~e studies demonstrated that thyroidectomy and adrenalectomy completely inhibited the production of hepatoma by this carcinogen in male rats. Hypophysectomy, as expected, was shown to prevent 2-acetylaminofluorene-induced hepatocarcinogenesis in rats of both sexes [103]. Later, Weisburger et al..[l04] demonstrated that weanling male or female rats fed 0.016% 2-acetylaminofluorene, a subcarcinogenic dose, resulted in tumor development in those animals receiving high levels of pituitary hormones from an implanted marnmotropic pituitary tumor, whereas the control rats showed only precan- cerous lesions. It has been known for some years that male rats are much more suscep- tible to the hepatocarcinogenic action of 2-acetylaminofluorene than are female rats [105]. Testosterone administration to young castrated male rats fed 0.025% 2-acetyl- aminofluorene in the diet caused a marked increase in the development of hepatic neo- plasms as compared with animals which received only the carcinogen and not the hormone [106]. Such studies suggest that the pituitary and the male sex hormones are acting as promoting agents in hepatocarcinogenesis. However, studies by Lotlikar, Enomoto and the Millers [107;108] showed that hormonal alterations such as those described above alter the metabolism of 2-acetylaminofluorene to its ultimate form, thus changing the effective dose of this carcinogen. As indicated earlier and as exemplified by some of the above, studies, the hormones could be acting as true co-carcinogens or as. anti-carcinogens in their manner of altering the initiation of hepatic cells. In contrast, promoting agents in the liver exert their action as promoters only when given after the administration of the carcinogen. In addition to the effects of altering riboflavin content [101], hepatocarcinogenesis by several different chemicals has also been shown to be modified by other dietary altera- tions. Rogers [109] demonstrated that male rats fed a lipotropic deficient, high fat diet along with 2-acetylaminofluorene developed hepatocellular carcinomas more rapidly and in higher incidence than in those animals fed a normal diet. Earlier studies by Miller, Rusch and their associates [1101 further demonstrated that the feeding to rats of un- PAGENO="0040" 36 202 saturated fatty acids along with a carcinogenic azo dye increased the incidence of tumors in these animals. One may attribute the effect of riboflavin to its action on the metabo- lism of the azo dye carcinogen since riboflavin is a known cofactor for several of the elect tron t~ansport systems in the microsomes where this carcinogen is metabolically acti- vated. On the other hand, the mechanism for the effect of dietary alterations in fatty acid composition and in lipotropes on hepatocarcinogenesis is not so clear. Some authors have demonstrated that inhibitors of the metabolic perox.idation of lipids also inhibit carcino- genesis when given with the carcinogen regimen [111,1121, but as indicated in Table II, at least one such antioxidant, butylated hydroxytoluene, when fed after carcinogen administration is an effective promoting agent for hepatocarcinogenesis [51]. IIIB-2. Biologic demonstration of the stages occurring during heparoc~rrcinogenesis Although earlier studies [62,63,67] had suggested that the hyperplastic nodule was an early or precursor lesion to hepatocellular carcinomas, the implication of this hypothesis for the demonstration of stages during hepatocarcinogenesis was not successfully exploited. The first clear demonstration that hepatocarcinogenesis could be separated into stages was reported by Peraino and his associates in 1973 [49]. In this study 2-acetylaminofluorene was fed at a level of 0.02% for 18 days to male weanling albino Sprague-Dawley rats. Following the feeding of 2-acetylaminofluorene, one group of ani- mals was maintained on the control diet while several other groups were placed on diets containing 0.05% phenobarbital. The phenobarbital diet was given at various times up to 30 days after the cessation of the 2-acetylaminofluorene feeding. The results of this study showed that, while only 20% of the animals fed 2-acetylaminofluorene alone devel- oped hepatic neoplasms within 180 days following cessation of the carcinogen feeding, ~ approx. 70% of the animals fed phenobarbital, including those receiving this barbiturate in the diet as late as 30 days after the cessation of the 2-acetylaminofluorene feeding, devel- oped neoplasms within the 180 day period. This study confirmed and extended an earlier and less detailed study by these investigators [1131 in which they demonstrated that when phenobarbital was fed simultaneously with 2-acetylaminofluorene, there was a sig. nificant decrease in the number of neoplasms induced by the carcinogen, thus indicating the critical nature of the timing of feeding the promoting agent in relation to initiation. Later studies by Peraino and his associates [114] further demonstrated that DDT, the common insecticide, exhibited a promoting effect similar to that exhibited by pheno- barbital, but other compounds including amobarbital and diphenyihydantoin, which have similar effects on liver structure and metabolism as does phenobarbital, had no `enhanc- ing' or promoting activity during hepatocarcinogenesis. This study also demonstrated that both phenobarbital and DDT increased DNA synthesis and caused an increase in liver weight while the other agents studied did not. In a more recent study by these workers [1151, the interval between the cessation of 2.acetylaminofluorene feeding and the start of phenobarbital administration was extended for up to 120 days. Even with this long interval, a significant promoting action of phenobarbital on hepatoma production was seen, and it was suggested by these investigators that the final yield of neoplasms was much more influenced by the duration of phenobarbital treatment than by the length of interval between carcinogen and promoter administration. In addition, Peraino and his. associates showed that as little as 20 days of phenobarbital feeding begun several ~ days after the cessation of 2.acetylaminofluorene administration was sufficient for a sig- nificantly increased production of hepatomas, while the feeding of phenobarbital for only five days beginning at one week after the cessation of carcinogen administration had no promoting effects. PAGENO="0041" 37 203 These studies on the promoting activity of phenobarbital in hepatocarcinogenesis have now been extended by other investigators. Weisburger et al. [1161 have demonstrated that dietary phenobarbital begun one week after the cessation of diethylnitrosamine administration increased the productiori of liver tumors nearly 5-fold. Kitagawa and Sugano [117] confirmed these investigations and also demonstrated that dietary pheno- barbital promoted the hepatocarcinogenic action of 3 `-methyl.4.(dimethylamino)azo. benzene. In addition, Peraino et a!. [118] showed that both male and female C3H mice, whose incidence of spontaneous hepatornas by one year was approx. 10% in females and 50% in males, exhibited in each case a 100% incidence of hepatoma when fed 0.05% phenobarbital in the diet for the same p~riod of time. It is thus apparent that this pro. moting agent is capable of accelerating the normal incidence of spontaneous neoplasia which would be expected to occur in a specific species. This fact is quite important when interpreting the data of Rossi et al. [119], who demonstrated that the feeding of 500 ppm of phenobarbital in the drinking water to rats for approx. 2 years led to a significant number of `liver-cell tumors' in the experimental animals. Since the control animals were maintamed for 2 or more years in order to determine the exact incidcncc of spontaneous hepatic neoplasms, these data could also be interpreted as demonstrating the promoting activity of phenobarbital on `endogenous' hepatoma formation just as reported in mice by Peraino and his associates [118]. More recently, Pitot et al. [50] clearly demonstrated the promoting action of pheno- barbital when this agent was given to rats which had previously received a single oral dose of diethylnitrosamine administered within 24 h after a 70% hepatectomy. The effect of administering low doses of diethylnitrosarriine shortly before or after a partial hepatec- tomy was first described by Scherer and Emmelot [120-122]. These investigators demonstrated with this regimen that there is a direct proportion between the number of enzyme-altered foci formed and the dose of diethylnitrosamine up to a dose of 30 mg/kg. At doses of 20 rng diethylnitrosamine/kg~ no neoplasms resulted even when animals were allowed to survive for 24 months [120,121]. In the study of Pitot et al. [50] the partial hepatectomy followed by diethylnitrosamine at doses of S or 10 mg/kg was subsequently followed 2 months later by the feeding ÓÝ 0.05% phenobarbital for another 6 months. While a significant number of enzyme-altered foci resulted from the administration of either of these doses of diethylnitrosamine after partial hepatectomy, the number of enzyme-altered foci was increased almost 5-fold and hepatocellular carcinomas occurred when animals were fed the diet containing 0.05% phenobarbital. Furthermore, this inves- tigation extended that of Scherer and Emmelot [120,121] in that serial cryostat sections were stained sequentially for the enzyme glucose.6-phosphatase, canalicular ATPase, and 7-glutamyl transpeptidase. As indicated earlier, these three qualitative markers had been used in the past to score enzyme-altered foci (Ref. 89, Table III). The experiments of Pitot and his associates clearly distinguish between the stage of initiation and that of promo- tion in that, at the doses of carcinogen employed, no carcinomas resulted in the animals for at least 24 months [120,121]. Only those animals receiving the promoting agent, phenobarbital, exhibited carcinomas within 6 months of continuous feeding of the drug [50]. The enzyme-altered foci themselves, which had earlier been shown to be clonal in origin [73], may be considered as being the immediate progeny of the initiated cell. The mechanism whereby the promoting agent actually appears to increase the number of initiated cells, which then express themselves as enzyme-altered foci, is not at the mo- ment understood [123]. IIIB-2a. Promoting agents for hepatocarcinogenesis. In addition to phenobarbital, a PAGENO="0042" 204 38 TABLE IV PROMOTING AGENTS IN HEPATOCARCINOGENESIS Promoter Dosage regim~n Initiating agents Species effective with: Refs. Phenobarbital 0.05% in diet con- tinuous for 4-5 months Same 9-11 months Same 12-24 weeks 500 ppm in diet for 19 weeks 0.05% in diet for 5-7 months Same 12-24 weeks 2-AcetylaminotluOrefle Rat . - Mouse Diethylnitrosamine Rat DiethyLn.itrosamine Rat Diethylnitrosamine Rat 3'-methyl-4(dimethyl- Rat amino)azobenzene [49,5 1] [1181 [117] [116] [50] [117] DichlorodiphenYltrichlOro 112 mg/kg daily 2-Acetylarnino- Rat [114] ethane (DDT) intraperitoneal injection fluorene Rat [51] Butylated hydroxytoluene 0.5% in diet 2-Acetylamino- fluorene Polychiorinated biphenyl(s) 400 ppm in diet 6 months 1000 ppm in diet S weeks 3'.Methyl-4(dimethyl- Rat amino)a.zobenzene 2-Acetylamino- Rat fluorene [1271 [128] 3.(3,5~dich1orophenyI).5.5' dimethyloxazoline-2,4~iOfle Estradiol-17-phenyl 2500 ppm in diet 8 weeks 645 ~g/100 g body wt. 2-Acetylamino- Rat fluorine Nitrosomorpholine Rat [128] [1291 propionate and estradiol benzoate number of other chemicals have now been reported to possess the capability of promot- ing hepatomas in rodents whose livers were previously initiated by chemical carcinogens. A summary of these chemicals can be seen in Table IV. It is of interest to note that of the chemicals described, at least three, polychiorinated biphenyis, DOT, and estradiol, have been found to exhibit significant carcinogenic effects in liver [124-1261. Since several efficient promoting agents for hepatocarcinogefleSiS have also been shown to induce significant numbers of hepatocdllular carcinomas after extended periods of administration (greater than 22 months), the question is again raised as to whether such agents as DDT, phenobarbital and polych.lorinated biphenyls are actually weak hepato- carcinogens or are highly efficient promoting agents. In the latter instance, the neoplasxns apparent after prolonged feeding of such chemicals might have been initiated by ambient environmental conditions such as those of diet, background radiation, airborne particu- lates, etc. Recent studies demonstrating the nitrosamine contamination of most com- monly used laboratory animal diets [1301 clearly indicate that there is no lack of initiat- ing agents in the ambient environment of laboratory animals. Pitot [1311 has further reported that while phenobarbital feeding alone to rats for 8 months resulted in no evi- dence of enzyme-altered foci, if the animals were initially subjected to a 70% hepatec- tomy followed by phenobarbital feeding for a similar period of time, a low level (5% or PAGENO="0043" 39 205 less compared with animals receiving diethylnitrosamine plus phenobarbital) of enzyme- altered foci could be demonstrated after seven months of administration of the promot- ing agent. Thus, it is apparent not o~ily that the opportunities exist for the initiation of hepatic cells by ambient environmental conditions, but also that one can demonstrate such initiated cells following the `fixation' [1321 of the initiating event by induced cell proliferation and their subsequent promotion by the administration of an efficient pro- moter of hepatocarcinogenesis, such as phenobarbital. In preliminary investigations bearing on this question, studies by Pitot and his asso- ciates (unpublished data) have demonstrated that the powerful environmental toxicant, 2,3,7,8-tetrachlorodibenzo-p.dioxin (TCDD), is an extremely efficient promoting agent for diethylnftrosamine hepatocarcinogenesis when administered according to the same regimen as described for phenobarbital [50]. In addition, these preliminary studies showed this agent to be on a molar basis one million times more effective than pheno- barbital as a promoter of hepatocarcinogenesis when given bi-weekly to rats by subcuta- neous injection at doses of 0.01-0.1 ig/kg per day. The study of Kociba et al. [133] also demonstrated both the hepatocarcinogenicity of TCDD in rats, as well as the induction of squamous cell carcinomas of the lung and upper respiratory passages when this agent was administered in the diet over a two-year period. On the other hand, Poland and Glover [134] have been unable to demonstrate any interaction of this agent with cellular DNA in vivo, thus indicating that TCDD is not acting as an initiating agent, by definition (subsec- tion IIC'). It should be pointed out, however, that while the evidence described above does sug- gest that efficient promoting agents for the liver may, under the appropriate circum- stances of prolonged administration, be capable of promoting cells initiated by ambient environmental conditions, there is no solid experimental evidence as yet reported which rules out the possibility that such promoting agents are also weakly hepatocarcinogenic or capable of a low level of initiation of hepatocytes. Experiments designed to test this question must revolve around the known characteristics of promoting agents such as their reversibility, threshold levels, and the demonstration of `background' hepatoma forma- tion as the sole source of carcinomas resulting from prolonged promoter administration in the absence of added known initiators. IIIB-2b. Systems demonstrating multiple stages during heparocarcinogenesis. In addi- tion to the studies of Peraino and his associates [49,114], Pitot et al. [50], and several others (Table IV) that showed the effect of promoting agents on hepatoma formation, other systems have been described which demonstrated by morphological and biological means the existence of multiple stages during hepatocarcinogenesis. The studies of Scherer and Emmelot [120,1211 as previously described, demonstrated that at low levels (1-30 mg/kg) of diethylnitrosamine, enzyme-altered foci were induced, but not hepato- cellular carcinomas. Increasing the dose of diethylnitrosarnine to above 40 mg/kg pro- duced no further increase in the number of enzyme-altered foci but did result in the pro- duction of hepatocellular carcinomas. These studies imply a two-stage phenomenon which is subject to the dose of the administered diethylnitrosamine. Since diethylnitrosamine is a complete carcinogen, one would expect that at sufficiently high doses both initiation and promotion would occur. More recently, Farber and his associates [135,136] have reported a system involving the administration of relatively high doses of diethylnitrosamine (200 mg/kg) followed by the short-term administration of a diet containing 0.02% 2-acetylaminolluorene and a partial hepatectomy. This regimen results in the relatively rapid appearance (within one PAGENO="0044" 40 206 to three weeks) of small nodules in the liver exhibiting many of the characteristics of enzyme-altered foci. The administration of the low level of 2-acetylaminofluorene was used to select for those cells (i.e., initiated cells) less affected by the toxicity of this agent Morerecently, Laishes et al. [137] have extended these studies by demonstrating that cells from the altered foci developed with the Farber protocol can be transplanted to recipient animals which were given a partial hepatectomy and which were maintained on a short-term feeding of 0.02% 2-acetylaminofluorene. Such cells can give rise to nodular lesions identical to those found in the protocol referred to above [135,1361. These studies clearly demonstrate the initiation and rapid promotion of cells in the livers of ani- mals placed on this regimen. However, this model has not as yet been developed for the efficient demonstration of promoting agents for hepatocarcinogenesis. LTIB-2c. Evidence for the sequential relationship of enzyme-altered foci to hepatocellu- lar carcinomas. While there is substantial evidence for the two-stage mechanism of hepato- carcinogenesis, evidence for the direct relationship between the enzyme-altered foci (i.e., the putative progeny of initiated cells) and hepatocellular carcinomas is less convincing. A good demonstration of this relation is the experiment of Laishes et al. [137] showing the apparent transplantability of the nodular lesions produced in the Farber protocol [135]. On the other hand, Scherer and Emmelot [122], utilizing small continuous doses of diethylnitrosamine administered after the initiation of hepatocytes by 20 mg/kg of diethylnitrosamine following a partial hepatectomy, did present evidence that enzyme- altered foci were directly related to hepatocellular carcinomas, although it was evident that not every focus developed into a frank neoplasm during the life span of the animal. Actually one would not expect this to be the case based on the data of Pugh and Gold- farb [89] who showed differential {3H]thymidine labeling of enzyme-altered foci exhibit- ing different phenotypes. Pitot and his associates [50] have pointed out that the pheno- types of hepatocellular carcinomas are extremely variable just as are those of the enzyme- altered foci. Utilizing the characteristic of resistance to iron accumulation, Watanabe and Williams [138] presented evidence that enzyme-altered foci are precursors of hepatocellu- lar carcinomas. Finally as already mentioned, Rabes and his associates [71] demonstrated by morphology and autoradiography that foci of cells which were histologically and biologically compatible with early carcinomas arise within some enzyme-altered foci. Pre- liminary studies (Sirica, A.E. and Pitot, H.C, unpublished data) have also demonstrated that at least some cells having the histological and histochemical characteristics of enzyme-altered foci in animals subjected to partial hepatectomy/diethylnitrosamine and phenobarbital feeding [501 may be transplanted into the livers of recipient rats following their short-term maintenance in primary cultures. Recipient animals were treated in a manner comparable to that described by Laishes et al. [137]. These investigations, although not absolutely conclusive, point to the fact that cells of enzyme-altered foci do represent not only the potential, but at least in some instances, the actual precursors of hepatocellular carcinomas in those animals in which the natural history of the neoplastic process is allowed to proceed unhindered. IIIB-3. Analogies between the stages of carcinogenesis in liver and in skin Earlier in this review (Table I) several characteristics of initiation and promotion in epidermal carcinogenesis were presented. Table V presents a comparison of these pro- cesses in epidermal carcinogenesis and hepatocarcinogenesis. In general, the two-stage pro- cess of carcinogenesis in these two systems is very similar. Initiation has been found to be irreversible both in the epidermis [3,4] and in the hepatocyte [120,131]. However, the PAGENO="0045" TABLE V 41 207 COMPARISON OF TWO-STAGE PROCESS OF EPIDERMAL AND HEPATOCELLULAR CAR- CINOGENESIS IN THE RAT Epidermis Hepatocyte Initiation: Irreversible Irreversible Initiated cells and their progeny not usually Enzyme-altered foci probably represent progeny of identifiable initiated cells `Pure' initiator (incomplete carcinogen) Incomplete carcinogens induce enzyme-altered foci causes no discernible change unless promoter applled Dependent on cell cycle and, for many chemi- Markedly enhanced by cell replication. Promoting cals, on the metabolism of the cell ~ agents administered with or before initiator inhibit initiation Promotion: Reversible, at least in early stages Reversibility not tested but does require certain time of promotion for full effect Papilloma resulting from promotion may Many hyperplastic nodules resulting from complete regress ~ carcinogen administration regress on removal of carcinogenic stimulus Promoting agents may promote cells Promoting agents do promote cells initiated by initiated by ambient environmental factors ambient environmental factors Promotion modulated by diet, hormonal and Modulation of promotion not yet studied directly other environmental factors : but hepaeocarcinogenesis is modulated by hor- monal and dietary factors immediate progeny of the initiated cell can be identified in the hepatocyte as enzyme- altered foci [50] whereas no comparable morphologically identifiable lesion has been described in epidermal carcinogenesis. The action of `pure' initiators or incomplete car- cinogens has been described both in the thin [20] and in the liver [131,139]. In addition, as was described earlier, initiation in the epidermis is clearly dependent on the cell cycle [21 ,22] and it may also be subject to the metabolic status of the epidermis at the time of initiation [23]. Such is also true for the liver since many promoting agents and a num- ber of other chemicals, most of which have not been tested for promoting activity, alter the metabolic capability of the liver in such as way as to inhibit the initiation process [113,1141. Demonstrable initiation is markedly enhanced by cell replication in hepato- cytes [120-122]. The phenomenon of promotion is also quite similar in these two systems insofar as indicated by the experiments which have been carried out to date. Certain critical experi- ments rethbin to be done. such as the test for reversibility, which has been clearly demon- strated in epidermal carcinogenesis [16J,but has yet to be satisfactorily tested in hepato- carcinogenesis. Neverthe1es~, Peraino and his associates [115] have demonstrated that tumor promotion requires a certain definite time before any effect is seen and a much longer time before a 100% effect may be demonstrable. In most experiments in epidermal carcinogenesis the immediate end point has been the formation of epidermal papillomas. Such papillomas have been demonstrated, to regress upon removal of the promoting stim- ulus [9,6,16]. An analogy may be seen in the liver in the case of hyperplastic nodules PAGENO="0046" 42 208 which result from the feeding of complete carcinogens but may completely regress on removal of the carcinogenic stimulus [641. As discussed above, promoting agents such as phenobarbital in the liver have been reported to be carcinogenic when fed for prolonged periods of time in the absence of any demonstrable initiating agent. However, there is rea- son to believe that such `carcinogenicity' may be the result of promotion of cells initiated by ambient environmental factors (subsection IIIB.2a). Such may also be true in the case of the skin although in neither situation has this question been answered unequivocally. While tumor promotion has been shown to be modulated by diet, hormonal and other environmental factors in epidermal carcinogenesis [4,6], there has yet to be a study demonstrating similar modulation of promotion in hepatocarcinogenesis. On the other hand, numerous examples exist of the effect of hormonal and dietary alterations on the entire- process of hepatocarcinogenesis [100-102,109,110]. However, the exact site of action of all of these environmental alterations has not yet been ascertained, although there is evidence given by several specific examples [107,140] that suggests that it is the process of initiation rather than that of promotion which is modulated by the environ- mental alteration. Thus, substantial experimental evidence points to a complete analogy between the stages of initiation and promotion as demonstrated in epidermal carcinogenesis and those seen in hepatocarcinogenesis. It is quite likely thz: a detailed investigation of the natural history of carcinogenesis in other :vstems (Table II) will reveal analogies similar to those found in these two systems. IV. Mechanisms of tumor promotion during hepatocarcinogenesis ( The definition of a promoting agent as stated earlier in this review, while very general in nature and scope, did suggest certain mechanisms of action for promotion. Specif- ically, the designation of promoting agent action as one of altering the expression of the genome without the necessity of a covalent interaction of the promoter with the genetic component of the cell relates the mechanism of action of promoting agents to the general area of the epigenetic regulation of gene expression. Boutwell was among the first to postulate that promoting agents altered gene expres- sion in cells [4]. Trosko and Chang [1411 emphasized those concepts focusing somewhat more on the sole action of initiating agents on the genome itself. Most of these concepts have originated from studies on the promoting action of phorbol esters in mouse skin car- cinogenesis and as such, the mechanisms of the phorbol esters have been reviewed in detail by Sivak [6]. Phorbol esters affect differentiation in specific systems [142-144], increase the frequency of mutants in cultured cells treated with ultraviolet light [145], and interact in some manner with the surface membrane of the cell [146,1471. All of these functions subsequently lead to or have a direct effect on the expression of genetic information. Classically, the principal mechanism whereby promoting agents were thought to act was their effect on cell replication. Phorbol esters exhibited such an effect in a variety of systems [6], but it has long been known that the induction of cell replication itself is insufficient to explain the tumor-promoting effects of such agents. Rather, in view of our knowledge of the requirement for cell replication to `fix' the stage of initiation [181, it becomes of great importance to separate the fixation of initiation by cell replication induced by promoters from the unique expression of gene function brought about by a promoter which is specifically involved in the stimulation of neoplastic transformation. PAGENO="0047" 43 209 Very few studies have been directed towards an understanding of the mechanism of action of promoting agents effective in hepatocarcinogenesis. Clearly many of the effects of phorbol esters in a variety of systems are largely not applicable to phenobarbital and the related compounds seen in Table IV. In fact, there is considerable question as to whether phenobarbjtaj has significant effect on cell replication in relation to the promo- tion of neoplasia (Refs. 49, 113 and Peraino, C., personal communication) in the liver or whether the increase in ornithine decarboxylase which is induced by phorbol esters in mouse skin [148] and during phenobarbital feeding [149] is a significant characteristic of tumor promotion during hepatocarcinogenesis. Phenobarbital has been shown to enhance teratogenicity and the production of chro- mosome aberrations induced by 2-(2-furyl)-3-(5.nitro.2.furyl) acrylamide [150]. In con- trast, phenobarbital was found not to increase the binding of 2-acetylaminouluorene to rat liver nuclear DNA when fed as a promoting agent [151]. As expected, when pheno- barbital was fed prior to the administration of this carcinogen, the binding 2-acetyl- aminofluorene to DNA was reduced by 80%. Further. butylated hydroxytoluene (Table IV) has been shown to inhibit excision repair synthesis in normal human lymphocytes damaged by ultraviolet light [152]. Earlier studies on phorbol esters [153] had also shown a similar effect, but as yet, what role, if any, such an effect of promoting agents plays in the carcinogenic process is unknown. However, Pitot [131] has demonstrated that when phenobarbital administration is begun one day following partial hepatectomy/ diethylnitrosamine, the yield of enzyme-altered foci is increased more than 50% over the number seen when administration of the promoting agent is delayed until one week or later following the administration of diethylnitrosamine in this system. This finding sug- gests that the promoting agent may prevent some repair mechanisms when given just after the initiating agent. Examination of the known promoting agents effective in hepatocarcinogenesis does reveal a characteristic common to all. That characteristic is the induction of xenobiotic (drug) metabolism accompanied by a proliferation of smooth endoplasmic reticulum. However, as was demonstrated earlier [115] such a function does not appear to be, in itself, sufficient to explain promoting action during hepatocarcinogenesis. The possibility, which was alluded to previously [37], that all hormones may act as promoting agents in specific systems coupled with the effect of estrogens as promoting agents in the liver [129], might suggest that the action of promoting agents to alter the expression of the genome depends on their interaction with specific receptor molecules within the cell. The concept that promoting agents exert their effect through receptor molecules has already been suggested for the action of phorbols [4], and in view of the studies by Poland and his associates on the TCDD receptor in liver [154], it may be possible that all promoting agents exert their organo-specific promoting effects through some type of receptor mole- cule. In this respect, the. analogy between the regulation of genetic expression affected by hormones and that by promoting agents is a useful one, but extensive experimental work is needed before such a hypothesis can begeneralized. V. Applications of the two-stage process of hepatocarcinogenesis to problems in human cancer VA. Stages in human liver cancer A number of studies have demonstrated many similarities between the pathogenesis and morphological changes of experiment~i and human liver cancer [155,156]. Evidence PAGENO="0048" 44 210 that specific environmental chemicals from industrial, medical, and dietary sources are carcinogenic to the human has now become quite clear [157]. Furthermore, there appears to be a role for at least one virus, the hepatic B virus, in the induction of human hepatic cancer [158]. Early lesions, possibly representing pre-neoplastic foci have been described in human liver [159]. An example of this is focal nodular hyperplasia, a lesion exhibiting the capability of regression similar to hyperplastic nodules in rodents. which was shown to be related to the administration of estrogenic preparations in oral contraceptives [160]. In addition, the association of cirrhosis and hyperplastic nodules with hepatoceLlular car- cinomas in the human has been described by several authors [16 1,1621. At present, it is not absolutely clear that hepatocarcinogenesis in the human can be divided into distinct stages on a biochemical and functional basis. On the other hand, the effect of ethanol ingestion in enhancing (promoting) hepatocellular carcinomas in the human has been known for many years [136] while at the same time there is no experi- mental evidence that ethanol is carcinogenic. However, ethanol is capable of stimulating xenobiotic metabolism in liver and is associated with an increased proliferation of smooth endoplastic reticulum [1641. Interestingly enouch, it has been shown that when ethanol was given simultaneously with a human carcinogen, vinyl chloride, the metabolism of the carcinogen was significantly inhibited [165]. Thus, while functional evidence for multiple stages in the natural history of human hepatocellular carcinoma is relatively sparce. some data point to a marked degree of similarity in hepatocarcinogenesis in the expenmental animal and in the human being. VB. Implications of the two-stage process of hepatocarcinogenesis in the monitoring and regulation of potential human carcinogens At the present time there is considerable world wide interest in identifying environ- mental agents which may and/or do pose a significant risk to the human population by their presence. Although a number of rapid assays, most notably the Ames test [166], have been devised to establish specific characteristics of chemical agents in our environ- ment (notably their mutagenic capabilities), the assay system which ultimately defines the carcinogenic potency of a specific agent remains the whole animal, usually a rodent species. Considerable efforts have been directed towards establishing the best possible rodent bioassay for carcinogenic potency of environmental chemicals [167]. That which is most commonly used in the United States is the assay in which the potential carcinogenic agent is administered at maximally tolerated doses either by chronic feeding over a two-year period or by other forms of administration to the test animal. A statistically significant increase in a specific type of neoplasm in the test animals is usually sufficient to label the agent as carcinogenic and, in many instances, potentially dangerous to the human. Although the usual end point of a whole-animal bioassay is the production of benign and/or malignant neoplasms, several investigators have employed the production of `pre- neoplastic' lesions in the liver as methods for monitoring environmental carcinogenic agents. Tatematsu et al. [168] utilized a modification of the Farber protocol [135] to assay the carcinogenicity of a variety of chemicals within 4 to 6 weeks. Good agreement was found between the carcinogenicity or noncarcinogenicity of nine agents tested and their effects in the system. Sirica et al. [169] have suggested that the protocol described by Pitot et al. [50] is potentially capable not only of assaying complete hepatocarcino- PAGENO="0049" 45 211 gens, but also of distinguishing between incomplete carcinogens (pure initiators) and pro- moting agents in hepatocarcinogenesis, The ability to assay for the latter two agents offers distinct advantages over the present utilization of the whole-animal bioassay test. Incomplete hepatocarcinogens are not monitored by the present whole-animal bio- assay, and thus an agent like ~ which is fully effective as an initiator of hepatocytes [139J would not be picked up by most assay sys- tems that are presently used. In addition, efficient promoting agents, which are capable of bringing out cells initiated by ambient environmental conditions and not by the agent ifself, can not be distinguished from complete hepatocarcinogens by the present bioassay systems. The importance of distinguishing promoting agents is clear from the fact that promoting agents exhibit threshold levels of effect in contrast to initiating agents [4,161 and, as has been shown in the skin system, tumor promotion is not only easily modulated but also is reversible under appropriate circumstances [1 6}. Thus, in the future it will be important to take into consideration the concept of the two-stage phenomenon in the natural history of neoplastic development, not only in the liver, but also in many tissues of the test animals. Such considerations will, in the opinion of these reviewers, go a long way towards alleviating much of the present public concern that almost everything in our environment is potentially carcinogenic as well as of equal risk at any dose level. VI. Conclusions It is evident from this review that the two-stage mechanism of initiation and promo- tion as first described for skin carcinogenesis is not a unique feature of this system, and that entirely comparable stages have now bÚeit demonstrated for the carcinogenic process in liver as well as in other tissue and organ systems (Table II). In this respect, the liver model of carcinogenesis, in contrast with that of the skin, provides the important advan- tage of being able to demonstrate the very early progeny of the initiated cell (the enzyme- altered focus) long before any actual appearance of hepatic neoplasia. This, in turn, makes it possible to quantitate precisely the effects of both initiating and promoting agents in hepatocarcinogenesis and offers the potential for making clear distinctions between tumor promoters and complete and incomplete carcinogens for the liver. Also, with the liver system, it is possible to demonstrate the ability of agents such as pheno- barbital to promote cells initiated by ambient environmental factors. These distinctions become quite important when attempting to extrapolate the potential carcinogenicity of an agent from experimental animal to man. In this respect, it is noteworthy that pheno- barbital by itself has not been shown to be associated with an increase in human liver can- cer [170J, although it may play a role in the production of certain human brain tumors [1711. The most critical factors in the evaluation of the potential carcinogenic risk to humans of initiating and promoting agents relate to their distinct properties; namely that initia- tion is an irreversible process whereas promotion can be modulated and even reversed. This latter property of promotion has already been shown for skin carcinogenesis [161, although the reversibility of the effects of promotion in the liver still needs to be clearly demonstrate4. However, it has been shown with both systems that promoting agents require a continued application in order to exert their effects. Thus, it is theoretically possible that the cessation of exposure to a promoting agent, even after a rather pro- longed interval of exposure, could completely reverse its effects, thereby drastically 22-143 O-83----4 PAGENO="0050" 46 212 reducing the risk of tumor development. On the other hand, from the standpoint of human risk, it becomes quite important to be able to detect and eliminate where possible the exposure of the human population to `pure' initiating agents (incomplete carcino- gens). These agents, unlike complete carcinogens, are not identified by whole screening systems as presently utilized, but because of their irreversible effects, they probably pre- sent as great a risk as complete carcinogens to the human population. Therefore, the application of our knowledge of the natural history of neoplastic devel- opment to the bioassay of environmental carcinogens is critical to the scientific determi- nation of the relative risk of such agents to the human population. References' 1 Bryan, W.R. and Shimkm, M.B. (1941) J. Nat. Cancer Inst. 1, 807-833 2 Rice, J.M. (1973) Teratology 8, 113-126 3 Berenblum, 1. (1975) in Cancer, a Comprehensive Treatise (Becker, F.F., ed.), Vol. I, pp. 323- 344, Plenum Press, New York 4 Boutwell, R.K. (1974) Cdt. Rev. Toxicol. 2, 4 19-443 S Slaga, TJ., Sivak, A. and Boutwell, R.K. (1978) Mechanisms of Tumor Promotion and Cocar- cinogenesis, Vol. 2, Raven Press. New York 6 Sivak, A. (1979) Biochim. Biophys. Acta 560, 67-89 \ 7 Mottram, S.C. (1944) 5. Path. Bact. 56, 181-187 - 8 Berenblum, I. and Shubik. P. (1947) Br. 3. Cancer 1, 379-391 9 Berenblum, I. and Shubik, P. (1949) Br. J. Cancer 3, 109-118 10 Van Duuren, D.L. (1976) Chemical Carcinogens, ACF Monograph 173 (Searle, C.E., ed.) pp. 24-5 1, American Chemical Society, Washington, DC 11 Hecker..E. (1971) Methods Cancer Res. 6,439-481 12 Saffiotti, U. and Shubik, P. (1963) Natl. Cancer Inst. Monograph 10, 489-501 13 O'Brien,T.G. (1976) Cancer Res. 36, 2644-2653 14 Rous, P. and Kidd, 3G. (1941) 3. Exp. Med. 73, 365-384 15 Friedwald,W.F. and Rous, P. (1944)3. Exp. Med. 80, 101-125 16 Boutwell, R.K. (1964) Progr. Exp. Tumor Res. 4, 207-250 17 Foulds, L. (1969) Neoplastic Development, Academic Press, New York 18 Pitot, H.C. (1978) Fundamentals of Oncology, Marcel Dekker, Inc., New York 19 Scribner, S.D. and SŘss, R. (1978) Int. Rev. Exp. Path. 18, 137-198 20 Berenblum, I. and Haran-Ghera, A. (1957) Br. 3. Cancer 11, 77-90 21 Iversen, OH. (1973) Proceedings of the Fifth International Symposium on the Biological Charac- terization of Human Tumours, Excerpta Med. mt. Congress Series 321, 21-29 22 Berenblum, I. and Armuth, V. (1977) Br. 3. Cancer 35, 615-620 23 Miller, J.A. and Miller, E.C. (1971) 3. Nat. Cancer Inst. 47, V-XIV 24 Conney, A.H. and Levine, W. (1974) Chemical Carcinogenesis Essays, IARC Sc Publ. 10, pp. 3-22, tnt. Agency Res. Cancer, Lyon 25 Boutweil, R.K.., Brush, M.K. and Rusch, H.P. (1949) Cancer Ret. 9,747-752 26 Tannenbaum, A. (1944) Cancer Ret. 4, 683-687 27 Verma, A. and Boutweli, R.K. (1977) Cancer Ret. 37, 2196-2201 28 Sporn, M.B. (1978) in Mechanisms of Tumor Promotion and Cocarcinogenesis (Slaga, Ti., Sivak, A. and Boutwell, RJC., edt.), VoL 2, pp. 545-551, Raven Press, New York 29 Roe, FJ.C. and Clack,!. (1964) Br.!. Cancer 17,596-604 30 Pierce, J.B. and Fennell, R.H. (1976) Nat. Cancer Inst. Monograph 44, 99-102 31 Moore, D.H. (1975) in Cancer, a Comprehensive Treatise (Becker, F.F., ed.), Vol. 2, pp. 131- 168, Plenum Press, New York 32 McGrath, C.M. and Jones, R.F. (1978) Cancer Res. 38, 4112-4125 33 Kotin, P. and Wisely, D.V. (1963) Tumor Ret. 3, 186-215 34 Korda, C.V., Schon, E. and Brown, A. (1969) Neoplasma 16, 485-490 35 Tannenbaum, A. and Silverstone, H. (1953) Adv. Cancer Res. 1,451-501 36 Hecker, E. (1976) Z. Krebsforsch. 86, 219-230 PAGENO="0051" 47 213 37 Berenblum, 1. (1978) J. Nat!. Cancer Inst. 60, 723-726 38 Radomskj, J.L., Krischer, C. and K.rischer, K.N. (1978) J. Nat!. Cancer Inst. 60, 327-333 39 Radomskj, J.L., Radomskj, T. and MacDonald, W.E. (1977) J. Nat!. Cancer Inst. 58, 1831-1834 40 Hicks, R.M., Wakefield, J.St.J. and Chowanjc, J. (1973) Nature 243, 347-349 41 Wang, C.Y., Hayashida, S., Pamukcu, A.M. and Bryan, G.T. (1976) Cancer Res. 36, 155 1-1555 42 Reddy, B.S. and Watanabe, K. (1979) Cancer Res. 39, 1521-1524 43 Lipkin, M. (1974) Am. J. Digest. Dis. 19, 1029-1032 44 Takayam, S. and Fujiwara, M. (1976) Acta Path. Jap. 26, 435-439 45 Mondsi, S., Brankow, D.W. and Heidelberger,C. (1976) Cancer Res. 36, 2254-2260 46 Mondal, S. and Heidelberger, C. (1976) Nature (London) 260, 7 10-712 47 Berenblum, I. and Haran, N. (1955) Cancer Res. 15, 510-516 48 Peirce, W.E.H. (1961) Nature (London) 189, 497-498 49 Perajno, C., Fry, RJ.M., Staffeldt, E. and Kisjeljskj, W.E. (1973) Cancer Res. 33, 2701.~~2708 SO Pitot, H.C., Barsness, L., Goldswortliy, T. and~Kitagawa, T. (1978) Nature 271,456-458 51 Peraino, C., Fry, R.LM., Staffe!dt, E. and Christopher, J.P. (1977) Food Cosmet. Toxicol. 15, 93-96 52 Kimura, N.T., Kinematsu, T. and Baba, T. (1976) Z. Krebsforsch. 87, 257-266 53 Witschi, H.P., Williamson, D. and Lock, S. (1977) J. Nat!. Cancer Inst. 58, 301-3 10 54 Nadina, D. (1976) Cancer Res. 36, 2589-2595 55 Armuth, V. and Berenblum, 1. (1974) Cancer Res. 34, 2704-2707 56 Levitt, M., Harris, C., Squire, R.. Wenk, M., Moilelo, C. and Springer, S. (1978) J. Nat!. Cancer Inst. 60, 701-705 57 Roebuck, B.D. and Longnecker, B.S. (1977) J. Nat!. Cancer Inst. 59, 1273-1277 58 Hail, W.H. and Bielschowsky, F. (1949) Br. 3. Cancer 3,534-541 59 Firminger, HI. (1955) J. Nat!. Cancer Inst. 15, 1427-1445 60 Stewart, H.L. and Snell, K.C. (1959) in Physiopatho!ogy of Cancer. 2nd edn. (Homburger, F., ed.), pp. 85-126, Paul B. Hoeber, Inc., New York 61 Rouffler, C. (1963) The Liver, VoIs. 1 and 2, Academic Press, New York 62 Farber, E. and Ichinose, H. (1959) Acta tjnio Intern. Contra Cancrum 15, 152-153 63 Goldfarb, Sand Zak, F.G. (1961) 1. Am. Med~ Assoc. 178, 729-73 1 64 Teebor, G.W. and Becker, F.F. (1971) Cancer Res. 31, 1-6 65 Farber, E. (1976) in Hepatoceilular Carcinoma (Okuda, K. and Peters, R.L., eds.), pp. 3-22, John Wiley and Sons, New York 66 Lin, J..C., Hiasa, Y. and Farber, E. (1977) Cancer Res. 37, 1972-1981 67 Farber, E. (1973) Methods Cancer Res. 7, 345-375 68 Squire, R.A. and Levitt, M.H. (1975) Cancer Res. 35, 3214-3223 69 Butler, V/H. and Hempsail, V. (1978) J. Pathol. 125, 155-16 1 70 Frjedrjch.Freksa, S.H., Gossner, W. and Bonner, P. (1969)*Z. Krebsforsch. 72, 226-239 71 Rabes, H.M., Schoize, P. and Jantsch, B. (1972) Cancer Res. 32, 2577-2586 72 Kitagawa,T. and Sugano, H. (1973) Cancer Res. 33, 2993-3001 73 Scherer, E. and Hoffman.n, M. (1971) Eur. 3. Cancer 7, 369-371 74 Bannasch, P. (1976) Cancer Rca. 36, 2555-2562 75 Williams, G.M. and Watanabe, K. (1978) 3. Nat!. Cancer Inst. 61, 113-121 76 Cheng, S., Nassau, K. and Levy, D. (1978) FEBS Lets. 85, 3 10-312 77 Harada, M., Okabe, K., Shibata, K., Masuda, H., Miyata, K. and Enomoto, M. (1976) Acts Histo- chain. Cytochean. 9, 168-179 78 PlaZa, S. and PlaZa, E.S. (1973) .1. Nat!. Cancer Inst. 51,151-158 79 Tateislal, N., Higashi, T., Nomura, T., Naruse, A., Nakashima, K., Shiozaki, H. and Salcamoto, Y. (1976) Gann 67, 215-222 80 Fiala, S., Mohindru, A., Kettering, W.G., Fiala, A.E. and Morris, H.?; (1976)1. Nat!. Cancer Inst. 57,591-598 81 Fiala, S., Fiala, A.E. and Dixon, B. (1972) 1. Nat!. Cancer Inst. 48, 1393-1401 82 Fiala, S. and Fiala, A..E. (1970) Experientia 26, 889-890 83 Cameron, R., Keilen, I., Kolin, A., Malkin, A. and Farber, E. (1978) Cancer Rca. 38, 823-829 84 Ka!engayi, M.M.R., Ronchi, G. and Desmet, VJ. (1975) 1. Nat!. Cancer Inst. 55,579-588 85 Taniguciii, N., Tsukada, Y., Mukuo, K. and Hirai, H. (1974) Gann 65, 381-387 86 Laishes, B.A., Ogawa, K., Roberts, E. and Farber, E. (1978) 3. Nat!. Cancer Inst. 60, 1009-1016 PAGENO="0052" 48 214 87 Pugh, T.D. and Goldfarb, S. (1977) Proc. Am. Assoc. Cancer Res. 18, 30 88 Taniguchi, N., Saito, K. and Takakuwa, E. (1975) Biochim. Biophys. Acta 391, 265-271 89 Pugh, T.D. and Goldfarb, S. (1978) Cancer Res. 38, 4450-4457 90 Kitagawa, T. (1971) Gann 62, 207-216 91 Daoust, R. and Molnar, F. (1964) Cancer Res, 24, 1898-1909 92 Karasaki, S. (1976) Cancer Res. 36, 2567-2572 93 Taper, H.S. and Bannasch, P. (1976) Z. Krebsforsch. 87, 53-65 94 Farber, E. (1956) Cancer Res. 16, 142-148 95 Ogawa, K., Minase, T. and Onoe, T. (1974) Cancer Res. 34, 3379-3386 96 lnaoka, Y. (1967) Gann 58, 355-366 97 Grisham, LW. and Porta, E.A. (1964) Expel. Mo!. Pathol. 3, 242-256 98 Goldfaxb, S. (1973) Cancer Rca. 33, 1119-1 128 99 Takahashi, T., Oril, T. and Kaneda, M.(1968) Tohoku 3. Exp. Med. 94, 203-224 100 Bleiwhowsky, F~ (1944) Br.!. Expel. Path. 25, 90-95 101 Miller, E.C., Miller, J.A., Kline, D.E. and Rusch, H.P. (1948) 3. Exp. Med. 88, 89-102 102 Bie!schowsky, F. (1955) Br.!. Cancer 9,80-116 103 Weisburger, LH. and Weisburger, E.K. (1963) Clin. Pharmacol. Ther. 4, 110-129 104 Weisburger, 3.H., Pai, S.R. and Yamamoto, R.S. (1964) 3. Nat!. Cancer Inst. 32, 881-904 105 Firminger, HI. and Reuber, M.D. (1961) 3. Nat!. Cancer Inst. 27. 559-572 106 Reuber, M.D. (1976) Eur. J. Cancer 12, 137-14 1 107 Lotlikar, PD., Enomoto, M., Miller, E.C. and Miller, J.A. (1964) Cancer Res. 24, 1835-1842 108 Miller, E.C. and MiUer, J.A. (1971) in Chemical Mutagens (Hollaender, A.P., ed.), pp. 83-119, Plenum Press, New York 109 Rogers, A.E. (1975) Cancer Res. 35, 2469-2574 110 Miller, .T.A., Kline, D.E., Rusch, H.P. and Baumann, CA. (1944) Cancer Res. 4, 153-160 - 111 Shamberger, R.J. and Rudolph, G. (1966) Experientia 22, 116 112 Wattenberg, L.W. (1972) 3. Nat!. Cancer Inst. 48, 1425-1431 113 Peraino, C., Fry, RJ.M. and Stadfe!dt, E. (1971) Cancer Res. 31, 1506-15 12 114 Peraino, C., Fry, R.J.M., Stadfeldt, E. and Christopher, J.P. (1975) Cancer Res. 35,2884-2890 115 Peraino, C., Fry, R.J.M. and Stadfe!dt, E. (1977) Cancer Res. 37, 3623 -3627 116 Weisburger, J.H., Madison, R.M., Ward, 3M., Viguera, C. and Weisburger, E.K. (1975) 3. Nat!. Cancer Inst. 54, 1185-1188 117 Kitagawa, T. and Sugano, H. (1978) Gann 69, 679-687 118 Peraino, C., Fry, R.J.M. and Stadfeldt, E. (1973) .1. Nat!. Cancer Inst. 51, 1349-1350 119 Rossi, L., Revera, M., Repetti, G. and Santi, L. (1977) mt. 3. Cancer 19, 179-185 120 Scherer, E. and Emmelot, P. (1975) Eur. 3. Cancer 11,689-696 121 Scherer, E. and Emmelot, P. (1976) Cancer Res. 36, 2544-2554 122 Scherer, E. and Emmelot, P. (1975) Eur. 3. Cancer 11, 145-154 123 Pitot, H.C. (1977) Am.!. Pathol. 89, 401-412 124 Kinibrough, R.D., Squire, R.A., Linder, R.E., Strandberg, J.D., Montali, R.J. and Burse, V.W. (1975) 3. Nat!. Cancer Inst. 55, 1453-1459 125 Reuber, M.D. (1978) Tumor! 64, 571-577 126 Vana, 3., Murphy, T.P., Aronoff, B.L. and Baker, H.W. (1977) 3. Am. Med. Assoc. 238, 2154- 2158 127 Kimura, N.T.,Kanematsu, T~ and Baba, T. (1976) Z. Krebsforsch. 87, 257-266 128 Ito,N., Tatematsu, M., Hirose, M., Nakanishi, K.. and Muraski, .1. (1978) Gann 69, 143-144 129 Taper, H.S. (1978) Cancer 42,462-467 130 Edwards, G.S., Fox, LG., Policastro, P., Goff, U., Wolf, M.H. and Fine, E.H. (1979) Cancer Rca. 39, 1857-1858 131 Pltot, H.C. (1979) in The Induction of Drug Metabolism (Estabrook, R.W. and Lindenlab, E., eds.), pp. 471-483, F.K. Schattauer-Verlag, New York 132 Kakunaga, T. (1975) Cancer Rca. 35, 1637-1642 133 Kociba, RJ., Keys, D.G., Beyer, I.E., Carreon, R.M., Wade, C.E., Dittenber, D.A., Calnins, R.P., Frauson, L.E., Park, C.N., Bernard, S.D., Homme!, R.A. and Huminston, C.G. (1978) Toxicol. App!. PharmacoL 46, 279-303 134 Poland, A. and Glover, E. (1979) Cancer Rca., in the press 135 Solt, D. and Farber, E. (1976) Nature 263, 701-703 PAGENO="0053" 49 215 136 Solt, D.B., Medline, A. and Farber, E. (1977) Am. J. Pathol. 88, 595-618 137 Laishes, BA., Ogawa, K., Roberts, E. and Farber, E. (1978) J. Nat!. Cancer Inst. 60, 1009-1016 138 Waeanabe, K. and Williams, G.M. (1978) J. Nati. Cancer Inst. 61, 1311-1314 139 Kitagawa, T., Pitot, H.C., Miller, E.C. and Miller, J.A. (1979) Cancer Res. 39, 112-115 140 Goodall, C.M. (1966) Cancer Res. 26, 1880-1883 141 Trosko, S.G. and Chang, C.C. (1979) Adv. Rad. Biol. 8, in the press 142 Yamasaki, H., Fibach, E., Nude!, U., Weinstein, I.B., Rifkind, R.A. and Marks, P.A. (1977) Proc. Nat!. Acad. Sci. U.S.A. 74, 345 1-3455 143 Huberman, E. and Callaham, M.F. (1979) Proc. Nat!. Acad. Sci. U.S.A. 76, 1293-1297 144 Diamond, L.,O'Brjen, T.G. and Rovera, G. (1979) Life Sci. 23, 1979-1988 145 Trosko, J.E., Chang, C., Yotti, C.P. and Chu,E.H.Y. (1977) Cancer Res. 37, 188-193 146 Mastro, A.M. and Mueller, G.C. (1974) Exp. Cell Res. 80, 40-46 147 Sivak, A., Ray, F. and van Duuren, B.L. (1969) Cancer Rca. 29, 624-630 148 O'Brien, P.G., Simsiman, R.C. and Boutwell, R.K. (1975) Cancer Rca. 35, 1662-1670 149 Russell, D.H. (1971) Biocheni. Pharmacol. 20, 3481-3489 150 Nishi, Y., Taketomj, M. and mu!, N. (1978) Japan J. Gen. 53, 59-62 151 Mushlin, P.S. and Peraino, C. (1974) Proc. Soc. Exp. Biol. Med. 145, 859-862 152 Daucherty, J.P., Davis, S. and Yielding, K.L.~(1978) Biochern. Biophys. Res. Commun. 80, 963- 969 153 Cleaver, J.E. and Painter, R.V. (1975) Cancer Rca. 35, 1773-1778 154 Poland, A.P., Glover, E. and Kende, A.S. (1976) J. Biol. Chern. 251, 4936-4946 155 Popper, H. (1977) Am. J. Pathol. 87, 228-265 156 Remmer, H., Bolt, H.M., Bsnnasch, P. and Popper, H. (1978) Primary Liver Tumors, MTP Press. Lancaster 157 Tornatis, L., Agthe, C., Bartsch, H., Huff. J., Montesano, R., Saracci, R., Walker, E. and Wilborn, J. (1978) Cancer Res. 38, 877-885 158 Blumberg, B.S., Larouze, B.. London, W.T., Werner, B., Hesser, J.E., Mlllman, I., Saimot, G. and Payet, M. (1975) Am. J. Path. 81, 669-682 159 Bannasch, P. and Klining, 0. (1971) Virchow's Arch. Abt. A Path. Anat. 352, 157-164 160 Knowles, D.M. and Wolff, M. (1976) Human Pathol. 7. 533-545 161 Anthony, PP. (1976) Cancer Res. 36, 2579-2586 162 Peters, R.L. (1976) in Hepatocellular Carcinoma (Okuda, K. and Peters, R.L., eds.), pp. 107- 168, John Wiley and Sona, New York 163 Rothman, K.J. (1975) in Persons at High Risk of Cancer (Fraumeni, J.F., eds.), pp. 139-150, Academic Press, New York 164 Rubin, E., Hutter, F. and Liber, C.S. (1968) Science 159, 1469-1470 165 Holtmark, A., Sundh, 3., Johanson, L. and Arratinous, E. (1979) Chem. Biol. Interactions 25, 1-6 166 McCann, J. and Ames, B.N. (1976) Proc. Nat!. Acad. Sci. U.S.A. 73, 950-954 167 Sontag, 3M. (1977) in Origms of Human Cancer, Book C (Hiatt, H.H., Watson, 3D. and Winston, J.A., eds.), pp. 1327-1338, Cold Spring Harbor Laboratory, Cold Spring Harbor 168 Tatematsu, M., Shiras, T., Tsuda, H., Nyata, Y., Shinohara, Y. and Ito, N. (1977) Gann 68, 499- 507 169 Sirica, A.E., Barsness, L., Goldswortlay, T. and Pilot, H.C. (1978) 1. Environ. PathoL Toxicol. 2, 21-28 170 Clemmeson, 1. (1977) Acta Path. Microbiol. Scand. Suppl. 261, 38-50 171 Gold, E., Gordis, L., Tonascia, I. and Szldo, M. (1978) .1. Nat!. Cancer Inst. 61, 103 1-1034 172 Kitagawa, T. and Pitot, H.C. (1975) Cancer Rca. 35, 1075-1084 173 Farber, E., Parker, S. and Gruenstein, M. (1976) Cancer Rca. 36, 3879-3887 PAGENO="0054" [CANCER RESEARCH 40. 3616-3620. October 1000] 0058,5472/80/0040-0000S02.00 50 Quantitative Evaluation of the Promotion by 2,3,7,8-TetrachlorOdibeflZO- p-dioxin of HepatocarCinogefleSis from DiethylnitroSamine1 Henry C. Pitot,░ Thomas Goldsworthy, H. A. Campbell, and Alan Poland3 McArdIe Laboratory for Cancer Research. Unr'ersity of Wiscon0~t. Maitooa. W,sconsat 53706 ABSTRACT In order to test the potential of 2,3,7,8_tetraChlorOdibenZo- p-dioxin (TCDD) as a promoter of hepatocarCinOgeflesis, rats which had received a single 1 0-mg/kg dose of diethylnitrosa- mine (DEN) following partial hepatectomy were given TCDD (0.14 and 1.4 pg/kg s.c. once every 2 weeks) for 7 months. Animals which received (a) only a single initiating dose of DEN after partial hepatectomy and no further treatment of (b) TCDD alone with no initiating dose of DEN exhibited relatively few enzyme-altered foci and no hepatocellular carcinomas. How- ever, animals initiated with DEN and then given TCDD had a marked increase in enzyme-altered foci. At Ihe higher dose of TODD, hepatocellular carcinomas were present in five of seven rats. By means of three different enzyme markers used to evaluate the phenotypes of the enzyme-altered foci, a distinct phenotype heterogeneity of the foci was noted with a shift towards phenotypes exhibiting a greater deviaticn from normal liver when TCDD was given following DEN-partial hepatectomy. Quantitation of the numbers of enzyme-altered foci was per- formed by relating measurements made from two-dimensional tissue sections to the numbers of foci per unit volume of liver using relationships established in the field of stereology. The total volume of the liver occupied by the enzyme-altered foci, but not their number, increased with the dose of TCDD admin- istered following DEN-partial hepatectomy. These studies dem- onstrate that TCDD is a potent promoting agent for hepatocar- cinogenesis. INTRODUCTION TODD', a trace contaminant formed in the commercial syn- thesis of the herbicide, ~,4,5_trichlorophenoxyacet1c acid, is an extraordinarily potent toxin and teratogen (33). TCDD is the prototype of a large series of halogenated aromatic hydrocat- bons including dibenzo-p-dioxins, dibenzofurans, a.zo- and azoxybenzenes, biphenyls, and naphthalenes which are all approximate isostereomers, produce similar biochemical ac- tions, and produce a similar and characteristic pattern of toxic responses (10, 26). These compounds are all thought to exert their toxic action by a common mechanism (24). Recently, TODD has been shown to be carcinogenic in chronic feeding expenments in rats and mice (1 5).░ Kociba et This study was supported in pan by Grants CA-OTt 75 and CA-22484 from the Nationui Cancer ivstitute and Grant E5-00985 from the Natiovoi instrtote of Enoironmentai Healttn Scences. `To whom tecoests for repents should be addressed. Recipient of Research Career De'celopment Award K-04E5-00t7. The abbreoiatioes used are: TCDO, 2.3,7.8~tetrecfrlorOd:benZwPsh0~n DEN. diethyinitrosamme i.g.. intragastsc. PA. Holmes. J. H. Rust. W. R. Richter. and A. M. Shefner. Long-term effects of TCDD and HcDO in mice and rats. Presented at the intereationoi Conference on Health Effects of Helogenated Aromatic Hydtocanbons. Nev' Yew City. free Yoris Academy of Sciences. June24 to 27. 1976. Receoed Febraary 22. 1960: accepted June 9. 1980. a!. (15) reported that a dietary intake of 0.1 pg/kg/day for 2 years resulted in an increased number of hepatocellular car- cinomas and squamous cell carcinomas of the lung, hard palate, and nasal turbinate in female Sprague-Dawley rats. Lifetime feeding of TCDD equivalent to 0.001 and 0.01 pg/kg! day produced no increase in tumor incidence in rats. At a daily dose of 0.1 pg/kg/day, TCDD produces nearly a 50% inci- dence in female rats of one of the 3 cancers listed above, making it one of the most potent carcinogenic agents known (23). The carcinogenic potency of TCDD is especially interesting in light of studies which have failed to demonstrate any covalent binding of TCDD (23. 29. 40) and have provided inconclusive evidence that TODD is a weak mutagen (41). Following the in vivo administration of [3H]TCDD to Sprague-Dawley rats, the radioactivity associated with purified DNA from liver, a maximal estimate of covalent binding, was 6 pmol of TCDD per mol of nucleotide (23). This is 4 to 6 orders of magnitude lower than the binding observed with most chemical carcinogens. It seems unlikely that TCDD-induced oncogenesis is mediated through the covalent binding of this compound to DNA and subsequent somatic mutation. Tumor development in the skin has been shown to occur in 2 stages, initiation and promotion (2, 3, 18, 35). Initiation is ati irreversible process, which results from the admintstration of a carcinogen, and it presumably involves the covalent binding of the carcinogen or an active metabolite to DNA (20). Promotion is a reversible process, influenced by many factors (3, 21, 39), which converts the irreversibly altered, initiated cell into a neoplasm. Studies by Peraino et al. (19) have demonstrated a 2-stage model of carcinogenesis in the liver by the phenobar- bital promotion of acetylaminofluorene-initiated hepatic cells. These studies have been confirmed by other investigators (14, 42) and have extended our knowledge of tumor promotion (13). Since TCDD does not appear to be an initiator (i.e., there is no conctusive evidence that it is a mutagen), we considered the hypothesis that the liver cancer associated with chronic administration of TCDD might arise from the promoting activity of the compound, presumably stimulating cells already spon- taneously initiated by dietary and other environmental carcin- ogens. To study this question, we used a 2-stage model of hepatocarcinogenesia developed in our laboratory (21, 34) based on studies of Peraino, Scherer, Emmelot, and others (1, 8, 19, 28, 31). Rats were subjected to a partial hepatectomy to stimulate cell division, and 24 hr later received a single low dose of DEN by intubation. The animals were then treated by chronic administration of a promoting agent, i.e., phenobarbi- tal, beginning 1 to 2 months later. After a single low dose of DEN, we demonstrated that chronic dietary administration of phenobarbital resulted in the hepatocellular carcinomas and a marked increase in an enzyme-altered foci (21). Such foci had 3616 CANCER RESEARCH VOL. 40 PAGENO="0055" been reported earlier to result from DEN administration (8) and are thought to be precursors of hepatocellutar carcinomas (28. 31). Animals subjected only to partial hepatectomy and DEN develop far fewer enzyme-altered foci and no liver carcinomas. For the quantitative analysis of the results per unit volume, we have used relationships established in the field of stereology (7) including quantitative stereology(38), quantitative metallog- raphy (36), and quantitative microscopy (5). The application of these established mathematical techniques to the computation of the number, volume, and size distribution of enzyme-altered foci per unit volume from measurements made on 2-dimen- sional histological sections was recognized and adapted for this purpose by H. A. Campbell. The method of Johnson (12) as extended and modified by Saltykov (30, 37) was very useful in calculating the foci size distribution of the foci permitting the estimates of foci number made from the method of Scherer Ŕt a!. (32) to be placed on a firm quantitative basis. In this report, we examine the promoting effect of TCDD on this 2-stage model of liver cancer and compare it with the effect of phenobarbital. MATERIALS AND METHODS Female Charles-River rats weighing 200 to 250 g wore subjected to a 70% hepatectomy according to the proceoure of Higgins and Anderson (11). A single i.g. intubation of DEN (10 mg/kg; Eastman Organic Chemicals, Rochester, N. Y.) in water was administered 24 hr later according to the protocol of Scherer and Emmelot (31). The animals were divided into treatment groups (Table 1); alt of the groups were subjected to partial hepatoctomy, but only 4 groups received DEN. The rats in Group 1 were given DEN and then maintained on standard RESULTS laboratory chow for the remainder of the experiment (32 weeks). The rats in groups 2 and 3 received no DEN, but starting 1 week after hepatectomy they received biweekly s.c. injections of TCDD (0.14 or 1.4 tog/kg, respectively) in corn oil~ for a period of 28 weeks. (The TCDD was provided by Dow Chemical Co., Midland, Mich., as Lot 851-144-2, and it wasP 98.6% pure as determined by gas-liquid chromatography.) Groups 5 and 6 received DEN, and 1 week later they were initiated on a regimen of 14 biweekly injections of TCDD (0.14 or 1.4 ysg/kg, respectively). The animals in Group 4 received 0.05% sodium phenobarbital in the diet starting 1 week after partial hepatectomy for 28 weeks, and the animals in Group 7 received DEN and then one week later also were given 0.05% sodium phenobarbital in the diet for the duration of the exper- iment. At the end of the experiment, the rats were killed, and sections of the liver were removed and frozen on solid CO2. Serial sections of the frozen blocks of liver were cut and stained consecutively for glucose-6-phosphatase, canalicular ATPase, and y-glutamyl transpeptidase (22) and with hematoxylin and eosin. The area of each liver section was measured using a planimeter. Photographs were taken of each histochemicall' stained section, and the number of enzyme-altered foci we determined from tracings of the photographic enlargement. made on transparent plastic. By appropriate overlaying of the 3 transparencies, one for each enzyme stain, the total number of enzyme-altered foci per cu cm of liver could be calculated by use of the formula: 1 1 1 1/ N,~-+-+-+....-/nA where N, is the number of foci per cu cm of liver; r, r2, r3, and r,, are the radii in cm of the foci transections; aed A is the area in sq cm of the liver sections evaluated. This relationship has been reported by Drinkwater eta!. (6)6 and by Fullman (9). The percentage of the enzyme-altered cells in the entire liver-pop- ulation was calculated from the fact that the ratio of the area of the foci sections to the area of the liver tissue sections is identical to the ratio of the volume of the foci to the volume of the liver (43). Hepatocarcinomas were diagnosed by standard histopathological criteria. As seen in Table 1, animals that were partially hepatecto- mizcd and then given only TCDD (Groupa 2 and 3) or pheno barbital (Group 4) developed relatively few enzyme-altere foci. Hepatectomized animals receiving only DEN developed a substantial number of foci (Group 1), but the number of such N. R. Dnnkwcter, M. R. Moore. E. c. Millei. and J. A. MOor. Mcthodsiortitn quantitatise estimatien at hintochnmicaiiy detectable foci ef aitered iivcr cells in Corcinoaen.frnoted animals, submitted for publioation. 51 TCDD Promotion of Hepetocar~Ó~J.~ Promoting effect of TCOD on 600atocarcinogenes,s by a dose of PH░ and Fnmoin rats 1200 a) were tntubated with DEN where Seven days iater. TCDD (injected s.cJ or phenobarbitai 10.05% tn the dm11 administration was begun and was continued for 28 weeks, at which time the onimais were sacnticed, and the itoers were examined. The tow high doses of TCDD were 0.14 and 1.4 yg/kg/2 weeks. respnctioeiy, admtnlstnred s.c. DEN was gionn at a dose 0110 mg/kt. See teat for further details. N.oi ~ Mean voi rats of enzyme-ai- of enzyme- % iiver cxi with cr68 foci/cu cm aitered foci occugied carci- Group Trexhnenf animais of iluer lou mmj foci name 1 PH + DEN 4 98b 309 ▒ 0.02 0.7 0 2 PH + TCDD yew done) 4 34 ▒ 17 0.05 0.2 0 3 PH + TCDD (high dose) 5 25 ▒ 7 0.04 0 1 0 4 PH + pheeobajtltej 4 se ▒ 13 0.01 0.1 0 5 PH + DEN + TODD (low dose) 5 1068 ▒ iee 0.08 9.0 00 PH + DEN + TODD (high dose) 7 ▒ ee 0.49 43.0 ~2 7 PH + DEN + pa~ix,& 10 533 ~ 103 0.15 * PH, par68 hepatectomy, Mean a 5.0. o Three rats exhibited noepiasbc nodules in the iinei. One rat exhibited 1.1 1- OCTOBER 1980 3617 PAGENO="0056" 52 H. C. Pitot eta!. foci increased approximately 3-fold when these animals then received TCDD (Groups 5 and 6) or phenobarbital (Group 7) for the next 7 months. Most significant was the production of well-differentiated hepatocellular carcinomas in the high-dose TCDD-DEN-lreated animals (Groups 5 and 6), whereas no neoplasms were observed in animals That received only DEN (Group 1) or only TCDD (Groups 2 and 3). Rats subjected to partial hepatectomy alone showed no foci (data not shown). As can be seen from Table 1, the number of enzyme-altered foci per cu cm liver in animals of Group 5 was significantly higher than that for those in Group 6. We do not feel that this indicates that the lower dose of TCDD (equivalent to 0.01 pg/ kg/day) is more effective in promoting liver tumors than the higher dose (0.1 pg/kg/day) but rather that the larger foci which occurred after 7 months on the higher dose of TCDD were the result of fusion of several of the foci, thua accounting for a lower total number in livers of animals on the higher dose for the same period of time. A comparable but somewhat lower number of foci than induced with the higher dose of TCDD was produced by the feeding of 0.05% phenobarbital following DEN-partial hepatectomy atthe dose used in these experiments (Group 7). Thus, TODD administration is similar to the promot- ing effect of phenobarbital in increasing the number of enzyme- altered foci but ikies not show any effect of dose on the number of foci produced in the ranges used in this study. Almost twice the number of enzyme-altered foci were produced by TCDD (0.01 pg/kg/day) as were produced by phenobarbital (0.05% in diet), but the total molar dose of the former was approxi- mately 1 million times less. While both dose levels of TCDD resulted in the same number of enzyme-altered foci, the higher dose level caused a marked increase in the total number of cells or volume occupied by the enzyme-altered foci in the livers of animals in Groups 5 and 6 (Table 1). This result is consistent with the concept that the promoting action of TCDD enhances the growth of the cells in the foci, which is reflected by a greater proportion of foci exhibiting phenotypes of ATPase and glucose-6-phosphatase deficiency combined with the presence of y-glutamyl transpep- tidase. Pugh and Goldfarb (27) have earlier shown that such a phenotype is characteristic of foci exhibiting the largest number of cells in DNA synthesis. It was evident from observation of the transparent overlays for the 3 enzymes that many of the foci as well as the larger carcinomas exhibited heterogenous phenotypes of altered en- zyme activities. The use of serial sections and composites of the transparent overlays for the enzymes allowed us to quari- titate the number of enzyme-altered foci of each of the 7 possible phenotypes. Table 2 presents these results. The par- tially hepatectomized animals receiving DEN only (Group 1) or the TCDD only (Groups 2 and 3) exhibited a greater percentage of their foci as a single enzyme alteration when compared to the PH plus DEN plus TCDD animals (Groups 5 and 6). The percentage of foci exhibiting alterations in all 3 enzymes was 3- to 7-fold higher in the DEN-TCDD-treated animals than in the other groups. DISCUSSION In the present study, we have found that enzyme-altered foci induced in rat liver by partial hepatectomy and DEN were greatly increased in number, total volume, and phenotypic a: `~ e e ~ 0 al `E -U:-' ~ ~ ~ `~ ~ ~a o ~ r. -:~`-~:d~ C ix 1- C 3618 CANCER RESEARcH VOL 40 PAGENO="0057" 53 ~ TCDD Promotion of Hepatocarcinoc'p~j,e~j~ heterogeneity by the administration of TCDD. A significant 9. Fuliman. R. L Measwement ofparticlesgzesin opaquebodies. Trans. AIME. incidence of hepatocellular carcinomas (5 of 7) was observed ~ ~ reIat~onsh,psofhaIogonat~ b~pheny~s in the DEN-treat~ij rats which were given the high dose of ~ ~ induces Ann N V Acad Sd . 320 164-178. 1979 TCDD (1 .4 ~zg/kg biweekly), but no carcinomas were seen in I 1. Hia9ins. G. M.. and Anderson, R. M. Eapenmentai pathoioay of the liver. i. the rats treated only with DEN (0 of 4) in confirmation of ~ ~ Ph i ~ 196 2g~h19~ht I I ii w a P vii i gcai m previous results (21 , 34). The rats and the TCDD dosage 12. Johnson, W. A. Estimation ot spaciai grain size. Metai. Prog.. 49: 87. 1946 regimen used in this study were chosen to resemble closely t3. Kimura, N. 7., Kanematsu, T.. and Baba, 1. Poiychlorinated bipherylisi as the conditions in the 2-year feeding study by Kociba eta!. (15). ~;romoter~n ~p7e6rimentai hepatocarc~nogenesg a rats. Z. Krebstorsch., The rats in the present study, initiated with DEN and then given 14. Kitagawa, T.. and Sugano, H. Enhancing enect ot phnnobarbitai on the TODD, developed a much higher incidence of liver cancer in a much shorter time period (28 weeks) than did those maintained Gann, 60 679-697, 1978 on a diet of TCDD for 104 weeks in the study by Kociba at a!. 15. Kociba, R. J.. Keyes. D. 0.. Bayer. J. E.. Carreon. R. M.. Wade, C. B., In the absence of convincing evidence that TCDD is a mu_ tagen or that it covalently binds to DNA to any appreciable and oncogenicity study of 2.3,7,8-tetrachiOrodibenzo.p,jjooin in rats. Tax- extent and in light of the present results that TCDD enhances icoi. April. Phanmacol., 46: 279-303. 1978. DEN-initiated hepatic carcinoma, it seems a reasonable hy-~ 16. M on J.andA s. B. N. The Sa.imvnelia/nrlcrosom. ~a~ncltytes~ pothesis that all the tumors associated with the chronic admin- J. A. Winsten (ads.), Origins ot Human Cancers, Book C. pp. 1431-1450 istration of TCDD arise from its promoting activity of cells 17 CoidSpnngHarftorNy Cold spring Harbor Laboratory, 1977. already initiated by exposure to the environment pot cy i H H Hi a J D W tso and J A WI ten (ad t OrIg I Boutwell (4) has suggested that promoting agents act to alter Human Cancer, Book C, pp. 1473-1481. Cuid Spnng Harbor, N. 7.: Coid gene expression, and studies of one of the best known pro- 18 ~rberabor:g in eopenmentai biastogenesru J Pathoi moting agents in skin, tetradecanoyl phorbol ester, have re- Bactenoi , 56 181-197, t944 peatedly demonstrated its relative metabolic inertness and lack 19. Peruino. C.. Fry. R. .1. M.. 5tatteidt, E.. and Christopher, J. P. Comparative of covalent interaction with DNA In conformity with this con ~ hi ph th b rbt mbarbt i~1 dph yihydt cept, TCDD has been shown to bind reversibly to a specific tumongenesis or the rut. Cancer Res., 35: 2884-2890. 1975 cytosol receptor, and the ligand-receptor complex initiates the 20. Pitot, H. C. Biologicai and enzymatic events or chemrcai carcinogenesis. Anna. Rev. Med., 30: 25-39, 979. coordinate expression of a number of genes (25). 21. Pitot, H. C., Barsoess, L, Goldsworthy, T., and Kitagawa. T. Biochemical The characteristic toxic responses of TCDD have recently characterication ot stages of hepatocarcinogenecsafteras~ngle dose of been shown to be mediated through the cytosol receptor (24), 22. promohon in and it is possible that its action as a promoter of hepatocellular hepatocarcinogenesis. Biochim Biophys. Acta, 605 191-215, 1980 carcinogenesia may also be mediated by its stereospecific 23. Poland, A., and Glover. E. An estimate of the maoimum iv `iu'o covalent - binding to the receptor and the coordinated gene expression ~ 4ONAC t bosom that ensues. The extreme effectiveness of this compound in its 24. Poised, A., and Giocer, 6. 2.3.7.9-Tetrachiorodibenzo.p.ijiooin segregatior, promoting action suggests that the relative strength of other of toxicity with the Ah locus. Moi. Phannacoi., 17: 86-94, 1980. promoting agents in the liver and probably other organs will 25 P~ ~ ~J7A~Giove ad b K04~ A 5 ~3te~eo ~ec ~ h?irj ~ vary by many orders of magnitude just as can be seen in the 25t: 4936-4946, 1976. potency of chemical carcinogens (16). Furthermore, such an 26. Poland, A., Greerlee. w ~ and Kend:. A. S. studies on tfre rnectranism ot effective promoting agent might well be expected to be able to Acad Sd., 320214-230, 1979 promote cells initiated by ambient environmental conditions 27. Pugh, T. 0.. and Goldtarb, S. Quantitative histochemical and autoradi- such as diet background radiat on or other factors (Table 1 r~IJ byph nori~tt~can R 38 44504457 1978 Groups 2, 3, and 4), giving its effects the semblance of a 29. Rabes, H. M.. and Szymkowiak, R Call kinetics of hepatocytes during the complete carcinogen. prxneoplasflc period 01 dixtvyln,trosam,ne-ingucer) liver carcinogenesis. Cancer Res.. 39:1299-1304, 1979. EFE E ES 29. Rose, J. 0., Ramsey, .1. C., Wentzier, T. H.. Hummel, 8. A., and Gehnrg, P. J. The fate of 2,3.7.8-fetrachlorodlbenzo.p.diooins following single and repeated oral doses to the rat. Toxicoi. Appi. Pharmacoi., 36: 209-226, 1. Bannasch, P. Cytology and cytogenews of neoplasiic(hyperpiagtic)hepafm 1976. nodules. Caacer Ran,, 36:1298-1304, 1976. 30. Saifykov, 5. A. Stereometnic Metailograpiry. Ed. 2. Moscow: Mefailurgszdat, 2. Berenblum, i., and Shubik, P. A new quanfltaove approach to the study of 1958. the stages of chemical carcinogenesis in the mouses skin. Br. J. Cancer, 1: 31. Scherer, E., and Emmelot, P. Kinetics of induction and growth of precan- 393-391, 1947, cerous liver-cell lxiv, and liver tumor tormafion by diethyinitrosamine in the 3. Boutweli, R. K. Some biological aspects of skin carcinogenevs. Frog. Eop. rat. Ear. J. Cancer. It: 689-696. 1975. Tumor Ran,, 4. 207-250, 1964. 32. 5cherer, 6., Hoffman, M., Emmeiot. P., and FrIedrich-Freksa. H. Quantitative 4. Boutwell, R. K, Thefunction and mechanismxfpromgfewofcarcinogenesis study on foci of altered liver cells induced in the rat by a single dose of Cdt. Rev. Tooicoi., 2. 419-443, 1974. diethylnihosamire and parbal hepatectomy. J. Nati. Cancer Inst., 49: 93- 5. DeHoff, R. T., and Rhinee, F. N. Ouanhtative Microscopy. New York: Mc- 106. 1972. Graw-Hill Bock Co., 1969. 33. 5Ch~et~, B. 5., Norris, J. M., Sparschu, 0. L, Rowo, V. K., Gehring, P. J., 6. Drinkwafxr, N. P. Physical and chemical eRects of carcinogen binding to Amerson, J. L. and Gerbig. C. G. Toxicology of chlorinated dibenzo-p- DNA in relation to biological activity and statistical problems in chemical dioxins. Environ. Health Perspect., 587-09. 1973 darcinogenesis. Ph.D. Disserrahon, university of Wisconsin at Madison, 34. Sirica, A. E.. Baroness, L, Goldsworthy, T., and Pilot, H. C. Dehnition t980. stages during hepatocarcinogenesis in the rat: potertiaf application to 7. Ellas, H. (Ed.). Stereology. Proceedings of the Second internatianai Con- evaluation of iritiating and promoting agents in the environment. J. Environ. grass for Sterexiogy, Chicago, 1967. Springer-Verlag Berlin, Heidelberg, Pathol. Toxicol., 2:21-28. 1978. New York. 35. Sivak, A. Co-carcinogeresis. Blochem. Biophys. Acta. 560: 67-89, 1979. 8. Friedncfr-Freksa, H., Papadopulu, 0., and Gossner, W. Hlstochemisohe 36. Underwood, E. E. Applicationsof quantitative metallography. Metals Han-lb., untersuchungen den Cancerogense in 4cr Rattenleber ouch zeitlich be- 8:37-47, 1973. grenzter Verabfoigung von Dinthyinirrosamin. Z. Krebsforsch., 72: 240- 37. Underwood, 6.6. Particle-size distribution. In: R. 7. DefloR and F N. Rhines 253. 1069. teds.), Quantitative Microscopy, pp. 149-199. New York: McGraw-Hill Bock OCTOBER 1980 3619 PAGENO="0058" H. C. Pitot eta!. 54 Co.. 1968. 38. Underwood. 8.8. Quantitative Steteology. Reading. Mass.: Addison-Wesley Publishing Ca.. 1970. 39. Van Duuren. 8. L. Smith. A. C.. and Melchionee. S. N. Effect of agmg tao-stage carcinogenesis on mouse skin seth phorbol mynstate acetate as promoting agent. Cancen Res.. 38: 865-866. 1978. 40 Vinopal. J. H.. and Casida. J. E. Metabolic octicity of 2.3.7.6-tetnoChlxrodc- benzo-p-diOo~n in mammalian liver microsomal systems and n l:ung mIce. Arch. Environ. Contamin. Toxicol.. 1:122-132. 1975. 41. Wasson. J. S.. Huff. J. E.. and Lopriano. N. A. A review on the genehc toxicology of chlorinated dibenzo-p-diooins. Mutat. Res.. 47: 141-160. 1977. 42. Weisbcrger,J. H.. Madison. R. N.. Ward. J. M.. Viguera. C.. and Wersburger. E. K. Modihcotion of diethylnitrosami ne livercatcinogenesis auth phenobar- bital but notoith immunosuppresgicn. J. NatI. Cancerlnst.. 54: 1185-1188. 1975. 43. Wicksell. S. D. The corpuscle problem. A mathematIcal stcdy of a blometcc problem. Biotrretnika. 17: 84-09. 1925. 3620 CANCER RESEARCH VOL 40 PAGENO="0059" 55 Journal of Supramolecular Structure and Cellular Biochemistry 17:133-146 (1981) Mechanisms of Chemical Carcinogenesis 141-154 The Natural History of Carcinogenesis: Implications of Experimental Carcinogenesis in the, Genesis of Human Cancer Henry C. Pitot, Thomas Goldsworthy, and Susan Moran McArdle Laboratory for Cancer Research, Departments of Oncology and Pathology, The Medical School, University of Wisconsin, Madison, Wisconsin 53706 Although the long latent period after administration of a carcinogen until development of a cancer has been recognized for more than a hundred years, until the last four decades little consideration had been given to the phenomena occurring during the latent period itself. Systems in which to study the molecular mechanisms underlying the phenomenon of latency have not been available to the investigator until relatively recently. Furthermore the importance of taking into account the natural history of neoplasia in whole animal lioassay procedures used for carcinogen testing is still not appreciated. The comparison of tumor- bearing test animals with controls and, in some instances, the time from the initial administration of the test agent until the appearance of the first neoplasm are the principal data from which conclusiOns about bioassays are drawn. It is clear that we do not understand all the biological changes that occur during the latent period before the development of any neoplasm. The beginnings of an experimental basis for the biological changes occurring during the latent period were initiated with the studies of Rous and Kidd [1], Mottram [2], and Berenblum and Shubik [3]. However, even these studies told little of the detailed biology and far less of the molecular biology of the earliest changes occurring in cells initiated by carcinogens, since the endpoint of these experiments was the appearance of grossly visible neoplasms. Although "preneoplastic" lesions had been described both in experimental [4, 5] and in human neoplasia [6, 7], not until the last decade was it experimentally feasible to quantitate the number of such lesions. Such quantitation was first successfully carried out with "preneo- plastic" lesions during hepatocarcinogenesis following diethylnitrosamine adminis- tration [8]. Furthermore, studies from human pathology [9-11] suggested that, with certain neoplasms, the progeny of some initiated cells never developed into neoplastic foci but rather regressed and disappeared into essentially benign, even normal, tissue cells. Received April 2, 1981; revised and accepted July 27, 1981. 0275-3723/81/1702-0L33$04.OO ę 1981 Alan R. Liss, Inc. PAGENO="0060" 56 134:JSSCB Pitot, Goldsworthy, and Moran STAGES IN THE NATURAL HISTORY OF CARClNOGENESIS-~ DEFINITIONS In order to consider the implications of the natural history of carcinogenesis in relation to our knowledge of human cancer, it is necessary to define and understand the natural history of the development of neoplasia. Since the l940s carcinogenesis in mouse skin has been divided into the stages of initiation and promotion. Later Foulds, largely on the basis of his studies of mammary carcinogenesis in the mouse [12], proposed the term progression for virtually all of the developmental stages following the initial event in the conversion of a normal to a neoplastic cell. While Foulds saw the natural history of carcinogene- sis as a continuous event that could be arbitrarily divided into several phases, modern oncologists take the position that the process of promotion is distinct from that of progression even though each of these phases has been divided into several steps by previous investigators [12, 13]. Furthermore it is now apparent that the two-stage concept of carcinogenesis as originally proposed [1-3] and reviewed and extended by Boutwell [14] is applicable to a variety of tissues during their conversion to malignant neoplasms [15]. Therefore we can consider the characteristics and definitions of each of the stages in the natural history of carcinogenesis as applicable to virtually any cell type. For the purposes of this discussion, we will divide the natural history of carcinogenesis into three stages: initiation, promotion, and progression. A simplified diagram of this process is given in Figure 1 [16]. The definitions of initiation and promotion listed below are excerpted from the same reference. Initiating Agent- a chemical, physical, or biological agent that is capable of directly altering irreversibly the native molecular structure of the genetic component (DNA) of the cell. Such alteration(s) may be the result of a covalent reaction of DNA with the initiating agent itself or with one of its metabolites, but this alteration may also include a distortion of the structure of DNA without covalent binding of the agent to DNA. Finally, the agent may cause one or more complete scissions of the DNA chain, an elimination of one of its component parts (eg, bases or sugars), or errors in DNA repair. All such capabilities of an initiating agent, however, do not in themselves prove that alteration of DNA is the only or the absolute requirement for the neoplastic transformation. Promoting Agent- an agent that alters the expression of genetic information of the cell. Examples of such agents include hormones, drugs, plant products, etc, which in themselves do not directly react with the genetic material but rather affect its expression by a variety of mechanisms, including their interaction with cell surface receptors or with cytoplasmic and nuclear components and functions. The definition of an initiating agent is made in reference to molecular species, especially DNA, because of the advances in molecular biology and our understanding of the mechanisms of mutational events. While the definition clearly hedges in stating that initiation may not always be the result of mutation, there is no question that the initiation of the neoplastic transformation and genetic mutation are closely related in the majority of instances of carcinogenesis. Unfortunately, the definition of a promoting agent given here is still relatively 142:MCC PAGENO="0061" 57 Natural History of Carcinogenesis JSSCB:135 inexact. However, one may hypothesize that promoting agents may be divided into specific and nonspecific classes. Specific promoting agents are those that interact with receptors or receptor-like molecules on or within target cells. Such specific promoting agents have a defined range of tissues susceptible to their promoting action. Examples of this class would be steroid or polypeptide hormones [15], which are known to be effective promoting agents in their target tissues, their metabolic effects being mediated by cellular receptors. Nonspecific promoting agents are those that do not act through receptor mechanisms but alter gene expression by a variety of nonspecific mechanisms. Examples of this class would be iodoacetate or detergents in the case of epidermal carcinogenesis. In his earlier work Foulds suggested that there are at least two basic characteristics of progression [12]. The first is the independent progression of neoplasms; ie, progression occurs independently in different primary neoplasms within the same host. The second characteristic is the independent progression of specific characteristics of the neoplasm, each of which undergoes progression independently of the others in any single neoplasm. These characteristics include growth rate, invasiveness, metastatic frequency, hormonal responsiveness, morphologic characteristics, etc. As can be seen from Figure 1, another crucial characteristic of progression, as defined herein, is karyotypic change. The following operational definition of the stage of progression in neoplastic development will be used here. -~ Progression - that stage of neoplastic development characterized by visible karyotypic alterations as evidenced by light microscopic techniques within a majority of the neoplastic cells that make up~the tumor. These karyotypic alterations in turn are associated with increased growth rate, increased invasiveness, metastases, and alterations in biochemical and morphologic characteristics of the neoplasm. IMPLICATIONS OF STAGES IN THE NATURAL HISTORY OF CARCINOGENESIS A variety of implications are derived from the above definitions of initiation, promotion, and progression. The definition of an initiating agent as one capable of covalent interaction with DNA, or any other macromolecule, implies but does not prove that one or more mutational events in the genome result in the conversion of a normal cell to an initiated cell. In contrast, evidence from chimeras produced by transplantation of malignant cells into blastocytes [17], the transplantation of nuclei from neoplastic cells into eggs which then exhibit normal development [18], and the forced terminal differentiation by chemical means of a variety of neoplastic cell lines (19, 20] argue that permanent genetic damage is not necessary for the initiation of neoplastic transformation. Various theories have been proposed that argue for a permanent alteration in the initiated cell resulting from extragenomic changes [21-25]. Despite these latter considerations, the most common working hypothesis is that initiation does involve covalent and/or structural changes in the genome. The exact molecular mechanism of action of promoting agents has not, however, yet been defined. The definition above clearly suggests a mechanism for promoting agents-that of altering gene expression [141-but the variety of such MCC:143 PAGENO="0062" 58 136:JSSCB Pitot, Goldsworthy, and Moran Fig. 1. The natural history of neoplastic development in relation to initiation, promotion, and progression in reference to cell karyotype. mechanisms is so great as to suggest that this definition is too general to deter- mine the ultimate action of promoting agents in carcinogenesis. Yet many of the characteristics of promoting agents (see below) correlate well with this definition. That progression is a result of karyotypic abnormalities may be construed as a somewhat arbitrary definition. However, virtually all neoplasms that exhibit characteristics of metastases, high rates of growth and invasiveness, high rates of glycolysis, and anaplastic morphologic characteristics are aneuploid, suggesting that such an operational definition is reasonable. The final common pathway of the natural history of carcinogenesis, the metastatic lesion, is nearly always the result of the growth of aneuploid neoplastic cells. The concept of Goldenberg [26] that progression may be due in part to cell fusion further supports the importance of aneuploidy and chromosomal abnormalities in progression. CHARACTERISTICS OF THE STAGES IN THE NATURAL HISTORY OF CARCINOGEN ESIS Until the last decade the characteristics of the stages of initiation and promotion were based exclusively on experiments carried out with mouse skin as the test tissue. Although it had become apparent from studies of the pathology of human cancer, as well as from several experimental situations [15], that the two- stage process also applies to the genesis of neoplasms other than those in the skin, only in the last decade have experimental systems amenable to study and in some ways superior to the mouse epidermis model been exploited. Also, Foulds' concept of tumor progression was based on studies of yet another experimental Liver metastases EUPLOID CELL POPULATIOM INCREASING AHEUPLOIDY 144:MCC PAGENO="0063" 59 Natural History of Carcinogenesis JSSCB:137 system, mouse mammary carcinogenesis [27]. Therefore, our knowledge of the natural history of neoplasia can now be developed from and applied to a wide variety of tissue systems. Although it is possible that the natural development of neoplasia in each tissues exhibits unique characteristics, it is to be expected that certain characteristics will be common to each stage during the development of all types of neoplasms. These characteristics are reviewed here. CHARACTERISTICS OF INITIATION AND THE INITIATED CELL The characteristics of initiating agents in skin and their comparison with liver have been previously reviewed [15]. The accumulated experimental evidence supports the concept that the effects of initiating agents on cells are essentially irreversible. Furthermore, agents capable of initiating the neoplastic transforma- tion in vivo or in vitro can be divided into two general classes. Those agents capable not only of initiating neoplasia but also of causing promotion and progression of the initiated cell are termed "complete carcinogens." Those agents capable only of initiating cells but not of promoting them are termed "incomplete carcinogens" or "pure" initiating agents. Once a cell has been initiated it will remain so throughout its life-span, and the characteristics of initiation will be transmitted to all daughter cells, unless the initiated lesion is repaired or elimina- ted by some other mechanisms [28]. At least theoretically, and in some instances experimentally [29], initiation can result from a single "hit" of the initiating agent on the target cell. The irreversibility of the effects of an initiating agent, the single "hit" concept, and the corollary - ie, the additive Of the effects of initiating agents-are applicable to both incomplete and complete carcinogens. Furthermore, these characteristics are identical with what one would expect for a mutagenic agent. The apparent efficiency of initiation varies widely, depending on the system employed. Sachs reported that treatment of a mixed cell population of normal hamster embryo cells with carcinogenic hydrocarbons resulted in a 3-20% incidence of transformation in these cells [30]. However, x-irradiation resulted in only a 0.5% transformation rate. In contrast, in the mouse skin system, the average number of tumors produced in any animal is usually less than 25 [13]. On the assumption that each neoplasm arises from a single neoplastic cell [311 and that all epidermal cells are targets for the carcinogen, this means that the incidence of initiation is in the range of 10' to 10-8. In the liver system, a single dose of diethylnitrosamine will, on the average, initiate 1 in l0~ to 10~ hepato- cytes [32]. Although it appears that all initiated cells in vitro and in the skin develop into neoplastic foci, this is not so clear in the liver system. Of the foci produced (about 1,000/g liver), less than 1% develop into histologically defined neoplasms. Preliminary investigations from our laboratory have indicated, however, that it is unlikely that all foci are capable of transplantation into syngeneic hosts [33]. If this is true, then, as the term "neoplasia" is defined and characterized [16], it is likely that all such enzyme-altered foci in rat liver are neoplastic or at least potentially neoplastic. Obviously, with low doses of a complete carcinogen or with the administration of incomplete carcinogens (pure initiators), no neoplasms result, although significant numbers of enzyme-altered foci may appear under the conditions of the experiments [32, 33]. MCC:145 PAGENO="0064" 60 138:JSSCB Pitot, Goldsworthy, and Moran The latter point then raises the issue of the ultimate fate of initiated cells. Although the apparent incidence of initiation in mouse skin is extremely low, on the basis of the findings with liver it is quite possible that many more initiated cells occur in mouse skin but, as in rat liver, that these do not become neo- plasms, even following promotion, and the initiated cells and/or their progeny remain in the animal for life. In liver such initiated foci can be identified by suit- able histochemical means [32]; once such foci are produced they also do not dis- appear during the life-span of the animal [34]. Incomplete carcinogens (pure initiating agents) have been identified both in skin [35] and in liver [36] carcinogenesis. Agents that are incomplete carcinogens for these tissues are either complete carcinogens in other tissues (eg, urethan for liver and lung, and dimethylbenzanthracene for skin) or are noncarcinogenic in the adult. However, incomplete carcinogens are themselves mutagenic or may be metabolized to a mutagenic form by liver. Urethan induces hepatocellular carcinomas, pulmonary adenomas, and lymphomas in mice but does not by itself cause epidermal carcinoma [37]. Promotion by croton oil of the skin of animals given urethan results in epidermoid carcinoma [35]. Similarly, polycyclic hydrocarbons that effectively induce epidermoid carcinoma do not induce any neoplastic response in the liver of adult rodents. However, if a mitotic stimulus is applied to the liver following parenteral administration of the hydrocarbon, with subsequent promotion by phenobarbital, then heptocellular carcinomas will result [34,38]. The induction of enzyme-altered foci following short-term promotion by carbon tetrachloride has been reported for alkylating agents such as N-methyl-N- nitro-N-nitrosoguanidine and the colon carcinogen 1 ,2-dimethylhydrazine [39]. Although these latter experiments were not carried to the formation of neoplasms, on the basis of earlier arguments that such foci are initiated hepatocytes [15,33], these agents exhibit the characteristics of incomplete hepatocarcinogens. It is clear, therefore, that incomplete carcinogens, capable only of initiating cells in specific tissues, do exist and appear to have most, if not all, of the characteristics of mutagenic agents. These compounds appear to differ from complete carcinogens in that they exert virtually no promoting action on the cells of the tissue in which they serve as incomplete carcinogens. The reason for this is not clear, but one possible component is their failure to induce an increase in cell division in the tissue in which their action is incomplete. THE ROLE OF CELL DIVISION IN THE INITIATION OF NEOPLASIA Borek and Sachs [40] were the first to demonstrate in a relatively unequivocal manner that cell replication was required for the "fixation" of the transformed state in cell culture. Although their studies were initially based on the time required for fixation rather than an actual demonstration of a require- ment for DNA synthesis and mitosis, later studies [41, 42] have supported their interpretation of earlier studies. Both in the liver and in the skin, the pro- cess of promotion involved cell division, although, as has been pointed out by Boutwell and others [14], cell division is a necessary but not sufficient charac- teristic of tumor promotion in the skin. In the liver the 2-stage system of carcinogenesis described by Peraino et al [43], Solt and Farber [44], and our laboratory [32] always involves a higher than normal level of cell division. In 146:MCC PAGENO="0065" 61 Natural History of Carcinogenesis JSSCB:139 Peraino's experiments weanling animals~having a relatively high rate of hepatic cell division were utilized, whereas both~Farber and our group include the stimulus of partial hepatectomy in the initiation-promotion sequence. Ying et al have reviewed the necessity for cell proliferation as an obligatory step in the induction of enzyme-altered foci in hepatocarcinogenesis [45]. Thus there is substantial evidence that fixation or cell division is required during a relatively early period following application of the initiating agent for optimal efficiency of initiation. If one views initiation as a mutagenic event, this interpretation is entirely plausible, since one would expect that repair of macromolecular damage, the initiating event, could occur if no cell division intervenes to perpetuate the initial damage produced. CHARACTERISTICS OF TUMOR PROMOTION AND PROMOTING AGENTS There is no evidence that promoting agents exert their effects by direct covalent interaction with the genome. The available evidence suggests that the effects of promoting agents are on one or more extragenomic processes, which may in turn influence genetic information or its expression (see above). The early studies of Berenblum and Shubik [3] indicated that administration of the promoting agent alone results in no neoplasms in the mouse skin system. However, later investigations have demonstrated that, following the prolonged administration of croton oil or its active component, tetradecanoylphorbol acetate (TPA), a small number of neoplasms is produced. This has led some investigators [46] to propose that promoting agents are merely weak complete carcinogens. If this were so, one would expect that increasing doses of a promoting agent would lead to increasing numbers of neoplasms. As we shall see below, this is clearly not so in the mouse skin system. Furthermore, substantial evidence has accrued since those earlier studies that the efficiency of promotion is a function of diet and of hormonal, environmental, and other factors in the host [15]. Most recently Van Duuren et al [47] have demonstrated that, with increasing age of the animals, the efficiency of tumor promotiOn in the mouse skin system is significantly decreased. DOSE RESPONSE TO PROMOTING AGENTS The evidence for a "no threshold" lßvel and the irreversibility of complete carcinogens has been well documented [48-50]. One of the earliest examples is that shown in Figure 2, which describes the results of Druckrey [48] with an aromatic amine carcinogen, 4-dimethylaminostilbene, in rats. As can be seen from line 1, there is a linear relation between dose and tumor response, which extends through the origin. At extremely low doses (line 2), however, the time until the first neoplasm appears extends beyond the life-span of the animal. In the case of promoting agents, earlier studies suggested that, although a dose-response relation occurred, the no-threshold effect did not necessarily apply to these materials. Recently Verma and Boutwell [51] have reported a dose- response curve for TPA in mouse skin carcinogenesis. In their studies a distinct MCC:147 22-143 O-83----5 PAGENO="0066" 62 140:JSSCB Pitot, Goldsworthy, and Moran Fig. 2. Dose-responSe relationships seen in the chronic feeding of 3,4~dimethylaminostilbene to rats. 1. Relationship between the daily dose and the median total dose of animals with carcinoma. 2. Relationship between daily dose and median induction time. The abscissa shows the daily dose, whereas the ordinate on the left is the total dose administered, and that on the right is the time from the beginning of the experiment. All scales are logarithmic. (Modified from Driickrey et al, 1967 [481). threshold was obtained below which no tumors occurred. At the two highest doses of promoting agent employed, maximal incidence of tumors was the same. Thus both a threshold and a maximum effect of the promoting agent were noted. Neither of these characteristics would be expected with a complete carcinogen. More recently Peraino and his associates [521 have established a dose-response relation for phenobarbital administration following initiation of hepatocarcino- genesis by a short feeding of acetylaminofluorene. In those studies a distinct threshold was also noted, as well as a maximum, when the total incidence of hepatic tumors was considered. In our studies on the quantitation of enzyme- altered foci, a maximal number of foci is achieved at doses of phenobarbital in the diet above 0.01%. At extremely low doses (.0001%) no difference in the incidence of foci was noted compared with control animals. Thus it is apparent from three separate studies, using three different endpoints of analysis, that promoting agents exhibit a threshold dose below which no effect is noted, as well as a dose above which no further effect on incidence of tumors or foci is noted. Both of these characteristics clearly distinguish promoting agents from complete carcinogens, whether weak, moderate, or strong. REVERSIBILITY OF THE PROMOTION STAGE Boutwell was the first to describe the nonpermanence of the effects of promoting agents [13}. Using the mouse skin system with croton oil as the promoting agent, he demonstrated that changing the format of the dosage regimen could alter the incidence of tumors finally produced. When the promoting agent was applied once every 4 weeks rather than 3 times per week, 148:MCC PAGENO="0067" 63 Natural History of Carcinogenesis JSSCB:141 but with the experiment extended until the same total dosage of promoting agent was given under both regimens, tumors resulted only in those animals receiving the promoting agent thrice weekly. These studies clearly demonstrated the re- versibility and non-additivity of the effects of croton oil in promoting epidermal carcinogenesis. Preliminary experiments in~the hepatocarcinogenesis system employed in our laboratory [33] have supported the findings of Boutwell in that the admin- istration of phenobarbital for 2 days every 2 weeks rather than continuously, as in the control animals, but with ~he same total dose in both groups, resulted in significantly fewer enzyme-altered foci in the animals that received the promoting agent at 2-week intervals. Thus in these two different systems one can demon- strate the reversibility of the effects of promoting agents, another characteristic clearly distinguishing them from complete carcinogens or pure initiating agents. MODELS FOR THE STUDY OF MOLECULAR MECHANISMS OF TUMOR PROMOTION Table I is a list of the model systems exhibiting a 2-stage mechanism of carcinogenesis. Of these systems, those most commonly studied are from the mouse epidermis and rat liver in vivo and mouse and hamster cells in culture. In addition, two other well-studied systems are those of the rat bladder and mam- mary gland. In the case of mouse epidermis, mechanistic studies were greatly facilitated by the isolation and purification of the active ingredient of croton oil, tetra- decanoylphorbol acetate (TPA), the classical promoter for mouse epidermis [2,3]. Much is now known of the biochemical actions of TPA, the highly active phorbol diester of croton oil. Its action and effects have been reviewed [53] and are the subject of many of the papers at this symposium. The difficulty in studying specific effects of TPA on a variety of cellular systems both in vivo and in vitro is the extrapolation of such results to the phenomenon of promotion in the mouse skin or other tissues in which TPA has been shown to promote tumor- igenesis. Since TPA also stimulates a significant inflammatory response in the epidermis, and the role of this in tumor promotion is unknown, the biochemical actions of TPA related to inflammation may or may not be important in the mechanism of promotion. While mouse epidermis is readily accessible to experi- mentation, the tumor-promoting effects of TPA and other agents used in this model system can only be ascertained through the induction of benign and malignant neoplasms seen grossly on the skin. Following the report by Berwald and Sachs [54] and later the establishment of the C3H 10 TV2 transformable cell line by Heidelberger and his associates [55], cell culture has offered some of the most promising systems for the study of the molecular mechanisms of tumor promotion. Unfortunately, the exact significance of morphologic transformation in cell culture by carcinogens is not fully understood, and in the few epithelial systems available it is not, of itself, sufficient to define the cell as biologically neoplastic [56]. However, the ability to manipulate the cellular environment and the ready access of the system to the investigator make this model one of the most promising now available for the study of the molecular mechanisms of tumor promotion. Unfortunately, because MCC:149 PAGENO="0068" cI~ ct~ C,) lB C,) 0 0. 0 53 0 0. 0 53 0 us TABLE 1. Initiators, Promoters, Tissue Initiating agent "Preneoplastic" lesions Promoting agent Dog bladder 2-Naphthylamine Alkaline phosphatase-deficient foci D,L-tryptophan Saccharin Rat bladder Methylnitrosourea Allopurinol Rat bladder N.[4]-(5-nitro-2-furYl)-2- Rat colon Rat bone marrow thiazolylformamide N~methyl~N'-nitrosogUaflidine N,N'~2,7-fluorenYlbisaCetamidC Proliferative foci Lithocholic acid Blood loss (leukemia) Tetradecanoylphorbol Mouse embryo 3-Methylcholanthrene acetate fibroblasts Tctradecanoylplsorbol in culture Ultraviolet radiation acetate Croton oil or tetradecanoyl- Mouse epidermis ~ 3-Methylcholanthrene, ~-propiolactone, urethan, etc phorbol acetate Croton oil or lime oil Mouse forestomach 3.Methylcholanthrene, benzo(a)pyrene, dimethylbenzo(a)anthracene diethyl- Hyperplastic nodules Phenobarbital, DDT, PCBs, Rat liver 2-Acetylaminofluorene, nitrosamine, azo dyes Enzyme-altered foci butylated hydroxytoluene, estrogens Butylated hydroxytoluenc Mouse lung * Urethan Ductular hyperplasia Phorbol, prolactin Rat mammary gland 7,l2-Dimethylbenzo(a)anthracene Hyperplastic alveolar nodules Rat pancreas Rat thyroid Azaserine N-Methyl-N-nitrosourea 2-Acetylaminofluorene * Acinar nodules Adenomas Methylthiouracil "Taken from Pitot and Sirica [15); the reader is referred to this source for details. PAGENO="0069" 65 Natural History of Carcinogenesis JSSCB:143 of the limitation of the use of mesenchymal cells in the most commonly employed systems, the cell culture transformation systems are relatively limited to the number of specific promoting agents that may be studied. Trosko and his associates [57] have recently demonstrated, however, that promoting agents inhibit metabolic cooperativity in cells in culture. Whether this will be an ubiquitous mechanism of all promoting agents and how such a mechanism can account for the biological action of promoting agents remains to be seen. The liver system in vivo offers the possibility of monitoring transformed cells shortly after initiation more~ so than is seen in other systems, even those in culture. The ease with which liver cells may be manipulated in vivo and in vitro, together with the extensive biochemical knowledge of this tissue, offers distinct advantages. However, the transformation system occurs only in vivo, and thus far it has not been possible to isolate in pure form the population of initiated cells for studies in culture. Furthermore, despite some reports, it has not been possible to transform adult or even fetal hepatocytes in cell culture into neoplastic cells. Thus each system has both advantages and disadvantages for the study of the molecular mechanism of tumor promotion. In the hepatocyte system, chem- ical agents such as phenobarbital, halogenated aromatics, and antioxidants are all effective as tumor promoters, and all act to regulate xenobiotic metabolism. The relationship between this effect and the actions of these compounds as promoters of hepatocarcinogenesis is not yet clear. Since this system has distinct advantages and disadvantages, the molecular biologist interested in studying the mechanism of tumor promotion must decide which aspect of tumor promotion to study- eg, the action of TPA, phenobarbital, or other promoters and the cell biology of tumor promotion in vivo. It is only through a concerted effort of studying aspects of all these systems that we will ultimately understand the molecular mechanisms of tumor promotion. IMPLICATIONS OF THE CHARACTERISTICS OF THE STAGES OF CARCINOGENESIS There are two principal implications of our knowledge of the characteristics of the stages of carcinogenesis. The first is the importance of determining the molecular mechanism of action of promoting agents. There is now overwhelming evidence that the initiation of neoplasia involves, in most instances, a direct alteration in the genetic material of the cell. \Vhether this alteration is ultimately repairable so that a neoplasm may revert to the normal state or whether the genetic alteration invariably results in malignant neoplasia is not critical to our understanding of the mechanism of initiation. However, tumor promotion appears to be regulated by environmental factors, even to the point of a reversal of the effects of such agents during the process of carcinogenesis. Promoting agents differ in their effects on different tissues and in different species, just as do initiating agents and complete carcinogens [15]. Such tissue and species specificities for complete carcinogens can be understood on the basis of the required metabolism to the active carcinogenic form in a specific tissue and the ability of the agent to induce cell replication and tumor promotion in target MCC:15l PAGENO="0070" 66 144:JSSCB Pitot, Goldsworthy, and Moran tissues. Promoting agents, however, are not readily metabolized but usually must be present in substantial amounts over prolonged periods to exert their promoting activities. The specificity of promoting agents for tissues may be related to their interaction with specific receptors in the target tissue. Substantial evidence now exists for surface receptors for the active phorbol esters, promoting agents for mouse epidermis and other tissues (58,59), and studies from our laboratory (60) have shown that tetrachlorodibenzodioxifl (TCDD) is an excellent promoting agent in liver. This latter compound interacts with a specific receptor molecule in liver and other tissues (61), an interaction that is necessary for the expression of its toxicity and, possibly, its promoting activity. These studies indicate that unless a receptor for a specific promoting agent is present, that tissue will not be promoted to a tumor by the agent. As pointed out earlier (62), virtually all hormones become promoting agents by this concept. Estrogens are effective promoting agents in liver (63), as is prolactin in mammary tissue [64]. However, some promoting agents, such as iodoacetamide, act in a nonspecific manner to alter gene expression or exert whatever other mechanistic effects are required of promoting agents for their action in carcinogenesis (vide supra). The second major implication is in relation to human carcinogenesis. Probably most important in this area is the question of testing environmental agents to determine their carcinogenicity. At present all such testing method- ologies do not distinguish among initiators, promoters, or complete carcinogens, so that all agents are treated in a similar manner. This approach is not reasonable in considering promoting agents, which on prolonged feeding may be expected to induce a significant number of neoplasms in test animals as compared with control animals. However, because of the reversibility of the effects of promoting agents during carcinogenesis and the existence of threshold levels of these agents, the risk of such agents for the human being is significantly different from the risk of mutagenic agents and complete carcinogens. More important is the fact that recognition of promoting agents important in the human environment will allow a rational control of such agents. Specifically, it may not be necessary to completely eliminate promoting agents from the environment but to control their level and the period of exposure of humans to such agents. Table II lists promoting agents known to occur in the human environment. Not all of these agents have been associated with neoplasms in the human. In fact, phenobarbital and saccharin appear to exert little if any effect on the incidence of human liver and bladder tumors, as judged by published epidemiologic studies [65-67]. In contrast, the importance of dietary fat and calories, cigarette smoke, asbestos, and alcohol as promoting agents in the environment has been documented (Table II). In fact, one may conjecture that the production of clinical cancer in the human is largely a result of the action of continued exposure to promoting agents rather than exposure to minute amounts of complete and incomplete carcinogens in the environment. Recently Weber and Hecker [76] have reported that chemicals structurally related to TPA and having a similar promoting action are frequently ingested by people living in areas of Curacao where there is a high rate of esophageal cancer. In addition, Kopelovich Ct a! [77] demonstrated that TPA will promote the transformation of fibroblasts from patients with hereditary adenomatosis of the colon and rectum, an autosomal dominant trait in which all affected individuals eventually develop adenocarcinoma of the colon or rectum. PAGENO="0071" 67 Natural History of Carcinogenesis JSSCB:145 TABLE II. Promoting Agents in the Human Environment and the Neoplasms Associated With Prolonged Contact With Those Agents Agent Resultant neoplasm References Dietary fat (calories) Increased cancer incidence in general with excess caloric intake [69] Mammary adenocarcinoma [68] Cigarette smoke Bronchogenic carcinoma [70] Esophageal and bladder cancer [71] Asbestos Bronchogenic carcinoma and mesothelioma [72] Halogenated hydrocarbons Livera [60,73] (TCDD, PCBs) Saccharin Bladdera [74] Phenobarbital Livera [32] Prolactin Mammary adenocarcinomaa [64] Synthetic estrogens Liver adenomas [63] Alcoholic beverages Oral cancer [75] Liver and esophageal cancer apromotloit demonstrated in experimental animals but not yet in humans. Our knowledge of the action of promoting agents and complete carcinogens thus becomes extremely important in relation to human cancer and the human environment. The demonstration of promoting agents as distinct from complete -- carcinogens will be necessary in brder for rational and valid decisions to be made concerning the regulation of these agents in our environment. Furthermore, the onus and fear that go with the labeling of a compound as a "cancer-causing agent" can be removed or modified in many instances when we understand better what that compound contributes to the natural history of carcinogenesis in the human, as well as in lower animals. REFERENCES 1. Rous P, Kidd JG: J Exp Med 73:365, 1941. 2. Mottram JC: J Pathol Bacteriol 56:181, 1944. 3. Berenblum I, Shubik P: Br J Cancer 1:379, 1947. 4. Friedrich-Freksa 8-I, G÷ssner W, B÷rner P: Z Krebsforsch 72:226, 1969. 5. Farber E: Methods Cancer Res 7:345, 1973. 6. Bannasch P, Clining 0: Virchows Arch Abt A Pathol Anat 352:157, 1971. 7. Anthony PP: Cancer Res 36:2579, 1976. 8. Scherer E, Hoffmann M, Emmelot P, Friedrich-Freksa M: J Natl Cancer Inst 49:93, 1972. 9. Berger BW, Hon Y: Arch Dermatol~ll4:l698, 1978. 10. Johnson LD, Nickerson RJ, EasterdÓy CL, Stuart RS, Hertig AT: Cancer 22:901, 1968. 11. Evaiss AE, Gerson J, Schnaufer L: Nati Cancer Inst Monogr 44:49, 1976. 12. Foulds L: In Muhlbock 0, Emmelot P (eds): "Cellular Control Mechanisms and Cancer." Amsterdam: Elsevier, 1964, pp 242-295. 13. Boutwell RK: Progr Exp Tumor Res 4:207, 1964. 14. Boutwell RK: Crit Rev Toxicol 2:419, 1974. 15. Pitot HC, Sirica AE: Biochim Biophys Acta 605:191, 1980. 16. Pilot HC: "Fundamentals of Oncology." New York: Marcel Dekker, 2nd Edition, 1981. 17. Mintz B, Illmensee K: Proc NatI Acad Sd USA 72:3585, 1975. 18. King Ti, DiBerardino MA: Ann New York Acad Sci 126:115, 1965. 19. Friend C, Scher W, Holland iG, Sato T: Proc Natl Acad Sci USA 68:378, 1971. 20. Sachs L: Nature 274:535, 1978. 21. Markert CL: Cancer Res 28:1908, 1968. 22. Finckh ES: Med i Aust 1:438, 1974. PAGENO="0072" 68 146:JSSCB Pilot, Goldsworihy, and Moran 23. Gronow M: Chem.BioI Interact 29:1, 1980. 24. Pitot HC, Heidelberger C: Cancer Res 23:1694, 1963. 25. Pitot HC: Perspect Biol Med 8:50, 1964. 26. Goldenberg DM, Pavia RA, Tsao MC: Nature 250:649, 1974. 27. Foulds L: J Nat! Cancer Inst 17:70!, 1956. 28. Cairns J: Nature 255:197, 1975. 29. Smets LA: Biochim Biophys Acta 605:93, 1980. 30. Sachs L: Colloq Gesellschaft Physiol Chem 17:242, 1966. 31. lannaccone PM, Gardner RL, Harris H: J Cell Sci 29:249, 1978. 32. Pitot HC, Barsness L, Goldsworthy T, Kitagawa T: Nature 271:456, 1978. 33. Pitot HC, Goldsworthy T, Moran S, Sirica AE, Weeks J: In Hecker E, Fusenig NE, Marks F (eds): "Cocarcinogenesis and Biological Effects of Titmor Promoters." New York: Raven Press (in press). 34. Pitot HC: In Estabrook RW, Lindenlaub E (eds): "The Induction of Drug Metabolism." New York: FK Schattauer-Verlag, 1979, pp 471-483. 35. Salaman NIH, Roe FJC: Br J Cancer 7:472, 1953. 36. Kitagawa T, Pitot HC, Miller EC, Miller JA: Cancer Res 39:112, 1979. 37. Mirvish SS: Adv Cancer Res 11:1, 1968. 38. Kitagawa T, Harakawa T, Ishikawa T, Neomoto N, Takayama S: Toxicol Lett 6:167, 1980. 39. Tsuda H, Lee G, Farber E: Cancer Res 40:1157, 1980. 40. Borek C, Sachs L: Nature 210:276, 1966. 41. Kakunaga, T: Cancer Res 35:1637, 1975. 42. Mironescu S. Love R: Cancer Res 34:2562, 1974. 43. Peraino C, Fry RiM, Staffeldt E, Kisieleski WE: Cancer Res 33:2701, 1973. 44. Solt D, Farber E: Nature 263:701, 1976. 45. Ying TS, Sarma DSR, Farber E: Proc 11th Intem Congress Biochem 1979, p 664. 46. Roe FJC, Clack J: Br J Cancer 17:596, 1963. 47. Van Duuren BL, Smith AC, Melchionne SM: Cancer Res 38:865, 1978. 48. Druckrey H: In Truhaut R (ed): "Potential Carcinogenic Hazards from Drugs. Evaluation of Risks. UICC Monograph Series, Vol. 7." Berlin: Springer-Verlag, 1967, pp 6.0-77. 49. Brown CC: Oncology 33:62, 1976. 50. Schneiderman MA. Decoufle P, Brown CC: Ann NY Acad Sci 329:92, 1979. SI. Verma AK, Boutwell RK: Carcinogenesis 1:271, 1980. 52. Peraino C, Staffeldt EF, Haugen DA, Lombard LS, Stevens FJ, Fry RJM: Cancer Rca 40:3268, 1980. 53. Diamond L, O'Brien TG, Baird WM: Adv Cancer Res 32:1, 1980. 54. Berwald Y, Sachs L: Nature 200:1182, 1963. 55. Rezttikoff CA, I3rankow DW, Ileidelberger C: Cancer Res 33:3231, 1973. 56. Borek C: Radiat Res 79:209, 1979. 57. Yotti LP, Chang CC, Ttosko JE: Science 206:1089, 1979. 58. Driedger PE, Blumberg PM: Proc NatI Acad Sci USA 77:567, 1980. 59. Fiorowitz AD, Greenbauns E, Weinstein IB: Proc Nat! Acad Sci USA (in press). 60. Pitot HC, Goldsworthy T, Campbell HA, Poland A: Cancer Res 40:3616, 1980. 61. Poland A, Glover E: Mol Pharmacol 11:389, 1975. 62. Berenblum I: J NatI Cancer Inst 60:723, 1978. 63. Taper HS: Cancer 42:462, 1978. 64. Medina D: Cancer Res 36:2589, 1976. 65. Clernmenson J: Acta Pathol Microhiol Scand Suppl 261:38, 1977. 66. Hoover RN, Strasser PH: Lances 2:837-840, 1980. 67. Kessler II, Clark JP: JAMA 240:349, 1978. 68. Tannenhaum A, Silverstone H: Ads Cancer Res 1:451, 1953. 69. Wynder EL: Cancer 43:1955, 1979. 70. Doll R, Peto P: Br Med J 2:1525, 1976. .71. Howe GR, Burch JD, Miller AB, Cook GM, Esteve J, Morrison B, Gordon P. Chambers LW, Fodor G, Winsor GM: J NatI Cancer Inst 64:701, 1980. 72. Selikoff I, Hammond C, Churg J: JAMA 204:104, 1968. 73. Nishizumi M: Gann 70:835, 1979. 74. Cohen SM, Arai NI, Jacobs JB, Friedell GH: Cancer Rca 39:1207, 1979. 75. Tuyns AJ: Cancer Rca 39:2840, 1979. 76. Weber J, Hecker E: Experientia 34:679, 1978. 77. Kopelovich L, Bias NE, Helsott L: Nature 282:619, 1979. PAGENO="0073" 69 Mr. FL0RI0. Thank you very much. Dr. Nelson. STATEMENT OF NORTON NELSON, PH.D. Dr. NELSON. Thank you, Mr. Chairman, for the privilege of ap-. pearing before you, and to speak with you and the members of your committee. With your permission, sir, I would like to submit a written state- ment for the record, and to summarize what I believe to be the sa- lient points. I am, as you introduced me, Norton Nelson. I have been for many years closely associated with the cancer problem as it relates to the environment, going back at least that long as chairman of the Committee on Safety, Protocols for Safety Evaluation of FDA in the sixties where I supervised the preparation of one of the first outlines of testing procedures. I have worked with my colleagues on this panel on a number of related issues, including a major one in 1978, in a special panel advising the National Cancer Board. I have been closely related to the International Agency for Re- search in Cancer, where I chaired a number of their monograph committees. These monographs are a kind of the gold standard for international evaluation of the chemical carcinogenicity of chemi- cals. I chair the Scientific Advisory Committee to which you referred earlier under OTA, the Office of Technology Assessment of Con- gress, having to do with an evaluation of the sources, and evalua- tion of cancer from environmental sources. I was asked by president press of the National Academy of Sci- ences in the autumn of this year to organize a panel on the health effects of hazardous chemicals. We assembled a panel and reported to Dr. Keyworth: at least two of the six initiatives were closely re- lated to the issues of cancer. I am currently chairman of the Scientific Advisory Board of the National Toxicology Program; this consortium of Federal agencies is involved in the toxicological evaluation of chemicals. I can tell you that on Tuesday of this week, the day before yesterday, the Na- tional Toxicology Program Scientific Advisory Board authorized the beginning of a major review of the principles for cancer testing and cancer evaluation. Now, at this point I would like to make a distinction, a distinc- tion that I use, it may or may not be widely accepted. Mr. FLORIO. That is fine. Dr. NELSON. I would like to make a distinction for my discussion now between what I call prin˘iples for testing and evaluation of chemicals from that of cancer policy. The principles relating to the testing and evaluation of the carcinogenicity of chemicals really re- lates to the purely scientific and technical aspects of such an evalu- ation: How do you test? What animals do you test? Do you use short-term tests? What should be the protocols? How should they be used? This relates to the evaluation of the evidence, which is a scientific issue. What is the meaning of a malignant finding? What is the meaning of benign tumors for human cancer? The objective PAGENO="0074" 70 of all of this, of course, is the relevance of test results to human cancer. Now, there was launched on Tuesday of this week, the day before yesterday, a major review of the testing issues. This 18-month study was authorized, not because of any serious concern in the mind of any scientist familiar with this field as to the validity of the current approaches, but because all would agree that issues of this sort need periodic review to be sure that the best of available science is incorporated in the national testing programs. I may say that our national testing patterns have become widely adopted throughout the world and have international standing. Indeed, they can be regarded as the gold standard for the evalua- tion of carcinogens. So, now a new effort has just been launched, not because of any uncertainty about it, but to be sure that it is updated and the best of current science is being used. This will be an independent panel reporting to the Board of Sci- entific Counselors of the National Toxicology program. In distinction to what I call principles for cancer testing and evaluation, I will refer to cancer policy. By cancer policy, I refer to how evidence of carcinogenicity and exposure is applied to and re- lated to the control or regulation of carcinogens, and I wish to speak here particularly, not so much to the technical details which my colleagues here on my left are better able to discuss than I am. Rather, I wish to speak to several simple, I think, to me, rather ob- vious points of how such information is used in the interest of public health, and how they are used by control and regulatory agencies. I am forced to the conclusion-Well, let me first say that I think there has been widespread acceptance throughout the industrial- ized world of the facts so far mentioned as to the impact of the en- vironmental sources of cancer and general principles for their test- ing. There has been further, I think, widespread acceptance of the fact that positive findings well established in the laboratory must be used as the basis for control. To wait for bodies to fall, for people to die of cancer, is inhumane and ethically unacceptable. These principles, then, have been, I think, rather thoroughly ac- cepted in the United States, and our sister industrialized countries, and I have been closely associated with these on the international scene, our friends abroad have followed somewhat behind us, but have joined quite uniformly in the policies used in the United States. Now, what I have perceived let me frankly say, under the present administration and, I think, with the direct encouragement of the present administration, is a relaxation of regulation. I find this to be not necessarily open, forthright, but step by step, perhaps in a manner that could even be regarded sometimes as covert. The point I wish to make with you is that I believe until we have a better policy, better procedures, we should stay with the policies and procedures that are well expressed in a number of documents, had been widely followed earlier. In this connection I will also urge a rereview of cancer policy, not so much that I am concerned with the earlier policy, but more as a matter of updating and to provide reassurance that it is being reviewed. PAGENO="0075" 71 Let me mention a few points. I will be brief. You have a full schedule, and my points, I think, are simple and, I hope, easily sup- portable and evident. I will mention briefly two examples of relaxed regulation. For- maldehyde has been clearly demonstrated to be carcinogenic for animals. It fully meets the definition of sufficient evidence of car- cinogenicity under the IARC, International Agency for Cancer Re- search, definition of sufficient. It meets not only one of the three criteria. It meets all three of the criteria. Certainly evidence for human carcinogenicity of formaldehyde is not established. There are suggestive findings. But as I have said, to wait for such evidence is wholly unethical. Now, there was briefly mentioned a memorandum from Dr. Tod- hunter which became, I believe, a forceful factor in the nonregula- tion of formaldehyde. I discussed that in some detail in my appear- ance before the Gore committee May 20, 1982. I will not repeat that statement here. It is in the Congressional Record and is avail- able to anyone to read. Let me simply say that that memorandum, was a remarkable document in the sense that it reviewed in perfectly sensible scien- tific terms a series of issues relating to how formaldehyde may act biochemically in the body. In each one of these instances, the inter- pretation, the judgment based upon that memorandum was that formaldehyde is not very important, perhaps not even carcinogenic. It is the sort of a document that one would not expect an objec- tive scientist to produce nor to be produced by a responsible public agency. I will not push that further. I will simply say that formal- dehyde to me clearly meets the requirements for contrOl, not ban- ning, but for selective control. EPA has not taken that step. I find that irresponsible. I would like to give you another example. Let us not single out EPA as the only example of, shall we say, slowness in regulation. Ethylene oxide has been suspected for many years of being a dangerous chemical on a number of grounds. It is very widely used as a sterilizing agent in hospitals and drug houses and many other places for a very convenient and easy way of sterilizing equipment and containers. Over the last several years, the evidence has become convincing and explicit that in animals in two separate studies, it produced leukemia on inhalation and mesothelioma of pleura. There is in this case supporting human evidence. OSHA has not responded to these findings. There was brought suit in Washington District Court on ethylene oxide. They conclud- ed that the evidence was clear, solid, and convincing, and instruct- ed OSHA to issue emergency regulation. In my statement, I say that has now gone to appeal. There was a note yesterday morning in the Post and in the New York Times that the appeal report has been out. They have taken the position that it is an urgent problem, that OSHA should get on with it and start procedures within 30 days, but they have taken the position that the emergency route may not be required, but they expect a regulation, not precisely stated, but implied, within a year at the most. PAGENO="0076" 72 I mention these two, and there are other examples one can find to indicate that the perceived administration mandate of "reducing the burdens on industry," to loosely paraphrase the statement, is taken, I think, incorrectly and too literally by some of our control agencies. I would like to say at this point that I believe that industry is far more responsible than they are being given credit for. It is my pre- diction that in the case of both formaldehyde and ethylene oxide, that the responsible industries have already moved forcefully toward control of these two agents, and-and this is speculation- are indeed probably ahead of the regulatory agencies in charge of our national public health problems. Now, I would like to make just one comment in terms of some of the technological issues that have just been discussed, and that is the possible simplification of cancer evaluation through classifica- tion into the genotoxic, nongenotoxic categories. I wholly support the colleagues on my panel that this is an oversimplification at this time, that science is not capable of making with confidence that kind of a distinction, and I believe that cancer evaluation and as- sessment should continue for the present not to use such a classifi- cation. I also agree that the use of a threshold, even though it is conceiv- ably possible scientifically that there may be one, is not scientifi- cally supportable at this time. It is not supportable on the basis of prudent public health practice. I would like to come now to the issue of national cancer policy and what I believe should be done in this respect. OSTP has in course the production of a two-stage review of what I would call principles of cancer testing, and second, a cancer policy statement. I have seen the first. I don't believe the second part has been circu- lated. I find, that first part is technically flawed, but that is not my primary concern. Cancer control, as has been mentioned, is an extremely impor- tant public issue. It is one that has caused widespread concern, probably - in some cases needless concern. There is a distrust not only of policymakers but, unfortunately, of scientists as well. I think that in the public interest, it is absolutely imperative that national cancer policy be based on the most credible and technical- ly competent advice and auspices. I have no doubt that there are scientists in the federal establish- ment that can do quite as well in terms of the science as scientists outside of the Federal establishment, but I do not believe that the credibility required in this important issue can be achieved by cancer policy formulated within a purely administrative frame- work of OSTP and the federal agencies. Accordingly, I wrote to OSTP when that draft was submitted to me, not commenting on the quality of the document, but stating that I believed that the auspices under which it was being prepared were unsound and would not achieve the kind of public acceptance that was absolutely imperative, and I suggested as a first choice that this assignment be made to the National Academy of Sciences. There are alternative routes. One would be to do what was done earlier, to have a subcommittee appointed at the National Cancer PAGENO="0077" 73 Board. Another would be to go to the Scientific Advisory Council of the National Toxicology program. I believe that of these three options, all of which are essentially outside the Federal establishment, the National Academy of Sci- ences is the best. This, I think, is an important issue, and one that is imperative to launch and to reassure the public that things are under way under technically competent and the most credible pos- sible auspices. Now, until that is done, I see no reason for not staying with what are widely accepted patterns and policies that are in place. There is no serious doubt at this time about testing procedures. They will be reexamined. I do not foresee major changes. There is no serious question about patterns and policies under which these have been operated. The IRLG document, which has been, I believe, disavowed by the present administration, is a perfectly good statement of many of the policies for cancer control., What I am asking is, why should we not stay with the accepted principles and policies which were es- tablished until we have undertaken an appropriate review, which I think is very important? And that review should be done, I sincerely believe, under aus- pices outside the Federal structure. Mr. RITTER. Excuse me. What was your comment on the IRLG, that latter comment? Dr. NELSON. There was, under the former administration, an interagency regulatory liaison group. I am not expert in this. You will have before you Dr. Anderson later today. She was very active- ly involved in it. This was a very reassuring innovation organized within the regulatory agencies. It involves the regulatory agencies and some of the research input agencies. They met regularly; I do not know precisely how often. Douglas Costle was the innovator in this. It brought the heads of the regulatory agencies together. They could discuss their, problems. They could adopt uniform approaches. They could avoid needless delays, needless overlap. It was one of the reassuring innovations in the regulatory policy. It had heavily to do with cancer, but was not exclusively related to cancer. That has been abandoned. I do not know whether it has been of- ficially abandoned, but it is not functioning. There are others here who can tell you the precise status of that now inactive unit. The point is, again, we have in place a general understanding and some documentation to sUpport it. Let us stay with that and not quietly, covertly move away from the public health posture which we need and which is needed for public reassurance. Thank you. [Dr. Nelson's prepared statement follows:] PAGENO="0078" 74 TESTIMONY BEFORE THE SUBCOMMITTEE ON INVESTIGATIONS AND OVERSIGHT OF THE COMMITTEE ON ENERGY AND COMMERCE My name is Norton Nelson. I am Professor of Environmental Medicine in the Institute of Environmental Medicine, New York University Medical Center. I was Chairman of the Department and Director of the Institute of Environmental Medicine for some 25 years, a position which I relinquished some two years ago. I remain, however, full time in teaching and research in the Institute. I have been intimately related to the development of the knowledge base in environmental cancer for many years, and have actively participated in many advisory roles in this country and internationally in environmental cancer. I will not cite all of these. I might merely say that over the years I have been continuously involved in a series o~ advisory roles to the National Cancer Institute; these included The Clearinghouse on Environmental Cancer, as a consultant to the National Cancer Board Panel on Environmental Cancer, which produced in 1978 a document cal1e~ "General Criteria for Assessing the Evidence of Carcinogenicity to Chemical Substances," and many other advisory relationships to NCI. I have been continuously related to the program of the National Institute of Environmental Health Sciences which has had a strong interest in environmental cancer since its inception. I chaired the Food and Drug Administration Committee on Protocols for Safety Evaluation, one of the Chief activities of which was the production of a document on cancer testing. I served on the Science Advisory Board of EPA from its organization until October 1981 and was Chairman of its health committee. I have been intermittently (and am currently) a member of the Armed Forces Epidemiological Board, which is frequently concerned with the problems of occupational and environmental cancer. I served on a special committee to the Office of the Secretary of DHHS in which I chaired the committee which led to the establishment of the National Center for Toxicological Research. I am currently serving on a committee advisory to the Secretary of DHHS relating to Agent Orange. I served on a series of advisory groups to the Office of Science and Technology and the PAGENO="0079" 75 President's Science Advisory Committee over several administrations. Many of the issues concerned were environmental cancer. I have served for many years on advisory committees to the National Academy of Sciences/National Research Council and was first chairman of the Board on Toxicology and Environmental Health Hazards which is regularly concerned with problems of environmental cancer. Several years ago I chaired the Scientific Advisory Committee to the Congressional Office of Technology Assessment, which led to the report, "Assessment of Technologies for Determining Cancer Risks from the Environment." Recently I participated in an effort of the National Academy of Sciences for the President's Office of Science & Technology Policy (OSTP) relating to new national research initiatives as Chairman of a briefing Panel on Selected Research Opportunities on the HealthEffects of Hazardous Chemical Exposures. Several intiatives in our report dealt with environmental cancer. Internationally, I have worked for many years with WHO in a variety of committee and special advisory roles, including the International Agency for Research on Cancer (IARC), which produces the widely accepted monographs on the evaluation of cancer from environmental and other sources. I have chaired a number of the monograph committees which have produced those reports as well as committees reviewing other aspects of IARC's program. In my comments today I would like to state my concern about what I see as an alarming tendency towards change in the national policy for regulating carcinogens which, unless examined, may go beyond prudent public health policy and good science. My apprehensions may be incorrect; I hope they are. I understand that there is underway in OSTP a re-review, as there should be periodically, of national cancer policies relating to environmental chemicals. I wish them well and await their deliberations. I will further suggest below that in addition to this "in house" review, another more "public" review should be undertaken. Until these reviews are completed, debated, and accepted, I belive we should stay with present policy. Society is, for very persuasive reasons, concerned about chemical `~arcinogens (that is, chemicals that cause cancer) that may potentially expose either special groups (as in occupations) or, in general, consumers. PAGENO="0080" 76 Cancer is a disease that is often irreversible, often fatal; once initiated, it cannot, at this time, be halted with certainty; the only known totally reliable method of prevention is to avoid exposure. Most chemical carcinogens, although not all, act by producing genetic alterations in the cells of the body, the somatic cells, so that they can still reproduce as cells but, in doing so, will lead to aberrant and uncontrolled growth; this is what we call cancer. Our understanding of the biology of this process has improved immensely over the last ten years so that, at this time, we have a rather coherent pattern of understanding of these events; these are rather generally accepted by the scientific community. That is not to say that there is not debate over many aspects. Nevertheless, this general concept is very widely accepted as an approximate description of the means whereby normal cells can become malignant and lead to uncontrolled growth when exposed `tO chemical carcinogens. In recent years there has been developed an elaborate pattern of testing chemicals for carcinogenicity (often pretesting before use) through the use of laboratory studies which generally extend over the lifetime of the animals. In most cases these tests are carried out on rats or mice. In these and my following comments I will in some cases not reference certain statements of principle. If it is the desire of the Committee, I will be glad to try to supply the appropriate citations; but I believe that anyone in the field will recognize the widespread acceptance of the principles that I will comment on. First, a well-documented positive outcome in a lifetime test on mammals should be accepted as potentially capable of producing cancer in humans. In the IARC definition, this fits the category of sufficient evidence; their definition is as follows:* `Sufficient evidence of carcinogenicity, which indicates that there is an increased incidence of malignant tumours: (a) in multiple species or strains, or (b) in multiple experiments (preferably with different routes of administration or using different dose levels), or (c) to an unusual degree with regard to incidence, site or type of tumour, or age at onset." *Some of the material that I will present today was covered in testimony May 20, 1982 before the Subcommittee on Investigations and Oversight of the Committee on Science and Technology, chaired by Mr. Gore. I will not, therefore, go into detail as I did there on the issue of formaldehyde. PAGENO="0081" 77 Only one of these criteria needs to be met; yet the formaldehyde data meets this requirement in respect to (a), (b) and (c). IARC has evaluated the animal data on formaldehyde and found it "sufficient," as have many other agencies. The IARC warning when evidence for the carcinogenicity of a chemical is "sufficient" is as follows: "In the absence of adequate data in humans, it is reasonable for practical purposes to regard such chemicals as if they presented a carcinogenic risk to humans." There is as yet no accepted human evidence of the carcinogenicity of formaldehyde, although there are some suggestive but unpublished data. Nevertheless, the precedent has been widely accepted that human evidence is not required for instituting control or regulatory steps. Were it to be otherwise, one would wait until actual human cases of cancer occurred. This, of course, would be totally immoral and unethical. In fact, the whole basis for undertaking animal tests on carcinogenicity is to avoid or reduce human exposure. This principle has been widely accepted throughout the world by public health agencies. For any agency, person or industry to insist that positive cases of cancer in humans must be proven before regulation is undertaken is a truly irresponsible and insupportable position to take. Another example of non-rgulation but in which human evidence is avail- able, is the failure of OSHA to regulate ethylene oxide. This is a chemical widely used for the sterilization of equipment and containers of various kinds. Evidence of the carcinogenicity of ethylene oxide has been accumulating for a decade and has been abundant and convincing for several years. There are two independent inhalation studies showing decisively that this compound produces mononuclear leukemia and peritoneal mesothelioma. This meets IARC's definition of "sufficient" evidence of carcinogenicity. There is also support- ing epidemiological evidence of human carcinogenicity. Yet OSHA chose not to regulate. A suit brought in the Washington District Court found that the evidence clearly supported the presence of a clear and significant risk, and instructed OSHA to issue an emergency regulation. Again we have evidence that cancer regulation has been relaxed. 22-143 0-83-6 PAGENO="0082" 78 On another issue, the assumption of a threshold is widely regarded in the scientific community as being imprudent for public health protection. The basic issues are rather simple. The demonstration of a threshold in animal studies would be a feat of major proportions, extraordinarily diff i- cult, expensive, and of uncertain outcome. We must accept the fact that a demonstration of a practical threshold in humans or in animals is likely to be difficult and elusive and, at the present time, probably not attainable. There has been substantial growth in our understanding of cancer biology in recent years. In some cases the new knowledge may suggest major changes in cancer evaluation in others it may only confirm earlier approaches. In some instances it may suggest directions for additional research but may not be mature enough to change cancer evaluation or alter test protocols at this time. Examples of the latter are recent proposals aimed at an attempt to simplify carcinogen evaluation by classifying them into two major groups, genotoxic and non-genotoxic. The proponents of this approach took the position that this was justified by current science and would permit less stringent control of "non-genotoxic agents. The Cancer Assessment Group of EPA, without advocating such a move, took the responsible step of soliciting outside opinion ~n this proposal. There were major objections from a number of scientists to this two category classification primarily on the basis that present science could not support it. In my view science is not sufficiently advanced to permit such an oversimplication. Although one may successfully group genotoxic agents for some purposes, the grouping of all the others together under the label "non-genotoxic" is not rational. This includes a large diverse group of agents working through a number of different mechanisms of action; in fact this group may include genotoxic agents because of the inadequacy of current methods for their detection. It should be parenthetically stated that the Board of Scientific Counselors of the National Toxicology Program has just launched a major review of cancer testing and evaluation procedures. This will, no doubt, examine issues such as these more fully. PAGENO="0083" 79 Finally, I am forced to the conclusion that agents which would normally gualify for regulation are not been regulated. Thus, I do advocate review of the scientific principles which provide the basis for carcinogen evaluation using the best available scientific expertise and conducted under auspices that will maximize credibility. A covert and haphazard case-by-case revision of cancer policy such as that perhaps now underway in several of our national regulatory agencies will not be acceptable. As noted above, OSTP now has underway a two-stage examination of cancer principles using federal scientists exclusively. This effort will not meet the credibility standards required in an issue of such widespread interest. Principles proposed by the regulators themselves will be suspect since the view is inevitable that the outcome may have been influenced by a desire not to reveal former errors or by eagerness not to block regulatory policies already quietly established within the agency. Eligible mechanisms for this review would be the National Academy of Sciences, a special panel appointed by the National Cancer Board, or perhaps a similar panel appointed by the Board of Scientific Counselors of the National Toxicology Program. In 1978, as I noted earlier, I participated in such an examination and publication in a special panel responding to the National Cancer Board. The principles there expressed have been very widely accepted. However, the substantial scientific progress that has been made since then justify an orderly re-review. Once these have been reexamined, established, and published regulatory agencies should be required to observe them, or where failing to do so be required to make public a detailed statement of the reasons why these principles were not adhered to. Thank you. PAGENO="0084" 80 Mr. FL0RI0. Thank you very much. Let me express my appreciation to this very distinguished panel for their assistance in trying to make clear to the committee what is a very obviously complicated subject, and something that is so important to us. Let me just draw what I think are a few conclusions from the testimony of all three of our witnesses, and I would like your com- ments as to whether my analysis of the conclusions is correct or not. On the three documents that you have made reference to, Dr. Todhunter's memo, the OSTP paper, and the water carcinogenic paper evaluation of EPA, is it fair to say that it is your thought that the philosophy or the approach embodied in these three docu- ments represents a rather significant change in policy with regard to evaluating carcinogens? And if this is the case, it represents significant change in the policy that has previously been followed in evaluating carcinogenic exposure and the implementation of regulatory systems. Do you feel that it is scientifically supportable? Whomever feels comfortable responding. Dr. Weinstein? Dr. WEINSTEIN. I could start. Yes, I think that the three documents cited suggest changes in policy which, as Dr. Nelson has emphasized, are unjustified. I see no recent changes in our scientific understanding of carcinogenesis to justify these policy changes. I would like to cite again three or four major issues. (1) The issue of the validity of rodent bioassays in predicting human carcinogens. I find no evidence that suggests that rodent assays are nonspecific or invalid. (2) The question of threshold. I know of no recent data that would change our opinion regarding the threshold issue. (3) The distinction between initiators and promoters. This is an important area of investigation at the laboratory level and is excit- ing in terms of understanding the action of promoters and initia- tors, but the findings should not be converted to public policy until our understanding has moved ahead. I see no reason to regulate "the genotoxic and epigenetic" agents differently. A fourth issue which I did not mention previously is the attempt to specifically examine the chemical structure of a compound, the so-called SAR approach, and to bypass actual experimentation, animal testing and human epidemiology, to predict whether a sub- stance is carcinogenic. There are advances in our understanding of the chemistry of carcinogens, but we are a long way from being able to just look at the chemical structure of a new compound and predict with confidence whether or not it is carcinogenic. So I think that these four areas are not ready for policy change at the present time. Mr. FLORIO. Is it the case that if regulatory policies were modi- fled in accordance with the philosophy embodied with this new ap- proach, that would justify greater human exposure to potential car- cinogens? Dr. WEINSTEIN. Yes, I think that there may be even a covert at- tempt to utilize these four issues to soften regulatory guidelines PAGENO="0085" 81 and to increase the permissible levels. I see no scientific basis for doing that. Mr. FL0RI0. Something that is very vivid and .1 think perhaps meaningful even to the nonscientific Congress is the question of dioxin. That is something everyone is concerned about. As I under- stand it, the attempt to create a division within carcinogens such that there would be less regulation and less concern about one clas- sification of carcinogens that you referred to, I think, as indirect acting, nongenotoxic, tumor-promoting carcinogens, as opposed to a classification that is regarded as, I think you described it, as initi- ating contaminants, genotoxic. And in fact dioxins would fall into the first class of carcinogens, which under this philosophy would be less scrutinized or less regulated. Can I ask if my analysis is correct and if in fact, as I understand your testimony, that that is evidence of the fallacy of trying to make hard and fast divisions between these two classifications of carcinogens for purposes of regulatory application? Dr. Pitot? Dr. PITOT. Yes, Mr. Florio, I think that it is very well said, be- cause dioxin, as I indicated in my remarks, has this unique capac- ity of it is extremely potent, its effectiveness in promotion. And al- though one always wonders about the extrapolation from animals to humans, one is dealing with a level of potency that, even if one considers two or three orders of magnitude difference, still in the human one must take into account the fact that this material, thus far we have certainly no evidence that there is a no-effect level. And second, which I think is as important or perhaps more im- portant, is the way it stays in the body, that if something such as a promoting agent continues to act over a long period of time at levels at which we have every reason to believe it is effective, then in essence it becomes a very moot point as to whether there is a no- effect level in this situation or not. Now, I think that is a very good example of where, if one at- tempts by a blanket situation to change policy at this time-and I agree with Dr. Weinstein entirely that our present state of knowl- edge of the actions of promoting agents versus complete carcino- gens, genotoxic, nongenotoxic, and so forth, is really not at a stage where we can, let us say, take the risk to attempt to suddenly in a blanket way take a whole group of chemicals and treat them differ- ently, although we know that under a variety of circumstances they will cause cancer in certain species, from another entirely dif- ferent type of chemical based on what they, for example, do with DNA. I just do not think that at the present time we have the knowl- edge to do this. We may in 5 Or 10 years, and I would certainly echo Dr. Nelson's point. But I think this is an ongoing phenom- enon. One cannot stop and say, OK, what policy we make today will be good until the end of the century. Science is moving much too rapidly, and the more we learn, I think, the more-the better we will be able to make rational poli- cies. And I would certainly echo Ihat of my colleagues. Mr. FL0RI0. Just let me conclude my questioning with the obser- vation that what I appear to be hearing is that the 1979 IRLG doc- ument, which is the basis for the existing policy with regard to reg- ulatory monitoring and enforcement on cancer policy, is what you PAGENO="0086" 82 regard as the most currently acceptable approach to this problem, and until there is a scientifically supported alternative approach, that it is your thought that we should continue in that way while we go on monitoring and evaluating alternative approaches, and that it would be premature to implement an alternative approach to this problem, particularly as embodied in the three documents, until we have better scientific knowledge. Is that a fair statement of your conclusions? Dr. PIT0T. Yes. Mr. FL0RI0. Dr. Nelson. Dr. NELSON. Let me say that I am wholly in accord with that. Going back to your original question, I will further state that by action and by written word, there has been clear evidence of a re- laxation of public health concern in the regulatory agencies, that is, some of the regulatory agencies, not all-I want to be sure that that distinction is made, some of the regulatory agencies-for the control of carcinogenic chemicals. Mr. FL0RI0. And Dr. Pitot and Dr. Weinstein, you would agree with my analysis as to what it is? [Witnesses nod in the affirmative.] Mr. FLORIO. Mr. Ritter. Mr. RITTER. I thank the Chairman, and I apologize for being a little late. I missed the opportunity to introduce an opening state- ment, so I would ask unanimous consent that I could. Mr. FL0RI0. Without objection, all the members will be permitted to include a statement into the record. [The material referred to follows:] STATEMENT OF HON. DON RITrER I commend the chairman for holding a fact-finding hearing on the timely topic of the Environmental Cancer Policy of the Administration. I would like to extend a warm welcome to each of the witnesses. I look forward to a fair exchange of infor- mation on this topic which is of importance in our society. It is a subject of personal interest to each one of us. Today, I hope that we can take cancer out of the political battleground where would-be knights have been jousting. We need to explore this subject using as firm a scientific base as we can find to help the American people best understand and deal with the problem. What do we know about problems relating to cancer and how do we in government help to solve them? With the testimony of such esteemed wit- nesses, I hope that we can keep today's discussion on a scientific and a medical rather than on an emotional political level. Cancer research is such a rapidly-expanding field that our learning curve in this area has been almost exponential in the last few years. Up until now, we've been looking at the individual pieces of the puzzle: separate studies on carcinogens. These days, and today in particular, we can begin to look at the way these pieces fit to- gether so we can see the larger picture. A picture which not only the expert, but the man on the street, can see. One in which he can understand the risks of contracting cancer based on what we know from scientific studies and seeking consensus from the scientific and medical community. We need to present a sound basis for evalua- tion of what is known about carcinogens and what may be found out, so that if a government policy is to evolve, it is a policy based on science. Our knowledge of carcinogens of their control has increased by leaps and bounds since the IRLG reported their findings a few years ago. The Office of Science and Technology has recently pieced together a report on the current "state-of-the-art" for identifying and characterizing potential human carcinogens. This could form a basis for developing a cancer policy. Today's hearings should further that process. With the constant concern about potential carcinogens expressed each day in the press, it is timely to review our progress in this area. I hope that through the testi- monies given today, we can make better use of science to look at the history and recent developments on the control of carcinogens in the environment. That we can PAGENO="0087" 83 take this information and examine the regulatory and legislative proposals relating to chronic health hazards. I am hopeful that in addition to our other witnesses, the Administration can share with us its short and long-term objectives, the guidelines used, the scientific bases employed, the successes, failures, and projections for the future. Mr. RITTER. Thank you. I would simply like to open my comments with some general ob- servations, which I would then like to question the witnesses on. First of all, I would like to talk a little about the studies of Sir Richard Doll and Dr. Peto. Do I have the pronunciation correct? Dr. PIT0T. Yes. Not me. That is Richard Peto. It is spelled differ- ently. Dr. NELSON. That was a commissioned study done for the Con- gress of the United States under an OTA contract. Mr. RITTER. The Doll-Peto study was a commissioned study? Dr. NELSON. A contractual study done for the Congress of the United States by the Office of Technology Assessment. In the study that we mentioned earlier conducted a little over 1 year ago, it ap- pears as an appendix. Mr. RITTER. I would like to know if that is the same study that the chairman referred to, an OTA study? I believe they were both in June 1981. Is that the same study? One of the features of the Doll-Peto study seems to be that the predominant risk of cancer in this country is lifestyle, and in lump- ing environment and lifestyle together I think we are misleading the American people. The Ameri˘an people look upon environmen- tal pollution as air and water pollution, and to some extent the in- dustrial workplace. I think we should try to separate out, at least to the best of our ability, these various contributors to the 400,000 estimated cancer deaths per year~ In the lifestyle piece of this cancer-causing pie chart, so to speak, tobacco comes across as a major cause; diet, not including food ad- ditives, comes across as a major cause; nutrition comes across as a major cause. You can lump alcohol in there. The Doll-Peto study talked about cancer caused by environmen- tal pollution-that is, basically air and water pollution-estimated at about 2 percent, with a range of 5 percent Or less; occupational, 8 to 2 percent range, about 4 percent. I just want to point that out because the data is showing all of us and the American people that the age-adjusted incidences of cancer have remained relatively con- stant if you factor out the smoking effects on cancer incidence, and in those age-adjusted cancers, the incidence seems to be independ- ent of massive introduction of new chemicals on the American in- dustrial scene. Now, I say this simply to put today's hearings in perspective. Im- portant though they are in getting to the bottom of what public policy can do to reduce the risk of cancer, it is important that the American people understand that: one, contrary to the political po- sitions of some, there is no cancer epidemic in this country; two, while there has been a major introduction of new chemicals to the American marketplace, this does not have corresponding incidences of age-adjusted cancer rates. And indeed, one can look at some very major achievements and major accomplishments in regulating chemicals, for example in the PAGENO="0088" 84 workplace, and you do not have some of these rather bizarre kinds of exposures you had several decades ago. I think it is important to make this kind of comment. Now, having said that, I would like to ask a few questions of the witnesses. I would like to talk about the dioxin situation for a moment. Are the witnesses familiar with the 1949 Monsanto acci- dent, where there was very substantial exposure to dioxin, and the followup studies that have been done on the exposed population? Dr. PIT0T. I am not, but I am familiar with the Sevesio incident in Italy. Mr. RITTER. The Sevesio incident is different in that it is a recent incident. And while, incidentally, the Sevesio incident of exposure to dioxin has caused symptoms of chloracne, of course, it is too recent to really evaluate for cancer. But on the West Virginia Monsanto accident there have been some extensive followup studies done on these individuals. The ac- cident was in 1949, as I understand it, and to date there has been no perceived difference in the incidence of cancer among those ex- posed in the West Virginia accident. I thought that some 150 work- ers were exposed and just for the record one should point that out. Dr. Nelson, you mentioned that the current governmental at- tempt to try to refine or at least expose some possibilities of coming to grips with a policy on cancer, that somehow it was not under the correct auspices; was not proper. I should like to point out that the IRLG, which you seem to hold in very high esteem, was also a gov- ernmental body, an interagency body. It did not bring in peer review. It had a rotating membership. As a matter of fact, the very roughest first document from the OSTP review went out to peer review to people like yourself and mem- bers of the environmentalist community. You know, people have said that the IRLG review occurred behind closed doors and was not open to extensive peer review. But I would like to know what your comments are. Dr. NELSON. Well, I am somewhat familiar with the development of that document. In fact, much of it was done by a colleague of mine, Dr. Albert. It had extensive participation from outside scien- tists. I think I would agree with you that it should have had out- side peer review. It did not. I have made such a recommendation to the preceding Adminis- trator, and I have made such a statement in the Science Advisory Board under the present EPA. The auspices did not include as much outside participation, but it included a substantial amount, but it was extensively reviewed by many scientists outside. I would simply say that it is a fairly good document. It could have been handled better in terms of outside and formalized peer review. I have told you that I do not say that the scientists within the Federal agency are incapable-I say they are as capable as those outside. I think that the question is one of assuring the public that this has the best stature -- Mr. RITTER. But I think the process, Dr. Nelson, of peer review is going to be the one element of evaluation which can evolve public trust. But it is absolutely essential that that occur, and I do not think you can point to IRLG as the paragon of peer review, because the history certainly shows-- PAGENO="0089" 85 Dr. NELSON. Sir, I have just said I do not regard it from the standpoint of peer review as a paragon. Mr. RITTER. One could point to that, I agree with you. Dr. NELSON. You may and you can, and I think that is fair. I am simply saying that there is at least something in place. Mr. RITTER. I have one question I would like to ask about the constitution of the panels. This is a comment and a question. In the constitution of the panels, Dr. Neal has been placed in the panel with industry and environmental representatives. The CuT, as we know, is funded by industry. But I would like to solicit your opinions as to whether Dr. Neal qualifies as, or CuT and Dr. Neal qualify as, an independent re- search organization or not, or whether Dr. Neal qualifies as a rep- resentative from industry. Dr. NELSON. I am in no position to say how independent the orga- nization is. I know Dr. Neal very well. He is a highly skilled scien- tist. Mr. RITTER. I think Dr. Weinstein and Dr. Pitot may have some greater familiarity. Dr. Weinstein, are you prepared to comment on that? Dr. WEINSTEIN. Both of us, sir, Dr. Pitot and I, serve on a scien- tific advisory panel of CuT. Speaking for myself, and I guess Dr. Pitot shares this, it is a superb organization. Dr. Neal is a superb director. Mr. RITTER. Is it an independent research organization or does it represent industry's views? Dr. WEINSTEIN. I think it pursues research in an objective and independent fashion. It is clearly supported by industry. Dr. PITOT. I would share Dr. Weinstein's views. I think a very good example of this is one of the compounds you are interested in today, and that is formaldehyde. It was CuT that first demonstrat-. ed the carcinogenicity and reported it immediately, even before the data was absolutely certain, just precautionary. Then their data very clearly showed the carcinogenicity of formaldehyde. It was, I think, reproduced by Dr. Nelson's group in New York and now I believe has been reproduced even further than that. Mr. RITTER. So I am curious as to whether Dr. Neal should have been sitting on your panel or on the panel with industry and envi- ronmental representatives. Dr. WEINSTEIN. Would it be appropriate to respond to a couple of points that you have made, Congressman Ritter? I would agree with you that there is no evidence of a massive ep- idemic of cancer related to exposure to industrial chemicals. On the other hand, I must emphasize that the Doll and Peto estimate is probably one of the lowest estimates. Experts differ on what the figure should be. But even if we take their estimates, which have a minimum of 5 percent and may go as high as 15 percent, and using the figure of approximately 500,000 cancer deaths per year than we would attribute, 50,000 to 150,000 deaths related to the types of. chemicals we are concerned with now. I do not think these are hard numbers, but they are large num- bers-50,000 is a large number. If we could prevent or antici- pate----- PAGENO="0090" 86 Mr. RITTER. Five percent of 500,000 is not 50,000. It is 25,000. Just to get the arithmetic on the record. Dr. WEINSTEIN. I think that if I could prevent 25,000 cancers in my lifetime I would be a hero, I would relax, and I would have done a good job. Mr. RIrrER. I agree completely. Dr. WEINSTEIN. So prevention moves on many fronts. Now, cer- tainly cigarette smoking is a major problem, and in another forum I think the three of us would express concern about the lack of Government efforts in that area. Diet is an important area. Many experts, however, feel that much more research is required before we give rigid guidelines to the public. In addition, we are talking about cancers occurring today as a result of exposures during the past several decades, because of the long latent period that I emphasized. What we cannot anticipate is what the effects of introducing chemicals today will be on the can- cers which may occur 40 years from now. So that is your estimate you are looking retrospectively, rather than prospectively. Mr. RITTER. I think it is important to recognize that this organic chemical industry is not a new endeavor. It is something that has grown with us by leaps and bounds since the end of World War II, and vast numbers of new chemicals have been introduced all along the way. I am not trying to diminish what the chairman and this committee is trying to do in seeking an intelligent science-based health effect policy on cancer. But I think, and many of you would agree, that fear of cancer has become perhaps in some cases irra- tional. I think we as legislators need to set a balanced tone to develop this science base. While we are focusing in a narrow area, we should nevertheless not lose touch with the larger realm, and that is really the gist of my comment. Dr. Pitot. Dr. PITOT. If I might, I would like to go back to the Monsanto incident. I was not aware of that, but I would like to again come back to my testimony. The data that Dr. Poland and I have put to- gether does suggest very strongly that dioxin is an extremely pow- erful promoting agent. We know that promoting agents in order to exert their effects cannot be given only once; they must be given over a long period of time. Mr. RITTER. Well, these people worked with the substance, as well as received significant exposure in one batch, so to speak. Dr. PITOT. Well, let me give you an example of the other side, where people have worked with dioxin-contaminated compounds for a long period of time now are coming up with certain numbers of cancers, and that is the Scandinavian experience with the phen- oxi acetic acid derivatives. Mr. RITTER. Are you familiar with the Finnish experience as well? The Finnish experience is similar to the Norwegian experi- ence. . Dr. PITOT. Except there are a whole variety of different circum- stances in that as well. And the point that you brought out earlier, I think, with the Doll-Peto report to me at least-and what I am saying here now is certainly an opinion and not a fact, and that is PAGENO="0091" 87 that promoting agents in our environment are extremely impor- tant. Probably diet-I mentioned that cigarette smoke, granted, is a complete carcinogen, that clearly epidemiologic evidence suggests very strongly its tumor promotion, which may in fact be a very strong component of the carcinogenic action in the human. Alco- holic beverages may well be very potent promoting agents, and cer- tainly there is now scientific evidence that asbestos is a promoting agent. So in fact in our environment, as the Doll-Peto report points out, it may turn out that things that we know scientifically and experi- mentally are promoting agents are going to be extremely impor- tant things in our lifestyle, in our environment and our lifestyle. Furthermore, again this thing of the promoting agent being present for a long period of time as an effective component, rather than a single-shot type of circumstance, is a very important one. Dr. WEINSTEIN. I would like to also comment on the limitations of the Monsanto study, which I think is an example of the general limitations or insensitivity of epidemiologic studies. If 150 workers were exposed, then that is a very small population at risk for gath- ering statistically meaningful data. First of all, I am not certain how well documented the extent of exposure was. So were all 150 really exposed to a known amount at the tissue level, and so forth? Let us grant that it is 150. If those 150 live out their lifespan, that would produce, given about one in five deaths from cancer, about 30 cases of cancer, occurring as a background. Since they have not lived out their lifespan, I would estimate that at most we have an opportunity to score for perhaps 20 cancers. If there was a 10 percent increase as a result of this exposure, we would see 22 rather than 20 cancers, which is not statistically sig- nificant. So the data do not rule out the possibility of a 10 percent increase in cancer in the exposed group. In addition, since, as Dr. Pitot has emphasized, dioxin is a pro- moter perhaps only a subset of that population who were exposed to an initiater and so the response might even be lower. But on the other hand, given a population of 200 million people, increasing cancer by 1 percent would be an epidemic in terms of the number of new people getting cancer. So I think we must weigh this apparent negative, and I do not think it is clearly negative-because of the lack of ser~sitivity of the method-against the overwhelming evidence that Dr. Pitot and others have cited that dioxin is one of the most potent toxic sub- stances tested in rodents and subhuman primates. So I am still not comfortable with having it at high levels in-- Mr. RIrrER. You are saying the subhuman primates show similar effects to the rodents? Dr. WEINSTEIN. Yes, in terms of their biochemical effects. It is also a teratogen. Mr. FLORIO. Dr. Nelson. Dr. NELSON. At the risk of some replication, let me simply say that Mr. Ritter's summary of the general disposition of cancer pat- terns I wholly agree with, and I would hope that I would be as active as the next one in attempting to control all the sources. In PAGENO="0092" 88 the case of smoking, we know it, we have been very inactive. In the case of diet, we do not know yet what to do. Now, although I was involved in the selection of the contractor for the Doll-Peto report, I would say that is a very conservative es- timate. That is what Sir Richard's position normally is, and I think is so viewed by many of the experts in the field. It represents a bottom, essentially, rather than a median or upper level of risk. But I think more important than this, the question of the size of the impact, is that preventable cancers should be prevented, and if this can be done without any significant economic disability I see no excuse for not doing so. Now, we must remember some of the warnings we have had. Reasonable estimates of deaths from asbestos exposure converged at about 8,000 to 10,000 per year throughout the last decade of the century. That is not a trivial issue, and we do not know how many sleepers there are out there. I think basically we should obviously prevent-- Mr. RITTER. I am not arguing against doing all we can to prevent the incidences of these cancers. I just get alarmed sometimes when I read about the issue in the media or watch it on television. It is enough to scare the living daylights out of the average citizen. I am sure that is not your intent. Your intent is to go out and solve the problem. But you are going to be diminished in your ability to solve the problem, because the response will not nearly be as medi- cal and as scientific as one would hope it would be. - Thank you very much. Mr. FL0RI0. The Chair has been very liberal with regard to the 5- minute rule that we adhere to in questioning because of the com- plexity of this subject and because this panel is particularly quali- fied to provide us with information. We will continue to do so for the balance of this panel. At this point I would recognize the gentleman from New Mexico. Mr. RICHARDSON. Thank you, Mr. Chairman. I would like to read a brief statement into the record, because I would like to try to get the discussion more to the policy issue. I will be the first one to state without reservation that scientifically I feel very deficient in this area, but possibly on the public policy I can offer some. insights. I think this hearing is very timely, because during the last couple of months there have been disclosures which have called into question this administration's commitment to protecting our citizenry from the effects of hazardous substances. I think it there- fore only prudent for the subcommittee to examine the administra- tion's policy on the very disease that hazardous substances have been shown to cause. In that regard, I believe that our cancer policy should be de- signed to prevent cancer from occurring rather than limiting the damage after this dreadful disease has appeared. Of course, we cannot live in a post-industrial society and have no risk of cancer, but that should be our goal, and the historic EPA risk tolerance of one in one million incide˝ces of cancer due to exposure to hazard- ous substances should remain our policy, and any attempts to reduce that risk to one in 100,000 or one in 10,000 must not be tol- erated. PAGENO="0093" 89 Any attempt to make that reduction in the name of cost benefit analysis or regulatory relief for industry must be labeled for what it is, economic Darwinism at its worst. How can anyone quantify the cost of even one American developing lung cancer? Mr. Chairman, I am troubled by what I have read recently in Science Magazine about the Reagan administration's cancer policy. The article says that this administration is in effect willing to in- crease the risk of cancer to the American public. The journal quotes John Todhunter, a leading EPA official, as saying. that this administration wants more flexibility in deciding when to act and when to tolerate hazards. I would be willing to give Mr. Todhunter the flexibility he de- sires if I believed the agency would err on the side of public health and safety, the sound public policy that we all desire: but to date this administration has erred on the side of the industry it has been charged with regulating. From the retraction of Eula Bingham's statement on a worker's right to be protected from cancerous materials to carcinogenic guidelines developed by the Albert group last June, in this admin- istration the winner has always been the manufacturers, and I would be very interested to hear from the administration's wit- nesses on this subject. Finally, gentlemen, the questiOn I would like to ask you, is, who makes cancer policy within this administration? And as a corollary to that question, is there a bureaucratic problem within the admin- istration in terms of focusing the proper policy on this issue? Dr. Nelson? Dr. NELSON. I am no expert in the sources and origins of Federal policies. I have served in an advisory role to every regulatory agency, I believe, and most of th~ research agencies, and I am often mystified as to how these come about. I would not pretend to answer that question. I do say, however, that the IRLG had the great merit of bringing together the regulators in a dialog and they kept a record of what they were doing. They had stated principles, and I think that is a very wholesome approach. Now, in my written submission, I have suggested that I would like to see a cancer policy developed under the National Academy of Sciences auspices. I would further like to see such a policy to be accepted or openly rejected, as the case may be, by the relevant Federal agencies, and if they choose to ignore those principles, they should be forced to explain the grounds on which they do ignore them. So, I think this problem of public acceptance and public trust is one of the most important and djfficult issues of the time. I there- fore plead for forthrightness, openness, and the most believable auspices for establishing this, and openness on the part of the Fed- eral agencies in deciding to accept or reject such guidance. Mr. RICHARDSON. I think the only point I am trying to make is that just as scientifically I felt deficient, those of you that we look to for our scientific advice must also recognize that part of the problem in terms of developing a scientific policy is the bureaucrat- ic structure that we probably now have. Now, what I am trying to get at is how, from a public policy standpoint, are you going to recommend to us that we define our PAGENO="0094" 90 cancer policy? I sense from Dr. Weinstein he has some reservations about this policy, and the orientation seems to be negative, and I would like to see from you some concrete recommendation. Should science policy be directed by the science adviser to the White House? Should it be a White House-related issue? Should it be done at EPA? I would like to get how can we magnify the ad- ministration's commitment to cancer policy in focusing at the bu- reaucratic level? What ideas do you have on that? Dr. WEINSTEIN. Well, I am not an expert in this area, but it would seem to me the basic principle is that you must stay in touch with the advances in science, and you must consult with the scientific community itself if you want to know where it is at, so that reports, for example, by the National- Academy of Sciences or by the National Cancer Institute have an authenticity because they are the best informed opinion of scientists. It is possible that some of these recent guidelines are faulty be- cause they do not emanate from the scientific community and have not taken advantage of consultation with the best of science that is ongoing. Dr. Nelson has mentioned previous reports of the Nation- al Academy of Sciences, and these are in general authoritative and reflect the current state of the art. The International Agency for Research on Cancer calls upon scientists at a worldwide level, be- cause this is a health problem of worldwide dimensions. And its advice is not being listened to. So, I think part of the problem is that the administration is not seeking the best advice. Mr. RICHARDSON. Mr. Chairman, I have one last question. In controlling carcinogens, is it not true that we must also ac- count for the synergistic effect of chemicals such as occurs with smoking and absestos in evaluating carcinogens for potential con- trol? Dr. PITOT. Perhaps, Mr. Richardson, I might speak to that, be- cause in a sense that is what both Dr. Weinstein and I spoke to this morning when we talked about promoting agents. The syner- gism of asbestos and the cigarette smoke, of alcoholic beverages and smoking, even of diet and certain aspects of what we may take in to cause initiating components, and I think that is what we were discussing here today. And I think I echo my colleagues in saying that although we know such a thing occurs, we are not in a position as yet to be able to distinguish out of different classes which have different levels, let us say, of risk or some other component. Mr. FL0RI0. Thank you. Mr. Dowdy. Mr. DOWDY. Thank you, Mr. Chairman. I have one question. I was informed that within the past couple or 3 weeks there was a segment on CBS News where a number of people related concerns that they have that the USDA, through either lax standards or for whatever reason, is allowing the sale of some meat products that might contain carcinogens or might con- tain a high level of cancer-causing agents. Do any of you have any opinion about this? Would this be a valid concern? PAGENO="0095" 91 Dr. PIT0T. I am not familiar with that particlar thing, Mr. Dowdy, so I cannot really respond to it. Mr. DOWDY. That is all I have. Mr. FLORIO. Mr. Eckart. Mr. ECKART. I have two lines of questioning I would like to pursue, and perhaps this panel can help me. I suppose the broad question that I am interested in~ is a spinoff of Mr. Richardson's question. In some of these tests, the conclusions are fairly clear, and in others they appear to be somewhat marginal. How do we, in trying to formulate public policy, properly assign the benefit of the doubt? In what direction, based on your best scientific medical esti- mates should it be skewed? Dr. NELSON. Basically, I think this is a problem for the repre- sentatives of the people to decide, namely, this Congress. The bal- ancing of risk against the benefit is not a purely scientific issue. The estimate of risk can be a purely technical issue. What is bene- ficial in a particlar segment of society may be harmful in another, and that tradeoff is something that is a sociopolitical problem. I throw that issue right back, sirs, into your laps. I think that you are delegated by the people of the United States to make these decisions, and you will make one set of decisions in wartime, an- other set of decisions in peacetime, another set of decisions depend- ing upon public climate and economic conditions of the country. So, I am avoiding the issue, sir, and telling you that I think you must make these decisions, and yOu must decide how much latitude you give to the regulatory agencies and under what kinds of sur- veillance they will be kept. Mr. ECKART. Spoken like a seasoned Member of Congress. Is there some point where we should overstate giving someone the benefit of the doubt, and we should understate it? Is there at least some generally recongnized minimal standard? I understand that in marginal cases we are always going to have trouble, and that clearly exterior factors will influence that. Dr. NELSON. At one extreme is,~ let us say, the Delaney clause. I see nothing very much wrong with that. If we have one way to make ice cream pink, do we need another one? In short, the bene- fits are trivial. The risks may be very minor. But in a case like that, I am very happy to see sort Of a blanket imposition of restric- tion. Let us take, however, a pesticide that is important for the food supply of the Nation. This is going to require the wisdom of a Solo- mon to make that kind of decision. Are there substitutes? Are there not substitutes? Can it be used in restricted ways? Can suffi- cient controls be applied? So, it is a mix of public health concern, of economy, and of tech- nical fixes. I know of no set rule, and I am sure that we will all agree that if the survival of the i~ation is at stake, we will take more risks than we will in a peacetime, healthy economy. I am sorry I cannot be more helpful, sir. Dr. WEINSTEIN. I would agree with Dr. Nelson. I think, as scien- tists, at least speaking personally,~ what disturbs me is that during this process, one is not allowed to distort the scientific facts, to bend to the economic or political needs. The scientific facts are that animal tests are valid and predictive, that there is no evidence of a PAGENO="0096" 92 threshold, that at the .present time we cannot distinguish promot- ers from initiaters in risk extrapolation. These are the facts. They, I think, are incontrovertible. Now, there is the additional problem that you raise of benefits versus risk. What is economically and politically feasible? That is a differ- ent debate, but one is not allowed to distort the scientific facts during the course of that debate. Mr. ECKART. Do you suspect, if I may pursue this last point, that there has been some distortion within this administration of scien- tific data in the context of issuing rules or regulations to restrain the use of certain substances in our society? Dr. NELSON. My answer is, from my limited post of observation, yes. In fact, as I said before-I am not sure you were here-I think that industry is in many cases very probably ahead of the regula- tory restraints put upon them in their own proper concern for the health of their employees. Mr. ECKART. The second point I would like to pursue is one that gets raised repeatedly. It was in the testimony of Dr. Weinstein and I would like to expand upon it just a little bit for the record. Whenever some new chemical or substance is discovered to be cancer-inducing, we hear that a rat would have had to drink 10,000 cans of saccharin-sweetened pop to obtain cancer if the results of the study had been extrapolated over to humans. The dose levels used by some to try to cast doubts on the conclusions of the study. Could you evaluate the validity of rodent bioassays? Can we ex- trapolate that information into relevant conclusions about humans? Should we in the Congress rely on these tests as a valid research tool in regulation? Dr. PIT0T. Well, perhaps I could say something to that. I think it is important to remember that in those tests, really one question is trying to be answered, and that is, is a particular material carcino- genic, that is, does it cause cancer to any species under any condi- tions? And if that is the only question that one is trying to answer, then I think the studies with saccharin and other things are clear- ly valid. And as Dr. Weinstein pointed out earlier, the real question is, what is our state of knowledge at this particular point with respect to how specific chemicals cause cancer? And I think we are still-many of us realize the difficulties in interpreting those types of results beyond just answering the one question that was posed, and I think that points out that our sci- ence at the moment is by no means perfect. It is in a stage of devel- opment, and clearly, with the points that were raised here, the question of promoting agents and their importance in our environ- ment, we have every reason to believe that this is going to become a very important factor. And that is why I mentioned to the chairman that as Dr. Nelson said, the review of this sort of thing is going to have to be done probably on a yearly basis in order to make it efficient and to use it to the best of all possible things to our society. So, I think that as long as we keep within our minds what is ac- tually being answered by these tests and what is our present state of knowledge, then there is a rationale, but granted, our present state of knowledge is still certainly deficient. PAGENO="0097" 93 Mr. ECKART. Thus, while it is imperfect and subject to continuous refinement it still is an appropriate tool to use in ascertaining what should or should not be~ allowed in the marketplace. Is that a fairly common conclusion? Dr. PIT0T. I would think so, because if you do not, what is the alternative? Mr. ECKART. I thank the chairman. Dr. WEINSTEIN. If I might-- Mr. ECKART. Please. Dr. WEINSTEIN [continuing]. Just to briefly review this thought, you are probably familiar with this. argument. Again, it has to do with the sensitivity of tests. When one does these rodent assays, there are usually, at most, 50 to 100 animals per group. Even so, once the test is fully designed, it may cost close to $1 million to test a single compound at one or two doses, and it will take 2 or 3 years or perhaps even longer to accumulate the data. Therefore, it is difficult to use larger numbers of animals or do precise dose-response curves. That is unfortunate, but that is the state of the art. One therefore increases the dose to the animals to increase the sensitivity. In the case of saccharin, the dose used would be equivalent to about 800 diet drinks per day per individu- al, which of course is ridiculous in one sense. It is at least 100 times greater than humans might be exposed to. On the other hand, it produced an incidence of tumors of about 20 percent, so we get into the question of how to extrapolate from high dose to low dose, a very difficult area. If we would do a simple linear extrapolation, we would say that humans consuming 1/100th of that amount would get-0.2 percent of the exposed group would get cancer over a lifetime. Given a popuation of 200 million, that would still be a large number of cancers. There are many assumptions here that we are uncomfortable with, but we do not know how to do it otherwise. What is incontrovertible is that those animals did get bladder cancer. There is no doubt about it. And it was clearly related to the saccharin exposure. Now we enter into the public arena, and public policy, and in that case, the decision was to put on labeled diet drink: "Use of this product may be hazardous to your health. This product con- tains saccharin, which has been determined to cause cancer in lab- oratory animals." And in that case it leaves the decision up to the public, but at least an informed public. It would have been inappro- priate to say that the animal tests were invalid and have no rel- evance to human hazards. Mr. FL0RI0. Ms. Mikulski. Ms. MIKULSKI. Thank you, Mr. Chairman. The purpose, I believe, of us reviewing the cancer policy of the administration is for the same reason we have created all these agencies, which is to in some way to prevent people from getting cancer. I have two concerns. One is the implementation of cancer prevention policy, and second, the administrative mechanisms for doing that. Let me share with you my concern, and then get your .reaction if you think there need to be administrative changes. A person eats 22-143 0-83-7 PAGENO="0098" 94 food, has a job, lives in a neighborhood, and all of that impacts upon him or her. Yet when we look at the investigation in carcinogenic research, we find that food is being investigated, or may be investigated by FDA, that EPA investigates those factories or sources of pollution that might cause adverse health effects. OSHA is supposed to take a look at what is happening to workers. The national defense exempts everybody, just about, who works for research for them and in some of their production of materials related to nerve gas and chemicals and other kinds of things. Now, the point that I am making is that as a person who lives in a neighborhood, buys food, all of this is impacting upon them, and they are their own test tube. One of the concerns that I have is that because every agency is looking at it from their own stand- point, no one is really looking at the synergistic impact that all of these have on a person. Therefore, do you think, one, our research is fragmented, and therefore unproductive, or do you think that just the way we have a Center for Disease Control, that there ought to be some type of one-stop shop that really does carcinogenic or cancer-producing re- search that would impact across all lines, whether it is workplace, community, diet, or all of the things that you have outlined? Dr. NELSON. The implication that multiple exposures such as you describe are not of a concern within the research or the Federal community is not quite correct. It is correct to some degree. Let me first make a response to a limited part of this. That is that when it comes to the testing of chemicals for what- ever use they are put under the national bioassay program, it is under the national toxicology program, and I am chairman of the board of scientific counselors to that unit, the choice of chemicals for tests is a very detailed process which goes through a series of steps with the opportunity for public input. The final decision is made by the executive committee of the na- tional toxicology program, which includes the heads of the four major regulatory agencies, and the heads of research agencies. So, at least there is this forum for bringing together these issues. Ms. MIKuL5KI. Does that happen? Dr. NELSON. In some degree. It does not happen as well as it could. I cannot answer that. Ms. MIKuLSKI. My question is, does that happen, and is there a mandate to make it happen? Dr. NELSON. As to a mandate one would have to go back to the legislative charter or the administrative charter. Ms. MIKuLSKI. But you served on that. Was that a clear mandate to you that there should be this type of coordination? Dr. NELSON. There should be coordination of regulation between agencies but there is not. There is available the machinery for co- ordination of testing requirements; there is no question about that. And formerly, under the IRLG, there was an opportunity for the administrative heads to come together to discuss interlocking regu- latory issues. That is not functional now. So, there are regulatory resonsibilities not now being met. There is a gap here which I think suggests to me that the IRLG or something like it ought to be reconstituted. OK? PAGENO="0099" 95 Ms. MIKULSKI. It is not functional now? Dr. NELSON. It is not functional now. Ms. MIKuLSKI. And when did it stop funtioning, and why? Dr. `NELSON. May I refer to~ a person who can give you the specif- ic date and time? Would that be in order? Dr. Anderson, when did the IRLG cease functioning? Early in the present administration, Dr. Anderson states. This unit provided what a minimum step, but at least it was a place where the regulatory agencies, Consumer Product Safety Commis- sion, OSHA, EPA-What have I missed?-FDA, came together with those concerned with basic science, the basic science going into toxicological evaluation, not only cancer but other issues. That activity was abandoned. Now, I told you about the chemical selection for testing. That is only a small part of the program. Ms. MIKuLSKI. Do you know why it was abandoned? Dr. NELSON. I can only utter a suspicion that the IRLG activity was contrary to the attempt of the present administration to reduce regulatory impact. Mr. RITTER. This is prior to the Reagan administration, prior to this present administration? Ms. MncuLSKI. Excuse me, Mr. Ritter. I have the time. Dr. NELSON. Again, I need more specific information. I am told that it expired in September 1981. Now, the concern you express, I think, is an appropriate one. Ms. MIKuLSKI. I want to stick to this point, because what we are trying to see is, what are the mechanisms in place for evaluating the issue of cancer, and one of the most important things is coodin- ation between agencies, and therefore not only getting a dollar's worth of research services for a dollar's worth of taxes, but assess- ing this impact. Now, if the main mechanism for coordination is not in place, that is what I wanted to find out. Dr. NELSON. I would say only a small component of the appropri- ate mechanisms are in place, and I speak here of the selection of chemicals for testing. That is only a small part. Much more at issue is the question of the coordination of regulatory policy. That is, to the best of my knowledge, not in place. There may be a sub- stitute. I am not aware of it. However, let me make one more point, and that is, at the more basic research level, this issue is very high up in the level of con- cern of the National Institutes of Health. Both the National Insti- tute of Environmental Health Sciences, and the National Cancer Institute are very much aware of these issues. And at the basic level, there is intense interest and concern for the recognition that carcinogens can be found in many places, and it is a scientific problem of complexity, but it must be tackled. That is under way, and I think the~coordination there is pretty fair. The big gap I see is not at that level, nor at the level of selection of chemicals for testing, but at the coordination of Federal policy. Ms. MIKuLSiu. Does anyone else wish to add to that? Dr. PIT0T. That was very well stated. Ms. MIKULSKI. Mr. Chairman, I have no other questions. PAGENO="0100" 96 Mr. FLORTO. Gentlemen, thank you very much. You were very generous with your time and very helpful to the committee. Thank you. Our next panel is comprised of Dr. Frederica Perera, Dr. Ellen Silbergeld, Dr. Robert A. Neal, and Mr. Jackson Browning. We would appreciate if it the individuals would come forward. Dr. Perera is the senior staff scientist, Natural Resource Defense Council. Dr. Silbergeld is the chairman of the toxic substance pro- gram of the Environmental Defense Fund. Dr. Neal is director of the Chemical Industry Institute for Toxicology. And Mr. Browning is the corporate director of the health, safety, and environmental affairs program of Union Carbide Corp., who is here representing the American Industrial Health Council. Ladies and gentlemen, we appreciate your participation this morning. Your statements will be made a part of the record in their entirety. You may feel free to proceed in a summary fashion. Dr. Perera, we appreciate your going forward. STATEMENTS OF FREDERICA P. PERERA, DR. P.H., SENIOR STAFF SCIENTIST, NATURAL RESOURCES DEFENSE COUNCIL; ELLEN K. 5ILBERGELD, PH.D., CHIEF TOXICS SCIENTISTS, ENVIRON- MENTAL DEFENSE FUND; ROBERT A. NEAL, PH.D., PRESIDENT, CHEMICAL INDUSTRY INSTITUTION OF TOXICOLOGY; AND JACKSON B. BROWNING, MEMBER, BOARD OF DIRECTORS, AMERICAN INDUSTRIAL HEALTH COUNCIL Dr. PERERA. Thank you. Good morning, or is it good afternoon? I am Dr. Frederica Perera. I am a senior staff scientist with the Natural Resources Defense Council. My testimony today will ad- dress a matter of great concern to the public health and public in- terest community, the de facto relaxation of Federal cancer policy as evidenced by the actions of the Environmental Protection Agency during the last 2 years. As you know, cancer is the second leading cause of death in the United States, resulting in over 400,000 deaths a year. During their lifetimes, more than one in four Americans will be diagnosed with cancer. One out of five will die of the disease. The great majority, about 80 percent of cancers, are believed to be caused by environ- mental factors, including lifestyle and chemicals in the air, water, and food. Hence, this large fraction is considered to be preventable by limiting human exposures to cancer-causing substances and agents. As you have heard, considerable uncertainty surrounds the com- plex and multistage process of cancer, as well as the available methodology for identifying and quantifying carcinogenic risks. Therefore, Congress and Federal agencies have in the past consist- ently stressed the need for a conservative and protective approach in carcinogen identification and regulation. It is thus with growing concern that we have watched the EPA informally adopt a less protective policy towards carcinogenic risks. These changes have not been published in the Federal Register as -~ were the EPA 1976 and the IRLG 1979 Scientific Report, but are evident instead in various recent actions by EPA which show: (1), a discounting of positive animal testing data; (2), an emphasis on the PAGENO="0101" 97 need for positive epidemiological or human data; and (3), assump- tion of lower risk and even thresholds for certain types of carcino- gens. I would like to discuss these three significant revisions in policy, as well as the agency's apparent willingness to tolerate higher cancer risks than were allowed by prior administrations. The first revision is the discounting of valid positive animal tests as presumptive evidence of carcinogenicity in humans. Federal agencies have hitherto operated on the principle that a valid posi- tive bioassay result in a single species constitutes evidence that the substance is a potential human carcinogen and that it is not neces- sary to await positive human data before taking steps to limit human exposure. The EPA Interim Cancer Policy adopted in 1976 by the agency and the IRLG policy, which EPA helped to write, along with the representatives from the major regulatory agencies and senior scientists from NCI and the National Institute of Envi- ronmental Health Sciences, were both based on this premise. EPA has recently disregarded these principles in assessing the human risk of formaldehyde. By way of background, formaldehyde is a chemical of seven billion pounds annual production. It is found in the workplace, the home, in numerous consumer products, and in the ambient air. It has been judged to be an animal carcinogen, and hence a presumptive human carcinogen, by three major na- tional and international panels, by the directors of major U.S. cancer research institutes, and by numerous other scientific ex- perts. In the spring of 1981, after receiving CuT data showing formal- dehyde to cause rare nasal tumors in animals, the EPA staff rec- ommended that formaldehyde be designated for priority review under section 4[f] of the Toxic Substances Control Act. EPA re- versed that decision last year, in a February 1982, memo written by Dr. John Todhunter, EPA Assistant Administrator for Pesti- cides and Toxic Substances, which constituted final agency action. Two of the reasons given in the Todhunter memo for deciding not to act on formaldehyde were the existence of negative studies and only two clear positive results-so that the memo referred to formaldehyde as a potential animal carcinogen-and the absence of positive epidemiological data. The second departure from previous policy is the reliance on the- ories concerning mechanism of action and the assumption of prob- able thresholds to justify low concern for human risk of cancer. In the case of formaldehyde, EPA decided that despite positive animal carcinogenicity data and extensive positive short-term test data, formaldehyde was probably acting as an epigenetic or a threshold type of carcinogen. Yet, as we have heard from the previous panel, the assumption of human population thresholds for carcinogens is in direct contradiction to all Federal and State cancer policies pro- posed or finalized to date. Another recent EPA decision which reflects greater reliance on assumptions concerning mechanism of action is expanded registra- tion of the insecticide permethrin based on its classification as an epigenetic carcinogen, and hence as one of presumed low risk. This emphasis on mechanism of action and classification of chemicals as genotoxic or epigenetic echoes the EPA draft adden- PAGENO="0102" 98.. dum of June 1982. This proposal was not published in the Federal Register, but did receive informal written comment by 21 review- ers. I would like briefly to describe the proposal and the responses it received, most of which were highly critical of its scientific basis. Up to now, I would remind you that EPA has conventionally used the linear, no threshold, multistage model to estimate the risk of all carcinogens. According to the addendum however, evidence of gene-specifically point-mutation would be required to classify a chemical carcinogen as genotoxic and hence to justify using the no threshold linear extrapolation model to assess risk in setting stand- ards. For non-genotoxic or epigenetic carcinogens, the no observed effect level or NOEL safety factor approach would be used to iden- tify a safe human exposure level. Of the 21 scientists who reviewed the proposal in written com- ments, the majority criticized the underlying assumptions as lack- ing scientffic justification. Some of the recurrent criticisms were: (1) not enough is known about mechanism of action in chemical carcinogenesis to distinguish between them as a basis for regula- tory decision-making; (2) dose-response curves are not necessarily different for genotoxic versus epigenetic agents nor can thresholds be assumed for the latter; and (3) was the use of the NOEL, no ob- served effect level, approach is inappropriate for carcinogenicity, and was rejected long ago for chronic, irreversible, self-replicating effects. Disturbingly, in light of the extensive criticism this proposal re- ceived, the last chapter of the Office of Science and Technology Policy Draft Federal Cancer Policy of October 1982 alluded to this approach as one to which consideration was being given. Lastly, I would like to discuss the acceptance by EPA of a higher tolerable level of cancer risks for the U.S. population-100-fold higher than was sanctioned under previous administrations. We have seen this change in EPA's pesticide program and in the agency decision on formaldehyde. In fact, the Todhunter memo pro- vides a rationale for relaxing regulatory goals from lifetime cancer iisks of no greater than one in a million to between one in 10,000 and one in 100,000. In doing so, the memo ignores the fact that risk estimates are crude and could be "off' by several orders of magnitude. It fails to distinguish between voluntary versus involuntary exposures to car- cinogens, and between identifiable and hence readily preventable and nonreadily preventable causes of cancer. Most importantly, it neglects the possibility of additive and possibly synergistic-that is, greater than additive-effects among carcinogens. In the draft EPA addendum as well there is an acceptance of substantially higher risks for all carcinogens. At the present time, EPA issues estimates of exposure associated with lifetime cancer risks of one in 100,000 up to one in 10 million to serve as guidance to the States for purposes of setting standards for waterborne car- cinogens. New York State and New Jersey, for example, have based water quality criteria for carcinogens on the one in a million risk level. The addendum, however, provides for EPA to issue guid- ance for genotoxic carcinogens based only on a risk of one in 100,000. PAGENO="0103" 99. In particular, the procedure proposed for setting standards for nongenotoxic carcinogens means that the two methods, that is, the no observed effect level and the conventional linear extrapolation to one in 100,000, will differ by a factor of 100 to 500. Compared with guidance based on one iii a million risk estimated by linear extrapolation, there would be a 1,000 to 5,000-fold difference. Table 2, given in the addendum itself, shows that for the potent carcinogen dioxin the difference between water quality standards arrived at by the two methods as proposed-that is, linear extrapo- lation to one in 100,000, and the no observed effect level-would be as much as 181-fold; for chloroform, 169-fold; for carbon tetrachlo- ride, 203-fold, and for DDT, 158-fold. Compared, of course, to the linear extrapolation to one in a million, the differences would be another 10-fold greater. Mr. FL0RI0. Could I interrupt at this point? Is what you are saying that the implementation of the policy that is being dis- cussed today would effectively result in permissibly expanding tol- erance or exposure levels in the way that you have talked about? Dr. PERERA. Yes, that is right. I am simply quoting you numbers that were found in Table 2 of the EPA addendum. Mr. FL0RI0. Thank you. Dr. PERERA. In conclusion, I fully support the use of more precise and reliable scientific methods for identifying and assessing human risks of cancer. However, it is clear that such methods should be validated and viewed as operational by the scientific community before they are incorporated into regulatory policy. As we have seen, the significant changes in policy described * above have little or no scientific basis. They rest largely on as- sumptions about the complex and still little understood process of cancer. As such, they reflect a turnaround in the formerly conserv- ative response to scientific uncertainty embodied in the environ- mental statutes and in established Federal policies to date. They reflect a willingness to accept the likelihood of substantially higher risks of human cancer than hitherto tolerated to reach political goals. I believe, therefore, that it is a mistake to view these revisions as a purely scientific matter, and to cloak them in science, so to speak. Instead, just as a thorough peer review of the scientific basis of policy is essential, so must the philosophical and political basis be openly discussed. In that debate, it may come out that the country is willing to accept higher risks of cancer as a tradeoff for near-term economic benefits, or it may be determined, as the poils certainly suggest, that health protection remains a closely held and preeminent value. In any case, in prior administrations there has been formal pub- lication and public review of Federal reports and proposals con- cerning cancer policy before they are put into effect. It is more im- portant than ever before that this tradition be maintained, and that we avoid de facto policies that ignore valid and weil-estab- lished principles. Thank you very much. [Testimony resumes on p. 157.] [Dr. Perera's prepared statement and attachments follow:] PAGENO="0104" 100 TESTIMONY OF FREDERICA P. PERERA, Dr.P.H. ON BEHALF OF THE NATURAL RESOURCES DEFENSE COUNCIL, INC. My name is Frederica Perera. I am a senior staff scientist with the Natural Resources Defense Council (NRDC). NRDC is a national, non-profit public interest organization with over 40,000 members across the country. One of the major goals of our scientific and legal staff is to prevent and limit human exposure to toxic and carcinogenic substances. specifically, we have advocated prevention of cancer through testing of new chemicals before manufacture to screen out potential human carcinogens; limits on the discharge of cancer-causing chemicals to water supplies, air and the work- place; and restriction of their use in food and consumer products. My testimony today will address a matter of great concern to the public health and public interest community -- the de facto relaxation of federal cancer policy as evidenced by the actions of the Environmental Protection Agency during the last two years. Cancer is the second leading cause of death in the United States, resulting in over 400,000 deaths a year. During their lifetimes, more than one out of four Americans will be diag- nosed with cancer; one out of five will die of the disease. The great majority (about 80%) of cancers are believed to be caused by environmental factors, including lifestyle, and chemicals in the air, water and food. Hence, this large fraction is considered to be preventable by limiting human exposures to cancer-causing substances. As you have heard, considerable uncertainty surrounds the complex and multistage process of cancer, as well as the available methodology for PAGENO="0105" 101 identifying and quantifying carcinogenic risks. Therefore, Congress and federal agencies have in the past consistently stressed the need for a conservative and protective approach in carcinogen identification and regulation. It is thus with growing concern that we have watched the Environmental Protection Agency (EPA) informally adopt a less protective policy towards carcinogenic risks. These changes have not been published in the Federal Register, as were the EPA 19761 and the Interagency Regulatory Liaison Group (IRLG) 19792 policies, but are evident instead in various recent actions by EPA which show (1) a discounting of positive animal testing data, (2) an emphasis on the need for positive epidem- iologica]. (or human) data, and (3) assumption of lower risk and even thresholds for certain types of carcinogens. Con- cerning the latter, in its clearest statement of intent to date the Agency has drafted a proposal3 to revise its 1976 policy by using linear extrapolation models only for qenotoxic carcinogens. This approach could relax by over a hundred-fold the regulatory standards for epigenetic carcinogens.*4 Although that document has not~ yet been finalized, it echoes across a number of Agency decisions. It echoes as well in the final chapter of the draft federal cancer policy of the Office of Science and Technology Policy (OSTP) of October 1982, which * The former are defined by EPA as causing gene mutations (specifically point mutations) while "epigenetic" embraces "all other" mechanisms. PAGENO="0106" 102 stated that consideration was being given to using this approach. ~ I would like to discuss these three significant revisions in EPA's policy towards identifying and regulating carcinogens: 1) The first is the discounting of valid positive animal tests as presumptive evidence of carcinogenicity in humans. According to the Office of Technology Assessment (1980), federal agencies have hitherto operated on the principle that a valid positive bioassay result in a single species consti- tutes evidence that the substance is a potential human car- cinogen.6 The EPA Interim Cancer Policy adopted in 1976 by the Agency and the Interagency Regulatory Liaison Group (IRLG) Cancer Policy of l979,* which EPA helped to write, were based on the premise that a substance should be presumed to be a human cancer risk if it causes a statistically significant excess incidence of tumors in humans or animals. EPA has recently disregarded this principle in assessing the human risk of formaldehyde. By way of background, formal- dehyde is a chemical of 7 billion lbs. annual production; it * The policy was published as a scientific report in the Federal Register which stated that the report `represent- ed the best judgements of these scientists [senior scien- tists at NCI and NIEHS} and those of the four agencies (CPSC, EPA, FDA, and OSHA) comprising the IRLG. . ." and that "after reviewing the comments received, the four agencies anticipate publishing a statement giving notice of whatever revisions to the document are appropriate, if any." (29858) No revisions were published. PAGENO="0107" 103 is found in the workplace, the home, in numerous consumer products, and in the ambient air. It has been judged to be an animal carcinogen and hence a presumptive human carcinogen by three major national and international panels79, by the Directors of the National Cancer Institute (NCI), National Institute of Occupational Safety and Health (NIOSH), National Center for Toxicological Research (NCTR), and by numerous other scientific experts. (See~ attached article.) In the Spring of 1981, after receiving Chemical Industry Institute of Toxicology (CuT) data showing formaldehyde to cause a statistically significant increase of rare nasal tumors in animals, EPA staff recommended that formaldehyde be designated for priority review~ under Section 4(f) of the Toxic Substances Control Act in order to investigate exposures of greatest concern and determine the need for regulation. EPA's reversal of that decision last year was based in large part on a questioning both of the validity of the animal data and of their relevance to~ human risk. A February 1982 memo written by Dr. John Todhunter, EPA Assistant Administra- tor for Pesticides and Toxic Substances, constituted final Agency action.'░ Two of the reasons given in the Todunter Memo (TM) for deciding not to act on formaldehyde were: 1) the existence of negative studies and "only" two clear positive results; 2) the absence of positive epiderniological data. Concerning the first point, as a matter of principle, negative studies for Carcinogenicity are not given the same PAGENO="0108" 104 weight as valid positive studies because of the possibility of "false negative" results, which have a far more serious conse- quence for human health than false positives." In the case of formaldehyde, prior negative bioassays were judged by expert reviews to be either seriously flawed or suggestive of a positive effect and hence not to detract from the positive studies. Thus, it was surprising to read in the Todhunter Memo that because "only" two studies have given clear positive results and other studies have been "generally negative", Concern that formaldehyde gas may induce tumors in humans should be tempered by this observation that formaldehyde appears to have a high degree of species specificity and a strong dependence on route of exposure. (TM, p.7) The memo's conclusion that formaldehyde is only a "potential animal carcinogen with mode and degree of exposure quite important to the toxic outcome" (p.8) stands in sharp contrast to all prior judgements of experts mentioned above. It is illustrative of the shift in policy that the Todhunter Memo roundly criticizes the former EPA staff review for downplaying the negative bioassays in concluding that formaldehyde is an animal carcinogen. (TM, p.13) Second, as a matter of principle, agencies have not prev- iously awaited or required positive human data before taking prudent steps to limit exposure to a carcinogen. Expert reviews have termed the available epidemiological data for formalde- hyde severely limited, hence inconclusive and possibly suggest- ive -- but not negative.11 It was therefore anomalous to read in the Todhunter Memo that the decision was partly based on the PAGENO="0109" 105 fact that there were no clearly positive epidemiological studies: If formaldehyde were a potent human risk, this would show up epidemiologically. There does not appear to be any relationship, based on the existing data base on humans, between exposure and cancer. (TM, p.11) 2) The second departure from previous policy is the reliance on theories concerning mechanism of action and the assumption of probable thresholds to just~ify low concern for human risk of cancer. In the case of formaldehyde, EPA decided that, despite positive animal carcinogenicity and extensive positive short- term test data, formaldehyde was not likely to pose a signifi- cant human risk because its carcinogenic effect was apt to be secondary to physiological events triggered only at high levels. Thus, the memo concluded that formaldehyde was prob- ably acting as an epigenetic or threshold type of carcinogen: When taken together with the observations reported by CuT on the reversibility of hyperplastic and meta- plastic effects.. . and an apparent cytotoxicity thres- hold the degree to which formaldehyde promotes its own carcinogenesis may lead to qualitative, as well as quantitative, differences in response within different regions of the dose-response curve. (TM, p.9) The implication that thresholds can be assumed to exist for carcinogens is in direct contradiction to all federal and state cancer policies proposed or finalized to date. These include the EPA (1976) policy, the IRLG (1979) Report, the OSFIA (1980) Cancer Policy,12 and California's Proposed Cancer Policy (1982_83).13 There is broad consensus that data are extremely limited concerning low dose-response relationships and that methods do not exist to identify human population thresholds. (See attached art~icle, p.1288-1289.) PAGENO="0110" 106 Other recent EPA decisions which reflect a greater reli- ance on assumptions concerning mechanism of action are the expanded registration of the insecticide permethrin for wider use in agricultural products. Despite positive animal carcinogenicity data in one study with mice and suggestive results in two additional studies, EPA issued as a final rule an Acceptable Daily Intake (ADI) of 0.05 mg/kg/day, based on noncarcinogenic effects.14 (This level of exposure, according to calculations by EPA staff, corresponds to an estimated excess lifetime cancer risk of 1 x lOs, or one in a thousand.)15 Part of the rationale for this decision was the categorization of permethrin as an "epigenetic" carcinogen (and hence as one of presumed low risk) using the systems proposed by Weisburger and Williams16 and Squire.17 [The] lack of positive evidence for mutagenic potential from a battery of tests, including DNA repair and unscheduled DNA synthesis, coupled with oncogenic evidence only at high dose levels for one sex in one mouse study among three studies, indicates that permethrin falls into the .epigenetic category. According to this system of classification [Weisburger and Williams], permethrin falls into the group where the risk from exposure may be of a quantitative nature. 8 Then using the theoretical Squire ranking system which gives 25 of 100 points for genotoxicity, EPA determined that permethrin clearly belongs in the class of lowest potential human risk of carcinogenicity. (Both of these systems are highly theoretical and as yet not validated;.yet here EPA appears to be using them systematically to factor mechanism of action into the risk assessment.) PAGENO="0111" 107 EPA has also evidenced a s~trong reliance on mechanism of action in deciding that not enouqh was known about the carcinogenic mechanism of action of the fungicides ethylene bisdithiocarbamates (EBDCs) to justify cancellation of their registrations, although animal tests have shown statistieally significant incidences of several types of tumors.19 As mentioned earlier, this emphasis on mechanism of action and classification of chemicals as genotoxic or epigen- etic echoes the EPA draft Addendum (June 1982) to the 1976 Cancer Policy.20 This proposal~ was not published in the Federal Register, but received informal written comment by 21 reviewers, including myself. I would like briefly to describe the proposal and the responses it received -- most of which were highly critical of the scientific basis of the proposed dichotomous approach to carcinogen regulation. According to the Addendum,~ evidence of gene (point) muta- tion would be required to classify a chemical carcinogen as genotoxic and hence to justify using a no threshold linear extrapolation model to assess risk and set standards. For non-genotoxic or epigenetic carcinogens, the No Observed Effect Level (NOEL)/Safety Factor approach would be used to identify a "safe' human exposure level. As will be discussed, the resulting standards could be several orders of magnitude less stringent than those derived by using the risk extrapola- tion methods. Of the 21 scientists who reviewed the proposal in written comments to Dr. Roy Albert, Chairman of the EPA Carcinogen Assessment Group (CAG), the majority criticized various of the PAGENO="0112" 108 underlying assumptions as lacking scientific justification.* Recurrent criticisms were as follows: 1) Not enough is known about mechanisms of action of chemical carcinogeflesis to utilize a distinction between them as a basis for deciding whether and to what degree they should be regulated. 2) It is incorrect to assume that the single, most plausible theory of carcinogenesis and the mechanism of greatest concern is somatic cell point mutation and that epigenetic agents such as promoters, for example, are "safer". 3) Dose-response curves are not necessarily different for genotoxic versus epigenetic agents; nor can thresholds be assumed for the latter. 4) Even if the distinction were useful, our ability to reliably and accurately classify carcinogens according to mechanism of action is limited. 5) The assumption that the linear non-threshold multi-stage model used by EPA's CAG is dependent on a one-hit scenario of mutagenicity/carcinOgefl icity is incorrect: linearity is, in fact, justified by the assumption of background additivity; i.e., that the individual is exposed to other carcinogens of which at least one is acting via the same mechanism. 6) The use of the NOEL is inappropriate for carcino- genicity and was rejected long ago for chronic irreversible self-replicating effects. Many of these criticisms were succinctly stated by Dr. Weinstein of Columbia University in a recent letter published in Science. 21 * These included Dr. Ron Hart, Director of NCTR; Dr. David Hoel, NIEHS; Dr. Norton Nelson, NYU, Dr. Umberto Saffiotti, NCI; Dr. Lorenzo Tomatis, Director of IARC; Dr. Arthur Upton, NYU; Dr. Sidney Weinhouse, Temple University; Dr. I.B. Weinstein, Columbia University; Dr. Elizabeth Weisburger, NCI; and myself. PAGENO="0113" 109 Disturbingly, in light of the extensive criticism this draft received, the last chapter of the Office of Science Technology Policy (OSTP) draft federal cancer policy of October 1982 alluded to its approach as one to which consideration was being given. (See appended comments.) That draft stated: Under certain circumstances the weight of biological information may raise questions about the appropriate- ness of extrapolation models. Traditionally, toxico- logical endpoints have been evaluated quantitatively by using margins of safety placed on information ob- tained in the observed region of the dose-response cure. Consideration is being given to applying this technique, in selected instance~ alone or in combina- tion with extrapolation models. 3) Lastly, I would like to discuss briefly the acceptance by EPA of a higher tolerable level of cancer risks for the U.S. population (lOOx or higher) than was sanctioned under previous administrations. EPA has permitted registration of pesticides (e.g., per- methrin and benomyl) which the Agency has estimated are assoc- iated with lifetime cancer risks~ of l0~ (one in a thousand) to l0~ (one in one hundred thousand) when previously the Agency regulated substances such, as DBCP, dieldrin, and hepta- chlor on the basis of estimated risks of 1 x 10-6 (one in a million) 23 The Todhunter Memo provides the following rationale for relaxing regulatory goals from 1 x 10-6 (one in a million) to between 1 x l0~ (one in ten thousand) and 1 xl05 (one in one hundred thousand): 22-143 0-83-8 . PAGENO="0114" 110 In terms of individual lifetime cancer risks, the various fe~eral agencies do not tend to regulate risks of 1 x 10 or lower ~nd tend to b~ ambivalent about risks between 1 x 10 and 1 x lO~. Certainly (as absolute risks) these risk levels could never ~e de- tected any normal way and would (using 1 x 10 as an upper bound) represent increments of 0.03% or less above the estimated individual risk of persons in the U.S. population as a whole. (TM, p.5) Commenting on this statement during a recent hearing, Dr. Norton Nelson, former director of the Institute of Environ- mental Medicine at New York University stated: At one point Dr. Todhunter.. .notes that certain levels of cancer are probably undetectable against the background of the normal occurrence of cancer... . The point to be made here is that detectability should never, under any circumstances, be grounds for disregarding a cancer risk. It is the absolute impact, not the percentage or relative impact, that society needs to be concerned with. This statement in the Todhunter memo also ignores the fact that risk estimates are crude and could be "off" by several orders of magnitude -- 30 that a true risk for an estimate of 1 x l0~ might be 1 x 10-2 or 1 x 10-6. It also does not distinguish between voluntary versus involuntary exposures to carcinogens and between identifiable, hence preventable, and non-preventable causes of cancer. Most importantly, it neglects the possibility of additive and possibly synergistic effects among carcinogens. Regarding additivity, if we assume that 3.6 million people are born and an equal number die each year in this country, exposure to one carcinogen with a lifetime cancer risk of 1 x l0~ could result in (3.6 x 106) x (1 x l0~) = 360 additional deaths per year. If there were exposure to ten such carcinogens, annual excess death would PAGENO="0115" ill number 3600 and so on. This is a~rough approximation, but helps to put lifetime cancer risks in human perspective.* In the draft EPA addendum as well, there is an acceptance of higher risks for all carcinogens. At the present time, EPA uses the CAG multi-stage model and a conservative species con- version factor (surface area) to Ŕalculate Upper Confidence Level (95%) estimates of exposures associated with lifetime risks of 1 x l0~; 1 x l06; and 1 x l0~, to serve as guidance to the states for purposes of setting standards. New York State and New Jersey, for example, have based water quality criteria for carcinogens on the 1 x 10-6 risk level. The Addendum, however, provides for EPA to issue guidance for genotoxic carcinogens based ~ on a risk of 1 x l0~. For non-genotoxic agents the proposal calls for estima- tion of 10% response in animals using the multistage model to derive the Maximum Likelihood Estimate (MLE) which is general- ly less conservative than the UCL, by a factor of 2-3. An additional 6-12 fold difference, depending on whether rat or mice data are used, is built in by use of a weight rather than surface area conversion factor. Then the proposal states that a safety factor of 1000 will be applied. According to Dr. David Hoel of NIEHS, this procedure means that the two methods * Dr. Marvin Schneiderman of Clement Associates calculates that if one assumes that these 3.6 x 10 newborns each year are exposed to only 10% of the 70,000 existing chemicals, that only 10% are~carci~ogenic, and that each conveys a lifetime risk of 1 x 10░, one can estimate that an ~dditional 2,520 cases of cancer per year would result. 2~ PAGENO="0116" 112 (NOEL and conventional linear extrapolation to 1 x lOs) will differ by a factor of lOO_500.26 Compared with guidance based on 1 x 10-6 risks estimated by linear extrapolation, there would be a 1000-5000 fold difference. Table 2 given in the Addendum itself shows that for the potent carcinogen TCDD (dioxin), the difference between water quality standards arrived at by the two methods as proposed would be 13-181 fold; for chloroform, 13-169-fold; for carbon tetrachloride, 14-203 fold; for hexachlorobutadiene, 106-fold; and for DDT, 158-fold. (The lower number for the first three compounds represents the midrange between the linear and NOEL estimates.) CONCLUS ION In conclusion, significant advances have been made in the understanding of chemical carcinogenesis -- particularly in the development of short-term screening tests to detect potential carcinogens and of potential markers of human exposure to such agents. NRDC fully supports the use of more precise and reliable scientific methods for identifying and assessing human risks of cancer. However, it is clear that such methods should be validated and viewed as operational by the scientific community before they are incorporated into regulatory policy. As we have seen, the significant changes in policy described above have little or no scientific basis. They rest largely on assumptions about the complex and still little understood process of cancer. As such, they reflect a turnaround in the formerly conservative response to scientific PAGENO="0117" 113 uncertainty embodied in the environmental statutes and in established federal policies to date. They reflect a willingness to accept the likelihOod of substantially higher risks of human cancer than hitherto tolerated to reach political goals. I believe therefore that it Is a mistake to view these revisions as a purely scientific matter and to "cloak them in science' as it were. Instead, just as a thorough peer review of the scientific basis of policy is essential, so must the philosophical and political basis be openly discussed. In that debate, it may come out that the country is willing to accept higher risks of cancer as a tradeoff for near-term economic benefits; or it may be determined -- as the polls suggest -- that health protection remains a closely held and preeminent value. In any case, in prior administrations there has been formal publication and public review of federal reports and proposals concerning cancer policy~ before they are put into effect. It is more important than ever before that this tradition be maintained and that we avoid "de facto" policies that ignore valid and well established principles. PAGENO="0118" 114 REFERENCES 1 Environmental Protection Agency. 41 Federal Register 21402 (25 May 1976.) 2 Interagency Regulatory Liaison Group, 44 Federal Register 39858 (6 July 1979.) 3 Environmental Protection Agency: Additional U.S. Environ- mental Protection Agency Guidance for the Health Assessment of Suspect Carcinogens with Specific Reference to Water Quality Criteria, Draft (June 21, 1982.) 4 Comment on reference 3 to Dr. Roy Albert by Dr. David Hoel, NIEHS (July 2, 1982), p.2. 5 Office of Science and Technology Policy: Potential Human Carcinogens: Methods for Identification and Characterization, Part I (October, 1982.) 6 Office of Technology Assessment: Assessment of Technologies for Determining Cancer Risks from the Environment. OTA, Washington, D.C. (June, 1981.) 7 Griesemer, R.A., Chairman: Report of the Federal Panel on Formaldehyde, National Toxicology Program, Research Triangle Park, N.C. (November 1980.) Published in Environmental Health Perspectives 43:139-168, 1982. 8 Selikoff, I.J. etal.: Carcinogenicity of Formaldehyde: Final Report. Report to the American Cancer Society (February 25, 1981.) 9 IARC Monograph, Volume 20, pp.346-389, International Agency for Research on Cancer, Lyon (May 1982). (Report of the Working Group which met 13-20 October 1981.) 10 Todhunter, J.: Review of Data Available to the Administrator Concerning Formaldehyde and di(2-ethylhexyl) Phthalate (DEHP). Memorandum to A.M. Gorsuch (February 10, 1982.) 11 References 7 and 8, supra. 12 Occupational Safety and Health Administration: 45 Federal Register 5002 (25 January 1980). 13 Carcinogen Identification Policy: Sections 1 and 2 (July and October 1982) and Cancer Policy (December 1982), Department of Health Services, State of California. PAGENO="0119" 115 14 47 Federal Register 45008. 15 Litt. B.: Memorandum to O.E. Paynter, Chief, Toxicology Branch, entitled Permethrin Oncogenicity Evaluation: FMC- Mouse II Study Liver and Lung Pathology Findings in Females, EPA (February 10, 1982), Table IV. 16 Weisburger, J.H. and Williams, G.M.: Chemical carcinogens. In Doull, J., Elassen, C.D., and Arndur, M.O. (eds.): Toxicology, The Basis Science of Poisons, 2nd ed. Macmillan Publ.Co., New York, 1980, pp.84-138. 17 Squire, R.A.: Ranking animal carcinogens: A proposed regulatory approach. Science 214:877-880, 1981. 18 Perrnethrin, Assessment of Chronic and Oncoqenic Effects: A Summary. Hazard Evaluation Division, U.S. EPA (September 1982, p.29.) 19 Testimony of A.K. Ahmed, Natural Resources Defense Council, Before the House Subcommittee on Department Operations, Research and Foreign Agriculture (February 22, 1983.) 20 Ref. 3, supra. 21 Weinstein, I.B.: Carcinogen policy at EPA. Science, 219:794-796, 1983. 22 Ref. 5, supra, Chapter VII, p.8. 23 Ref. 19, ~gpra. 24 Nelson, N.: Testimony Before the Subcommittee on Investigations and Oversight of the Committee on Science and Technology (May 20, 1982.~) 25 Schneiderman, M.A.: Cost-benefit, social values and the setting of occupational health standards. In D.S. Lee and W.N. Rom (eds.): Legal and Ethical Dilemmas in Occupational Health. Ann Arbor Sci. Publ., Ann Arbor, 1982, p.204. 26 Reference 4, supra. PAGENO="0120" 116 DEPARTMENT OF HEALTH & HUMAN SERVICES Pubhc Health Servrce Food and Drug Adm~n~stration Washington DC 20204 December 2, 1982 Dr. Frederica Perera National Resources Defense Council 122 East 42nd Street New York, NY 10168 Dear Dr. Perera: Enclosed is my review of the chapter on short-term testing. It is not as detailed as I would like it to be but it would require more time for an in depth evaluation. Nevertheless, I hope that what I have done will be helpful. If you have any questions about the review please call me. Si~erely yours, ~ ~LL~4~ Virgj~nia C. Dunkel, Ph.D. Chi4~, Genetic Toxicology Branch, Bure&u of Foods, HFF-l66 PAGENO="0121" 117 Overall, Chapter IV entitled Current Views on Short-Term Testing of the Federal Cancer Policy is poorly written and does not present a balanced and informative perspective of the field of genetic toxicology. The document should be rewritten with a logical presentation of the types of endpoints that can be measured, the critical aspects of metabolic activation, the extend, limitations, and pitfalls associated with correlations, the uses to which the tests can be applied and finally future research needs. More specifically problems in the various sections are as follows: Background-Rationale: An omission in this section involves references to the early work of Malling and Smith. Prior to the work reported by Gabridge and Legator on the host mediated assay, Malling (Mutat. Res. 3:537-1274, 1966) demonstrated that DMN and DEN could be converted to mutagens for Neurospora crassa and Smith (Science, 192-1272-1274, 1966), at the same time, reported that cycasin yielded a mutagenic product (methylazoxymethanol) on enzymatic hydrolysis that was mutagenic for Salmonella. These are important in the development of the field of genetic toxicology and should be included in any background discussion. As is evident~ from the references, the work of Gabridge and Legator then followed these reports. In addition, there are really no mechanistic hypotheses (page 2) to explain the correlation between mutagenicity and carcinogenicit~. What the author(s) may be striving for is justification for the use of tests with genetic endpoints based on the somatic-mutation theory OL carcinogenesis. The somatic mutation theory however i~ not a total explanation since there PAGENO="0122" 118 are a number of aspects of carcinogenesis that are inconsistent with the somatic mutation theory such as the higher incidence of tumors in immunologically impaired individuals and solid-state carcinogenesis. The latter appears to be indicated as "local irritation". However, the chronic irritation theory to explain tumor induction has been disproved (Berenblum I. Carcinogenesis as a Biological Problem North Holland Publishing Co., 1974, p. 276-277). Test Currently Available Under this heading are subparts dealing with tests for gene mutation, chromosomal aberrations, DNA damage and repair and cellular transformation. Discussion of the various endpoints that can be measured is absolutely necessary in any discussion about short-term tests but in this document there is no overall presentation of what is actually measured. For example under the section on tests for chromosomal aberrations there is no definition or description of the types of aberrations that can be evaluated. It would be beneficial to state that chromosomal aberrations are characteristic of damage sustained in C1 cells which are translated into breakage/exchange figures prior to chromosome replication and that these are detected as breaks, terminal and interstitial deletions, rings, translocations and dicentrics. Something similar should be done for all the endpoints discussed and this should be followed by a short description of the systems most widely used. Referencing should also be to primary citations rather than to the Gene-Tox reports. All information on correlations needs to be deleted since `this is very misleading and provides no really valuable information. The merged PAGENO="0123" 119 carcinogen list used for establishing the correlations between carcinogenicity and mutagenicity results does not contain well known carcinogens. Dimethvlnitrosamine, 3-methylcholanthrene and 2-acetylaminofluorene are some examples of carcinogens that are not on the list. In addition, the percentage correlations do not provide information on specificity and sensitivty of a particular methodology. On page 7, for example, it is stated that "19/23 carcinogens (83%) were correctly identified in V-79 cells." This is totally misleading since a major portion of the compounds tested were polycyclic hydrocarbons. This fact is acknowledged in the text of the Gene-Tox Report and under Recommendations for further study (page 138) it is stated that "Nore classes of chemicals need to be studied." Finally of what value are statements such as "83% (5/6) of the noncarcinogens tested in human diploid fibroblasts were correctly identified.. .. (page 13) or "of 164 chemicals tested in the micronucleus assays, 27 were also tested in in vitro cytogenetics" (last assays," etc, paragraph on page 9) only to be followed by disclaimers? There are also many incorrect statements made in this section. Under the section on Tests for DNA damage and a air it is stated that "No noncarcinogens were tested in primary rat hepatocytes". This is not the case. Williams (Cancer Research 37:1845-1851, 1977) reported on tests with aflatoxin G2 N-4-fiuorenvlacetamide, anthracene and dimethylforma~i~5 in the primary rat hepatocyte/ONA repair system. There may be none according to the Gene-Tox carcinogen list but asI have indicated previously there are problems in using this list. On page 14 the statement is made that "The endpoint of ce~lular trans- formation assays is the development of foci of transformed cells, PAGENO="0124" 120 recognizable by visible alterations in growth pattern, against a background of normal cells'. This is one of the endpoints measured in transformation assays but not the endpoint. In the hamster embryo clonal assay, isolated colonies are evaluated and in the BHK-21 assay the endpoint is growth in soft agar. Although BHK-21 is listed as a transformation assay it is not in character with the other assays and there is a prevailing opinion that it may represent a mutagenic event. Use and Limitations In general the statements made in these sections are correct but are not sufficiently detailed. On page 17 there is a statement about assessment of nonunanimous test results as a limitation of the methodology. This is a problem, but it must be considered within the context of why it happens. This is turn requires a discussion of the false negative and false positive results and how batteries need to be constructed in order to account for the fact that certain assays are not useful for testing chemicals in certain classes. For example, metals, hormones and chlorinated hydrocarbons do not induce mutations in Salmonella whereas other systems can be used to detect chemicals in these classes. The use of Salmonella in a test battery for such compounds would of course result in nonunanimous test results. Another example is the statement on page 19 about the difference in response (response to what?) between epithelial cells (rather than tissue) and fibroblasts. Such a situation occurs with in vitro transformation, that is, a definite identifiable morphological change occurs after treatment of fibroblasts with carciongens but this does not occur with epithelial cells in culture. I think one of the major problems with this document is the spewing of information by rote without serious integration of the information at hand and how at this point in time short-term tests can be used most valuably. The overall tone is somewhat negative and there is a need for careful revision of the document. PAGENO="0125" 121 Canrrents on Potential Huren Carcinogens: Methods For Identification Pnd Characterization; Part I: Current Views. A Discussion Draft Prepared by the Regulatory Work Group on Science and Technology, Office of Science and Technology Policy, Executive Office of the President, October 1, 1982. I. CE%PI'ER I: T%~ PREAMBLE Ccxtparod to the docusents referenced 1-7, the Preamble tends to downplay the problem of environrrental carcinogenesis. A rore rreaningful statertent of the statistics relevant to cancer than that on pages 1 and 2 of the draft veuld he to say that cancer not only "often has a lengthy and progressive rourse" (p.1) but that it is fatal in rrnst cases. In 1981, 421,000 ~rrericans died of cancer; 58 million Arnaricans now living (25% of the population) will eventually suffer from cancer. 8 There is substantial agreeias~t that the majority of cancers are attributable to environrrental factors9 and by irrplication are preventable. (See p.2). The Preamble also fails to rrention the progress cede to date in iden- tifying husen carcinogens and cotential carcinogens. For exarrple, of 152 carpor~ds or industrial processes evaluated by LEPC, 30 have been identified as himnan carcinogens; and stout a~ dozen sore as having airrost sufficient data to be tenrad carcinogenic for hurkans. 10 For the "huron carcinogens" for PAGENO="0126" 122 which adequate testing data exist, virtually all have teen positive in animal bio- assays for cancer.ll (Nor is this infonnation included in any of the subsequant chapters.) The Preamble should have discussed the ninny reports of national and international panels during the last decades (for a coirplete list, see TSSC, p. 126-7) and the fact that they eirphasize the need for a conservative approach in light of the uncertainties surrounding available data and methodology.5 In particular, the U.S. Congress has stressed the need for a preventive strategy throogh controlling hirren exposure to carcinogens. There is a confusing and sorrevihat misleading discussion of the philosophy behind different environrrental/health statutes arid of recent court decisions. One questions whether this topic belongs in the Preamble to Part I, which is meant to be the scientific basis for cancer policy. In any case, the discussion of the Delaney Arren&nent and the subsequent statement on page 3 that "prevailing opinion" has shifted from the view that any exposure to carcinogens entails some level of risk and that all risk should be eliminated" is not accurate. First, it must be errphasized that despite various efforts to repeal it -- amid much debate - the Delaney Amend- sent is still in effect today, reflecting the view that it is prudent not to inten- ticoally add new chemicals Imown or suspected of causing cancer to the food supply. By contrast, mast other statutes pertain largely to substances and agents already in - casrercial use and frecru~ntly pervasive in the environment. In these cases, some bal- ancing of ithe need for health protection and the costs involved in regulation is required. Three statutes, for exairple, provide for regulation of "unreasonable risks." These are TSCA, FIFRA, and CPSA; but even these statutes do not require a precise risk/benefit assessment prior to regulation. (See, for example, the House Report on TSCA, p.14.) The result of regulation under the various statues is rarely a ban -- but a permissible level of exposure that is associated with the PAGENO="0127" 123 lowast level of risk achievable given econanic and technological considerations. This requirement for balancing does not in any way, howaver, invalidate the concept that increased exposure to carcinoqens conveys an additional risk of cancer. Bather, it is based on this assumption, but recognizes that other factors (cost and feasibility) must be considered in setting standards. The statement (on p.4) that the Suprerre Court decision on benzene "points out the need for scientific judgernents which are sound and objective" misleadingly implies that prior decisions in general and OSHA s decision to regulate benzene specifically ware not sound and objective. In fact, the Suprerre Court required OSHA to prove only that its best estimate indicates that exposure to the agent will result in a significant risi~ ~32, 13 Very importantly, the Preartele states that there have bean "significant advances in science which should affect the manner in which the regulatory bodies deal with suspected carcinogens." (p.4) Although this cryptic remark is echoed in the last chapter on risk assessment, it is unclear from tha intervening chapters what those "significant advances' are. Certainly, tbey are never discussed as such -- except perhaps as areas of research with possible future application to regulatory policy. Sore specific points and questions with reference to pages of the Preamble: 1) Page 7, Paragraph 3: - What "first chapter" is being referred to that discusses trends in cancer incidence and rrortality and evidence for a "cancer epidemic due to the growth of the - chemical industry in this country"? This chapter/discussion appears to have been dropped. In fact, it should be mentioned that while few are predicting an epidemic, there is concern about possible future increases in cancer rates due to recent increases in chemical production and use.14 2) Page 8, Paraqranh 2: Regarding use of the tens "genotoxicitv," the point should b~' made that there is no sirmnle operational definition of the tens or clearcut distinction between "gerotoxic" PAGENO="0128" 124 and "epigenetic" or "non-genotoxic" carcinogens. (See discussion in VII). The fact that cost available in vitro tests detect only those substances that affect the DNA does not mean that other types of carcinogens are not of iirportance. 3) Page 9, Paragraph 1: The 5th chapter discusses long-term animal bioassays. The fact should be lien- tioned here that all major federal and international policies have relied on results from these bioassays as reasonable qualitative predictors of effects in humans because of the alirost perfect correlation between positive epidamiological data and animal testing results for particular chemicals for which both types of data exist. Thus, the IARC has determined that "in tie absence of adequate data in humans it is reason- able, for practical purposes, to regard chemicals for which there is sufficient evi- dence of carcinogenicity in animals as if they presented a carcinogenic risk for humans.15 This conclusion is sgeported by the fact that, empirically, tie responses of humans and test animals have been generally similar)-6 4) Page 9, Paragraphs 2 and 3: Chapters 6 end 7 (not 5 and 6) discuss excosure end risk assessment respectively. Also, there are puzzling allusions to Part II. 5) Page 9, Paragraph 2: It should be acimowledged here that, despite progress iii the field of exposure assessment, there remain serious limitations in our ability to estimate husen exposure to carcinog~~nic substances. 6) Page 10, Paragraph 2: The last sentence stating that by first agreeing to this set of principles, the agencies will be able to generate policies which will ". . .provide reasonable assur- ances that human health will be protected from usreasonable risks posed by carcino- gens" requires revision in light of the discussion above about the specificity of the term to three statutes. PAGENO="0129" 125 II. CFThPIER II: CURRENT VIEWS ON TIlE MECIThNISMS OF CARCINOGENESIS This chapter contains a detailed account of recent research findings and major areas of research relating to chemical carcinogenesis and well illustrates the gaps in knowledge. As the author states: "What we do not know, hit need to know is equally as irrportant a qeestion as what we already know and are currently learning about." (p.35). The relevance and application of much of the new and developing information on the irechanisms of carcinogenesis to risk assessrrent and regulatory decision-making is unclear. In particular, there is a growing body of data that suggests that it is not even possible to make a clear-cut distinction between initiating and late-stage carcinogens on the basis of whether they affect the DNP~ or not. (See discussion of Chapter 7.) In this regard, the statanent on page 31, that "[blasic scientific data on the rrechanisrns of carcinogenesis and turror indirotion are continuously being used by regulatory agencies in the decision making process" requires clarification. In fact, current policies do not require knowledge of the rrachanism of action or state how to factor this information in to risk assessment. Considerable debate about this issus is found in the record accurrsilated during the developrrent of the OSHA cancer policy (1980) ~17 OSHA's conclusion was that "it would not be practical or justifiable to establish different criteria for the identification, classification, or regulation of initiating and praroting agents. OSHA agreed with the NCRE that - "any factor or combination of actors which increases the risk of cancer in hunrans is of concern regardless of its mechanism of action." (p.5152). Similarly, it is not established (nor is it clear fran the discussion) how one can relate diet, age, stress, sex, horrronal status to the outcorro of safety evaluations of specific chanicals. 22-143 O-83---9 PAGENO="0130" 126 1) Page 3, Paragraph 3: The statement that "[pJ hase 1 netabolites are somewhat sore water soluble than the substrate and thus sore easily transported through cells," nends qualification. In general, the rate of passage across the cell membrane is proportional to lipid solubility. For cost rretabDlites, there is no special transport system. 2) Page 4, Top: Regarding the conjugation of carcinogens with glutathione, sulfates, and glucuronidas, there are cases in which conjugates may be deconjugatad, making available the active fone of the carcinogen for macrmrolecular binding. 3) Page 5, Top: This sentence states that chemicals are detoxified by binding to target cellular rracrcziolecules, which is incorrect. The further statement that "if detoxification by glutathione is extremely efficient, rrmcrasolecular binding may not occur" slould be revised to "macronolecular binding will be reduced." 4) Page 8, Top: The reference (20) to a paper by yeber is incorrect. 5) Page 10, Bottczn: It wauld be better to say that the covalent interaction of electrophiles with nucleophilic centers within calls is believed to be a sigaificant initial event. 6) Page 11, Par~graoh 1: I believe Pitot and Heidelicerger, Cancer Res., 1963, should he cited with reference 28. 7) Pare 23, Paraqraoh 1: The point should be made that cvtotoxicity (e.g., as manifest in induction of hyperolasia) is generally viewed -- nor as a preremisita for carc'4 nogenicitv - but as a contributing factor. ~ PAGENO="0131" 127 8) Page 26, Top: A note of caution about the possible use of natural or synthetic inhibitors of carcinogenesis as a supplemental strategy with possible merit in the future. It is inportant to put recent research studies on inhibition of carcinogenesis-- by substances such as retinoids - in perspective. First, these are largely at the research stage and far from providing a means of cancer prophylaxis for tie general population. Second, in contrast to certain minor codifications of diet, administration of pro- phylactic doses of synthetic or manmade chemical inhibitors ~ould not be witteut danger of toxic effects. For this reason, the cost appropriate future application of this approach may be in cases of high accidental exposure to a known carcinogen. This discussion is not to irtply that current research is not of considerable interest but to underline the limitations of this approach. 9) Page 29, Top: It is not that neoplaxns exhibit high rates of growth, but that cells dn not stop dividing when they should. 10) Page 29, Paragraph 1: According to Nagasawa and little, they have confinned the results of Kinsella and Radman "indicating that exposure to tumor promoters alone can induce sister chrcxnatid exchange (SCE), and that this induction is suppressed by protease inhibitors." (p. 1946) ~l9 Furthermore, Birnboim has shown that the promoter P~ induces extensive DNA strand breakage in human peripheral lynphocytes (possibly by generating oxygen radicals) as dces another potent tunor promoter, benzoyl peroxide, by direct action.2░ According to the author, a combination of mutational damage by the initiator and DNA strand break damage by the prorroter may be necessary for the developeent of cancer. These and other studies put in question the statement that the effects of TPA on a cell are reversible. PAGENO="0132" 128 11) Pages 31-32: Under the "Pole of Science in Regulation," there should be a brief discussion of the substantial knowledge aucrued on the testing of potential carcinogens using bioassays and in vitro methods. Research goals 1-4 would he more appropriate onder "flierging Areas of Science. 12) Pages 32-34: This section on errerging areas of research should include short-tern tests for pronotion; research on factors controlling cellular differentiation; and tests for chrarosanal rearrangements or oncogene activation. 13) Page 35: This section, entitled "Research Areas Peqtiring Pophasis," illustrates the many gaps in knowledge concerning the rrechanians of carcinogenicity including: which factors rrodulate the neoplastic process? What are the precise iretabolic patts'ays for activation and detoxification? W'nat irethods can be used to detect and rronitor damaged DNA? What are the rralecular events resulting from exposure to multiple or conbined chemical agents? What physiolocic factors influence neoplasia? etc. III. CHAPTER III: WRRERT VIEWS ON ERID~flOLOGICAL ERTEODS This chapter is an exceptionally brief overview of epidemiology. There are several irrportant omissions. One is the statistical power of studies to detect causal associations. Gensrally, epiderniological studies cannot detect lower levels of risk (e.g., en OP. or P.R of less ti~-~ 1.5). In particular, very large niznters of people must be studied in order to identify a causal relationship when the effect size is ernall and the backgroond incidence is large; therefore, because of practical limits on the size of study populations - not to rrention difficulties in exposure classification - it is especially difficult to detect PAGENO="0133" 129 increases in corrrron types of cancer (e.g., cancer of tim lung, colon, rectus or breast). To illustrate this point, the authors should provide exarrgles of the statistical power of cohort or case-control studies (of varying n ) to detect lung or rectal cancer. The point might also be made that Ia~cause of difficulties such as the identification of appropriate study populations, epidemiological data do not exist - nox are they likely to be developed for the great majority of chemicals that have been found to cause cancer in enimals.2~- Thus a mix of animal tests, in vitro rrethods, and epidemiology is necessary to support regulatory decisionmaking. The usefulness of epidemiology (and its major advantage of allowing study of a real life situation) should be enhanced, for purposes of carcinogen identification and risk assessment, by a combination of laberatory rrethods and epidemiology. Several other consideratious that deserve to be included in this cheoter are: the irrportant problems of misclassification of individuals regarding prior exposure that arises from the inadequacy of ~r'osure data; and the healthy verker effect leading to difficulty in relating :~ilts of occupational studies to the general population. Finally, regarding the discussion of costs of different types of studies: the cost depends on the number of peoplO studied. In reality, the costs of case- control and cohort studies are not very different. Also, on page 5, the case-control - study gives one the ability to study rare cancers, not rare exposures. PAGENO="0134" 130 IV. CHAPIER IV: CUPPEPI VIEWS ON SHOEW-FDPZ1 ERSTING* This chapter presents a mass of information drawn largely from the EPA Gene-Tox Program (1979), which carniled extensive validation data on individual assays as ameans of screening of chemicals for potential carcinogenicity. As the Background section acknowledges, the Gene-Tox data base stops at 1979 and omits the International Collaborative Program. Therefore, the chapter should be revised and ppdated to include recent studies. ~bre irrportantly, the real issue of interest in this chapter is not how individual tests perform, since a significant nuiiber have teen validated, but how they can be used in carcinogen identification and risk assessment. There is little or no discussion of the major advantages of in vitro tests in this regard, i.e., that they permit tho use of different systems in corrbination to measure a range of endpoints. Unfortunately, the chapter and accorroanying Table are set ~ro to ireasure each individual short-term test against a standard of: can it detect all carcinogens? and can it provide metabolic activation? and advantages and dis- advantages are framed accordingly. A mare usefu~ approach wnuld be to lay out I what the tests can do in corrbination and what thc perceived future application is in: 1) rapid screening for potential carcinogenicity; and 2) risk assessment (e.g., in inter-species corrnarisons, evaluation of hisrian esuosure, ireasures of early biologic response, etc.) ~2la In a nrraber of cases, generic problems in testing are treated as if thay were unique to in vitro tests or specific to a particular assay; and characteristics which could - be viewed as advantages are presented as limitations. For exasple, generic problems such as the need for metabolic activation, toxicity of S-9, genetic drift etc. are described as limitations of specific tests. The need for metabolic activation is presented as a limitation of many in vitro tests, although the ability to corrpare results in systems using different metabolic activation methods in a controlled setting might as well be viewnd as a major advantage of these te: s. Particularly *Follo~~g rgy request for review of this chaDter, Dr. Virginia Dunkel, Chief, Genetic Toxicology Branch, Bureau of Foods, FDA, prepared detailed corrments on Chapter IV, appended. PAGENO="0135" 131 when one considers that individual in vivo tests generally provide information only about rretabolic activation and effects in one type of tissus in a particular species. The state of the art in testing for `non-genotoxic" effects is not presented despite the considerable interest in current and developing tests for prorroters and codifiers. Nor is there a discussion of various published rrethods for statistical rnetbods for analyzing shert-term test data. 22,23 Rather, the stateirent is made that "there is no generally accepted statistical rrethodology available for evaluating short-term tests for genotoxicity." (p.26). The following specific carrrents may clarify the core general discussion above: 1) Page 2, Paragraph 1: The discussion on mechanisms of carcincgenesis states that non-genotoxic carcinogens such as horremnes, metals, etc. are believed to act by rrechanisrns other than interaction with DNA or other cellular macro-molecules. This conclusion is premature. For example, chrcrnate is a rsutagen, as are various selenitsn derivatives.24 2) Page 5, Paragraph 3: Regarding the correlation between carcinogenicity and mutagenicity as detected by the Z~rrres assay, Aries and McCann have reported that, if results for chemicals tested are coithined, the correlation is about 83%. 2o 3) Page 9, Paragraph 1: Regarding the statement about "problems of toxicity, especially with S-9 systems and in vitro huasn lymphocyte cultures," altimugh huasn lymphocytes are more sensitive, there is a generic problem with S-9 toxicity in any rrairmalian cell essay, which can be controlled by proner dose selection. 4) Page 11, Paragraph 2: The ability of tests for DNA dsmage and repair to distingmish between error- free and error-prone repair is presented as necessary if they are to be used for carcinogen identification. This reflects a misunderstanding on the part of the PAGENO="0136" 132 author(s). Such a distinction ~vuld be necessary only if it were known that mutations arise only from error-prone repair. In fact, there is little if any evidence that error-prone repair is a factor in marmralian cells. Altinugh the term "error-prone repair" has been used to refer to a series of inducible ("SOS') functions observed in E. Coil, these presusably arise from a sacrifice in fidelity of DNA synthesis to preserve viability of the organism and are not believed to be trus repair systems. The paint appears to have been missed that the purpose of theme tests such as 13DB is to detect damage to the DNA - damage which can never be assumed to be corrgletely repaired prior to the cells' replication. The mistake is not necessarily in the excision process itself; rather, the occurrence of excision is a red flag that DNA damage has taken place.26 5) Page 17, Paragraph 3: The discussion of inter- and intra-laicoratory variation in enzymatic activity of S-9 and effects from length and tenmerature of storage, and test chemical solvent are not unique to S-9 or to the Ames assay, but can be viesed as generic problems. 6) Page 19, Paragraph 3: Before presenting the data from the de Serres study, the author(s) should have painted out that each laboratory was using its own protocol so that agreement ~ould not have been expected to be as close as if a single prescribed assay had been used. The study should be referenced.27 7) Page 20: The discussion about genetic drift raises concern without giving the estent of the problem or providing references. If in fact it is a serious problem, a remedy ~ould be to obtain strains and cell lines from the principal investigator or a single source. Hovever, an international collaborative study has shown that genetic drift has not been a problem regarding Ames tester strains. 28 ~s regards PAGENO="0137" 133 genetic drift in certain tissue culture lines, rrore information should be supplied. 8) Table: There are a number of inaccuracies in the table. For example: Page 2: The E-Coli WP2 assay probably detects frarre shift as vell as base-pair rrrrtagens; availability of strains is not currently a problem. The fact that assays in lovereucaryotes possess irore ability to activate prcmutagens/procarcinogens is cited as an advantage. However, this is only true if the system is relevant to humans. V. CHAPTER. V: CURPEST VIEWS ON LONG-TERM BIOASSAYS This chapter is framed as a series of issues regarding the conduct and interpretation of animal bioassays. It ci{es many "unresolved questions" without including major references in which these are discussed in depth and to various degrees resolved. 29-33 Therefore, one is left with the overall impression that there are overwhelming problems with bioassays. This irrpression is created largely because the chapter fails to mention the errpirical data showing excellent quali- tative agreement for human carcinogens which have been tasted adequately between results in animal bioassays and epidemiology or the fact that in the case of many carcinogens, bioassays provided the first evidence of carcinogenicity (e.g., BCME, vinyl chloride, DES, etc.). Also ignored are the observations based on existing data from rodent bioassays that mast chemicals have been shown to cause ca~icer in more than one s~cies,29'33 and the substantial agreement on interpreta- tion of carcinogenicity data accrued over the past decades. 29-33 For example, the absolute statement is made that there are "no universally agreed upon ways of interpreting carcinogenicity data." (p. 12). A mare correct statement would be that there is substantial agreement on the basic principles and criteria for date interpretation around which individual sc~ientific judgerrents are to be made; and a long history of animal testing provides a basis for resolving nany of the questions that arise. PAGENO="0138" 134 In addition, this chanter reouires substantial revision regarding the technical discussion of testing protocol, good lahnratorv practices, interpretation of data, and statistical analysis of long-term bioassays for carcinogenicity. Specific problan areas include the revie~i of dose levels, duration of experirrents, signifi- cance of cytotoxicity, probability of false positives, significance and grouping of ttrrors, and historical controls. Furthercore, a nrxrtar of the references given are incorrpletely cited. In particular, the discussions of high dose levels, false positives, and sicmificsnce of mouse liver tutors are flawed. On Page 9, it is concluded that there cay be possible misidentification of a substance as a carcinogen as a result of the use of the high dose levels necessary to provide rnaxirrrun sensitivity. Much has been written on this subject and, as mentioned earler, there is no convincing evidence that cytotoxicity alone (or any other effects observed at high dose levels) either do not occur at later dose levels or are either a necessary or sufficient condition for development of tutors. For exarrele, according to a recent report by a federal panel, a nuirter of agents are reported to induce epithelial hyperplasia but have no carcinogenic or tutor praroting activity; the panel found no evidence that the induction of epithelial hyperplasia is a sufficient condition for the industion of cancer, although it may contribute to sate extent to eresression of carcinogenic activity.18 The work of Barrcws and Shank is mentioned on Page 9 in the context of the argu- ment that cancer, which otherwise would not occur, can be produced or prorroted when normal physiology, iroceostasis, detoxification and repair rrechaniarns are over- wheLmed. The observation of intirect rrethvlation of DN~34, 35 is presented as sucrgestive supporting evidence. However, as shown in Figure 1 of reference 35 these studies do not dearly indicate a biphasic response (or thr~shold) because of the limited nuhrrer of data points at tire low dose end the general scatter of the data. PAGENO="0139" 135 The discussion on Page 10 relating to the significant probability of false positive results should be revised in light of recent articles indicating that the false positive rate is far lover than that described. 36,37 Furthernore, the relatively rrore serious problem of false negatives, resulting from the inability of these tests, in most cases, reliably to dabect an increase of less than 15%, has not been presented. This is a critical omission, since it is for this reason that a negative result in a bioassay cannot establish that the agent is not carcinogenic. Regarding the significance of rouse liver turors - dismissed as "controver- sial on the basis of a syrrposium held by the Society of Toxicologic Pathologists (1982), it is notroorthy that the majority of chemicals that cause rouse liver tutors also produce tumors at other sites in rodents. 38 Tho additional points: regarding Page 23, in addition to mutagenicity, DNA repair end neoplastic transformation, evidence of covalent binding to DNA, and chrorrosomal aberration can be used in aiding judgerrents on carcinogens. Second, this chapter should include a discussion of short-term in vivo animal tests. VI. CHAPTER VI: CURRENT VIEWS ON EXPOSURE ASSESSMENT As with all previous chapters, it is not clear what new methods and approaches can now be used routinely by federal agencies in assessing human carcinogenic risk; that is, which are operational and which are still in ti-er research stage. The - chapter provides en overly general review of "items to be considered in developing en exposure assessrrent without addressing key relevant and practical questions such as: How easily are the exposure data obtained (i.e., what are the data access and collection problems)? How do/can agencies estimate the degree of human exposure (numbers of individuals, levels, duration, etc.) to new as vell as to existing chemicals? How can multiple exposures be evaluated? Very importantly, the chapter does not adequately stress the limitations of PAGENO="0140" 136 the monitoring data themselves and the general cnideness of cost available modelling systems. Sons mention should be cede of the danger of false security from precise mathematical estimates of exoosure which are in fact frecmantly based on very limited monitoring data. In this omntext, more discussion of the potential usefulness of biochemical rrethods to assess biologically relevant exposure to chemical carcinogens would be appropriate in this chapter. VII. C}S~PTER VII: CUP~T V~S ON ?SSESSRI~T OF C~RCINOGE2~IC RISK This chapter as written bears little or no relationship to those preceding it. Instead of providing a review of currently available methods for risk asseserent and the major issues involved, the chanter offers a sketchy, and in several places inaccurate, discussion of risk assesenent cathode followed by a conclusory statement that consideration is being given to the safety factor (vs. risk extrapolation) approach to selected carcinogens. (p.8). No further clarification of this state- ment is provided, except for the subsecuent remark that "biological factors such as rnutagenic activity associated with chemicals might help to select anong the various ways of ggantifying risk." (p. 9). These sentences irrnly that a significant departure from presently practiced and accepted carcinogen risk assessment policy is being considered. As such, they are not appropriate to Part I -- the purpose of which is to provide the scientific basis for carcinogen identification and risk assesetant. - I~re inportantly, there are serious qrastions about the scientific basis for this suggested approach as will be discussed below. A general recorrnendatiori is that the chapter be revised to centain a review of the biological and statistical basis and assurotions underlying: 1) the various available risk extrapolation models; 2) the safety factor or no observable effect level (NOEL) approach hitherto used only for non-carcinogenic effects. In addition, the relative advantages and disadvantages of theme two approaches -- as well as of the individual models -- should be evaluated. Past and present experience with the PAGENO="0141" 137 various nodels to both husan and epidemiological data should be presented. The chapter should emphasize the recognized need to present a range of estimates (incluting an upper cunfidence level) and a clear statmrent of the uncertainty surrounding the estiirates. Specifically, the statecent on Page 7 that "tha choice of the nodel is limited only by the imagination" is misleading, in that it implies that there is little or no basis for the various mathematic nodels used for risk extrapolation. It s~ould be more accurate to say that in mast cases, enpirical data are not available with which to determine the nature of the low dose-response relationship; thus many rrodels fit the data equally well. As stated in tte California Draft Cancer Policy, Section 2, because erroirical data are not sufficient to specify the shape of the dose-response relationship for any specific carcinogen, the choice of the mast appropriate nodel is a matter of scientific judgement. This judgerrent must be made bearing in mind the present inability to determine population thresholds for an individual carcinogen; the substantial background incidence of cancer; and the multiplicity of husan exposures to carcinogenic agents and substances. As the authors point out, unless there is 1) instantaneous and orrrrplete (100%) deactivation of a carcinogen at low doses, 2) complete (100%) repair of initial damage, 3) a population with uniform sensitivity, 4) and, unless the carcinogen being evaluated is shown to be operating by a machanism different from all others to which the individual is exposed, there can be no assumption of a threshold. The discussion on Pages 10-20 of tier docmrent clearly sets forth the reasons for this conclusion and its corollary: that "low-dose linearity applies equally well to agents which are thought to act by either genetic or epi- yenetic nerchanians if one makes the reasonable assumption that similar rrechanions are already operating and contributing to the background incidenca of cancer."7 PAGENO="0142" 138 As mentioned earlier, the multi-stage model as modified by Crunp, deserves special consideration by virtus of its extensive use by federal agencies, including EPA' s Carcinogen AssesaTent Grog (CAG), CPSC, etc. This model is apparently referred to in Chapter VII as a "linear' model (pp. 8-9). In point of fact, this model does not assune a one-hit, linear biological scenario of carcinogenicity, but is modified to incorporate linearity at low dose in order to account for background additivity. In this regard, it is not unique among models. According to Crouch and Wilson, [il f the background cancers are similar to tha cancers produced by the chemical, many authors have shown that most models, including that of Mantel and Bryan, becorre linear at low doses."39 (p.107). The authors of the California Draft Policy have concluded that dose additivity is an appropriate assizaption for all carcinogens: The current Iciarledge of the rrechenisms of carcinogenic action provides little guidance as to the appropriate choice of dose-response models. Although the assunption of low-dose linearity is most generally accepted for carcinogens that interact directly with DNA. . . dose additivity (with the assunptions discussed previously) will lead to low-dose linearity for car- cinogens that act at any stage. 7 (Section 2, p.20). As already mentioned, probably the most serious problem with this chapter is the absence of discussion of how various theories regarding the mechaniem of carcinogenesis can be related to the various models and approaches. Instead, there is an allusion to the need to use "current biological thinking in developing proedures for the estimation of risk at low doses" (the statement is accorrpanied by an unpublished reference) end to consideration of biological factors (e.g., mutagenic activity) to help select among ways of quantifying risk. (p. 9). Most surprising in light of the absence of supporting discussion in the preceding chapters, cares the statement that consideration is being given to "using margins of safety placed on information obtained in the observed region o~ the dose-response curve.. . in selected instances, alone or in corbination with extrapolation models." (p. 8). There appears to be inadequate scientific justification for this approach. PAGENO="0143" 139 Instead, the statement found in the 1982 California docurrent, that "the differences between agents that act at different stages are of great scientific interest, but the precise mechanisms of action are still poorly understood," is an accurate one and was errphasized in discussion during a recent EPA-sponsored conference on the application of Biological Markers to Carcinogen Testing, November 15-19, 1982, in Washington, D.C. Although there have been several proposals to dichotomize chemical carcino- gens as "genotoxic" and "epigenetic," these proposals are largely based on theoreti- cal assumptions in the absence of a solid body of data regarding low dose-response relationships and a clear understanding of the critical rrechanimns by which even mutagenic carcinogens actually cause cancer. In addition, the definition of the tents `~enotoxic" and "epigenetic" poses problems, in that recent studies have tended to blur the distinction between the tao terms. For example, Weisburger end William's have defined "genotoxic" chemicals as those "which appear to exact their effects by interacting directly, or after conversion to an ultimate carcinogenic form, with DNA."4░ They include in this category direct-acting electrophilic agents, procar- cinogens, and inorganic carcinogens that are not directly toxic but affect DNA replication. Therefore, in their view, evidence for genotoxicity can be provided by in vitro tests for nsrtagenicity, DNA damage and repair, chranoscsnal effects, and cell transformation. By contrast, the autimrs define "epigenetic" carcinogens as - those which exert their effects throrph incompletely known rrechanisms and which "presumably are incapable of causing conversion of normal cell to a neoplastic one, but permit the expression of pre-existing latent neoplastic cells."41 They cite as examples of epigenetic carcinogens solid-state carcinogens, hormones, inrnunosuppressors, co-carcinogens, and prorrotors. Recent studies illustrate problems with these definitions and exarrrles. As mantiorred earlier, the phorbol esters are generally viewed as classic exannies of cocarcinorrens and PAGENO="0144" 140 prorrotors. Hosever, Birnboim has sIDwn that the pratotor, phorbol myristate acetate induces extensive DNA strand breakage in hisnan peripheral lynphocytes (possibly by generating oxygen radicals) as does another potent tuitor promoter, benzoyl peroxide, by direct action.42 According to the author, a combination of rants- tional damage by tho initiator and DNA strand break damage by the prorroter may be necessary for the developnrent of cancer. Nagasawa and Little have canfirired earlier reports that exposure to the tizror praroter TPA alone can induce sister chromatid exchange, suggesting that one nrechanimn for the promotion of carcinogenesis induced by physical and chemical agents may involve the facilitation of espression of mutational damage in cells by mitotic segregation.'43 These are not isolated reports.44 Similarly, DOT is generally considered to be a "late stage" carcinogen. Yet a number of studies have demonstrated that DOT and its derivatives induce st~tistically significant chroomsomal abnormalities and cytogenic effects in rodents.4547 as well as chramatid aberrations in lyrmpbocytes of individuals occugationally excosed to DDT.48 The distinction is further blurred by evidence that various otirer agents that do not interact directly with ERA may lead indirectly to misrrethylation of DNA bases. 34,35 Recently, ~A circulated a draft document entitled "Additional U.S. Environ- irental Protection Agency Guidance for the Health Assesonent of Suspect Carcinogens With Specific Reference to Water Quality Criteria" which proposed to distinguish between `~enotoxic" and "non-genotoxic" carcinogens for purposes of risk assessment and standard setting. Adopting a different, far narrower definition of "genotoxic," the draft ecuated the term with "mutagenic" and defined adecnnate evidence of muta- genicity as positive results in two different in vitro systems for point mutation, ob one positive in vivo assay' for mutagenicity. According to the proposal, the the multi-stage model modified to incorporate linearity at low dose laretofore used PAGENO="0145" 141 by CA3 for risk assesmient of all carcinogens wauld be used in risk assessrrent only for mutagenic CarCinogens, while the no observed effect level (NOEL) approach conventionally used for threshold-type toxic effects would he applied to non-mutaganic CarCinogens. This proposal was widely circulated and received roirrrent from over 20 scientists. Consistent criticisms were that this policy was based on a number of questionable asslrrptions including the following: 1) the distinction between genotoxic and non-genotoxic carcinogens is clear cut; 2) the dose-effect curves and rrechanisms of action are necessarily different for the two classes of agents, and mutation is necessarily the cost significant mechanism in carcinogenesis; 3) a threshold or no effect level can be identified which is applicable to the total human population at risk. The consents to R.E. Albert, Chair-ran, Carcinogen Assessment Group from Dr. Arthur Uoton (September 27, 1982); Dr. Elizabeth Weisburger; Dr. 1.13. Weinstein, Dr. Sidney Weinhouse, Dr. Ron Hart, Dr. Umber-to Saffiotti, Dr. Norton Nelson, Dr. Frederica Perera and Dr. David Noel disduss some or all of these points; and their authors did not sunport the pronosed scheme. Several other reviowers were opposed for other reasons. These responses are discussed in a recent article in Science.49 In a relevant section entitled "lack of predictable thresholds for en exoosed population," in 1979 the IRLG policy emphasized that the self-replicating nature of cancer, the multiplicity of human seposnres, the additive and possibly synergistic combination of effects, and the wide range of individual susceptibility make it currently irroossible reliably to predict a "threshold below which human population erposure to a carcinocen has no effect on cancer risk."50 In particular, "varia- bility among individuals makes it very difficult to have confidence that an observed no-effect level of exoosure in animals or even in a specific human population (for which individual variation may be small in borrpanison to the total population) will 22-143 O-83---10 PAGENO="0146" 142 be applicable to the total hwran population at risk."51 Three years later, in its review, the California taparbrent of Health Services has concluded: It is not aporooriate to apolv the concept of thresholds to carcinogen- esis unless convincing evidence is presented to demonstrate the existence of thresholds for a smocific carcinogen in specified circumstances. 7 (Section 2, p.19) Several additional points tear mention: 1) There is no discussion of the interactive effects of chemical carcinogens and the fact that currently available models do not take account of potential synergism. 2) There is no elaboration on the ca~stion of how "knowledge of metabolism aral pharmacokinetics can inprove the quality of does inforrration in the risk assessment." (p.9). 3) The statement on Page 6, that risk assesmnent may be used "to estimate risks at given chemical exoxosures, at given risk levels (e.g., virtually safe dose)" requires clarification. The purpose of risk assessment is to estimate the magnitude of the risk that is posed to an exposed population by specified levels of exposure. "Virtually safe dose" is not a tens generally applied to carcinogens. Mantel end Bryan have specifically defined "virtually safe dose" to mean a specified small (10-6 - 10-8) increase in lifetime risk. PAGENO="0147" 143 REFERENCES mi~ Scientific Bases for Identification of Potential Carcinogens and Estimation of Risks. Federal Register, 44:39859-39879, 1979. 2 Ibid., p.39858. Health Risk and Econmaic Ingact Assesmnents of Suspected Carcinogens, Interim Procedures and Guidelines. Federal Register 41:21402-21405, 1976. National Emission Standards for Identifying, Assessing, and Regulating Airborne Substances Posing a Risk of Cancer. Federal Register 44:58642-58670, see p.58647. Toxic Substances Strategy Ccmnittee (IESC), Couscil on Environriental Quality: Toxic Chemicals and Public Protection: A R~port to the President. U.S. Govern- sent Printing Office, Washington, s.c., May 1980. 6 U.S. Congress, Office of Technology Assessnmnt COlA): Assessrrent of Technologies for Detennining Cancer Risks frau the Environrrent. U.S. Congress, Washington, D.C., 1981. State of California, Departrrent of Health Services: Carcinogen Identification Policy: A Statenent of Science as a Basis of Policy. July 1982. 8 Aserican Cancer Society: Cancer Facts and ~`igures, 1981. H.H. Hiatt; Watson, J. D.; and Winsten, J.A. (eds.): Origins of Human Cancer. Cold Spring Harbor Laboratory, New York, 1977. 10lnternational Agency for Research on Cancer (IARC): Working Group Draft, February 1982. ~IARC Mnnographs on the Evaluation of the Carcinogenic Risk of Chemicals to Man, Volumes 1-26, Lyon, France, 1972-1982. `2D I)nniger: Defeat in benzene exposure case no death knell for 0511k standards, National Law Journal, September 15, 1980, pp. 26-27. 13P.F. Stone: The significant risk requirement in 0511k regulation of carcinogens: Industrial Union Deparinrent, AFL-CIO v. Anrerican Petroleum Institute. Stanford Law Review 33:551-566, 1981. 14DL Davis, Bridlyjrd, K.; Schneidenran, M.: Cancer prevention: Assessing causes, exposures, and recent trends in rrortality for U.S. males 1968-1978. Terat. Carcin. and Mutag. 2:105-135, 1982. 15lnternational Agency for Research on Cance~ (IARC): Chemicals and Industrial Processes Associated with Cancer in Hnnrsns~ IAPC Manograph Suppl. 1, Lyon, France. 16E Croush and Wilson, R.: Intersoecies corrparison of carcinogc ic potency. J. Toxicol. and Environ. Health 5:1095-1118, 1979. PAGENO="0148" 144 17š~g~: Regulation Covering the Identification, Classification and Regulation of Potential Occ~ationa1 Carcinogens. Federal Register 43:5002-5296, 1980. of the Federal Panel on Formaldehyde. Environ. Health Persp. 43:139-168, 1982. 19H Nagasawa and Little, J.B.: Effect of turor pranoters, protease inhibitors, and repair processes on X-ray induced sister chrorratid exchanges in rrouse cells. P.N.A.S. 76:1943-1947, 1979. 20HC Birnboim: DNA strand breakage in hunan leukocytes exposed to a tutor prorroter, phortol myristate acetate. Science 215:1247-49, 1982. Karstadt: A survey of epidemiologic data on hunans exposed to animal car- cinogens. In Quantification of Occutational Cancer; Danbury Report No.9, R. Peto and Schneidennan, M. (eds.), PP. 223-245. Cold Spring Harbor Laboratory, New York, 1981. 2la~ ~Cann: Short-term tests and cancer policy. In J.C. ~bDrnald (ed.): Recent Advances in Occupational Health. Churchifl Livingstone, 1982, pp. 149-161. 22L Bernstein; Kaldor, J.; McCann, J. and Pike, M.: An empirical approach to the statistical analysis of rs.itagenesis data from the sa:Irronella test. Mut. Pea. 97:267-281, 1982. 23RD. Snee and J.D. Irr: Cesign of a statistical urethod for the analysis of mutagenesis at the hypoxanthine-guanine phosohoribosyl transferase of cultured Chinese hamster ovary cells. Nut. Ens. 85:77-93, 1981. 24Hollstein, M.; McCann, `~tr.; Angelosanto, F.; and Nichols, W.: Short-term tests for carcinogens and rnutagens. Nut. PIes. 65:133-226, 1979. 25B Aires and McCann, J.: Latter to the Editor. Cancer PIes. 41:4192-4203, 1981. 26Personal Corrrnunication: Dr. G. Teetor, New York University, Dr. J. Mccormick, and Dr. V. Maher, Michigan State, December 1982. 27F. deSerres and J. Ashby (eds.): Evaluation of Short Term Tests for Carcinogens. Elsevier-North Holland, New York, 1981. 28B.H. Margolin; Risko, K.J.; Shelby, M.D.: Analysis of variability in the inter- national collatorative study on genetic drift" in Arias tester strains. (Manuscript), December 1982. 29R.A. Grieserrer and Cuato, C.: Towards a classification schema for degrees of experirrental evidence for thefl carcinogenecity of chemicals for animals. In Malecular and Cellular Asoects of Screening Tests, P. Montesano, Bartsch, H., and Tornatis, L. (eds.): IAE, Lyon, France, 1980. 3ONARC: Long-Tern and Short-Term Screening Assays for Carcinogens: A Critical Appraisal. Suppl. 2, Lyon, France, 1980. 3~.C. Chu; Cueto, C.; and Ward, J.M.: Factors in the evaluation of 200 National Cancer Institute Carcinogen Bioassays. J. Toxicol. and Environ. Health 8:251-280, 1981. PAGENO="0149" 145 Sontag; Page, N.P.; and Saffiotti, U.: Guidelines for carcinogenic bioassay in esall rodents. NCI, Bethesda, 1976. 331.F.H. Purchase: Interspecies msrparison of carcinogenicity. Brit. J. Canner 41:454-468, 1980. 34L.R. Barrows and R.C. Shank: aberrant methylation of liver DN7~ in rats during hepatotoxicity. Toxicol. and 2~pp1. Phannacol. 60: 334-345, 1981. 35R.A. Becker; Barrows, L.R.; and Shank, R.C.: Methylation of liver DN~ guanine in hydrazine hepatotoxicity: Dose-response and kinetic characteristics of 7-rrethyl guanine and O6-rrethylguanine formation and persistence in rats. Carcinogenesis 2:1181-88, 1981. 36TR Fears; Tarone, R.E.; and Chu, K.C.: False-positive and false-negative rates for carcinogenicity screens. Canner Res. 37:1981-1945, 1977. Gart; Chu, K.C.; and Tarone, N.E.: Statistical issuas in interpretation of chronic bioassay tests for carcincgenicity. JNCI 62:957-974, 1979. ward; Weisburger, E.K.; and Grieserrer, R.A.: The rrouse liver tonor as an endpoint in carcinogenesis tests. Toxicol. and 2~ppl. Pharrnacol. 51:389-397, 1979. 39E.A.C. Crouch and Wilson, N.: Reply to canrents on the regulation of carcinogens. Risk Pnalysis 1:107-111, 1981. 40J.H. Weisburger end Williams, G.M.: Carcinoqen testing: Currant problems and new approaches. Science 214:401-407, 1981. 41Ibid 4~ii.C. Birnboim: DN~ strand breakage in hmran leukocytes e~sed to a tirror prarater, phorhal myristate acetate. Science, 215:1247-1249, 1982. Negasawa and Little, J.B.: Effect of ttrror prceoters, protease inhibitors, and repair processes on X-ray-induced sister chranatid exchanges in mouse cells. P.N.A.S. 76:1943-1947, 1979. Reerit end Cerutti1 P.: Turrrour pranot~r phorJroi-12-rrwristate-13-acetate induces chrnromatml damage via incnirect action. Nature 293:144-146, 1981. 45F. Kellv-Garbert and Legator, M.F.: Cytogenetic and rnutagenic effects of DDT and DDE in a Chinese hamster cell line. Mutat. Pes. 17:223-229, 1973. 46K.A. Palrrer, Green, S.; and Legator, M.S.: Cytogenetic effects of DOT end derivatives of DOT in a cultured rrerrrnalian cell line. Toxicol. and AppL Phanracol. 22:355-364, 1972. 47D. Mshr, and Mitenburger, H.G.: The effects of insecticides on Chinese hamster cell cultures. Mutat. Baa. 40:107-118, 1976. Pairallo; Becak, N.; IDe Alrreida, W.F.; Pigati, P.; Ungaro, M.T. Murata, T.; and Pereira, C.A.B.: Cytcxgenetic stody on individuals occupationally esposed to DOT. Nut. Res. 28:449-454, 1976. E.: ~A's high-risk carcinogen policy. Science 218:975-978, 1982. 50]~pjj3 p.39876.. 51 Ibid. PAGENO="0150" In what mac constitute a test case for a nesv federal cancer policy, the Formalde- hyde Institute, an association of formal- dehyde producers and users. has advo- cated that formaldehyde not be regulated by the federal government despite recent studies shosving that the substance causes tumors in animals and despite evidence that there is considerable hu- man exposure to formaldehyde. The in- stitute has argued that the animal data do not provide a sufficient basis to regard formaldehyde as a likely human carcino- gen and that federal regulatory agencies SCtENCE. VOL. 216. 18 IUNE 1982 should asvait the development of conclu- sive human (epidemiological) data before taking protective action. This position contradicts principles for assessing carcinogenic risk that have been ssidelv accepted by the scientific community for over a decade and em- bodied in policios of regulator.' agencies follooving deliheralions of broad-based scientific panels. These principles assert that confirmed positise animal data are presumptis e evidence of carcinogenicity in humans: that svilh current information and methods it is not possible to estab- lish threshold or no-effect levels that can be reliably applied to the human popula- tion; and that positive human epidemio- logical data are not necessary to con- clude that a chemical substance poses a significant human risk (1). In fact, feder- at agencies have regulated such sub- stances as pesticides, hair dyes, food additives, and industriat carcinogens (for example. 3-propiolactone and ethylelie- imine( in the workplace primarily on the basis of results in experimental animals (2). These principles are consistent with the accepted social policy that it is pref- erable to err on the side of caution in interpreting the available scientific data in order to avoid failure to regulate a serious health hazard. Thus. acceptance by federal agencies of the industry position regarding the risk posed by exposure to formaldehyde could overturn established procedures fo rasses smg end regulating cat'einogenic Ftedenec Pcteca is t `enict t~tf scientist it the Natasil Retcaceet Defence Ceaticil. New Yew 10168. cud .icuiuteuut eliniecl ptcfestcn. Ilusi-.icn ci Eo,itennuenual tte;s(th Sd inset. Celuuwbis Unisen- toy Scheel cf Pablic Hnelth. Cethenine Petite it sit he Natasul Rcteuuneet Defente Ccancil. Nest Yuutk 10(618. 146 Formaldehyde: A Question of Cancer Policy? Frederica Perera and Catherine Petito u8v9yur~ 820sl8.t20590l0o 0 Ccpynighu c 1982 └bAS 1285 PAGENO="0151" substances in general. During the last 8 months, Environmental Protection Agency (EPA) and Occupational Safety and Health Administration (OSHA) offi- cials have reversed prior staff recom- mendations to initiate regulatory action to limit human exposure to formalde- hyde (3-5). In February 1982, EPA de- clined to regard formaldehyde as a prior- ity candidate for regulation under the Toxic Substances Control Act (TSCA) on the basis that the animal data may not be relevant to humans, that there is an absence of positive human data, and that it has not been established that, at hu- man exposure levels, the risk of cancer is "probable and would be high" (6). Ac- cordtng to sources quoted in Inside EPA, not only does the agency's decision not to move quickly on formaldehyde reflect a "clear divergence from current federal poltey," but "EPA Deputy Administra- tor John Hernandez has made tentative plans to totally revamp the agency's can- cer policy." (4). Our purpose in this article is to reviesv in detatl the data on the carcinogenicity of formaldehyde in light of established guideltnes for assessment of carcinogen- tc substances to see the extent to svhich the recent federal agency decisions rep- resent a major policy change. Backgruund Formaldehyde (HCHO) is a versatile chemical used in the manufacture of such products as particle board, ply- wood, paper, home insulation, material polymers and resins, leather and agricul- tural products, permanent-press fabrics, preservatives, embalming fluids, drugs, and cosmetics. About 7 billion pounds of formaldehyde are produced each year, making tithe 26th largest s'olume chemi. cal in the United States (7). An estimated 1.4 million people are exposed to formal- dehyde in the workplace: I 1 million peo- pte may breathe vapors in the home released by construction and insulation materials: and virtually the entire popu. lation comes into contact svith the chemi- cal because of its ubiquitous presence in polluted air and in consumer products (8). Concentrations of more than 8 parts per million (9), 0,02 to 4.2 ppm (8), and 0.1 to 3.4 ppm (10) have been measured in the workplace, in mobile homes, and in U.S. houses insulated svith urea-form- aldehyde foam, respectively, while lev- els frequently range above 0.1 ppm in urban air (24-hour average) (8). The pres- Cut U.S. occupational standard for form. aldehyde is 3 ppm (time-weighted 8-hour average) (1/). Thus, in the fall of 1980 there was svtdespread recognition of the signifi- cance of an interim report from the Chemical Industry Institute of Toxicolo- gy (CuT) that formaldehyde seas carci- nogenic in rats (/2). According to the final report (/3), at the end of a 30-month period squamous cell nasal carcinomas were observed in 103 of 232 rats exposed by inhalation to 14.3 ppm of formalde- hyde, in 2 of 235 rats exposed to 5.6 ppm of formaldehyde, and in 2 of 225 mice exposed to 14.3 ppm of formaldehyde. No such nasal cancers were found in 236 rats exposed to 2 ppm of formaldehyde or in the control animals. Polypoid ade- nomas svere reported in all exposure groups and in one male control rat. By February 1981, sarious groups of experts had reviewed the CuT interim (7) data, including a federal panel convene d cit the request of the Consumer Product Safety Commission )CPSC( under the aegis of the National Toxicology Program (NTP( (/4) and the Environmental Cancer In- formation Unit of the Mount Sinai School of Medicine (/5). The NTP report stated that "Formaldehyde should be presumed to pose a risk of cancer to humans" in agreement svith the Mount Sinai conclusion that HCHO is a carcinogen in rats and, data sug- gest. i umice ate sposure levels comparable to those found in some home and scorE environ. ments. Those hndcugs indicate that ettectise controls should bb initiated to reduce ore limi- nate human exposure to HCHO (15. p. 91. Meanss'btle, experiments at Nesv York Unis-ersity (NYU) (/6, 17) showed that exposure of groups of tOO mate rats to formaldehyde and hydrogen chloride separately and combined, at average concentrations of 14 ppm and tO ppm. respectively, resulted in an excess of histologically confirmed nasal squamous cell carcinomas in rats exposed to HCHO alone and none in the controls or in the rats exposed to HCI alone. Corn' bined exposure to HCHO urns HCI pro- duced about the same number of histo- logically confirmed nasal squamous cell carcinomas as HCHO alone (/6), As in the CuT study, no grossly visible spon- taneous nasal tumors of this type had been observed in control rats at that laboratory over a period of many years (/7, 18), In the spring of 1981, on the basis of the CuT study and a review of available data on formaldehyde use and human exposure to the chemical (/9), EPA stuff drafted a Federal Register notice under (4(f) of TSCA designating formaldehyde as a priority chemical for regulatory as- sessment (20). The draft 4(f) notice stat- EPA has determined that there may be a reasonable basis to conclude that some expo- sures to formaldehyde present a significant risk of widespread harm to humans. There- fore the Agency is initiating action to investi- gate those exposures of greatest concern and determine sshether they lead to unreasonable risks [20]. The notice was not signed by the Admin- istrator of EPA. Rather, during the sum~ mer of 1981 EPA Deputy Administrator John Hernandez convened a series of unannounced meetings-termed "sci- ence courts-primarily attended by EPA and Formaldehyde Institute repre. sentatives in order to reviesv the scien- tific data on formaldehyde (21). A con- gressional subcommittee svas critical of this significant departure from the ac- cepted peer reviesv process (3, 22). On 4 3epinmbcr 1981, the Natural Resources Defense Council (NRDC) re- quested an explanation of EPA's failure to act on formaldehyde under ž4(f) of TSCA and notified the agency of -bs intention to seek judicial reviesv of that failure under ž20 of the act. On II Sep- tember 1981, a memorandum from Don Clay, Office Direclor of the EPA Office of Toxic Substances (OTS(, to John Tod- hunter, then Assistant Administrator Designate for Pesticides and Toxic Sub- stances, recommended against treating formaldehyde as a priority for assess- ment under ž4(f) of TSCA pending addi- tional epidemiological information (23). In parallel devetopmenls at OSHA, in July 1981 an OSHA official recommend. ed res-ersal of a priu)r decision to release a bulletin on formaldehyde jointly svith the National Institute for Occnpaiional Safety and Health (NIOSH) 1241. The NIOSH Curre',tt Intelligencu' Bui//eei,t had stated that formaldehyde should be SCIENCE. VOL. 216 147 Summary. Thin article dencribes recent events concerning the asnesnmenf and regulation of formaldehyde, and evaluates the scientific data pertaining 10 Ihe carcinogenicily of thin substance in the contexl of established cancer policies and guideltnes. The conclusion is that recent decisions by several federal agencies 10 defer action to limit human exposure to formaldehyde may be a "test case" for a new, less proteclive policy concerning the regulafion of carcinogenic subslances in general. 1286 PAGENO="0152" Ithas come to ourattention that EPA. OSHA. and possibly other federal regulatory agen- cies. may be planning nut to take immediate protective action on formaldehyde, in spite of substantial evidence for its carcinogenicity from animal bioassays. We are concerned about the possibility of such a departure from established public health policy. ttsvould con- flict svith the prevailing viesvs ofib escientific community and would set a precedent svhich could hamper fat are regulatory action on oth- Their is general agreement among esperts in chemical carcinogenesis that a substance which cause scarcer in significant numbers of experimental animals in svell-conducted as- says poses a presumptiv ecarcinogenic risk to some humans. even in the absence of confir- matory epidemiological data. ~Vhile negative human data can define the upper limit of risk to man, there is no recognized method as yet for establishing the existence of a threshold for a carcinogen in the human population. These principles. sihich are accepted through- out the world, have served for many years as 18 JUNE 1982 the basis for sound public health policy and regulatory action on carcinogens. To compare our views on this subject with those of our colleagues. sve have consulted several of the siorld's leading authorities on chemical carcinocenesis for their opinions. The repliessve have received from them thus far are unan imous in supporting the principle that definitive demonstration of carcinogenic- ily in siell-conducted animal bioassays suf- fices to provide evidence of presumptive car- cinogenicity for the human population [28]. A week later, the American Cancer Soci- ety issued a statement nrging regulatory agencies "to set appropriate standards to minimize occupational and public expo- sure to the chemical, ins industrial prod- ucts and applications" 129). On 10 February 1982, Todhunter, EPA Assistant Administrator for Pesticides and Toxic Substances, formally recom- mended against considering formalde- hyde as a priority candidate for regula- tion. Characterizing formaldehyde as a "potential animal carcinogen" he ob- served that concern about human carci- nogenicity should be "tempered" by the observations that quantitative and possibly qualitative results of esposure to formaldehyde appear to depend highly on exposure level, species, and route: that rats seem to be particularly sensitive to formaldehyde: and that long human experi- ence does not seem to indicate any pressing concerns - . - 181. By contrast, the CPSC voted on 22 February 1982 to ban urea-formaldehyde foam insulation (30). Canada and the states of Massachusetts and Connecticut had previously banned the use of urea- formaldehyde foam (31). Questions have been raised about the validity of the animal data (26). Hosves'- er, the CuT study (13) svas rigorously peer-reviessed and is considered to be valid (14, 15, 27, 32). The possibility of a viral respiratory infection confounding the data in the CuT study ss'as considered unlikely in the NTP and Mount Sinai reports (14, 15). Control animals had also shown signs of viral infection but did not devel- op tumors. Further, in some of the rats, nasal cancers had formed by the lime respiratory viral infection occurred (33). In the NYU study confirming the CuT findings, a sample of the animals was tested for the virus and found to be negative 134): in the CuT stud)' mice svere not affected by the virus yet the)' developed tumors. Although it is unlike- ly that the transient viral infection con- tributed to the carcinogenic response of formaldehyde, people exposed to the chemical may also experience viral in- fections of the upper respiratory tract (14, 15). Addressing the criticism that the CuT sludy is flasved because "ulcerative inflammatory lesions" svere present in nasal mucosa, CPSC staff scientists have written that pathologists u-ho examined the slides from the CuT stud)' did not observe such changes (35). The NYU studies provide confirma- tion of the CuT results in a different strain of rats (16, 17). According to Up- ton (17), the studies "provide decisive confirmation of the Chemical Industry Inslilule of Toxicology findings that formaldehyde induces squamous cell carcinoma in rats." Partly because the type of tumor seas not that associated svith bis(chloromethyl)elher (BCME), it is judged unlikely that the formation of BCME as a result of combination of formaldehyde and HCI was responsible for the excess squamous carcinomas in the nasal cavities (9, 14, 36). The fact that formaldehyde alone produced about the same number of tumors as svhen combined sviih HCI also argues against an etiologic role for BCME. The prior reports of negative results in three long-term inhalation studies of formaldehyde do not detract from the significance of the CuT and NYU stud- ies. According to the NTP report (14), all three had shortcomings in experimental protocols and execution (for example, high mortality, inadequate exposure, or deficient histopathology). Bioassays in which other routes of exposure svere -used have been similarly limited; howev- er, some give definite clues that formal- dehyde may be carcinogenic to a variety of target tissues and animal species (14). Formaldehyde Data in the Context of Established Cancer Policies A number of principles have been elaborated in reports svritten during the last 10 years by various scientific com- mittees concerning the assessment of human risk from environmental carcino- gens. Composed of scientists affiliated svith academic institutions, industry, and government, these committees svere broadly representative of the scientific community. Their reports included those of the National Cancer Ads'ivory Board, the Interagency Regulator)' Liaison Group (IRLG), the National Research Council (NRC) the Food Safet)' Coun- cil, the Office of Technology Assessment (1), and the Occupational Safety and Health Administration (OSHA) as svell as publications by the IARC and the 1287 148 handled as a potential occupational car- cinogen and Ihat appropriate controls should be imposed to reduce svorker exposure (9). OSHA also denied a 26 October 1981 petition b)' the United - Auto Workers (25) for an emergenc)' standard for formaldehyde. OSHA's ac- tion in pulling aside nesv standards on formaldehyde and ulher substances lng~ gered concern that ness- officials at OSHA ss'ere likely to revise the agency's cancer policy (5). On 17 August 981, Arthur Upton, Chairman of the NYU Medical Center Institute of Environmental Medicine, wrote to the heads of federal agencies that formaldehyde is "decisive])' carci- nogenic in animals" and "if the carcino- genicity of formaldehyde is ignored, it would mean that no agent could be re- garded as carcinogenic in the absence of positive evidence in humans" (17). This letter prompted a response from Joel Bender, Chairman of the Medical Com- mittee of the Formaldehyde Institute, that "to regatd formaldehyde as a likely carcinogen in man is not supportable" (26). B)' contrast, in October 1981 a working group of the International Agen- cy for Research on Cancer (IARC) con- cluded, on the basis of the CIIT and NYU studies, that formaldeh)'de gas is carcinogenic to rats and should be con- sidered "for practical purposes," in the absence of adequate data in humans, as if it represented a carcinogenic risk to man (27) On 29 January 1982, a detailed letter to the assistant secretary for OSHA, the EPA administrator, and the chairman of the CPSC from Upton of NYU and I. B. Weinstein of Columbia Unis'ersity rec- ommended prudent measures to restrict Validity of Ihe Dala exposure to formaldeh)'de: PAGENO="0153" New York Academy of Sciences (37-45). Developed cooperatively by representa- tives of all federal agencies concerned with toxic substances control, the IRLG gutdeltnes were based on an extensive review of the scientific literature on carcinogen assessment. EPA formally adopted (in 1979) the IRLG policy for purposes of evaluating evidence regard- ing suspect carcinogens as a supplement to its own Interim Guidelines for Carci- nogenic Risk Assessment (46, 47). In particular, the report of the Toxic Sub- stances Strategy Committee (TSSC) published tn 1980 is significant because it was based on a review of 23 major re- ports written between 1956 and 1979 (48). The TSSC, composed of represen- tatives from federal agencies with re- sponsibility for controlling toxic sub- stances, identified "principles and tech- nical considerations underlying federal policies for the identification of potential human carcinogens": Although they have been the.subject of con- siderable public misunderstanding, these prin- ciples are widely supported in the scientific community and in the deliberations of rule- making and adjudicatory bodies, the courts. expert committees, and international agencies 148, p. 1251. Even more recently, the Office of Tech- nology Assessment has reviewed and reaffirmed the basic principles of carci- nogenicily assessment (1). These princi- ples are as folloxvs. 1) Animal testing data from properly designed and well-conducted tests are adequate for concluding that a chemical substance is a likely carcinogen in ha- Over 30 chemicals or industrial pro- cesses are judged by the IARC to be carcinogenic or probably carcinogenic to humans on the basis of epidemiological evidence (49). Of those for svhich animal data exist, all (svith tsvo possible excep- tions) have been positive in experimental animals. The isvo possible exceplions are benzene and arsenic. However, there is evidence from two recent bioassays that benzene is carcinogenic in animals and that arsenic may be a cocarcinogen capa- ble of inhibiting DNA repair (50, 51). Hence, the IARC has concluded: In the absence of adequate Oaia on humans. it is reasonable, for practical purposes, to re- gard such chemicals for which there is suffi. cieni evidence of carcinogenicity in animals) as if they presented a carcinogenic risk for humans 52, p. 14). Thus: All Federal agencies accept a positive bioas. say result in a single species as evidence that the substance is a potential human carcinogen 1/, p. 13). t2nu 149 a) Animal testing at high dose levels is a valid and necessary procedure for iden- tifying potential human carcinogens. The exposure of experimental animals to tox- ic agents in high doses is a necessary and valid method of discovering possible carcinogenic hazards in man 39, p. 71. Since carcinogenic response is usually dose related, the biological and statistical sensitiv- ity of a bioassay may be enhanced by increas- ing the exposure levels of the test substances rather than the less feasible alternative of increasing the number of test animals to match the human population at risk 42, p. 5093). The basis for such extreme doses, most sim- ply stated, is to maximize the sensitivity of the text and its capability of detecting irre- versible molecular events leading to neoptas- tic transformations of cells which could also occur as the result of tow level exposure 140. p. 1271. This method is valid as svell as practical and necessary because The intrinsic carcinogenicity of a chemical does not depend on dose level although the proportion of animals developing cancers and the earliest time that tumors are detected are usually related to dosage 148, p. 1311. Therefore, were environmental levels considerably lower than animal dose lev- els, this would not invalidate the results of the testing for purposes of human risk assessment. Hosvever, in the case of formaldehyde, OTSstaff (7), the Mount Sinai committee (/5), and the NTP panel (14) have all pointed out that the tumors have occurred in rats at levels compara- ble to those encountered by humans. b) Results in laboratory animals are qualitatively relevant to humans since the overall patterns of metabolism are generally similar, although the type and site of cancer induced may differ. Basic biological processes of those molecular, cellular, tissue and organ functions that con- trol life are strikingly similar from one mam- malian species to another 50, p. 85). IMletabolic studies have shown that most differences between humans and experimen- tal animals are quantitative rather than quali- tative and support the idea that animal results can be used to predict human responses 1/, p. 1261. For example, the large body of infor- mation on the metabolism of benzo[eiJpy- rene shows that the overall pattern of metabolism in all species and systems tested is the same (although the carcino- genic potency may differ) (53). It should be noted that several scientific panels have raised the possibility that humans may, in fact, be more vulnerable to cer- tain carcinogenic substances than labo- ratory animals (39, 50, 5!). Weighing the usefulness of metabolic studies in human risk assessment, OSHA concluded: Illu general, for the purposes of negating the identification or classification of potential oc- cupational carcinogens, information on me- tabolism and pharmacokinelics is of extreme- ly little practical value at the present time 142, p. 51581. There is also scientific consensus that a negative human effect cannot be con- cluded from evidence that the target of the carcinogen differs in humans from that in experimenial animals. In experi- mental carcinogenesis, the type and site of cancer seen may or may not be the same a~ that recorded in human studies. For example, 2-naphthylamine induces bladder cancer in man, monkeys, dogs, and hamsters but hepatic cancer in the rat (43). Present knowledge indicates that . - - the re- sponsive target tissues or organs and the types of tumors induced in different species may vary greatly. Therefore - - - the finding of negative results in some other species gener- ally does not detract from the validity of a positive result as evidence of carcinogenicity for the test substance 38, p. 398661. Specifically, as regards formaldehyde, the CPSC states: There is no evidence of biological differences between the laboratory animals tested and humans that would decrease the potential for humans to develop cancer when exposed to formaldehyde 1351. This is in agreement with the findings of the OTS staff and the NTP panel, that formaldehyde metabolism and its reac- tion with cellular components is qualita- tively the same in all mammalian species examined to date, including man (7, 14). The reports also concurred that, al- though formaldehyde caused nasal can- cer in rats, this may not necessarily be the site affected in human beings. 2) It is not now possible to extrapolate from animal data a "safe" population threshold for any carcinogen regardless of the mechanism of action. IThel position that there is no presently ac- ceptable way to reliably determine a threshold for a carcinogen for any given population is amply supported by the evidence presented and also represents, to a large extent, a con- seusus of scientific opinion 142, p. 51371. Methods do not now exist for determining a safe threshold level of exposure to carcino- gens. The major obstacle to determining whether there are safe threshold exposure levels for carcinogens is the lack of data on the effects of tow exposure levels. - . - Be- cause there is not definitive evidence of the existence of thresholds and because not all cancer variables have been identified, pru- dence requires that no safe level thresholds be assumed to exist. - - - Exposure to any amount of a carcinogen. however small. must be regarded as an addition tothe total carcino- genic risk 48, pp. 133-134). The self replicating nature of cancer, the multiplicity of causative factors to which mdi- SCIENCE. VOL. 216 PAGENO="0154" viduals can be exposed, the additive and possible synergistic combination of effects, and the svide range of individual susceptibd' ities ssork together in making it currently unreliable to predict a threshold belosv xvhich human population exposure to a carcinogen has no effect on cancer risk 38. p. 398761. [Iii is generally accepted that a popula- tion threshold svhich svould define a "risk- free" dose for a group of people composed of diverse individuals, ifit exists. cannot now be demonstrated 1!. p. 121. This principle applies to any agent that contributes to the carcinogenic process: lilt would not be practicable or justifiable to establish different criteria for the identifica- tion, classification, or regulation of initiating and promoting agents. OSHA agrees with the NCAB Subcommittee that "any factor or combination of factors which increases the risk of cancer in humans is of concern regard- less of the mechanism of action" 142. p.51521. The mechanisms by svhich individual carcinogens induce tumors are not easily understood. Even svhere it max be possi- ble to definitis'ely classify a substance or agent as an initiator, promoter, or com- plete carcinogen, these distinctions are not practically useful for purposes of regulation. This is because of the impos- sibility of identifying "population" thresholds for any carcinogen regardless of the mechanism by svhich it operates. Under defined experimental conditions, a given substance may appear to shosv a threshold les'el (that is, a dose level beloxs svhich it does not increase tumor incidence). Hosvever. as indicated by a large-scale bioassay in svhich a threshold could not be identified for aceivlamino- fluorine (54), such observed thresholds may simply reflect the limited ability of the test system to detect effects at losv dose levels, Es'en svhere experimental thresholds could be established svith cer- tainty, it svould be impossible with cur- rent information and methods reliably to predict from experimental data the threshold level in humans (55). Such predictions are precluded by the difficul- ty of obtaining quantitative data at very loss dose levels in small numbers of animals, variations in human host re- sponses. and possible additive or syner- gistic effects of other agents that individ- uals might be exposed to. Thus. regu- latory agencies has-c sought to reduce human exposure to carcinogenic sub- stances to the lowest possible level con- sistent svith reles'ant social and economic considerations. From the abase discussion it is clearly not necessary that the mechanism of action be definitively established before formaldehyde is identified as a carcino- gen. Hoss'ever, the Formaldehyde Insti- tute (26) contends that formaldehyde is 18 JUNE i982 likely to act primarily via an epigenettc mechanism-causing tumors only at high doses through its cytotoxic or irritant action-and is therefore probably a "threshold" carcinogen. This s-iesv is also implied in the Todhunter memoran- dum (6). The epigenetic mechanisms suggested include Ii) cell destruction and rapid cell proliferation triggered at high dose levels that prevent DNA repair and detoxification systems from operating ef' fectively and (ii) induction of ulceration, irreversible hyperplasia. or metaplasta only at high dose levels svhich are prema- lignant in themselves or serve to pro- mote tumor formation 16. 26). The OTS, the NTP panel, and CPSC staff have rejected this interpretation. citing the substantial body of evidence that shosvs formaldehyde tobe genotoxic and the absence of factual support for the epigenetic or "threshold" theory (7, 14. 56). While the promoting effect of form- aldehyde may play a part in its overall carcinogenicity, formaldehyde is a po- tent alkylating agent (57); is a mutagen in a svide variety of test systems including microbial, insect, and mammalian sys- tems (8); induces sister chromatid ex- change in human lymphocytes (58); and causes unscheduled DNA repair in HeLa cells (59). Formaldehyde is able both to transform mammalian cells in culture at losv concentrations and to initiate cell transformation in vitro (60. 61). Formal- dehyde also enhances the genoloxic ef' feet of peroxides and radiation (14). Thus, svith regard to the hypothesis that the "carcinogenic effects of formalde- hyde are indirect, termed epigenetic," OTS concluded, "There is - . - absolute- ly no scientific evidence for this hypoth- esis in the published literature" (7, p. A-6). The NTP panel further noted that "a number of agents ssere reported to in. duce epithelial hyperplasia in several types of tissues but they had no carcino- genic or tumor promoting activity associ- ated sviih them" (14, p. 34). The panel "found no es'idence that the induction of irritation or. more specifically. of epithe- hal hyperplasia is a sufficient condition for the carcinogenic activity of an agent" (14, p. 34). The NYU studies (17) sup- port the NTP panel on this point. If severe irritation and resultant rapid cell turnover svere either a sufficient condi- tion or a necessary prerequisite for carci- nogenicily of formaldehyde, one svould expect an increased effect in animals exposed to a combination of formalde- hyde and a strong irritant. Hosvever. in the NYU study HCHO and HCI svere administered singly and in combination: there ssas no increased response to the combination of agents although HCI is a gaseous irritant; nordid HCI alone cause tumors although it did induce hyperpla- sia (16. 62). Scientists from the CPSC responded to the Todhunter memorandum, citing evidence against the "threshold" theo- ries above and rejecting the assertion that the observation of endogenoas lev- els of formaldehyde in animals svilhout spontaneous tumors indicated a thresh' old (63). They reaffirmed their prior con- clusion that "there is no evidence dem- onstrating that there is a threshold for formaldehyde, or a dose les'el below which it is certain that formaldehyde will not cause cancer" (35). 3) Positive human data are not neces- sary to regard an agent as a likely human carcinogen svarranting protective action. Although epidemiological evidence is a necessary prerequisite to actually call- ing a substance a "human carcinogen," this is a point of terminology rather than a criterion for taking protective action. The most posverful reason is that the usefulness of epidemiology for the den- tification of carcinogens is limited by a number of constraints. These include cost, the usual long lag betsveen expo- sure and appearance of cancer. the con- founding effect of multiple exposures to carcinogens, and difficulties in ideqtify- ing an appropriate control group Isee (48, pp. 125-128)1. A practical problem is that very large samples must be com- pared if the risk in the unexposed popu- lation is loss' and the number expected to show the effect is small. Because of the possibility of a false negative result, the absence of positive results cannot prove an absence of risk; however, an absence of positive results may be useful in plac' ing upper bounds on the magnitude of the risk (64): - . . lNlegaiive epidemiological data, .agies- tionable because of limitations in the power of deteclion of such siudies. do not deny the conclusion of carcinogenicily on the basis of animal assays 138, p. 398711. Thus: When a tosic substance is identified in a mammalian test system (in svhich the critena listed in the standards are used) as a pmdent health policy matter this substance is to be treated as posing a carcinogenic risk to human beings. . . - Because public policy mandates prevenlive health care, svailing for epidemio- logic data is unacceptable, since ii means waiting to "count dead bodies" 14!, pp. 14- 161. Specifically, as regards formaldehyde: epidemiological studies completed to date have not been specifically designed to evaluate the carcinogenicity of formal- dehyde in human populations; thus they 150 1289 PAGENO="0155" ~r)-~ ~r ~~ri 3~-~ ~ ~ c,~ ~ ~ `<~ rD ~ ~ :~ ~ E;~ ~. ~ OQ~ ~ z ~ ~ .3 ~~LH CT' z PAGENO="0156" z 2 C,' PAGENO="0157" 153 Natural Resources Defense Council, Inc. 122 EAST 42ND STREET NEW YORK, N.Y. ioi68 212 949-0049 IIo.shietgloo Office 11'esleot Office SI.. ITE 6oo SAN FRANCISCO C.I.LIF. 64108 WASIIINGTON, D.C. 20006 415 421-6361 202 223-8210. December 7, 1982 Denis J. Prager, Ph.D. Assistant Director Of fice of Science and Technology Policy Executive Office of the President Washington, D.C. 20500 Dear Dr. Prager: Thank you for the opportunity to comment on the discussion draft Potential Human Carcinogens: Methods for Identification and Characterization, Part I," prepared by the Regulatory Work Group on Science and Technology, of the Office of Science and Technology Policy (OSTP), October 1, 1982. I have reviewed the draft document in detail and find it to be in need of major revision in order to provide a Óoherent account of the state-of- the-art of the science underlying methods of carcinogen identi- fication and risk assessment. The chapters are of widely vary- ing quality, scope, and level of detail. Although certain areas and topics are adequately covered, there are numerous important omissions. To give a few examples: the progress made to date in identifying potential and known human carcinogens is ignored, as is the substantial biological and statistical basis for the various available risk axtrapola- tion models. Although there is an extensive (and inaccurate) discussion of the probability of false positives in animal bioassays, no corresponding attention is given to false nega- tives -- a far more serious problem from the point of view of public health. The limitations of in vitro tests are given undue emphasis while the re1evance~nd usefulness of combina- tions of these tests for purposes of carcinogen identification and risk assessment are not presented; and relevant practical questions related to exposure assessment and limitations of available methods and models are not adequately discussed. Most important, however, although the Preamble and Chapter 7 allude to advances in science' and "current biological thinking" that should affect regulatory policy, - ."ew E'glatul Of/ill: 7 ERIE DRIVE NATICK. SIA. III 71)() 9t 7 Public La'uL Ius1tt,~c: 72(1 RACE STREET KNEES. CO. 80206 ~j3 377-1)740 (RSI2RSclc((paper PAGENO="0158" 154 these advances are nowhere specifically identified, discussed, and evaluated. In particular, no scientific justification is offered in the draft for statements in Chapter 7 that consider- ation is being given to using a nargin of safety (or no observed effect level) approach for certain types of carcinogens based on their (presumed) mechanism of action (e.g., lack of evidence of mutagenic activity). Yet these statements imply a major departure from prior reportsl7 that have supported the use of risk extrapolation models to assess risk for all carcinogens. Therefore, the scientific basis and underlying assumptions of the safety factor or NOEL approach deserve extensive discussion and review. (It should be noted that much recent research does not tend to support this type of dichotomous approach to carcin- ogen risk assessment.) (See discussion on pp.19-22, imfra.) There are reservations about the need for this document in the first place. It is apparently intended to serve as the scientific basis for a regulatory policy to guide all federal agencies concerned with regulating cancer-causing substances and agents. As such, the document would replace the 1979 Imteragency Regulatory Liaison Group (IRLG) policy.1 The IRLG report was prepared by scientists from CPSC, EPA, FDA, OSHA, NCI, and NIEHS, was published in the Federal Register, and received wide public comment. It was described as representing the best judgement of the scientists in the agencies comprising the IRLG on the scientific concepts and methods currently in use to identify and evaluate substances that may pose a risk of cancer to humans. 2 Regulatory agencies have since relied on the IRLG for guidanße on identifying and assessing human risk from car- cinogens. EPA, for example, announced that it would supplement its formally adopted 1976 Interim Cancer Policy3 by the IRLG release for comment as a basis for assessing substances present in the air as well as in all other exposure media.4 Since 1979, several federal panels have reevaluated and confirmed the interpretation and general principles in the IRLG policy con- cerning carcinogen identification and risk assessment.5'6 Host recently, the State of Californiats Department of Health Services has issued a comprehensive draft Carcinogen Identification Pol icy: A Statement of Science as a Basis of Policy, which re- ceived review during its preparation from almost 50 scientists.7 Although these more recent documents refine and update the IRLG in some areas, they do not contradict the basic scientific interpretations and principles contained therein. The present OSTP draft itself does not present new scientific data that contradict the IRLG report. PAGENO="0159" 155 For this reason, one must question whether a comprehensive reevaluation -- entailing a substantial commitment of resources and expertise -- is in fact justified. As discussed in the enclosed comments on individual chapters of the draft, the present document is very far from serving as a useful resource, and would require major, if not total, revision. I would recom- mend, therefore, that the OSTP request the National Academy of Sciences to convene a panel to determine whether, and to what degree, and in what areas, the scientific basis of the IRLG policy requires updating. The NAS could then itself conduct any necessary review. Given the verydistinct disciplines in- volved -- in vitro testing, animal bioassay, epidemiology, risk assessment, etc. -- it would~ probably be appropriate to convene subgroups or working groups as necessary to be coordin- ated by an executive panel. If you have any questions on, the enclosed comments, please do not hesitate to contact me. Sincerely, ~ 2 c~,r. Frederica P. Perera, Dr.P.H. Senior Staff Scientist FPP : crp Enclosure PAGENO="0160" 156 REFERENCES 1IRLG: Scientific Eases for Identification of Potential Carcinogens Sand Estimation of Risks. Federal Register, 44:39858-39879, 1979. 39858. 3Health Risk and Economic Impact Assessments of Suspected Carcin- ogens, Interin Procedures and Guidelines. Federal Register 41:21402-21405, 1976. 4National Emission Standards for Identifying, Assessing, and Regulating Airborne Substances Posing a Risk of Cancer. Federal Register 44:58642-58670, see p.58647. 5Toxic Substances Strategy Committee (TSSC), Council on Environ- mental Quality: Toxic Chemicals and Public Protection: A Report to the President. U.S. Government Printing Office, Washington, D.C., May 1980. 6U.5. Congress, Office of Technology Assessment (OTA): Assess- ment of Technologies for Determining Cancer Risks from the Environment. U.S. Congress, Washington, D.C., 1981. 7State of California, Department of Health Services: Carcinogen Identification Policy: A Statement of Science as a Basis of Policy. July 1982. PAGENO="0161" 157 Mr. FL0RI0. Thank you very much. Dr. Silbergeld. STATEMENT OF ELLEN SILBERGELD, PH.D. Dr. SILBERGELD. Thank you very much for this opportunity to present testimony at these extremely important hearings. I am Ellen Silbergeld, the chief toxics scientist at the Environ- mental Defense Fund. My training and professional experience are in basic toxicology research. Most recently, before joining the EDF, I was chief of neurotoxicology at the National Institutes of Health, and I am presently a guest scientist in the Reproductive Toxicology Section of NICHD, involved in research on the effects of chemical carcinogens on reproduction. Over the past 2 years, we, with you, have watched with dismay the unraveling of a scientifically-based Federal policy on carcino- gens. What we observe is nothing less than a radical revisionism which uses a purported scientific basis to alter the regulatory im- plementation of carcinogen policy. The clear purpose of this revi- sionism is to justify a deregulation of exposure to carcinogens. Efforts have been made to develop a scientific basis for this regu- latory reform. These efforts have largely failed. Nevertheless, regu- lation by regulation, a de facto policy of revisionism, has been pushed through, and that is what I want to outline chiefly today. This revisionism has adversely impacted upon the protection of our health in the face of exposure to carcinogens in the workplace, in our food, our air, and our water. Moreover, there is a significant connection between these revisions in the hands of the EPA, and that agency's misconduct of the Superfund program, which has been of extensive concern to this subcommittee. As you have heard, a revisionist cancer policy is unnecessary and scientifically inappropriate. The Federal Government has or had a scientifically-derived cancer policy when the present administra- tion took office in 1981. This is the IRLG document which was de- veloped in 1979 through a rigorous process primarily of internal peer review, and published in the Journal of the National Cancer Institute. As you have heard, an early act of the new administration was to abolish the IRLG in 1981. We have heard repeated declarations and promises, declarations that a radically revised approach to car- cinogen identification and regulation is required, and promises that we shall be provided such a revision expeditiously. Now, what did this administration find unacceptable in the cancer policy it inherited when it took office? Basically, it can be summed up as follows: too many, too low, and too quick. That is, using the established guidelines, too many chemicals at too low doses were being identified as carcinogens too rapidly. This is not surprising in light of the widespread incidence of cancer in our society and the recent evidence published by Davis and associates for continued increases in age-specific incidence of certain cancers associated with increases in production of some synthetic organic chemicals which are human carcinogens. Nevertheless, for those whose constituency is the industry pro- ducing such chemicals, and for those whose responsibility is the 22-143 0-83--li PAGENO="0162" 158 cleanup of chemical wastes, this razor of cancer policy cut too close, too deep, and too often. Thus, clearly, the objections to the work of the IRLG and other extant cancer policies in place in 1981 were not scientific, but political. However, the administration has preferred to disguise its objec- tives in terms of science, with references to recent advances in mo- lecular biology made, interestingly enough, not by molecular biolo- gists but by economists in the administration, as a basis for the need to revise policy. You have heard this most frequently as a need to base regulations upon purported distinctions in the mecha- nisms of carcinogens. The Under Secretary of HHS, Dr. Brandt, the Administrator of OSHA, the Acting Administrator of the EPA have all made refer- ence to this so-called principle. However, it is interesting to note that as more light is brought to bear on this hypothesis, fewer and fewer advocates can be found to defend it, with reference to Dr. Todhunter's letter of partial recantation to Science which we have attached to our testimony. Another aspect of this revisionism which Dr. Perera touched upon, which has received, unfortunately, rather less attention has been at the level of applicability rather than so-called basic science. This type of revisionism relates to attempts to alter drastically the definition of acceptable risk of cancer. Now, I would like to pass over much of the material which has been touched upon by others before me, and go very directly to the examples which we are concerned about as they define a de facto cancer policy in the hands of many agencies. Clearly, dioxin is one issue of tremendous concern when we con- sider current Federal policy related to carcinogens. As Dr. Pitot has presented to you, what we know of the mechanisms of action of dioxin clearly undercuts any simplistic approach to carcinogen reg- ulation based upon distinctions, for example, between initiators and promoters. Indeed, in earlier investigations, even the EPA recognized the ex- traordinary potency of dioxin. In its Draft Ambient Water Quality Criteria document on dioxin, which was prepared in final form in July 1981, but for some reason has never been given final agency approval, in a 1982 speech delivered by Dr. Anderson, dioxin has been repeatedly described by the EPA as the most potent, most hazardous carcinogen yet investigated. In 1982, the EPA's Carcino- gen Assessment Group ranked dioxin eight orders of magnitude more potent than arsenic or vinyl chloride, and five orders of mag- nitude more potent than benzoapyrene. It also has very significant reproductive effects. Thus, those who would anticipate the more sensitive regulation of carcinogens according to their mechanism of action must con- front the example of dioxin and consider how appropriate regula- tory vigor can be brought to bear to prevent human exposure to this agent. On the issue of the relevance of animal experimentation and the results of animal bioassays to cancer policy, in addition to the very elegant statements made by Dr. Weinstein, I want to emphasize a point that Dr. Nelson made. That is for regulating new compounds, which is a burden of responsibility upon the EPA which it has PAGENO="0163" 159 failed to take up, no human data will ever be forthcoming on which to determine carcinogenicity or indeed any other type of hazard. To comply with the requirements of the Toxic Substances Con- trol Act, the EPA and the administration must rejoin the rest of the scientific world. They must accept the importance and preemi- nence of animal data in toxicity testing. What have they done? The opposite. On the advice of the~ Department of State, the EPA, in Oc- tober 1982, rejected such a policy of premarket evaluation when it was proposed by our trading partners in the Organization of Eco- nomic Cooperation and Development. I would now like to turn to the issue of risk assessment. This is the second part of any carcinogen policy. It is the derivation of quantitative values for the likely human health impacts of expo- sure to carcinogens by consideration of their toxic properties, dose- response relationships, and the types of likely human exposure which may be encountered. Currently there are many attempts under way to standardize and, we fear, in the process to degrade the sensitivity of different approaches used by the several regulatory agencies. However, the National Academy of Sciences recognized that the use of risk as- sessment may well be specific, to an agency's mandate and area of oversight, and on March 2 of this year declined to recommend a single or centralized mechanism for performing the task of risk as- sessment. We have agreed with this recommendation in our earlier com- ments to the Administrative Coun˘il of the United States on the same subject. With respect to risk assessment, as you have heard from Dr. Perera, the EPA has attempted to force into the risk assessment process used for carcinogens the approach of general or convention- al toxicology based on a no-observable-effects level. With this ap- proach the EPA has begun to propose standards based on accept- able daily intake basis, ADI, which is related to the establishment of a no-observable-effect level. Regardless of what models are used to provide a basis for risk assessment, whether you wish to argue over the existence of thresholds or the exact shape of linear extrapolations from data down to environmental levels of exposure and regardless of the role through which one wishes to incorporate mechanistic differences in calculating these risks, there remains the most important policy phase of risk assessment, which Mr. Richardson alluded to. This is the utilization of the calculated risks in order to guide regulatory or other policy action. It is on this point that the cur- rent actions of this administration in many agencies speak the loudest. One of the most egregious instances was the acquiescence of the National Cancer Institute to industry pressure to actually remove risk assessment on benzene. Apparently bowing to the large U.S. leverage exerted by our funding, the International Agency for Research on Cancer actually removed the section from its report on benzene, which showed an increased risk of 17 leuke- mia deaths per thousand persons exposed at exposures as low as 10 parts per million. We have appended a report on this subject from Science magazine. PAGENO="0164" 160 Other examples can be found in EPA's misconduct of the Super- fund program. In addition to the problems which you, among others, have brought to light in its enforcement, settlement and cleanup, EPA has deliberately brought into the processes of site as- sessment and remedial action its unacceptable strategies of risk as- sessment. There are two examples to consider. First, one which we all know of-the EPA's well-publicized actions and inaction on dioxin in Missouri and now apparently also in Michigan, and, second, the less well publicized attempts to discount hazards from trichloroeth- ylene leaching from Price's landfill in New Jersey. In Missouri, after some delay, the EPA undertook a series of as- sessments on the risk of cancer associated with exposure to dioxin at some of the infamous Missouri sites, and we now know risk as- sessments were done in connection with assessing the hazard of eating fish from the Tittibiwasi River and the Great Lakes in the Michigan area, which were contaminated with dioxin from mysteri- ous sources, according to the final document, but most likely from Dow Chemical Co. Despite a deplorable record of struggling with concepts of deter- mining media of exposure, the amount of exposure, duration of ex- posure and the types of people likely to be exposed in Missouri, the EPA has proposed using their Missouri-type risk assessments as a model for determining action at Superfund sites. Let's examine the risk assessment they did in Missouri in some detail. They used the standard no-threshold model and determined an upper limit of risk. What were these upper limits of risk found in Missouri? For stable workers at one of the most contaminated horse arenas, the risks ranged from 24 in a million to 15 in 1,000. That is an increased risk of cancer after exposure to dioxin in the soil. For riders, the range was between 5 in 1 million to 17 in 10,000. For only occasional spectators at the horse arenas, the risks ranged from 1 in 10 million to 3 in 100,000. Now let's move to children living at the contaminated Minker resident site in Missouri. The risks there were calculated from 26 in 10,000 to 72 in 100. What is most disturbing about these risks is the apparent lack of reaction provoked in policymakers of the EPA. These are extraordinary increases in risks, even given the relative insensitivity of any assessment technique. As a matter of fact, under some possible conditions of exposure and duration, as described by the EPA and included in our testimo- ny on page 19, the doses would start to approach the range at which tumors have been reported to be induced in animals. There is no need for extrapolation under such conditions. In the matter of trichloroethylene, another connection between EPA's revisionist risk assessment and the Superfund program comes to light. Documents that we have recently obtained from the EPA show that the former assistant administrator, Rita Lavelle, hired as her toxicology consultant Dr. Arthur Pallotta, with the ex- press purpose of revising the risk assessments of some of the more important carcinogens and other chemicals found leaking from many Superfund sites. Price's landfill in New Jersey is a large, 22-acre abandoned dump site which is high on the Superfund priority list. Chemicals cur- PAGENO="0165" 161 rently escaping from it threaten the drinking water supply of At- lantic City and have already contaminated wells of persons living near it. Trichloroethylene, TCE, is one of the compounds promi- nently identified in the ground water near this site, at levels ap- proaching a daily dose of 10 milligrams per kilogram per day, as admitted in a memo from Dr. Pallotta. The EPA documents show a concerted effort, coordinated with Dr. John Hernandez and the, Office of Research and Development, to apply conventional, no-observable-effect level, ADI risk-type as- sessments to TCE rather than carcinogen assessment. Mr. FL0RI0. Would you clarify the point that you made about the type of assessment? Dr. SILBERGELD. There is very clear evidence tracked through these memos, which we can submit for the record if you would desire them, stating a strong interest in the Office of Solid Waste and Emergency Response in revising human health risks associated with the solvent trichloroethylene in drinking water. And this is very clearly associated in statements by Dr. Hernandez with its ap- pearance in municipal water supplies and in ground water ac- quifers near hazardous waste sites. The agency goes on to say, indeed to admit, that a determination of what levels of TCE are acceptable in the ground water could affect the extent of remedial actions at such sites. And we have here a memorandum from Dr. Pallotta to Ms. Lavelle stating that it is "with great pleasure" that he presents "the first toxicology ADI," that is, a reassessment of the carcinogen, using an accept- able dose approach rather than a no-threshold linear extrapolation, "all signed off for TCE." Other EPA decisions speak with equal eloquence on the new ac- ceptability of high risk. In their recent final rules for the cleanup of abandoned uranium mine tailing sites, exposures will be permit- ted, as the document admits, which can be expected to cause an in- creased risk of lung cancer from radon exposure of 1 in 140. This very large increase in relative risk is buried in attempts to mini- mize its importance by citing currently low population densities around such sites. I would like to turn to the FDA. Rebuffed in its attempts~ to modify the Delaney clause, the Administration has neverless at- tempted to reduce regulatory viligance over carcinogens in our food supply by promulgating rules exempting from Delaney regulation food additives which contain carcinogenic chemicals. This is a dis- tinct departure from earlier regulatory guidance where carbon black, D&C Yellow No. 2, Reds No. 10, 11, 12, and 13, were all re- moved from use on the basis of containing carcinogenic constitu- ents. The principle behind this revision appears to be based on intent. That is, when constituents in an additive are intentionally present they can be regulated. But if constituents which are carcinogenic are only incidentally found, they cannot be regulated. This princi- ple is masked in the Federal Register notice in a discussion of de minimis risk. In the preamble it is stated that "the estimate of risk may be ex- aggerated by these conservative linear, no-threshold extrapolation models." PAGENO="0166" 162 Mr. FLORIO. I have been informed that we are going to have to leave here by 1:45. Ironically enough, Mr. Waxman is having a hearing on cigarette smoking, so I would appreciate it if you might summarize and then we will ask all of the other witnesses to keep that in mind. Dr. SILBERGELD. Fine. I would just like to quote from one of the regulations promulgated by OSHA with respect to exposures to in- organic arsenic, which again in the Federal Register notice it is ad- mitted that the standard which was accepted against the recom- mendations of NIOSH will permit an increased incidence of up to 61 cancers per thousand exposed workers. Incredibly, OSHA admits the discrepancy with the Supreme Court ruling on its early regulations regarding benzene. I think it is extraordinary to see this admission of the incongruity between current regulatory practice and all types of guidance from past his- tory, from court decisions and from the scientific community. I quote: "OSHA concludes that the new inorganic arsenic stand- ard does not reduce the risk of the exposure below the level of sig- nificance. The OSHA Act was enacted in order to reduce signifi- cant risk insofar as feasible. It should be noted that the Supreme Court"-this is the benzene case-stated as to a one in 1,000 level of risk of fatality that, and this is quoting the Court, "a reasonable person might well consider that risk significant and take appropri- ate steps to eliminate it." This is not what is going on in many of the regulatory fora we see before us today. This is not reasonable policy and, as you have heard, it is not scientifically-based policy. The administration has been unable to formulate a scientifically-based rational cancer policy. Both the White House and the several agencies have been repeatedly rebuffed by the scientific community. The self-serving equation of scientific complexity and regulatory laxity has been exposed as baseless and self-serving of special inter- ests. Nevertheless, as I have tried to outline to you, we can discern an accretion of cancer policy on a case-by-case basis as the adminis- tration deregulates and disarms our protection against carcinogens. This policy, as being covert and diffuse, is much less easy to recog- nize, debate or control. It is, therefore, imperative for Congress to provide oversight of this process to draw out from these various regulatory actions and proposals the continuing momentum toward revisionism. Science has provided you the tools to reject the explicit statements of this policy. You must defend us from the continuing attempts to imple- ment a bankrupt revisionist policy. Thank you. [Testimony resumes on p. 265.] [Dr. Silbergeld's prepared statement follows:] PAGENO="0167" 163 TESTIMONY OF ELLEN K. SILBERGELD The Environmental Defense Fund (EDF) is grateful for the opportunity to present testimony today at the hearings on carcinogen policy before the House Subcommittee on Commerce. Transportation, and Tourism. The Environmental Defense Fund is a national non-profit environmental organization of more than 47,000 members. Since its founding in 1967, EDF has been involved in identifying and commenting on issues of chemical carcinogenesis and the regulation of human and environmental exposure to such agents. I am Dr. Ellen K. Silbergeld, Chief Toxics Scientist of EDF; my training and professional experience are in basic toxicology research. Most recently. I was chief of neurotoxicology at NIH; I am presently a guest scientist in the Reproductive Toxicology Section, involved in research on the effects of chemical carcinogens on reproduction. Over the past two years we have watched with dismay the unravelling of a scientifically based federal policy on carcinogens. it is this process of decay and destruction which I wish to comment upon today in my testimony. What we observe is nothing less than a radical revisionism which uses a purported scientific basis to alter the regulatory implementation of carcinogen policy. The clear intent cf this revisionism is to justify a deregulation of exposure to carcinogens. Efforts have been made to develop a scientific PAGENO="0168" 164 basis for this regulatory reform. These efforts have failed. Nevertheless, regulation by regulation, a ~ facto policy of revisionism has been pushed through. This revisionism has adversely affected the protection of our health, in the face of exposure to carcinogens in the workplace. in our food, air and water. Moreover,* there is a significant connection between these revisions, in the hands of the EPA. and its misconduct of the Superfund program, which has been of concern of this Subcommittee. A revisionist cancer policy is unnecessary and inappropriate. The federal government has-or had-a scientifically based cancer policy when the present Administration took office in 1981. There was in existence a lengthy document on carcinogen policy which had been developed in 1979 through a rigorous process of internal and external peer review by the Interagency Regulatory Liaison Group (IRE.G) (attachment I). An early act of the new Administration was to abolish this integrative workgroup of government scientists. We have heard repeated declarations, and promises--declarations that a radically revised approach to carcinogen identification and regulation is required. and promises that we shall be provided such a revision expeditiously. What did the administration find unacceptable in this policy? Basically, it can be summed up as follows: too many. too low, and too quick--that is. using the IRLG guidelines, too PAGENO="0169" 165 many chemicals, at too low doses, were being identified as carcinogens too rapidly. This is not surprising, in light of the widespread incidence of cancer in our society and the recent evidence for continued increases in age-specific incidence of certain cancers which is associated with increases in production of some synthetic organic human carcinogens. Nevertheless, for those whose, constituency is the industry producing such chemicals and for those whose responsibility is the cleanup of chemical wastes, this razor cut too close, too deep, and too often. Scientifically, the tenets of the IRLG policy remain remarkably consistent with the important advances in our basic understanding of chemically mediated carcinogenesis which you have heard today from earlier witnesses. The IRLG policy is based on several concepts: (1) chemicals cause cancer by permanently altering the genetic nature or other control mechanisms of cells; (2) cellular carcinogenicity and the alteration of cell growth may involve very few cells as targets, and a very limited time of ~xposure; (3) a conservative approach to health protection is to assume a close relatiÓnship between dooe and carcinogenic effect, down to very low doses; PAGENO="0170" 166 (4) *because of problems in detecting tumorigenic events associated with exposure to low doses of a specific chemical, it is very difficult to challenge positive results with negative results on either epidemiologic or experimental evidence; (5) mutation is significantly related to carcinogenesis. ~such that observation of mutagenic activity is grounds to suspect carcinogenicity; (6) it is reasonable to extrapolate from data derived by objectively defined principles from other than studies on humans; such data nay be derived from long or short tern animal bioassay tests or from studies of cells exposed to chemicals j~ vitro; and (7) unit risk cannot be extrapolated to population risk without adequate attention to sources, population variability, modes of exposure. - None of these principles has been overthrown by recent advances in molecular biology. Clearly, the objections to the work of the IRLG are not scientific, but political. However, the Administration has preferred to disguise its objections in terms of science, with vague reference to recent advances in molecular biology as a basis for the need to revise policy. This has been most frequently expressed as the need to base regulations upon purported distinctions between carcinogens PAGENO="0171" 167 based on their mechanisms of action. The Undersecretary of HHS, the Administrator of OSHA, and the Acting Administrator of the EPA have all made reference to this so-called principle. However, it is interesting to~ note that as more light shines on this hypothesis, fewer and fewer advocates can be found to defend it (see Dr. Todhunter's letter of partial recantation to Science, attached as II). Another aspect of this attempted revisionism which has received less attention has been at the level of applicability rather than basic science. This type of revisionism relates to attempts to alter drastically the definition of acceptable risk of cancer. II. NEW APPROACHES TO CHEMICAL CARCINOGENESIS Basic to the so-called scientific revisionist approach is the hypothesis that carcinogens should be regulated according to their mechanism of action. It is conceded that some (few) substances may react with cellular DNA; these types of carcinogens require rigorous control, and for their regulatioi~ the standard approaches of no-threshold, linear dose-response can be applied. Presumably, radiation is such a carcinogen. These substances are sometimescalled `genetic' carcinogens. However, according to the theory, most carcinogens do not act directly on DNA--these are the so-called `epigenetic' carcinogens described by the Chemical Manufacturers Association: PAGENO="0172" 168 "One can separate chemical carcinogens into two groups. Once group obeys traditional toxicological principles which the other does not. This classification of carcinogens has been described by Weisburger and Williams in Casarett and Doufls' TOXICOLOGY: THE BASIC SCIENCE OF POISONS. These authors separate carcinogenS into genotoxic (initator) and epigenetic (promoter) carcinogens, the distinction being that genotoxic carcinogens induce cancer by initiating a permanent change in DNA. Epigenetic chemicals comprise those chemicals which alter the manifestation of cancer by other than genetic means . . . This distinction is extremely important because genotoxic carcinogens are irreversible and may not have thresholds; therefore, some risk is present at any dose. Promoting agents (epigenetic chemicals). on the other hand, do have demonstrable threshold levels; thus, there are exposures for which the risk is minimal, or the risk exists only for some finite interval after exposure." (Summary of the Health Effects of PCB5. prepared for ~MA by Ecology and Environment, Inc., November, 1981, pp. 2-1-2). The distinction is not always easy to pin down. Sometimes, it is used to distinguish carcinogens which do not show positive results in mutation assays; sometimes it appears to describe those compounds which require metabolic conversion to active substances; sometimes it seems to refer to the distinction between initiators and promotors of carcinogenesis, or between complete and incomplete carcinogens. These definitions are not equivalent to a relative description of potency or strength of evidence. The actual descriptive classification by Weisoburger and Williams is more complex and moEe precise: *epigenetic carcinogens are not only limited to promoters, but also include such solid-state carcinogens as asbestos, hormones like PAGENO="0173" 169 diethylstilbestrol, immunosuppressants, cocarcinogens like phorbol esters, as well as the classic promoters like croton oil which act sequentially with genetic carcinogens or procarcinogens. The potency of some of these compounds is well known, and should discourage easy equivalence of epigenicity with reduced hazard or potenÓy. In fact, the specific distinction of mutagenicity, is not considered of overriding importance even by Squires, who, in an attempt to develop a ranking system for regulating carcinogens, places limited emphasis on evidence of genotoxicity, allowing only 25 points out of 100 for this factor (see attachment III). But the intent of the revisionists is clear: nongenetic carcinogens do not require the same control as genetically acting substances. This can be better understood by looking at an example which has been proposed as deserving less stringent control. One of the most vigorously argued is the group of chemicals known as polychlorinated biphenyls (PCBs). These compounds are not clearly mutagenic in bacterial bioassay tests; it is claimed that epidemiological studies have failed to find evidence of human carcinogencity. But much research has been directed towards understanding the mechanisms of action of PCBs and related compounds. As discussed at these hearings by Dr. Pitot, one of the leaders in this area of research, these compounds are known to be in~ucers PAGENO="0174" 170 of the monOXygenase enzyme system. They act, with varying degrees of potency. by binding to a specific cell receptor (sometimes called the TCDD or dioxin receptor). As a consequence of this interaction, there is an increase in activity of the enzyme aryl hydrocarbon hydroxylase (AHH). This event is of considerable importance for several reasons: first, many chemical procarcinogens are metabolized by these enzymes to substances which interact directly with DNA; second, individual sensitivity to many classes of carcinogens has been related to activity of AHH (and its genetic control); and third, inhibition of this enzyme is related to inhibition of carcinogenesis under certain circumstances. Thus, ii~ is clearly a simplification to claim that because compounds like PCBs act through such mechanisms, their importance or hazard is degraded. The fallacy of this reasoning is most clearly exposed by consideration of dioxin, which acts through the same AHH mechanism. Dioxin is, according to many scientific studies, among the most potent carcinogens yet studied. Indeed, in ealier investigationS. the EPA has recognized the extraordinary potency of dioxin. In its draft Ambient Water Quality Criteria document on dioxin, prepared in July, 1981 (but for sone reason never given final Agency approval). in a 1982 speech by Dr. Elizabeth Anderson, Director of the PAGENO="0175" 171 Health Assessment Office, dioxin is repeatedly described as the most potent, most hazardous carcinogen yet described. In 1982. the EPA's Carcinogen Assessment Group ranked dioxin eight orders of magnitude more potent than arsenic or vinyl chloride and five orders of magnitude more potent than benzo(a)pyrene (see draft Health Assessment document on acrylonjtrile, pp. 13-158 to 13-163. attachment IV). A real area of complexity relates to understanding the conversion of chemicals into carcinogenic substances, the detoxification of reactive metabolites, and the opportunites available to the cell for DNA repair. When these processes occur, there may be non-linearities in the relationship between dose and effect (here defined as covalent binding of the reactive metabolite to DNA. or formation of so-called DNA adducts, to use Weinstein's term). Full understanding of carcinogenesis will depend on information on the concentration of these DNA adducts within the organ where tumors develop. Also, appreciation of the ability of cells to excise adducts, or otherwise suppress genetic damage from becoming permanent will be of importance. These concepls may appear tantalizing to those who would deregulate carcinogens, because the. types of kinetic relationships which deßcribe processes of activation, inactivation, and repair involve non-linearities. PAGENO="0176" 172 However, to incorporate this new knowledge into carcinogen regulation requires. as the IRLG recognized in 1979, exclusive reliance on animal studies. Human experiments, unless coupled with exhaustive biopsy study, will never develop such information on such precise events as toxicokinetic relationships and formation of DNA reactive compounds. Those who would anticipate the the more sensitive regulation of carcinogens according to their mechanism of action must acknowledge the preeminence of those methods capable to detecting such differences. But this is not the case. The strongest proponents of a tiered approach to carcinogen policy are often the strongest opponents of reliance upon experimental data. Recent scientific advances support a shift to increased use of experimental data on which to base regulation. Recent advances in molecular biology, based on such data, strongly endorse the causal connections between early mechanistic events and tumorigenesis. Policy and regulation should be based on these pretunorigenic events--ia vitro mutagenesis, alkylation. adduct formation, chrotnosomal alteration, or enzyme induction. These methods will permit not only detection of those nonlinear processes of interest to the revisionists. but also the protective screening of new chemicals envosaged by Congress in passing the Toxic Substances Control Act in 1976. Clearly, for PAGENO="0177" 173 new compounds, no human data will be forthcoming on which to determine carcinogenecity or any other type of hazard. To comply with TSCA, EPA and the Administration must rejoin the rest of the world and accept the importance of animal data in toxicity testing. On advice of the Department of State, the EPA in October, 1982, rejected such a policy of premarket evaluation as proposed by our trading partners in the Organization of Economic Cooperation and Development. This act of secession is part of am unscientific approach to regulation. III. NEW APPROACHES TO RISK ASSESSMENT Risk assessment is the second part of carcinogen policy, the derivation of quantitative values for the likely human health impacts of exposure to carcinogens by consideration of their toxic properties, dose-response relationships, and types of human exposure. The process is at best only approximate because almost never is there human evidence directly related to providing information on the health effects associated with the types and amounts of exposures under consideration. Most human data are derived from epidemiological studies where exposure and dose are only approximate and where sufficient number of `~ases nay not be available. Before the development of formal risk assessment. intuitive estimations of human health effects were implicit in much regulatory policy. 22-143 O-83----12 PAGENO="0178" 174 Currently, there are many attempts underway to standardize -- and in the process degrade their sensitivity -- the different approaches used by regulatory agencies. However, the National Academy of Sciences recognized that the use of risk assessment may be specific to an agency's mandate and area of oversight, and declined to recommend a single or centralized mechanism for performing this task (see the National Research Council report, "Risk Assessment in the Federal Government," March 1, 1983, attachment V). We have agreed with this recommendation, in our comments to the Administrative Council of the U.S. Areas of contention in risk assessment are primarily three: 1) the integration of animal studies (usually long-term bioassay studies of whole animal tumorigenesis, but other mutagenesis tests, such as short-term mutagenesis testing or measurement of DNA adducts, may also be considered); 2) the types of models used in order to predict events over a large range of exposure or dose; and 3) the actual judgment which is made of the results of risk assessment. An important aspect of risk assessment is the utilization * of animal studies. As recognized by the IRLG, and formerly by most other agencies of government, well-conducted animal studies are a critical part of the responsible and protective regulation of carcinogens. The alternative, which has PAGENO="0179" 175 sometimes been demanded by industry, and in some instances by the EPA. is to wait for human results, in effect to wait for human experimentation to occur by serendipity. Risk assessment requires the integration of information obtained over the range of exposures measured, in order to predict events occurring at much~ lower levels of exposure. To accomplish this, decisions concerning mechanisms are essential in order to consider events associated with very low doses. One important point of assumption is whether similar mechanisms are involved at high and low doses of carcinogens. Carcinogenic events at the level of the genome are probably similar; however, other important processes may well be affected by dose. This is discussed in detail by Hoel and co-workers in their article in Science. EPA The EPA has attempted to force into the risk assessment models used for carcinogens the approach of general or conventional toxicology, that of a no observable effect (or no observable adverse effect) level (NOEL or NOAEL]. With this approach, the EPA has begun to propose standards based on an acceptable daily intake basis (ADI), related to the establishment of a NOEL. The differences in this approach, as compared to classic models, can be seen in considering some of PAGENO="0180" 176 the changes in regulatory approaches to carcinogens in water proposed by Dr. Roy Albert, chairman of the EPA'S Carcinogen Assessment Group. Interestingly, he, too, is less willing to be identified with revisionism in the light of scientific and Congressional attention (see attachment II).. Albert's draft offers some graphic examples as shown in his Table 2. For dioxin, the water quality criteria set by the conventional technique would be 181 x higher than that set by the conservative cancer model; for aldrin, it would be 285 x. These is regulatory reform of an unparalleled savagery. Regardless of which models are used to provide a basis for risk assessment, and the role to which mechanistIc differences are incorporated in calculating risks, there remains the most important phase of risk assessment. This is the utilization of the calculated risks to guide regulatory or other policy action. It is on this point that the current actions of this Administration speak loudest. One of the most egregious instances was the acquiescence of the National Cancer institute to industry pressure to actually remove a risk assessment on bemzene. Apparently bowing the large U.S. leverage exerted by funding, the International Agency for Research on Cancer (IARC) removed a section from its rc'pert on benzene which showed an increased risk of 17 leukemia deaths per 1000 at exposures of loppm (see Science ~j~:914-5, September 7, 1982; attachment IV). PAGENO="0181" 177 Other examples caff be found in EPA's misconduct of the Superfund program. In addition~to the problems in enforcement, settlement, and cleanup, EPA has deliberately brought into the processes of site assessment and remedial action an unacceptable strategy of risk assessment. There are two examples to consider: first, the EPA's well-publicized actions --and inaction-- on dioxin in Missouri; and second, the less well-publicized attempts to disÓount hazards from trichioroethylene leaching from Price's Landfill in New Jersey. In Missouri, after some delay, the EPA undertook a series of assessments of the risk of cancer associated with exposure to dioxin at some of the infamous Missouri sites. These assessments involved several steps: estimation of likely concentrations (external dose) in the environment; conditions of exposure to that external dose, including media of exposure. amount, and duration; people likely to be exposed (age, sex, occupation); and application of relevant mechanistic models for assessment. Despite a deplorable record of struggle with these concepts, particularly those related to exposure, the EPA has proposed using the Missouri-type risk assessments as a model for determining action at Superfund sites. In Congressional testimony, for example. EDF had to convince the EPA that because children do indeed eat dirt, often in large amounts. PAGENO="0182" 178 ingestion constitued a significant route of exposure to dioxin. WG had to dissuade the EPA from its first concept, as enunciated by Rita Lavelle in a letter to Congressman Hammerschmidt. which was that exposure to carcinogens In soil would only occur with deliberate, and probably unusual and unlikely, ingestion of dirt. After settling this, we debated with EPA bow to determine the amount of dust which persons in the arenas might be expected to breathe. Risk assessments then were conducted using standard no threshold models, and the upper limit of risk found to be as shown in Table 1. For stable workers at one of the most contaminated horse arenas. the risk ranged from 2.4 x 10-5 to 1.5 x 10-2. For riders, the range was from 4.7 in a million to 1.7 in a thousand. For occasional spectators. risks ranged from 0.1 in a million to 3 in 100,000. For children living at the contaminated minker residence site, risks ranged from 26 in 10,000 to 72 out of 100. (Source: November 22, 1983. assessment of risks to persons coming into contact with contaminated areas at Imperial. Missorui, from Dr. E.L. Anderson to G.A. Lucero). What is most disturbing about these risks is the apparent lack of reaction they provoked in policy makers. These are extraordinary increases In risk, even given the relative insensitivity of assessment techniques. As a matter of fact. undnr some posr~ib1e conditions of exposure and duration, daily dose would start to approach the range at which tumors have been reported to be induced in animals. (December 14, 1983, exposure assessment of Missouri horse arenas and related areas. from Dr. Anderson to Lucero, attachment VII.) PAGENO="0183" 179 TABLE 1 Estimated Exposures to Dioxin at Some Missouri Sitesa Group Average Daily Dose Horse Arena 5.7 x 10-5-3.6 X Stable Workers mg/kg/day Riders 1.1. x lO-~--4.1 x Spectators 2.7 x 1O-~-7.O x 1O-~ Residences: Children 1.7 x iO-~-i.O ng/kg/day Adults aSource: December 14th Exposure Assessment, attachment V. Estimated Risk Assessmentb Concentration Dose Risk Water 2.1 x 1O-9ug/l 4.2 x i~_6 ng/kg/da 2.1 x lO-10ug/l 4.2 x i~-7 i~_6 ng/kg/da Other 1 ng/kg/da NOAEL Level (for general toxicity 1 x lO-3ng/kg/da API (for general toxicity) bwater Quality Criteria for 2,3.7,8-TCDD (draft) July, 1981. EPA. PAGENO="0184" 180 Yet, as we know, the EPA took no action in Missouri until political pressure, particularly following the Tines Beach Flood, made its continued inaction impossible. Thus, an implicit decision was made that risks as high as 72 ou~ o~ .100 were not significant. In the matter of trichloroethylene (TCE), another connection between EPA's revisionist risk assessment and the Superfund program can be seen. Recently obtained EPA documents show that the former Assistant Administrator Rita Lavelle hired as her toxicology consultant Dr. Arthur Pallotta, with the express purpose of revising the assessment of some of the more important carcinogens found leaking from many Superfund sites. Pri:e's Landfill in New Jersey is a large abandoned dumpsite, on the Superfund priority list. Chemicals escaping from it threaten the drinking water supply of Atlantic City. TCE is one of the compounds identified in the groundwater near this site. The EPA documents show a concerted effort, coordinated with Dr. John Hernandez, Dr. Elizabeth Anderson, and Frederic Eldesness, to apply conventional -- ADI -- type risk assessments to TCE. Other EPA decisions speak with equal eloquence on the new acceptability of high risks. In the recent final rules for cleanup of abandoned uranium mine tailings sites, exposures wil1 be permitted, as the document admits, which can be PAGENO="0185" 181 expected to cause an increased risk of lung cancer fron radon exposure of 1 in 140 (p. 55. Final Environmental Impact Statement for remedial action standards for inactive uranium processing sites (EPA 520/4-82-013-1. October. 1982). - This large increase in relative risk is buried in attenpts to ninimize its importance by citing currently low population densities around such sites. This raises another aspect of revisionist trends in carcinogen risk assessment. Risk is customarily expressed as increased probability of cancer. on the basis of orders of magnitude change. While risk is calculated using a representative hypothetical individual (for example, the stable worker at a horse arena or the child at a residence in Missouri). statistical extrapolations can be made to the population as a while, such that a 10~ risk may be considered equivalent to one additional cancer case in a population of 100.000. However, this is not a smple extrapolation. As pointed out by Park and Snee, in a paper for the American Industrial Health Council (1982). "it is an overly simplistic expectation to represent the entire carcinogenic process (in populations) by one tolerance distribution" (p. 15). The EPA is guilty of just such simplistic assumptions--for example, in the Final Environmental Impact Statement on uranium nine tailings, the agency admits thatits PAGENO="0186" 182 risk assessment is based on equating children and adults in terms of lifetime risk of radiation exposure. Populations clearly vary, and children differ from adults in at least the expected lifetime of their exposure, as well as more complex differences in pharmacokinetics. Extended discussion has already been presented to Congress concerning EPA's unscientific regulatory approach to carcinogenic pesticides at earlier hearings this year (February 21. 1983). At these hearings. Dr. Frederica Perera will review the implications of EPA's descision not to regulate formaldehyde. Similarly, there are important policies implicit in EPA's continued failure to regulate ethylene dibromide. These actions imply that evidence based on animal experiments alone are not sufficient to force regulatory action. Tn athlit.ion. the EPA staff has recently sent up for SAB review several documents on hazardous air pollutants, which are suspect carcinogens. Examination of the acrylonitrile document is of initc~rest. Aflur c~xtc~nnive discussion of the evidence for carcinogenicity in animals and in some epideniological studies. the document inserts a new approach to the estimation, and regulat.ion of carcinogens, based on potency. (This section is included as attachment IV to this testimony.) It presents relative rankings for 54 carcinogens, based on the so-called upper-limit risk estimates for persons exposed to 1 ug/m3 (or PAGENO="0187" 183 1 ugh) of the substance. This estimate is then converted to a daily dose, and divided by molecular weight to derive a potency index. This approach is curious: it has nothing in common with the proposals of IARC or of Weissburger and Williams or of Squires to rank carcinogens in terms of weight of the evidence. or in terms of mechanism of action. What is interesting about this approach, aside from its lack of justification in the text, are the results it provides as a relative scale to judge EPA's own tolerance levels. Compounds with equal ratings--ethylene dibromide and heptachlor, for example--have received very different regulatory at.t.ent.ion. Toxaphene receives a higher ranking than vinyl chloride or benzene. What is the "cutoff' or action level for EPA? FDA * Rebuffed at its attempts to modify the Delaney clause, the Administration has nevertheless attempted to reduce regulatory vigilance over carcinogens by promulgating rules exempting from Delaney regulation food additiveswhich contain carcinogenic chemicals ~ ~g. j~. p. 14464ff.. April 2. 1982). This is a distinct departure from earlier~regulatory guidance, where carbon black, D&C Yellow No. 2, D&C Red Nos. 10, 11, 12. and 13 were all removed from use on the basis of contamination by PAGENO="0188" 184 carcinogenic constituents. The principle behind this revision appears to be based on intent; that is, which constituent(s) of an additive are intentionally present (these can be regulated), aiid which constituent(s) are incidentally found (these are not be regulated). This principle has been masked in a discussion of de minimis risk, but in the preamble it is stated that `the estimate of risk may be exaggerated by these conservative [i.e., linear, no threshold] extrapolation models' (p. 14466). Risk assessment is to be used to caluculate an "acceptable maximum exposure level to the chemical . . . using procedures that would include an evaluation of pharmaco-dynamic data . . . and studies of the mechanism of action of the carcinogen" (p. 14468). In ruling specifically to allow continued use of D&C Green No. 6, the FDA used in part a no obuerved adverse effect level approach (p. 14141 Fed. ~ 47. April 2, 1982). Dr. Bramdt's testimony os February 21, 1983, alludes to FDA policy of making "a distinction between genotoxic and nongenotoxic carcinogenic effects" (p. 24). What this will mean for regulation is not yet clear. OSHA Following the Supreme Court decision on its benzene standard. OSHA attempted an interim revision of its carcinogen policy to reflect court-ordered considerations of risk. The PAGENO="0189" 185 Court stipulated that risk must be evaluated using "c:onuc~rvat.ivm assumptions in interpreting the data with respect to carcinogens, risking error on the side of over protection. rather than under protection." Clearly, this does not sit well with Reagan's OSHA. But they have been unable to formulate an escape from the court's stringent ruling, although in January. 1982. the new OSHA Administrator was still trying to insert cost-effectiveness and cost-benefit factors into carcinogen regulation (Fed. ~ 47, pp. 187-190, January 5, 1982).. In addition, OSFIA has evaded its obligation to establish and publish candidate andpriority lists of carcinogens. Nevertheless, as with EPA a ~ ~g~o cancer policy is being made. In its final rule on inorganic arsenic (January 14, 1983; Fed. Reg. 48. p. 1864 ff.)OSHA set a standard of 10 micrograms/rn3, despite the recommendation by the National Institute of Occupational Safety and Health that the exposure 3 limit should be set at 2 micrograms/rn . By OSHA's own risk assessment, this higher standard would permit an jnp~.~.~ed incidence of up to 61. cancers per 1000 exposed workers (using World Health Organization data). The lowest estimate of increased cancer incidence was 2 per 1000. This cannot be what the court intended as serious attention to significance of risk, or over protective interpretation. Incredibly, OSHA blatantly admits this discrepancy, in its notice: PAGENO="0190" 186 Finally. OSHA concludes that the new inorganic arsenic standard . . . does not reduce the risk of the exposure to inorganic arsenic below the level of significance. The OSHA Act was enacted in order to reduce significant risk insofar as feasible. It should be noted that the Supreme Court stated as to a 1 in a 1000 level of risk of fatality that `a reasonable person might well consider the risk significant and take appropriate steps to eliminate it' (IUD V. API. 448 US 665)' (p. 1867) IV. SUMMARY Over the past two years. several agencies of government and the White House Office of Science and Technology Policy (OSTP) have offered for external review proposals to revise carcinogen policy. All have been rejected with rather harsh criticism by ~c:ic~r,t.if Ic peer review, much of which has been presented to Congress at earlier hearings. In response to this, in one instance, the primary author of the EPA document, Dr. Roy Albert, as well as Assistant Administrator Todhunter, has attempted to distance himself from his effort (see attached letter. II), and in another, the primary author, Dr. Dennis Prager of OSTP has resigned. Several other revisionist documents are said to be in the process of preparation--from OSHA. EPA. and the Department of PAGENO="0191" 187 Energy. It is likely that they will be similar, at least in intent, to the rejected proposals already put forward. All these attempts are betrayed by their transparent commitment to regulatory revisionism. Their process is too clearly retrograde, whereby science is butchered to fit a Procruntean bed of regulatory revisionism. Clearly this approach tends to obfuscate science. We hope that the revisionists will learn that policy should follow science, rather than the reverse. At this point we appear to be at a stalemate. No new carcinogen policy has met with peer approval. But the 1979 IRLG policy etands without final acceptance. Nevertheless, carcinogens are being regulated--or deregulated--by this Administration. As evidenced in the discussion above, there is being created a ~ facto carcinogen policy, despite the rejection of the revisionist attempts. In conclusion, the Administration has been unable to formulate a scientifically based, rational cancer policy. Both the White House and the several agencies have been repeatedly rebuffed by the scientific community. The self-serving equat.ion of scientific complexity with regulatory laxity has been exposed as baneless and self-serving of special interests. *Novertheleos, we can discern an accretion of cancer policy, on a case-by-case basis, as the Administration deregulates and disarms our protection against carcinogens in food, the PAGENO="0192" 188 workplace, and the general environment. This process, being covert and diffuse, is nuch less easy to debate and control. It is therefore imperative for Congress to provide oversight of this process, to draw out from these various regulatory actions and proposals the continuing nonentum towards revisionism. Science has provided you the tools to reject. the explicit statements of this policy; you must defend us from the continuing attempts to implement a bankrupt revisionism. For this reason, your hearings today are very important. PAGENO="0193" 189 REFERENCES 1. American Industrial Health Council, 1980. Chronic Health Hazards: Carcimo_g~j~esis, Mutagenesjs, Teratogenesis. 2. Anderson, E.L. (1982). "Quantitative methods in use of the U.S. to assess cancer risks." Paper presented at the workshop on Quantitative Estimation of risk to Human Health from Chemicals. 3. Davis, D.t.., Bridbord, K., and Schneiderman, M. (1982). "Cancer Prevention: assessing causes, exposures, and recent trends in mortality for U.S. males 1968-1978,~ Teratogen. Carcinogen. Mutagen. a:105-135. 4. Chemical Manufacturers Association (1981). ~~a~gf the Health Effects of PCBs. 5. Environmental Protection Agency (1982). Health Assessment document for Acryloritrile. 6. Environmental Protection Agency (1981). Ambient Water Quality Criteria document for Dioxin. 7. Goldstein,. J.A. (1980). "Structure-activity relationships for the biochemical effects and the relationship to toxicity." In Kimbrongh. R.D. (ed) Haloqenated flj_eqyj~ ~g.~phenyls. Dibenzodioxims, and Related Products, Elseier-Aorth Holland, pp. 151-190. 8. Hoel, D.C., Kaplan, N.L., and Anderson, M.W. (1983). "Implication of nonlinear kinetics on risk estimation in carcinogenesis," Science 219:1032-1037. 9. Kouri. R.E., Schechtinam, L.M.. and Nebert, D.W. (1980). "Metabolism of chemical carcinogens." In Kouri, R.E. (ed) Genetic Differences in Chemical Carcirioqenesis, CRC Press, pp. 21-61., 10. National Research Council (1983). "Risk Assessment in. the Federal Government." 11. Park, C.N. * and Snee. R.D. (1982). "Quantitative Risk Assessment, State of the Art for Carcinogenesis" prepared for AlEC. 12. Squires, R.A. (1981). "Ranking Animal Carcinogens: a proposed regulatory approach," Science ~j:877-880. 13. Weissburger, J.H. * and Williams, G.M. (1980). "Chemical Carcinogenesis." fl~ Casarett and Doull's ~~~šicology~ the Basic Science of Poisons, pp. 84-138. - 14. Whittemore, A. * and Keller, J.B. (1978). "Quantitative Theories of Carcinogensis." SIAM Rev. ~:1-30. 22-143 0-83---13 PAGENO="0194" 190 AttachTnents I. tRr~G Cancer Policy, J. Nat. Cancer Institute 63:244-268 (1979). TI. T&LL~rs La ~gjen.gg~ by Dr. John Todhunter and Dr. Roy Albert. III. Article by Squires. Science. IV. CAG relative potency ranking of carcinogens. V NAS-NRC (1983) Risk Assessment in the Federal Government. VI. IARC and Benzene - Science. VII. EPA Risk Assessment on Missouri dioxin sites. December, 1982. PAGENO="0195" 191 ATTACHMENT I ~D s7 `4L i',~oc~ Environmental Protection Agency Food and Drug Occupational Administration Safety & Health Administration Scientific Bases for Identification of Potential Carcinogens. and Estimation of Risks Report of the Interagency Regulatory Liaison Group, Work Group on Risk Assessment Received February 5, 1979; accepted February 17, 1979. Address reprint requests to Executive Assistant, Interagency Regulatory Liaison Group, Room 500, 1111 18th Street N.W., Washington, D.C. 20207. PAGENO="0196" 192 242 InteragenCy Regulatory Liaison Group During the preparation of this document, the Interagency Regulatory Liaison Group consisted of four agencies: the United States Consumer Product Safety Commission (CPSC); the United States Environmental Protection Agency (EPA); the Food and Drug Administration (FDA) of the United States Department of Health, Education, and Welfare; and the Occupational Safety and Health Administration (OSHA) of the United States Department of Labor. Work Group Members' Eula Bingham. IRLG Principal (Assistant Secretary of Labor for Occupational Safety and Health) Joseph V. Rodricks. Chairman (Food and Drug Administration) Elizabeth L. Anderson (Environmental Protection Agency) David W. Gaylor (Food and Drug Administration. N~tional Center for Toxicological Research) Richard A. Metier (Consumer Product Safety Commission) Anton H. Keller (Occupational Safety and Health Administration) Frank Rover (Environmental Protection Agency) Joseph McLaughlin (Consumer Product Safety Commission) Additional Participants in the Work Group Roy E. Albert (Environmental Protection Agency) Richard R. Bates (National Institute of Environmental Health Sciences) David G. Hoel (National Institute of Environmental Health Sciences) Umberto Saffiotti (National Cancer Institute) Marvin A. Schneiderman (National Cancer Institute) Valuable guidaore was received from Arthur C. Upton (Director. National Cancer Institute) and David P. RaIl (Director, National Institute ot Environmental Health Sciences). The Work Group acknowledges the, assistance ol Edward Altera (Food and Drug Administration); Ann Barton. Steven Jellinek. Rirhard Hilt, and Elten Siegler (Environmental Protection Agency); Steven Bayard, Donald Clay. and Raymond SV.t)tman (Consumer Product Satrty Commission): Osarles C. Brown and James Sontag (National Cancer Institute); Carl Gerber (Oilier ot Sri mr and Technology Policy. Executive OIlier ol she President); Nathan J. Karch (Council on Environmental Quality. Executive OIlier ot hr President); and j. William Lloyd (Occupational Safety and Health Administration). PAGENO="0197" 193 identification of Potential Carcinogens and Risk Estimation 243 ABSTRACT-Three types of evidence can be used to identify substances that may pose a carcinogenic hazard; these types are designated in Part I of this report cc 1) epidemiologic evidence derived from atudies of exposed human populations, 2) experimental dvidence derivod from long-term bloccacyc on animals, and 3) supportive or suggestive evidence derived from studioc of chemical structure or from short-term or other tests that are known to correlate with carcinogenic activity. Part II delineates the scientific bases for accepting evidence from these three sources and describes their relative contributions to the determination that a substance may pose a carcinogenic hazard. Further, It details the factors that should be considered In the evaluation of experimental and epidemiologic data for ascertaining the reliability and scientific merit of each source of evidence. it also specifies how certain types of limitations In data may require qualification of conclusions. Because data on experimental animals are currently the major source of information for assessing ccrclr'ogenlcity, they receive the greatest emphasis. Features of experimental design and conduct that influence the evaluation of such studios are discussed, as are the criteria for making evaluations. The report Is not Intended to specify how such studies should be designed and conducted; rather, it discusses how data from experimental animal studies of widely varying content and quality should be evaluated for purposes of identifying carcinogens. Epidemiologic data and some of their limitations are discussed In less detail. Chemical structure and the short-term tests that correlate with carcinogenic activity are briefly described, as are their roles In providing suggestive or-if coupled with positive dOta on animals or humans-sup- portive evidence of carcinogenlclty. In Part II are presented the criteria used to ascertain the adequacy of evidence purporting to shbw that a substance does not pose a risk of cancer. Part Ii also includes discussions of some types of experimental evidence that, If the extent and quality are adequate, may be used to show that certain carcinogenic responses observed in experimental animals may not be predictive of human response. Part ill sets forth current methodologies for quantification of risk, Included are discussions of mathematical models available for extrapolation, within a biologic system, of cancer incidence data observed at experimental dose levels to estimate risks at the (usually much lower) levels that are of concern for humans. Also presented are the factors that should be considered In attempts to Identify the human population(s) at risk and to define their conditions and levels of carcinogen exposure. Part Iii also deals with correlation of the magnitude of effects observed In one human population group or in experimental animals (under their condi- tions and level of exposure) with the magnitude of effects in the human population for which the estimate of risk Is being made. Limitations in current risk estimation methodologies are described, as are the problems of ensuring that human risk is not underestimated. The issue of thresholds for carcinogens Is discussed In the final section of Part lll.-JNCI 63: 241-268, 1979. PAGENO="0198" 194 TABLE OF CONTENTS Part I. INTRODUCTION 245 Part II. THE QUALITATIVE DETERMINATION THAT A SUBSTANCE POSES A CARCINOGENIC HAZARD 245 DEFINITION AND EXTENT OF THE PROBLEM 245 Nature of CarcinogefleSiS and Carcinogenic Responses 245 Estimation of the Number of Carcinogenic Substances 246 Enhancing Factors 246 Variability of Effects of Carcinogens 246 EPIDEMIOLOGIC EVIDENCE 247 Types of Epidemiologic Evidence 247 Disease Ascertainment 248 EVIDENCE FROM EXPERIMENTAL ANIMALS 248 Criteria for Evaluation of Experimental Design and Conduct 248 Experimental Design 248 Choice of the Animal Model 248 Number of Animals 249 Route of Administration 249 Identity of the Substance Tested 250 Dose Levels 250 Age at Treatment 251 Conduct and Duration of Bioassays in Animals 251 Criteria for Evaluation of Pathology 252 Pathology Examination 252 Evaluation of Pathologic Results 252 Internal consistency of the data 252 Reproduciblzty of test results 252 Evidence of a positive dose-response relationship 252 Concordance of results 253 Evaluation of tumor incidence 253 Evaluation of tumor morphology 253 General evaluation of neoplsstic pathology for carcznogenesis bioassays 254 Statisticat Analysis of Results 255 SHORT-TERM TESTS FOR CARCINOGENS 256 Methods Based on Genetic Alterations 256 Methods Based on Neoplastic Cell Transformation 257 Evaluation of Short-Term Test Results 257 MOLECULAR STRUCTURE AS SUPPORTING EVIDENCE IN IDENTIFICATION OF CARCINOGENS.. 257 QUALITATIVE JUDGMENTAL FACTORS IN EVALUATION OF TOTAL EVIDENCE 258 Part II. THE QUANTITATIVE ESTIMATION OF RISK 258 MATHEMATICAL MODELS FOR HIGH-TO-LOW DOSE EXTRAPOLATION WITHIN A SINGLE B1OLOGIC SYSTEM 260 The Models 260 Procedures 261 CHARACTERIZATION OF POPULATION EXPOSURE 261 Sources of Human Exposure 262 Analytical Methods for Detection and Measurement of Exposures 262 Routes and Conditions of Exposure 263 Duration, Frequency, ond Intensity of Exposure 263 Size and Characteristics of Exposed Populations 283 EXTRAPOLATION FROM OBSERVED EFFECTS TO ESTIMATES OF RISKS FOR EXPOSED POPULATION 263 Correlations From ObservEd Human Population Groups to Others 263 Animal-to-Human Correlations 264 LACK OF PREDICTABLE THRESHOLDS FOR AN EXPOSED POPULATION 264 SUMMARY OF RISK ESTIMATION 265 REFERENCES 265 Jscl. VOL 63. NO. 1. JULY 1979 244 PAGENO="0199" 195 Part I. INTRODUCTION This document describes the best judgments of the scientists in the agencies comprising the Interagency Regulatory Liaison Group (IRLG) on the scientific concepts and methods currently in use to identify and evaluate substances that may pose a risk of cancer to humans. These are fundamental steps in any program regulating carcinogens. The document was prepared by the Risk Assessment Work Group of the IRLG agen. cies and senior scientists from the National Cancer Institute (NC!) and the National Institute of Environ. mental Health Sciences. The document describes I) the basis for qualitative evaluation whether a particular substance presents a carcinogenic hazard and how the results of epidemio. logic studies and animal bioassays, along with other types of information, are used in making that evalua. tion; and 2) the methods used for quantitative estimates of the carcinogenic risk posed by the substance, if such risk estimates are appropriate or required. This document will provide a valuable scientific tool, to be considered with other information, in the evaluation of risk and ascertainment of the adequacy of experimental and epidemiologic methods used in that evaluation. It is an important step in ensuring that the regulatory agencies evaluate carcinogenic risks con~ sistently. The four IRLG agencies caution, however, that this document presently has no regulatory status. The methods used for regulatory purposes in making a qualitative determination that a substance poses a carcinogenic hazard to humans are based on a sub. stantial scientific consensus that has emerged from experience, research, debate, and review. Aithotigh some points need further clarification and definition, substantial agreement exists among the Federal regula- tory agencies on criteria for evaluating the carcino- genicity of a substance. In addition to determining that a substance may pose a hazard of cancer, regulatory agencies must consider other possible health hazards, and in some instances they are required to balance considerations of risk with other factors (such as possible health benefits or economic costs and benefits) in reaching regulatory decisions. DEFINITION AND EXTENT OF THE PROBLEM Nature of Carclnogeneals and CarcInogenIc Responses The characteristic toxicologic event in carcinogenesis is a change in the regulatory mechanism of the target cells, resulting in self.replicating cell lesions. The car- Its use will, of course, depend upon the statutory requirements of the individual agencies. The agencies have subjected this document to scien- tific peer review through the submission of the docu. ment to the Journal of the National Cancer Institute. In addition, a public notice and comment procedure has been initiated by publication in the Federal Regis. icr. Since the Occupational Safety and Health Ad. ministration (OSHA) has already received extensive public comment on these and other issues regard. ing the development of its cancer policy rule.mak. ing and will soon promulgate its policy, only the Consumer Product Safety Commission (CPSC), the Environmental Protection Agency (EPA), and the Food and Drug Administration (FDA) will participate in the public notice and comment procedure on this docu- ment. At the condusion of the notice and comment procedure, OSHA will consider whether revisions to its final cancer policy are appropriate. The four agencies emphasize that the goal of this process is to articulate a consistent policy on the scientific principles applicable to the identification and evaluation of substances that may pose a carcinogenic risk to humans. Part II discusses the qualitative determination that a substance poses a carcinogenic hazard. Part III dis- cusses quantitative estimation of risk. cinogenic event so modifies the genome and/or other molecular control mechanisms in the target cells that. these can give rise to a progeny of permanently altered cells. This progeny of cells constitutes the basis of the neoplastic disease. The expression of the toxic injury therefore does not derive from the same cells originally hit by the toxic agent nor from their functional products but rather from the proliferation of a new population of altered cells. The critical molecular injury caused by specific car- cinogens may be quantitatively extremely limited- even to a few cells-and may therefore not be de- tectable. What will make it manifest, through the subsequent growth of a clinically detectable neoplasm, is the proliferation of the altered cell population. The intensity of the pathologic response in a subject (i.e., the growth rate and spread of a cancer) depends on conditions of the host subsequent to the initial carcino- genic event and can be modified by other factors, such as enhancing agents and dietary factors. The continued* progression of clinical manifestations of the carcino- genic process can occur in the absence of continued exposure to the carcinogen. Carcinogenic effects are therefore self-replicating toxic effects different from the common terminal toxic effects in which the manifesta. 245 JNcI, VOL. 63, NO. I. JULY 1979 Part H. THE QUALITATIVE DETERMINATION THAT A SUBSTANCE POSES A CARCINOGENIC HAZARD PAGENO="0200" 196 246 interagency Regulatory Liaison Group* tions of toxicity are due to altered functional products, degenerative changes, or death of the target cells them- selves (1). A rigorous methodology must be followed in obtain- ing, reviewing, and documenting the data required for a determination of carcinogenicity from observations on humans and experimental studies. Both epidemics. logic observations and experimental studies need to be correlated with informatio~n on the chemical and physi- cal nature of the agents under consideration, their reactivity, and their fate in the environment and in the exposed organisms Evidence of carcinogenicity can be obtained from three sources: 1) epidemiologic evidence from exposed hu- man populations; 2) experimental evidence from long-term bio- assays in animals; 3) suggestive evidence derived from studies of chemical structure, reactivity, DNA damage acid repair, mutagenicity, neoplastic tans- formation of cells in culture, induction of preneoplastic changes, or from other short- term tests that correlate with carcinogen- In the evaluation of the results of carcinogenesis studies, the evidence obtained from epidemiologic ob- servations or from experimental bioassays does not necessarily fall sharply into the two categories of positive and negative: In many instances tlse evidence may be insufficient for a definitive assessment. Estimation of the Number of CarcInogenIc Subatancert Relatively few chemicals have been found to be carci- nogenic. In fact, available evidence indicates that most substances do not cause cancer. The NCI's "Survey of Compounds Which Have Been Tested for Carcinogenic Activity" (2-8) and other literature surveys and reviews provide results of long-term animal.bioassays on about 7.000 chemicals. Evidence of carcinogenicity on the basis of currently accepted experimental testing meth- od.s is available for less than 1,000 chemicals and possibly for as few as 600-800 (9-31). Many of these substances were selected for testing because of their structural similarity to known carcinogens. Thus these data considerably overstate the true proportion of card- nogentc substances in the human environment. A critical review of the literature on carcinogenicity of chemicals has been undertaken by the International Agency for Research on Cancer (IARC) with the support and collaboration of NC! (9-25). Of 368 chemicals evaluated in volumes 1-16 of the IARC monographs, some evidence of carcinogenicity was found for 247 (35). A small number of chemicals has been adequately studied by epidemiologic methods to determine whether a carcinogenic hazard exists. By one recent estimate, 26 chemical substances or processes have been identified as responsible for cancer induction in humans (9-25, 35). Of those 26 substances, 6 were fimt identified as carcinogenic by tests in animals, whereas 20 were first identified by epidemiologic evidence, Of the 368 substances for which carcinogenesis data were reviewed by the IARC, 221 showed some evidence of carcinogenicity from tests in animals, but these sub- stances had not received adequate epidemiologic study to evaluate their effects in humans (35). In addition, 15 occupational categories have been reported to be asso- ciated with excess cancer incidences without identifica- tion of a specific etiologic agent (36-50). EnhancIng Factors * Experimental and epidemiologic data suggest that some agents may not be carcinogenic alone but sub- stantially contribute to the development of cancer in subjects that have been exposed to carcinogens- De- pending on experimental circumstances, these agents have been referred to as cocarcinogens, promoting agents, syncarcinogens, or more generally, modifying or enhancing factors (51, 52). Research on this category of agents suggests that they may work through a number of mechanisms of action, including (51. 52): a) alteration of the uptake and/or distribution of carcinogens, b) modification of the metabolic activation of carcinogens, c) enhance- ment of the susceptibility of target tissues, and d) ac- celeration of neoplastic progression. Current evidence suggests that some of these agents act by a mechanism that may be specific for particular organs or conditions of exposure. Because of the possible specificity of their mechanisms of actions, the activity of these agents may not be recognized by conventional bioassays. Since no common general pathway of action has been recognized, it is not expected that tests based on a single-mechanism end point will be applicable for the identification of a broad range of these substances. Enhancing mechanisms may be a major factor in the development of human cancers; therefore, their identi- fication and control may be important in cancer prevention. Since no general methodology yet exists for testing and evaluation of this entire group of sub- stances. the special circumstances under which each may act must be carefully evaluated. Interpretation of a positive effect in a carcinogenesis bioassay as being due to one of these mechanisms would require rigorous documentation that a full carcinogenic process is not involved. VariabIlIty of Effects of Carcinogens Variability in the action of carcinogens may be due to inherent differences in susceptibility among species and strains of test animals and within populations of humans, and also to variability in the intrinsic differ- ences in carcinogenic reactivity of individual agents. For example, aflatoxin B1 is strongly carcinogenic in rats but is ineffective in several strains of adult mice (53). ~-NaphthyIamine is carcinogenic for humans, dogs, and several other species; but this compound has JNU. VOL 63. NO. t. JIYLY t579 PAGENO="0201" not produced tumors in rats (51). With some other carcinogens, there is a greater concordance of results among species: Dimethylnisrosamine has been found to be carcinogenic in all of the strains of vertebrates tested (55). Species and strain differences in susceptibility to carcinogens may be due to factors that affect transport and metabolism, which in turn determine the effective dose of the ultimate form of the carcinogen delivered to target cells. These differences may also be due to inherent variations in susceptibility to neoplastic trans- formation of different organs in different species (56). Differences in the level of carcinogenic effect of individual agents can only be compared with precision under strictly defined conditions of dosage and biologic end points. Frequently the level of effects, even under strictly defined conditions, will show marked variabil- ity depending on the test system used. Nevertheless, in the extreme, some carcinogens are clearly more effective than others by several orders of magnitude (9-25). However, such comparative potency estimates must be made with caution. EPIDEMIOLOGIC EVIDENCE Evidence of carcinogenic activity of an agent can be obtained from epidemiologic studies when evaluation of the observations shows that the test agent causes an increased incidence of neoplasms or a decrease in their latency period. Evidence from studies of human populations identi. fies carcinogenic chemicals to which those popula- tions were exposed in the past. Many substances that have been identified as carcinogens in humans were discovered by epidemiologic studies of exposed workers; this evidence dates from 18th-century observa- tions of cancer in chimney sweeps to more recent observations on dye workers, asbestos workers, and workers in certain chemical industries (31). It was noted early that clinical signs of cancer are delayed for a long time after initial exposure to carcinogens. This period of latency-often 5-40 years from initial ex- posure until the disease appears-makes prompt deter- tion of newly introduced carcinogenic substances by epidemiologic studies nearly impossible. As more substances are introduced into the human environment and as more are tested experimentally, it is expected that a larger proportion will be identified as carcinogenic; this will be followed by adequate control measures, so that epidemiologic confirmation may become impossible, Types of Epidemiologic Evidence Types of epidemiologic evidence of carcinogenicity in humans include neoplastic response directly related to duration and dose of exposure, incidence or mor- tality differences related to occupational exposure, incidence or mortality differences between geographic regions related to environmental rather than genetic differences, altered incidence in migrant populations, time trends in incidence or mortality related to either the introduction or removal of a specific agent from the environment, case-control studies, and the result of rerrospective.prospective and prospective studies of the consequences of human exposure. Clinical case reports may also provide early warning of a potential carcino- gen (57). The two main types of epidemiologic studies used to establish evidence of a carcinogenic hazard are cohort studies and case-conirol studies (58). Epidemiologic cohort studies involve the comparison of groups differ- ently exposed to a substance, The comparison may include a) totally unexposed versus exposed groups, b) groups having distinctly different levels of exposure, or c) rates in exposed groups versus rates prevailing in the general population. The groups need to be comparable for demographic factors such as age, sex, and race, and controlled for exposure to known carcinogens. Epidemiologic case-control studies involve compari- son of people with a given cancer type versus people without the disease but otherwise comparable with respect to appropriate demographic variables, to ascer- tain if they differ in exposure to the cancer hazard under investigation. Epidemiologic findings gain greater force with in- creasing numbers of well-conducted studies that show similar effects from a given substance under different circumstances. Absence of a positive statistical correlation does not by itself demonstrate absence of a hazard. Whereas negative epidemiologic data usually do not adequately establish the noncarcinogenicity of suspected materials, such negative data obtained for a given agent from epi- demiologic studies of sufficient extent and duration may indicate the .upper limits for the rate at which a specific type of exposure could affect the incidence and/or mortality of specific human cancers under the conditions of observation, The detectability of a carcinogenic effect in a group of humans depends on several factors, including the duration and extent of exposure, size of the exposed population, and background rate of cancer in the target organ. Evaluation of epidemiologic studies re- quires a knowledge of the smallest possible increase in tumor incidence detectable under the conditions of each study. Such information has rarely been included in published reports. This information is, however, of critical importance in the evaluation of apparently negative studies. The larger the number of persons in the exposed and control groups and the greater the similarity of these groups for factors other than exposure to the suspect carcinogen, the more likely will an effect be detected. Often, only a small number of humans exposed to a substance can be studied, conditions of exposure are inadequately defined, and records are incomplete. Thus a carcinogenic effect can be easily missed by epidemio- logic methods, especially when common types of can- cer (such as cancer of the lung, breast, colon, or rectum) are itudied, inasmuch as these types often require a large excess of risk before a causal relation- JNa. VOL 63. NO. I, JULY 979 197 Identiflcctlon of Potential Carcinogeno and Rick EstImatlorL 247 PAGENO="0202" 248 Interagency Regulatory Lialeon Group ship can be identified for the exposure to a particular substance. Substances distributed widely in commerce or in the environment are particularly difficult to study by epidemiologic methods unless high risk ratios are observed, because it is often impossible to identify unexposed groups as controls or to separate groups with high and low exposure. The problem of adequate controls is further compounded by the long latency of cancer, during which multiple opportunities exist for exposure to other potentially carcinogenic substances and modifying factors. The effects of such other exposures on rates of cancer are rarely known, al. though in some instances they were found to be more than additive (22). DIsease AscertaInment Because the effect under consideration is cancer morbidity or mortality, it is important to establish the validity, consistency, and reliability of the methods used to ascertain that neoplastic disease is clinically present or that it causes death. Disease classification is also important, and uniform criteria of tumor nomenclature are needed. Some types of cancer may be classified under a generic name in such a way that changes in their frequency may be missed if only the generic classification is used. Some members of a population may be "lost" to a study if their disease conditions cannot be adequately ascer- tained. Specific uniform procedures are not recommended here, but careful attention needs to be given to the extent to which these problems may affect comparison of relevant characteristics between groups. In the statistical evaluation of cancer incidence or mortality differences, there has been a strong tendency for particular confidence levels (e.g., 95%) and particu- lar probability values (e.g., P'0.05 Or P'O.Ol) to be used as standard points for a finding of statistical significance. It is recognized that probability values fall along a continuum and should be so reported. The uniform use of a standard probability value is not suggested. Regulatory needs are best served by accurate estimates of the possible role of chance in accounting for observed differences. The most important parameter in the assessment of an epidemiologic study is the magnitude of the effect measured; its interpretation is tempered by considera- tions of biologic plausibility, bias, confounding fac- tors, and chance. EVIDENCE FROM EXPERiMENTAL ANIMALS Evidence of the carcinogenic ;ctivity of an agent can be obtained from bioassays in experimental animals showing that the test substance causes either an in- crease in the incidence of neoplasms or a decrease in the latency period. The experimental design and conduct should be reviewed for quality and accuracy, and the results should be evaluated statistically for significance, with JNCt. VOL 63. NO. I. JULY t979 198 the only major experimental variable between control and experimental groups being the presence of the test substance. Positive results observed in more than one group of animals or in different laboratories and the demonstration that the occurrence of neoplasms fol. lows a dose-dependent relationship provide additional confirmation of carcinogenicity. Determination that a causal relationship exists between a test treatment and the responses observed in a bioassay is a complex judgmental activity that includes evaluation of the identity of the test agent and the biologic test system, the conditions of exposure, the methods of observation, and the qualitative and quantitative nature of the pathologic response. The assessment of cardnogenicity therefore relies upon the judgment and experience of professionals. The following discussion refers to ~ pects of experimental design and conduct that concern evaluation of results. They are not intended as a prescription of protocols. CrIterIa for EvaluatIon of ExperImental DesIgn and Conduct ExperImcntal Design Commonly recommended requirements for a thor- ough assessment of carcinogenic potential in experi- mental animals generally include a) two species of rodents, b) both sexes of each, c) adequate controls, d) a number of animals sufficient to provide an adequate resolving power to detect a carcinogenic effect, e) treatment and observation extending to most of the lifetime of the animals at a dose range including one level likely to yield maximum expression of carcino- genic potential. f) detailed pathologic examination, and g) statistical evaluation of results (9-25, 27, 31, 32, 57, 59-73). Positive results obtained in one species only are considered evidence of carcinogenicity. Positive results in more limited tests (e.g., when the observation period is considerably less than the animal's lifetime), but by experimentally adequate procedures, are acceptable as evidence of carcinogenicity. Negative results, on the other hand, are not considered evidence of lack of a carcinogenic effect, for operational purposes, unless minimum requirements have been met. Choice of the Animcl Model The animals used most often for carcinogenesis bioassays are mice, rats, and hamsters. These animals are used extensively because 1) their natural life-spans are short; 2) they are easier to breed and handle in large numbers than larger animals; 3) they are inex- pensive and easy to care for; 4) inbred strains exist that are genetically homogeneous for such traits as "back' ground" cancer rates, susceptibility to carcinogens at specific organ sites, longevity, and response to hus- bandry systems. Adequately designed and performed studies in other mammalian species may also provide useful information on carcinogenicity. For human risk evaluation, data obtained from bioassays with the use PAGENO="0203" of nonmammalian species can presently provide only suggestive evidence if positive but permit no cotsclu- sion if negative. Experience on the background incidence of tumors in the colony of animals used for testing, obtained over a period of years by extensive observation of untreated animals under the same general maintenance condi- tions (historical colony controls), is useful in assessing the relevance of experimental findings, such as the appearance of rare tumors. Rodents with different types of genetic homogeneity have been used for carcirsogenesis bioassays. These include a) inbred strains, it) first-generation hybrids of parents of inbred strains, c) randombred animals frÓm a closed colony, d) noninbred animals, and e) animals of unspecified strains or origins. As the genetic and/or environmental variation increases, so does the need for concern about the variation of background tumor in- cidence. A particular problem is posed by the use of certain strains of rodents in which particular tumor types reach a high frequency, often well above 50%, in untreated controls. Examples of such strains include mice of sirain A for lung adenomas, strain AKR for lymphomas, strain C3H/HeN males for liver cell - tumors and C3H females for mammary tumors, and females of several rat strains for mammary fibroade- nornas. Although viral factors have been identified in the etiology of mouse AKR leukemia and CSH mam- mary tumors, no such factors are known to be at work for the othertypes mentioned above. The effect of car- cinogens has been clearly demonstrated in all of the above strains by detection of substantial decreases in the latency period, by definite increases in incidence or multiplicity of these tumor types, and by the induction of tumors of other histologic types in the same or other organs (2-25). Caution must be used, however, in evaluating the significance of a higher incidence of these tumors in a treated group compared with concur- rent controls when the incidence in the treated animals falls within a range commonly seen in historical controls from the same colony. Background incidence rates for tumors of the lung, liver, mammary gland, and hematopoietic tissues are much lower in many other strains of mice, and for tumors of the mammary gland in other strains of rats. tn these other strains, no unique biologic trait distinguishes the types of tumors mentioned above from many others, and no reason has been demonstrated for considering that they have any dif- ferent significance than tumors in other organs as indicators of a carcinogenic response, under otherwise appropriate test conditions. Number of Animals The number of animals in each group to be effec- tively considered for the evaluation of carcinogenesis test results is the number in which detection of carci- nogenic effects could be expected. This number :j5 obtained by subtracting from the number of animals started on the test the number of those lost to adequate observation (e.g., by intercurrent death followed by cannibalism or autolysis). The number of animals on which complete pathologic examination is conducted is important in the evaluation of tumor pathology. Positive results can be obtained in tests with the use of a small number of animals if the test is otherwise adequately designed and conducted and if the tumor response is significant. For example, in a group of 15 animals, if 12 show a syell-defined neoplastic lesion of a kind rarely seen either in matched or historical controls, the finding is positive. However, a negative finding in a group of 15 animals is not adequate evi- dence that the test agent is not carcinogenic. Ideally, the number of animals required to provide adequate negative evidence would be such that an excessive risk would not arise if the test failed to detect carcinogenicity. The likelihood that such a risk would not arise increases both with the number of animals on test and the extent to which human exposure levels are exceeded. The probability of a false negative finding also depends on the background tumor rate in the control animals. For example, if a one-sided level of statistical significance of 5% is used with 55 animals, there is an 80% chance of detecting a tumor rate of 20% in the treated animals for whom the control rate is 5%, whereas 130 animals are required to detect the same difference if the control rate is 30%. The number of animals tested may need to be increased if the number of humans exposed is large or if a small margin of safety exists between the animal dose and the human exposure. In practice, resource limitations often require a trade-off between the number of animals used and the number of substances tested in order to control the total cancer burden resulting from chemical carcino- gens. This is particularly true with substances whose toxicity limits the test dose to a lose multiplicity of human exposure levels. In those instances, it may be necessary to accept a lower than ideal degree of "negative evidence." Route of Administration A key factor in the comparison of an experimental result to the human situation is to assess whether cells capable of malignant transformation are exposed to the reactive carcinogenic agent(s) in both the human and the experimental animal, regardless of whether trans- formation occurs in identical organs and cell types. Although this comparison is most readily made from experiments with animals in which the route of administration is the same as that in humans, other routes of administration may also be comparable and provide results useful for evaluation of the human hazard. For example, some chemicals are rapidly ab' sorbed by inhalation, circulated through the body, and metabolized by the same pathways that occur following intravenous exposure (74). Some routes of administration in animals may fail to provide adequate metabolic activation or exposure of 199 identification of Potential Carcinogens and Rick Estimation 249 Jesci. voL. 63. NO. I. JULY 979 PAGENO="0204" 250 interagenCY Regulatory Unison Group 200 target tissues and therefore may lead to false-negative results. This possibility should be assessed in evaluat- ing negative results obtained when the route of admin- istration in animals differs from the route of human exposure. Generally, the route should be one that leads to absorption and distribution of-the test substance. The induction of tumors at a remote site in the animal is evidence of absorption, distribution, and possible meta- bolic activation of the test substance. If exposures of both humans and animals involve absorption of the substance, any route of administration in animals may be regarded as relevant for a qualitative demonstration of human hazard unless there is evidence that the route of administration in the test species results in the production of carcinogenic substances (from degrada- tion or metabolism) which does not ever occur with human exposure. When tumors appear only at the site of injection or implantation, careful review is necessary. If there is reason to believe that the tumors occur as a result of "solid state" carcinogenesis (75, 76), the results may be inappropriate for extrapolation to human exposure. If, however, the test material produces tumors at the site of injection or implantation as a result of its chemical reactivity, this response is an indication of carcino- genicity. There are a number of practical reasons for studying certain substances in animals by a route of administra- tion different from the expected route of human exposure. If a substance under test is highly volatile, accurate administration in food may be difficult be- cause of evaporation; often feeding through a stomach tube is used so that the dose may be measured with greater accuracy. Es-en for nonvolatile test substances, a stomach tube may be uted when it is important to know the exact amount of a substance administered to the test animals. The administration of high doses of a test substance with a disagreeable odor or taste may require the use of routes other than ingestion. Thus experimental exposures need not necessarily be by the route of human exposure in order to be meaning' ful, but possible physiologic and metabolic differences related to routes of absorption and distribution should be considered in the assessment of their relevance. Identity of the Substance Totted Substances to which humans are exposed through their occupations, the environment, and the products they use vary widely both in the number and the proportion of contaminating impurities. A full assess- ment of the carcinogenicity of an impure mixture ideally requires that each component be tested indi- vidually at an adequate dosage and that the mixture itself be tested in order to detect cumulative or syner- gistic effects. Limitation of resources makes this ideal approach impractical as a routine. It is common, therefore, simply to rely on tests either of the product to ss'hich humans are exposed, including the impurities present, or of the purified principal chemical sub- stance(s). Because the products may vary according to procedures used in manufacture and processing, tests for one commercial product may not be applicable to -another product containing a different set or level of impurities. Change in the manufacturing process of a product may require additional tests to confirm the safety of the new product if the change involves the introduction of different impurities or a substantial increase in the amount of any single component of the product. Even though it is accepted practice to test mixtures, the nature of any impurities known or likely to be present as a result of the manufacturing process is important and may require separate examination or testing. Information on the carcinogenicity of any single chemical in a mixture is an indication of potential hazard of the entire mixture. However, nega- tive results obtained on a component of a mixture may not reflect the potential carcinogenicity of the entire -mixture. Doze Levels "Testing should be done at doses and under experi- mental conditions likely to yield maximum tumor incidence." This recommendation of an FDA advisory committee summarizes the issue of test doses (68). Bioassays with the use of a few dozen or even a fess hundred animals have relatively lose sensitivity for detection of carcinogenic effects. Millions of people of varying degrees of sensitivity or exposure may be exposed to the substances under evaluation. Although a test animal cannot be strictly viessed as a "surrogate" of a large number of people without oversimplifica- tion, the role of animal tests is to pros-ide maximum detectability of carcinogenic effects within the already narrow confines of test sensitivity. Under otherwise identical conditions, the greater the ratio of test expo- sure to human exposure, the greater is the safety margin provided by a negative result in a carcinogene- sis bioassay. It is generally recommended that more than one dose level be tested. Most carcinogenic effects show a posi. tive dose-response relationship, but maximum tumor incidence in test animals may not occur at the highest dose when competing toxicity prevails. The highest test dose that can be effectively used in a cardnogenesis bioassay is limited by the conditions of absorption, by the amount that the animal can tolerate during life- time administration without unwanted toxic side ef- fects, and by the effects on nutrition sehen the chemical constitutes too large a proportion of the diet. Results of bioassays done at doses and under condi- tions permitting maximum expression of carcinogenic- ity provide a sound basis for the identification of a car- cinogenic hazard or its absence. It is important to estimate the highest dose level that will be tolerated by the test animals during lifetime ad- ministration, i.e., the estimated maximum tolerated dose (EMTD). The EMTD is defined as the highest JNti. VOL 63. NO. t. jt1.v 1979 PAGENO="0205" dose that can be administered to the test animals for their lifetime and that is estimated not to produce a) clinical signs of toxicity or pathologic lesions other than those related to a neoplastic response, but which may interfere with the neoplastic response; b) alteration of the normal longevity of the animals from toxic effects other than carcinogenesis; and c) more than a relatively small percent inhibition of normal weight gain (not to exceed 10%) (71). The EMTD is determined on the basis of prechronic tests and other relevant information. If the test reveals that the EMTD is too high to meet the conditions defined herein, positive results obtained above the EMTD are acceptable as evidence of carcinogenicity unless there is convincing evidence to the contrary. Alternatively, negative results obtained above the EMTD are considered inadequate unless particularly strong and specific scientific reasons justify their accep. tance as negative. Positive results obtained at or below the EMTD provide evidence of cardnogenicity. Ago at Treatment Because of the long latency period required for induction and manifestation of tumors, treatment should be started in young animals, and the animals should be observed for a carcinogenic response through most of their expected life-spans. The older the age. at first treatment, the shorter is the remaining life.span available for tumor development; consequently, the smaller is the chance of detecting delayed carcinogenic * effects. Although treatment is often started in young adult animals soon after weaning, some protocols call for treatment soon after birth (neonatal) or during fetal development (transplacental). The rationale for expos. ing test animals transplacentally or neonatally is based on the greater susceptibility of certain organs to carcin- ogens during early development. Such susceptibility has been demonstrated in several species, including those commonly used for bioassays (77, 78). Animals firsi treated during the perinatal period must be also treated and observed throughout their life.spans to obtain a valid negative response. Virtually any agent that is carcinogenic in adult animals can be expected to have some carcinogenic effect when administered to young animals including the neonate and the fetus. Unless a substance is demonstrated to be exclusively carcinogenic when ad. ministered to the fetus or neonate, enhanced perinasal susceptibility to carcinogens should be considered not a separate and distinct toxicologic property; rather, it should be a means for increasing the sensitivity of conventional bioassay procedures by extension of the exposure period to these earlier and more susceptible portions of the life-span. It should be emphasized that these protocol modifi. rations greatly complicate dose selection and experi. mental design. An agent may be significantly more toxic to the fetus, the neonate, or the pregnant or lactating female animal than to the normal young adult of either sex. This requires independent deter. mination of the toxicity and EMTD. Furthermore, individuals in the litter of a treated pregnant animal cannot be considered independent units for statistical evaluation of effects. Conduct and Duration of Bioauaya in Animeia A long-term bioassay for carcinogenesis in animals is a complex procedure requiring control of many var- iables for several years. Professional experience and knowledge of the relevant biologic parameters are needed for adequate quality control. Detailed guidance on procedures is provided by reports such as the FDA's "Good Laboratory Practice Regulations" (79) and the NCI's "Guidelines for Carcinogen Bioassays in Small Rodents" (71). Review of the observations made during the bioassay (on food intake, weight, clinical course, and patho- logic conditions of the animals) provides a basis for determining whether these experimental variables are recorded in sufficient detail and are internally con- sistent to permit independent assessment of their valid- sty. The purpose of these bioassays is primarily to provide maximal opportunity for detection of a neo- plastic response; therefore, the longer the period of observation the better is the chance of detecting delayed effects. A "point of diminishing return" can be reached when intercurrent disease and/or survival considera- tions make the observation or evaluation of old ani- mals particularly difficult. It is expected that the animals will be observed for most of their life-spans. The best negative evidence for the carcinogenicity of a substance is obtained from tests in which both expo- sure and observation last through all or nearly all of the expected life-spans of the animals under study. Negative results decrease in value as the exposure and observation periods are shortened, and they be- come practically meaningless if these periods are shorter than half the life-spans of the animals, When some animals die early in the course of a test, the value of the test is reduced as a function of the percentage of animals dying without tumors at periods markedly shorter than the life-span of the species. Sometimes, a positive carcinogenic response may be definitely dem- onstrated in a shorter- period of observation if the experiment is adequately controlled; in such cases the test is considered valid even if it is shorter than usual (80). Accepted procedures include a) the observation of all animals in the study (treated and control groups) until their spontaneous death,.b) the sacrifice of animals that show clinical signs of severe illness or impending death (sacrifice of moribund animals prevents losses due to autolysis and provides better observation of tissue pathology), and c) terminal sacrifice at a sched- uled date near the end of the life-span (e.g., after 24 mo on test). 201 identification of Potential Cnrclnogena and Rlak Ectlmetlon 251 JN~. VOL 65. NO. I. JULY 979 PAGENO="0206" 202 252 InteragenCy Regulatory Lialeon Group Criteria for Evaluation of Pathology Pathology Examination The evaluation of carcinogenesis bioassay results rests on the extent and accuracy with which organs and tissues of both treated and control animals are ex- amined for morphologic changes. After the termination of a bioassay, the only physical evidence that can be used to permit reevaluation of results, even years afterwards, is represented by the written descriptive and diagnostic records, the graphic or photographic records of gross or microscopic observations, and most im- portantly, the original slides of tissue secticns for micro- scopic examination. The histologic slides are of critical importance as a lasting direct documentation of the conditions of normal and abnormal tissues and organs, both for scientific and regulatory purposes. Quality and extent of pathologic documentation are therefore major factors in establishing the validity of bioassays in animals (71, 79). Although a well-conducted pathologic examination cannot generally rescue a poorly designed or badly conducted bioassay, inadequate pathologic examina- tion can significantly reduce or eliminate the value of an otherwise well-conducted experiment. Among the factors to be considered in evaluation of the pathologic examination are: 1) the care and thoroughness of gross tissue ex- amination and the qualifications of the per- sons conducting this examination to recog- nize abnormalities; 2) the quality of preservation, sectioning, and staining of tissues; 3) the accuracy of the record-keeping system used for labeling tissues as they are moved from the animal through slide-processing to finai diagnosis and reporting; 4) the extent of selection of normal and abnor- mal tissues for microscopic examination; and 5) the qualifications of the pathologist making the microscopic examination. The numbers of tumors or other lesions diagnosed by the pathologist are not a thorough assessment of incidence unless each factor is adequately considered, controlled, and documented. The strength of evidence provided by a bioassay also depends on the number of tissues examined. Failure to observe excess tumors in treated animals cannot be considered evidence of the absence of a carcinogenic hazard unless all organs have been examined grossly and all grossly visible suspect lesions have been ex- amined microscopically. In a large organ, the taking of a single random section for histologic examination can result in failure to detect small tumors. Thus multiple cuts through such organs should be made. It is also important to open and search the entire cavity of all hollow organs for abnormalities. For example, the entire length of the.gastrointestinal tract should beopened and inspected. Grossly visible lesions should be selected for itistologic examination, and if they are not subsequently observed on tissue slides, preparation of additional sections may be necessary until the gross lesion is verified histologically. Furthermore, histopathologic examination should be made of major organs in the treated groups and matched controls, and specific organs should be studied in detail in all dose groups and controls in which there is either gross or microscopic evidence of lesions. Major organs are defined in the NU's Guidelines for Carcinogen Bioassays in Small Rodents" (71). Positive evidence of carcinogenicity may be valid for a particu- lar organ if it has been adequately examined in both treated and control groups. Negative reports are inade- quate for any organ that has not received careful gross examination in all animals and histologic examination of suspect lesions. The more limited the number of organs examined grossly and microscopically, the less the value of the experiment in providing evidence of a negative result. Evaluation of Pathologic Reoulta The evaluation of bioassay results and their quality requires a detailed review and expert judgment of all the experimental conditions and observations, includ- ing the identity of the test substance; the conditions of administration; the identity, source, and characteristics of the test animals; the accuracy and systematic record- ing of observations; the extent of pathologic examina- tion; and the competence of the investigators. Meticu- lous and detailed documentation is of great impor- tance. Several criteria are applied in the evaluation of bioassay results. I) Internal consistency of the data is important in reviesring the conduct of the test. Apparent incon- sistencies should be investigated by analysis of records. 2) Reproducibility of test results can be demon- strated within a single experiment (in different groups of similarly treated animals or in different dose-level groups) or in separate bioassays conducted with the same experimental design in the same or in different laboratories. Evidence of reproducibility adds greater confidence to the evaluation of results. Statistical considerations provide an estimate of the level of detectability of an effect and the consequent level of probability that the effect may be missed in a repetition of the test in a given number of animals. Apparent contrary results in any two tests may be simply an effect of chance variation and may be fully compatible with an identical mechanism and level of activity of the test compound. 3) Evidence of a positive dose-response relationship adds further confidence to the evaluation of a positive test, but lack of it may be due to testing in a portion of the dose-response curve with a shallow slope or even with a declining slope due to competing risks. In the presence of positive results in well-designed, well- conducted tests, evidence of reproducibility and posi- tive dose-response relationships is not necessary to reach a conclusion of carcinogenidty. Jsu. vot at. NO. t. it-I_v sin PAGENO="0207" 4) Concordance of results obtained under differin~ test conditions (e.g., different species, different routes of administration, or markedly different basal diets) pro- vides greater confidence in the evaluation of both positive and negative studies, but it has a different meaning from "reproducibility" within the same tests or under the same conditions. Lack of concordance from tests performed under different conditions does not, in itself, detract from the validity of the positive test. Reasons for a discordance in observation may be identified by evidence obtained during a test or may be sought through further research. The response to carcinogens in different animal species and even strains is known to vary greatly because of genetic, metabolic, nutritional, and other factors that affect susceptibility in a given test animal. Present knowledge indicates that a substance that is clearly carcinogenic in one test species is likely to be carcinogenic in other species, that it may take extensive tests in several species to demonstrate this correlation, and that the responsive target tissues or organs and the types of tumors induced in different species may vary gr~atly. Therefore, although concordance of positive results (even if different tumor types are involved) adds support to an evaluation of carcinogenicity, the find- ing of negative results in some other species generally dpes not detract from the validity of a positive result as evidence of carcinogenicity for the test substance, In this respect, positive results supersede negative ones. The assessment of such apparent discrepancies in results requires consideration of all experimental vÔr~ iables, since apparently negative results may derive from limitations in the sensitivity of the test (e.g., early scheduled sacrifice, limited extent of pathologic exam- ination, and statistical probability). If the positive result is itself not fully conclusive or if reasons exist for questioning its validity as evidence of carcinogenicity, the result is generally classified as "inconclusive" or "only suggestive" even in the absence of other negative test results. b) Evaluation of tumor incidence is made on the basis of thepathologic findings and therefore depends on professional diagnostic judgment. Tumor incidence is evaluated `by consideration of all tumors of specific organ sites or anatomically or physiologically related systems. At present there is considerable uncertainty about the interpretation of carcinogenic responses in terms of the total tumor yield in contrast to the response in terms of a statistically significant increase *of tumors in specific target organs or tissues. Ti-a- ditionally, carcinogens have been recognized in studies on humans and animals by a decisive increase in tumors of target organs. However, it is conceivable that a general increase in total tumor yield, in he absence of an excess incidence in one or more target tissues, could occur-for example, by a promoting effect that generally increases the spontaneous incidence of tu- mors in test animals or by the action of a multipotent carcinogen whose response did not reach statistical significance in any one organ even at the maximum tolerated dose. Its some instances, however, control animals may have a high frequency of tumors at certain sites (e.g., testicular tumors in F344 male rats), In such instances, a simple cumulative count of tumor- bearing versus tumor-free animals may fail to reveal carcinogenic effects in the treated groups, Prudent judgment is needed on the appropriate categorization of tumors used to evaluate induced effects. A positive' result in a carcinogenesis bioassay can be based on evidence of the induction of an increased incidence or a substantially decreased latency period, The latter is often difficult to establish. Determination of the latency period can be made by various tech- niques of observation during a bioassay, If both test and control animals are sacrificed at a fixed time, only the early part of a temporal distribution curve may be observable; consequently, the estimate of the average latency period for all tumors or tumor-bearing animals may be artificially altered. If the test and control groups are allowed to live out their life-spans, the comparison of latency periods must take info account the relative survival and the number of animals at risk, particularly in the case of competing risks. The methods used in estimating the latency period must be defined in the context of each bioassay. It is always difficult to determine the exact onset of a neoplasm. Morphometric criteria may be used for tumors (e.g., skin or subcutaneous tumors) detectable during clinical observation of the animals and a minimum size may be established as a criterion for identification. For neoplasms of the internal organs it is practically impossible to determine an adequate time of onset: Methods such as palpation of the abdomen are highly subjective and generally unreliable. Serial sacrifice studies provide excellent data on time to tumor induction, but they should not be substituted for adequate numbers of animals under lifetime observa- tion. In most instances, what is referred to as latency period is the time between the beginning of the exposure and the observation of a tumor at death. This parameter is obviously influenced by all the factors that determine time of death, e.g., intercurrent dis- eases, other tumors, or growth rate of individual tumors. Here too, the judgment of experienced pathol- ogists may provide critiral evaluation of such aspects as tumor size, location, cell differentiation, and inva- sion; these factors may contribute to an estimate of temporal sequence The observation in treated groups of tumors that are considered rare in untreated and historical controls may raise considerable suspicion even when their incidence is below the required level of statistical significance, Careful review and cautious judgment are necessary in their evaluation; often the rarity of a tumor type is estimated on the basis of a small control population. The occurrence of one or a few neoplasms of a kind, however rare, is not necessarily evidence that a substance is carcinogenic in the absence of other supporting evidence. 6) Evaluation of tumor morphology in the final analysis of bioassay results is highly dependent on the way in which pathology data are categorized. It is JNa, VOL 63. NO. I. JULY 19t9 203 Identlllcatlon of Potential Carcinogens and Risk Estimation 253 PAGENO="0208" 254 Interagency Regulatory Llalaon Group 204 incorrect, for example, to subdivide diagnoses into so many individual categories based on different stages of disease or different morphologic features that no single category is large enough to be statistically significant. At the other extreme, it is incorrect to group unrelated end points in a way that maximizes the opportunity to find statistical significance, whether or not such group- ings are biologically meaningful. Carcinogenic and chronic toxic effects of a chemical on an organ, tissue, or cell develop through a series of stages from minimal changes to advanced and possibly fa;al end points (81). The stage reached at any particu- lar time is related to the dose of the substance, the conditions of exposure, the time elapsed since be- ginning of exposure, and host susceptibility factors. Early lesions that are pathognomonic of a disease process resulting from toxic chemicals should be grouped with more advanced lesions, whether or not the animal has survived long enough for the process to develop to the latest stages. The carcinogenic process may go through early stages including atypical hyper- plasia, carcinoma in situ, and/or histologically benign tumor before progressing to a clearly malignant stage. Although the stage of development is of critical im- portance in clinical oncology for assessing the prog- nosis of a patient at the time of therapy, it is not relevant in deciding whether a chemical is capable of inducing cancer as long as the induction of lesions recognized as neoplastic is conclusively demonstrated. The induction of preneoplastic lesions in the process of cancer development is an indication that the test substance is capable of inducing cancer in a suscep- tible host given sufficient exposure and tiree for cancer to arise. Care must be taken, however, to distinguish atypical hyperplasias that are pathogriomonic of neo- plastic progression from other nonspecific or reactive hyperplasias. In the evaluation of bioassays, the concern is with the capability of a test substance to react with a biologic system to give rise to a neoplastic response that may develop through all stages to malignancy. One issue is whether or not the response is the kind that stops at the benign stage and never evolves further to the invasive and metastasizing stage. Few if any tumor types are presently known to belong to this category, which could be called "permanently benign" tumors. For benign tumors, no specific mechanism of induction is known that can be distinguished from the mechanisms of induction of other neoplasms. More- over, no established body of evidence exists showing that certain substances or groups of substances are capable of inducing exclusively permanently benign tumors without ever inducing any more malignant ones. The mammary fibroadenoma isgene.-ally con- sidered to be a benign tumor in both the human (82) and the rat (83), and it has been suggested that its ex- perimental induction provides little evidence that the inducing agent can cause cancer. X-rays or carcino- genic polycyclic aromatic hydrocarbons, however, which principally induce fibroadenomas in some rat strains, induce mostly malignant adenocarcinomas in other strains; the genetic characteristics of the animal rather than the inducing agent determine whether benign or malignant tumors develop (84). Thus the induction of benign tumors, even of a type that rarely progresses to a malignant stage, must be considered a warning that the inducing chemical may be capable of causing cancer in some humans. The induction of benign neoplasms, even if they were demonstrated to be of a permanently benign type, would therefore be considered evidence of carcinogenic activity unless definitive evidence is provided that the test chemical is incapable of inducing malignant neoplasms. Neoplasms at a benign stage may jeopardize the health and life of the host. Furthermore, it is ex- tremely difficult to rule out the presence of malignant changes simply on the basis of a limited histologic ex- amination of the primary tumor, because focal malig- nant change or local invasion may have occurred in other areas of th~ tumor that were not examined microscopically. Similarly, it is very difficult to rule out the metastatic spread of a neoplasm that may be biologically capable of metastasizing without an ex- tremely detailed search for metastases, which can begin as small foci of one or a few cells lodged in the arteriolar walls of peripheral organs (85). The fre- quency of observation of such metastases depends directly on the amount' of peripheral tissue that is examined (86). Another case to be considered is the combination of neoplasms diagnosed as benign and malignant. This may include instances in which the incidence of histologically malignant tumors is only a relatively small fraction of the total tumor incidence but repre- sents the most advanced stages of the neoplastic re- sponse. Although the number of tumors diagnosed as malignant may not reach statistical significance as such in the number of animals at risk, the total neoplastic response (benign and malignant) may be clearly significant. Some common types of neoplasms found in carcino- genesis bioassays in laboratory rodents are among those often diagnosed as being at a benign stage when observed in test animals. Examples inclttde lung ade- nomas, skin and bladder papillomas, liver cell ade- nomas (hepatomas), and hemangiomas in various or- gans. All of these tumor types are known to progress to frank malignant stages. No pathogenetic mechanisms have been identified thatcould demonstrate that the induction of such tumors, whether in a benign or malignant stage, in otherwise appropriate, comparable, and well-controlled experimental conditions, provides any different kind of evidence for carcinogenesis than - the induction of other tumor types. In the evaluation of tumor incidence, therefore, neoplasms in~different stages of progression are counted together. 7) General evaluation of neoplastic pathology for - carcinogenesis bioassays includes consideration of the - total number of animals with tumors in each group, the total number of individual tumors, and the index of wmor multiplicity in tumor-bearing animals. The tumor response can be further characterized by a jscs. rot., 65. NO. I. JULY sm PAGENO="0209" detailed observation of the tumor morphology and related preneoplastic changes. The extent of tumor growth and spread and special morphologic character- istics may give useful indications of the time of develop- ment of the neoplastic response The quality of the pathologic response is determined by a comprehensive evaluation of all the pathologic changes observed in both treated and control animals. Special attention is required in the evaluation of toxic effects other than carcinogenicity, because their pathologic manifesta- tions have to be distinguished from those due to the neoplastic response. The organs and tissues that are the targets of carcin- ogens may vary greatly in different species and even under different exposure conditions; therefore, no di- rect analogy of morphologic response can be expected from a carcinogen in animals of different species and in humans. Examples are known both of widely different target sites [e.g., benzidine induces bladder carcinoma in humans and cholangiomas and liver cell carcinomas in hamsters and rats (87)] and of similar responses [e.g., vinyl chloride induces the same type of angiosarcomas of the liver in humans, rats, and mice (88)]. Special conditions of tissue exposure or reaction may result in a tumor response by mechanisms that appear due to physical rather than chemical properties of the test material. The following conditions-are evaluated differently in this respect: a) The induction of sarcomas around a "solid state',' implant of the test substance into a connective tissue is not considered an indication of the carcinogenicity of that substance when it is administered in another physical form (75, 76). b) The induction of a carcinogenic response by asbestos and other fibrous materials by a mechanism linked to certain physical characteristics such as fiber length and diameter is recognized as a basis for categorizing the exposure to such fibrous materials as a carcinogenic hazard (22). c) The effect of particulate materials in the induction of respiratory neoplasms, when they are administered jointly with certain carcinogens (probably through their capacity to absorb and retain carcinogens, to penetrate the respiratory tract tissues, and to stimulate early cellular responses) is not recognized as evidence of carcinogenicity of these substances but rather as an indication of their role as cofactors in carcinogenesis; particulate materials require careful but separate con- sideration as a potential hazard (89, 90). d) The induction of a neoplastic response by a sub- stance because of its radioactivity is recognized as a cancer hazard. Other factors are sometimes suggested to be sufficient to refute the presumption of positive evidence of a car- cinogenic effect. These factors must be critically ex- amined to avoid false.negative judgments based on unsubstantiated hypothetical explanations of the cir- cumstances of tumor induction. The following factors are considered in this respect: a) Indirect mechanisms of action requiring special exposure levels or conditions. An example has been suggested in the case of substances that may induce bladder neoplasms only in the presence of bladder stones resulting from high levels of intake and urinary excretion of the test substance (91). Support for such a mechanism as an explanation for development of bladder tumors is provided by determination of a specific association of tumors with stones, a dose- response correlation between stones and tumors, and the absence of other chemical or biological indications that the substance might be carcinogenic by other mechanisms. In evaluation of the relevance of such ex- perimental observations to the assessment of human hazard, special consideration is needed for mechanisms by which exposures or intercurrent diseases in the human may act as the cofactor (e.g., in bladder stone induction), thus producing a susceptible state for the possible carcinogenic activity of the test substance. b) The action of promoting agents only on tissues previously initiated by carcinogens (51, 52). Few ex- amples are well documented, such as the phorbol esters in epidermal carcinogenesis in mice. Criteria of risk evaluation need to be defined and dose-response rela- tionships considered. Any claim that a substance acts only by this mechanism and thus is of less concern to humans needs to be supported by experiments showing the mechanism of action and demonstrating that the effect does not occur at human exposure levels. c) Metabolic pathways of carcinogen activation (92). which are suggested as occurring exclusively under certain test conditions in experimental animals but not under other test conditions or in other species. This situation would be important if thorough studies demonstrate that the metabolic pathways for carcino- genic activation of a substance in animals do not occur in humans. Another important situation would be the demonstration that the metabolic pathways of activa- tion of a particular carcinogen identified by studies at high levels of exposure are exclusively formed at such high levels but are absent at lower dose levels. Statistical Analysis of Results Statistical hypothesis testing provides an estimate of the likelihood that an experimental observation may or may not be a result of chance alone. The 95% confi- dence level is widely accepted as a reasonable assurance that the observed effect is real, but confidence that an increased incidence of tumors is a true indication of the carcinogenicity of a substance increases with in- creasing statistical significance of the results. Thus the level of statistical significance should be reported rather than the fact that a result is statistically signifi- cant or not significant at a singlš preassigned level of confidence. Failure to detect an increase of tumors in a bioassay may be due to an insufficient number of ani- mals tested and does not unequivocally prove that a substance does not pose a risk of cancer. Tumors rarely seen in experimental animals may raise considerable suspicion even if the statistical sig- nificance is well below the 95% confidence level. JNCI. VOL. 63. 50. I. Jt'Lv 979 205 ldentlflcetlon of PotentIal CarcInogens and RIsk EstImation 255 22-143 0-83---14 PAGENO="0210" 206 256 Interagoncy Regulatory Liaison Group Because of the frequent use in chronic studies of both sexes, more than one species or strain, and more than one dosage level, and because many different tissues are examined, a large number of statistical comparisons are possible between control and treated animals. Thus the results from a chronic study must be interpreted cautiously to control the rate of false positives arising from the large number of possible statistical comparisons (93). Lifetime animal experiments are often difficult to interpret because of competing causes of death, which may alter the pattern of the observation period of the tumor type under study. A common but inadequate form of presenting tumor data is a report only of the proportion of animals in which particular tumor types were observed during the study. This proportion may contain a mixture of three types of observations: I) The tumor causes the death of an animal and is subse- quently observed upon necropsy; 2) an animal dies due to some cause other than a particular tumor and the tumor is observed upon necropsy; or 3) the tumor is observed when an animal is necropsied at the time of a scheduled sacrifice, generally at the termination of an experiment. Simply combining tumors observed under these three situations makes interpretation difficult, and in fact the data may be misleading if the mortality pattern is altered by the toxicity of the substance. Serial or terminal sacrifices pros-ide an opportunity to compare the prevalence of tumors in various groups of animals unperturbed by mortality. However, sacri- fice data do not pros-ide an opportunity to study the effect of a substance on survival or on causes of death. The analysis of a bioassay is limited by the quantity and quality of data. Such studies must include the age of each animal at the beginning of the experiment, its age at time of removal from th.~ -periment, reason for removal (death, moribund conditio., scheduled sacri- fice, or others), and all clinical and pathologic obser- vations, including gross and microscopic examination. When survival curves of control and treated animals differ due to competing causes of death, adjustment of the number of animals at risk may be necessary. For a tumor type generally leading to the death of an animal, statistical analyses of survival experiments should incorporate life-table or competing risk tech- niques in order to estimate and test tumor incidence. This approach requires assumptions concerning the independence of the competing causes of death. If all the animals are utilized from a survival study, includ- ing sacrificed animals, the net probability of death due to a tumor type can be estimated as though that were the only cause of death of a group of animals. Sta- tistical tests for differences between control and treated groups can be performed on the adjusted tumor inci- dence rates (94-96). For a tumor type that is unlikely to kill the animals, methods of analysis based on life-table techniques are not appropriate for adjusting the number of animals at risk. These tumors are observed conditionally as a result of other events occurring first: death of the animal or a scheduled sacrifice. To estimate the preva- lence rate of these tumors, mortality is assumed to be unrelated to the presence of the tumor. Statistical methods for the analysis of tumors that are not generally life-threatening are discussed by Hoel and Walburg (94) and Peto (95). SHORT-TERM TESTS FOR CARCINOGENS Carcinogenesis tests have traditionally been based on the experimental induction of tumors in laboratory animals. Such tests usually involve the observation of treated animals for most of their life-spans. Recently, short-term methods have been developed to provide more rapid markers for the tentative identifica- tion of carcinogenic effects. These methods are di- rected toward the study of mechanisms underlying neoplastic transformation as well as toward provision of reproducible and rapid methods for testing chemi- cals and physical agents for potential carcinogenic activity. The use of short-term methods for the evalua- tion of carcinogens was the subject of a recent review' (97) from which the following discussion is largely derived. Methods Based on Genetic Alterations The analysis of mutagenic effects has been developed mainly to assess the ability of a substance to induce genetic alterations. The resulting information can be used for estimating the genetic hazard of chemical agents for man. Because of the similarities of basic molecular mech- anisms by which chemical mutagens and most chemi- cal carcinogens appear to induce genetic effects (i.e., molecular alterations of DNA), it has been postulated that mutagenic effects can be used to predict cardno- genicity. The use of a battery of short-term genetic tests is usually recommended in order to minimize false- negative and false-positive results and to select com- pounds that require further long-term investigations. This battery of tests may include: a) tests for mutations in bacteria and eukary- otic microorganisms; b) tests for mutations in somatic mammalian cells; c) tests for effects on chromosomes in higher eukaryotes, including mammals; d) evaluation of DNA repair synthesis. For screening purposes, preference has usually been given to tests that have already been validated with a large sample of compounds belonging to different chemical groups. Among the mutagenicity tests on microorganisms, the one most widely used and validated is the Ames reversion test in Salmonella. Tests in Escherichia coli, Saccharomyces, Neurospora, and Aspergillus are also being used. Mutagenicity testing is also being con- ducted in Drosophila. Several other methods currently being evaluated may be used to monitor genetic damage in mammalian cells JMS. VOL 63. NO. t. JPLY t979 PAGENO="0211" 207 identification of Potential Carcinogens end Risk Estimation 257 by carcinogens in vivo and in vitro. These methods given to the effectiveness of metabolic activation lunc- include the production of sister chromatid exchanges ions in each test system used. as well as measurement of the induction of direct damage to DNA and its subsequent repair. Evaluation of Short-Term Test Results Various short-term mutagenesis tests, some of which are used to provide supportive evidence of carcino- The study of carcinogenesis at the cell level presently genicity, are discussed in (98). offers an effective means to identify carcinogenic effects and mechanisms. In vitro mammalian cell transforma- Methods Based on Neoplastic Cell Transformation tion systems are simple models for the study of the mechanisms of chemical and physical carcinogenesis: Several systems are now available at the mammalian As these systems bes~'ome more widely used as test cell level for the identification and study of substances methods, they will lead not only to better develop- that represent a possible cancer hazard (99). ment and definition of screening techniques but also to In recent years a number of systems have been, better understanding of the underlying mechanisms of developed to test for neoplastic cell transformation by. carcinogenesis. chemical and physical carcinogenic agents. Some of Short-term tests for chemical carcinogens presently these systems are being used in several laboratories do not, in the absence of animal bioassays and epide- with good reproducibility; other systems are still being miology data, constitute definitive evidence that a sub- developed. Those that have been most seidely studied stance does (or does not) pose a carcinogenic hazard to are a) the golden hamster embryo cell system and b) humans. However, positive responses in these tests are the mouse embryo fibroblast cell line systems. considered suggestive evidence of a carcinogenic In the golden hamster embryo cell system, primary hazard. and/or secondary cultures of normal embryo cells are Such positive results also supply supporting evidence used. Transformation is determined 7-10 days after to positive animal bioassays or epidemiology results. In treatment of cells seeded for colony formation. Quanti- some instances results from short.term tests may con- tation is based on the frequency of morphologically flict with animal bioassay data. If an animal bioassay altered colonies, shows a positive response, it cannot be dismissed In the mouse embryo fibroblast systems, established because a negative response was observed in these tests. homogeneous cell lines are used. Thus cloned popula- However, positive responses in such short.term tests are tions of cells can be grown in large quantities and used ordinarily sufficient to provide suggestive evidence of by many laboratories. Transformants are identifiable carcinogenicity, even if the substance tested has shown 4-6 weeks after exposure to the carcinogen. They may only negative responses in some animal bioassays. As be scored quantitatively by morphologic criteria (focus the degree of certainty attached to the negative re- assay), which correlate highly with tumorigenicity in sponses in animal bioassays increases because the. animals. Among these established lines, the CSH 10Th observation is reproduced in other animal species and Clone 8 cell system has been the most widely studied. strains or under more rigorous test conditions, the In these tests for neoplastic transformation, the cells suspicion about the chemical as a result of short.term derived from transformed colonies or foci, when inocu- tests may be reduced and eventually eliminated. These lated into syngeneic or immunosuppressed animals, conclusions are in accord with those of the National can grow as malignant tumors. Although the definitive Cancer Advisory Board's Subcommittee on Environ- evidence for neoplastic transformation of cells in cul- mental Carcinogetsesis (57): ture remains their tumorigenicity in animals, a number As the present, none of the short-term tests can he used of phenotypic changes of the cultured target cells are to establish svhether a compound taut or tisti not he carci- commonly used as indicators. nogentr so humans or experimental animals. Positive Other in vitro systems are being developed with the results obtained in these systems suggest extensive testing use of spec I ed cell types u h as ep th I I cell h cIt g It te rn m I b sa pa f om I p d m nd o h o g n Neopla t h se m h d tabl h h ml y t h transformation of well-defined epithelial cells by chem- . icals has been achieved in vitro; conditions for quanti- tative studies are under development. Such systems may be needed to identify critical target cell populations MOLECULAR STRUCTURE AS SUPPORTING EV1 within target tissues closely correlated with carcinogen- DENCE IN IDENTIFICATION OF CARCINOGENS esis in vivo. To be effective, most chemical carcinogens require Information useful in identifying possible carcino- metabolic activation by cell enzymes to an ultimate gens is provided by their molecular structures. Ii is reactive metabolite. In mammals metabolic activation , well established that certain groupings of atoms (func- of carcinogens takes place in many organs and tissues. tional groups) in some molecules may import carcino- Cells in culture can retain enzyme activities, but genic properties-e.g., some polynuclear aromatic sys- specific culture systems or preparatiotss may lack or tems, hydrazine groups, N-nitroso groups, and a, lose the enzyme activity necessary to activate certain p-unsaturated lactones. There is a moderately sub- chemicals. Therefore, adequate consideration should be . stantial base of empirical data that permits conclu- JNCI. vOL a. NO. I, jVtX 5979 PAGENO="0212" 208 258 interagency Regulatory Liaison Group sions about carcinogenic potential on the basis of molecular structure (33, 100). Similarly, some functional groups have never been shown to impart carcinogenic properties to molecules, although the data base for such negative correlations is much smaller and probably inconsequential. The rea- son for the absence of a strong empirical data base for noncarcinogens is that structure has frequently been used as a guide to testing chemicals for carcinogenicity, and priorities for testing have often been based on the suspected cancer-inducing properties of chemicals. In some instances, the predictive posser oI molecular structure of functional groups known to be correlated with carcinogenic properties has proved unsatisfactory. Therefore, the general consensus of the scientific com- munity appears to be that chemical structure has limited value, in identifying carcinogens and is to be used in carcinogenesis hazard assessment only as cor- roboratise supporting evidence. In the absence of other data, however, there are instances in which structure may provide suggestive evidence that a risk of carcinogenesis exists. When structure is to be used as suggestive evidence, well- documented support should be presented and qualified where necessary by complete notation of substances of similar structure that have been adequately studied for carcinogenic activity. QUALITATIVE JUDGMENTAL FACTORS IN EVAL- UATiON OF TOTAL EVIDENCE Evidence that a substance poses a carcinogenic haz- ard is contributed by each source discussed in the preceding sections of this report: epidemiologic studies, studies on experimental animals, and studies based on~ short-term and other tests that have been shown to correlate with carcinogenicity; this includes studies of biochemical pathssays and chemical structure, For some substances data may be available from all three sources; for others, there may be data from only one or two sources. Each source of relevant data needs to be critically evaluated by consideration of the many as- pects discussed in this document. The judgment that a substance poses a carcinogenic hazard derives from the evaluation of the total evidence provided by all of the sources. Different data sources may not contribute equally to the cumulative evalua- tion, depending on the specific nature and extent of the data, the scientific quality of the studies, and the adequacy of their documentation. Conclusions on the carcinogenicity of a substance may be reached on the basis of evidence provided by epidemiologic studies, bioassays in animals, or both. Suggestive evidence is provided by the other types of studies. In the absence of adequate epidemiologic or animal evidence, a positive response in any of the short-term in-vitro tests that correlate with carcinogenicity is considered suggestive of a carcinogenic hazard. Sugges- tive evidence may also derive from considerations of chemical structure or biochemical pathways. Ordinarily, if a substance has produced positive results in a single adequately designed and conducted animal bioassay and no other data ace available, the conclusion is that the substance is likely to pose a risk of cancer to humans. These results may be further confirmed by data on chemical structure, in vitro testing, or relevant biochemical studies that suggest a carcinogenic potential. However, negative data from the latter three sources do not override adequate positive data from an animal bioassay. Further con- firmation that the substance poses a carcinogenic hazard to humans is obtained from bioassay data shoving reproducibility of results, positive dose-re- sponse relationships, and concordance of results (see "Evaluation of Pathologic Results"), Because of biologic variability among species, the conclusion that the evidence is positive on the basis of results obtained in one animal species is not altered by negative data obtained in other species or strains of test animals. Moreover, negative epidemiologic data, ques- tionable because of Iimitatidns in the power of detec- tion of such studies, do not deny the conclusion of carcinogenicity on the basis of animal bioassays. Nega- tive evidence from properly designed and conducted ep- idemiologic studies may, hossever, be used to set an upper limit on human risk to comparable populations under analogous conditions of exposure, It should be stressed that the qualitative judgment whether a substance poses a carcinogenic hazard is based on the evaluation of cumulative evidence from all pertinent data sources. The reasons for specific conclusions need to be clearly detailed. The terms "strong" and "weak" have been used in the literature to describe both the nature of the hazard *or risk and the extent and quality of the evidence. A certain confusion may have ensued, since one could refer to weak evidence of a strong effect or to strong evidence of a weak effect. The two categories are clearly not equivalent and should not be confused. Part III. THE QUANTITATIVE ESTIMATION OF RISK' The previous section of this document dealt with the estimate quantitatively the cancer risk of sstch a sub- issue of the likelihood that a substance poses a carcino- stance in exposed humans if the compound is assumed genic hazard to humans. In some instances a regulatory to be a human carcinogen. agency may be required, or may find it useful, to Quantitative assessment of human cancer risk may JNu. rot.. 6~. so. t. it-tx 979 PAGENO="0213" 209 IdentIfIcation of PotentIal Carcinogena and Rielt Ecilmatlon 259 be based on epidemiologic or animal data. In either instance, methodologic problems arise because of the need to extrapolate from effects observed under one condition and level of exposure and in one population group or biologic system to arrive at an estimate of the effects expected in the human group or individual. Because extrapolations are involved, uncertainties are necessarily attached to the cancer risk estimates that can be made with current methodologies. Furthermore, uncertainties arise from other sources, particularly from attempts to identify accurately conditions and lesels of exposure of the human group or individual. Despite the uncertainties, risk estimates can be and are being made, not only by some regulatory agencies but by otherscientific bodies. Because of the uncer- tainties, however and because of the serious public health consequences if the estimated risk were under. stated, it has become common practice to make cau- tious and prudent assumptions wherever they are needed to conduct a risk assessment. This approach has a precedent in other areas of public health protection where similar problems arise because of gaps in knowl- edge (101, 102). Thus current methodologies, which permit only crude estimates of human risk, are de- signed to avoid understatement of the risk. It must be recognized, however, that in some circumstances this cannot be guaranteed because of other factors that may enhance human response, such as synergistic effects. Thus risk assessments should be used with caution in the regulatory process. If data on animals are used as the basis for estimat- ing human risk, data obtained from the most sensitive animal species or strain tested are commonly recom- mended as the starting point for extrapolation. Of the available data, these are clearly the least likely to understate human risk. Use of data from less sensitive species or strains is justifiable only if there are strong reasons to believe that the most sensitive animal model is completely irrelevant to any segment of the exposed human population. A limited comparison of human and animal data for carcinogens is contained in a report of the National Academy of Sciences (103). Data were compared for benzidine, chlornaphazine, diethylstilbestrol, aflatoxin B1, vinyl chloride, and cigarette smoke. The authors~ stated that ". . - as a working hypothesis, in the ab- sence of countervailing evidence for the specific agent in question, it appears reasonable to assume that the~ lifetime cancer incidence induced by chronic exposure in nan can be approximated by the lifetime incidence induced by similar exposure in laboratory animals at, the same total dose per body ss-eight." These pre- liminary observations suggest that current methodolo- gies may not lead to serious errors. Whether quantitative risk assessment is based on data from animals or humans, there is uncertainty about the shape of the dose-response relationship at the (usually low) levels of actual human exposure. Mathe- nsatical extrapolation models are discussed in detail later in this section. The linear nonthreshold dose- response model is most commonly used at the present time. Of the various models, it appears to have the soundest scientific basis and is less likely to understate risk than other plausible models. It has, for many of the same reasons, a long history of use in protection against radiation (10!, 102). The most favorable foundation for quantitative risk assessment is based on well-characterized responses in human populations with well-defined exposures. Un- fortunately, the exposure estimates are often unavail- able or crude. Negative epidemiologic studies on popu- lations for which usable exposure estimates are avail- able can be valuable in conjunction ssith animal data; the studies on animals provide evidence for carcino- genic hazard, and the epidemiologic data may provide upper limits of response for cross-comparison with the animal data. Although extrapolation from the observed human population group to other groups carries less uncertainty than extrapolations from animals to hu- mans, the possibility of significant differences in the characteristics and conditions of exposure of the two population groups must be recognized. Any such differences that may affect the estimate of risk should be noted, although information is rarely available that will permit specific integration of these factors into the risk assessment methodology. To the extent currently possible, the methods dc- scribed in the folloss-ing section permit a crude order- of-magnitude estimate of risk for substances that may pose a cancer hazard to humans. As more knowledge develops, risk assessment methodologies should be improved. Some of the kinds of information and knoss.ledge that ss'ill likely prove useful in the future are discussed in the sections to folloss', At present, most such information is not available and thus cannot ordinarily be used in risk assessment ssithout the imposition of numerous assumptions. Caution is needed in risk assessment as long as these gaps in knowledge exist. Much has been written about threshold doses for car- cinogenic effect, but unfortunately, there is no recog- nized method for determining their existence. A model recently proposed by Cornfield (104) permits the inclu- sion of thresholds. How-ever, as Cornfield stipulated originally and again recently (105), a threshold could be derived from this model only if there were instanta- neous and complete deactivation of the material before any carcinogenic effect occurs-an improbable event. Since threshold doses for carcinogenesis have not been established, a prudent approach from a safety standpoint is to assume that any dose may induce or promote carcinogenesis, Some of the mathematical models proposed to describe the dose-response relation- ship for carcinogenesis are discussed in the following With the present state of knowledge, the quantitative assessment of cancer risks provides only a rough estimate of the magnitude of the cancer risks; this estimate may be useful in setting priorities for control of carcinogens and in obtaining a very rough idea of the magnitude of the public health problem posed by a given carcinogen. jvci. vol. et. vo. . JuLY 979 PAGENO="0214" 210 260 InteragenCy Regulatory Liaison Group MATHEMATICAL MODELS FOR HIGH-TO-LOW DOSE EXTRAPOLATION WiTHIN A SINGLE BIOLOGIC SYSTEM Mathematical models were developed in the last two decades for estimating the effects of exposure levels well below levels for which test data were available, with the goal of ensuring that the risk will not be underestimated. These models of dose-response rela- tionships make use of data obtained in a given biologic system to extrapolate from high to low doses. Consid- eration must be given to the many biologic variables that influence the level of response in different species or under different exposure conditions. The Models In order to extrapolate outside the experimental range of exposure levels, some mathematical formula- tion relating response to dose must be available. The two categories of mathematical models commonly used to depict the relationship betsseen response and dose are dichotomous-response models and time-to-response models. In the dichotomous-response sttuation the response of interest is the presence or absence of some specified condition. Time-to-response models attempt to relate dose level to distribution of the time until the occurrence of a given event, such as tumor observation or death. (Both categories of models are completely specified except for a few unknown parameters, which are typically estimated from a given set of experimental data.) A variety of different approaches have been proposed to deal with the problem of low-dose extrapolation involving a dichotomous response. Included are the Mantel-Bryan procedure, the one-hit model, linear extrapolation, and various extensions of the multistage model developed by Armitage and Doll (106). Mantel and Bryan (107, 108) proposed an exuapola- tion technique based on the log-probit model, which had long been used for bioassays to estimate median lethal doses. They selected this model because it seemed to provide a reasonable fit to a large body of experi- mental carcinogenesis data and not because of any mechanistic arguments in its support. Under this procedure, extrapolation is conducted from the upper confidence limit on the observed experimental response along a probit log-dose line with a preassigned slope of one to some specified low level of risk. By using the upper confidence limit and fixing the slope at one (a shallower slope than they had typically seen with their experimental data sets), Mantel and Bryan hoped to generate an upper bound on the estimated dose asso- ciated ss-ith the predetermined risk level, regardless of the true form of the underlying and unknown dose- response curve, Host-ever, subsequent theoretical and applied research has demonstrated that the Mantel- Bryan procedure is not as conservative as once thought attd may underestimate risk in some situations (109, 110). The one-hit model is based on the concept that a tumor can be induced after a single susceptible target or receptor has been exposed to a single effective dose unit of a substance (109, 110). Thus, unlike the Mantel- Bryan procedure, there is an assumed biologic mech- anism of action for the carcinogen underlying the one- hit model. This action implies that the probability (P) that a tumor will be induced by exposure to a chemical at dose d is given by the equation P(d) I-exp(-Ad), where A is an unknown non-negative constant. When Ad is small (i.e., in the loss--dose region), it can readily be shots-n that P(d).oAd, i.e., for lose dose levels the one-hit model is well approximated by a simple linear model in which the probability of tumor observation is directly proportional to dose. If the unknown (true) dose-response curve is assumed to have a sigmoidal shape-an assumption supported by a wealth of toxicologic data-then the response will curve upward in the low (or, typically, environmental) dose region. Thus a linear model will provide an upper bound to curves of this shape and, it is hoped, a conservative estimate of the dose associated with any specified level of risk (111). A line connecting zero tsith a point on the dose-response curve for the excess tumor rate above background will always lie above the true dose-response curve for the convex portion of. the curve. An additional degree of conservatism is intro- duced by extrapolating back to zero from an upper confidence limit (UCL) for the net excess tumor rate above the background rate. In the linear model the tumor rate is assumed io be proportional to dose: P(d)░Ad. The upper confidence limit for the slope A is UCL+experimental dose. Thus the maximal risk for a given dose d may be estimated by the equation maxi- mal risk=(UCL/d,)Xd, where d, is the experimental dose. Conversely, the equation for a predicted dose for a maximal level of risk is: predicted dose=(riskXd,)/ UCL. A number of investigators have published papers (112-115) based on the Armitage and Doll (116) formu- lation of the multistage model of carcinogenesis. Under the multistage model it is assumed that the cancer originates as a "malignant" cell, which is initiated by a series of somatic'Iike mutations occurring in finite steps. It is also assumed that each mutational stage can be depicted as a Poisson process in which the transi- tion rate is approximately linear in dose rate. Then the lifetime probability of tumor induction can be ex- pressed approximately as P(d) l-exp(-At-Aid-. - - _Asdk). where A~╗=O for all values of i, and k corres- ponds to the number of transitions or mutational stages. (Highly sophisticated computer algorithms have been developed for fitting the multistage model to laboratory data with the use of a restricted maximum likelihood approach which does not require that the value of k be prespecified.) Both the total incidence of tumors and the time at which tumors occur are important. Tumors leading to early death and life-shortening need to be considered. Time-to-tumor is the time at which a tumor is detected or observed by palpation . or by. gross or microscopic examination of an animal at the time of death or sacrifice. Time-to-tumor is not used here to jvct. sot.. us. NO. 1. JULY ttts PAGENO="0215" 211 identification of Potential Carcinogens and Risk Estimation 261 indicate the instant at which a pretumorous condition proach, no assumptions are made a priori about the becomes a tumor. Time-to-observance is better ter- exact form for the mathematical extrapolation. Instead, minology. the experimental dam are used to estimate the shape of Some hope for improving risk estimates has been the dose-response curve. However Crump et al. (109, based on use of the time-to-observance of tumors in 114) and Guess et al. (110) have shown that the upper addition to use of the proportion of animals possess- confidence limit on estimated risk becomes essentially ing tumors. On the basis of Druckrey's work (117), the linear for generalized polynomial extrapolation in the median time to tumors appeared to increase as the dose low-dose region. This approximate linearity holds even decreased. It was hoped that low doses could be found when the maximum likelihood estimate of excess risk that would result in median times-to-tumor observa- does not contain a linear component (estimated). tion well beyond the expected lifetime; this might Therefore, there is some question whether the mathe- result in the identification of `practical thresholds." matical refinements of generalized polynomial extrapo- Albert and Altshuler (118) expanded on the use of lation are justified for application to animal bioas- median time-to-tumor observance by employing distri- says, which may be only crude approximations to the butions of time-to-tumors for individual animals. human situation (109). Chand and Hoel (119) showed that use of a log-normal As an interim p~ocedure, it has generally been time-to-tumor distribution leads to a probit-log dose recommended (106) that whenever quantitative risk relationship, and use of a Weibull time-to-tumor distri- analysis is deemed necessary, linear extrapolation bution leads to an extreme value model for the propor- should always be included among any methods used lion of animals with tumors: P(d) l-exp[-exp(a+$ unless there is reason to believe that the experimental log d)]. sshere alpha and beta are constants. Schneider- (observed) response does not fall in the convex portion man et al (120) demonstrated that even though the of the dose-response curve. If the response is in the median time-to-tumor may be well beyond the expected concave portion of the curve, the one-hit model is lifetime, a significant proportion of animals or hu' suitable. At low observed responses the linear and one- mans may still develop tumors within the normal life- hit models yield nearly identical results. An added span. Peto (121) examined human data and questioned degree of protection can be achieved by starting the cx- the concept that lower doses result in longer latency. trapolation from the upper confidence limit of the Whittemore and Altshuler (122), analyzing data on response. cigarette smoking, concluded that it seas not possible The mathematical procedures per se are intended to to distinguish between the log-normal and the Weibull provide upper limit estimates of risk from a statistical models, standpoint. However, the risk estimates as applied to The available data do not permit a conclusion as to humans should not be regarded as upper limit esti- sshether lower doses lengthen the latency periods, mates because of large biologic uncertainties (see "Ex- Animal experiments at high doses may induce more trapolation From Observed Effects to Estimates of Risk tumors resulting in easier and therefore earlier detec- for the Observed Population"). tion, and this may not be due to an actual decrease of latency period. CHARACTERIZATION OF POPULATION Ttme-to-observance response models have not re- EXPOSURE cetred the same degree of attention as dtchotomous-~ response models in carcinogenesis risk extrapolation. The estimation of total population exposure to a One of the major factors underlying this relative lack given substance (and/or to its decomposition and meta- of emphasis may be that studies in which animals were bohr products) requires consideration of the following given the substance in their feed have not generated~ aspects: sufficient information to determine the relationship a) sources of human exposure (occurrence, pro' between age and cumulative cancer incidence. duction, uses, and environmental distribu- Procedures b) analytical methods for detecting and meas- uring exposures in the environment and in In the preceding section it was noted that the~ the population; Mantel-Bryan procedure is essentially empirical and c) routes and conditions of exposure; lacks biologic relevance with respect to current knowl- d) duration, frequency, and intensity of expo- edge about carcinogenesis. Since risk extrapolations sure; and developed by the Mantel-Bryan technique tend to zero e) size and characteristics of the exposed popu. much more rapidly in the low-dose region than do cx- lations. trapolations based on somatic mutation models, the During examination of exposure data, important Mantel-Bryan procedure would certainly not be ap- qualitative and quantitative factors beyond definable propriate if the carcinogen under study were thought numerical values of dose level and population size will to act directly on cellular DNA (109). emerge; although such information may not be usable Initially, extrapolation based on a multistage model directly in a mathematical calculation of risk estimate, appears to offer significant advantages over linear cx- it will frequently provide additional perspective and trapolation procedures. Under the multistage ap- insight during risk evaluation. Because of the great JNct. VOL. 63. NO. t. JULY 979 PAGENO="0216" 212 262 InteragenCy Regulatory LiaIson Group diversity in sources and estimating procedures available in various situations, it does not seem practicable at this point to set minimum detailed specifications for the reliable estimation of exposures or to identify recommended or approved methods and procedures for producing exposure estimates. The following general considerations indicate the kind of data useful for assessment of population exposures. The better defined these data are, the higher will be the confidence that a realistic estimate of risk for the exposed populations has been made (123). Sources of Human Exposure Two types of exposure sources are considered: pri. mary sources and human contact sources. Primary sources of exposure to a chemical are those that determine its release into the human environment, and they include natural occurrence, extraction from natural products. mining, chemical synthesis, manufac- ture or production, and specific uses. Human contact sources are those that bring about the contact of the substance with the human body, and they include items or preparations containing the chemical (such as foodstuffs or consumer products), vchic!ec. or a medium in which the chemical is present (such as ambient air or drinking water). Some substances may originate from a single pri- mary source and be present in a cside range of human contact sources; conversely, a specific human contact source may be traced to several different primary sources. It is important that for each substance the entire range of sources and environmental distribu- tion be examined. Frequently, there is more than one source of human exposure, and an individual may be exposed to a sub- stance of concern from an array of sources depending upon the circumstances. Analysis of environmental distribution and exposure pathways allows identifica- tion of the most significant sources, so that both the size of the population exposed and the intensity of exposure can be established. In some instances, it is possible to estimate com- bined exposures to the same substance from different sources, primarily where the populations affected by these different sources are the same. Frequently, how- ever, differences in the populations exposed from various sources are so large that any attempt to combine the estimates may produce an unrealistic or unclear description of the actual human exposure conditions. Then it is preferable to con:ider each source separately and subsequentl) use whatever knossledge is available on multiple sources of expo- sures to interpret these observations. Estimates of the total level of production of a substance can be useful indicators of the extent of exposure, particularly over time. Dates of first synthesis and commercial production of a substance are useful in the evaluation of delayed toxic effects and allow an estimate of the time before cshich no human exposure could have occurred. The accuracy of data on national production and foreign trade of individual substances (which are often difficult to obtain) needs to be ascertained. Uses of a substance are important descriptors of its environmental distribution and the extent of human exposure. Whenever possible, all uses of carcinogenic substances should be identified. An important distinction is that between uses for which human exposure is intended (intentional expo- sures) and that for which it is not intended (uninten- tional exposures). Individual exposure or consumption of a substance may be voluntary or involuntary. The sociologic bases and implications of these definitions are beyond she scope of this report. Anslytlcaf Methods for Detection and Measurement of Exposures The specificity and limit of detection of analytical procedures for the identification of many carcinogenic substances, both in she environment and in exposed organisms, have been remarkably improved in recent years. Progress in analytical chemistry is expected to undergo further refinement and improvement in the near future. The limit of detection of analytical methods varies considerably for different substances and different con- ditions of analysis, and this is a critical factor in assessing a source of exposure. It is important to consider that the agent may not be measurable but may still be present below the minimum detectable level. The minimum detectable level of a substance may vary depending on different vehicles, media, and conditions of exposure. Quantitative determinations of the level of a sub- stance in various exposure sources should consider time and space distribution and variations, and ranges of values may be useful to estimate the conditions of exposure. The chemical and physical properties of the sub- stance should be identified. Such characteristics as particle size distribution for aerosols and dust should be determined insofar as possible. Analytical determination of the levels of a substance in exposed organisms, particularly in the exposed population, is of great value but not always obtain- able. Available data on the levels of substance (or its metabolites) in the target tissues or body fluids should be considered. The dose of an ultimate carcinogen at the site of action in the tissues or cells, which is measured at all times alter its introduction ("target tissue dose") is ideally the dose that should be estimated and correlated with expected effects. This target tissue dose usually cannot be closely estimated because of many variables and uncertainties (102). The relationship between tar- get tissue dose and exposure dose may vary con- siderably under different conditions. To the extent practicable, documentation of the analytical methods, the sampling conditions, the limits of detectability, and the range of observed values is desirable. jxrt. vnt~ 63. NO. t. JULY 579 PAGENO="0217" 213 IdentIfIcatIon of Potential Carcinogens and Risk Estimation 263 Routes and Conditions of Exposuro the extent of risk related to that substance. Because All possible routes of exposure associated with each combinations of exposures to different carcinogens may source should be identified. If any routes of exposure contrtbute to the cancer rtsk in the same population or are considered irrelevant for estimation of effective individual, and because no threshold level for exposure doses, the circumstances should be specified. Careful to a carcinogen can presently be reliably determined for consideration of sources of exposure-e.g.. product use a population, a contributory risk level from any expo- potterns. environmental or occupational situations, and sure level, however small, must be assumed. background-may suggest or reveal routes of exposure Age of exposure should be considered. i.e., whether not immediately apparent. For example, a chemical exposure is essentially lifelong (at more or less constant may also be absorbed through the skin or by ingestion rates? or is concentrated in certain age ranges. The when inhalation is apparently the primary route. relatsonship between tOtal lifetime exposure in each cx- For estimation of animal-to.human correlations in posure pattern and the amount of this exposure that the evaluation of test data on animals, it is necessary to may be concentrated in any specific age mnges should obtain the human dose level in units consistent witli be identified. Wherever fmsible, the degree of stratifi' those used to describe the effective dose in the animal catton of exposed populations should be identified to bioassay being used for comparison. In some instances permit distinctions between effective exposure amounts any necessary conversion from the actual measurement by age f e.g.. childhood, working age, and elderly age at the source to the needed units describing exposure groups) and by sex. As noted above, populations dose can be straightforward (e.g., by simple application having high-risk age groups should be identified. At- of observed or estimated food ingestion rates to a tentton should be given to exceptsonal exposure groups chemical's concentration in a food). In other instances of special concern, such as infants, children, and complex calculations or modeling procedures may be pregnant seomen, as well as to groups with special necessary, as in the estimation of effective exposure genetic conditions or concurrent disease. In addition, distributions from ambient air on the basis of monitor- in descriptions of certain population subgroups, the ing data or emission inventories for point sources. This! smoking habits, dietary and alcohol consumption patS conversion or translation step, often necessary in the terns, and other cultural and environmental charac- estimation of human exposure, should always be cx- teristics should be considered if possible. ~ ~er.~ttt~tt~ :~: t~ EXTRAPOLATION FROM OBSERVED EFFECTS TO conditions of exposure in tmt animals and in humans, ESTIMATES OF RISKS FOR EXPOSED it is necessary to rely on estimates of comparability and POPULATION to attempt to establish an acceptable equivalent dose. The quantitative estimation of risk from a carcino- In the absence of satisfactory equivalent dose data, only genic substance for the entire exposed or potentially defensible conservative assumptions should be used in! exposed population may be conducted with the use of such a ss'ay that the possible risk is not underestimated. observations on the effects of the substance in 1) a defined human population group and 2) experimental DuratIon, Frequency, and Intensity of Exposure ! animal tests. An important factor in the quantitative evaluation of In both situations the extrapolation will take into population exposure is the length of time during~ account the factors that characterize and distinguish svhich exposures occur. Although the time of exposure~ the groups observed and the factors to which the cx- may vary considerably within a population, there trapolatton applies. are cases in which it can be reasonably well-defined~ These include cases of specified duration of exposure! Correlations From Observed Numan Population (e.g., to certain drugs or certain occupational carcino. Groups to Others gens( or continuous lifetime exposure to seidely dis. The problem to be considered here is the estimation semtnated environmental carctnogens fe.g.. polyryrtie of present or potential risks for all people exposed to a aromatsc hy rocarbons). . . given substance by means of data obtained from Effective exposure rates correspoodtng to typtcal observations in a defined population group. The ob' patterns of indtvsdual exposure, whether short-term or served group may be small and its exposure conditions long-term temporal trends, must be reported ssherever may be well defined, as for certain studies of drugs or ssgnsfscansly different patterns exsst. The tseo com- for occupational exposures. In other situations the ponents of the, estimated level or amount of expo- observed group may be poorly defined, even if larger. sure-the effective rate per untt time or per incident of ! In analyzing the correlation betseeen observed and exposure and the frequency-dumtton pattern-should rstimated population eflects, it is desirable where be evpltcttly identtfted for each exposure pattern con- feasible to reviese the critical differences between the stdered. ! two conditions, such as age and sex distribution of the SIxe and Characteristics of Exposed Populations population; genetic, racial, and ethnic differences; en- vironmental differences and migration patterns; dietary The total number of people exposed to any level of a ! and cultural habits; smoking patterns; alcohol con- carcinogenic substance represents a major indicator of ! sumpcion; patterns of intercurrent disease; and particu- pci. vo~ 65. NO. t. JutS 979 PAGENO="0218" 214 264 Interagency Regulatory LIaIson Group lar susceptibility states including pregnancy and fetal and neonatal exposures. Many of these complex var- iables are considered under "Epidemiologic Evidence" in Part II and "Characterization of Population Expo- sure" in Part III. Animal-to-Human Correlations Although a close qualitative similarity has been established in the nature of the response of laboratory animals and humans to carcinogenic substances, a quantitative correlation is more uncertain because of the marked variation of susceptibility in different animal species and among individuals in the human population It is not possible to reduce the variables to a single safety factor for general use (106). Several species-conversion factors should be consid- ered in estimating risk levels for humans from data obtained in another species. Species-conversion factors are affected by many variables, such as body surface, body weighs, metabolic pathways, nutritional condi- tions, genetic variability, and bacterial flora as well as tissue distribution and the retention and fate of the chemical. In evaluating exposures to the general popu- lation, one should consider all ages, transplacental ex- posures, concurrent disease conditions, and special susceptibility states. Other conversion factors should also be considered when observations are obtained for test species under exposure conditions markedly different from those in the population (e.g., different routes or modes of expo- sures, vehicles, modifying factors, variations in age, sex, perinatal exposures, disease states, and single vs. multiple exposures). The limits of uncertainty should be stated whenever possible (102, 106). Different carcinogens tested under comparable exper- imental conditions show a wide range of response; if extreme cases are included, the range of variation is more than one millionfold. Changes in experimental conditions, particularly ones that alter the effective dose, can markedly affect the observed level of effect of a carcinogen within the same genetic strain of animal. Exposure of experimental animals to certain other chemicals in addition to a carcinogen under test may change the observed effect in either direction and at the extremes up to one hundredfold or even one thousand- fold (121). Differences between species can be even greater. On the other side of the correlation, the human response to carcinogens as well as to many other chemicals and drugs may also shose great quat;ti. tative variations among indis-iduals. Studies on the metabolic activation and chemical interaction of car- cinogens in human tissues in vitro have shown inter- individual quantitative variations of about one hun- dredfold in relatively small population samples (125- 127). Individuals resistant or sensitive to one carcino- gen may not be equally resistant or sensitive to another carcinogen or to combined effects of several exposures. Such wide interindividual variations are also well known from many pharmacokinetic studies. A number of variables are relevant to the correlation of animal and human conditions. Some problems inherent in the use of animals must be kept in mind when animal studies are used for estimation of the quantitative carcinogenic potential of a substance for humans. A concise statement of some of these factors is contained in "Drinking Water and Health," prepared by the Safe Drinking Water Committee, Advisory Center on Toxicology, National Research Council, National Arademy of Sciences (56). Factors discussed in this document include the rate of chemical absorption, distribution within the body, metabolic differences among exposed animals, effect of intestinal bacteria, rates of excretion and reabsorption, differences in molecular receptor sites for the carcinogen, environ- mental and genetic differences, and number of exposed animals and susceptible cells. Metabolism and pharmacokinetics account for major differences in sensitivity to chemical carcinogens be- tween species. In principle, this information could be used its estimating the relative sensitivity of humans compared to experimental animals. In practice, de- tailed metabolic pathways in humans are not known for many carcinogens; moreover, the marked variation in metabolism and sensitivity among individuals of different ages, states of health, and other biologic conditions require more information on the heteroge- neity of human metabolic and pharmacokinetic re- sponses than is usually available. It is hoped that future research will clarify these important correlations in much greater depth. Such information, if available, should be used to correct for an underestimate of human risk, but it should be used to correct for an overestimate of human risk only when there is sub- stantial information on diversity of human response. The contribution of animal test data to the estima- tion of the risk level for humans should be based on experiments with the most sensitive species available. Confidence that this procedure will not underestimate the human risk increases with the number of experi- ments and the number of species and strains studied. LACK OF PREDICTABLE THRESHOLDS FOR AN EXPOSED POPULATION The self.replicating nature of cancer, the multiplicity of causative factors to which individuals can be ex- posed, the additive and possibly synergistic combina- tion of effects, and the wide range of individual sus- ceptibilities work together in making it currently un- reliable to predict a threshold below which human population exposure to a carcinogen has no effect on cancer risk, Observation of the marked individual differences in the response of human subjects to carcinogens shows that some individuals do not develop cancer in their lifetime, whereas others develop it readily after the same exposure to a carcinogen. Although these obser- vations are compatible with the existence of different `thresholds" for individual subjects in certain condi- tions, they are not a basis for predicting a no-effect level of a carcinogen in other individuals or under JNCI. vOL 63. NO. I. Jv'Lv 979 PAGENO="0219" different conditions. There is no presently acceptable way to determine reliably a threshold for a carcinogen for an entire population. Individual human subjects in the population are exposed throughout life to a number of carcinogens, which may be considered to provide a background of carcinogenic risk; exposure to any amount of a single carcinogen, however small, is regarded as capable of adding to the total carcinogenic risk (109). Cancer susceptibility varies greatly among individual members of human populations due to genetic racial, and ethnic factors; to environmental and dietary exposure; and to other modifiers. Variability among individuals makes it very difficult to have confidence that an observed no-effect level of exposure in animals or even in a specific human popu- lation (for which individual variation may be small in comparison to the total population) will be applicable to the total human population at risk. A large number of factors (e.g., age, sex, race, nutritional status, im- munologic status, general state of health, previous cx- posure to the substance in question or to other sub- * stances) could affect individual susceptibility. Even if thresholds for. carcinogens could be demonstrated for certain individuals or for a defined population, no reliable method is known for establishing a threshold that could apply to the total human population (67). SUMMARY OF RISK ESTIMATION For a given substance, the usefulness of dose-re- sponse data obtained from a specific human popula- tion group or from animal tests for estimation of risk in the general population is limited by the considera- tion that general population exposures to one sub- stance are usually only a component of the total carci- nogenic burden derived from multiple sources, with their possible interactions. Recognition of these limitations, however, does not imply that no attempt should be made to develop reasonable risk estimates for different conditions of human exposure. The several components of quantita- tive risk assessment include the following: a) definition and quantification of exposures; b) characterization of the exposed populations in quantitative terms; c) chemical and physical properties of the sub. stance and its chemical reactivity in relation to exposure; d) prudent quantitative mathematical extrapo- lation of the responses from observed to esti- mated exposure ranges within the observed biologic system; and e) qualification of the estimated risk in light of identifiable biologic and toxicologic dii- ferences that may be present in theexposed human population. REFERENCES (1) Svrrtorrt U: Scientific ba,r, of environmental earcinogene,is and cancer prevention: Devetoping an interdisciptinary science and facing its ethicat implications. J Tosicol En- viron Health 21435-1447, 1977 (2) HAtTWtLL JL: Survey of Compounds Which Have Been Tested for Carcinogenic Activity. Nail Cancer Inst Public Health Serv PubI No. 149. Washington. D.C.: U.S. Govt Print Oft, 1951 (3) Sncntc P HARTWcu. 3L Survey of Compounds Which Have Been Tested for Carcinogenic Activity. NatI Cancer Inst. Pubtic Health Sees Puht No. 149 (suppl I). Washington. D.C.: U.S. Govt Print Off. 1957 (1) Sovntc P. HARTWLLL JL. Pvrccs JA. edt: Survey of Com- pounds Which Hate Been Tested for Carcinogenic Activity. NatI Cancer Inst. PSblic Health Seer PabI No. 149 (suppl 2). Washington. D.C.: U.S. Govt Print Ott, 1969 (5) National Cancer Institute: Survey of Compounds Which Have Been Tested for Carcinogenic Activity, 1961-1967 vol. Public Health Seer PubI No. 149. Washington, D.C.: U.S. Govt Print 011, 1973 (6) -: Survey of Compounds Which Have Been Tested for Carcinogenic Activity, 1968-1969 vol. Public Health Seer PobI No. 149. Washington, D.C.: U.S. Govt Print Off, 1971 (7) -: Survey of Compouods Which Have Bern Tested for Carcinogenic Activity, 1970-1971 vol. Public Health Bern' PubI No. 149. Washingtoo, D.C.: U.S. Govt Print Off, 1974 (8) -: Survey of Compounds Which Have Been Tested for Carcinogenic Activity, 1972-197) vol. Public Health Bern' Publ No. 149. Washington. D.C.: U.S. Govt Print Off, 1975 (9) International Agency for Research on Cancer: Inorganic sub- stances, chlorinated hydrocarbons, aromatic umines, N- nitroso compounds, natural products, miscellaneous. IARC Monogr Es-al Carcinog Risk Chem Man 1:1-184. 1972 (10) -: Some Inorganic and Organometallic Compounds. IARC Monogr Es-al Carcinog Risk Chem Man 2:1-181, 1973 01) -: Certain Polycyclic Aromatic Hydrocarbons and Hetero- cyclic Compounds. IARC Monogr Es-al Carcinog Risk Chem klan 3:1-271, 1973 (02) -: Some Aromatic Amiors, Hydrauine and Related Sub- stances. N-Nitroso Compounds and Miscellaneous Alkylat. ing Agents. IARC Moougr Es-al Carcioog Risk Chem Man 4: 1-286. 1974 (03) -: Some Organochlorine Pesticides. IARC Monogr Es-al Carcinog Risk Chem Man 5:1-241. 1974 (ii) -: Sos Hormones IARC Monogr Es-al Carcinog Risk Chem Man 6:1-243, 1974 (05) -: Some Anti.thyroid and Related Subsurces. Nitrofurans and Industrial Chemicals. IARC Monogr Es-al Carcinog Risk Chem Man 7:1-326. 1974 (16) -: Some Aromatic Ato Compounds. IARC Monogr Eval Carcivog Risk Chem Man 8:1-357. 1975 (17) -: Some Aainidires. N-. S.. and O.Mustords and Selenium. IARC Monogr Eval Carcinog Risk Chem Man 9:1-268, 1975 (00) -: Some Naturally Occurring Substances. IARC Monogr Es-al Carcinog Risk Chem Man 10:1-353, 1976 1191 -: Cadmium, Nickel. Some Eponides, Miscellaneous In- dustrial Otemicals, and General Considerations on Volatile Anesthetics. IARC Monogr Es-al Carcinog Risk Cltrm Man 11:1-306, 1976 (20) -: Some Carbamates, Thioearbamates and Carbaaides. IARC Monogr Es-al Carcinog Risk Chem Stan 12:1-282 1976 (21) -: Some Miscellaneous Pharmaceutical Substances. IARC Monogr Es-al Carcinog Ri,k Chen~ Man 13:1-255, 1977 (22) -: Asbestos. IARC Morogr Es-al Corcinog Risk Cltem Man 14:1-106. 1977 (23) -: Some Fumigants. the Herbicides 2,4-D and 2,4,5-1', Chlorinated Dihrnaodiovins and Miscellaneous Industrial Chemicals. IARC Monogc Es-al Cas-cinog Risk C7tem Man 15:1-345. 1977 (26) -: Some Aromatic Amines and Related Nitro Com- pounds-Hair Dyes. Colouring Agents and Miscellaneous Industrial Chemicals. IARC. Monogr Es-al Carcinog Risk Chem Man 16:1-400, 1978 (25) -: Some NNitroso Compounds. IARC Monogr Es-al Car' cinog Risk Chem Man 17:1-365, 1978 (26) National Institute for Occupational Safety and Health: Sus- 215 identification of Potential Carcinogens and Risk EstimatIon 265 JNCI. VOL 63. NO. I. JULY 979 PAGENO="0220" 266 interagenCy Regulatory Liaison Group 216 pected Carcinogens. A subfik of she NIOSH Tonic Sub- stances List. DHEW Publ No. (NIOSH) 75-188. Rocksille. Md.: U.S. Dept Health, Edoc. Welfare, 1975 127) Aucon JC. Aoccs HF. WOLF C: Chemical Indocrion of Can- cer: Strocoaral Bases and Biological Mechanisms, sol I. Ness Yosk and London: Academic Press, 1968 (28) AocoS JC, Mucs HF: Chemical Induction ol Cancer So~c.- total Bases and Biological Mechanisms, sol hA. Nest York and London: Academic Press, 1974 (291 -: Chemical Induction of Cancer. Srroctoeal Bases and Biological hlechanisms, sol IIB. Nets York and London: Academic Press, 1974 (301 Bcncsstcr-s I: Caecioogenesis as a Biological Problem. Vol 34. Frontiers of Biology. Amsterdam: North-Hollaod, 1969 (33) HucecB WC, CONWAY WD: Chemical Carcinognnrsis and Con- cern. Springlield. IlL: Thomas. 1964 32) Q,rnnora DO: Chemical Carcioogenenis. Boston: Little, Seotcn & Co.. 1962 (33) Sc.sot.c CE, ml: Chemical Carcinogens. American Chemical So- ciety Monogcaph 173. Washington, D.C.: Am Chem Soc. 1976 31) TrtcstsrAean B, ScossAno T: Substansen mis kanserogener Winkong. Berlin-Bach: Zentralinstitut Or Kcebslorschung der Akademie der W'issenschalten dee DDR, 1973. 1351 TOnsATts L, AGT#tE C, BARTSCH H, es al: Ecaloation of the carcinogeoicity of chemicals: A reciess of she monograph pmgram ol she International Agency foe Research on Can- cer (1971 to 1977). Cancer Ret 38:877.885, 1978 (36) National Cancer Institute. National Institute of Ensironmental Health Sciences. National Institute for Occupational Salesy and Health: Estimates ol she Fraction of Cancer a she United States Related so Occupational Factors. U.S. Dept Labor. Occupational Saletyand Health Admin, Docket No. H-09O. Washington, D.C.: 1978 (37) R0cKETTc HE: Caose specific mortality ol coal miners. J Os- cup Med 19:795-801, 1977 301 U PP. Fo,stsrcNt JF Jo. MANTeL N, es al: Cancer mortality among chemists. J Narl Cancer Inst 43:1159-1164, 1969 (39) KosKEt.A R-S, Heosrroc S. SOARAVA R, es al: A mortality study of foundry mothers. Stand J Work Entiron Health2)suppl 11:73-89, 1969 10) Gtnsost ES, MARTIN RH, LOCEINCTON JN: Long cancer sality in a steel foundry. 3 Occup bled 19:807-812. 1977 111) Moss E, Let WR: Occurrence ol oral and pharyngeal cancers in tensile scorkers. Br 3 Ind bled 31:224-232, 1974 112) Lt.oso JW, DccourLc P. SAI.5SN LG: Unusual mortality ex- perience of printing pressmen. J Occup bled 19:543-550, 1977 13) 9%'AGONsit JR. l.ltLLEo RW. LeNDs.'. FE Jo, et al: Unusual can- cer mortality among a group of underground metal miners. N EngI J Med 269:284-289, 1953 (41) RrDsroso CR, STonrtNo BY. Cter.ss RH: Cancer experience among coke by-pmduct scorkres. Ann NY Acad Sci271:102- 115. 1976 (45) I_ctroN RA, LIE JO, WAGONeR JR. es al: Cancer mortality among cadmium predoction saorkrrn. Ann NY Acod Sd 271: 274-279, 1976 (.96) MoNsox RR, NAL5N0 KR: i.lortality among rubber monkers. I. S%'hite male onion employees in Akron, Ohio. Am 3 Epi- demiol 103:2)4-295, 1976 47) Acrsr.sos ED: Natal cancer in she furnitttreaod bootand shoe manulacturing industries. Poet bled 5:295-315, 1976 (88) BOtNT0N LA: A death cersilicare analysis ol nasal cancer among furniture stockers in North Carolina. Cancer Res 37: 3473-3474, 1976 -. (19) Ausor M. Enncsr S. DtNšoL C: Leukemia in shoertcorkers en- posed chronically to krozrre. Blood 44:837-841, 1974 (90) COLE P. GOLnst.sN 515: Occupation. In Persons at High Risk of Cancer (Fraumeni JF Jr. ed). Nesc York: Academic Poets, 1975, pp 167-184 581 Ss..xcA TJ, Stvao A, BoRTscctL RK. eds: Mechanisms of Tu- mor Promotion and Cocarcinogenesis. Carcinogenesis, a Comprehensise Sursey, sol 2. Ness York: Eaten Press, 1978 (52) COLntEN NH: Tumor promotion and peeneoplassic progres- non. In Modifiers of Carcinogenesis. Carcinogenesis. a Corn- prehensise Surcey (Slaga T3, ed). sal 7. Nt-ru York: Rasten Press. In press (33) International Agrocy for Research on Cancer Aflatosins. 1ARC Monogr Esal Carcinog Risk Chem Man 10:51-72, 1976 (51) -: 2-Naphshylammne. IARC Mooogr Esal Carcinog Risk Chrm Isfan 497-111, 1974 (551 MAccc PM. fslostrcnuxo R. PREUSSMAN R: N-Nisroso com- pounds and related carcinogens. In Chemical Carcinogens (Seasle CE. ml), American Chemical Society Monograph 173. 9%'ashingson. D.C.: Am Chem Soc. 1976. pp 491-625 (56) Bale Drinking Water Committee, Adcisory Canter on Tosicol- tegy, National Research Cotincil. National Academy ol Sri- ercec Drinking Water and Health. Washington. O.C.r NatI Acad Sci. 1977 (77) National Cancer Adcisory Boar& General criteria for assessing she evidence foe carcinogenicity of chemical substances: Re- port of she Subcommittee on Eovironmensal Carciongenesis. National Cancer Adsisory Board. j Nash Cancer Inst 58:46!- 465, 1544. 1977 (58) FcAcncNt JF Jo. ed: Pmsons as High Risk of Cancer: An Ape pmach so Cancer Etiology and Control. Proceedings of a cool erence sponsored by she National Cancer Institute and she American Cancer Society, Key Biscaysse. Fl a., Dec. 10-12, 1974. Ness York: Academic Press, 1975 (59) Sos.nts P. Stci J: Chemical corcinngenesis as a chronir sonic. ity test. A resiew. Cancer Ret 16:728-742, 1956 (60) International Union Against Cancec Report of symposium on potential cancer hasards from chemical additises and con- taminants so foodstuffs. Arsa Un Ins Contra Canon 13:170- 193, 1957 (61) Subcommittee on Cascinngenesis. Food Protection Committee. Food and Nutrition Board, National Academy of Sciences- National Research Council: Prublems in she evaluation of carcinogenic hacard from one of food additives. Cancer Ret 21:429-456, 1961 (62) Joint FAO/WHO Enpert Committee on Food Additises: Fifth report of the Joint FAO/WHO Expert Committee on Food Additives. Esuluation ol Carcinogenic Haaards ol Food Ad- ditites. . WHO Tech Rep Sec 220:1-32. 1961 (63) WHO Expert Committee on she Pretension of Cancer: Report of she WHO Expert Committee on she Pretension of Cancer. WHO Tech Rep Sec 276:1-53, 1964 (61) BvooNrs.s.ss I. ed: Carcinogenicisy Testing. Union Irserna- sionale Consre Ic Cancer Tech Rep Ser, sol 2. Geneva, Stuita. erland: UICC, 1969 (65) WHO Scientific Group: Report of the WHO Scientific Group on Principles for she Testing and Evalaasion of Drugs for Carcinogenicity. WHO Tech Rep See 426:1-26. 969 (66) Advisory Panel on Carcinogenicity of Pesticides: Carcinogensc- isy of pesticides. in Report of she Secretary's Commission on Pesticides and Their Relationship so Environmental Health, U.S. Dept Health, Educ, Welfare. Washington, D.C.: U.S. Costs Print 011, 1969, pp 459-506 (67) Ad Hoc Committee on she Esaluasion of Lost Lesels of En- slronmensal Carcinogens: Esaluation of environmental car- ciongens-repors so she Surgeon General. In Chemicals and she Future of Man. Hearings before the Subcommissce on Executive Reorganication and Covemment Research ol she Committee on Cosrrnmens Operations of the U.S. Senate. 92d Congress. 1st session. Washington. D.C.: US. Cots Print OIf. 1971. pp 171-183 (68) Food and Drug Administration Adcisory Committee on Prottv colt for Salesy Evaluation: Panel on rarcinngenesis report on cancer testing in she safety evaluation of fuod addisirrs and pesticides. Toxicol AppI Pltars6iurol 20:419-438, 1971 (69) Health used Wrflare, Cassada: The Testing of Chemicals for Car- sinogenicisy, Mutagenicisy. and Terasogrnicisy. Ottasta, Canuda: hlinissry Health Welfare, 1977 (70) 96'HO Scientific Group: Report of she WHO Scientific Groop for she Assessment of she Carcinogenicisy and Mutagenicity of Chemicals. WHO Tech Rep 6cr 546:1-19, 1974 71) SONsAG JSI. Pscc NP, SAsasorrs U: Guidelines for carcinogen bioassays in small rodents. Nash Cancer Inst Carcinogexesis JNC5. VOL 65. NO t. JELY 1n79 PAGENO="0221" 217 Identification of Potential Carcinogens and Risk Estimation 267 Tech Rep See No. I. NatI lost Health. DHEW Pub) No. (NIH) 76-80). Washington. D.C.: U.S. Coos Print 011. 1976 (72) ALBcRT RE. Tones RE. A8aocxsora Er Rationale deceloped by the Enviroomeosal Protection Agency for she assessment of carcinogenic risks. J NatI Cancer Inst 58:1537-1541. 1977 (73) Htarr HH. WATSON JO, WtosTcn JA, edt: Origins of Human Cancer. Cold Spring Harbor Conferences on Cell Prolifera. ion. vol 4. Cold Spring Harbor, N.Y.: Cold Spsing Harbor Laboratory. 977 (71) KLAASSEN CD: Absorption, distribution and enceetion of toni- caots. irs Toxicology. she Basic Science of Poisons (Ca saress U. Doll J, eds). Ness York: Macmillan. 1975, pp 26-44 (75) Btsctsorr P. Boosoc C: Carcinogonenis through solid state sue- laces. Prog Exp Tumor Ret 5:85-97, 1964 (76) Boson KG: Foreign body induced sarcomas. in Cancer, a Compteheosive Treatise (Becker FF. ed), so) I. Nets York and London: Plenum Pt-ens, 1975. pp 485-511 (77) Rtcc JM: Carcinogertesis: A lute effect of irrecersible sonic dam- age during decelopment. Environ Health Perspecs 18:133- 139, 1976 78) Rtcc JM. ed: Perinatul Carcinogenesis. NasI Cancer Inst Monoge 51:1-282. 1979 (79) Food and Drug Administration: Nonclinical Laboratory Stud- ies. Coed Laboratory Practice Regulations. Fed Register 43: No. 247, 59986-60025, 1978 (80) SArrtorrt U, PAGc NP: Releasing carcioogenesis sent results: Timing and extent of reporting. Med Pediuto Oocof 3:159- 667. 1977 (81) Fsxoctc E. St'ooo MB. cochairmen: Symposium: Early lesions md she development of epishelial cancer. Cancer Ret 36: 2475-2706, 1976 (82) STE56ART FW: Tumors of she breast. in Atlas of Tumor Pa- thology, sect IX, fats 34. Washington. D.C.: Armed Forces Inst Pathof, 1950. pp 7-10 (83) Yyvvc S. HsLt,otvr.s RC: Tumoars of the mammary gland. in Pathology of Tumors in Labora:ory Animals (Tcrusov VS. ed). vol 1, port I. IARC Sci PuhI 5:31-74. 1973 (81) Socu..snaxcxx CJ: Mammary neoplastic response of Lewis and Speague-Datcfey female rats so 7,l2-dimethylkent(a)anthra. or X-ray. Cancer Ret 32:883-885, 1972 (85) BxScstcA R. Sarrtorrt U: Experimeotal studies on histogeoesis of blood-borne metastases. Arch Pathol Chicago) 59:26-34, 1955 (86) Kvntszts AP, Kov.x 51, Vssssitsovtrt:ts SD: Metastatic rate of liter tumors induced by diethyloitrosamine in mice. Cancer Res 34:288t-2886, 1974 (87) International Agency for Research on Cancer: Bensidine. IARC Monoge Esal Carcioog Risk Crem Man 1:80-86, 1972 (88) -: Vinyl chloride. IARC Monogr Ecal Carcinog Risk Chem Man 7:291-318, 1974 (89) H.xxxs MC Jo, Nct-rcssactvt F, CtLncrcr JR. edt: Inhalation Carcinogenesis. Atomic Energy Commission Symp See No. (8 (CONF.69t001). Oak Ridge, Tenn.: U,S. Atomic Energy Comm. Div Tech Information Extension, 1970 (90) K.xnno E, Patio JF. eds: Experimental Lung Cancer. Carcioo- genesis and Bioassays. Berlin, Heidelberg, Nets York: Springer-Verlag. 1974 (91) CLxYsoN DB: Bladder carcioogenesis in rats and mice: Possi- bility of artifacts. J Nat) Cancer Inst 52:1685-1689, 974 (92) MILLO5t EC, MtLLEB JA: The metabolism of chemical carcino- gens so reactive elevteupkiles and their possible mechanisms of action in carcirogenesis. in Chemical Carciocgens (Searle CE, cdl. American Chemical Society Monograph 173. 8S'ash- iogton, D.C.: Am Chem Soc. 1976, pp 737-762 (93) Foans TR, Taxooc RE. Crtc. NC: False positive and false neg- ativc rates for carcioogenicity screens. Cancer Res 37:1941- 1945, 1977 (91) ItooL DC, Waa.rcrc HE Jx: Statistical analysis of survival ex- periments. J. Nat) Cancer Inst 49:361-372, 1972 (95) Psro R: Guidelines on she analyses of sumout rates and death rates tn et.periment,tl animals. Br J Cancer 29.101-lOS, 1974 (96: Tstonas DC, Brost.tttv N. Cxxi JJ: Trend and homogeneity analyses of proportions acrd life sable d:tta. ~omput Biomed Ret 10:373-381, 1977 (97) Methods for Carcinogenesis Tests us she Cellular Level und Their Evaluation for the Assessntens of Occupational Cancer Hazaedn. Proceedings of she meeting ol she Scientific Com- mittee, Milan, Italy. Dec. 4-6. 977. Milan: Fondaaione Carlo Erba, 1977 (98) Workteg Croup on Mutagenicity Testing, Subcommittee on Envtronmental Mutagenesis, US. Department of Health, Education, and Welfare Committee to Coordinate Toxicol- ogy and Related Programs: Approaches so determining she mutagenic properties of chemicals: Risk to future genera. stuns. J Environ Pathol Toxicol 1:301-352, 1977 (99) Sorrtorrt U, At-race H, eds: In Vitro Carcinogeresis. Guide so she Literature. Recent Advances and Laboratory Pro- cedures. NasI Cancer Inst Carcinogenesis Tech Rep See No. 44. Natl lost Health, DHEW PubI No. (NIH) 78-844. Wash- ington, D.C.: U.S. Govt Print Off, 1978 (ZOO) Aut-tro IM, Zcxvox C, eds: Structural Correlates of Carcinogen. esis and Muragenesis: A Guide to Testing Priorities? Pro- ceediogn of the Second Food and Drug Administration Of. ftve of Science Summer Symposium, Annapolis. Md., Aug. 31-Sept. 2, 1977, Rocktille, bid.: Food Drug Admin, 1978 (101) International Commission on Radiological Protection: Radia- tion Protection-Recommendations of she International Commission on Radiological Protection. ICRP Pub) 9. Ox- ford: Pergamon Press, 1966 (102) Task Group: Air pollution and cancer: Risk assessment meth- odology and epidemiological evidence. Enc)ron Health Per- spect 22:1-12. 1978 (103) Environmental Studies Board, National Research Council, Na- tional Academy of Sciroves: Carcinogenesis in man and lab- oratory animals. in Pest Control: An Assessment ol Present and Alternative Technologies. Vol I. Contemporary Pest Control Practices and Prospects: The Report ul she Executite Committee. Washington, D.C.: NasI Acad Sd, 1975, pp 66- 82 (101) Coxvrtrut J: Carcinogenic risk assessment. Science 198:693- 699, 1977 (105) -: Modrls for carcinogenic risk assessment. Science 202: 1107-1109. 1978 (106) Hoot. DC, C.xxLox DOS', KtxxcstuTctn RU, es al: Estimation of rcsks of isrex-ersible, delayed toxicity. J Toxicol Ensiron Health 1:133-151, 1975 - (107) MANTOL N, Bxvos OVR: "Safety" testing xl carcinogenic agents. J NutI Caocee Inst 27:455-470, 1961 (108) hlxsTcL N, Boostx.xx NR. Bcowx CC, ct al: An improved Slaotel-Bryan procedure for `safety" setting of carcitrogeon. * Cancer Ret 35:805-872, 1975 (109) Oovvte KS, Hoot. DC, LxracLco CH, et al: Fundamental carci- nogenic processes and their implications foe Into dose risk assessment. Cancer Rex 36:2973-2979, 1)76 1180) Cccxx HA, Cot-ore KS, Pcro K: Uncertainty estimates for Ions- dose-rate extrapolations of animal carcinogenicity data. Can. ocr Res 37:3475-3483, 1977 (ill) Cooxx MA, Prrztsrjcn 0G. hlANTcL N: Evaluation of safety for food addisicex: An illustration involtiog she inlI uenve of methyl salicylate on rat erproduction. Biometrics 26:181.184, 1970 I i12) Cuts OR: Regression models and life-sables. J R Stat Soc [B] 34:187-202, 1972 (ill) Cccxn HA. Couxx' KS: Lots-dose-rate extrapolation of data from animal carcioogeoicity experiments-analysis xl a new statistical technique. Math Biosci 32:15-36, 1976 (iii) Coctte KS, Cvc.xx HA, DEal. KL- Confidence intervals and test of hypotheses coot-pen ing dote response relatians inferred from animal carcinogenicity data. Biometrics 33:437-451, 1)77 (115) HxxTLoy HO, StEL000 RU: Estimati~A of "safe doses" in rae- ctnugenic experiments. Biometrics 33:1-30. 1977 I (116) Axxtirac.c P. DOLL R: Stochattic models foe carcinogenesis. itt Proceedings of she Fourth Berkeley Symposium xx Mathe' musical Statistics and Probability, Berkeley. Calif., June 20- July 30. 960 (Neymun J, ed), vol 4. Berkeley. Calif.: Univ California Press, 1961. pp 19.38 (117) Dxccouov H: Quantitative aspects of chemical carcinogenesis. JNC.I. VOL 63. NO. I. JULY 979 PAGENO="0222" 268 InteragenCy Regulatory LIaIson Group 218 In Potential Carcinogenic Harards Front Dtugs (Es-aluarion of Risks) (Truhact R. ed) Unio Inreonationale (~-ra-e Ic Cancer Monograph 5cr, vol 7. Borlin: Springer-Verlag. 1967. pp 60-78 (118) ALSERT RE ALTSHELER B: Considerations relating to she foetnulation of limits for unavoidable populatioo esposures to environmental catcinogeus. In Rodionuclide Corcinogene. sis (Sanders CL. Batch RH, Ballou JE, et al edt), Atomic Energy Commission Symp Ser No. 29 (CONF.720505). Springfield. Va.: U.S. Atomic Energy Comm Ollice of In- formation Ser,:ices, 1973. pp 233-253 (089) Cvsuno N, Hors. DO: A comparison of models for determining safe Icons of environmental agents. in Reliability and Biom- coy: Statistical Analysis of Lifefength (Pnoschan F. Serfling RJ, eds). Philadelphia: Soc Indust AppI hfath, 1973. pp 681- 700 (120) Scote.csucsssAea hI A, DncocrLc P. Bootco CC: Thresholds for environmental cancer-biological and statistical considera' lions. NY Acad Sci. In press (121) Pc~o Ba Epidemiology, multistage models, and short-term mu- sagenicity tests. In Origins of Human Cancer. Cold Spring Harbor Conferences on Cell Proliferation (Hiatt NH, Wat- son JD, Winsten JA, eds). vol 4. book C Cold Spring Har- bor. N.Y.: Cold Spring Harbor Laboratory, 1977. pp 1403- 1428 (122) 56'rrtrrcvronc A. As.rsut'LcR B: Lung cancer incidence in cigarette smokers: Forth en anal~sis of Doll and Htff's data for British physicians. Biometrics 32:805-81 6, 1976 (123) Second Task Force for Research Planning in Ensironmentaf Health Science: Environmental measurements of chemtcals for assessment of human esposure, chaps 7. in Human Health and the Environment. Some Research Needs. Narl lost Health, DHEW PubI No. (NIH) 77-1277. Washington, D.C.: Dept Health, Edac, Welfare, 1977, pp 217-242 (121) Btvct-rAvr E, FALE H: Environmental carcinogens: The modtf~' ing effect of carcinogens on she threshold response. Arch Environ Health 19:770-783, 1069 (125) HARRIS CC. Acsuce H, Ss-o',cn G, es af: Metabolism of ben- ao(a]pyrrne and 7,12.dimethylhen4ujanrhnacene in cultured human bronchus and pancreatic duct. Cancer Ret 37:3349- 3355, 1977 (126) MantIs CC, APTRcP H, ~ovvou R, et al: Interindividual var)- ation in binding of henaola(pyrene so DNA in cultured hu- man bronchi. Science 194:1067-1069, 1976 (127) HARRIS CC, AUTRUP H, STONER C: Metabolism of bensu)alpy. eeoc in cultured humus tissues and cells. In Polycycfic Hy- drocarbons and Cancer fT'so P0, Celboin HN'. eds). vol 2. Neu- York and London: Academic Press, 5978, pp 331-342 j760I. VOL. 63. NO. I. Jt'LY 1579 PAGENO="0223" 219 ATTACHMENT II Carcinogen Policy at EPA I would like to correct some errors in Eliot Marshall's article, "EPA's high risk carcinogen policy" (News and Com- ment, 3 Dec., p. 975). Marshall was given a lengthy interview during which he was told 1) That I was not a proponent of the genotoxic versus nongenotoxic segrega- tion of carcinogens for regulatory pur- poses. The article alleges that I am. I consider there to be a spectrum of activi- ty between these two extremes, and the same compound might influence both genetic and epigenetic events under ap- propriate circumstances. Accordingt~~,j,~ feel it is premature to make hard and fast E~?6Hice. "-2J'fli~ the Environmental Protection Agency (EPA) had no requirement for positive human data on carcinogenicity. The art&le says we do. The rodent bioas- say remains the basis of our progran9 to detect chemicals with carcinogenic po- tential, and good animal evidence (to- gether with evidence on exposure) is enough to trigger action. We need to think beyond a "black box" interpreta- tion of the rodent bioassay and, for chemicals on which there is good epi. demiotogical data, human experience should be considered as part of the over- all database. The article's treatment of statistical modeling of risk assessment also omits a fundamental point: These models do not provide estimates of absolute risk. The numbers these models generate are most properly treated as rough risk indices that can allow one to compare the risks from different carcinogens or different activities with the same chemical. To treat them as absolute risks is incorrect. Making an issue of whether a risk is l0~, 10-6, or l0~ is equivalent to ask- ing how many angels can dance on the head of a pin. These numbers take on meaning only when referenced to the model used, the confidence limits, the reliability and nature of the underlying 794 Letters data, and in comparison with other car- cinogens to which the modeling is ap- plied. As David Rat pointed out during his 1981 testimony, on the National Toxi- cology Program, it would be inappropri- ate to use such risk numbers as point estimates of absolute rink and make them the turning point of a regulatory deci- sion. Unfortunately, we have seen a ten' dency to do this in thg past. In his inset article "The odds on can- cer: EPA's recent bets" (p. 976), Mar- shall makes much of M. Adrian Gross' concern over permethrin, presenting it as a case of EPA versus Gross. This is inaccurate, In March 1981, the science advisory panel for the Federal Insecti. cide, Fungicide, and Rodenticide Act (F1FRA) concluded that permethrin did not appear to be a potential human onco- gen, The subsequent review by EPA's Hazard Evaluation Division (BED) of the oncogenicity data on permethrin was led by Orville Paynter (a Board-certified toxicologist and chief of BED's Toxicol- ogy Branch), and scientists from the Canadian government participated. The Canadian scientists concurred with HED's conclusion that permethrin was not likely to be a human carcinogen. When Gross raised his concerns, BED asked two former members of EPA's FIFRA scientific advisory panel-John Doull and Edward Smuckter-.-to review HED's assessment., Neither of these gentlemen can be considered light- weights in toxicology. Both concurred with HED. Why toxaphene is listed in Marshall's article is something of a mystery. Cer- thinly EPA had concerns about the carci- nogenic potential of toxaphene. Howev- er, a more immediate problem was the accumulation of toxaphene (and toxa- phene-like materials) in the aquatic en- vironment and the imminent endanger- ment of fish. Solving the fishes' prob. lems also solved, the human health threats, but we emphasized that this should have been dealt with whether or not a human health threat was involved, as EPA's mandate is to protect human health and the environment. Contrary to Marshall's assertions, EPA estimates that there is, at most, sufficient toxa- phene in distribution stocks for only one growing season. JoHN A. TODHUNTER Office of Pesticides and Toxic Substances, Environmental Protection Agency, Washington, D.C. 20460 Marshall discusses what t believe is a very disturbing trend in government reg- ulatory policies, namely the attempt to establish separate guidelines for evaluat' ing the health effects of "genotoxic" and "epigenetic" carcinogens, with empha. sin on softening the restrictions for the latter class of agents. The distinction between these two classes of agents is largely theoretical and has no factual basis in terms of our current knowledge of mechanisms of action of carcinogens, for several reasons. I) We do not know svith certaintv_thgs certai~i~i~cinogens act throu h eno- toxic mec am ot era throu h epigene ic mec anisms. ndeed, recent fihildins ~~i˘Fecu1ar genetics,~gygt~fl fidntdibiology, and immunology tend to ~TUr the classical cistinCffons"between ghiiiiic and epigenetic mechanisms,. ec1~iormat biologic proc,g~g,3(]J. 2) Even if this distinction were true, our current methods for assessing whether or not a given agent is likely to be genotoxic in humans have very seri- ous llmitations (2); and what is worse, at the present time we do not have well- validated short-term tests for assessing agents that might act through nongeno- toxic mechanisms, thai/is, tumor pro- moters, hormones, and Io forth. Identifi' cation of the "epigenetic" agents must, therefore, often be done by exclusion, a risky approach. 3) Most of the known carcinogens produce multiple effects. In fact, when given at sufficient ~ chemicals are usually complete carcino- ~n,ter~5re~rd53b~prnduve- ~Br-iniian~umm~promot-- ~ tcoidit~iiiBy fail to assess the promoting capacity of these compounds. This, and other factors, severely limit attempts to predict the mechanism(s) of action and relative potencies of carcinogens, when findings based simply on genotoxic ac- tivity are used. The paradigm of random point mutation as a basis for understand. ing the carcinogenic action of agents that display genotoxic effects may itself be antiquated, in view of the multistage aspects of the carcinogenic process, probable synergistic (and sometimes in- hibitory) multifactor interactions, and the possibility that carcinogenesis in- SCIENCE. VOL. 219 PAGENO="0224" volves more complex genomic changes (gene rearrangements, chromosomal translocation, oncogene activation, al- tered DNA methylation, and so forth) (1). 4) Certain tumor promoters (such as the phorbol esters and TCDD) can in- duce a significant number of tumors in animals, even without prior application of an initiating carcinogen (3). In addi- tion, there are a few studies suggesting that, although the primary target of the phorbol ester tumor promoters is cellular membranes rather than DNA (1), these compounds may indirectly inflict chro- mosomal damage, perhaps via the gener- ation of activated forms of oxygen (5). If this is the case, then these compounds also have genotoxic activity, albeit through an indirect effect. 5) It is often assumed that tumor pro- moters and other agents that might act through epigenetic mechanisms will, in contrast to initiating and genotoxic car- cinogens, display a threshold in their dose response. The data on dose-re- sponse relationships with tumor promot- ers are skimpy, and I know of no evi- dence Ihat clearly establishes a threshold for tumor promoters in humans or in experimental systems. Even if this were the case, how would we know how to extrapolate from a specific set of data the actual threshold level in a heterogeneous human population? 6) It is true that the known tumor promoters require repeated application to exert their tumor-promoting effect, whereas the single application of certain initiating carcinogens is sufficient (3). This does not necessarily imply a com- fortable margin of safety for tumor pro- moters, because for many substances that are of concern (such as water pollut- ants, industrial chemicals, and food addi- tives) there is likely to be repealed and prolonged human exposure. Moreover, some of these substances are only slowly degraded and, therefore, will persist or even accumulate in body tissues or the general environment. 7) There is the impression that tumor promoters are much less potent than initiating carcinogens and, therefore, are less hazardous. This is not necessarily the case. On a molar basis TPA is about two orders of magnitude more potent in exerting biologic effects than benzo[a]- pyrene, and TCDD is about four orders of magnitude more potent than benzo[a]- pyrene (3). 8) We know that nature has evolved specific defense mechanisms against some of the genotoxic agents, including conjugation and detoxifying mechanisms and DNA excision repalr. We do not know to what extent humans have evolved protective mechanisms against tumor promoters. I do not doubt that such mechanisms exist, but at the pres- ent time we do not know their properties or relative efficiencies. 9) A final reason for being concerned about the potential health hazards of tumor promoters and various carcino- genic cofactors that do not appear to act by directly damaging cellular DNA is the evidence that a major fraction of human cancer is due to "lifestyle factors" and that many of these may not act as simple genotoxic agents (6). ft is essential, therefore, that we not overemphasize our concern with genotoxic agents, downplay the potential health hazards of other types of agents, and thus distort priorities in our efforts at primary cancer prevention. In summary, although there has been exciting progress in our understanding of the mechanism of action of environmen- tal carcinogens (1), the field is in a suffi- cient state of flux that at the present time it would be premature to alter the exist- ing, well-established guidelines for risk extrapolations of potential hazards to the human population. Specifically, I see no justification for assuming a nonlinear dose response and threshold model for certain carcinogens simply because they do not give a positive response in certaln currently used assays for genotoxicity. I. BERNARD WEINSTEIN Division of Environmental Sciences, School of Public Health, and Cancer Center/Institute of Cancer Research, Columbia University, New York 10032 Rofresesces I. I. B. Weinstein, I. Sopratnol. Strove. Cell. Rio- chore. 17,99(1931). 2. P. B. Fisheraod I. B. Weinstein,in Carcinogens bslodostryondEnoirotonent,J.M.Sontag.Ed. (Dekker, NewYork, 1981). p. 113; M. Hollatein, 3. McCann. F. A. Angetosanlo. W. w. Nichols. Morse. Re:. 68. 13 (1979). 3. T. 3. Slaga. A. Sinak, R. K. Boutwell, Eds., Carcinogeneois, vol. 2, Mechanic,,: of Too-or Pronotiort sod Cocarcinogenesi, (Raven, New York, 1978). 4. V. tvaennic and I. B. Weinstein. Corcbtogene. sis 3.505 (1981). 5. A. R. Kinsello and M. Radnvs,,, Proc. Not). Acad. Sci. U.S.A. 78,6149(1978); H. Nagasawa and 3. B. Little, ibid. 76, 1943 (1979); H. C. Birnboirn, Science 215, 1247 (19)2). 6. R. Doll and R. Peto. The Cooneo of Co:cer (Dxtord Unix. Press. Oxford, 1931). I wish to correct the possible implica- tion from Marshall's article that I am an uncritical supporter of the current Ad. ministration's carcinogen regulations. This is not the case. As chairman of the Environmental Protection Agency's (EPA's) Carcinogen Assessment Group, I merely solicited opinions from scien- tists outside the EPA on possible alterna- tive approaches to carcinogen risk as- sessment without expressing my own position. In my view, the real concern ought not to be whether this Administration is more or less conservative than other administrations in its approach to regula' (ion of carcinogens, but rather that there has never been a federal cancer regula- tory policy that really works. In looking back over the last dozen years, the one thing that stands out most forcibly is the lack of accomplishment in the area of carcinogen regulation. When one consid- ers the tremendous amount of effort ex- pended by the regulatory agencies, re- markably few carcinogens have been regulated. For example, fewer than a half dozen carcinogens have been regu. lated by the Air Office of the EPA since 1970. The reason for this poor record is that every attempted regulatory action is fought bitterly. There is no consensus in this country on hosv and to what extent carcinogens should be controlled. There is a hodgepodge of laws passed over many years by different Congresses which have different philosophies of con- trol and very inadequate guidance as to how to carry them out. These regulatory philosophies include banning carcino- gens, regulation by the best avallable technology, regulation on the basis of weighing risks and benefits, regulation to protect everyone with a margin of safety, regulation to the extent possible by tak- ing economic and technical consider- ations into account, and so forth. With all of these different approaches, the regulators are given little actualguidance on how to regulate. We have learned a great deal over the years about the prob- lems of regulating carcinogens, and I think that we are now in a msch belier position to develop a simpler, more com- prehensive, and unified approach to car- cinogen regulation. What I mean by uni- fied cancer policy can be illustrated by the suggestion 1 recently made to the Canadian Ministry of Labor, namely, to use economic and technical consider- ations for all carcinogen regulatory deci- sions together with annual cancer,jj~ guidance invels (based~"ii~e linear ~ 10~fot occupaliotsal exposure and 10 Tposu~~gen~I~tboab are lifetime cancer risks of lO~ an4 to'', respectively). The ALARA (as low as reasonably achievable) principle nhould also be part of the regulatory approach. This is an example of a unified approach which applies to ionizing radia- tion as well as to chemical carcinogens and brings occupational and environ- mental standards into balance. Paren- theticalty, the current carcinogen stan- SCIENCE, VOL 219 22~ 796 PAGENO="0225" dards of the Occupational Safety and Health Administration entail lifetime can- cer risks as high as 1 percent to 2 percent (10-2), which incompletely out of balance with the attempts to control environmen- tal exposure to lifetime risk levels of 10-6. Regardless of the acceptability of this par- ticular approach, the main point is that we need something like it. * I think the federal regulatory agencies under the aegis of the Office of Science and Technology Policy will have great difficulty in effectively formulating an overall cancer regulatory policy because they represent only one of the many groups that are involved with dancer regulation. I suggest that Congress com- mission the National Academy of Sci- ences to develop a comprehensive and unified program for the regulation of carcinogens of all types and by all modes of exposure: food, waler, air, drags and cosmetics, consumer goods, and so forth. The Academy is the only body with sufficient stature and detachment to carry out the task; the effort should include the participation of all the con- cerned parties: academia, labor, indus- try, the environmental groups, regula- tory agencies, and so forth. The program should deal with all aspects of regula- tion, including risk assessment and the mechanisms required to separate scien- tific evaluations from the regulatory de- cision process. This program could be translated by Congress into appropriate legislation that would override all other legislation in the area of carcinogen regu- lation. If we cannot achieve a unified and comprehensive system that reflects a reasonable balance among the various views about carcinogen regulation, the whole regulatory enterprise will continue to be bogged down in endless polemics and legal warfare. Roy E. ALBERT Institute of Environmental Medicine, New York University Medical Center, 550 First Avenue, New York 10016 In evaluating government regulatory policies, it is often difficult to separate scientific judgments from policy deci- sions. Marshall's article addresses sever- al good examples. A congressional staff investigation of the pesticide regulatory program in the Environmental Protec. lion Agency (EPA), under way since last June. analyzed the scientific basis for several recent regulatory actions taken by the EPA in an effort to sort out legitimate scientific refinements in regu- latory decision-making from changes in policy. The investigation's findings, con~ 798 221 clusions, and recommendations are con- tained in a staff report presented in De- cember 1982 to the members of the de- partment operations, research, and for- eign agriculture subcommittee of the House Committee on Agriculture ("Reg- ulatory procedures and public health is- sues in the EPA's Office of Pesticide Programs"). Chapter 6 of, the report focuses on regulation of pesticides shown to pro- duce cancer in laboratory animals. An in-depth review of several case studies, along with dozens of interviews with staff scientists responsible for analyzing available data on pesticide oncogenicity, led subcommittee staff to conclude that significant changes had indeed been in- corporated in the way the EPA balances and juxtaposes experimental evidence under the aegis of "weight-of-evidence" decision-making. The unstated, but ob- servable, changes from past risk assess- ment policies and procedures described in the report are comparable to those discussed by Marshall-that is, less con- cern for oncogenic pesticides thought to be nongenotoxic, markedly higher levels of tolerable risks, and greater skepticism in evaluating whether toxic effects ob- served in animal experiments pose suffi- cient hazard to man to warrant consider- ation of restrictive regulatory actions in light of the benefits from use of the pesticide. Officials of the EPA have disputed the notion that cancerpolicy has changed in the pesticide program. In a letter dated 22 December 1982 to subcommittee chairman George E. Brown (D-Catif.), Assistant Administrator for Pesticides and Toxic Substances John Todhunter argued that recent decisions are a logical extension of policies established in past pesticide regulatory decisions involving suspect carcinogens. Independent scien- tists contacted by the subcommittee are currently evaluating these issues and will be catted upon to help the subcommittee determine the advisability of alternative risk assessment procedures. Because of his desire to widen the debate on generic cancerpolicy issues to include the exper- tise of scientists outside the regulatory community, Chairman Brown plans to hold hearing focusing on the cancer poli- cy issues addressed in the report early in the new session of Congress. CHARLES M. BENBR0OK Staff, Subcommittee on Department Operations, Research, and Foreign Agriculture, Committee on Agriculture, U.S. House of Representatires, 1301 Longworzh House Office Building, Washington, D.C. 20515 Fluidized Bed Techno!ogies Hans Landsberg's article "Relaxed energy outlook masks continuing uncer- tainties" (3 Dec., p. 973) provides the incidental information that "fluidized bed technologies" are an example of "nonpolluting ways of coal combus- lion." This is simply not true. There are some indications that low levels of pollutant emissions with fluid- iced bed combustion may be achieved at somewhat lower cost than competing technotogies, Even this remains to be proved in commercial applications. Although no method of coal combus- tion can be considered nonpolluting, emissions of significant pollutants can be reduced to acceptable levels by installing expensive control equipment. JERRY L. SHAPIRO Bechtel Group, Inc., 50 Beale Street, San Francisco, Cal(fornia 91105 Shapiro is correct in saying that the description of fluidized bed technologies as "nonpolluting ways of coal combus- tion" overstates the performance offiuid beds with respect to reduction in air pollutant emissions, His statement that fluidized beds can be operated with low- er pollutant emissions than other com- peting technologies is a more accurate description of the present state of the technology. Fluidized beds do have low. er nitrogen oxide emissions and can be operated so that sulfur oxide emissions can be greatly reduced. Particulate con- trol should also be less costly than for conventional pulyerized coal boilers. To date fluidized beds have received only limited application and then only in relatively small installations. The com- parative economics of combustion of coal in fluidized beds and in convention- al large boilers, both meeting air pollu- tion emission standards, is yet to be demonstrated. I appreciate Shapiro's calling attention to these facts. HANS H. LANDSBERG Resources for the Future, 1755 Massachusetts Avenue NW, Washington, D.C. 20036 - En-anon,: In the article "Breast-feedinypatterns in tuw-sncume cuontrirs" by B. M. Poplin en at. (t8 Dec., p. 1098), Tabte 2 was printed iucorrrctty. The data for "Peru, 978' and "Guyana, 1975' shoatd have been listed under "Latin Ainet-Ie-a." The data for "Nepal. t976" and "Bangladesh, 1976" shnutd have been listed under `Asia and ahe Pacifiv." The data for "Lesoitto. 977' shnutd bane been listed under "Africa and the Near East" Eriatie,,: In the report "Taste lashes: Reaction times, intensity, and quality" by 5. T. Ketting and B.P. Hatperit (2OJaa., p. 412), an eerorappeared in Table 2 en page 413. The magnitude estimate for the 1000-millisecond sodins, saccharin pulse obtained daaegthe tast tOOmittisecondsofibe putsedamtian was 16 ▒ 1.4, nat 1.6 ▒ .4. SCIENCE, VOL 219 22-14B O-83----15 PAGENO="0226" 222 ATTACHMENT III The regulation of chemical carcino- gens is based on several types of scien- tific evidence. Among these, well-con- ducted human epidemiological studies are the most persuasive and least contro- versial. Short-term studies in vivo and in vitro, including genotoxicity. neoplastic cell transformation, and chemical struc- ture-activity relationships. may provide supportive or suggestive evidence. How- ever. carcinogenicity testing in labors- tory animals remains the primary basis for most regulatory decisions (1). In light of our ignorance of carcinogenic mecha- nisms and our inability to determine no- effect or threshhold levels, public health concern has required that all animal car- cinogens be considered as potential hu- man carcinogens. In the past decade, daring which there v~ere many animal carcinogenicity tests, it became evident that the nature and extent of positive evidence varied widely among different chemicals, as is true in other toxicologi- cal testing. Yet the existing all-or-none approach to carcinogen regulation re- quires that all animal carcinogens be treated as if they pose equal risk to humans. This is a difficult position to The author is esroulule profettur cf comparatire evedaite or the Johvc Hapuiva Uviceesity Schuct uf .trd:ive avd a oonruttartr in eeriwvvrevtat ~vd ~~icaiagy. H: was formerly uurin5 dieectv eftha Carcinafeeesis TcsOv~ Prn~acve and head the Tartar Pothototy Section at the NariovatCavcee defend, and questions have been raised in the scientific and lay communities regarding the relevance of animal evi- dence to human risk There has also developed a fatalistic disregard for ex- perimental evidence, even among some of the best informed members of society. The assertion that all animal carcinogens pose equal threats to human health can- not cOntinue without risking a greater skepticism for regulatory decisions. Illustrative of the problem is the fact that there is no acceptable regulatory procedure-particularly with regard to food additives-to permit distinctions to be made, based on the weight of evi- dence, about the potential cancer risks posed by such diverse substances as saccharin, 2-naphlhylaminc, nitrilotri- acetic acid (NTA). chloroform, DDT. dimethylnitrosamine, afiatoxin, chlor- dane, vinyl chloride, and tris(2,3-dihro- mopropyllphosphale (Tris). In this group are included exarnrles of chemicals that vary svidely in their carcinogenic poten- cy and chemical characteristics. For ex- ample, NTA is not biotransfarmed, is biologically nonreactive, and is promptly excreted in the urine. It is carcinogenic only to the urinary tract of mice and rats at doses of .5 percent of the diet or above, administered for 2 years (2). This level is also very toxic to the kidneys. Similarly, saccharin, chloroform, chlor- dane, and DDT has-c shown positive results in only a few of several tests for genotoxicity, and they are carcinogenic in laboratory rodents only at very high levels of exposure through major por- tions of the animals' life-spans. By con- trast, 2-naphthylamine, dimethylnitros- amine, afialoxin. vinyl chloride, and Tns are genotoxic in vivo and in vitro in several different test systems. They are carcinogenic in multiple tissues, in more than one species. at very low doses, and-in some cases-relatively brief cx' posures. For example, liver tumors can be induced in rats with aflatoxin B5 at a level of 0.00000000) percent in the diet (3)- A system by which selected animal carcinogens could be ranked semiquanti- tatively may be a useful regulatory alter- native to current methods. It may not be possible to rank all animal carcinogens in a scientifically s-slid manner, even if this were desired, Many have had limited or inadequate testing. Others, such as hor. mones, may be considered to operate through a relatively unique (though un- clear) mechanism, and a system based on traditional toxicological measure- ments may be considered inappropriate In many instances, however, where test- ing has been adequate and equivalent, a ranking scheme may assist in interpret- ing toxicological data for health risk as- sessment and regulatory policy. In this article, I propose a possible approach to ranking animal carcinogens based on evi- dence derived primarily from the test animals themselves. Other ranking sys- tems could be developed, and other fac- tors and types of data will ultimately be included in an overall human risk assess- ment. Horvever, test animals are the human surrogates in toxicology and will continue to be the basis for regulatory decisions for some time to come, The system is proposed on the assumption that the chemicals under consideration have not already been shown to be hu- man carcinogens. Current Efforts to flank Cercinogros The carcinogen standards of the Occu- pational Safety and Health Administra- tion have separated potential carcino- gens into three categories according to the level of evidence (4). However, most of the pertinent data are omitted from the criteria and all of the chemicals men- tioned above would be classified as be- longing to category I, This is misleading in light of present knowledge and the nature of the available evidence. A recent paper by Griesemer and 077 Ranking Animal Carcinogens: A Proposed Regulatory Approach Robert A, Squire Summary, The nalure and extent of positive evidence associated with animal carcinogens vary widely, yet present regulatory policy does not permit adequate dincriminalion among the many carcinogenic nubntances. Most are treated an it they pose equal potential rink to humans, and thin is not consistent with the available data. Without knowledge of carcinogenic mechanisms, the evaluation at renponnea in intact mammalian nurrogates best relleclsthe potential levels of human rink. An example of a scoring system is proposed by which animal carcinogens are ranked according 10 Ihe moat relevant toxicological evidence derived from animal and genoloxicily studies. Different classes of animal carcinogens could thus be recognized and would permit several regulatory options and provide a means to establish priorities for public snd scienlilic concerns. tittx~ru77SttI.xrtr C,ateueichr t t5ttt OAthS PAGENO="0227" Cueto 5) offers a more detailed classil- cation of animal carcinogens from data derived from the National Cancer Insti- tute's Testing Program (now the Nation- al Toxicology Program). The criteria used were based on those recently adopted by the International Agency for Research on Cancer (IARC) for use in the IARC monograph series that evalu- ates the carcinogenic risk of chemicals to humans (6). This method classifies the evidence of carcinogenicily in animal experiments as either "sufficient" or limited. Sufficient evidence requires that animal experiments show an in- creased incidence of malignant tutrnors: (ii in multiple species or strains, and/or (ii) in multiple experiments (routes and! or doses): and/or iii) to an unusual de- gree (with regard to incidence, site, type. and/or precocity of onset). Additional evidence may he provided by data con- cerning dose-response. mulugenicity. or structure. Limited evidence is not pre- cisely defined, hat incluties induction of "certain neoplasms. including lung tu- mors and hcpatomas in mice, which are considered of lesser significance than neoplasms occurring at other sites for the purpose of evaluating the carcinoge- nictty of chemicals." This is a significant step toward ranking animal carcinogens according to the strength of experimental evidence. It does not. however, include the biological factors lobe considered in assessing carcinogenic potency or poten- tial human risk. Other efforts to recognize the apparent differences among animal carcinogens base been expressed through proposals that two distinct categories he recog- nized: genetic and nongenctic (7). Stich carcinogens would he considered either as initiators or as promoters (or modifi- ers). The danger in this dichotomous approach is that it may result in substitu- lion of one rigid policy for another Inamely. the Delaney clause 8)1, both based on theoreticat assumptions and yielding only two possible categories. Although the opinion prevails that a mu- tagenic-like event-that is, DNA dam- age-is the ultimate mechanism of neo- plastic transformation 19), this remains hypothetical. Further development of genotoxicty tests and understanding of carcinogenic mechanisms may ultimate- ly permit short~lerns studies to largely replace long-term animal bioassays. However, at present, the high correla- lion between genotoxicity and carcino- genicity is empirical and should not gov- ern regulatory policy. At certain exposure levels, most-if not alt-animal carcinogens are toxic to the target cells. Neoplaslic trensforma- lion ~~tittld therefore be either all genetic or all ns'ngenetic. and the differences observed in animal studies may be epi- phenonten:s associatet) vs lb detoxttica- lion. repair. or other adaptive mecha- nisms. Furthermore, if so-called modifi- ers or promoters act on initialed cells, a promotional effect on tissues such as breast. colon, or lung, where there exists a high background of cancer in humans, could produce a greater risk than would he produced by an initiator acting on the liver, for example. where human cancer rates in the United States are very low. Perhaps the most important consider. ation in the mechanism of neoplastic transformation is whether it is direct or indirect. If transformation is secondary to certain levels of toxicity, then no- effect or threshold levels would exist irrespective ~f whether carcinttge'ttesis is genetic. nongenetie. or both. Sever.t) ~n~t'' `~r.'i"-"." ~::`: ::.: been shown to he mutagenie and give positive results in only a few of a large battery of other genotoxicity tests, ac- cording to current methods (10, ii): there is no firm evidence to explain the mechanism by which they induce can- cer. Furthermore, there remains contro- versy over which short-term tests should constitute an appropriate battery for de- termination of mulagenic orcarcinogenie potential (1/, 12). While genotoxicity tests should be a significant part of the total assessment of carcinogenic poten- tial. they should he considered as provid- ing suggestive or supportive evidence- as are other short-term and in vitro meth- ods-which may or ma~ not add to tlte evidence derived from animal stttuliex. Extrapolation from animal resutts to potential human risk has recently cen- tered on the use of mathematical tnodels. partially in an effort to obviate the debate over threshold or no-effect levels. Sever. al models have been developed, some of which are said to reflect certain biologi- cal events at low levels of exposure (/3). Hosveser. no models can actually he based on the biological es'ents. since these are not known for any carcinogens. For the same animal data. different mod- els may predict levels of risk that vary widely 114), indicating the potential error involved in estimating carcinogenic po- tency or human cancer risks by such methods. Because of the uncertainty, regulatory agencies have tended to em~ ploy conservative models, for which low-dose linearity is assumed. These models are based on theoretical one-hit mechanisms, as in radiation-induced mu- tagenesis: such models may also bejusti- fled by assuming the additivity of back. grotind and indttced tumors, which would yield low~dose linearity regardless of the mathematical model employed. Other less cirttsersalive models. sttclt as SCIENCE. VOl. ~I4 223 Table I. Proposed system for ranking Inintal carcinogens. Factor Score A. Number of different species affected Two or more ts One B. Number of histogenetically different types of neoptasns in one or more species Three or more IS Two ID One C. Spontaneous incidence in appropriate control groups of n'optasms induced in treated groups Less than t percent Is Ito tO percent to 10 to 20.percent 5 More than 20 percent I D. Dose-response relationships (cumulative oral dose equivalents per kilogram of body weight per day for 2 yearsl' Less than I microgram IS 1 microgram to I mittigram to 1 milligram to I grant More than I gram E. Mutigrtancy of induced neoptasmv More than 50 percent Is 25 to 51) percent I)) Less than 25 percent No matignancy F. Genotoxicity. measured in an appropriate battery of tests Positive 25 Incompletely positise to Negative o `Based en estimated consumption of tOO From. of diet per titc~vrvt of (r.rjv urtig)). Se.rrint could also Inrt desnkrpod for inhalation er ether appropriate rouser. PAGENO="0228" 224 `table 2. Ranking oniiiv.i( oircioiycns into ice past did wit proviilc data fur the tise of classes according to total tutor score. such a ranking system, recent govern- - crc~t-' inent and nongovernment guidelines rcc- factor oren Regulatory omniend or require protocols that pro- score class ~P ions vide the necessary information. tiitaIOOtRestrict or sin Six factors (lahlc I) arc proposed in 710 ~ this example of a ranking system. They 56 / \ are based on evidence from long-term to / \ carcinogenicity studies in animals and 41 to 55 IV / -s from genotoxicity tests, and there is bio- Less than 4t V Snver.ut optuon\ logical justification for including each of the factors. Some carcinogens that have ne public cults been tested tn several animal species c,itiuini have produced clearly positive results in - two species or more (factor Al. and some also induce more than one type of neo- plasm (factor B). Examples include 2- the multi-hit, do not presuppose losv. napthylitntine. nitrovamines. aflatoxin. UOsC itteattib. but instead, depend on and vinyl chloride. At other extremes. the shape of the dose-response curve in chemicals have been positive in only one the observed range in animal tests, tissue of one species and. sometimes. Mathematical models also neglect only in one sex, as with saccharin or most of the biological information rele- DDT. Metabolism. pharmuucokinelics. vant to human extrapolation. They re- and detoxification may vary qutititatively duce the risk assessment to counting and quantitatively among different spe. animals with neoplasms on the unwar- cies. and universality of toxicological ranted assumption that human response responses is more likely to indicate an will be quantitatively comparable to that inherent properly of a substance. It is in lest animals. As slated by Munro and biologically reasonable to assume that Krewski 115). "We must not lose sight of the greater the number of mammalian the fact that animal studies serve primar- species and tissues that are affected in a lb as qualitative surrogates for humans similar manner by a toxic substance. the and that any attempts to quantify re- more likely it is that the human response sponse beyond the realm of biological will also he similar. Comparative mela- certainty are open to serious question.' bolic and pharmacokinetic studies can be Extrapolation from anintal data to poten- equally or more revealing. but they are tial human risk requires consideration of often not available when decisictns must many factors, including biological data be made. and the nature of the substance in ques- The natural incidence in control ani- tion. rnals of the type of neoplasm induced in treated animals must be considered (fac- tor C). The high susceptibility of lahora- Proposed Method for Ranking tory rodents to several types of curcino- Animal Carcinogens genie effects provides a sensitive indtca- tor for regulatory purposes and is proha- The identification of an animal carcin- bly based on genetic susceptibility. as ogen requires long-term exposure of lest suggested by the very high incidence of animals. usually mice and rats, to the spontaneous tumors. All laboratory ro- chemical in question. The design. con- dents have tumor rates far exceeding duct, and evaluation of the experiments those of humans at most sites (16). Even are complex procedures. and these ire so-called low tumor incidences of up- discussed at length in several recent pub- proximately I percent in animals would licalions (1, 11, 12). II is assumed for the he major epidemics in the human popula- purpose of this discussion that animal lion. Whatever the mechanism. there carcinogens have been identified by test- appear to be large populations of so- ing at ntultiple doses in at least twit called initiated or latent neoplastic cells species and that the adequacy or validity in certain tissues in laboratory rodents. of the experiments and conclusions are Thus. the experimental induction of to. not in serious question. If such testing mors that have high natural occurrences requirements have not been met, this in the test animals is less relevant to ranking system should not be applied. In human risk than is the induction of tu- fact, it is difficult at present to conceive mars that are normally rare in the test of a method for comparing carcinogenic animals. It can he argited that tissue p.utcatials of chemicals inadequately or ~spectficity is not alivays correlated he- unequally tested. Although testing in the ~tween man and test antnals. No biologt- 20 ~OVt.hlBIiR tOOt `t':ihle 5. Apprirsintata rank sit ten animal carcinogens based on the proposed s)~tem. - Carcinogen Score Rank Atatosin IflO I Dirneth~lnitrosatnine 95 I Vinyl chloride UI) 1 Trisi.5.dihtomaprOPyli' 90 1 phosphate tTris) 2.Naphthylamine fit II Chloroform 65 Ill PiTA ~I IV Chlurrdane . `iii V Saccharin Sh V l)t)T 31 V 7 V ~j cal rules are absolute. Hossevercoccord- I ing to the ARC documents to date. there is an 1111 percent correlation between tis- sue site stisceptihilities in humans and in test anintals among the 15 known carcin- ogens that have been adequately tested in animals by routes comparable to those of human exposure 117). Dose'recponse relationships (factor D) must also he considered. The amount of chemical required to induce a neoplaslic response and the latency, or time before the tumor appears. are generally consid' ered to reflect the potency of a chemical for the species being tested. A chemical that must he administered in massive or overtly toxic doses throughout a large proportion of a test animal's life.span in order to induce a neoplasm should be regarded differently than one for which low doses for relatively short periods of time are carcinogenic. Latency as such is not included in this scoring system be- cause this determination requires that large numbers of animals be killed at various intervals to detect the onset of most neoplasms. and this is not a routine procedure in most testing programs. La' tency is reflected, however, in the use of cumulative dose equivalents. The implications of dose-response for the mechanism of action may he equally Ifl.2t7 important. If genotoxic as opposed to nongenotox)c properties are directly Iated to a carcinogenic mechanism. those substances that induce neoplasms only'-~_ after severe and prolonged tissue dam.š~,,. ~ age make subtle, irreversible one'hit ` type effects implausible..ź=~- The induction of malignant rather than benign neoplasms generally provides persuasive evidence for carcinogenic po- tential (factor E). As indicated earlier. this is a major criterion in the recent IARC approach. The inclusion of benign neoplasms in evaluation continues to be controversial. However, since there is evidence of progression from benign to malignant stages in the mulli'stage devel- PAGENO="0229" opmcnt of several epithelial cancers in hum.ann and other animals (18), it is prudent from the regulatory aspect to include benign neoptasms. Consequent- ly, they are included here but are weight- ed less than malignant neoplasms. The criterion of genotoxicity (factor F) takes into consideration the prevailing theory of neoplastic transformation and the possibility of subtle, irreversible ef- fects at low, nonloxic exposure levels that cannot be assessed in animal tests. Positive findings in all or in some tests in an appropriate genoloxicity battery could be required. Another approach is to assign a lowerscore to substances that give incompletely positive results, as il- lustrated in Table I - In the final report of the Scientific Committee of the F°od Safety Council (12), chemicals were clas- sified as belonging to category A, B, or C, depending on the strength of the evi- dence for mutagenic potential; such s scheme could be developed to score chemicals in this proposed ranking sys- tem, Regulatory Applications As shown itt Table 1, application of the scoring system to the six factors will result in total scores varying from 13 to 100, If results are positive in more than one species, sex, or experiment, data from the most sensitive responders would be used for scoring categories C, D, and E. There may be different-.. perhaps equally defensible__assign. ments of numerical values, and these scores may be grouped to rank animal carcinogens into any number of classes. The development of an ultimate scoring system would probably require the coor- dinated effort of a multidisciplinary panel or committee, As recommended here, however, five classes would permit an adequate spread and several regulatory options (Table 2). In Table 3. the ten chemicals listed above were scored by this method (19) These chemicals were chosen somewhat arbitrarily to present a range of scores, and because adequate experimental data were available. Of the several established human carcinogens, relatively few~ have been appropriately tested in animals to permit applying this system, and consequently the ultimate test of its validity is lacking. Regulatory options would be influ- enced by the nature or intended use ofa chemical, the estimated types and levels of human exposure, she number of per. Sons exposed, and by considerations of health and economic benefit. In this sys- tem. a class I substance would represent the greatest potential hazard and may. in the case of an intentional food additive. trigger a total ban. Class I and 11 chemi~ cats would alsohave the highest priority for regulation. Chemicals in classes Ill to V may permit many options including no action, approvals for limited uses, labeling, or public education programs. Carcinogen class may also influence the selection of mathematical models if quantitative risk assessement is to be performed. Chemicals classified I or II, for example, might prompt a more con- servative approach than chemicals clas- sified III, IV, or V. regardless of other considerations. ConclusIon In this article, I have considered only one aspect of carcinogesesis risk assess- ment and cancerpreventios, that is, the evaluation of animal carcinogens, Con. tinued epidemiological research and de- velopment may provide greater health benefits in the future. Also, educational efforts by government and the scientific community to create public awareness of the importance of life.styte and the vol- untary aspects of environmental control should be expanded. At present, howev- er. and presumably for some time so come, testing in animal surrogates will continue to influence our cancer prevcn- lion efforts. The proposed system is based on available data and the current stale of knowledge for rational control of animal carcinogens. The emphasis is on test animal data, since without further knowledge of mechanisms, this informs. lion is the most relevant to human risk. Whatever experimental data are to be included, however, the weight of scien- (ific evidence should be considered in an appropriate system ot carctnogen ctasst. fication. Concerns about animal carcino- gens may thereby be put into better perspective. 225 Oofnneos, sad Notes 1. tnterauency Regutatory Liaison Group. Anna. Rev. Public Healih 1. 345 (t950). 2. R. And,rson, Food Cosme:. Taxied. t6. 569 (t970). 3. 0. N. Wogan, S. Pagtiotunga. P. N. Nemborse, ibid. 12. 60) 11974). 4. Department of Labor, Oceupational Safely and Health Administration, Fed. Regiot. 4.tlNo. 5). 5001 (1980). 3. 5. A Gnrsrmer and C. Cueto. Jr.. in Molr,,dar and Cello),,, Aspects of Carcin,n~en Screeninj. Tests. R. Montos.ann et a!., Eds. tScientitc Publication No. 27, International Agency for Research on Cancer. Lyon. 9801. It. 259. 6. IAIOC monographs on The Evaluation of the CarcinogenIc RisC i! Chc'micals to Human: (International Agency for Research os Cancer. Lyon. 980), no). 22. 7. 3. H. Weishur er and G. N. Williams, in Coon- ceo and Docfls Tooico!ogy. I. Doull, C. D. Ktaassen. M. 0. Amdur. Eds. (Macmillan. Now York, ed. 2. 9001. p. 04. 8. Federal Food. Drug. and Cosmetics Act. Sec- tion 409(c) (31 (A). 9. 5. Miller. Cancer Re:. 33, 1479 (19701. Is. F. A. Dr La tglnnis. 5. S. Lake. 3. E Fitegrr. aId, Drug Metab. Rae., II, (031(9501. II. (ARC monographs on Tlte E'ahcalion of the Carcinogenic Riot of Chemical: to Honor: In. teroaliosat Agency for Research on Cancer, Lyon. t9tO). vol. 22. sop lement 2. 2. Food Safety Council. l'roposed System fitr Food Safer Assessment (Food Safety Council, Washington. D.C., 1900). (3. 3. van Spain. J. Occup. Med. 22, 321 ((9001. 14. National Research Council-National Academy of Sciences, Saccharin: Ta-clinical A osrasment of Risk.s and Benefits National Technical Infor- marion Seroice. SpringOeld. Va., 970). (5. 1. C. Munro and D. 5. Krrmski, Food Comet. Tonical., in press. (6. R. Squire, D. Goodman. N. Vaterio. I. Harsh. banger, C. Dame, in Patho.lggt of Lahinrcutort. Animal:, K. Benirschkr, F. Gamer, T. Jones, Edt. (Springer-Verlag. New York, 197(1, p. (7. L. Tomaris an :1., C'arrcer Re:. 33. 077 (19701. The carcinogenr are aflatoxin, aminobiphenyl. asbestos. aunumine, beoaidire. bislchtotometh. ytloiher. chromium. stillnestnol. minyl chloride. nickel. cyclophosphamidc, melphat.cu. mustard gas, pherytoin. and chloromethyl melhyl ether. It. E. Parlor, ibid. 36. 2705 (9761. (9. The NTA data are from the Naliorat Cancer Ientiiute, Carcinogenesis Technical Report St- oics No. 6, DHEW PubI. No. 77-895. All other data are taken from the IARC monographs on The Etalcsaslan of she Carcinogenic Risk of Chemicals to Hantans. PAGENO="0230" 226 ATTACHMENT IV acrylonitrile is P = 1 - exp (-3.35 x 102 x 4.53 x iO~) = 1.5 x iO-~ In summary, the upper-limit unit risk estimates ~9y'humans breathing 1 ug/m3 of acrylonitrile in ambient air (equiv~2crf'to 0.45 ppb) are 6.8 x 10~ based on the occupational study),X~ x ~ based on the rat drinking water study, and 1.5 x ~ ~ź=ed on the rat inhalation study. Parenthetically, it should be no)~d'~hat if the human equivalent dose assumption were changed to dose per bo)Y.~~i~ht. the unit risk for inhalation based on the rat drinking water st Ŕ~would be 1.2 x 10~ x 1/5.8 = 2.1 x ~ a value which is close ~ other two estimates. Although this estimate is considered unreliable ~d~se of the inappropriate route of administration, It is included here ajA'~atter of interest. The upper-limit unit risk for 1 ugh of a~p1i'nitrile in drinking water is estimated to be 1.2 x i0~. Relative Potency One of the uses of unit risk is to compare the potency of carcinogens. To estimate the relative potency, the unit risk slope factor is divided by the molecular weights and the resulting number expressed in terms of (mMol/kg/day)~. This is called the relative potency index. Figure 13-7 is a histogram representing the frequency distribution of ootencv indices of 54 suspect carcinogens evaluated by the CAG. The actual data summar- ized by the histogram is presented in Table 13-36. When human data are available for a compound, they have been used to calculate the index. When no human data are available, animal oral studies and animal inhalation studies have been used in that order. 13- 158 PAGENO="0231" :227 The potency index for Ócrylonitrile based on the O'Berg study of Dupont workers is 4.5 x iO~ (mMol/kg/day)4. This is derived as follows: the slope estimate from the O'Berg study [6.8 x 10-5(ug/m3)-lJ, is first converted to units of (mg/kg/day)1, assuming a breathing rate of 20m3 of air per day and a 70 kg person. 6.8 x 105(ug/m3)1 x X ~j~3Uj~g x 70 kg = O.24(rng/kg/day)-l Dividing by the molecular weight of 53.1 gives a potency index of 4.5 x ~ Rounding off to the nearest order of magnitude gives a value of i0~ which is the scale presented on the horizontal axis of Figure 13-7. The index of 4.5 x i~-~ lies in about the middle of the third quartile of the 54 suspect carcinogens. PAGENO="0232" 228 14 12 10 0 -4 4th cuarile: .I3ex~5X1~ 3rd ouartile~ ~nd~ 5x3Q - and -3 I 6<10 -3 2nd quartile: Index > 6x10 and -2 -2 1st quar~i1e: Index > 6)CLO * -6 --5 -4 -3 -2 -1 ~ +1 +2 +3 +4 10 10 10 10 10 10 10 10 10. 10 10 Potency index Figure 13-7. Histogram representing frequency distribution of the potency indices of 54 suspect carcinogens evaluated by the Carcinogen Assessment Group. PAGENO="0233" 229 Slope Compounds (mg/kg/dÓy) Molecular Weight Potency. Index Order of Magnitude (expc~nent of base 10) Acrylonjtrjle 0.24 53.1 4.5x103 -3 Aflatoxin B1 2924 312.3 9x10'░ G) 0 Allyl Chloride 1.19x102 76.5 2x10-4 -4 Aldrin 11.4 369.4 3x102 -2 Arsenic 14(H) 14g.8 9x10-2 -2 B[a]P 11.5 252.3 5x10-2 -2 Benzene 5.2x10-2 78 7x10-4 -4 Benzidine 234(W) 184.2 lxlO'░ 0 Beryllium 4.86 9 5x101 -1 Cadmium 6.65(I) 112.4 6x10-2 -2 Carbon Tetrachioride 8.28x10-2 153.8 5x10-4 -4 Chlordane 1.61 409.8 4x103 -3 Hexachlorobenzene 1.67 284.4 6x103 -3 1,2-dichloroethane 3.70x10-2 98.9 4x10-4 -4 1,1,2-trichloroethane 5.73x10-2, 133.4 4x10-4 -4 l,l,2,2-tetrachloroethane 0.20 167.9 lxlO-3 -3 Hexachioroethane 1.42x10-2 236.7 6x10-5 -5 2,4 ,6-trichl orophenol 1 .99x10-2 197.4 1x104 -4 Bis(2-chloroethyl)ether 1.14 143 8x10-3 -3 *Bis(chloromethyl)ether 9300(I) 115 8x1O~1 (~7~' 1 Chloroform 0.11 119.4 9x104 -4 Chromium 63(W) 104 6xlO1 -1 Dichlorobenzjdjne 1.69 253.1 7x103 -3 DOT 8.42 354.5 2x102 -2 TABLE 13-36. RELATIVE CARCINOGENIC POTENCIES AMONG SUSPECT CARCINOGENS EVALUATED BY THE CARCINOGEN ASSESSMENT GROUP (continued on the following page PAGENO="0234" 230 TABLE 13:36. (continued) Order of Magnitude Slope Compounds (mg/kg/day) Molecular Weight Potency Index (exponent of base 10) 1,1-dichioroethylene 1.04 97 1x102 -2 Dieldrin 30.4 380.9 8x1O2 -2 Dinitrotoluene 0.31 182 2x10-3 -3 `~ LY Tetrachlorodioxin 4.25x105 322 lxlOl3 3 Diphenyihydrazine 0.77 180 4x103 -3 Epichlorohydrin 7.69x104(I) 92.5 8x106 -6 Ethylene Dibromide (EDB) 8.51 187.9 5x10~2 -2 Ethylene Dichloride (EDC) 1.44x102 99.0 lxlO-4 -4 Ethylene Oxide 1.86x102(I) 44.0 4x1O4 -4 Formaldehyde 2.14x102(I) 30 7x104 -4 Heptachlor 3.37 373.3 9x1O3 -3 Hexachlorobutadiene 7.75x102 261 3x104 -4 Hexachi orocycl ohexane . technical grade alpha isomer beta isomer gamma isomer 4.75 11.12 1.84 1.33 290.9 290.9 290.9 290.9 2x102 4x102 6x103 5x1O~3 -2 -2 -3 -3 Wickel 6.30(W) 58.7 1x101 -1 t~itrosami nes Dimethylnitrosarnine Diethylnitrosamine Dibutylnitrosarnine fl-nitrosopyrrolidine fI-nitroso-W-ethylurea U_nitroso-U-methylurea H-nitroso-diphenylamine 25.9(not by 43.5(not by 5.43 2.13 32.9 302.6 4.92x103 q*) q~) 74.1 102.1 158.2 100.2 117.1 103.1 198 4x10-1 4x10-1 3x102 2x10-2 3x101 3x10~░ 2x1O~5 -1 -1 -2 -2 -1 0 .5 PCBs 4.34 324 lxlO2 -2 (c6˝tTnued on the following page PAGENO="0235" 231 TABLE 13-36. (continued) Order of Magni tude Compounds Slope (mg/kg/day) Molecular Weight Potency Index (exponent of base 10) Tetrachloroethylene 5.31x10-2 165.8 3x104 -4 Toxaphene 1.13 414 3x10-3 -3 Trichloroethylene 1.26x102 131.4 IxlO4 -4 Vinyl Chloride 1.75x102(I) 62.5 3x104 -4 (?) Vinylidene Chloride 0.13(I) 97 lxlO-3 -3 Remarks: 1. Slopes (q*) in (mg/kg/day)4 are calculated based on animal oral studies, except fo~ those indicated by I (Animal Inhalation), W (human occupational exposure), and H (human drinking water exposure). 2. The potency Index is a rounded-off slope In (mMol/kg/day)-l and Is calculated by dividing the slopes in (mg/kg/day)-1 by the molecular weight of the compound. 3. Not all the carcinogenic potencles presented in this table are final. Some are - subject to change as the CAG Is getting the individual risk assessment documents approved. PAGENO="0236" 232 ATTACHMENT V news from the NATIONAL RESEARCH COUNCIL Fhe Notional Research Council was orgoni_Wd by the National Academy of Sciences in 1916 in order to provide for a broader participation by American scientists and engineers in the work of the Acodemy. The Academy was chartered by the U.S. Congress in 1863 as a private organization with a responsibility for examining questions of science and technology at the request of the Federal Covernozent. The National Academy of Engineering wos organized in 1964 under the original NAS charter. The Nctional Research Council now serves at the agent of both Academies in the conduct of studies and investigations in the public interest. ins CONSTITUTION AVENUE, NW.. WASHINGTON, D.C. 20418 AREA CODE 202 EX 3-8100 Date: February 28, 1983 Contact: Gail Porter or Barbara Jorgenson, (202) 3315-2138 `The basic problem...(is) sparseness and uncertainty of scientific knowledge." COMMITTEE PROPOSES USE OF UNIFORM GUIDELINES FOR RISK ASSESSMENT; ADVISES AGAINST CENTRAL SCIENTIFIC AGENCY FOR RELEASE: 12 noon EDT, Tuesday, March 1, 1983 WASHINGTON - A National Research Council committee today advised the federal government to adopt a package of standard procedures for estimating health effects from hazardous substances, rather than centralizing this function in a single scientific agency. At the heart of the committee's proposal are uniform risk-assessment guidelines, which would provide standard methods for estimating health effects from hazardous chemicals, drugs, or food additives. The committee also recommended that: 9 Risk assessment be clearly distinguished from regulatory decision-making within 1ndividual agencies; 9 Major risk assessments be reviqwed by independent scientific panels; and 9 Written explanations of the soientific rationale used in risk assessments be available for public review and comment before regulatory action is taken. The committee said these procedures would improve the quality and consistency of federal risk assessments and regulatory decision-making without the disadvantages of centralization. It pointed out that centralization would not increase basic understanding of health effects--the central problem in assessing public health risks-- and "would surely reduce" the timeliness and relevance of risk assessments for agency decision makers. (OVER) *The committee's report, Risk Assessment in the Federal Government: Managing the Process, is available for $11.75 (prepaid) from the National Academy Press at the letterhead address. Reporters may obtain copies from the Office of Information, also at the letterhead address. PAGENO="0237" 233 To ensure that new knowledge is continually incorporated into the government's isk-assessment guidelines, the committee called for creation of a Board on Risk ;aseasment Methods. The proposed Board would develop the initial guidelines for adoption by federal agencies, recommend revisions, and identify risk-assessment research ~eeds. It would not "adjudicate disputes arising from regulatory actions," or perform 3cientific reviews of individual risk assessments. After considering a number of ptions, the committee suggested that the Board be established within the National :-esearch Council. RISK ASSESSMENT VS. REGULATORY POLICY Much of the criticism of risk assessment, according to the committee, stems prom confusion between risk assessment and regulatory decision-making. Risk assessments ise factual information from laboratory experiments and other sources to estimate ~ossible human health effects from exposure to hazardous substances. Regulatory ~ecision-making, or "risk management," on the other hand, considers any expected health affects along with social, economic, and political factors. A risk assessment, therefore, is only one of many factors in risk management. Blurring the distinction between risk assessment and risk management jeopardizes the credibility of both functions, the committee said. However, it concluded that physical separation of risk assessment and risk management staffs is not necessarily the answer. While risk-assessment and risk-management are conceptually distinct, said the committee, they do interact. Risk assessors need to know what regulatory policy options are under consideration in order to determine expected exposure to hazardous substances both with and without regulatory action. Risk managers need direct communication with risk assessors to appreciate fully the quality and uncertainty of scientific data upon which risk assessments are based. (MORE) PAGENO="0238" 234 In addition, the committee found that the programs of the federal regulatory agencies `differ markedly in structure, procedures, personnel characteristics, administrative history, and statutory direction." An organizational structure that may work well for one agency, it said, may not work well for another. Therefore, the committee recommended that each federal regulatory agency "commit itself to safeguarding the distinction between.. .risk assessment and risk management," but that no one organizational structure be prescribed for all agencies. KNOWLEDGE GAPS "The basic problem with risk assessment is not its administrative setting, but rather the sparseness and uncertainty of the scientific knowledge of the health hazards addressed," concluded the committee. While "evidence of health effects of a few chemicals, such as asbestos, has been clear, in many cases the evidence is meager and indirect," said the committee. Often available evidence consists entirely of data from animal testing with no direct information on human health effects. "To make judgments amid such uncertainty, risk assessors must rely on a series of assumptions." For example, because the significance of benign tumors in animal tests as indicators of carcinogenicity is unknown, one agency could decide to include such tumors in calculating cancer risk, while another agency may not. Or in the absence of complete dose-response data for a particular substance's hazardous effects, one of several possible models--some more conservative than others--may be used to calculate health r~ cks. The reasons for choices among such options should be explicitly stated and made publicly available before any regulatory action is taker, the committee advised. These written explanations, combined with use of uniform guidelines, the committee concluded, will help distinguish between the factual basis and the scientific assumptions that go into a risk assessment, will enable industry and other regulated parties to anticipate government decisions, and will ultimately improve regulatory decision-making. (OVEB) PAGENO="0239" 235 The corrmittee's report was undertaken at the request of the Food and Drug Administration in response to a Congressional mandate to: determine the merits of separating risk assessments from regulatory policy-making; evaluate whether a single organization should be created to carry out risk assessment analyses for all federal regulatory agencies; and consider the feasibility of uniform risk-assessment guidelines. The committee's study focusedprimarily on assessing cancer risks due to exposure to chemicals in the environment and detailed the risk-assessment programs of the Environmental Protection Agency, the Food and Drug Administration, the Occupational Safety and Health Administration, and the Consumer Product Safety Commission. Its recommendations, however, cover other environmental health areas as well. The Committee on the Institutional Means for Assessment of Risks to Public Health was chaired by Reuel A. Stallones, School of Public Health, University of Texas, Houston. Other members were: Morton Corn, department of environmental health sciences, The Johns Hopkins University School of Hygiene and Public Health; Kenny S. Crump, Science Research Systems, Inc., Ruston, La.; J. Clarence Davies, Conservation Foundation, Washington, D.C.; Vincent P. Dole, department for the biology of addictive diseases, Rockefeller University; Ted P.1. Greenwood, department of politioal science, Massachusetts Institute of Technology; Richard A. Merrill, University of Virginia School of Law; Franklin E. Mirer, department of health and safety, International Union, United Auto Workers, Detroit; D. Warner North, Decision Focus, Inc., Los Altos, Calif.; Gilbert S. Omenn, department of~environmental health, University of Washington School of Public Health and Community Medicine; Joseph V. Rodricks, Environ Corp., Washington, D.C.; Paul Slovic, Decision Research, Perceptronics, Inc., Eugene, Ore.; H.M.D. Utidjian, American Cyanamid Co., Wayne, N.J.; and Elizabeth Weisburger, National Cancer Institute, National Institutes of [Health, Bethesda, Md. Lawrence E. McCray of the Research Council's Commission on Life Sciences served as the project director. # gp: 1,11,10,13 PAGENO="0240" 236 SCIENCE, SEPT. 3, 1982 ATTACHMENT VI Risk Estimate Vanishes from Benzene Report Critics charge the National Cancer Institute with meddling after meeting with industry In October 1981, a group of scientists meeting at the World Health Organiza- tion's cancer agency concluded that workers regularly exposed to small amounts of benzene might contract leu- kemia at three times the expected rate. The estimate would have had wide regu- latory implications in the United States for the chemical industry and the I mil- lion workers currently exposed to ben- zene. But last month, when the Interna- tional Agency for Research on Cancer (IARC) published a report on benzene based on the October meeting, the group's conclusion had vanished from print. The deletion of the estimate has drawn harsh criticism from many scientists, la- bor leaders, and Representative David Obey (D-Wisc.). They charge that Na- tional Cancer Institute (NCI) officials, after meeting with chemical industry rep- resentatives, pressured the international cancer agency to back off from quantita- tive risk assessment. In apparent acqui- escence to NCI, IARC withdrew its risk estimate of benzene. Critics allege that the actions taken by the cancer institute constitute undue interference in the af- fairs of a widely respected health agen- cy. At the very least, IARC's handling of the deletion represents a serious proce- dural blunder, according to more moder- ate critics. The risk estimate calculated by the advisory group was derived from an ex- trapolation of human data. It was signifi- cant because, for the first time, IARC was addressing the scientifically contro. versial area of quantitative risk assess- ment in its reports or monographs, as they are called. The documents are wide- ly used by many countries as the basis for regulation of chemicals. Indeed, the publication of the risk estimate on ben- zene was important because it would have bolstered arguments in the United States to limit exposure to very low levels. In 1980, the U.S. Supreme Court rejected attempts by the Occupational Safety and Health Administration to reg- ulate benzene below the current stan- dard of 10 parts per million (ppm) in air, saying that the agency had not demon- strated significant risk. Labor leaders would have ammunition to reargue the need for benzene regulation, if a docu- ment svere published bearing the impri- matur of IARC. But the risk assessment failed to ap- pear in the monograph. Last October, IARC published a draft of the benzene monograph shortly after its advisory group had met. In January, representa- tives from the Shell Oil Company, Ex- xon Corporation, and Chemical Manu- facturers Association held a meeting with NCI official Richard Adamson to complain that the risk estimate in the IARC draft was faulty partly because it did not state the limitations of the data and suffered from "procedural flaws." According to an internal memo written by Curtis Smith of Shell Oil, they met with Adamson "to solicit his sup- port. - Dr. Adarnson understands the regulatory impact of risk assessment and does not believe that the IARC should be engaged in thisactivity." Two months later, Adamson wrote a letter to IARC director Lorenzo Toma- tis, admonishing him to refrain from quantitative risk assessment, although he did not specifically mention the ben- zene case. Adamson warned Tomatis that the area is fraught with scientific and societal difficulties. "I wish to make sure that no discredit comes to the NCI or IARC as a result of possibly going into the area of risk assessment. I know that it is an area that the regulatory agencies are heavily involved with, but it is an area that also involves national policy." Adamson added, ". . ,I recommend that no change in policy be made as a result of a unilateral decision by you" and that the issue be addressed by the agency's governing council. In the same letter, Adamson said he did not foresee any difficulties in the renewal of NCI funding of IARC's monograph series. Adamson, who has been at NCI for almost 20 years, is director of the divi. sion of cancer cause and prevention. He controls the $1.5 million that the NCI contributes to IARC's $13-million annu- al budget. About $500,000 of the NCI money is allocated for the monograph program. In July, when the benzene report was published, it was apparent that Tomalis had expunged the estimate. Specifically, he deleted a calculation predicting that exposure over a lifetime to 10 ppm of benzene daily might result in 17 excess leukemia deaths per 1000 workers. His action was highly unusual because changes in IARC's monographs are cus- tomarily cleared with the original discus- sion group before publication. But in this instance, members of the IARC group were not polled for a consensus. Adamson denied that he placed any pressure on Tomatis to block publication of the benzene risk assessment. "I have no objection that the group did a risk assessment," he said ic an interview. But he pointed out that he believes one of the studies which was important to the conclusions about risk is "debatable," Adausson said, "I don't know if it was scientifically sound to extrapolate down to 10 ppm. That question is best ad- dressed by the working group, Tomatis and peer review," Adamson notes that his letter to To- matis did not mention benzene and only discussed risk assessment in general. IARC, however, at that time, was con- sidering risk estimates for only two chemicals, benzene and a less common substance, benzidine. When asked why NCI did not raise objections at the Octo- ber meeting to which the institute sent a representative, Adamson said that NCI generally maintains a "hands off' poli- cy. The only reason he wrote the letter was to follow up a conversation in which Tomatis first raised the issue of risk assessment, Adamson claimed. Tomatis said, however, that the letter "came from out of the blue." He re- marked in a telephone interview from IARC headquarters in Lyons, France, "I was upset by its style. I was upset be- cause he told me how to behave. I would not tell the director of NCI how to be- have." But he said that Adamson's letter had "nothing to do with benzene," He insists that NC! did not force him to change the monograph. Tomatis said he pulled the 10 ppm risk assessment from the monograph because the data were inadequate to support the estimate. "I wanted it to be solid and defensible," Tomatis sfuted. But other scientists believe that the benzene data are good enough to justify an estimate of risk. David Hod, director of the division of biometry at the Nation- al Institute of Environmental Health Sci- ences and a key member of the working group that drafted the risk estimate. says, "I thought we were cautious. We felt we were on safe ground. I was sur- prised to see the changes." Hod and other epidemiologists find it odd that SCIENCE. VOL. 217.3 SEPTEMBER 1902 914 0o36-g07512/09u3-09t4501.oan Cupynght C 1982 AAAS PAGENO="0241" 237 Tomatis removed the 10 ppm extrapola- estimate initially appeared was the ap- tion but retained a 100 ppm estimatethat pendix and therefore not subject to the was also based on the same extrapola- same procedural tradition as the actual tion. Hod said that the inconsistency monograph itself. was "peculiar." In retrospect, Tomatis said, the quan- Tomatis said that he had reservations titative risk assessment ahould have about the estimate once the draft was been published separately from the published. He then had an agency statis- monograph. He concluded he should tician contact the scientists who calcu- also have sent a written confirmation of lated the risk estimate. Hod said that his changes to the scientists involved, is true, but he told the statistician But he said he is unsure what he would that if the 10 ppm calculation was have done if they had objected to his dropped, then a paragraph should be actions. "I wish 1 could go back in added to explain the deletion. Hod history," he said with frustration. requested that a written draft of any The toughest critics of NCI and IARC changes be circulated among the work. in this matter point out that Tomatis is ing group members and another consen- caught between a rock and a hard place sus reached. because the agency is financially sup- Philip Landrigan, a member of the ported by the cancer institute. "Tomatis working group and director of the sur- is a good man," but he "must have felt veiltance, hazard evaluations, and field threatened by NCI," said Roy Albert, a studies at the National Institute for Oc- member of the October working group cupational Health and Safety, fired off a and a professor at the Institute for Envi. telegram to Tomatis in July saying that ronmental Medicine at New York Uni. he was "surprised and chagrined to see a versity. "I know for a fact that Tomatis critical portion of the benzene risk as~ was leary of quantitative risk assess- sessment altered. , . , I fail to under. ment," and that he has legitimate scien~ stand why the working group was not tific reasons. On the other hand, the consulted in regard to this important circumstances leading up the deletion change.,,," He said the deletion "goes "look perfectly awful," Albert said, against the text agreed upon by the group At a National Cancer Advisory Board ,and also appears to run completely meeting in May, Tomatis reiterated that counter to that stated policy of IARC" the institute had not pressured him to that working groups' conclusions are im- refrain from risk assessment. Board mutable. members seemed satisfied with his deni- Tomatis wired back, arguing that the al. But Sheldon Samuels, a board mem- section of the monograph in which the her and director of health, safety, and environment for the American Federa- tion of Labor and Congress of Industrial Organizations (AFL-CIO), proposed that a board subcommittee investigate the matter further, He was voted down II to 2. (William E. Powers, chief of radiology at Detroit's Harper Grace Hospital, sided with Samuels.) The problems with the benzene mono- graph apparently troubled the special review group at the cancer institute that evaluates funding proposals, including IARC requests. Thisgroup, comprised of scientists outside the cancer institute, told Nd, in effect, to mind its own business and slop meddling in the agen- cy's affairs. According to a memo writ. ten by a cancer institute official who attended the meeting, the committee "believes that IARC should remain open to suggestions from NCI , ,, , but it would be a mistake for NCI to use its financial leverage to influence unduly the selection of topics or the choice of indi- viduals to participate in thereviews." Obey, who is a member of the House Appropriations Committee that oversees the NCI budget, has promised to contin' ue investigating the matter. He said in a recent statement that he finds it "diffi' cult to believe that the extraordinary steps taken by IARC staff in altering the findings of a scientific panel without ap- proval from that panel were not at least partially a result of pressure from the National Cancer Institute officials who control IARC fttnding.'-MAtuotuE SUN 22-143 O-83----3fl PAGENO="0242" LETTERS IARC Benzene Report Following the article "Risk estimate vanishes from benzene report" by Mar- jorie Sun (News and Comment, 3 Sept., p.9l4) I should like to clarify the follow- ing points. 1) In the heat of the argument con- cerning benzene, it seems to have gone unnoticed that volume 29 of the IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Hu- mans (1) published by the International Agency for Research on Cancer (IARC) contauss 18 monographs on individual chemicals, one of which is benzene. The monograph on benzene (1, pp. 93-148) contains a critical review of all available information on this chemical and ends with an evaluation which reads: "there is sufficient evidence that ben.zene is carci- nogenic to man." 2) The annex to volume 29, with the title "Some aspects of quantitative can- cer risk estimation" (1, pp. 391-398), is the first attempt the IARC has made to explore the possibility of making quanti- tative cancer risk estimations. While such quantitative estimates may eventu- ally become incorporated into the IA.RC Monographs, the present annex is in - way an integral part of the Monographs. The Working Group which met in Octo- ber 1981 "recommended to IARC that a special monograph be prepared on quan- titative risk estimation" (1, p. 391). 3) During the October 1981 meeting that prepared volume 29, there was no detailed or in-depth discussion about methodologies for extrapolation of can- cer risks. As is stated in the summary remarks of the annex (1, p. 396), "the Working Group restricted their analyses to data available in published form and kept extrapolations to a minimum." With this in mind, the annex was later revised with the aim of making it a solid scientific document that did not attempt to provide risk estimates beyond what the data permit and would represent a sound initial step on which a program to explore the feasibility of making quanti- tative risk estimations could be built. The text as it now stands reflects the quantitative data derived from published epidemiological studies. On pages 395 and 396, there is a com- plete and objective summary of the avail- able evidence of risks derived from ex- posure to benzene. It is clearly indicated that, at 100 parts per million, the estimat- ed relative risk for leukemia is increased more than 20-fold. Risks of this magni- tude should attract attension to the possi- 2t4 238 .Sc.it~, OcF,š~i bility of significant risks at much lower levels. The IARC felt, however, that the data were insufficient to quantify pre- cisely risks at lower levels. ttis possible that many people have not read the an- nex in its entirety. 4) Several officials of the National Cancer Institute have expressed their concern about the issue of quantitative risk assessment. It is clear that, at least in part, this concern is related to the considerable adverse reaction caused within the scientific community by a document on quantitative risk estimates from occupational exposures prepared some time ago by scientists at a number of federal agencies, including the Nation- al Cancer Institute (2). The opinions of national institutes have been, and always will be, taken into consideration, but this in no way implies that IARC has been or will be ready to accept any interference itt its activities and decisions. LORENZO T0MArts International Agency for Research on Cancer, 150 cours Albert Thomas, 69372 Lyon, France Retereao's t. IARC Mono~top1ct on the Ecalaotiote of the Carcinogenic Risk of Chemicals no fiamass. vol. 29. Some lndaotnisl Chemicals and Dye- staffs (tnternatior.at Agency for Research ore Cancer, Lyon. France. t982). 2. St. Pent and M. Sohneidecmao. Ed,.. Qaanr9l. cation of Occcpsnior-.al Cancer (RanburyReport 9, Cold Spring Harbor Laborutony, Cold Spring Harbor. N.Y.. 991). American Participation in HASA We welcome the editorial by Jurgen Schmandt (10 Sept., p. 987) on the Inter- national Institute for Applied Systems Analysis (IIASA). We are glad to take this opportunity to let the American sci- entific community know that American participation will not cease with the end of this calendar year, when the National Academy of Sciences withdraws as the National Member Organization. The American Academy of Arts and Sciences in Boston will take up the role as Nation- al Member Organization for the United States as oft January 1983 and is seeking to raise from foundations and corporate sources enough money to cover the dues and related administrative expenses for the next several years. We cannot agree with Schmandt's view that the lesson to be drawn from the Administration's decision to withdraw from IIASA is that the "design was too complex and the goals too ambitious." We think the lesson is a simpler one: short-term and ideological consider- ations were given too large a weight in the government's decision, to stop fund- ing. We expect to demonstrate by con- tinued participation of abte American scientists that IIASA's work is of sub- stantial intellectual value and that it can continue to make an indispensible contri- bution to the understanding and resolu- tion of problems that go beyond the ideological divisions and barriers in the world today. CAR!. KAYSEN Program in Science, Technalogy, and Society, Massachusetts Institute of Technology, Cambridge 02139 ROGER LEVIENC Xerox Corporation, Stamford, Connecticut 06904 - HowAnen RA!FFA Harvard Business School, Boston, Massachusetts 02163 `Ameeican Academy of Acts and Scieocos Special Comoaooo onttASA. PAGENO="0243" 239 ATTACHMENT VII UNITED STATES ENVIRONMENTAL PROTECTION AGENCY WASHINGTON, D.C. 20460 RESEARCH AND OEVtLOPMENT SUBJECT: Assessment of Risks to Persons Coning into Contact with Contaminated Areas at Imperial, Missouri FROM: Elizabeth L. Anderson, Directot~.,~ Of fice of Health and Environmental Assessment (RD-689) TO: Gene A. Lucero, Acting Director Office of Waste Programs Enforcement (WH-527) As noted on the cover of the report entitled "Exposuure Assessment of Missouri Horse Arenas and Related Areas" distributed to you on Novembber 17, 1982, the material was not peer or administratively reviewed. As part of the review process, we have updated estimates presented in the November 17 version of the report. As a result of changes in exposure estimates, one change was also made in a risk estimate. Changes and correctic~ns were made in the following tables: o Dioxin Contamination in Missouri Cancer Risks o Table 4 - Exposure from Dust Inhalation at Shenandoah Stables o Table 5 - Exposure Resulting from Ingestion of Soil by Children with Pica Attached is a revised version with the corrected tables and minor editing changes. Non&of these changes affect the original risk estimates, except the high estimate for reproductive risk to stable workerS where the ratio of exposure to ADI was reduced from 310 to190. Please substitute this version for the one issued on November 17, 1982. One more revision is scheduled upon completion of peer and administrative reviews. A copyof the final report will be sent toyou. Attachments - 3 PAGENO="0244" ESTIMATE OF REPRODUCTIVE RISK* FROM INHALATION OF DIOXIN CONTAMINATED DUST AT SHENANDOAH HORSE ARENA Low Estimates High Estimates Estimated Number of Ratio** Estimated Number of Ratio Popul ation Segment Exposure Days Exposure/AD! Exposure Days Exposure/AOl ng/kg/day Exposed ng/kg/day Exposed 1. Stable workers 3.0 x iO~ 500 3.0 1.9 x 3000*** 190 2. Riders 3.0 x i0~ 94 3.0 1.8 x i~1 560 180 3. Spectators 1.1 x i0~ 6 1.1 7.1 x i~2 25 71 *Reproductive Risk depends on both daily exposure level, relative to the AD!, and the number of days exposed. Estimated Exposure Total Exposure/Exposure Duration **Acceptable Daily Intake (AD!) of 1 x if)~ ng/kg/day was calculated from Murray et al. (1979) study using a safety factor of 1000 applied to the lowest observable effect level. The application of a 1000 safety factor can be further justified because TCDD can bioconcentrate in body tissues. In addition, there is Some evidence that TCDD may bind to macromolecules (DNA) and could be a potential mutagen. ***Although the high exposure estimat~ for stable workers was based on 5001) days (p. 12), the duration was reduced to 3000 days here to reflect the reduced period during a person's life when reproductive effects are a concern (i.e., ages 15~45). PAGENO="0245" DIOXIN CONTAMINATION IN MISSOURI Equivalent Lifetime Average Daily Exposure Estimates and Upper-Limit Lifetime Cancer R'isk* Associated-with TCDD Exposure in Shenandoah Horse Arena and Contaminated Soil of Two Residences. Exposure (ng/kg/day) Upper-Limit Risk Estimates Based on Low High Low Exp, High Exp. Estimate Estimate Estimate Estimate A. Inhalation of Dust from Shenandoah Stablest 1. Stable Workers 5.7 x i~-5 3.6 x i~-2 - -- 2.4 x ~°-~ 1,5 ~ 1~-2 - -- 2. Rider -- - 1.1 x 1O~ 4.1 x iO~ 4.7 X 1O~ 1.7 x iO~ 3. Spectator 2.7 x iO~ 7.0 x 1O~ 1.1 x iO~ 3.0 x B. Child Ingestion of Dirt from Residences~ 1. Minker Residence (soil 6.2 x iO'~ 3.0 2.6 x103 0.72 concentration 550-900 ppb) 2. Stout Residence (soil 2.3 x ~ 7.0 x 10-2 9.8 ~ i~-6 3.0 x 10-2 concentration 2.1-21 ppb) *Upper_ljmit slope estimate of Qi= 4.25 x101 (ng/kg/day)4 tDust levels between 0.4 and 4 mg/rn3 with dioxin contamination approximately 100 ppb. žAverage 17.5 kg child ingests contaminated dirt between 50-2500 mg/day, 100 to 200 days/year, 1 to 3 years0 PAGENO="0246" 242 ~~RAFT EXPOSURE ASSESSMENT OF MISSOURI HORSE ARENAS AND RELATED AREAS November 22, 1982 ,John Schaum, Environmental Engineer Exposure Assessment Group This document has not been peer or administratively reviewed within EPA and is for Agency use/distribution only. PAGENO="0247" 243 TABLE OF CONTENTS 1.0 INTRODUCTION 2.0 BACKGROUND 3.0 SHENANDOAH STABLES 4.0 MINKER AND STOUT RESIDENCES~ 5.0 CONCLUSIONS AND RECOMMENDATIONS PAGENO="0248" 244 EXPOSURE ASSESSMENT OF MISSOURI HOURSE ARENAS AND RELATED AREAS 1.0 INTRODUCTION The purpose of this report is to analyze the human exposure to 2,3,7,8-tetrachlorodibenZO-P-diOXin (abbreviated to simply TCDD In remainder of report) occurring at the following sites near St. Louis, Missouri: o Minker Residence o Vern Stout Residence o Shenandoah Stables The report is in response to a request from the Office of haste Program Enforcement who is evaluating the need for enforcement actions relating to these sites. The report discusses the release routes by which TCDD may escape from the sites, the size of these releases, the resulting levels of human exposure and possible number of people exposed. The health and environmental effects which these releases may cause and possible remedial actions are not discussed. 2.0 BACKGROUND The TCDD contamination at the horse arena and related sites involves a complicated series of events which will not be described in detail here. The facts most essential to this study are as follows (Harris 1982): Shenandoah Stables o - On May 25, 1971 2,000 gallons of TCDD contaminated oil was sprayed on the horse arena at the Shenandoah Stables for purposes of controlling dust emissions. The oil contained 306-356 ppm TCDI) (approximtely 6 lb). PAGENO="0249" 245 o On August 22, 1971 the top 6-8 inches of soil (350 yards) In the arena was removed and replaced with clean soil. The contaminated soil was used as fill under a highway. o On April 1972 790-1060 yards of soil were removed from the arena and used as fill in a slough area adjacent to the stables. Bubbling Springs Area o On June 11, 1971 the Bubbling Springs horse arena was sprayed with TCDD Contaminated oil to control dust emissions. The oil contained an estimated 3 lb TCDD. o In March 1973 850 yards of soil~ were removed from the arena and used as fill at the Minker and Stout residences. o The Romaine Creek is located near the Minker residence and Bubbling Springs Arena and has apparently been contaminated with dioxin which eroded from the Minker residenc~. The monitoring data collected at these sites is summarized in Table 1 and Figure 1. The physical and chemical properties of TCDD are presented in Table 2. PAGENO="0250" 246 TABLE 1. MONITORING DATA Date TCDD Concentration (ppb) Shenandoah Stables o Arena August 1971 32,000-33,000 o Middle/Perimeter of Arena: 0-6' April-May 1982 75/127 6-12" " 20/101 12-18" 1 12/101 18-24" H 3/23 24-30" " 1/14 o Slough: - 0-6" 113 6-30" 1750 Minker Residence o Fill Area August 1974 85, 380, 740 o Fill *Area (2-18') April-May 1982 550, 660 o 75' Downgradient from Fill ` 900 Stout Residence o Fill Area August 1974 170, 430, 440 o Fill Area April-May 1982 10, 21 Source: Harris, 1982 3 PAGENO="0251" 247 FIGURE 1. ROMAINE CREEK Meramac River Romaine Head (1) Creek (2) (3) (4) Waters -~ ~rii~i Minker Shenandoah Residence Stables Scale: 1 inch = 1 mile Sampling Point TCDD Level in Sediment (ppb) 1. Headwaters <0.02 2. 800 ft upstream of Shenandoah Stables 0.32. 3. 1/3 mi. downstream of Shenandoah Stables 0.79 4. 2 mi downstream of Shenandoah Stables 0.30 Source: Harris, 1982 PAGENO="0252" 248 TABLE 2. PROPERTIES OF.TCDI) cl~O~l Structure of 2,3,7,8-TCDD: Molecular Weight: 322 Vapor Pressure: 10-6 rnmHg (estimated) * Sblubility in Water: 0.2 ug/l Octanol-Water Partition Coefficient: 6.9 x 10-6 (calculated) Source: Maybe, et al . 1981 5 PAGENO="0253" 249 3.0 SHENANDOAH STABLES Fate of TCDD Difficult to Explain The original TCDD containing matÚrial was sprayed on the arena in May 1971. A few months later 15-20 cm of son was removed and replaced with clean soil. TCDD's low solubility and tight adhesion to soil particles indicates It is not likely to migrate very far, especially in such a short time. Thus, one would expect that the TCDD would have remained near the surface and have been removed with the top 15-20 cm of soil. In April 1972 another 46-61 cm were removed and replaced with clean soil which even further suggests that all of the initial TCDD should have been removed. However, the 1982 sampling clearly shows that considerable TCDD still remains in the soil. Furthermore, the current TCDD contamination is highest near the surface where supposedly clean dirt was placed. We cannot offer any definitive explanation concerning the fate of TCDD In the arena soil. The facts do suggest that the TCDD appears quite persistent in this particular situation. The following exposure estimates are based on the assumption that the mostrecent monitoring results reflect conditions that will continue for a number of years. However, the uncertainty in our understanding of TCDDs fate, emphasizes the need to continue monitoring to confirm If TCDD levels change. Air Route Probably Causes Highest Exposure Levels A large barn type structure encloses the arena, stables and club area. The arena has a dirt floor measuring 190 x 75 feet. Workers, riders, or spectators who spend time in or near the arena could be exposed to TCDD via a number of routes: 0 Inhalation of TCD0 vapor or TCDD contaminated dust - Dust and vapors PAGENO="0254" 250 generated in this area will accumulate in the air to higher levels than would occur in an outdoor area where much better ventilation occurs. Site visitors report that high dust levels and odors are readily detected. o Direct ingestion - Dust which settles on food or dirt on hands which Is transfered to the mouth could be ingested. This route would be of particular concern where children with pica tendencies are Involved, however these children are normally under 5 years old and unlikely to be spending much time in the arena. The exposure occurring via this route will vary considerably among individuals depending on their habits regarding protecting food, washing hands, etc. Due to difficulties in making accurate estimates and belief that the exposure resulting from this route is small compared to others, the direct exposure route was not analyzed further in this study. o Absorption through the skin - Dust which settles on exposed skin or direct contact with the dirt provides the opportunity for TCDD to be absorbed through the skin. This form of exposure is reduced by several factors: 1) Much of the skin area is covered by clothing; 2) the contact time may be short; and 3) the TCDD must be removed from the particles before passing through the skin. o Others - Releases to the groundwater and associated exposure via drinking water consumption are unlikely. The arena surface is ocassionally watered to reduce dust emissions, but probably not enough to produce any significant infiltration. Furthermore, TCDD has a very low water solubility and is tightly bound to soil making it very immobile in the subsurface environment. Similarly, releases to surface waters and associated exposure are virtually impossible since no 7 PAGENO="0255" 251 hydraulic connections exist.~ The above points suggest that the air route would probably cause the highest exposure levels and this study was limited to evaluate only it in depth. However, some level of exposure via ingestion and direct contact probably does occur and should be considered in future analyses. Dust Levels Probably High But Difficult to Estimate Dust levels inside a building depend on the following four factors: o Source - The dirt floor of the arena is the primary source of *particulates. Site visitors report that the arena is covered with a light, fine, and usually dry soil which would tend to maximize dust emi ss ions. o Activity - The activity of people and horses cause air turbulence and physical agitation of the dirt increasing dust emissions. o Ventilation - The building is~ ventilated exclusively by the natural flow of air through doors, windows and cracks in the wall. The corrugated metal walls and roof are not insulated and would not provide a tight seal against drafts. Thus, the ventilation rate would probably exceed that of most homes which are typically more tightly constructed. The actual rates will vary depending on wind direction/velocity and whether doors/windows are open or closed. o Controls - As reported earlier, stable operators spray water or oil on the arena floor to reduce dust emissions. The effectiveness of spraying, particularly with water, diminishes rapidly with time. Although the ventilation and spraying controls will reduce dust levels, the ample particulate source and high level of activity suggest that the levels will be high. The fact that spraying is necessary and observations of PAGENO="0256" 252 dusty conditions by visitors support this opinion. Ideally, the dust levels should be determined through actual measurements. In the interim, however, - some idea of the possible levels can be judged from the following: o The OSHA standard for the respirable portion of nuisance dust is 5 mg/rn3 (Committee on Indoor Pollutants, 1981). o Crichiow et al. (1980) measured dust levels in a riding stable and - found them to Óverge 0.41 mg/rn3 (30 to 40% of this total were determined to be in the respirable range). o NIOSH has estimated that dust levels could reach 10 mg/rn3 in horse arenas (personal communication from Michael McCowley to Judy Bellin, November 1982). Since ventilation rates, activity levels, and soil types may differ between the Shenandoah Stable and one studied by Crichi ow et al., the dust evels may differ also. The NIOSH estimate suggests that the Crichiow measurement represents the low end of possible dust levels, but it seems unlikely that the dust could exceed 10 times an actual measurement. Thus, for purposes of this study, it was assumed that the total dust levels range from 0.4 to 4.0 mg/rn3. We strongly recommend that a monitoring program be initiated soon to confirm the actual levels. The TCDD levels in the top 6 inches of soil were found to range from 75-127 ppb (Table 1). Assuming that the TCDO levels in the soil are the same as in the dust particles, the concentration in the air can be calculated as fol 1 ows: TCDD concentration in air = TCDD concentration in soil x total particulate concentration in air This method suggests that the total TCDD levels in air will range from 3 x i08 to 5 x 10~ mg/m3. 9 PAGENO="0257" 253 In order to calculate what porlion of the total dust causes human exposure, estimates must be made of the respirable fractjon* and fraction retained in the body. Based on analysis of the particle size distribution, Crichiow et al. (1980) estimated that 32-42 percent of the dust was In the respirable range for nose-breathers. Some of the exposed population may breathe through their mouth, particularly workers or riders during times of heavy exertion. For mouth-breathers none of the dust is filtered out in the nasal passages. Thus, this study assumes that the respirable fraction could vary from 0.32 to 1.0. After entering the lungs, approximately 25 percent of the smaller dust particles which would not be filtered In the nasal passages Is exhaled. The remainder of the dust either stays in the lungs or is expelled by ciliary action and swallowed (International, Commission on Radiological Protection, 1968). Accordingly, this study assumes the 75 percent to 90 percent (for mouth breathers)** of the dust respired is retained in the body. Three population groups were selected to represent a variety of exposure times: stable workers, riders and spectators. Based on assumptions about the duration and frequency of exposure events, a range of exposure times were derived for each group (summarized in Table.3). Mr. Robert Shiflet of the Palomino Horse Breeders of M~erica (Personal Communication, October 20, 1982) confirmed the reasonableness of these assumptions. Based on the above assumptlons,~exposures were calculated for each of these groups (summarized in Table 4). The exposures were expressed in three ~As used here, this rÚfers to the portion of the total dust in air which enters the lung. **Assumjng that 37% of the dust is composed of the smaller particle~ (an average of the 32-42% measured by Crichiow), then mouth breathers will retain 75% of the smaller particles (0.75 x 0.37 = 0.28) and all of the heavier particles (1 - 0.37 = 0.63) for a total retention of 90% (0.28 + 0.63). 22-143 O-83------17 PAGENO="0258" 254 ways as defined below: Total = fExposure\ (~nhalation\(Total TCDD~\(~espirable'~fFraction\ /` 1 Exoo~ Duration) \, Rate )~Level in Air)L~FractionJ~Retained)~ˇd3' Weight)' in Body Actual Daily = Total Exposure Exposure Exposure Duration Average Daily = Total Exposure Lifetime Exposure (70 yr)(365 days/yr) Figure 2 ShOWS how the exposure varies with dust levels and allows one to quickly determine the exposure level associated with different dust levels than used to make the exposure estimates in Table 4. Vapor Emissions Require Further Study The vapor pressure of TCDD has not been measured, but estimated at 10-6 mm Hg by Maybe et al. (1981). This low vapor pressure has been widely assumed to indicate that very little TCDD would evaporate from contaminated soil. However, many investigators are now discovering that low vapor pressure compounds which also have low water solubility evaporate more readily from soil. Thus TCDD may have an enhanced volatilization rate from the arena soil particularly after it has been sprayed with water to control dust. This phenomena has been confirmed in field measurements with similar compounds such as DDT which has a vapor pressure of 1.9 x mm Hg and water solubility of 5.5 x ppm (Maybey et al. 1981). For example, Spencer (1975) measured the flux rate of DDT vapors from soil as 5-10 kg/ha/yr. The methods for predicting such volatilization rates have not been widely reviewed or well validated, so this study does not provide any quantitative estimates. However, these facts suggest that TCDD vapors could cause exposure inside the arena and that any future monitoring efforts attempt to measure the vapor levels in the air. PAGENO="0259" 255 TABLE 3. EXPOSURE TIMES FOR THREE POPULATION SEGMENTS Total Population Segment Hours/Day Days/Week Weeks/Yr. Years Exposure Time * . (days) Stable Worker A 50 500 B 50 5000 Rider A 50 94 B 50 560 8 6 8 6 3 3 3 3 Population Segment Hours/Show Spectator A 3. B 3 Total Years Exposure (days Time 10 5 6 10 30 25 5 50 5 30 Shows/yr. PAGENO="0260" E ci) Figure 2. Dust Levels vs. lixposure Levels at Shenandoah Ac'en.a .1 1.0 0.1 01 Average flaily Lifetime Exposure (ng/kg/day) 10-i PAGENO="0261" TADLE 4. EXPOSURE P0011 OUST INHALATIOR AT SHENANDOAH STADLES Stable Workers Riders Spectators Royal ation Segment Low High Low High i.ow 11gb Exposure1 Duration (days) 500 5000 94 560 6 25 Inhol ation2 Rate (m3/day) 29 29 29 29 11 11 Total TCDO level in Air 0.03 0.5 0.03 0.5 0.03 0.5 (ngfn3) Respirahle 0.32 1.0 0.32. 1.0 . 0.32 1.0 F actim Fraction Detained in 0.75 6.9 0.75 0.9 15.75 0.9 Rudy Rudy Weight3 (kg) 7Q 70 70 70 70 76 Total Exposure4 (ng/kg) 1.5 9.3 X 102 2.8 u 101 1.0 x 102 6.8 a i~-~ 1.8 Average Life- time Daily Exposure5 (ng/kg/day) 5.7 io5 3.6 0 10-2 1.1 X ~ 4.1 x iŘ~ 2.7 5 ~ 7.0 5 1fl~ Actual Daily Exposure6 (nq/k~/day) 3.0 u 1O~ 1.9 x 16-1 3.0 u 163 1.8 u 10~ 1.1 o ~ 7.1 x 1. See Table 3. 2. Snyder 1975. Adult male, light activity - 20 1/mm o~ 211 m3/day Adult nale, resting - 7.5 1/mix or 11 rn/day 3. Abraham 1979. Adult rn~le - 70 kg. 4. Tutal (Eopasure\(lnhalotion~fiotal TCD1f~(1iespirable\fFraction\( 1 \ Eopnsure lpuration)(~ Rote /~in Air )(~raction 4OetamnedJI~l~?iYWeight) S. Aeeragn Daily Lifetime Exposure Total Expasore `~ ` 6. Actual Daily Eoposiire Total Eopasure/Exposore Duration PAGENO="0262" 258 4.0 MINKER AND STOUT RESIDENCES Direct Ingestion Probably Causes Highest Exposure As discussed earlier, TCDD containing soil was removed from the Bubbling Spring horse arena and used as fill at the Minker and Stout residences. The filled areas at both sites are located on steep slopes, are covered sparsely with weeds, and show obvious signs of erosion (Personal communication from Daniel Harris, EPA Region VII, September 1982). The similarity in site conditions suggest that the possible exposure routes are also similar. Accordingly, the following analysis of the possible exposure routes are equally applicable to either site: o Ingestion of the contaminated soil by children, particularly those with pica tendencies, is quite possible since children live at both residences and in the surrounding neighborhoods. o Direct contact with the soil provides the opportunity for absorption of TCDD through the skin. As discussed in 3.1, this appears.unlikely. o Inhalation of dust or vapors could also occur. The emission rates and dispersion would vary greatly depending on meteorlogic conditions. However, the ambient air would dilute any emissions much faster than occurs inside the Shenandoah horse arena where ventilation is much more restricted. o Exposure via drinking water consumption could occur if dioxin leached into the groundwater. However dioxin's low water solubility and strong adsorption to soil particles suggest that l~ching would be very slow. o Finally, eroded fill material could be carried by run-off to local streams, where the TCDD C~uld7bioconcentrate through the aquatic food chain and ultimately%esult in human exposure via fish consumption. The Minker residence is located near the headwaters of PAGENO="0263" 259 the Romaine Creek where dioxin has been detected in the sediment. The above points imply that the potential for exposure is highest via direct ingestion of contaminated soil. Accordingly, this study examines only this route in depth. However, the others should be analyzed more carefully in any further studies. Exposure Via Ingestion Depends on Soil Consumption Rates The amount of TCDD contaminated dirt which children may ingest as a result of playing in the filled areas at the Stout and Minker residences is very difficult to estimate. The ingestion rates will depend on: o the pica tendencies of the children; o the time which children spend in the contaminated areas; and o the soil conditions (i.e., if the ground is frozen less dirt will get on to children's hands). Based on measurements of the amount of dirt found on children's hands and observations of mouthing frequencies, Lepow (1975) estimated that children with pica ingest at least 100 mg of dirt per day. Since children probably spend less than a full day playingat the contaminated areas, for purposes of this study it was assumed that the low estimate of ingestion rate would be 50 mg/day. The above estimate does not account for direct ingestion of soil which could increase daily ingestion rates to 5 g/day (personal communication from Julian Chisolm, Baltimore City Hospital, US EPA, November 1982). Consequently, the upper estimate for ingestion accounting for length of exposure/day, was assumed to be 2.5 g/day. In order to estimate the duration of exposure, it was further assumed that: o the number of days which children played on the site when the ground was not frozen would vary from 100-200 days/yr; and PAGENO="0264" 260 o a child would spend 1-3 yr playing in the area while exhibiting pica tendencies. Finally, assuming that TCDD concentrations measured in the soil during the most recent monitoring, reflect the conditions present throughout a child's exposure time, the exposure can be calculated as follows: Lifetime Average Daily Exposure = (Ingestion Rate)(TCDD Level in Soil) Exposure Duration)(Fraction Absorb~] (70 yr life)(365 days/yr (17.5 kg oodyweight) For this exposure route it was assumed that 100% of the TCDD ingested was absorbed into the body. The bodyweight esti ate of~ 17.5 kg is based on an average 5 year old (Snyder 1975). Table 5 summarizes the exposure occurring via this route. Since much of the uncertainty is derived from assumptions regarding ingestion rate, Figure 3 was included to show how the exposure varies with ingestion rate. PAGENO="0265" 261 TABLE 5. EXPOSURE RESULTING FROM INGESTION OF SOIL BY CHILDREN WITH PICA - Parameter Ingestion Rate (mg/day) TCDD Concentration in soil (ppb) Duration of Exposure' (days) Lifetime Average Daily I Exposure2 (ng/kg/day) 6.2 x ~ 3.0 2.3 x i0-~ 7.0 x 10-2 `Low estimate assumes child is exposed lOOdays/yr for one year and high estimate assumes child is exposed 200 day/yr for 3 yr. 2Lifetime Average Daily Exposure = (Ingestion Rate)(TCDD Level in Soil)(Exposure Duration)(Fraction Absorbed) (70 yr) (365 days/yr) (17.5 kg) 50 2500 50 2500 550 900 2.1 21 * 100 600 100 600 Minker Residence Low Estimate HighEstimate Stout Residence Low Estimate High Estimate 100% absorption was assumed. PAGENO="0266" Figure 3. Soil Ingestion Bate vs. r~osiire 0 C 0 .~1 U) a) C .~ c5~ 1 ~ io2 Average I)aily Lifetime Bxposuro (ng/kg/day) PAGENO="0267" 263 5.0 CONCLUSIONS AND RECOMMENDATIONS This study focussed on two potential exposure routes: o Inhalation of TCDD dust or vapors is the exposure route of most concern at the Shenandoah Horse Arena. We estimate that dust inhalation could result in the following range of lifetime average daily exposures for the various population segments: Stable Workers: 5.7 x - 3.6 x 10-2 ng/kg/day Riders: 1.1 x iO-~ - 4.1 x 1O~ ng/kgfday Spectators: 2.7 x 1O~ - 7.0 x iO~ ngfkg/day For the duration of the exposure, we estimate the following range of average daily exposures for the various population segments: Stable Workers: 3.0 x 103 - 1.9x 1O~ ng/kg'day Riders: 3.0 x 1O~ - 1.8 x io1 ng/kg/day Spectators: 1.1 x iO~ - 7.1 x 102 ng/kg/day o Direct ingestion of TCDD contaminated soil, particularly by children * with pica, represents the exposure route of most concern at the Minker and Stout residences. We estimate the following range of lifetime average daily exposures could occur to children with pica: Minker Residence: 6.2 x iO~ to 3.0 ng/kg/day Stout Residence: 2.3 x 1O~ to 7.0 x 10.2 ng/kg/day We recommend that future analyses consider the other exposure routes in more depth. In particular, the exposure due to vapor inhalation at the Shenandoah Horse Arena; and exposure resulting from consuming fish caught from the Romaine Creek or downstream waters which may contain TCDD eroded from the Minker residence should be examined more closely. Finally, we strongly recommend further monitoring of TCDD dust and vapor levels in the Shenandoah Horse Arena; and TCDD levels in fish from Romaine Creek and downstream waters. PAGENO="0268" 264 REFERENCES Abraham. S. 1979. Weight and Height of Adults 18-74 Years of Age, U.S., Committee on Indoor Pollutants. 1981. Indoor Pollutants, National Research Council, Washington, DC Crichlow, E.C., Y. Yoshida, and K. Wallace. 1980. Dust Levels in a Riding Stable, Equine Veterinary Journal , Volume 12, Issue 4. Harris, Daniel J. May 3, 1982. Memorandum to Scott Richey: Briefing on Horse Arena Investigation. Lepow, Martha, et al. 1975. Investigation into Sources of Lead in the Environment of Urban Children. Environmental Research 10:415-426. Maybe, W.R. et al. 1981. Aquatic Fate Process Data for Organic Priority Pollutants, Final Draft Report. EPA Report No. 440/4-81-014. Snyder, W.S. 1975. Re2ort of the Task Group on Reference Man. International Commission on Radiolugical Protection No. 23. Pergamon Press. Spencer, W.F. 1975. Movement of DOT and It's Derivatives Into the Atmosphere. Residue Reviews, Vol. 59. 21 PAGENO="0269" 265 Mr. FL0RI0. Thank you very much. Dr. Neal. STATEMENT OF ROBERT A. NEAL, PH.D. Dr. NEAL. Let me preface my remarks by saying that the CuT as an institute is independent of its funding companies by virtue of the timing and release and content of its scientific deliberations so that the statement I am presenting today is my own personal view and may not represent the official view of the chemical industry as a whole or individual companies that support the Institute, and I would like to proceed on that basis, if I might. The letter inviting me to appear today named a number of issues on which they would like me to address, and I will address those, some brief remarks to each of those. One of the issues was mechanisms of carcinogenesis. I think it is clear from the research at our Institute and others that chemicals do cause cancer by a variety of different mechanisms. A number that we have examined produce cancer by in fact causing irreversa- ble changes in the genetic information and code in the DNA of so- matic cells. Among other compounds we are studying in considerable detail is one that produces tumors in the livers of rats and mice, but does not appear to produce irreversable changes in the structure of DNA as measured in a number of systems designed to detect these genetic changes. A third example are chemicals that cause cancer~of the urinary bladder in rodents as a result of precipitation of large crystals or stones of that chemical in the bladder. These stones .provide a con- stant source of irritation to the lining of the bladder and they even- tually lead to the appearance of bladder tumors. To :form the bladder stones, the amount of the compound that must be fed to the rodents is very high and stone formation dis- plays a sharp threshold. In other words, there is an amount of the chemical in the diet above which bladder stones form that eventu- ally cause a precancerous and cancerous lesions to appear and below which no stones form and no precancerous or cancerous le- sions are seen. This could be referred to as nonspecific promotion. To sum up this segment, then, there are, as I have noted before, a number of mechanisms by which chemicals can cause cancer in experimental animals and man. The progress in the last few years in understanding of carcinogenic mechanisms and in experimental analysis technique have allowed us to begin to differentiate be- tween the mechanisms of chemical cancer induction. Another issue was the predictability of rodent bioassays. The re- sults of rodent bioassays have been* the major source of data for es- timating human risk from exposure to potentially carcinogenic chemicals. In my opinion, the rodent bioassay will continue for the foreseeable future to be the major means of identifying potential human carcinogens. However, the rodent bioassay as currently designed is believed to be only approximately predictive of the potential of the chemical to produce cancer in humans. In a recent publication, Purchase exam- ined data on 250 chemicals that had been tested for carcinogenicity PAGENO="0270" 266 in rats and mice. The utility of these data is they allow us to deter- mine how accurately the mouse bioassay can predict for cancer in rats and vice versa. Thus, a rat from this data-the rat was a pre- dicter of cancer in the mouse with the specificity of about 85 per- cent and the mouse predictor of cancer in the rat with a specificity of about 82 percent. I think it is reasonable to assume that the ability of the bioassay in the rat or mouse to predict for the potential or lack of potential of a chemical to cause cancer in man is probably no better than their ability to predict cancer in each other. Thus, the rodent, as a model for detecting potential human carcinogens, has an implied degree of unreliability. However, it is the most sensitive and reli- able method currently available and we must continue to rely upon it. Short-term tests were another issue. Substantial evidence indi- cates that one step in carcinogenesis is damage or mutation of the DNA. Many bacterial and mammalian cell culture assays can detect such chemically induced alterations. Because there is some correlation between activity of chemicals in these assays and car- cinogenic activity of these chemicals in rodents, these assays have been termed short-term tests for potential carcinogenicity. Thus, these short-term assays can be useful in qualitatively pre- dicting the potential carcinogenicity of a chemical to humans. The major shortcoming of short-term tests is that they do not provide reliable data on the relative potency of carcinogenic chemicals in experimental animals or man. It is apparent that for all carcinogenic chemicals there is a level of human exposure which is without significant risk. Thus, infor- mation on the carcinogenic potential of chemicals is necessary in - regulating human exposure and those data must come from animal studies. The major utility of the short-term tests currently available is to prioritize chemicals for commercial development and/or subse- quent animal testing. The short-term tests are also extremely valu- able in studies of mechanisms by which chemicals may cause an increase in tumors in experimental animals and humans. Turning now to the issue of human risk assessment for carcino- gens, it is currently the practice of Government regulatory agen- cies to use mathematical modeling to estimate cancer incidence in man from data in tests of rodents. For toxic effects of chemicals other than cancer, the risk assessment procedures is to determine a dose of the compound in experimental animals which does not pro- duce an effect, a no-observed-effect level (NOEL). To that NOEL a safety factor is applied to account for uncertainty in estimating ef- fects in man from data in experimental animals. There is a percep- tion on the part of some members of the regulatory and scientific community that there is a level of exposure to a carcinogen that is not without some risk. In other words, there may be an exceedingly sensitive individual within an exposed population. An extension of this reasoning is that the no-effect-level plus a safety factor ap- proach does not protect very sensitive individuals from cancer risk. To statistically estimate the number of persons at risk to small exposures, mathematical models have been developed. These models typically rely on carcinogenicity information from rodent PAGENO="0271" 267 bioassays. The major difficulty with mathematical models current- ly in use is that there is no experimental verification of the valid- ity of the models at low exposures where they are predicting cancer risk. In addition, the various models often fit the dose cancer inci- dence data generated in rodent bioassays equally well, but vary by orders of magnitude in their prediction of tumor incidence in man at low levels of exposure. Thus, in my opinion, there is currently no scientific justification for choosing the mathematical modeling approach over the conventional NOEL-safety factor approach for preventing excess cancer risk in humans at low exposure levels using cancer incidence data generated in experimental animals. In spite of these reservations 1 believe that an understanding of the toxicity and carcinogenicity. mechanisms operating in experi- mental animals and man will eventually provide the scientific basis for use of appropriately modified mathematical extrapolation models. Of particular interest in this regard are the recent publications which suggest that cancer incidence should be related to the con- centration of active carcinogen in the target organ rather than to the amount of the chemical to which the animal was exposed. In this way, more meaningful estimates of cancer incidence at low levels of exposure may be obtained. Some techniques for measuring the concentration of active car- cinogens at the target site are currently available and more are under development, a number of them in Dr. Weinstein's labora- tory. On the subject of genetic versus epigenetic carcinogens, which has been the subject of most of the presentations today; as I noted previously I am convinced that chemicals can cause an increase in cancer in experimental animals and man by a variety of mecha- nisms. I believe there is substantial evidence supporting the con- cept that some chemicals can cause an increase in cancer incidence by mechanisms other than directly causing an irreversible change in the structure of DNA. In other words, they can act as an epi- genetic carcinogen. With some of these compounds there appears there may be a threshold dose level, for example those compounds that form stones in the bladder. With others, the. biological effect leading to an in- crease in tumors is, under cŰrtÔin conditions of exposure, revers- ible. However, examining the data concerning genetic and epigenetic mechanisms of carcinogenicity I do not believe that our ability to differentiate between these two mechanisms of carcinogenicity is well enough established at this time for unqualified use of these data in regulating human exposure to potential carcinogens. However, I do believe that sufficiently well-validated in vivo and in vitro techniques are currently available to determine with rea- sonable certainty whether a compound is causing an increase in rat liver tumors by a genetic or epigenetic mechanisms. When techniques have been developed and appropriately validat- ed, it may be also possible to determine with reasonable certainty whether a compound is increasing the incidence of tumors in other tissues and in other species by a genetic or an epigenetic mecha- PAGENO="0272" 268 nism. Human exposure to those chemicals adequately identified as epigenetic carcinogens should, in my opinion, be regulated less con- servatively than compounds shown to increase tumor incidence in experimental animals by a genetic mechanism. The last issue is thresholds for carcinogens. I have also been asked to comment on that subject. As I noted earlier, there is a perception on the part of some of the regulatory and scientific com- munity that there is no threshold for chemicals which cause an in- crease in cancer incidence in animals by a genetic mechanism or, for some individuals, by any mechanism. It is argued that cancer may result in some individuals from the exposure to a single molecule of the chemical. This statement may be correct since as yet there is only limited and suggestive experi- mental evidence for the existence of thresholds for cargcinogens in experimental animals. However, it must be kept in mind-and this is important-that there is also no experimental evidence in ani- mals which disputes the existence of a threshold for cancer induced by carcinogens. Stated another way, there is no experimental evidence which either supports or refutes the existence of thresholds for cancer caused by exposure of experimental animals to a carcinogen. In my view, the scientific uncertainty concerning both the absence or ex- istence of thresholds should be kept in mind in the process of scien- tifically regulating human exposure to carcinogenic chemicals. I believe that for those compounds which can cause an increase in tumors by a mechanical rather than a chemical mechanism, i.e., bladder stone formation, a clear threshold can usually be demon- strated. For those compounds which cause an increase in liver tumors in rats but not other sites, in vitro and in vivo techniques are available which can show with reasonable certainty whether the compound is acting by a genetic or epigenetic mechanism. For those compounds where the mechanism of carcinogenicity has been shown to be epigenetic by appropriate tests, I believe a good case can be made for the existence of a practical threshold dose level under conditions where the exposure is for a limited por- tion of the animal's life span. I would like to thank the committee for the opportunity. [Testimony resumes on p. 281.] [Dr. Neal's prepared statement follows:] PAGENO="0273" 269 STATEMENT BEFORE THE SUBCOMMITTEE ON COMMERCE, TRANSPORTATION AND TOURISM OF THE HOUSE COMMITTEE ON ENERGY AND COMMERCE March ~7, 1983 Robert A. Neal, Ph.D. My name is Robert Neal. I am a toxicologist by profession. I am currently President of the Chemical Industry Institute of Toxicology, a non-profit research institute funded by 34 chemical manufacturing companies. CIIT occupies a 60,000 square foot research facility in Research Triangle Park, North Carolina. The current research and administrative staff at the Institute now numbers approximately 125. The major research focus at CUT is to determine the potential of widely used non-proprietary chemicals to adversely affect human health. CIIT is independent of the companies that fund it in terms of its research agenda and the content and release of research results. Therefore, the statement I am presenting today may not coincide with the official views of the chemical industry or any individual chemical manufacturing company that supports the Institute. The letter inviting me to appear today listed a number of issues about which I was requested to comment. I will address some brief remarks to each of these issues. Mechanisms of Carcinogenesis A substantial part of the research effort at CIII is directed toward a better understanding of the mechanisms by~which chemicals induce cancer in 22-143 O-83--18 PAGENO="0274" 270 experimental animals and humans. It is clear from research at our Institute and others that chemically induced cancer in mammals can occur by a number of different mechanisms. A number of the compounds which s~ie have examined at CIII appear to cause cancer by producing an irreversible change in the genetic information encoded in the DNA of somatic cells. We believe these compounds cause an increase in neoplasia because they are complete carcinogens. That is, they can cause an irreversible change in the structure of DNA in a cell or cells which then have the potential to produce a tumor. In addition, these compounds cause other, as yet unidentified changes in that initiated cell and/or surrounding cells which allow the cell to divide in an uncontrolled fashion. Among the compounds which we are studying in considerable detail is one that produces tumors in the livers of rats and mice but does not appear to produce irreversible changes in the structure of DNA as measured in a number of systems designed to detect these genetic changes. We have reached the tentative conclusion that this compound might be causing an increase in tumors by producing changes in the livers of rats which stimulate existing cells whose DNA has been altered by some other factor than the chemical under study (another chemical,. irradiation, etc.) to grow and produce a tumor. In other words the compound may be acting as an epigenetic carcinogen or promoter. Additional research, which is underway, will be necessary to establish with greater certainty that this is the -- mechanism by which this compound causes rat liver timors. Another example includes chemicals which cause cancer of the urinary bladder in rodents as a result of the precipitation of large crystals PAGENO="0275" 271 (~tones) of the chemical in the bladder. These stones provide a constant source of irritation to the lining of the bladder and may eventually lead to the appearance of bladder tumors. To form the bladder stones the amount of the compound which must b~ fed to the rodents is very high and stone formation displays a sharp threshold. In other words, there is an amount of the chemical in the diet above which bladder stones form that eventually cause precancerous or cancerous lesions to appear and below which no stones form and no precancerous or cancerous lesions are seen. Using Dr. Henry Pitot's nomenclature (1), the mechanism by which these compounds cause cancer in rodents can be referred to as nonspecific promotion. Thus, as noted previously, there are~ a number of mechanisms by which chemicals and other agents cause cancer in experimental animals and man. The progress in the last few years in our understanding of carcinogenic mechanisms and in experimental analysis techniques has allowed us to begin to differentiate between different mechanisms of chemical cancer induction. Predictive Value of Rodent Bioassays for Carcinogenicity. Theresults of rodent bioassays have~ been the major source of data for estimating human risk from exposure to potentially carcinogenic chemicals. ~n my opinion, the rodent bioassay will continue, for the foreseeable future, to be the major means of identifying potential human susceptibility *to carcinogens. However, the rodent bioassay, as currently designed, is believed to be only approximately predictive of the potential of a chemical to produce cancer in humans. In a recent publication (2), Purchase examined data on 250 chemicals which have been tested for carcinogenicity in rats and mice. Of the 126 chemicals found to be positive for cancer in rats, only 87% were PAGENO="0276" 272 positive for cancer in mice. Of the 119 chemicals found to be negative for cancer in the rat, only 82% were negative for the mouse. Conversely, of the 130 chemicals found to be positive for cancer in the mouse, only 84% were positive for cancer in the rat. Of ~he 115 chemicals which were found to be negative for cancer in the mouse, only 85% were negative for cancer in the rat. These data allow us to determine how accurately the mouse can predict cancer in the rat and vice versa. Thus, the rat as a predictor of cancer in the mouse has a specificity of about 85%. The mouse as a predictor of cancer in the rat has a specificity of about 82%. It is reasonable to assume that the ability of the rat or mouse to predict for the potential or lack of potential of a chemical to cause cancer in man is no better than their ability to predict for cancer in each other. Thus, the rodent as a model for detecting potential human carcinogens has some implied degree of unreliability. However, it is the most sensitive and reliable method currently available and we must continue to rely upon it. "Short-Term Tests Substantial evidence indicates that one step in carcinogenesis is damage to or mutation of the DNA. Many bacterial and manTnalian cell culture assays can detect such chemically induced alterations. Because there is some correlation between the activity of chemicals in these assays and carcinogenic activity in rodents, these assays have been termed "short-term" tests for potential carcinogenicity. Substantial progress has been made in the development of these assays so that one can now measure chemically induced alteration of DNA in bacteria, insects, maninalian cells in culture, human cells in culture and, in some cases, in cells from living animals exposed to these chemicals. Thus, these short-term test assays can PAGENO="0277" 273 be useful in qualitatively predicting the potential carcinogenicity of a chemical to humans. The major shortcoming of short-term tests is that they do not provide reliable data on the relative potency of carcinogenic chemicals in experimental animals or humans. It is apparent that for all carcinogenic chemicals, there is a level of human exposure which is without significant risk. Thus, information on the `carcinogenic potency" of chemicals is necessary in regulating human exposure. The major utility of the short-term tests currently available is to prioritize chemicals for commercial development and/or subsequent animal testing. The short-term tests are also extremely valuable in studies of the mechanisms by which chemicals may be~causing an increase in tumors in experimental animals or humans. Such tests in whole experimental animals are of more value in this regard. This is because these in vivo tests allow the chemicals to be subjected to the same absorption, distribution, metabolism and excretion processes as is the case in lifetime rodent bioassays for chemically induced cancer. Moreover, any damage to cells caused by the chemical in the short-term test would be subject to repair in the same manner as when the animals are exposed to the chemicals in the lifetime bioassay. What is clearly needed is additional information that will allow us to better assess the reliability of the rodent models as predictors of human cancer. The information most useful in this regard would be knowledge of the mechanisms of carcinogenicity of a subject chemical as determined by studies in the experimental animal species and strains in which the cancer occurred. Short-term tests with the compound that examine possible mechanisms of alteration of DNA structure~ would also provide valuable PAGENO="0278" 274 information. However, without question the most valuable information for determining the reliability of a particular rodent model in predicting human cancer would be comparative studies of the pharmacokineticS, metabolism and genotoxicity of the' compound in both humans and rodents. In these studies humans occupationally and accidentally exposed to the compound could be used. If no exposed human populations were available in which it was ethical to carry out these studies, experimental studies with human tissues obtained at autopsy or surgery might also provide valuable information, especially in comparison with identical studies in rodent tissues. After a substantial body of information has been accumulated comparing pharmacokinetics, metabolism, genotoxicity and other biological effects of chemicals in humans and rodent models, we would be in a better position to determine the predictability of our rodent models for chemically induced cancer in man. Human Risk Assessment for Carcinog~fl~ It is currently the practice of government regulatory agencies to use mathematical modeling to estimate cancer incidence in man from data in tests of rodents. Of course the rodent studies are conducted using exposures which are often orders of magnitude higher than the exposures which occur in humans. For toxic effects of chemicals other than cancer, the risk-assessment procedure is to determine a dose of the compound in experimental animals which does not produce an effect (no observed effect level; NOEL). To that NOEL a safety factor is applied to account for uncertainty in estimating effects in man from data in experimental animals. A safety factor of 100 is usually used but the magnitude of the safety factor varies with the PAGENO="0279" 275 * strength of the data and the toxic effect observed. There is a perception on the part of some members of the regulatory and scientific community that there is not a level of exposure to a carcinogen that is without some risk; ~in other words there may be exceedingly sensitive individuals within an exposed population. An extension of this reasoning is that the "no-effect level" plus a safety factor approach does not protect very sensitive individuals from a cancer risk. To statistically estimate the numbers of persons at risk to small exposures, mathematical models have been developed. These models typically rely on carcinogenicity information from rodent bioassays. The major difficulty with the mathematical models currently in use is that there is no experimental verification of the validity of the models at low exposures where they are predicting cancer risk. In addition, the various models often "fit" the dose-cancer incidence data generated in rodent bioassays equally well, but vary by orders of magnitude in their prediction of tumor incidence in man at low levels of exposure. Thus, in my opinion there is currently no scientific justification for~choosing the mathematical modeling approach over the conventional NOEL-safety factor approach for preventing excess cancer risk in humans at low exposure levels using cancer incidence data generated in experimental animals. In spite of the reservations I have stated above, I believe that an understanding of toxicity and carcinogenicity mechanisms operating in experimental animals and man will eventually provide the scientific basis for use of appropriately modified mathematical extrapolation models. Of particular interest in this regard are the recent publications of Hoel (3) and Starr (4) which suggest that cancer incidence should be related to the PAGENO="0280" 276 concentration of `active" carcinogen in the target organ rather than to the amount of chemical to which the animal was exposed. In this way more meaningful estimates of cancer incidence at low levels of exposure may be obtained. Some techniques for measuring the concentration of "active" carcinogens at the target sites are currently available and more are under development. Genetic versus Epigenetic Carcinogens In my comments I will use the term genetic carcinogen to describe a compound which directly interacts with the genetic material and causes a change in the structure of DNA that leads to a cell with the potential to grow and produce a tumor. The term epigenetic carcinogen refers to compounds that increase the tumor incidence in experimental animals and humans by some mechanism other than an irreversible change in the structure of the DNA of the cells in question. As noted previously, I am convinced that chemicals cause an increase in cancer in experimental animals and man by a variety of mechanisms. I believe there is substantial evidence supporting the concept that some chemicals can cause an increase in cancer incidence by mechanisms other than directly causing an irreversible change in the structure of DNA. In other words they can act as epigenetic carcinogens. With some of these compounds there appears there may be a threshold dose level and with others the biological effects leading to an increase in tumors is, under certain conditions of exposure, reversible. However, examining the data concerning genetic and epigenetic mechanisms of carcinogeneSiS, I do not believe that our ability to differentiate between these two mechanisms of carcinogenicity is well enough established at this time for the unqualified use of these data in regulating human exposure to PAGENO="0281" 277 potential carcinogens. I do believe that sufficiently well validated in vivo and in vitro techniques are available to determine with, reasonable certainty, whether a compound is causing an increase in rat liver tumors by a genetic or epigenetic mechAnism. Although short-term tests provide some indication of genotoxic potential, in vivo and in vitro techniques for differentiating between genetic and epigenetic mechanisms for cancer induction in organs and species other than rat liver have not yet been developed. When these techniques have been developed and appropriately validated, it may also be possible to determine with reasonable certainty whether a compound is increasing the inńidence of tumors in other tissues and other species by a genetic or an epigenetic mechanism. Human exposure to those chemicals adequately identified as epigenetic carcinogens should, in my opinion, be regulated less conservatively than compounds shown to increase tumor incidence in experimental animals by a genetic mechanism. Thresholds for Carcinogens I have also been asked to coment on the concept of thresholds for carcinogens and the implications of this concept for control of exposure to formaldehyde, 2,3,7,8-tetrachlorodibenzo..p.dioxin and other animal carcinogens. As I noted earlier in this statement, there is a perception on the part of some of the regulatory and scientific corrinunity that there is no threshold for chemicals which cause an increase in cancer incidence in animals by a genetic mechanism or, for some individuals, by any mechanism. It is argued that cancer may result in some individuals from the exposure to a single molecule of the chemical. This statement may be correct since, as yet, there is only suggestive experimental evidence for the existence of thresholds for carcinog~ns in experimental animals. PAGENO="0282" 278 However, it must be kept in mind that there is also no experimental evidence in animals which disputes the existence of a threshold for cancer induced by carcinogens. Stated another way, there is no experimental evidence which either suppJrts or refutes the existence of thresholds for cancer caused by exposure of experimental animals to a carcinogen. Cancer incidence in animals exposed to low-level radiation or the variation in mutation frequency with the concentration of chemicals in bacterial and marrrnalian cells in culture is cited as evidence for the lack of a threshold for chemical induced cancer in experimental animals and humans. However, the living marrmal is a much more complex system than in vitro cell cultures. The living animal provides opportunities for selective excretion or metabolic inactivation of a compound and repair of precancerous lesions caused by a chemical or its metabolites to a greater degree than is the case with most, if not all, in vitro systems. Also, other defense mechanisms such as inmiune surveillance are active in the whole animal. In addition, chemically induced cancer is, procedurally, a much more complex process than radiation-induced cancer. Thus, it is highly speculative to relate conclusions from low-dose radiation exposure and from these in vitro model systems exposed to chemicals to experimental animals and humans exposed to carcinogenic chemicals, particularly for those chemicals which must be first metabolized by the animal before they can cause changes in the cell which eventually lead to cancer. Thus, the existence or absence of a threshold for chemicals that causes cancer in experimental animals and man has not yet been experimentally verified. In my view the scientific uncertainty concerning both the absence or existence of thresholds should be kept in mind in the process of scientifically PAGENO="0283" 279 regulating human exposure to carcinogenic chemicals. In the specific case of formaldehyde, the data is currently insufficient to demonstrate the presenceor absence of a threshold for this compound for the productio~ of nasal cancer in rats. It is also unclear at this time whether formaldehyde causes an increase in nasal tumors in rats by a genetic or epigenetic mechanism. There is currently evidence which suggests that a number of processes are operative in the nose of rats which serve to limit the interaction of the formaldehyde with the target cells. The effect of these processes may be a practical threshold for exposure to formaldehyde below which the compound does not reach the target cells. While the data are insufficient at this time to unambiguously indicate the presence or absence of a thresho~d dose of formaldehyde for nasal cancer in rats, the presence of formaldehyde as a normal cell constituent indicates that a practical threshold must~ exist. In the case of TCDD, the available data suggest that this compound probably increases tumor incidence in rats and mice by acting as a potent epigenetic carcinogen or promoter. If further experimentation proves this to be the case, a threshold for the carcinogenic effect of this compound may exist. However, TCDD causes an increase in tumors in a number of tissues, tissues for which techniques for~ differentiating between enetic * and epigenetic mechanisms of carcinogenicity are not yet developed. Thus, the question of the existence of a threshold dose for the carcinogenic effect of ICOD must, in our opinion, await the development and validation of these techniques. * As concerns other carcinogens, I believe for those compounds which cause an increase in tumors by a mechanical rather than a chemical PAGENO="0284" 280 mechanism (e.g. bladder stone formation), a clear threshold dose can usually be demonstrated. For those compounds which cause an increase in liver tumors in the rat but not at other sites, in vitro and in vivo techniques are available to show with reasonable certainty whether the compound is acting by a genetic or an epigenetic mechanism. For those compounds where the mechanism of carcinogenesis has been shown to be epigenetic, I believe a good case can be made for the existence of a practical threshold dose level under conditions where the exposure is for a limited portion of the animal's lifespan. I wish to thank the Subcommittee on Coirmierce, Transportation and Tourism for the opportunity to present this statement. References 1. H. G. Pitot, T. Goldsworthy and S. Moran, The natural history of carcinogenesis: Implications of experimental carcinogenesis in the genesis of human cancer. J. Supramolecular Structure and Cellular Biochem., 17:133-146 (1981). 2. I.F.H. Purchase, Inter-species comparisons of carcinogenicity. Br. J. Cancer, 41:454-468 (1980). 3. 0. G. Hoel, N. L. Kaplan and M. W. Anderson, Implication of nonlinear kinetics on risk estimation in carcinogenesis. Science, 219:1032-1036 (1983). 4. T. B. Starr, Mechanisms of formaldehyde toxicity and risk evaluation, In FORMALDEHYDE TOXICOLOGY, EPIDEMIOLOGY AND MECHANISMS, Clary, 3. 3., Gibson, 3. E. and Waritz, R. S., e~ds., Marcel Dekker, Inc., (1983), in press. - PAGENO="0285" 281 Mr. FL0RI0. Thank you very much. Mr. Browning? STATEMENT OF JACKSON B. BROWNING Mr. BROWNING. Thank you, sir. I am pleased to appear before the committee today as a member of the board of directors of the American Industrial Health Council to discuss the importance of an adequate science base for regulatory decisions. This organization was formed in 1977 for the purpose of evaluat- ing regulatory or legislative proposals involving scientific issues re- lating to chronic health hazards and in supporting legislation and administrative actions which promote the use of the best science and regulatory decisionmaking, assure the adequacy of the science base for regulatory decisions. It does not act or appear as an advocate for any particular sub- stance and does not become involved in regulatory actions involv- ing particular substances. AIHC believes strongly that regulatory decisions should be based on the best science. Scientific evaluation necessarily involves not only the application of the latest scientific developments, but expert scientific judgment on science matters. Policy in this con- text does not mean predetermination of scientific issues or any ar- bitrary dictation to the scientists as to how to evaluate scientific data. Rather, sound policy involves the legal and regulatory frame- work to assure objective expert scientific evaluations. The national policy refers to a coordinated approach to the scientific identifica- tion and evaluation of risk to enable sound regulatory distinction between trivial and significant risks. AIHC believes that interagency cooperation is a critical part of a national policy as I have described it. It is essential that there be a coordinated approach to the use of the best science in the identifi- cation, characterization, and management of risk. AIHC has welcomed the efforts under the leadership of the Office of Science and Technology Policy to develop the national policy on cancer assessment. We have reviewed and commented on the discussion draft entitled "Potential Human Carcinogens, Meth- ods for Identification and Characterization," which was prepared by the Regulatory Work Group on Science and Technology. The draft is part 1 of a document which we understand is to cover the state of the art discussion-that is part 1-and a discus- sion of general principles applicable to individual agency regula- tory decisions, which will be part 2. AIHC has not seen a draft or outline of part 2 of the OSTP document. Nevertheless, AIHC be- lieves that the procedure for interagency cooperation is on the right track, and we expect to comment and be actively involved as the procedure moves ahead. The recent report of the National Academy of Sciences under- lines the importance of objective scientific evaluations as the indis- pensable basis for sound regulatory decisions. This report, which was issued on March 1, is entitled "Risk Assessment in the Federal Government: Managing the Process."~ PAGENO="0286" 282 A basic premise of the report is the distinction between the sci- entific risk assessment and risk management. Risk assessment is the scientific evaluation of facts and data and the characterization of the human risk. Risk management is the process of integrating the scientific evaluation with technical data and the broad econom- ic, social, and political statutory objectives to weigh policy alterna- tives and reach a regulatory decision. As described in the NAS report, the scientific function of risk as- sessment involves four steps: Hazard identification, the scientific process of gathering and evaluating data; dose response assess- ment, the evaluation of the magnitude of exposure and the prob- ability of adverse health effect; exposure assessment, which is data validation and evaluation; and risk characterization, the evaluation and magnitude of human risk. Similar reviews are expressed in a report of the Subcommittee on Environmental Carcinogenesis of the National Cancer Advisory Board, which is quoted in my full statement. AIHC has used somewhat different words to describe the scientif- ic functions, but it is in basic agreement with the NAS report on the nature of the scientific function and the distinction between and the relationship of the scientific evaluation to the regulatory process of risk management. The regulator must depend on the soundness of the scientific evaluation to provide a sound basis for the complicated and impor- tant judgments the regulator must make. Good regulatory deci- sions, therefore, require a scientific evaluation of the underlying scientific data, including toxicology, exposure, and other scientific parameters which are not biased by social, political, and economic value judgments of the scientists or by administrative or policy dic- tates to the scientists which would distort a full, objective scientific evaluation. This distinction between the scientific function and the regula- tory function is essential to what AIHC believes is the scope and content of a scientifically sound national policy on cancer assess- ment. It is axiomatic that we do not live in a risk-free society. The regulatory job is to separate the significant from the trivial risks. Regulatory focus on trivial risks is, of course, a waste, and takes away from attention to significant risks. Science has steadily increased its analytical capability to meas- ure minute quantities of substances in the environment. In the space of two decades, the ability to detect minute quantities has in- creased from parts per million to parts per trillion. Testing of sub- stances has reached a high level, and a continued increase in test- ing can be expected. The result is that we are discovering constitu- ents or contaminants in food, air, water, and the environment for which there is some evidence of carcinogenic potential. The problem was stated dramatically by FDA in 1979: "Indeed, a requirement for warning on all foods that may contain an inherent carcinogenic ingredient or a carcinogenic contaminant, in contrast to a deliberately added carcinogenic substance, would apply to many, perhaps most of the foods in the supermarket." Identification by itself is not enough. Science can also perform the function of analyzing the signals and evaluating all the data to characterize the potential human risks. We know that carcinogens PAGENO="0287" 283 vary in potency a millionfold or more. Substances, for example, like saccharin are at the low extreme of the potency range, and the range extends to the very potent carcinogen aflatoxin, a natural mold on corn and peanuts which leaves residues in meat, milk, peanuts, and food products made from them. Advances are being made in the understanding of comparative metabolism and other areas which have been described quite ade- quately by other speakers here this morning. Scientific evaluation should be based on most likely estimates. The scientific evaluation, to be useful, must be based on most likely or most probable values or estimates. Assumptions which produce only upper limits of risk are of very little utility, and if relied upon alone for regulatory judgments can be positively misleading. Again, in the words of the FDA: Upper limit estimates of risks using worst case assumptions cannot be used to predict with mathematical precision what will actually occur. Worst case estimates are factored in to reach a conclusion with reasonable certainty of what will not occur. Mathematical models can perform a useful but limited role in risk assessment. These have been discussed, and I will not go into them in detail, but it is our view that mathematical models do have a place in the regulatory scheme of the scientific evaluation; but it is important to understand that mathematical models are statistical procedures or tools. None of the models currently in use has been biologically validated. When properly understood, the function of scientific evaluation is substantially the same for all the regulatory agencies dealing with chronic health hazards. The statutory basis for regulatory de- cisions vary among the agencies, but in each case the regulator needs a sound, unbiased, objective scientific evaluation upon which to base the regulatory decisions. For these reasons, AIHC has strongly favored interagency coop- eration in the methods and procedures for data development and evaluation and for scientific risk assessment methodology and cri- teria. We have, as I have indicated before, presented material to the National Academy of Sciences on developing procedures for the implementation of science panels. I have two documents here which I would like to offer for inclu- sion in the record that deal with this subject in greater detail. These are documents which we have developed. Mr. FL0RI0. Without objection, they will be put in the record. [See pp. 303, 332.] I. AIHC Science Panel Proposal (Summary) March 12, 1981; AIHC Proposal for a Science Panel-March 26, 1980; AIHC Recommended Framework for Identifying Carcinogens and Regulating them in Manufacturing Situations-Oct. 11, 1979. II. Independent scientific peer review of the AIHC proposal-Report of the Scien- tific Workshop-Aug. 15, 1982; Critical Evaluation of Proposals by the American In- dustrial Health Council to Strengthen the Scientific Base for Regulatory Decisions; Report to the American Industrial Health Council, "Evaluation and Criticism of the American Industrial Health Council's Proposals to Strengthen the Scientific Base for Regulatory Decisions. Mr. BROWNING. I would like to comment on the IRLG process and the differences we see between that and the OSTP work group, and I must tell you that we see much, in both of the efforts to com- mend them. PAGENO="0288" 284 We did feel that the IRLG was administered according to-was administered by rotation. It was a deficiency. Rotation among the agencies meant no continuing accountable responsible administra- tion. We thought that was a weakness. The group was not a policy group. They developed common positions, but had no authority to adopt or implement a policy, as we understood it. The IRLG effort in our view was largely conducted behind closed doors. I think the absence of adequate peer review has been men- tioned by others this morning. The one major report which the IRLG developed on cancer was published in July 1979. Public com- ments were sought, but to our knowledge, those comments were never reviewed, and a promised revised document taking those comments into account was never published. We believe the OSTP group was established on a sounder basis. The focus is scientific. There is a central direction under the OSTP which is responsible and accountable for the functioning of the work group. The OSTP procedure appears to be more open, and the public has been given an opportunity to comment early in the pro- ceeding. Public comments are being reviewed, and a new draft of part 1 is being prepared which will be subject to further scientific peer review. We believe that independent science panels provide the best as- surance of sound science. The utility and validity of scientific risk estimations requires an evaluation of all the relevant data through the application of the best scientific judgment in the light of the latest scientific developments. This seems like a truism. On occasions, Congress has stated this intention expressly. I submit, however, that this concept of using the latest scientific developments is inherent in all of the statutes where science is a basic component of the regulatory decisions. To skip along here. As a step to insure that the latest developments are considered by the regulators, AIHC welcomes the NAS report, previously re- ferred to, which strongly supports the institutionalization of peer review by imminent independent scientists of the adequacy of the science base for regulatory decisions. We agree with the NAS report that there should be such independent science panels in each agency. The NAS report also advocates the creation of a central board of risk assessment to be created by Congress. We favor the creation of a central panel. Although we differ somewhat from the NAS as to the functions and role of a central panel or board, there is strong agreement that both agencies' science panels and the central panel or board should be made up of independent scientific experts with the highest qualification, thus assuring input of the latest scientific developments. I believe, gentlemen, that that concludes the essence of my re- marks, and I thank you for your time. [Testimony resumes on p. 341.] [Mr. Browning's prepared statement follows:] PAGENO="0289" 285 Statement of Jackson B. Browning Representing the American Industrial Health Council Mr. Chairman, my name is Jackson Browning. I am Ccrporate Director-Health, Safety and Environmental Affairs of Union Carbide Corporation, Danbury, Co~necticut. lam pleased to appear before your Committee today as a member~of the Board of Directors of the American Industrial Health Council to discuss the importance of an adequate science base for regulatory decisions. The American Industrial Health Council(AIHC) is an organization formed in 1977 for the purpose of evaluating regulatory or legislative proposals involving scientific issues relating to chronic health hazards and supporting legislation and administrative actions which * promote the use of the best science in regulatory decision making * assure the adequacy of the science base for regulatory decisions. AIHC does not act or appear as an "advocate' for any particular substance and does not become involved in regulatory actions involving particular substances. AIHC's interest in appearing today is to urge the Committee to support the im- portance of sound science and unbiased scientific evaluation in re~ulatory actions. I would also like to discuss some AIHC proposals for accomplishing that objective. 22-143 O-83----19 PAGENO="0290" 286 AIHC is made up of some 90 industrial companies.~ The members represent a broad spectrum of industry ranging from basic producers of chemicals and metals to producers of pharmaceuticals, and consumer products. AIHC's role has been to facilitate the mobilization of the s.šientific resources of its members to present positive proposals to regulatory. ag- encies, proposals whose objective is the improvement of the science base for regulatory decisions involving potential chronic health hazards. AIHC has participated in regulatory efforts to develop policies on potential carcinogens and has urged that a national policy on cancer assessment be developed. I will discuss some details later but I think it is helpful to under- stand what AIHC is urging as "policy." AIHC believes strongly that regulatory decisions should be based on the best science. Scientific evaluation necessarily involves not only the application of the latest scientific developments but expert scientific judgment on science matters. "Policy" in this context does not mean pre- determination of scientific issues or any arbitrary dictation to the scientists as to how to evaluate scientific data. Rather sound policy involves the legal and regulatory framework to assure objective, expert scientific evaluations. A national policy refers to a coordinated approach to the scientific identification and evaluation of risk to enable sound regu- latory distinction between trivial risks and significant risks. PAGENO="0291" 287 AIHC believes interagency cooperation is a critical part of a national policy, as I have described it. It is essential that there be a coordinated approach to the use of the best science in the identification, characterization and management of risks. AIHC has welcomed the efforts under the leader-ship of the Office of Science and Technology Policy to develop a national policy on cancer assessment. We have reviewed and commented on the Discussion Draft entitled "Potential Human Carcinogens: Methods for Identification and Characterization" prepared by the Regulatory Work Group on Science and Technology, Office of Science and Technology Policy in the Executive Office of the President. The draft is Part I of a document which we understand is to cover the state-of-the-art discussion (Part I) and a discussion of general principles applicable to indi- vidual agency regulatory decisions (Part II). AIHC has not seen a draft or outline of Part II of the OSTP document. None- theless AIHC believes that the procedure for interagency co- operation is on the right track and we expect to comment and . be actively involved as the procedure moves ahead. The Recent Report of the. National Academy of. Sciences Underlines the~Importance of Objective Scientific Evaluations as the Indispensible Basis for Sound Regulatory Decisions On March 1st the National Academy of Sciences pub- lished the report entitled "Risk Assessment in the Federal Government: Mahaging the Process." A basic premise of the Report is the distinction between the scientific risk assess- ment and risk management. PAGENO="0292" 288 Risk Assessment is the scientific evaluation of facts and data and the characterization of the human risk Risk Management is the process of integrating the scientific evaluation with technical data, and the broad economic social and political statutory qb- jectives to weigh policy alternatives and reach a regulatory decisionS As described in the NAS Report~ the scientific function of risk assessment involves four steps Ci) hazard identification: the scientific process of gathering and evaluating data; (ii) dose response assessment; the evaluation of the magnitude of exposure and probability of adverse health effect; (iii) exposure assessment; data validation and evaluation; (iv) risk characterization: evaluation and magnitude of human risk. Another way to describe the scientific function is contained in the report'of the Subcommittee on Environmental Carcinogenesis of the National Cancer Advisory Board. The NCAB Subcommittee said: "The complexity of the problem dictates that the evaluation of potential human hazard of a given agent be individualized in terms of the chemical and metabolic aspects of that agent, its intended use(s), and the data available at the time the de- cision must be made, and other factors pertinent PAGENO="0293" 289 to the case under consideration. Each case must be considered on its own and the criteria appropriate for one agent may not necessarily apply to another." (The Report of the NCAB Subcommittee appears in the Journal of the National Cancer Institute for February 1977, Volume 58, pp. 461-465. I have a copy of the full report for the information of the Committee.) * - -~ AlEC has used somewhat different words to describe the scientific function, but is in basic agreement with the N~S Report on the nature of the scientific function and the distinc- tion between, and the relationship of, the scientific evaluation to the regulatory process or risk management. The regulatory function - what the NAS Report calls "risk management" - is to determine in light of the scientific findings: * the social significance of the risk involved and * the scope of regulatory action which is appro- priate under the particular statute. The regulator must depend on the soundness of the scientific evaluation to provide~ a sound basis for the compli- cated and important judgments the regulator must make. Good regulatory-decisions, therefore,~ require a scientific evaluation of the underlying scientific data, including toxicology, exposure. and other- scientific parameters which are not biased by social - political and economic - value judgments of the scientist or by administrative or policy dictates to the scientists which would distort a full objective scientific evaluation. A biased scientific evaluation will lead to unsound regulatory decisions. PAGENO="0294" 290 This distinction between the scientific function and the regulatory function is essential to what AIHC believes is the scope and content of a scientifically sound national policy on cancer assessment. Science Provides the Foundation for a Determination of the Significance of the Risk It is axiomatic that we do not live in a risk~ free society. The regulatory job is to separate the significant risks from the trivial risks. Regulatory focus on trivial risks is a waste and takes away from attention to significant risks. The need to identify significant risks is central to the regulatory task. Science has steadily increased its analytic capabilities to measure minute quantities of substances in the environment. In the space of two decades the ability to detect minute quantities has increased from parts per million to parts per trillion. New technology will undoubtedly further increase our measurement ability in the future. In the same two decade period testing of substances has reached a high level and a continued increase in testing can be expected. The number of substances for which there is some evi- dence of potential carcinogenic action or some other chronic toxic response in some test system at some dose level is rapidly mounting. The result is that we are discovering constituents or contaminants in food, air, water and the environment for which there is some evidence of carcinogenic potential. The problem was stated dramatically by FDA in 1979. PAGENO="0295" 291 "Indeed a requirement for warning on all foods that may contain an inherent carcinogenic in- gredient or a carcinogenic contaminant (in contrast to a deliberately added carcinogenic substance) would apply to many, perhaps most, foods iiithe supermarket." Federal Register Vol. 45 at p. 59513, October 16, 1979. (Emphasis added.) The FDA obviously did not conclude that we should stop eating. What the FDA was saying was that if we were tc5 label foods which have a component or constituent (not an additive) for which there is some evidence of potential carcinogenicity, "many, perhaps most, foods in a supermarket" would have to be labelled as hazardous. Science has helped us to identify these substances which gave some "signals" of carcinogenicity. But identification by itself is not enough; science also can perform the function of analyzing the "signals" and evaluating all the data to character- ize the potential human risk. We know that carcinogens vary in potency a millionfold and more; substances like saccharin are at the low extreme of the potency range and the range extends to the very potent carcinogen aflatoxin, a natural mold on corn and peanuts, which leaves resi- dues in meat, milk, peanuts and food products made from them. In addition, advances in the understanding of com- parative metabolism have significantly increased our ability to translate experimental data when we extrapolate those data across species to estimate human risk. To give practical content to these comments, I would like to suggest that much of the controversy relating to the PAGENO="0296" 292 hazards from waste dump sites stems from uncertainty. If the enormous resources of the scientific community were mobilized, the risks could be characterized and the range of uncertainties could be significantly reduced. The risks and hazards could then more properly be evaluated by the Congress and the public. These major social and political decisions need a soundS base in science. Scientific Evaluation Should Be Based on Most Likely Estimates Because the concept is crucial, may I emphasize that in this increasingly complex field, sound regulatory decisions must be based on objective arid unbiased scientific evaluation. If the scientific evaluation is biased by predetermined value judgments - whether the bias is labelled "liberal", "con- servative" or administrative "policy" - the regulator has an unsure basis for the important social judgments the regulator must make. This means that the scientific evaluationS to be useful must be based on "most likely" or "most probable" values or esti- mates. Assumptions which produce only upper limits of risk are of very little utility and if relied upon alone for regulatory judgments can be positively misleading. Let me use the words of the FDA in pointing out that* risk estimations based on conservative assumptions do not predict actual risk "[U]pper.limit estimates of risks using `worst case' assumptions cannot be used to predict with mathematical precision what will actually occur.... "[W)orst case estimates"... .are factored in to reach a conclusion with reasonable certainty of PAGENO="0297" 293 what will not occur." Federal ~gj~ster, Volume 46, page 15501, March 6, 1981. (Empliiiis added.) To be useful, therefore, the scientific evaluation must inform the regulator in an unbiased way. In addition, the regu- lator must be informed of any uncertainties so he understands the full significance of what science can t~ll him. Mathematical Models Can Perform a Useful But Limited Role in Risk Assessment Two crucial steps in the scientific risk evaluation are what the NAS Report called "dose-response assessment" and "risk characterization." The first - dose-response assessment - in- volves a determination of the magnitude of expo~t~re and the probability of an adverse effect. Risk characterization is an evaluation of the magnitude of human risk. When experimental data are used these two steps involve: First extrapolation of the dose response curve from what are normally very high experimental levels of exposure to the low levels to which man is exposed. It is in this stage that mathematical models are used. Second extrapolation across species using comparative metabolism and Óther data to evaluate the relevance of animal data to man and to characterize the risk. There has been much discussion and debate in the scientific community about the use of mathematical models. Some have assumed that the models are a discipline for risk assess- ment which is separate from the biology involved. It is important to understand that mathematical models are statistical procedures or tools. None of the models currently in use has been biologically validated. One can PAGENO="0298" 294 discuss the merits of the assumptipp~ built into the models but that does not take away from the fact the models are statistical procedures which are not experimentally validated. The purpose of the statistical procedure - mathematical model - is to assist the scientists who are evaluating the data to assess or estimate the dose response relationship un~er ex- posure conditions usually much lower than the experimental con- ditions. I have underlined assist because the models at present utilize only one biological observation - tumor incidence - and deal only with an assumed or hypothetical dose response function. Models do not yet deal with the full range of experimental data nor are they developed to handle the complexity of extrapolation across species. A logical scientific evaluation should include, in addition to tumor incidence, mechanism, pharmacokinetics, metabolism, positive and negative studies, tumor types and relevance to man, and epidemiologic results. There are many developments and much thinking about mathematical models. Three distinguished scientists of the National Institute of Environmental Health Sciences published an important and interesting article in the March 4th issue of Science (the Journal of the American Association for the Ad- vancement of Science) "Implications of Nonlinear Kinetics on Risk Estimation in Carcinogenesis." (David G. Hoel, N.L. Kaplan and A. W. Anderson. The article points out that extrapolation would be more meaningful if based on DNA adducts - the combina- tion which a substance or its metabolite makes with cellular PAGENO="0299" 295 DNA - rather than on general exposure levels - the applied dose. The authors conclude: "The mathematical models typically used for low dose extrapolation are shown potentially to overestimate risk by several orders of magnitude when non linear kinetics are present." Science March 4, 1983, p. 1032. Dr. Robert Squire of Johns Hopkins, a disting~iished cancer scientist, developed a system for ranking animal car- cinogens based on the weight of the negative and positive experimental evidence (Science November 20, 1981) recently commented on the role of mathematical models: "As indicated by [Dr. Kenneth)Crump, a weakness of the model-fitting approach is the lack of information at low dose exposure. However, a greater weakness as indicated above, is that models ignore much of the relevant biological information derived from animal tests. I am not opposed to the use of mathematical models. However, they are currently based on too limited data, and I would prefer their use in conjunc tion with the weight of biological evidence." Science January 21, 1983, p. 238. (Emphasis added.) In short, models are useful to inform the scientists on probable or possible dose response function, but because of their limitations models are not by themselves a separate basis for risk, assessment. Because of these shortcomings and the lack of bio- logical basis models do not provide an actual estimate of human risk. Models should not be assigned a greater role than they should have at the present state-of-the-art. PAGENO="0300" The Importance of Interagency Cooperation in the Scientific Function This has been a long introduction whose purpose was to provide perspective and understanding of the role of science in the regulatory process. When properly understood, the function of scientific evaluation is substantially the same for all the regulatory agencies dealing with chronic health hazards. The statutory basis for regulatory decisions vary among the agencies. But in each case the regulator needs a sound, unbiased, objective scientific evaluation upon which to base the regulatory decisions. For these reasons, AlEC has strongly favored inter- agency cooperation in the methods and procedures for data de- velopment and evaluation and for scientific risk assessment methodology and criteria. A national policy on cancer assess- ment which AlEC advocates is designed to accomplish that ob- jective. This conclusion means that there can be agreement on methods and criteria for evaluation. If the science is to be sound, however, this also means that there must be no constraints on the scientific evaluation of all the data. Ultimately the scientific evaluationJnuSt be determined on the basis of expert objective scientific judgment applied to a review of all relevant data in each particular case. These same reasons have lead AIHC to oppose as `un- scientific' attempts to develop a "generic cancer regulation." These generic regulations and any other device which dictates the outcome of a scientific evaluation are in fact an effort by the agency to avoid the need for careful scientific appraisal PAGENO="0301" 297 of all the facts. Any pre-deterinination of the weight to be given certain kinds of evidence is an attempt to reduce a coin- plex judgmental scientific process to simple cook-book type of analysis. Attempts to force fit: science into rigid generic rules do not expedite regulatory decisions, as it is sorqetimes argued. Rather, inflexible rules stand in the way of the use of the latest scientific developments and lead to distortions which interfere with sound regulatory~ decisions. Reducing the complex judgmental and scientific criteria for identifying and evaluating a potential carcinogenic hazard (or other chronic health hazard) to simple, short-cut statements is no more valid or useful thanmaking an attempt to reduce complex economic or legal matters to simplistic short-cut "generic", rules. If it were possible to accomplish such an objective, there would be little role for universities or for research.' Computers would takeover. Opposition to "short-cut" or generic science regula- tions does not stand in the way of interagency cooperation on* scientific matters. On the contrary, AIHC supports the principle that there should be a'common approach to the assessment of po- tential carcinogenic or other chronic health hazards. This avoids `confusing differences among agencies on science issues and, if the science is properly applied and used, leads to regulatory decisions based on the latest scientific developments. In the last Administration there was an effort to develop interagency cooperation in the IRLG - the Interagency PAGENO="0302" 298 Regu1ator~r Liaison Group. AIHC supported this effort in principle but was critical of the way the cooperative effort was was handled * The IRLG was administered by rotation among the agencies with no continui-~ig accountable responsible administration. * The IRLG was not a `policy" group but did develop common positions. * The effort to develop common positions on science issues was largely behind closed doors. That approach is the opposite of the normal open peer exchange of views in the scientific community. The IRLG functioned effectively to cut off the government from scientific expertise offered by qualified members of the public. * The one major report which the IRLG developed on cancer was published on July 6, 1979 (44 Federal Register 39858). Public comments were sought on the document. But those comments were never reviewed and a promised revised document taking those comments into account was never published. AIHC believes the OSTP Work Group was established on a sounder basis * the focus is scientific there is a central direction under the OSTP which is responsible and accountable PAGENO="0303" 299 for the functioning of the Work Group. o the OSTP procedure is more open and the public has been given an opportunity to comment early in the proceedings. * public comments are being reviewed and a new draft of Part I is being prepared which will be subjected to further scientific peer review. AIHC. believes the OSTP procedure is soundly based and we anticipate the same procedure will be followed for Part II on principles to be used by the regulatory agencies. We urge the Committee to support this effort at interagency cooperation on science matters. Independent Science Panels Provide the Best Assurance of Sound Science The utility and validity of scientific risk estimations requires an evaluation of all the relevant data through the application of the best scientific judgment in the light of the latest scientific developments. This seems like a truism. On occasions Congress has stated this intentionexpressly, ~ the OSHA Statute S6(b). I submit, however, that this concept of utilizing the latest PAGENO="0304" 300 scientific developments is inherent in all the statutes where science is a basic component of the regulatory decisions. I am certain the Committee appreciates the great strides being made in the science of carcinogenesis. Many new developments have impacted scientific evaluations in the immediate past and will continue to do so in the near f~.iture. Recent discoveries include the development of monoclorial anti- bodies, the discovery of oncogenes, the disclosures about the role of vitamins in inhibition of cancer - and indeed in the reversal of the oncogenic process, the role of oxygen in the form of free radicals in the oncogenic process and other information on the mechanism by which carcinogens operate. Part I of the OSTP document is an attempt to set out these new developments, which provide information on the characterization of risk, information which we would be remiss not to consider. On the other hand, to rely on yesterday's "sophistication" when it has been supplemented or modified by new scientific developments, raises serious social questions. As a step to ensure that latest developments are con- sidered, AIHC welcomes the NAS Report which strongly supports the institutionalizatibfl of peer review by eminent independent scientists of the adequacy of the science base for regulatory actions. AIHC agrees with the NAS Report that there should. be such independent science panels in each agency. The NAS Report also advocates the creation of a Central Board of Risk assessment be created by Congress. AIHC also favors the creation of a central science panel. Although we PAGENO="0305" 301 differ somewhat from the NAS as to the functions and role of a central panel or Board, there is strong agreement that both agency science panels and the central panel or Board should be made up of independent scientific experts with the highest qualifications, thus assuring input of the latest scientific developments. The Committee has seen the NAS Report, and may be interested in the AIHC proposal for a central science panel. tinder the AIHC proposal the central science panel would address only (1) those few major scientific issues which concern more than one agency and (2) the significant science issues in regulations of national importance (such as the "major" rules as defined in EQ 12291.) These~few but important scientific issues could be evaluated expeditiously. The panel could also perform other functions such as those envisaged by the NAS Report. :.. AIHC does not propose an appeal panel nor a science court. The proposal would not delay regulation. To the contrary, it would speed up regulation and reduce controversy by assuring an adequate science base. I would like' to submit two papers which set forth the AIHC proposal together with a report prepared by an independent panel of distinguished scientists who reviewed the AlEC proposal. (The papers are bound together in the document submitted to the Committee.) The ringing endorsement of a central independent scien- tific review panel by these leading scientists, together with the recommendations of the NAS Report, provide strong support for 22-143 O-83--20 PAGENO="0306" 302 action by this Committee to support `the creation and strengthen- ing of independent science panels. Strong independent science panels provide a reasonable way to assure that the latest scientific thinking will be used in the scientific evaluations. The resul~~will also be. to tend by example to strengthen agency science capability. Until a central science panel or Board is created to perform the functions, we urge the Committee to support the OST? Work Group program for scientific matters. Conclusion AIHC urges the Committee to support the good beginnings on scientific interagency cooperation which have been initiated under OSTP. We hope we can also have the Committee's support in AIHC's continuing efforts to improve the science base for regu- latory actions in two ways: supporting good agency science and scientists and supporting the development of independent science panels. PAGENO="0307" __________ 393 AIHC AMERICAN INDUSTRIAL HEALTHCOUNCIL ______________ 1612 K STREET, NW:, SWTE 308. WASHINGTON. D.C. (202) 659.0060 March 12, 1981 AIBC Science Panel Proposal The Problem: The soundness of regulatory decisions on chronic health problems is a function of (1) sound scientific state-of-the-art identification and evaluation of the human health hazard, and (2) separation of the scientific evaluation from the regulatory decision to assure scientific objectivity. Prpposed Solution: Create a Science Panel made up of distinguished scientists of several disciplines to provide scien- tific evaluations to regulatory agencies concerned with chronic health to assure es~cellenše and uniformity of scientific evalua- tion by competent, objective scientists. Location: To assure objectivity and availability of the Science Panel to all regulatory agencies, the Panel should be located in an agency other than a regulatory agency. Make-up: The Panel would be composed of 15 scientists appointed for staggered three year terms. The Panel would select other scientists to participate on particular sub-panels. Panel- ists and support scientists would be part time to avoid career interruption. Appropriate conflict of interest rules would apply. The Panel would be supported by its own staff. Panel Decisions: The Panel would address significant scientific issues with provision for public input. It would not address routine matters nor regulatory questions ~ costs, benefits, etc.). The Panel evaluation would not be binding, but an agency not utilizing the Panel's evaluation would be obliged to explain its decision. The Panel's evaluation would be part of the record on review of ~he regulation. Agency Scientists. The Panel would not change the need for agency science capability. Because of the excellence of the * Panel review, agency scientific staffs would tend to be upgraded. This would lead to improvement of the science base for routine matters.not referred to the Panel. Interaction With Agency. The Panel and the agency will * interact as the scientific evaluation proceeds to assist the agency in data gathering objectives and in setting priorities. The Panel evaluations would provide a sound base for identifying regulatory alternatives and selecting the most cost effective. How Created: AIHC recommends creation by legislation. NEW YORK OFFICE: 075 CENTRAL- PARK AVENUE, SCARSDALE, NY 0583 (914) 725-1492 PAGENO="0308" 304 * AMERICAN iNDUSTRIAL HEALTH COUNCIL, INC. _______________ 075 CENTRAL ?ARK AVENUE ~CARSD~LE. NEW YC~K 0533 9I~ 725-;~c2 March 28, 1980 AlEC PROPOSAL ~0R A SCIZ~7CZ PA~TEL Introduction. The purpose of this document is summarize AlEC's views on the role of.an in endent Science Panel and the proper constitution and functioning of the Panel in making essentially scientific judgments about qualitative and quantitative aspects of human chronic health risks associ- ated with industrial activity. AlEC advocates use of a panel of eminent scientists to perform these functions separate from the regulatory agencies. ALEC proposes that the Science Panel perform risk estimations on a case-by-case basis for the benefit of all federal regulatory agencies and that, largely separate front, but utilizing, such risk estimations, the agencies then make regulatory or social policy decisions. The need for soundness of major regulatory programs requires that they be based on the best science. The purpose of this proposal is to provide the best assurance that the science upon which the agencies rely will achieve scientific expertise and freedom from conflict of interest. The proposal is intended to attain the level of excellence achieved by the ~ational Academy of Science or the equivalent. Summary. The essence of AlEC's proposal for achieving a more cohesive approach to the development of federal carcinogen and other chronic health control policies is to recognize that the determination of whether a material is likely to cause cancer, or to induce other adverse chronic health effects involves scientific rather than regulatory judgments. AlEC advocates that in the development of carcinogen and other federal chronic health control policies scientific determinations should be made. separate from regulatory considerations and that such determina- tions, assessing the most probable human risk, should be made by the best scientists available following a review of all relevant data. These-determinations should be made by a Panel of eminent scientists located centrally somewhere within government or elsawhe~e as appropriate but separate from the regulatory agencies whose actions wou1d~e-affected by the determinations. These determinations yc▒~- limited to scientific issues and would not intrude yp~n the regulatory responsibilities of the individual agencies involved. AlEC recognizes that these WASHINGTON OFFZC& 1612 K STREET, N. W., WASHINGTON, 0. C. 20006 (202):5S9-~0O60 PAGENO="0309" 305 regulatory responsibilities, cuite ~rooer1y, do differ from one agency to another. Panelists would.be selected on the basis of their eminence and would convene periodically to assess materials submatted for their consideration by agencies on a case-by-case basis. The Panel would have its own staff. The strength of the Pan~l $ s determinations would depend upon the eminence of the scientists, the soundness of their judgments, and the objectivity of their review process. Recommendation. AlEC recommends that a Science Panel be created by the Cong~iss through appropriate legislation. A. Pu~ose and Function of the Science Panel. In order to ensure the separation of the scientific function of risk estimation (~uali~ativ~ and cuaneitative) from the regulatory function of making societal decisions, AlEC proposes creation of a Science Panel. This Panel would fuhction on a continuing basis. and be located within the government or elsewhere as appropriate to achieve the excellence of ~IAS or ecuivalent. Eowever, the Panel would not be located in the agencies whose actions would be affected by the Panel's determin- ations. The Panel would perform the scientific risk assessment for regulatory agencies dealing with substances posing a potential carcinogenic or other chronic health risk. These determinations would be made on a case-by-case basis by the most eminent scientists available following a review of all relevant data. The reasons for the creation of such a Panel are: (1) To underline the izrmor-tance of the scientific evaluation and, therefore, - attract the nose experienced scientists. (2) To provide the scientific basis for a national cancer policy and eliminate incon- sistencies in scientific evaluation by scientists in different agencies. (3) To ensure that the regulators have evaluations from the most cootoetent, experienced, and objective professionals. (4) To ensure that the risk e~timations are objective and free from conflicts of interest. Experience has shown that there is a significant over- lapping interest in particular substances among the regulatory PAGENO="0310" 306 agencies and that the Panel will expedite completion of the regulatory process by providing the best possible risk assessments for use by all relevant agenc~.es without duplication. One of the benefits of the Panel.would be to minimize controversy on basic scientific issues. The Panel would function essentially to review and assess existing data, and would not itself be conducting or controlling research. The Panel staff would try to obtain as broad a data base as possible. ~1uch of the relevant information would be presented to the Panel by the regulatory agencies based on their regulatory priorities. Other government agencies would be expected, and industry and others outside of government would be encouraged, to submit appropriate data, including positive and negative results, for consideration by the Panel. The Panel would make a definitive assessment of the relevant scientific cuestions, reflecting the best scientific thinking and all of the evidence available at the time that the assessment is made. The Panel's conclusions would be set forth in a definitive report explaining in detail the data. believed relevant and how the Panel arrived at its conclusions. It is not appropriate to establish specific time limits on the operation of the Panel. It is anticipated that the Panels procedures and staff support would be such that the Panel would act expeditiously so that the regulatory process is not delayed. The regulatory agencies would utilize the Panel's report~ in any appropriate rulemaking proceedings. The Panel's determinations would not be legally binding upon an agency, but it is expected that the agencies would accept the Panel `s determinations as the basis for planning any subsequent regulatory action. The true force or strength of the Panel's determinations would depend upon the eminence of its members, the soundness of their judgments, and the objectivity of the review process. The Panel's aisessnents should be akin to those of the ad hoc panels of ~Tational Academy of Science ("~1AS") with respect to their quality and definitiveness; the Panel should make a definitive scientific assessment at a given tine on the basis of all of the data then available. If the agency decides to reject the con- clusions of the Panel, the basis for rejection must be set out in detail. final agency action, whether based upon and accepting the Panel's conclusions or otherwise, would be subject to judicial review. 3. Selection of the Panel. AIBC advocates that the very best scientifically qualified individuals be selected to serve on the Panel. 4ember- ship on the Panel should not be determined or influenced by special interest group "representation'. Such groups could of PAGENO="0311" 307 course continue to participate in regulatory proceedings within the various federal agencies. But the Panel should be as indepen- dent as possible from non-scientific constraints; what is desired from the Panel is the best scientific assessment of an issue that can be made at any given time AlEC recoi~ends that procedures equivalent to those of ~TAS should be applied to the selection of Panel members and operation of the Panel. ~AS, through the achievement of scien- tific excellence and ob~ectivity, has demonstrated the appro- priateness of these procedures for their application to the Panel concept. C. Comoosition and Operation of the Panel. AlEC recommends that the Panel have fifteen members who regularly participate in its activities. The membership should be interdisciplinary. AlEC believes the relevant disciplines include, but would not be limited to, veterinary pathology; human pathology; epidemiology; metabolism/pharrnacokinetics; toxicology/pharmacology; biostatistics; industrial hygiene; and physical sciences. AlEC proposes that Panel members would serve three-year terms. The initial Panel would be designated in three classes. Thereafter one-third of the Panel would be selected each year. Panel members would be eligible for reappointment. Panel members would select the Chair and Vice-Chair of the Panel. The Chair would preside at Panel meetings and would serve as the principal Panel contact with the regulatory agencies and with the staff. The Vice Chair~ would perform these functions when the Chair was unable to. Panel members would also from time to time select addi- tional scientists and technologists to participate in working groups assessing issues on which the additional participants have valuable expertise. Procedures coomarable to current ~TAS procedures regarding disclosures of conflict of interest would be applicable to the Panel and working groups. Scientists in the full-time ec~loy of the government would be eligible for membership on the Panel or for participation in one or more working groups. t is estimated that as many as fifty matters would be assessed. by the Panel or specific working groups annually. It is believed that a Panel, meeting with a frequency and duration as required by the work load, and utilizing specific working groups drawing upon the general scientific commmnity, and supported by a high caliber. stnff, could handle this workload. PAGENO="0312" 308 The estimated workload corresponds to estimates of likely regulatory activity. The Panel members would have quality control functions, including for exazr~le selection of working groups and providing direction to the staff, and would be responsible for managing the overall activities of the Panel and the working groups. tn view of these responsibilities and the substantial cor~itznent of tine that service on the Panel would entail, rerm.ineratiom beyond customary per diem allowances may be appropriate. Ideally, Panel members should enjoy an iunity from litigation arising from subsequent discoveries or data not made available to the Panel. It is recognized that legislation may be necessary to achieve such immunity. Under normal circumstances the report of the Panel would stand on its own. Zowever, Panel members may be called upon to explain the Panel's report during the regulatory process. D. Procedures for Making Risk Estimations and Retorts. En the interests of optimum utilization of the Panel, its consideration of a scientific issue or question could be initiated only by referral from a regulatory agency. An agency would refer on a case-by-case basis matters of high priority, statutory interest to it, and would do so only when certain preliminary evidentiary criteria agreed upon by the agency and the Panel were satisfied. In these cases, regulatory agencies would be required to utilize the Panel for the making of qualitative and quantitative risk estimations or other scientific determinations, although doing so would not preclude the agencies from making their own estimations. An agency could act in an emergency or imminent hazard situation without resort to the Panel, but the scientific basis of any such regulations would thereafter be subject to Panel review. AlEC proposes that the Panel and the regulatory agencies would interact, in an iterative fashion, as necessary or appropriate during the regulatory process. The Panel thus would continue to speak to the basic scientific questions. The regulatory agencies would consider the Panel's scientific conclusions along with matters such as statutory mandates, costs, benefits, economic immact, and social policy. AlEC's detailed views on how this interactive and iterative process should work appear in its October 11, 1979, "Recommended Framework for Identifying Carcinogens and Regulating Then in ~`Ianufacturing Situations," a copy of which is Appendix A hereto. The numbers that appear in parentheses below correspond to the numbered steps in the "Framework', which is in general applicable to chronic health hazards in addition to carcinogenicity. PAGENO="0313" 309 (L& 2) As the initial step in the Panel process, a regulatory agency should determine that a substance appears likely to Dresent a chronic health problem such as, but not necessarily limited to, carcinogenici~y. Pursuant to the criteria agreed upon by the agencies and the Panel, agencies should make comprehensive and systematic efforts to select appropriate substances for con- sideration by the Panel and for possibleregulation. 3efore referring a substance to the Panel for review, an agency should have assembled and analyzed a sufficient scientific data base to indicate that it is highly likely that there are serious chronic health effects and the agency would be likely to proceed with promulgation of regulations if the Panel confirms that serious chronic health effects would be likely to occur at given exposure levels. The regulatory agencies are strongly encouraged to estab- lish regulatory priorities before referral to the Panel, and to defer or reject frivolous or unsubstantiated requests or demands for regulatory action. When a matter is referred to the Panel for its con- sideration, the members of the Panel will decide whether a detailed assessment should be made by all or part of the Panel or whether a smaller working group should be constituted to carry out the particular assessment. Membership on specific working groups would not be restricted to Panel members, so that other scientists with recognized expertise in a particular area could be asked to serve on a working group in that area. Each working group should, however, be chaired by a Panel member. To the extent feasible (as determined by the work loads of the Panel members and the need for specific expertise), the working group should be comprised of Panel members. At least one Panel member would be fully involved in the preparation of each report. The Panel must have access to all relevant information that any regulatory agency has. Where the regulatory agencies have received trade secret or otherwise confidential material, arrangements should be made for the Panel members to have access to and consider that information, and for. the confidentiality of such data to be maintained. The Panel should have access to negative as well as positive data. (2-3) Upon referral of an issue or question to the Panel by a. regulatory agency, the agency would publish a notice of the referral in the Federal Register, inviting the public and Óther agencies to submit all available and relevant data to the Panel. In advance of such publication, the regulatory agency would have conferred with the Panel so as to define and agree upon the question or questions to be referred to the Panel for its consideration. The agreed statement of the questions would define the task of the Panel. The Panel would develop its response to the questions without further guidance or controlling action by the regulatory agency. PAGENO="0314" 310 (3) Pror~tly ucon. referral of a matter to the Panel, the staff of the Panel should affirmatively seek to elicit all available additional relevant information from the academic com~nity, independent research organizations, regulatory and government research agencies, international agencies, and industry. The staff should also conduct its own literature review. The public should be given at least sixty days from publication in the Federal Register of referral to the Panel in which to submit relevant data. As pror~tly thereafter as possible, the staff should present all of the relevant data to the Panel. * - The public would be entitled to audit deliberations * of the Panel, other than closed Panel sessions in which trade secret or other proprietary data are considered. (4, 5) At this point, the Panel itself would take on. of the following actions or would decide upon the constitu- tion cf a working group which, if constituted, would also take one of the following actions: (a) conclude that the available * data indicates specified health hazards, (b) conclude that the available information indicates that specific health hazards are unlikely to be experienced, (c) conclude that the available information is insufficient to warrant any scientifically sound conclusion on the health hazards. Generally, if this last conclusion is reached, the Panel or working group would indicate what kind of additional research or what kind of additional data, if any, should be sought. If at all possible, a quantitative risk estimation should be prepared, and the Panel's report, including the risk estimation, would be submitted at the end of Step 9 in the Frame- work. If it clearly would not be possible to prepare a meaning- ful quantitative risk estimation, the Panel would submit its report at the end of Step 5. A conclusion bythe Panel that no sound scientific conclusion is warrÓ.nted would not necessarily mean that no regula- tory action would be taken. At tines, evidence that falls short of warrantimg a definite scientific conclusion might be sufficient to warrant some regulatory activity, given a particular regulatory agency's statutory mandate, the possible adverse effects that night be anticipated if the potentially incriminating evidence should be confirmed or substantiated by additional inquiry or research, and similar considerations, would warrant some regulatory activity. The Panel would have still performed a very useful function for regulatory agencies and for reviewing courts in accurately assessing the available data and its strengths and weaknesses. PAGENO="0315" 311 (6, 7) tJpon receipt fr~ia the Panel of a brief report, perhaps oral, as to qualitative hazard assessment (or upon receict of a definitive assessment where no quantitative risk assessment is possib.le) the regulatory agency will then consider further its regulatory priorities and will assemble data on actual or potential exoosure, according significance to the degree of potency of the substance being considered and to the extent of potential exposure. All agencies with regulatory jurisdiction should be involved by this point. (8, 9) If the agency decides to proceed toward regulation, it should develop alternative regulatory approaches or levels of control and the Panel, upon receiving notification of this decision, should solicit additional data useful in making quantitative risk estimations. The Panel should then prepare the most probable risk estimation using all quantifiable data, selecting the most appropriate model or models based on the scientific judgment of the Panel members. The Panel should prepare a definitive assessment of qualitative and quantitative aspects of hazards and risks, clearly identifying uncertainties, assumotions, confidence levels. If for any reason the Panel includes safety factors directly or indirectly in its assessment, these factors should be explicitly identified. The report should develop the most probable or most likely estimate of risk, including the range of confidence of most probable values. The time during which the Panel is making its risk estimations should result in no delay. During that period the agency could, for its own purposes, assemble or obtain data relevant to the regulatory alternatives and the preparation of a regulatory analysis. Public input could be involved at this as at many other stages to ensure that both direct and indirect effects, including intangible ic~acts, such as the desirability of individual choice, are considered in the proceedings, as appropriate. Definitive assessments by the Panel could come at the end of step 5, where no quantitative risk assessment was possible, or at the end of step 9, following quantitative risk estimation. The agency that requested the Panels action should publish in the Federal Register notice of the availability from the Panel of its definitive assessment, and should include it in the administrative record of any rulemaking proceeding in which it is relevant. As noted above, the Panel's assessment should be reflected in a definitive written report for publication, reviewing all of the available relevant evidence and explaining the basis for the Panel's conclusions. If new data submitted to the Panel result in a change in the scientific assessntent, a modified report would be prepared and issued. There the report is prepared by a work~.ng group, it would undergo peer review by the full membership of the Panel prior to its publication. PAGENO="0316" 312 3. Orranizat.ion and Staff of the Panel. The staff would have size and cot~etence aooropriate to the work load and needs of the Panel and working groups. The staff would be adzain~.stered by an Executive Director and be so structured that Panel nembers would be assured of staff suoocrt in performing their functions. A~RI~A~T I USTRL?~L ~ALTH COUNCIL References A. "Recoended Framework for Identifying Carcinogens and Regu- lating Them in Manufacturing Situations', AI3C, October 11, 1979. 3. "Identification, ~aracterization, and Control of Potential Euman Carcinogens: A Framework for Federal Decision-4laking", Office of Science and Technology Policy, Executive Office off the President, February 1, 1979. (supporting separation off scientific and regulatory functions). C. "Assessment of Estimated Risks Resulting from Afla~oxins in Consumer Peanut Products and Other Food Co=odities", Bureau of Foods, Food and Drug Administration (January 19, 1978) (an exarle of a scientific risk assessment, used for regu- latory purposes). D. "Statement to the Interagency Regulatory Liaison Group, re Request for Public Coent on Scientific Report, `Scientific Bases for. Identification off Potential Carcinogens and Esti- mation of Risks', as ptthlished in the Federal Register of July 6, 1979 (44 FR 39858)", American Medical Association, September 29, 1979 (endorsing the Science Panel concept). 3. "Does. Nirite Cause Cancer? Concerns ?hcut Validity of FDA- Sponsored Study Delay Answer", Report by the Comptroller General of the United States, January 31, 1980 (illustrative of the- need for eminent ~ltidisciplinary scientific assess- ment of experimental data). Aomendix "Recorr~ended Framework for Identifying Carcinogens and Regulating Them in Manufacturing Situations", October 11, 1979. "General Information for Members of Cor~ittees of the National Research Council", The National Research Council, January 1977. PAGENO="0317" 313 * AMERICAN 1NDUSTR!AL HEALTH COUNCIL, NC. I 107E CENTRAL PARK AVENUE SCARSDALE. NEW YORK 0583 * ~ 725.~2 October 11, 1979 AlEC RECOt~4ENDED FRAMEWORK FOR IDENTIFYING CARCINOGENS AND REGULATING TELM IN hAETTJFACTtJRING SITUATIONS Introduction and Puroose. These Recommendations start with recognition of the need for improved and efficient federal regulation of carcinogen.ic risks, including risks associated with industrial activity. It is most important that the regulatory process proceed on sound scientific and cost effective bases. This process involves both scientific and societal functions,- and these should be recognized as s eparate functions. Since its inception in the fall of 1977, the American Industrial Eealth Council has been engaged in a continuing pro- cess of trying to make improvements in the various federal approaches to identifying and regulating carcinogens. The purpose of the "Sun~ary that appears below is to present AlEC's current viegs on a regulatory framework for carcinogens, recognizing that our present thinking has evolved with the need to address the problems of several agencies operatinq under different statutes. The Suzm~ary is a step-wise depiction of the evaluation and regulation process intended to highlight: 1. - The separation of scientific from social decision. 2. The role of risk estimation in assigning priorities for regulation action. (Risk estimation includes both identifying potential carcinogenic hazards and determining the probability of adverse occurrences, essentially scientific functions.) 3. The role of risk estimation in selecting among various approaches in addressing matters assigned regulatory priority. * 4. The sequential interplay between a regulatory agency and an expert scientific panel. The entire process presupposes an-independent scienti- fic revi~ panel to assure the quality of science and maintain the recuisite separation of social and technical issues. 22-143 4.O~ * WASHINGTON OFFICL~ 1612 k STREET, N. W., WASHINGTON, 0. C. 20006 (202) 659-0060 PAGENO="0318" 314 ~ote that congressional and regulatory decisions are se~arate from scientif~c determinations, but sometimes simul- taneous and imterdeper.den~. To utilize regulatory and scien- tific resources effectively, close interaction is needed. The scientific and reculatory processes should not be treated as independent regulatory `stages'. Where this framework addresses the cost/benefit rela- tionship, it is recognized that this is a shorthand description of a complex process. Costs include not only monetary costs imposed on those subject to regulation but also indirect burdens such as impact--on research and development. Senefits include not only improvement in health from reduced exposure but also other signi- ficant non-health values such as the social and economic value of the substance to be regulated and, in appropriate cases, the preservation of the individual choice and preferences of the consumer. The framework is not intended to be all inclusive, but suggests a process for the development of regulations which might apply in a manufacturing situ?tion. I. SU~!AP.Y OF AIHCS R~COMMENDATI0~S FOR IDZNTIrfIFG A~TD P~GtJ~TING CARCINOGE'TS Scientific Determinations Reculatorv (Social) Determinations 1. Determine that a substance * appears likely to present some carcinogenic problem. Generally consider reports of laboratory experiments, human experience, or scien- tific theories. Screen press reports, petitions, and other demands for action. 2. Collect and screen data that are readily available. Zx- dude clearly unsubstantiated claims. Prioritize substances for review and referral to scientific panel for expert determination. 3. !valuate quality and signifi- cance of all available data and science relating to poten- tial carcinogenic effects of candidate substances. 4. Qualitatively identify sub- stances Likely to pose a car- cinogenic hazard. PAGENO="0319" 315 Scientific Determinations ~egulatorv (Social) Determinations 5. Perform preliminar-t estimation of carcinogenic hazards to deter- mine relative potency or severity of effects and to permit reasoned and consistent setting of priorities. a. number of mammalian spe- cies for which evidence of carcinogenocity exists, and for each such species: (1) characteristics of tumors pro- duced (2) dose-response data (3) biological variation (4) route of exposure (5) dose (6) exposure charac- teristics (7) sex (8) Latency, or tine- -, to-tumor development b. quality of evidence of car-P cinogenic activity, whether epidemiologic or experiment- al animal evidence. c. dose-response relationships and associated metabolic and pharmacokinetics data, if available evidence of positive and/or negative clinical or epidemio- logic experience e. whether only experimental evidence suggests carcino- genicity, and the character of the evidence (e.g. PAGENO="0320" 316 Regulatory (Social) Determinations 6. Establish preliminary prior- ities for regulation, if war- ranted by the outcome of scientific determination in Step 5, and assemble data on actual or potential cx- poaure, according significance to high potency and extensive potential exposure. All ageri- cies with regulatory jurisdic- tion should be involved by this point. a. ~umber of people exposed b. Levels of exposure c. Conditions of exposure (j~., voluntary or in- voluntary; frecuency; duration) 7. Determine need for and priority of regulation. a. Qualitative carcinogenic hazard estimation from Scientific Determinations. b. Exposure evaluation (i.e., determine likeli- hood of significant ex- posure under current or anticipated conditions) 8.1. Develop alternative regu- latory approaches or levels of control. (tJtilize cuan- titative risk estimation if appropriate.) 9.1. Assemble (or obtain) benefit data for both substance and regulation, with public input. Scientific Determinations whether dosages were excessive) 8. Solicit additional data for quantitative risk estiation. (NOTE: If this step reguires data generation, a substantial delay could occur.) 9. Perform most probable risk estimation using all quantifiable data. a. Select most appropriate model or models based - on scientific judgment a. Economic data - e.g., uses of substance; PAGENO="0321" 317 Reculatorv (Social) Determinations volume and value of sales number of employees ; com- petitiveness of domestic industry; productivity; Consumer benefits; balance-of-payments advantages. b. Social value informa- tion - ~ health benefits; reduction in health care costs psychological and emo- tional benefits of proposed regulation. c. Social cost information - ~ increases in energy and other raw materials attributable to compliance with regulation; adverse effects of shifts to alternate suk stances; adverse environ- mental, health, and safety effects of corroliance; economic and technological feasibility; employment dislocation. 10. Commence analyses of cost- effectiveness of methods of control and analyses of regulatory alternatives, including alternative of not regulating. a. E~azard data from Scientific Determinations. (see 5 and 9 above) b. Exposure data from Regu- latory Determinations. c. Benefit data d. Comparisons of relative risks, before proceeding further. Scientific Determinations after review of all scientific data. b. ClearLy identify un- certainties, assumptions, confidence levels, and inherent safety factors. c. Develop most probable or most likely estimate of the risk, including the range of confidence of most probable values. 22-143 O-83---21 PAGENO="0322" 318 Scientific Determinations Regulatory (Social) Determinations e. Reasonable relaticnshi~, varying with the parti- cular statute invoked, between costs and benefits. 11. Risk evaluation, i.e., determine Level of control based on "socially acceptable level of risk" as derived from risk estimation and risk benefit analysis. tnco~orate appropriate margin of safety or degree of conservatism at this point, not in the risk estimation steps. Use relative risk analyses, cost and benefit information, and incremental cost-benefit comparisons. 12. Provide for most cost-effec- tive methods of achieving or maintaining Level of control. II. DSCUSSIO~ A. Determining That Regulatory Action ~4av Be Amorooriate. The framework for regulation of carcinogens has three components: a) Determination of the existence of hazard. (Qualitative) b) Performance of the most probable risk estimation. (Quantitative estimation, in light of hazard and and extent of potential human exposure.) c) Determination of degree of control (if any) required (Risk Control). The first two deteimaticns, although judgmental are basically scientific efforts recuiring interdiscipLinarY, scientific expertise. The degree of control is a social-value judgment, which should be influenced by the scientific determinations but in the end is political. ~HC maintains that the most protective and cost-effective regulations will require a new approach to this scientific-regulatory framework. - - PAGENO="0323" 319 First, the very best available scientific judgment should be utilized to evaluate a-ll the evidence relevant to whether the sub- stance is carcinogenic, including human as well as animal evidence. Such judgments should be reached on a consistent basis, for all regulatory agencies. It is also highly desirable to obtain a well-informed scientific. judgment on the relative carcinogenic strength or potency of a substance. AIHC believes that the scientific functions can be best evaluated by a panel of scientists, drawn largely from academia but in part from government, industry and elsewhere in the private sector, which would operate independently of any regulato~ agency. Having a single truly expert panel make these judgments for all federal agencies should provide a significantly greater degree of ex~ertise than the various agencies could themselves assemble. Having the panel operate free from control of a regulatory agency would also add to the credibility and thus acceptability of its judgments. In every respect, such a panel should be, in regulatory terms, extremely cost-effective. Second, the panel should not be confined to assessing only animal evidence, which is fraught with uncertainties in extrapolating from laboratory species to man, but should be free to consider all relevant evidence, including in particular, evidence of human experience with the substance in question. Such experience may be clinical or epidemiological, and may be favorable or adverse. Such evidence is the *iery best evidence in the sense that the focus of regulatory concern is protection of man. It is true that it is impossible to prove a negative, and thus human exnerjence cannot conclusively disorove the proposition suggested that a given substazi~e, at a given dose, is carcinogenic. At the very least, however, such evidence may often be significant in determining the maximum probabilities of adverse health effects and in aszessj.ng levels of risk. The refusal of many federal regulatory agencies to attri- - bute any significance to human evidence, unless it implicates a material as a carcinogen, very strongly discourages the substan- tial undertakings that would be required to compile and assess data and draw meaningful conclusions. An open mind with respect to such evidence could produce very beneficial results. AIHC contemplates that in addition to scientific iden- tification and estimation of carcinogenic hazards, the setting of priorities would be influenced by data gathered by the regulatory agencies that describes the numbers of persons who may be exposed, the levels of exposure, and the conditions of such exposure, to be used to arrive at an estimation of carcino- genic risk. PAGENO="0324" 320 .3. Basis For Determininc Regulatocy Priorities The prioritizatior. of suabect carcinogens is an essen- tial stern in both the selection of substances for scientific review by the expert panel and in preparing to regulate substances posing unacceptable carcinogenic risks. AIHC maintains that priorities must be based upon both the potengy of the substance and the degree of actual or potential exposure. For the purpose of establishing priorities, relative potency may be assessed by the use of simple, mathematical models (i.e., prelimina~' hazard estimation). A dialogue between the regulatocy agencies and the scientific panel should be established at this point. In determining the degree of exposure, the agency should enlist the aid of trade associations, other federal agencies and the public. The public sector should then be afforded the opportunity to correct government exposure estimates with actual data. While an orderly general approach in setting priorities and evaluating risks is needed, the AlEC recommended fr~mework is intended to be flexible, so that in emergency situations the process could be accelerated and abbreviated. The sane general approach is appropriate, however, even for emergencies. C. Procedures for Determination and ~egulation of Carcinogens rinder AlEC's recoendations, the regulato~j agency would conduct an initial screen of evidence that tends to suggest that a substance may be a human carcinogen. The agency would collect and review readily available data, and, where such data presented a reasonable theoretical and factual basis for concern, refer the matter to the scientific panel. The involvement of the panel is crucial because in the great majority of cases the available evidence will be neither clear nor conclusive, and informed scientific judgment is essen- tial to assess the evidence. Such assessment generally requires the talents of a variety of disciplines, talents which a panel could provide but which individual scientists and indeed indivtdual agen- cies are generally not Likely to possess. AlEC contemplates that the panel would, based on a thorough review of the evidence, make a qualitative estimation whether a material poses a carcinogenic hazard for man, and would also attempt to assess the severity of that hazard. The panel would report its conclusions to the regulato~ agency. The panel would report not only on carcinogenicity, but also on routes of human exposure of concern, and any other known health effects. PAGENO="0325" 321 Members of the public would be informed of the regulatory develo~ments no later than the tine of referral of a cuestion from a regulator', agency to the scientific panel. The public would have an opportunity to provide information and views to the panel, and a process of dialogue during the evaluation of evidence would be encouraged. tJpon completion of the qualitative estimation of hazard by the panel, and perhaps earlier, an agency would assemble exposure data and other pertinent data so that it could, upon receipt of the panel's views, promptly make an informed judgment as to the appropriate priority for regulation of the particular substance. tf it appears that the~matter is one of high regulatory priority, the scientific panel should be asked to prepare a comprehensive quantitative risk estimation and the agency should prepare and analyze alternative regulatory approaches or levels of control. In performing quantitative~risk estimation, the scienti- tific panel should, in light of the present science, select the most appropriate theoretical model or models, preferably ones that best fit the available data, as the basis for quantifying risk to man at low doses. The aim should be to produce the most likely accurate predictions of actual consequences of exposure of a stated population at a given level of exposure, and to enable estimates of the cost-effectiveness of various reductions in exposure. Levels of uncertainty, or degrees of confidence, should be expressed, and generally can be in mathematical terms. The use of safety factors or conservative assumptions should be avoided, and where unavoidable, should be explicitly identified. At present, many efforts at quantitative risk estimation have so many safety factors and prudential judgments built into them that the results are virtually meaningless in terms o~ setting priorities or ascertaining levels of control that are appropriate; such extremist apprbaches produce the same conclusion for all aubstances,, as a practical matter, whereas in fact the actual risks will vary widely. Following the determination of the quantifiable parts, all other known informa- tion should be incorporated in the estimation of the most probable risk. Following quantitative risk~ estimation by the panel, the regulatory agency must face up to the very difficult task of evaluating, under applicable statutory criteria, all of the information and arrive at a socially acceptable level of risk. Broad public involvement, including, in particular, involvement by those most likely exposed to the risks, is extemely desirable. There will always be some risks associated with any human activity; these should be quantified as clearly as possible and addressed frankly. In general, an acceptable Level of risk can be translated into an exposure level. Assuming that an appropriate exposure PAGENO="0326" 322. level has been set, there remains the desirable function of ascertaining the most cost-effective methods of achieving or maintaining that level-of control. The most cost-effective method may vary from one sector of regulated activity to another, and often a combination of techniques will provide the most cost-effective method. Flexibility should be Drovided for those regulated, to allow for the operation of ingenuity in devising more cost-effective methods of complying over a period of tine. D. Role of Benefit and Economic Evaluations AIHC's recommendations do not proceed on the illusory basis that a risk-free enviror~emt is attainable. Rather, they deal candidly with assessment of risk, benefit, and cost effective-- ness of control measures. The first step, noted above, is the assessment of car- cinogenic risk, that is, assessment of the likelihood of a carcino- genic event at a particular level of exposure. While such assess- ment will be somewhat imorecise, it can provide useful guides for action, particularly in comparing risks of exposure to a particular chemical with other risks including risks commonly encountered and accepted in. our society, including risks from other chemicals. The second step in this process is the assessment of the benefits derived from the chemical which raises a carcinogenic risk, and the cost involved in reducing that risk to an acceptable level. ~o one questions that any step which would reduce a car- cinogenic risk at little or no cost. In many instances, however, the means to achieve a given risk level may involve substantial costs that cax drastically reduce or eliminate important societal benefits. The third step in this process involves the balancing of benefits and risks in determining an acceptable exposure level AIHC recognizes that there is at present no generally applicable formula for balancing the two or arriving at the ultimate decision of an acceptable level of control for any particular carcinogen. AlEC therefore recommends that the regulatory agencies specifically recognize all of the various factors which comprise benefits and risks, that in every case the agencies take both into account, and specify how each element of benefits and risk is taken into account in reaching final regulatory decisions. As a step toward achieving a consistent methodology in this area, AlEC strongly supports current Administration and Congressional efforts to require agencies to perform regulatory analyses of proposed regulations and of alternatives tO the proposal, early in the regulatory process as. well as (in consider- PAGENO="0327" 323 ably more detail) at the time of promulgation of a regulation. As indicated at the outset, AIHC fully agrees that there is a need for federal regulation of chronic health risks. This need vill most effectively be met if the regulatory process proceeds along the foregoing lines. AMERICAS INDUSTRIAL E~EALTH COUNCIL References The f~llowing AlEC documents contain detailed discus- sions of the foregoing procedures: 1. AlEC Recommended Alternatives to OSHA's Generic Carcinogen Proposal, February 24, 1978. (Particularly, discussion of scientific panel; adequacy of bioassay; analysis of risks and benefits; Scientific Statement and Bibliography). .2. AlEC Draft comments onA Report of the Interagency Regulatory Liaison Group (IPLO) Workgroup on Work Assessment May 5, 1979. 3. AlEC Guidelines for Evaluation and Use of Occupa- tional Epidemiologic Cancer Studies, September 11, 1978. 4. AlEC Proposed Procedure for Prioritizing Substances on ~1IOSH List Tentatively Classified in OSHA Category I by Clement Associates, October 17, 1978. - PAGENO="0328" 324 GENERAL INFORMATION for Members of Committees of the National Research Council (Throughout this docsrnreat the term `csrnmittee is used to mean any duty appointed body of the National Research Council. whether actually desig- nated a board, committee, panel, woriloig group, or by a term indicating a sub-unit of any of these) nt NATIONAL REsE.kRCH COUNCIL 5 the princi- pal operating arm of the corporate institutiOfl. often known simply as "the Academy," that tncl,udes the Na- tional Academy of Sciences, the National Academy of Engineering, and the Institute of Medicine, The origin of this complex was an Act of Congress in 1363 incorporating the National Academy of Sciences as a private body to concerts itself with the furtherance of science and technology and to advise the agencies of the federal government upon request. As the de- ands and opportunities espanded, the National Academy of Sciences creamed the National Research Council in 1916 to be its operating instrument concerting the strength of the scientific/erigincer medical community in setwic to those purposes. Today responsibility for the National Research Council is shared by the National Academy of Sci- ences with the National Academy of Engineering and the Institute of Mcdicine. both of which have also boen established by tctions of the National Academy of Sciences under its congressional charter. Because of its charter the Academy has a close relationship with the fcdcral govcrnment, But it is legally a private, non-profit corporation, independent of the government e.'rcept for its charter obligatton to respond to requests for advice from federal agencies. A large proportion of its activities are thus under- taken at governmental request. Many studies, how- ever, originate in its own or other private initiatives. PAGENO="0329" Governance of the National Research Council Representatives of the two Academies and the Itesu- use of Medicine comprise he Governing Board of the National Research Council. The President of the National Academy of Sciences and the President of th~ National Academy of Engineering, reseectively, are the Chairman and Vice Chairman of the Govern. ing Board and of the National Research Council. Most programs of he National, Research Council are administered by its four Assemblies and four Commissions, shown on the inside front cover of this booklet. Each Assembly and Commission is headed by a Chairman and has a staff headed by an Executive Director. As their names imply, the As- semblies are concerned with areas or disciplines of science or engineering, white the Commissions are oriented toward the interactions of science and tech- nolo~r with society. These distinctions are general ones, sot invariably followed in detsil, Studies in the general field of the delivery of health care are carried on directly by the Institute of Medicine. Assemblies and Commissions Assemblies and Commissions prepare annual pro- grams for review by the Governing Board, These include the complex of studies and other activities already going forward, studies that cart be clearly anticipated by existing boards and committees, and new undertakings that are viewed by the Assembly or Commission as desirable if the necessary financial support can be found, When new queries are received from governmental agencies, or when new activities are proposed by either internal or other external sources, they are considered by the appropriate Assembly or Commis- sion, or the Institute of Medicine, which may then authorize a study by some element of its existing structure or propose th~ formation of a new cons- mireee if necessary. Committees Nearly all the substantive tasks of the National Re- search Council are performed by committees orga- 325 sized under one or other Assembly or Commission. or under the Institute of Medicine. or under an intermediate board already so organized. Those who serve on these committees thus comprise the Na- tional Research Council in action. On their abilIty and effort depend the quality and usefulness of its work, Most committee undertakings originate either in requests from external agencies having funds or in initiatives taken within the National Research Coun- cil that result in agreements with outside agencies for financial support. In either circs.tmztance, the tasks to be performed are ordinarily specified in negotiated contract or grant instruments. These are formal commitments authorized by the Governing Board of the National Research Council on behalf of the National Academy of Sciences, which bears the legal corporate responsibility. Committees are of many different kinds. Some prepare reports of an advisory nature, reflecting he thorough review of a problem or an issue and ens- bodying the considered judgment of the commitsea members; some plan and conduct studies, confer- ences, workshops, or symposia that may or may not result in substantive reports; some are responsible for U.S. representation in nsertsational organizations or for expert participation in foreign aid programs; and there are a variety of others. Ta a typical year several hundred different groups are at work, Appointments are made after a careful process of search and selection in an effort to assemble com- mittees of the highest competence, appropriately tailored to their functions. ~lembers art appointed on the strength of their professional qualifications; they may come from the academic, industrial. gay- srnmensal, or another sector of society, but they do not serve as rcpresctslatives of any agency, grrtsp, or institution unless they are specifically so desig- nated upon appointment. As the final step in he appointment procedure, the membership of every committee that will formulate a position, take an action, or prepare a report in the name of the Na- tional Research Council must he approved by the Chairman of she National Research Council. At the outset the members of a committee can expect so receive a clear presentation of: (a) the nature of its task, (b) the limitations, e.g., of time, funds, or other resources, within which it must work. (c) the staff support that will be provided, (d) the PAGENO="0330" service facilities that will be available. Ce) the rela- tionahip of its soak to other pertinent studies. (I) relevant institutional po)ic:es governing the conduct of studies, and (g) arrangements for rcimburscrnent of members' expenses. If during she undertaking she committee should conclude that, for some reason, it cannot adecuately perform its task the chairman ann staff should bring the situation to the timely attention of the parent board or committee, if any, and of the Executive Director of the Assembly or Commission. Committees may generally organize their delib- erations its whatever way they find to be most con- ducive to the successful accomplishment of their tasks, within the resources available to them and within Notional Research Council policies. They may invite participation in their meetings by those who they belIeve can be of assistance. They may also decide to meat in executive session at any time, with only mesitbem and staff present. and may insist upon the conddentiality of their discussions. Committees' prcparing advisory reports of immediate critical con- cone no a governmental body or to the public must carefully consider what preeeusiorrs may be neces- sary among themselves to ensure that their delibera- tions and draft reports remain privileged to the committee and staff until their conclusions have crystallized and their final reports have been osfictaily released. This caution is sometimes important both to save embarrassment through premature and pos- sibly misleading publicity and, where them ore soy- era.t parties at interest, to avoid giving any of them a special advantage through foreknowledge of the contest of the report. Committee members are reimbursed for their travel and other expenses ire connection with tom- mittee tacks, but are not compensated for their so:- vice except in special circumstances when the com- mittee's work demands are unusual prciportion of their time. Camnittee stuffs are expected to servo the nceds and wishes of ,osrsilsittees ir.sofar as possible. But staff nsembers have a multiple role, as assistants so the committee, as "institutional representatives" so interpret the relevant policies and procedures of the National Research Council, as inforniasions channels to and from tho entire NRC structure and relevant external bodies, and as monitom to assure that con- tract or grant requirements are met. In the first capacity they are responsible to the committee arid its chairman. In the others hey are resportssble, some- times through intermediate staff levels, to she Etecra- ice Director of the .ssscnsbly or Commission, who so turn has direct access so she Executive Office of the National Research Council and to the Chatrtnan of the National Research Council. Reports By "report" is meant any statement of the findings, conclusions or recommendations of a committee, intended for distribution outside the National Re- seo~ch Council, whether contained in a formal or si- formal report, or in letter correspondence, or in some other embodiment. The substantive content of a report that will carry a committee's name is of course the responsibility of that committee. lit the interest of producing a result of maximum usefulness, committees should ordinartly strive for consensus in their findIngs, particularly in the explication of difticult or obscure issues and in the resolution of points of diffurenca. But members must be free to state their individual reservations or dissents: explicit expressions of minority views should not be avoided when significant differences of con- viction or conclusion can best be clarified thereby. Soundness and completeness of reports are the principal considerations. As an aid to authoring cone mittens. and to help assure the highest scientific ant. expository standards, a report review procedure is an integral part of she operations of the National Research Council. This is a matter of institutional r~sp~nsibility. sttšported by she repeated experience that a report often benefits fronts critical scrutiny by scftsstifieally or technically sophisticated individuals other than the authors. Reviewers chosen for this sack are riot secxs'arily expert in the subiect master. nor are they asked to "approve" she report. Rather. they are asked so indicate whether they find she re~ port to be clear and concise, whether its argument rests on adequate data appropriately presented. snd whether the report appears so be complete, fair, and responsive so the committee's charge. Review procedures function stnder the uitthority of art institutional Report Rcviecv Committee, which works with the Assemblies and Commissions. The 326 PAGENO="0331" review process approved in a given case may in.olve reviewers sclected by the Asscnthly or Commission. or reviewers selected by the Report Review Cam- mitten. or both. Criticisms by reviewers are referred to the author- ing committee for consideration. Differences of opinion may require further interchanges. Irrecon- cilable differences have been extremely rare; they are referred to the Chairman of the National Research Council. For some reports, such as those limited to factual summaries or those embodying the proceedings of a symposium, review is ordinarily not required. But in every case, the Report Review Committee is charged with determining whether a prospective report is to be reviewed and, if to. with approving the proposed review process. Reports requiring review, whether by an Assembly or Commission or by the Report Review Committee, may not be released before the review has been completed, and must be protected from premature disclosure. Editorial standards to be sought its all reports are those generally required in first-class scientific and engineering journals. Editorial assistance is available to committees, and professional editors review re- ports intended for substantial circulation, Publication It is the intention of the National Research Council that the reports of its committees shall prompdy be mode available to the public. except for those Caw that may be classified (or reasons of national te- curtty. Publication may take many forms. The National Rmearch Council issues several hundred publications a year, including advisory findings, symposium rec- ords, literature summaries, bulletins, specialized di- rectories, periodicals, and diverte miscellaneous items. Ti's results of a committee's work may vary from a report prepared-in only a few copies and of limited interest beyond the agency asking for the study, to a publication of wide interest to the scien- tific, engineering, or other specisl~.zed community, or to the public more generally. A Printing and Publishing O~ce, offering a va- riety of services, is available to committees for ad- vice and assistance on any questions relating to the publication of their reports. That otBcn provides expertise on such matters as format, design, illustra- tions,. costs, audiences, and promotion. To be of maximum assistance it should be consulted early in the dovelopmettt of a report. Rel~lion~ with the Executive Agencies and the Congress The effectiveness of the National Research Council through the years has rested in no small measure on its relationship with the federal government. This relationship is at the same time clove, because of the need for our committees to be fully informed about the problems and issues they are asked to address. and yet sufficiently distant to protect the freedom and independence of judgment of those committees. Some of the `,vidn variety of studies conducted by the National Rcs~~rch Council at the request of governmental agencies are of such a nature that par- ticipation by o~cial representatives of the govern- ment is essential, lit such casts, formal liaison representatives are sometimes appointed to assist a committee in maintaining the necessary contact. But the committee itself is free to meet in executive session at any time, without external representation. Its other studies, it may be particularly important that the relationship with government be kept more distant and formal in order to protect the committee from both the reality and the appearance of undue intluence on its jttdgrsscttts. As toted earlier, com- mittees must bo especially prudent in such cases to safeguard the intrgrity of their deliberations and the confidentiality of their reports prior to their final review and release. External pressures for a `pre- view" of the comitsittee's thinking should be turned away, and can always be referred to the Chairman or Executive Director of the Assembly or Commis- sion. Bcsidct studies undertaken at the request of a committee or other body of the Congress. there are many occasions for congressional contact, either because of incoming inquiries from Congressmen or because an approach to Congress appears to be de- sirable. It is important that such contacts, whether they are to be oral or by correspondence or by 327 PAGENO="0332" formal testimony, should be pursued only otter con- sultasion with the Chairman or Esectitive Director of the As~trnhly or Commission concntmed. and with the office of the Chairman of the National Research Council. I'reas Relations Some studies generate considerable interest in the news media, with attendant pressures on committee members and staff to discuss the current thinking of the committee, Such information is regarded a.s priv- ileged: a committee member can always respond to an external inquiry with the explanation that mess- hers are not at liberty to discuss the committee's findings and conclusions before the report baa op. peared. Persistent inquiries cats be referred to the Assembly or Commission. Committee chairmen and staff officers should be kept informed of approaches of this kind. The Office of Information is available at any time for advice and assistance in press reia' tions. The timing and manner of public release of a report are of institutional concern, as are press rela- tions generally. It is important. therefore, that any plans for press announcemcttts be made in consulta- tion with the Office of lnformation...All such activity, conducted at the proper time and in the proper manner, is encouraged as it expands the dissemina' tion of the output of the work of our committees. Public Access The policy of the National Research Council with respect to public access is hased on two precepts: (1) as a general practice. there should be public notice of studies undertaken, of the personnel so engages4., and of the procedures and information utilized; but (1) committees must be able to pursue their tasks under conditions that encourage full and free discussion without undue external pressures and without fear of partial or out.of.context quotatioe. or of tetisinterpretatioti or misuse of information ott the cotntstittCe'O progress. Whets a study has been completed, a public access file is routinely prepared, containing a record of the 328 meetings of the committee other than esezuttve sessions, copies of all information that came before the committee from external -sources, and of any in- ternal reports from panels or suhcommtttens. Thts file is maintained in a location where access may be arranged by any member of the public. In the case * of reports classited for reasons of national securtty, the title and an unclassified abstract are accessible to the public. * When substantial public input is desired, open meetings are scheduled early in he committees de- lroeratioss, and are publicly announced well in advance. In some cases, after a report has been officially released, it is desirable to hold a public session its which the findings are presented and dis- cussed. Both types of meettng are encouraged. Reports of committees of the National Research Council often involve technical questions directly relevant to issues of public interest or policy, and frequently contain conclusions and reconsmendattons that necessarily rest upois professional judgments as well us upon quantitatively evaluated data. For some studies, nearly all the identifiable individuals of high relevant competence have a background of connec- lions and experience that constitute, or can be con- * strued by others as constituting, poteittial sources of bias irs one direction or another. It is therefore jim' * portant that members of any comrttittes dealing with such a study be- selected so as 0 comprise a care- fully balanced group in this respect. Potential bias is a factor considered in the initial process of selection of the members of all commit- tees for studies where this factor is of concern. Sc- teetios of individuals with unacceptable conflicts of interest is. to the extent possible, avoided in the first iitvtaiice. Nevcrshclnss. as an integral aspect of the appointment process. committee members arc asked to complete a short `l'otential Sources of Bias' form, listing relevant connections and indicating any rele- vant positions taken in public statements. These * forms are reviewed within the Assembly or Commis- sion, svhere a determination is made as to whether * they support the view- that the desired balance has * indeed been achieved. Adjustments may then be Potential Sources of Bias PAGENO="0333" 329 made, for osample. by an additional appointment or two, oven after the committee has begun its ask. If previously unknown connections revealed by the forms raise Oucstions, hey are taken. an with he individual. \Vhen required. fnal decisions rest with the Cisairmnan of he Assembly or Commission and the Citairman of he National Research Council. As a matter of institutional policy. every csm~ mitten, whether potential bias is a current concern or not, is asked to discuss the general question of bias and the circumstances of it5 individual members at its first meeting, and annually thereafter if its ask is prolonged. This proc~dure helps to maintain our alertness to the issue of possible bias, and also en- abies committee members to remain aware of esch ether's relevant connections, if any. It must be emphasized that to ask that the "bias forum'" be completed, and that committees discuss the master, is in no sense to question the integrity of any individual. Rather, it provides a basis for assuming balance in a committee and it furnishes a record available for use in defense of the committee if any allegation of bias is raised during its deliberations or after its report is released. PAGENO="0334" 330 On Potential Sources of 3ias Peporns of appointed committees and other bodies of the ~TationaJ. Academy of Sciences/National Academy of gineer~ng/Inst~tute of Nedicine/ National Research Council which consider technical matters directly rele- vant to issues of public interest or policy frequently contain conclusions and recormendat.ions that necessarily rest upon professional value judgments as well as upon findings arguable on purely scientific or tschnica). grounds. When this is the case, some instances will arise in which it is inappropriate to appoint to membership an individual who has a substantial professional or financial interest that would be affected by the outcome of the deliberations. In other instances it may be necessary, in order to en- sure that the cot~.ttse is highly competent, to appoint z~mbers in such a way as to represent a balance of potentially biasing backgrounds or interacts. It is for these reasons that you are requested to complete the form on the reverse hereof, showing: (1) remunerative affiliations over the last five years, as an emoloyee, director, officer, or consultant: (2) sources of research support in excess of $10,000 over the last five years; (3) any company in which yo~a or your spouse or minor children hold a financial interest in an aunt exceeding $10,000 in market value, which also represents mmre than 10% of your or their current holdings. (~ore subtle is the question of other potential sources of bias. These might be, for exale, prejudgments implicit in views to which you are publ~c1y coitted, or conclusions giv~ ~s an expert witoess in admin- istrative or legislative proceedings. you are asked on the reverse hereof to indicate any such factors that -in your opinion tight reasonably be con- strued as potential~y~ cosoromising your independence of judgment in matters within the assigued task of the group to which you have been appointed. If, during your term of service, any of these conditions should change, a letter explaining the circumstances should be provided for the file. Each of our committees and similar bodies is asked to discuss the matter of potential sources of bias at its first meeting and once - annually thereafter. On these occasions the chairman will share with the other members such excerpts from these statements as appear rele- vant. A record of the discussion will be made, for use at the discre- tion of the ~airm.an of the National Research Council in defense of the committee against any allegation of bias. The forms themselves will be treated in accordance with the statement of privilege on the other side of this page. NATIONAL ACACZN~( OF SCIENCES - NATIONAL ACADE~(Y OF ENGINEERING INSTITO' OF )~DICINE NATIONAL ESEARCN COUNCIL February, 1978 PAGENO="0335" 331 ~sSC:L~LY or CO~*C~1l5SlQ~ ~ DiViSION, or OFrICE PRESENT ~IPLOThENT (title) Co:C~1TTEE (or~anjzacion) suE-UNIT (Above lines to be coonleced by. cogni:anc office before fors is sent to aecointee) 1. R~IUNZT.ATIVE ORGAi~IZATIO~AL AFFILIATIONS DURING LAST FIVE YEARS (if other than present employment shot~-n above) (a) As an.~nolovee (state ~osicion1 (b) As a Director or Corporate Officer (c). As a consultant (sore than five days in any year) cOIRCES OF RESEARCH S'?PORT OTHER THAN ENPLOYZR (any source of sore than ~]O,OOO total in th~ last five years) 3. C~iPANIES IN ~4l:H Vt~U C~R YOUR .~?OLS~ OR HINOR CH1LflR~N HAVE FINANCIAL iNTERESTS (any cutr~:t holdin~ vith market v~1ue in ~c~ss of S1O,000 vhich also anounts to note than ]O~ of your or their t~r ii. investsents. Please do not list ~ctt~al untS.) 4. ADDITiONAL INFORHATION (see fourth pora~raph on reverse side). U there are other circumstances in your background or present connections that in your npinion might reasonably be construed as unduly affecting your judg- nenc in matters within the assigned task of the ~rcup to which you have beco ap~ointed, please describe tham briefly. ___________________________ - Signature This si er~nt is prtvile~ed to those offices whose propmr business It is. It may he rolen:cd, on a privileSed basis, c~ the head cf an a;oocy sponsoring t~e aru~y in which the comniccom is ~ ii that official so requests in wriUn~ nd if the Chairman of the Naticnal Resoarch C~u~~jl concurs. It wIll n~C S~ ~ r~1sosc-d by the ~RC or the agency except with the apprcval.of PAGENO="0336" 332 Report of the Scientific Workshop July 23-25, 1982 CRITICAL EVALUATION OF PROPOSALS BY THE AMERICAN INDUSTRIAL HEALTH COUNCIL TO STRENGTHEN THE SCIENTIFIC BASE FOR REGULATORY DEC IS IONS John Higginson, M.D. Chairman August 15, 1982 PAGENO="0337" 333 The Purpose of the Scientific Workshop The American Industrial Health Council (AIHC) is an organization with some 120 corporate members and 80 association members with a wide range of industrial interests from production of basic products to distribution and sale of consumer products. AIHC does not become involved with issues on particular products; its central concern is the strengthening of the science base of regulatory decision making. AIHC has convened this Work Group of distinguished scientists to evaluate ~nd criticize its pro- posals to accomplish that.objective. AIHC has made three inter-related proposals which it invites this scientific workshop~to evaluate critically. 1. The creation of a Central Science Panel made up eminent scientists selected solely for scientific expertise. The Panel's jurisdiction would be the scientific function of analyzing scientific data and evaluating risk, as distinguished from the societal function of managing risk. 2. Strengthening science panels established for each regulatory agency, which perform the function of scientific review and risk assessment. 3. Strengthening of the agency science programs. AIHC documents setting out these proposals have been widely distributed and the AIHC proposals have been discussed in academia, in government, and with private public interest groups including unions and environmentalists. AIHC has invited this Work Group of distinguished scientists with a wide range of expertise and background in- cluding scientific, legislative experience to conduct an in. depth. review and criticism of the AIHC proposals. AIHC requests the Work Group to evaluate the de- sirability of the AIHC proposals from the scientific perspective and to criticize these proposals based on experience with the formation and operation of independent science review panels. AIHC requests that the conclusions of the workshop be embodied in a summary report which AIHC intends will be made public. AIHC will use comment from this Work Group and others to refine the AIHC proposals. The redrafted position paper ~il1, of course, be AIHC's sole responsibility. AIHC believes that certain Of its proposals, par- ticularly the creation of a Central Science Panel, will require legislation. AIHC plans to bring the comments of this Work Group to the attention of the Administration and the Congress when proposals such as those made by AIHC strengthen the science base of regulatory decisions are under consideration. AMERICAN INDUSTRIAL HEALTH COUNCIL 22-143 O-83--22 PAGENO="0338" Dr. Arnold L. Brown Dean, Medical School University of Wisconsin Seventh Floor, WARF Building 610 North Walnut Street Madison, Wisconsin 53706 Dr. John E. Cantlon Vice President for Research and Graduate Studies Michigan State University East Lansing, Michigan 48824 Dr. Phil Cole Professor and Head Department of Epidemiology Tidwell Hall, Room 203 University of Alabama, Birmingham University Station Birmingham, Alabama 35294 Rappor teur Dr. Virgil H. Freed Professor and Head Dept. of Agricultural Chemistry Oregon State University Corvallis, Oregon 97331 Dr. Joe W. Grisham Professor and Chairman Department of Pathology University of North Carolina School of Medicine Chapel Hill, North Carolina 27514 Dr. John Higginson Senior Scientist Universities Associated for Research and Education in Pathology, Inc. 9650 Rockville Pike Bethesda, Maryland 20814 Chairman Professor Joshua Lederberg President Rockefeller University 1230 York Avenue New York, NY 10021-6399 Honorable James Martin, Ph.D. House of Representatives 341 Cannon House Office Bldg. Washington, D. C. 20515 Dr. Tom S. Miya Dean, School of Pharmacy Chairman, Curriculum in Toxicology School of Medicine University of North Carolina Beard Hall 200H Chapel Hill, North Carolina 27514 Dr. Emil Mrak Chancellor Emeritus University of California, Davis Davis, California 95616 Observer Dr. Lawrence McCray Committee o n Institutional Means for Assessment of Risks to Public Health National Academy of Sciences 2101 Constitution Avenue, N.W. Washington, D. C. 20418 334 Scientific Workshop To Evaluate AIHC Proposals to Strengthen the Scientific Bases of Regulatory Decisions Participants PAGENO="0339" 335 Report to the American Industrial Health Council Deliberations of the Scientific Work Group Aspen, Colorado July 23-25, 1982 Evaluation and Criticism of the American Industrial Health Council's Proposals to Strengthen the Scientific Base for Regulatory Decisions set out in: "Chronic Health Hazards: Carcinogenesis, Mutagenesis, Teratogenesis; a Framework for Sound Science in Federal Decision Making." October 30, 1 981 "Proposals for Improving the Science Base for Sound Science in Federal Decision Making." December 2, 1981 PAGENO="0340" 336. Report of the Scientific Work Group 1. General Statement The work group strongly supports the view that a mechanism be established to improve the scientific basis for regulatory decision-making. It is also strongly in favor of the concept of the proposed central Science Panel. However, in the interest of improving the functioning of the Science Panel, the work group wishes to express concern about certain aspects of implementing the Panel's activities. The work group's perception of the Science Panel is that it would be a group of scientists selected strictly for their expertise. The group will be charged solely with the scientific assessment of the risk to human health of imputed environmental hazards relative to proposed or enacted regulation or legislation. That is, the fundamental objectives of the proposed panel is to establish the credibility and limitations of the best scientific information available on the particular imputed health hazard under consideration. The Science Panel would not address questions of risk management and would have no responsibility for final regulatory judgments. The work group believes that the proposed Science Panel will have a number of advantages beyond the primary one of addressing specific regulatory and legislative problems. These would include a general improvement of the regulatory process, enhanced scientific input by agency scientists, and an improved public image of regulatory decision-making resultant from reduced. advisory confrontation as agreement between panel and agencies on substantive issues developed. The work group believes the Science Panel would best be associated with the National Academy of Sciences (NAS). A crucial element in the success of the Science Panel will be the creation of an appropriate mechanism free from non-scientific considerations to recruit the most competent scientists to serve on the Science Panel and its Task Forces. Further, it was emphasized that the Science Panel must have continuity of member- ship.to function effectively. 2. Nature of Chronic Health Hazards A "chronic hazard" is an exposure which may cause acute or chronic adverse health effects. Such effects result from short or prolonged exposures and the induction periods may be very variable. Further, the hazard may exert its effect only in the presence of some other interacting factor or chemical agent. PAGENO="0341" 337 It is clear that the documents prepared by AINC were intended to be relevant to other environmental health hazards where decision-making is difficult in addition to cancer, muta- genesis and teratogenesis (e.g. neurotoxicology). This typically occurs when etiological associations are weak or controversial, such as in the case of diseases with low incidence and/or long latent periods, or with other inherent biological limitations to etiological analysis. 3. A Fundamental Difficulty. in Regulatory Decision-Making The AIHC documents are intended to enhance the extent to which decision-making is based upon all pertinent scientific data that are available and to minimize the biases introduced by inappropriate bases for scientific risk assessment. It must be acknowledged that a wide variation of viewpoints are assumed for outside interests and similar differences in views may be found among scientists, individuals and even entire agencies involved in formulating and advocating regulations. Moreover, even in the absence of entrenched views, action may be precipitated or in- ordinately delayed for political, socioeconomic or other con- siderations. It is the reduction or elimination if possible of these latter factors on scientific risk assessment that is the object of the proposals put forward by the documents. 4. Nature of Science Panel A number of complex and interwoven issues are pertinent to the effective functioning of the Science Panel. I. Impartiality: The need to insulate the Science Panel from political and economic pressures as much as possible and to ensure that this is clear to all is of fundamental importance. Since the Science Panel will be required to exist for an indefinite period, both the appearance and reality of impartiality must be continuously guarded. II. Authority: While it is desirable that~the Science Panel be free from political bias, it is essential that its judgments have impact in appropriate quarters. This requires that: i. The Science Panel should have recognized scientific prestige and authority. 1/ The `AIHC documents" referred `to are identified on the cover page of this Report. PAGENO="0342" 338 ii. That Congress, in creating such a panel, should require that on any issue reviewed by the Science Panel subsequent regulatory decisions relating to that issue take into account the Science Panel's report and identify and explain disagreements. 5. Operation of Science Panel I. Task Forces: i. The Science Panel would operate through a series of ad hoc scientific Task Forces charged with conducting a full and objective assessment of the issues referred to it. Members of the Task Forces would be selected by the same criteria as the parent panel; II. Reports: The Science Panel would consider scientific issues related to risk assessment and would render opinions, with appropriate scientific justification in a written report. Each report should provide for minority viewpoints and supportive data. All reports of the Science Panel should be published in a form which will allow them to be readily available and suitable for use as reference material. The issue of confidentiality of information must be resolved. In arriving at an opinion, the Science Panel and its component Task Forces would use all available and pertinent information. While using all such information, the Task Forces should focus on readily available documentation. The evaluation of the quality of the scientific information available for risk assessment will also form a major component of the Task Forces' assignments. Where hitherto un- published material is used in a Task Force's or the Panel's study, every effort should be made to ensure that the same information can be made readily avail- able to other scientists and to the public. It is imperative, however, that the Science Panel and its Task Forces operate expeditiously but not precipitiously. It is essential that the de- liberations of the Science Panel and the Task Forces not delay regulation. 6. Location of the Science Panel Many members of the working group supported the view that the Science Panel and its component Task Forces be asso- ciated with the NAS. This opinion was based on the importance of functioning within the discipline and prestige of the NAS and the fact that the NAS has a historic mandate to do work of this nature. Lastly, such a Science Panel should fit reasonably well within the administrative structure ~of the NAS. PAGENO="0343" 339 On the other hand, certain aspects of NAS admini- strative procedures are seen as less suitable for the Science Panel's activity and would probably~ require changing. These considerations include: I. The existing NAS rules regarding "conflict of interest" should be changed to ensure routine public disclosure of all sources of potential conflicts. II. A mechanism to compensate the Science Panel and Task Force members should be developed. III. The appointments procedure of the NAS will re- quire simplification and greater flexibility for the Task Forces to ensure expeditious handling of human health risk assessment. IV. It will be important to separate Science Panel and Task Force functiÓns from other functions of NAS, as for example the role of NAS on risk management. Other reservations regarding the suitability of NAS were discussed. These related to the desirability of having the Science Panel located outside the Washington, D.C. area, support in part by non-governmental sources and the general slowness of reaction that has characterized NAS. Other possible locations, however, within the scientific community were not excluded. 7. Balance of Science Panel and Task Fo~ce Membership The science Panel and Task forces should be composed of eminent scientists selected from multiple disciplines for their expertise and competence, width of views, and objectivity. It is desirable that the appearance of representing a non-scientific constituency be avoided. To attain these objectives, great care must be given to procedures~for selecting Science Panel members. 8. Criteria for selecting Issues to be Addressed by the Science Panel or Task Forces While any person, agency, organization or group could submit an issue for the Science Panel's consideration, such pro- posals would be considered by the Science Panel as "discretionary". That is, the Science Panel would decide whether or not to under- take to study the issue at question. On the other hand, the heads of Executive Agencies or Congress could bring "mandatory" issues to the science Panel, in effect requiring the forming of a Task Force and the issuance of a report. It would seem reasonable for the Science Panel to use broad sets of criteria to determine which discretionary issues warrant study, such as: PAGENO="0344" 340 I. Public health significance; ii. Meaningful challenge to a regulation, especially to a pending or recently enacted regulation; and III. Economic impact of making (or failing to make) a regulation. 9. organizational Structure It was recognized that some of the suggestions that follow might not be acceptable to NAS. Nonetheless, the working group believes that, whether within or outside the NAS, the following organizational structure would facilitate the achiev- ement of the Science Panels goals. There should be a permanent, full-time Executive Secretary of the Science Panel who would be charged with pro- viding continuity, overseeing the administration and operation of the Science Panel and its Task Forces. The Science Panel should consist of not more than 15 scientists of the highest scientific repute appointed for periods of up to 3 years. The Science Panel would probably meet about once per month for 2-3 days to: I. Evaluate issues for possible referral to an an hoc Task Force; II. Review (and approve or refer back) the reports emanating from the Task Forces; and III. Conduct other business as appropriate. Task Forces should contain at least one Science Panel member whenever possible and be strictly ad hoc. Since issues coming before the Science Panel will be of great public interest, procedures should be developed to permit public input as appro- priate. 10. Suggestions to AIHC Relative to Proposal The AIHC documents should acknowledge that issues outside the realm of health risk assessment are important, even crucial, to sound regulatory decision-making. While the ex- clusion of such issues from documents devoted to scientific health risk assessment is appropriate, this exclusion should be made in a way which better acknowledges their importance. The document would do well to acknowledge the great importance and substantial difficulties involved in health risk management. Finally, the working group emphasizes that regulatory decision-making incorporate two reasonably distinct areas: scientific health risk assessment and health risk management, and that the AIHC documents relate only to the first. This is appropriate and the second should be excluded from the Science Panel's area of responsibility. John Higginson, M.D. Chairman PAGENO="0345" 341 ~ Let me, and this is a very nonscientific observation, but Dr. Neal, as you were giving your conclusions to each of the points that we asked about, I noticed that Dr. Weinstein was nodding his head af- firmatively in agreement to the points that you were making. Is it fair to say that your conclusions are pretty much the same? Be- cause I believe you were here for the previous panel. Your conclu- sions are pretty much the same as the conclusions of Dr. Wein- stein? Dr. NEAL. I think in a general sense our two positions relative to these issues are similar. There are a number of issues that I com- mented on that he did not, and there are some that he commented on that I did not, but there is certainly, I think, some similarity in our points of view of what initial science needs to be done before one can proceed in a general way to control chemicals as being ge- netic or epigenetic. Mr. FL0RI0. I would like to ask, and perhaps Dr. Neal and Mr. Browning would be the appropriate people to ask, the main point that I have tried to develop today is that there has been some criti- cism of the existing policy, particularly outside peer review experi- ence, but the main point that seems to be in agreement, agreed to by all parties is that, yes, this is not a static, stagnant process. There will be new developments taking place over the next years, and yes, they should be monitored, but that as of this point, there is yet no scientific certainty that exists with regard to each and every one of these points that we have asked you to comment on so as to justify reliance upon those new approaches in a way that dictates that policy considerations be implemented on the basis of these new considerations with regard to less reliance upon animal tests and things of that sort. Is that something, my description of what I think of as a consen- sus that has developed, do you feel comfortable with that? Mr. BROWNING. Yes. I do. A part of-well, the thrust of our work would be that we would like to have this uncertainty brought out into the open. We think that the scientific panels can perform a. very useful function in this respect, and they can let the adminis- trator and Congress and the public know that in this particular area, we are required to regulate, but we do so with great uncer- tainty, and here are the limits of our knowledge. I think we might work toward a better public acceptance of the process if we let them know that today we just cannot be certain in these areas. Mr. FLORI0. Thank you very much. Dr. Silbergeld, you mentioned something about TCE and Price's landfill, which is obviously something I am concerned about, being in my State, but it is also something I am concerned about as a manifestation of what we see in the implementation of Superfund that some of us have been pointing to for the last almost year. That is the subjective terms of the national contingency plan in the plan, the aspect of Superfund that defines what is clean. PAGENO="0346" 342 It has been our hope, at least some of our hopes, that that would be fairly objective, so that we can have a universal application of those standards. The feeling on the part of some, including myself, is that it is entirely too subjective, too generalized, and therefore allows interpretations that in a sense can define away problems. Now, you have made reference to a specific problem in Price's landfill where TCE [trichioroethylene] has been found and initially tested. In your testimony, I think you say in May of 1982~ or June of 1982, to have contained in the water supplies there inappropri- ately large amounts of TCE. And you then indicate that applying these new standards contained in the cancer policy position that seems to be evolving, that the impact of that was downward defini- tion of significance, which allowed the new standard in the nation- al contingency plan, that is, the somewhat subjective standards, to be applied so as to say that there was not a need for cleanup in as urgent a way as there would be under the old definition of what constitutes a carcinogenic hazard. Can you just elaborate on whether my observation is correct? And I am specifically interested in any documentation you can pro- vide to us on this specific site and how this reevaluation took place. Dr. SILBERGELD. We would be very happy to discuss with you the documents we have and which ones may be relevant to this partic- ular issue, of course. Mr. FL0RI0. And those documents you will provide to us will, without objection, be made a part of the record for sharing with all the members of the committee. Dr. SILBERGELD. As you know, the Environmental Defense Fund has a suit against the Environmental Protection Agency over the National Contingency Plan, because we share with you a distress as to its lack of definition. Moreover, when the agency has been pushed to define what most would consider minimally acceptable standards for cleanup, such as reliance on those drinking water standards which have been published by the agency or guidance on air pollutants or other materials for which standards have been set using the best science available at the time of their setting, when pushed even to adopt these as part of the National Contingency Plan, the agency has so far refused to do so. And there are in these documents allusions to the fact that there really is no standard to determine how clean is clean. We see in the example of Missouri, rather than a good faith effort based on the best science available to the Federal Government to determine really what to do about dioxin in Missouri, that we had several months of attempting to redefine the problem out of existence, and indeed developing a science that would support remediation at a level of 100 parts per billion, which was one of the action points discussed in September within the agency. In Price's landfill we see apparently the same strategy going on. Rather than determining with good faith on the basis of analytic results showing that there was indeed a high level of TCE in drink- ing water, already found in drinking water in New Jersey and threatening the drinking water supply oi Atlantic City, Dr. Her- nandez, in cooperation with the Office of Research and Develop- ment, appears to be, as in Missouri, redefining the risk of TCE. PAGENO="0347" Mr. FL0RI0. Dr. Perera said in her presentation, and I would like to ask both of you to respond, that under this new definition, and she gave some indications with regard to vinyl chloride and so on, but focusing on TCE, because that is what the subject is at Price's landfill, that the new definition would allow for a 172-fold increase in TCE exposure under this new definition. Is that your understanding as tÓ what the implications of this re- defining in accordance with the new policy would be? Either of you. Dr. SILBERGELD. More than that. I think it would impact very di- rectly with respect to Superfund, it would very much define action. Price's landfill, as you know, is a site for which responsible parties have been identified, and therefore there is the potential of a set- tlement and payment being exacted out of those companies which are involved in the contamination of that site. Now, very clearly, if the ground water and drinking water and soil and air can be cleaned up to some factor several-fold less pro- tective of human health based on a new inventive way of assessing the risk to humans, then the financial burden on those responsible parties will be also considerably less. Mr. FL0RI0. I think you have crystallized our apprehensions. That is, the subjectivity of the cleanup standards allows the agency to go forward. Should it see fit to go forward in the settlement mode, settling for amounts for cleanup that are far less than would be required under more acceptable standards of cleanup to meet safety considerations. Is that a fair comment? Dr. SILBERGELD. It is, and I would like to draw one parallel with something that Dr. Nelson said earlier, and that is what I think is alarming to scientists. It is one thing for problems in enforcement and settlement to take place. We in science, I guess, expect those of you in the law to run into these kinds of problems. But when the scientific data base, when the evidence upon which an admittedly disputable settlement may be based, when that data base is tampered with by altering the evidence for carcinogenicity of a compound, or by expunging from a study, as in the Dow dioxin issue in Michigan, relevant and vitally important data for deter- mining all sorts of actions, when that is removed, you have now meddled with science. And no reasonable policy or legal action can be expected to flow from that altered and debauched base. Mr. FL0RI0. And when you say altered, you are talking about reevaluating in accordance with standards that, there seems to be a consensus, are not universally accepted and acceptable to the ap- propriate community, the scientific community? Dr. SILBERGELD. Yes. Mr. FL0RI0. Thank you very much. Mr. Ritter. Mr. RIrFER. Thank you, Mr. Chairman. When you say tampered and altered, are you referring to a dif- ference between using human health effects as opposed to animal health effects which were considered in the past, or is this some- thing criminal that you are talking about? PAGENO="0348" 344 Dr. SILBERGELD. In the case of TCE, all we are discussing is animal data. The revision is based upon using the animal data in a different way. Mr. RITTER. You have made a charge that this is altered or tam- pered, as if there were some kind of criminal action going on. What do you mean by that? Dr. SILBERGELD. Well, in some of the material which I have sub- mitted for the record, you will see some explicit statements about suppressing data made by ORD. Mr. RFrTER. ORD is? Dr. SILBERGELD. Suppressing data which would be relevant to an appropriate risk assessment for TCE. I am not a lawyer, so I do not mean to make legal charges, but in the scientific community that would be considered tantamount to the kinds of scientific fraud that greatly disturb the scientific community, if you read Science magazine. Mr. RITTER. In other words, they have altered this data to show what they wanted to show? They have changed the data points from what they were to what they want them to be? Dr. SILBERGELD. They have taken a document which was pre- pared showing a certain type of risk assessment based on assump- tions of carcinogenic activity of a compound and turned it into an acceptable daily intake level assessment, with no justification for doing such a radical change. There is in these memorandums an explicit statement that the real data have been suppressed. Mr. RITTER. Wait a second. They have actually changed the data as to the amounts present? Is that what you are saying? Dr. SILBERGELD. No. I think what is going on here-we have, par- tial material from the agency, not the complete, and I advise you to consider subpenaeing the agency for all the relevant documents in this matter-is an omission of data which would be critical for dis- tinguishing between mechanisms of action. Mr. Rrvn~R. I see. In this Price's landfill, you have stated that the NOEL safety factor approach is not permissible. Is that basical- ly what you and Dr. Perera have stated? That there is no accept- able level applicable here? Is that what you are saying? Dr. SILBERGELD. I think it is the consensus of every witness you have heard from today, sir, that at the present time, the use of NOEL levels which then become ADI approaches is not acceptable on a scientific. basis. Mr. RIrn~R. Dr. Neal, could you please comment on that? Dr. NEAL. I think there is really no scientific basis for choosing between mathematical risk extrapolation as currently done in reg- ulatory agencies and the NOEL-safety factor approach. Mr. RITTER. You see, I think that is a major point in this hear- ing. With the onslaught of chemical detection equipment which does not seem to keep up with our ability to evaluate the impact of what we find, is it possible to come to some kinds of conclusions where there is no-observable-effect level? This is, I think, the key point of the hearing thus far. Can you begin to make decisions on these exceedingly small amounts of obviously harmful substances on the basis of what effect they might have on people? Could you expand on that, if you PAGENO="0349" 345 would, because I do not think we have heard anything on that other side. We have heard the first side all morning. Dr. NEAL. The data are generated in experimental animals at relatively high dose levels relative to the levels at which human risk extrapolation is usually done, and those extrapolations at the low levels just are not verifiable scientifically, so that they are, in fact, speculation as to what the incidence rate may be, and it is not really a scientific matter, it is more a societal, political, societal kind of judgment that is made with some minimal scientific sup- port for any particular number that is generated. You just cannot scientifically verify the validity of these models down at the low levels, so a NOEL approach with a safety factor, the safety factor usually varies in magnitude depending upon the severity of the end point for cancer, you would have a much larger safety factor than you would for, say, chalenesterase inhibition. But scientifically, I think there is no more scientific justification for mathematical modeling than there is for the NOEL approach. You could also argue that the NOEL-safety factor approach does not have any more scientific balance or validity than does the mathematical risk extrapolation. Mr. RITTER. Dr. Silbergeld, do you have a comment on the compe- tition between a no-observed-effeŰt level and the mathematical model extrapolation in terms of scientific validity? Dr. PERERA. Perhaps I might add ,a few words. Mr. RITTER. Could you answer that first question? Can you com- ment on the scientific validity of a NOEL approach, a no-observ- able-effect level, and the mathematical model extrapolation to very, very minute amounts? Which is most scientifically valid? What is the consensus there? Dr. SILBERGELD. In part, you are asking a question on two levels, and I agree with you, this is a very critical question. There is an extremely interesting article on this subject in the current issue of Science by David Hoel from the National Institute of Environmen- tal Health Sciences. I think part of what Dr. Weinstein said earlier, and I am sure if he came back to the stand he could say it more eloquently than I, is that the issue of mechanism and the issue of the particular stage of carcinogenesis becomes critically important in distinguishing this point, in that at the stage at which a reactive metabolite is formed, which goes on to bind to DNA and cause heritable disrup- tion in the function of that material, that is, I think everyone would agree, a process that does not have a threshold. Mr. RITTER. Dr. Neal has stated in his testimony that there is a difference, and that he went so far as to say that the epigenetic should probably be dealt with somewhat less conservatively than the-- Dr. SILBERGELD. I am speaking of genetic. Dr. NEAL. Could I address the question she is asking me? And that is, Is there a threshold for a genetic carcinogen? I think there is just no scientific way of establishing whether there is or is not a threshold for a genetic carcinogen. I think there is not evidence supporting nor evidence to refute it. So I think we are in that scientific quandary of is there in fact a threshold? The often quoted justifications for there not being a PAGENO="0350" 346 threshold is based on low-level radiation effects and effects in in vitro systems for mutagenesis, neither of which is really applicable to the situation of in vivo mammalian exposure. Mr. Rrrn~R. What about the NOEL extrapolation? Dr. SILBERGELD. In the Hoel model, given that one critical step in the formation of the initiated cell, there are other processes also impinging upon this cell and its fate with respect to the eventual formation of a tumor or induction of leukemia. Some of those proc- esses may involve enzymes. They may involve excision. They may involve deactivation of the reactive metabolite, and at that point, discernible nonlinearities which may or may not be describable outside of very defined systems in vitro may be possible to be ad- duced. That may be what you are referring to, because I know some of that work has come out of your institute. Mr. RITTER. I am trying to get an idea as to whether the no-ob- servable-effect level can be used as a basis for regulation or wheth- er it must always be the mathematical extrapolation to the very low levels. What Dr. Neal was saying is that there is as much sci- entific justification for one as there is for the other. Dr. NEAL. I think I would state it another way. There is no more scientific justification for one than there is the other, and I think both of them are really speculation. Mr. RITTER. Could this have been a way to judge Price's Landing? I mean, I just heard about Price's Landing. We just got Dr. Silber- geld's testimony. But is this the kind of thing that can impinge on a Price's Landing decision involving TCE? I do not know what the amounts are. We do not have any data on it. Dr. PERERA. Mr. Ritter, may I interject? I think there is a consid- eration missing. You are asking the question, why is not NOEL ap- propriate as a way of estimating safe levels of human exposure to a carcinogen? I would add to what has already been said that the NOEL has been rejected for chronic irreversible effects such as cancer up to now because of the reality that there is no practical way, there is no available way at this time to identify a safe level of exposure or a threshold to a carcinogen in the human popula- tion, even for a single agent. Second, we must remember-- Mr. RITTER. There is a zero risk--- Dr. PERERA. There is no currently available accepted way of iden- tifying a safe level of exposure in the human population to a car- cinogen, regardless of the mechanism of action. Mr. RITTER. But at this point, would that not mean with our surging chemical detection abilities, we can understand that parts per trillion are present? And in the presence of the 20th century, would this not mean that we are going to have to outlaw and ban every single substance from ethyl toxin to nitrites to all of these things which we have been told have carcinogenic effects? At that point, that does not allow us in the Government the flexibility to deal with the problem. Dr. PERERA. I should finish, because I have left you hanging a bit in the air. I say there is no method to identify human population thresholds. This is because of the wide variability in susceptibility, or sensitivity of the human population and because of the multi- PAGENO="0351" 347 plicity of exposures, the fact that we are adding on a new exposure or an additional exposure to a preexisting background. I have some calculations in my testimony which bring into human perspective the effective additivity of carcinogenic risk. For example, if you have 10 carcinogens, each with a lifetime risk of 1 in 10,000, the annual excess risk could be as high as 3,600 excess cancer deaths a year. That is a crude approximation, but I think it helps to put this issue into perspective. So, for that reason, although Dr. Neal is correct in saying that there is not a scientific basis to prove or to disprove thresholds, the point is that because of this uncertainty and because of these other human factors that I have mentioned, we do not assume thresh- olds. And that has been the preventive, conservative policy to date. In other words, it is really a response to scientific uncertainty. But it is based on the realities of human living. Mr. RITTER. But it is also advocating a zero risk approach, be- cause-- Dr. PERERA. No; that is not true. You asked a very good question. Does it mean we ban everything? No; it does not. As you know, very few substances have been banned, even under the Delaney amendment, so it is a question, after the human health risk is as- sessed using appropriately conservative models it is generally a question then of setting standards to reduce exposures to the lowest level possible given economic and technical considerations. Mr. RITTER. What Dr. Neal has stated is that with the scientific capability that we have, there is some rationale, or at least equal rationale, or lack of any different rationale, for some kind of no- observable-effect level to be part of a regulatory scheme. Dr. PERERA. I listened carefully to what Dr. Neal said, and I do not believe that he would suggest that we can identify that safe level at this time, that we know enough to identify that safe level in a human population exposed to a carcinogen. Mr. RITTER. Can we identify sufficiently strongly the low, unsafe level in the way that it is scientifically more valid than the NOEL? That is one question that I would like to leave, because I do not think we will solve it. Mr. BROWNING. Could I make a comment on this as one who is more concerned with policy implementation than with the techni- cal details that my colleagues are, addressing? We are very much concerned that the cleaning up of these hazardous dump sites is being unduly inhibited because of this uncertainty which you have just heard articulated before you today, and there are many mem- bers of the chemical industry through the Chemical Manufacturers Association who are in the same suit that the EDF has just re- ferred to, asking HHS and other agencies to get on with the job which you have asked them to do under section 104(i). We believe that there is enough scientific information around today if properly marshaled to have a considerable impact on this problem, to shed a lot of light on it, and we have taken steps through the Chemical Manufacturers Association to get a group of independent scientists to look at the data that is available in indus- try and in the literature to help us put some kind of quantitative PAGENO="0352" 348 assessment on the hazardous materials that we are dealing with here, and give us some guidance as to how we can proceed. Mr. FL0RI0. You are making reference, of course, to the lawsuit that has been filed by a number of the environmental groups and the Chemical Manufacturers Association to start accumulating the health data on those who have been exposed to toxic waste dumps? Mr. BROWNING. Right, and even to go beyond that, to look at these materials and evaluate-- Mr. FLORIO. Thank you very much. Ladies and gentlemen, we appreciate your participation this morning. It is very helpful to us. We are now pleased to have with us Dr. G. A. Keyworth, science advisor to the President, and the director of the Office of Science and Technology Policy of the White House. 1~r. Keyworth, we appreciate your presence, assuming you are present. Thank you for your patience in what has been a long but, I think, a very productive hearing. Your statement will be made a part of the record in its entirety. You may feel free to proceed in a summary fashion. STATEMENT OF DR. G. A. KEYWORTH II, SCIENCE ADVISER TO THE PRESIDENT, DIRECTOR, OFFICE OF SCIENCE AND TECH- NOLOGY, EXECUTIVE OFFICE OF THE PRESIDENT Dr. KEYWORTH. Thank you very much, Mr. Chairman. Chairman Florio and members of the subcommittee, I am pleased to have this opportunity to describe the progress we have made in establishing a scientifically sound basis for identifying and charac- terizing potential human carcinogens. This is the first step in a process that we expect will lead to new scientific guidelines to pro- vide a better basis for decisionmaking on potential carcinogens by regulatory agencies of the Federal Government. Let me begin by placing this effort in the context of the adminis- tration's objective to reduce the excessive burden of Federal regula- tions by improving the rational basis on which those regulations are made. Now, this is an important element in President Reagan's ap- proach to revitalizing the productivity and international competi- tiveness of our industrial sector. Its goal is a very practical one, to ensure that Federal regulations that sustain environmental quality and protect human health and safety achieve their objectives safely and efficiently and with a better scientific basis. The President demonstrated his commitment to regulatory relief during the very early days of the administration. On January 22, 1981, he established the President's Task Force on Regulatory Relief, chaired by the Vice President. This task force reviews major regulatory proposals by executive branch agencies, assesses poten- tially burdensome regulations already on the books, oversees the development of legislative proposals required by congressional mandates, and makes recommendations to the President on how to reform the Federal regulatory apparatus. This process has provided a strong central* focus to the tradition- ally fragmented process of regulatory policymaking in which major PAGENO="0353" 349 health, safety, and economic policies grew up through thousands of discrete rules, often with little coordination among agencies or in- tegration with the general economic policies of the administration. This administration's commitment to regulatory relief reflects a broad consensus among all segments of society-liberals, conserva- tives, Republicans, Democrats, those who favor more regulation, those who favor less-that Federal regulatory processes should be more rational, uniform, and predictable. A key to improving the process whereby environmental health and safety regulations are developed is strengthening the scientific foundation on which such regulations are based. Accordingly, fol- lowing my appointment as science advisor to President Reagan, Vice President Bush invited me to become a member of the Task Force on Regulatory Relief, focusing on task force activities involv- ing science and technology. Shortly thereafter, the Vice President appointed me Chairman of a new interagency panel consisting of the heads of the FDA, the EPA, the Consumer Product Safety Commission, OSHA, and the Department of Agriculture's Marketing and Inspection Services. Called the Regulatory Work Group on Science and Technology, the panel is concerned with initiatives among these agencies to strengthen the scientific basis of their regulatory decisions. To express this objective another way, I will borrow from the recent study of the National Academy of Sciences on risk assess- ment in the Federal Government. They wisely differentiate be- tween the scientific assessment of a hazard and the social manage- ment of it. Those two activities are very different, and our role is restricted to providing policy guidance so that the best science available is used to characterize or assess adverse health effects of human exposure to environmental hazards. The follow-on process, that of regulatory risk management, then evaluates alternative regulatory actions and selects among them. I believe it is fundamental to keep these equally important activities separate to achieve the overall gOal of effective and efficient hazard regulation. As its first order of business, the work group asked that I estab- lish and manage a process to develop and publish scientifically sound, uniform guidelines to assist their agencies in determining whether or not a chemical substance is a human carcinogen, and if it is, estimate its risk to human populations. The decision to pursue development of improved scientific guide- lines reflected the desire among those agency heads for a common guidance which would, within the constraints imposed by their dif- fering statutory authorities and mandates, permit increased uni- formity and rationality in decisions concerning potential human carcinogens. Several of the agencies already had either formal poli- cies or ad hoc policies that had been defined by specific regulatory decisions. Another complicating factor was that Federal guidelines pub- lished in draft form by the Interagency Regulatory Liaison Group, the IRLG, in July of 1979, had never been revised to reflect exten- sive comments and criticism as well as subsequent significant ad- vances in the science. 22-143 O-83--23 PAGENO="0354" 350 Last year, an assistant director of my office convened a group of scientists from those five agencies comprising the Regulatory Work Group as well as the National Institutes of Health. This inter- agency staff group chose first to make a thorough assessment of the current state of the science. Thus, for the past 12 months, the group has been preparing an indepth scientific review of our knowledge of the mechanisms of carcinogenesis, and of the meth- ods scientists use to identify and characterize potential human car- cinogens. By November of 1982, the group had compiled a rough first draft of this scientific review. It was entitled "Potential Human Carcino- gens: Methods for Identification and Characterization." Its contents included: mechanisms of carcinogenesis; epidemiological, animal bioassay, and short-term testing methods; exposure estimation; and risk characterization. Each chapter was written by a different individual or group, and the draft contained a variety of styles and formats and a wide range of specificity and depth. At that stage no effort had yet been made to edit or unify the document. Despite the very rough nature of that draft scientific review, we sent it out immediately for peer review by a limited number of academic, industrial, and other in- terested scientists knowledgeable about various aspects of carcino- genesis. This was based on my strong conviction that external input to our process at the earliest possible time was critical to the eventual development of a document which would represent to the degree possible the very best possible scientific consensus. Accordingly, copies of the first draft were mailed to approximate- ly 50 scientists selected for their knowledge of the general subject or of specific scientific topics. This list of individuals to whom the draft was sent, a copy of the letter requesting their review, and their contents are among the documents submitted to the Subcom- mittee prior to this hearing. The cover letter made the following points: One, early peer review was being requested to help guide later efforts and to begin a dialog on the most appropriate approaches to assessing carcinogens. Two, we were seeking constructive criticism and specific recom- mendations to improve the document's accuracy and completeness. Three, the comments provided would be used to refine and im- prove the draft document so that, to the degree possible, it is an objective, accurate reflection of the state of the science in 1983. Four, the revised draft document, reflecting the comments, would form the scientific basis for the guidelines on carcinogen assess- ment methods. \T~T~ received some 35 substantive reviews, representing a tremen- dous commitment of tune and energy by a large number of distin- guished scientists. By and large, the responses are supportive of our effort, while directing criticisms to the initial result. In general, the criticisms fall into three categories, those ques- tioning the ability of the Federal Government to develop truly ob- jective guidelines and urging that the effort be undertaken instead by an external body, such as the National Academy of Sciences; those related to the organization and format of the document and PAGENO="0355" 351 its obvious lack of uniformity and coherence; and those concerning the specific scientific content of the individual chapters. I reject completely the notion that the Federal Government is in- capable of developing scientifically sound carcinogen guidelines. We are managing the scientific guidelines development process di- rectly out of my office, to provide central direction, to provide an objective scientific focus, and to insulate it to the degree possible from time constraints and excessively structured procedures. In addition, through an interactive process of review and revi- sion, our approach is designed to maximize external contributions to the state of the science review and eventual formulation of sci- entifically justified guidelines. In fact, the response to our request for comments reinforces my belief that it is possible to openly pursue the development of a broad scientific consensus even in an area as complex, fast moving, and critical as chemical carcinogens. The criticisms .we received of organization, format, uniformity, and coherence were anticipated, given the preliminary nature of the draft we circulated. Usually, the Government is criticized for waiting to release a document until it is so close to final form that comments are likely to be neither desired nor welcome. At least by circulating a very early draft of our document, warts and all, our genuine desire for comments was clear. The criticisms concerning the substance of the document general- ly addressed balance, questioning the relative prominence and cre- dence given various scientific points of view. Considering the con- troversial nature of the policy issues implicit in the scientific issues, we expected and welcomed such criticisms. The draft document is now being revised to reflect the substan- tive comments and recommendations received as appropriate. At the same time, the interagency staff group has begun to address the scientific guidelines that will comprise part 2 of the final docu- ment. We intend to have a first~ draft of the complete carcinogen policy document ready for limited external peer review by June, a revised, total document ready for Federal Register review by Octo- ber, and the final document ready for publication as a Federal policy by January, 1984. Mr. Chairman, I would not pretend that this is a simple task, nor will it result in a cookbook that answers all the questions. We are dealing with issues that elicit understandable emotional reactions, as well as issues in which science~ is unable to give us unambiguous answers. In one way, modern science has made the job of regulating poten- tially hazardous substances more difficult, because each year we can detect increasingly smaller traces of chemicals in our environ- ment. So we can have a situation where a physically unchanged en- vironment nonetheless becomes legally more hazardous. I would not suggest that we relax our vigilence, but we really must devise better ways of responding aggressively to the truly se- rious threats instead of dissipating our energy and resources and public confidence by unnecessarily controlling substances repre- senting threats that are negligible. While it would be rash of me to make promises, I have great con- fidence that the work being done by the panel will provide us with an up to date, sound scientific basis for wise regulation. PAGENO="0356" 352 Thank you. Mr. FLORIO. Thank you very much. We appreciate your very suc- cinct statement. The only point that troubles me a bit is, you were very candid in conceding that your initial paper was preliminary and very early and indicated that there was some validity to the criticism that perhaps it was not as well organized, because you were in a very early stage. But then I look at the Todhunter memo, which appears actually to be a policy implementation memo on a particular aspect of this whole question, and the comments we have had from others who have looked at it say that it is quite compatible with the major thrust of what your preliminary policy statement is. Are you comfortable with policy decisions being made off of your early policy statement approach in the way that the Todhunter memo appears to be an actual policy implementation declaration? Dr. KEYWORTH. Mr. Chairman, I am not sure I can attest to the full complement of perceived regulatory policy implications that have occurred from this draft, but what I would say is the follow- ing. To the extent that our draft is designed to achieve the very best possible science, the best possible description of the state of knowledge of a very rapidly evolving field of science today, then, yes, I welcome its use in the regulation processes that are before us today. But to assume that what is really a set of guidelines to assist in the assessment of cancer risks today that has generated a per- ceived new cancer policy, I think, is grossly premature. Mr. FL0RI0. Dr. Todhunter wrote the risk assessment chapter in your preliminary paper, and then of course he is the author of the memo that in fact dealt with the implementation of the policy under the Toxic Substances Control Act dealing with formalde- hyde, which is a specific judgment on his part which one presumes was in accordance with his risk assessment philosophy that is em- bodied in your total policy in which you have acknowledged, I think, quite candidly, is somewhat premature for actual implemen- tation. Now, what I am suggesting is that if one says it is early, in a very early stage for implementation of the overall policy, the com- ponent part, that is Government doctrine, at least on the TOSA relative to formaldehyde, that might be premature as well. Dr. KEYWORTH. Mr. Chairman, at the risk of trying to hedge the question, let me just say this. We are making in this discussion, I think, great extrapolations from assemblage of pure science. I have tried to emphasize in my testimony the necessity to distinguish be- tween development of the scientific guidelines and the actual regu- lationmaking process in which societal considerations must domi- nate. Dr. Todhunter was responsible for using his judgment and the guidance that is provided within the law to make a regulatory, a societal decision. Mr. FJ~oRrO. I understand that. Dr. KEYWORTH. I am responsible for assuring that the very best possible science will be assembled by the administration to assist those regulations. PAGENO="0357" 353 Mr. FL0RI0. I think you put your finger on it, and I will ask for your observation on this. The testimony this morning, evaluating it as fairly and objectively as I can, the testimony this morning has been that the new approaches embodied in the document which you have circulated, which you say you hope to have go into effect in January of 1984, embodies some scientific assessment procedures that have not been uniformly accepted in the scientific community at this point. And the feeling is that it is therefore premature to be basing those economic regulatory decisions upon that new approach, and I just wonder if you feel strongly that the testimony we have had this morning, that those new assessment mechanisms are inappro- priately testified to, and that they are not sufficiently scientifically verified so as to authorize economic regulatory decisions to be made off of them. Dr. KEYWORTH. I think there are two things we have heard this morning that pertain to that. One of them is the fact that the sci- ence is not adequate today to clearly define generic guidelines by which regulationmaking can be made a simple, codified process. That is true. Fortunately, the field is moving very, very rapidly and we are trying to keep abreast of it in this document. But I do not-believe that we are incorporating in our study a radically new approach to risk assessment. Mr. FL0RI0. The testimony of the first panel, who I think most people perceive of as objective, with no particular axe to grind, was that this was a significant departure from the assessment proce- dures of the past. Dr. KEYWORTH. Mr. Chairman, because we have not yet distribut- ed nor completed the draft document of part 2, we are listening and have listened for many months to preconceptions of what has been our objective, what will be the next step. I think what we are hearing here is as much preconceptions of our ultimate objectives as testimony to the facts and observations in part 1. We have attempted to outline the philosophy and to assemble the science. Our sole goal has been to develop the very finest possi- ble and most comprehensive scientific basis, not a new process for regulationmaking. Mr. FL0RI0. Would you conclude that the Todhunter memo and the EPA water contamination study are not compatible with what it is that your preliminary work has demonstrated is the direction you are going in, because they are beyond-you are dealing with the more cosmic approach to this problem. These are specific mani- festations of implementation of what most people regard as an ap- proach that is compatible with the direction you appear to be moving in. Dr. KEYWORTH. I presume that Dr. Todhunter and other mem- bers of EPA pursued a line of reasoning that was consistent with our philosophy, but where they drew their judgment is the respon- sibility of a regulationmaker, with those societal considerations. And I cannot responsibly attest to what went through their consid- erations in detail. Mr. FL0RI0. Thank you. Mr Ritter. PAGENO="0358" 354 Mr. RITTER. Thank you, Mr. Chairman, and I would like to wel- come the President's science advisor and commend him on his tes- timony. Dr. Keyworth, the recent National Academy of Sciences report on risk assessment called "Risk Assessment in the Federal Govern- ment: Managing the Process"-emphasizes the need to separate sci- entific fact and judgments from the regulatory process. I think that what the chairman was talking about is a generic problem we have here in that we are constantly mixing the science of assessing risk and hazard with the socioeconomic-political as- pects of managing the process. Could you comment on this and how it may relate to carcino- gens? Dr. KEYWORTH. Yes, to this extent, Mr. Congressman. There is no question that the task before a regulator, as well as before the leg- islator, is a very difficult task in the light of the burden of the re- sponsibility and the limited science that we have today. On the other hand, I do believe that we are, and the National Academy panel also recognized this, really approaching biomedical sciences at an unprecedented period. The science is changing ex- tremely rapidly. How we understand the process of the formation of cancer, carcinogenesis, how perhaps promoters and initiators work, the question of genotoxicity and nongenotoxicity, as we have heard discussed here today-these are all questions that, although we cannot answer them now are quite likely around the corner for us. I believe that the regulator is going to have to better and better understand the scientific basis. That is only one of the consider- ations, his decisions must be based on the best factual information that he possesses. Mr. RITTER. We have heard about the IRLG-the Interagency Regulatory Review Group-and its determination. Why didn't the OSTP simply complete the existing Interagency Regulation Liaison Group work on carcinogen policy rather than abolish it and com- mission a new effort? Dr. Nelson commented on this. He had wonderful things to say for IRLG. Why change from IRLG to OSTP? Dr. KEYWORTH. Because as I undertook this responsibility I asked both the present heads of the regulatory agencies, as well as others in previous years, as to their recommendations. Almost universally I received the recommendation that there was one mechanical problem with the IRLG-that there was no overall assigned respon- sibility for concluding the document. Second, the peer review comments that had been received had not been incorporated in a document at that time, and the scientif- ic data upon which the document rested was generally pre-1977 and had been very much passed by at that time. So what we have done is to create something that is really very similar to the IRLG, except we are including and incorporating the peer review com- ments at the very outset, and my office has accepted responsibility for assembling the document. Mr. RITTER. Dr. Keyworth, just one other point. I commented on it earlier. I think we are in a critical stage in 1983, the latter part of the 20th century. We can detect down to minute amounts, parts PAGENO="0359" 355 per trillion, and perhaps even lower in certain other kinds of sub- stances and radioactive substances. How do you sense the public reaction to some of this? What is OSTP doing that helps to educate the public about these new-at least potentially new-dangers and certainly perceived new dan- gers that arise every day as we are able to detect lower and lower levels? What is the role of OSTP here? Dr. KEYWORTH. Mr. Congressman, I resist the temptation to be overly general in the reference to the more scientific and techno- logically intensive era upon which we are embarking, but we have taken a very serious effort within our administration to emphasize aggressively the support of science, teaching, and science educa- tion, as well as science as understood by the average American citi- zen. This is an immensely complex issue, no question. Risk assess- ment is not a trivial concept. We have been very, very supportive with the biomedical sciences because I think it is fair to say they rest today where the physical sciences were perhaps five to seven decades ago in their very rapidly evolving and moving stage. There is an educational process. We are going to have to address very responsibly the basis upon which we make regulatory deci- sions, the risk assessment statements, and I think that is a very powerful suggestion of the National Academy. Mr. FL0RI0. Would the gentleman yield? Mr. RITTER. Certainly. Mr. FL0RI0. I just think that really is the heart of the issue. You are talking about monitoring in a responsible way and going for- ward, and the real point of our being here today is that there ap- pears to be policy implementation in a premature way before we have evaluated these new potentially more effective risk assess- ment methods. Do you acknowledge the validity of the concern that some have that in fact we are seeing in at least two instances policy decisions being made on what you appear to concede is not a fully advanced risk assessment, and a new type of technology that would dictate shifting off of a more conservative approach to this problem? Dr. KEYWORTH. Mr. Chairman, certainly public concerns about anything as dreaded as cancer, something that can strike young or old is unquestionably appropriate and justifiable. But regulatory decisions have to be made upon the best data available, although any scientist would rather have a total understanding of the proc- ess of carcinogenesis before he had to make any regulatory deci- sion. But without having that data, these decisions had to be made, and the best judgment possible had to be taken. That judgment was made by people who I personally know to be highly responsible, and all I can say is that I trust their judgment and I think they used the best science that they had at their hands. Mr. FL0RI0. Thank you. Thank you very much. We appreciate your help today. Our next witness, we are pleased to have the Director of the Presidential Task Force on Regulatory Relief, Mr. Christopher DeMuth. PAGENO="0360" 356 Mr. DeMuth, welcome to the committee. As with all of our wit- nesses, your statement will be made a part of the record in its en- tirety. You may feel free to proceed in summary fashion if you see fit. STATEMENT OF CHRISTOPHER C. DeMUTH, ADMINISTRATOR FOR INFORMATION AND REGULATORY AFFAIRS, OFFICE OF MAN- AGEMENT AND BUDGET AND EXECUTIVE DIRECTOR, PRESI- DENTIAL TASK FORCE ON REGULATORY RELIEF Mr. DEMUTH. Thank you very much, Mr. Chairman. It is a pleas- ure to appear before you today to describe the administration's ef- forts to improve the scientific basis of regulations involving car- cinogenic substances and similar health risks. Cancer is a grave and dreaded disease and has become a relative- ly more prominent health issue in recent decades, as other causes of death have declined significantly and major diseases have been substantially eradicated. During this same period, and especially in the last decade, first our ability to detect the presence of minute quantities of substances in food, other products, and the environ- ment has improved by several orders of magnitude, while at the same time our ability to achieve confident conclusions about whether small quantities of particular substances contribute to cancer in humans has not advanced nearly so rapidly due to sever- al intractable problems in laboratory testing and epidemiology. In these circumstances, regulatory officials are often left to make decisions amidst great uncertainty and to make the best inferences they can from extremely limited and often conflicting information. When President Reagan took office 2 years ago his administra- tion inherited an array of efforts begun during the previous admin- istration to come to grips with these dilemmas. It was, of course, to be expected that the policy approaches of different regulatory agen- cies would differ somewhat since the presence of hazardous sub- stances in consumer products, work places, and the general envi- ronment present different policy issues and since the various regu- latory statutes contain different legal standards. But there is no reason that the scientific basis for assessing risk should vary among regulatory programs. Science consists of the construction of general principles. Scientific risk assessment should be devoted to describing what is know about the risks of a given substance as objectively as possible, leaving publicly accountable of- ficials with the responsibility for framing appropriate policies toward those risks, taking into account such further considerations as control costs, public opinion and whether risks are voluntary or involuntary. To address these problems, one interagency group in the Carter administration, the IRLG discussed earlier, had published a pro- posed set of risk assessment principles and although this document was never revised and made final following the public comment period, many of its principles were subsequently incorporated into a publication of a second interagency group, the Regulatory Coun- cil, and also in a document proposed by the Environmental Protec- tion Agency at the end of 1979. And separately the Occupational PAGENO="0361" 357 Safety and Health Administration had issued a generic cancer policy in 1980. While these various documents exhibited significant differences and while each were too detailed to be easily summarized, they have obviously provoked a great deal of controversy among scien- tists. Three aspects of this controversy in particular were apparent to nonscientists. First, the documents adopted consistently conservative assump- tions about such uncertain matters as the etiology of cancer and the inferences that should be drawn from the results of high dose laboratory experiments in determining the risks of cancer in humans from long-term, low-dose exposure. The criticism of this approach was that it confounded science with policy-importing judgments about appropriate margins of safety, which ought to be left to policymakers, into scientific assessments, which ought to be limited to identifying and characterizing risks as objectively as pos- sible. Second, the documents appeared to exclude use of valid scientific information that could be important in some circumstances, such as the heavy discounting of the results of negative as opposed to positive test results, especially in~ the OSHA carcinogen policy. And, third, the documents were based on scientific thinking no more recent than 1977, in an area where theoretical and empirical research has been burgeoning in subsequent years. Those of us concerned with regulatory policy within the adminis- tration were aware of these controversies and also aware of the im- portance of developing a coherent and consistent scientific basis for regulation of carcinogenic and similar health risks, one that would merit broad support from among both scientists and the general public. Our increasing ability to detect the presence of potential carcino- gens in extremely small quantities had rendered increasingly un- tenable what is sometimes called the zero risk regulatory ap- proach-under which a substance is limited to the maximum extent technologically feasible and measurable once it has been shown to cause cancer in any dose in man or animal. In many cir- cumstances, this approach was clearly untenable as a legal and policy matter, as well as scientifically. In the late 1970's, for example, the Congress acted decisively to prevent the FDA from banning saccharin, apparently a weak car- cinogen, notwithstanding the Delaney clause on food additives. This was not a singular episode even in the food, drug, and cosmet- ic area. As we know, the FDA sets tolerance limits, rather than bans, aflatoxin, a natural contaminant in corn, grains, and peanuts which is generally agreed to be a potent carcinogen. Coal-tar hair dyes, containing the carcinogen 24 diaminoanisole, are altogether exempt from FDA regulation under a statute passed by Congress in 1935. Then, in 1980, the Supreme Court held in the Benzene case that carcinogenicity at high exposure levels is not a sufficient basis for OSHA to limit exposure to the lowest levels technologically feasi- ble, heedless of the cost or actual risk reductions of going to the lower levels, holding instead that workplace exposure limits for PAGENO="0362" 358 hazardous substances must be addressed to demonstrated signifi- cant risks. These and other recent decisions demonstrate that the difficult issues of scientific risk assessment cannot be avoided where sub- stantial social and economic impacts are involved. For all of these reasons, the administration committed itself over a year ago to developing a new and updated set of scientific princi- ples for the assessment of cancer and other health risks in regula- tory decisionmaking-principles which we hope will enjoy a broad base of support from the scientific community, will assist in distin- guishing between issues of health science and health policy, and will encourage a more informed and sophisticated public debate on the broader policy issues of health and environmental regulation. Dr. Keyworth, who just appeared before you, has described his efforts on behalf of the Task Force on Regulatory Relief. I can say only in addition to that that the administration attaches the high- est importance to this process and that Dr. Keyworth is conducting it in a spirit of openness and objectivity that bodes well for our hopes of developing somewhat broader consensus on these difficult and contentious issues. The Keyworth group plans to finish its work by the end of 1983, and as it includes the heads of the major health, safety, and envi- ronmental regulatory agencies, it is our expectation that the scien- tific guidelines it produces may then be incorporated into the risk assessment efforts of each of the individual agencies. In the mean- time, I would like to mention the excellent recent study prepared for the Congress by the National Research Council which has been mentioned before. This report recommends the development of uniform guidelines for the assessment of cancer risks, which is exactly the purpose of the Keyworth project. The report also endorses the point I have emphasized in my own testimony, that there should be a clear de- marcation between scientific risk assessment and what the report calls risk management, regulatory decisions concerning appropriate controls of objectively stated risks. We also agree with the National Research Council that before an agency decides whether to regulate a substance a written risk as- sessment should be prepared and made publicly accessible. It is also desirable that an agency's risk assessment should receive inde- pendent scientific review before any major regulatory action is taken. Finally, we also agree with the NRC that formal organizational separation of the risk assessment and risk management functions in a central risk assessment body had drawbacks in limiting neces- sary communications between these two functions. However, we do believe this option deserves more study because of the potential for economies and increased objectivity in a risk assessment body inde- pendent of regulatory responsibilities. Thank you, Mr. Chairman. [Mr. DeMuth's prepared statement follows:] PAGENO="0363" 359 Statement of Christopher C. DeMuth Administrator for Information and Regulatory Affairs, Office of Management and Budget and Executive Director, Presidential Task Forceon Regulatory Relief before the Subcommittee on Commerce, Transportation and Tourism of the Committee on Energy and Commerce of the United States House of Representatives (March l7~ 1983) Chairman Florio and members of the Subcommittee: It is a pleasure to appear before you today to describe the administration's efforts to improve the scientific basis of regulations involving carcinogenic substances and similar health risks. The topic is among the most controversial and difficult of all areas of regulatory policy. Cancer is a grave and dreaded disease, and has become a relatively more prominent health issue in recent decades, as other causes of death have declined significantly and some major diseases have been substantially eradicated. During the same period: (a) our ability to detect the presence of minute quantities of substances in food, other products, and the environment has~improved by several orders of PAGENO="0364" 360 magnitude, while (h) our ability to achieve confident conclusions about whether small quantities of particular substances contribute to cancer in humans has not advanced nearly so rapidly, due to several intractable problems in laboratory testing and epiderniological surveying. In these circumstances, regulatory officials are often left to make decisions amidst great uncertainty, and to make the best inferences they can from extremely limited and often conflicting information. Moreover, they must make these judgments on behalf of a public where there are wide variations in the attitudes of individual citizens towards health and safety risks--even where serious cancer risks are concerned, as sharply differinq attitudes towards cigarette smoking demonstrates. When President Reagan took office two years ago, his administration inherited an array of efforts begun during the previous administration to come to grips with these dilemmas. It was, of course, to be expected that the policy approaches of different regulatory agencies would differ somewhat, since the presence of hazardous substances in consumer products, workplaces, and the general environment present different policy issues, and since the various regulatory statutes contain different legal standards. But there is no reason that the scientific basis for assessing risk should vary among regulatory programs. Science indeed consists of the construction of PAGENO="0365" 361 general principles. Scientific risk assessment should he devoted to describing what is known about the risks oF a given substance as objectively as possible--leaving publicly accountable officials with the responsihilit~ for framing appropriate policies towards those risks, taking into account such further considerations as control costs, n'ihlic ooinion, and whether risks are voluntary or involuntary. To address these problems, one interagency group in the Carter administration, called the Interagency Regulatory Liaison Group, had published a proposed set of risk assessment principals ("Scientific Bases for Identification of Potential Carcinogens and Estimation of Risks," 44 Federal ~ 39858, July 6, 1979), and although this document was never revised and made final following the public comment period, many of its principles were subsequently incorporated into a publication of a second interagency grouo, the Regulatory Council (`Statement on Regulation of Chemical Carcinogens," 44 Federal ~gister 60038, November 17, 1979) , and also in a document proposed by the Environmental Protection Agency ("Policy and Procedures for Identifying, Assessing and Regulating Airborne Substances Posing a Risk of Cancer", 44 Federal ~qister 58642, October 10, 1979) Separately, the Occupational Safety and Health Administration had issued a generic "Carcinoqen Policy" ("Identification, Classification and Regulation of Potential Occupational Carcinogens," 45 Federal ~q~ister 5001, January 22, 1980.) PAGENO="0366" 362 While these various documents exhibited significant differences, and were each too detailed to be easily summarized, they had obviously provoked a great deal of controversy among scientists. Three aspects of this controversy in particular were apparent to non-scientists: First, the documents adopted consistently "conservative" assumptions about such uncertain matters as the etiology of cancer and the inferences that should be drawn from the results of high-dose laboratory experiments in determining the risks of cancer in humans from long-term, low-dose exposure. The criticism of this approach was that it confounded science with policy--importing judgments about appropriate margins of safety, which ought to he left to policyinakers, into scientific assessments, which ought to be limited to identifying and characterizing risks as objectively as mossible. Second, the documents appeared to exclude use of valid scientific information that could be important in some circumstances, such as the heavy discounting of- the results of "negative" as opposed to "positive" test results, especially in the OSH~ Carcinogen Policy. And third, the documents were based on scientific thinking no more recent than 1977, in an area where theoretical and empirical research has been burgeoning in subsequent years. PAGENO="0367" 363 Those of us concerned with regulatory policy within the administration were aware of these controversies; indeed, in my own case, they had been the focus of considerable attention at the faculty research program I directed at Harvard University before coming to the administration. We also were aware of the importance of developing a coherent and consistent scientific basis for the regulation of carcinoqenic and similar health risks, one that would merit broad support amonq both scientists and the general public. Our incr~asing ability to detect the presence of potential carcinogens in extremely small quantities had rendered increasingly untenable what is sometimes called the "zero risk" regulatory approach--under which a substance is limited to the maximum extent technologically feasible and measurable once it has been shown to cause cancer in any dose in man or animal. In many circumstances, this approach was clearly untenable as a legal and policy matter as well as scientifically. In the late 1970s, for example, the Congress acted decisively to prevent the Food and Drug Administration from banning saccharin, apparently a weak carcinogen, notwithstanding the Delaney Clause on food additives, which is as cl6se to "zero risk" regulatory standard as appears in the United States Code.* And, in 1980, * This was not a singular episode even in the food, drug, and cosmetic area. The FDA sets tolerance limits, rather than bans, aflatoxjn, a natural contaminant in corn, grains, and peanuts which is generally agreed to be a potent carcinogen. Coal-tar hair dyes, containing the carcinogen 2,4 diaminoanisole, are altogether exempt from FDA regulation under a statute passed by Congress in 1935. PAGENO="0368" ~64 the Supreme Court held in the Benzene case (Industrial Union Department, AFL-CIO v. American Petroleum Institute, 448 U.S. 607 (1980)) that carcinogenicity at high exposure levels is not a sufficient basis for OSHA to limit exposure to the lowest levels technologically feasible, heedless of the costs or actual risk reductions of going to the lower levels. The Court held that workplace exposure limits for hazardous substances must be addressed to demonstrated "significant risks." These and other recent decisions demonstrate that the difficult issues of scientific risk assessment cannot be avoided where substantial social and economic impacts are involved. For all of these reasons, the administration committed itself over a year ago to developing a new and updated set of scientific principles for the assessment of cancer and other health risks in regulatory decisionmaking--principles that we hope will enjoy a broad base of support from the scientific community, will assist in distinguishing between issues of health science and health policy, and will encourage more informed and soohisticated public debate on the broader policy issues in health and environmental regulation. PAGENO="0369" 365 In the fall of 1981, Vice President Bush, who is Chairman of the Presidential Task Force on Regulatory Relief, asked Dr. George Keyworth, the recently appointed Science Advisor to the President, to become a member of the Task Force and chair a high-level Regulatory Work Group on Science and Technology consisting of the heads of the five major health, safety, and environmental regulatory agencies. Since that time, the principal effort of Dr. Keyworth's group has been to develolD a new set of guidelines for identifying and characterizing potential human carcinogens. Non-regulatory agencies, such as the National Center for Toxicological Research, have been actively involved in the effort. Dr. Keyworth is appearing before this Subcommittee today and he is, of course, far more qualified than I to discuss this project. I can say, however, that the administration attaches the highest importance to the project, and that Dr. Keyworth is conducting it in a spirit of openness and objectivity that bodes well for our hopes of developing a broader scientific consensus on these difficult and contentious issues. The Keyworth group plans tO finish its work by the end of 1983, and as the group includes the heads of the major health, safety, and environmental regulatory agencies, it is our expectation that the scientific guidelines it produces will then be incorporated 22-143 O-83--24 PAGENO="0370" 366 into the risk assessment efforts of each of these agencies on a consistent basis. In the meantime, I would like to mention the excellent recent study prepared for the Congress by the National Research Council, National Academy of Sciences on the management of risk assessment in federal regulatory programs (National Research Council, Risk ~ssessrnentJn the Federal Government~ Managing the Process, Washington, D.C.: National Academy Press, 1983). The report recommends the development of uniform guidelines for the assessment of cancer risks for the use of federal regulatory agencies, which is exactly the purpose of the project Dr. Keyworth and I have described for you this morning. The report also endorses the point I have emphasized in my testimony: that there should he a clear demarcation between scientific risk assessment and what the report calls "risk management": regulatory decisions concerninq appropriate controls of objectively stated t~aks. We also agree with the National Research Council that before an agency decides whether to regulate a substance, a written risk assessment should be prepared and made publicly accessible. It is also desirable that an agency's risk assessment should receive independent scientific review before any major regulatory action is taken. Finally, we also agree with the report that formal organizational separation of the risk assessment and risk management functions in a central risk assessment agency has drawbacks in reducing necessary communication between the two functions. However, we believe this option deserves more study because of the potential for economies and increased objectivity in a risk assessment body independent of regulatory responsibilities. Mr. Chairman, this concludes my prepared testimony. Thank you for inviting me to appear before you. PAGENO="0371" 367 Mr. FL0RI0. Thank you very much, Mr. DeMuth. Being as super-candid as we can, the apprehension that many have is that the administration is clearly working from a philo- sophic belief that governmental regulation has gone beyond what it should be and that in fact to enhance the productivity of our whole economy there should be movement toward less governmental reg- ulation, toward the end of a higher degree of productivity in the evaluation of the administration. The apprehension that some have is that in some way health concerns are being subordinated to economic concerns and that when we are charged with making choices that if we make the less conservative choices with regard to regulatory implementation, risk assessment, and all of these other questions we have been talk- ing about this morning, we are subordinating, particularly when, as we have had testimony all morning, the scientific base for one approach versus another approach is not clearly fixed, not univer- sally agreed to. It would seem that, then, good judgment might dictate that we take the more conservative approach, tilting toward the health aspect rather than the economic aspect. Yet that does not seem to be what it is that is taking place, particularly in regard to the two specific policy papers that we made reference to through the course of the morning-Mr. Todhunter's paper as well as EPA's ground water paper. The very fact that yourself and the previous witness come for- ward in the context of regulatory reform just sends out the impres- sion that we are talking about deregulation as part of improving the economic health of the Nation, working from a particular philosophic vantage point. I think it almost reinforces the concerns of some that the health aspects have taken second place to the economic aspects. I am not saying that-and I do not subscribe to the theory of zero risk-that economics have no relevance to health decisions. But what I do be- lieve-and I would like your thoughts on this-that absent justifi- cation for changing the more conservative standards in risk assess- ment, absent a unanimous or a virtually scientific justification for changing off of the existing system, how does one justify that change in risk assessment, particularly other than conceding the fact that health is less of an important consideration than are the economics of the decisions that are being made. Mr. DEMUTH. Mr. Chairman, I agree with you emphatically that caution and conservatism is often highly appropriate when we are making major public health decisions. The point I was making in my testimony is a more limited one. It seems to me that a cautionary approach could depend upon many, many circumstances-the extent and nature of populations at risk and so forth. And the issue I would take with some of the guidelines that are on the books now is not that they are conserv- ative per se, but that they purport to apply the conservative as a matter of science rather than policy. I believe it is bad science to state what is known about a given health risk in any other than the most objective fashion possible, and that it should be for officials of the executive branch and Mem- bers of Congress to determine the appropriate controls in each cir- PAGENO="0372" 368 cumstance. But it is bad science and bad policy to be informing the American public that the facts are more definite or more threaten- ing than they are. We should state the scientific facts objectively and then, I be- lieve, in many cases extreme caution would be appropriate. Mr. FLORTO. We are not talking about the PR aspect of this. None of us control what it is that is reported. What we are talking about, and I think we have this from most, if not all, of our scientific wit- nesses today, was that there really is not legitimate justification for incorporating the new risk assessment approaches that are em- bodied in the document that is being circulated by Dr. Keyworth, that there is not sufficient justification for accepting that at this point. Now certainly, and I think it flows inevitably from that, there is certainly, therefore, no justification for implementing policy deci- sions on regulatory implementation off of that very early, newly developed approach. And then, of course, when we look at the Tod- hunter memo and we look at decisions that have actually been made with regard to formaldehyde, that just seems to be totally in- appropriate, and I ask for your-- Mr. DEMUTH. Yes; I would like to make two points. First, it is obviously the case that policy decisions cannot stand still. Decisions have to be made by members of this administration before the Keyworth group finishes its work. I am not intimately familiar with the particular decisions you raise, but the scientific basis for them are surely no less controversial among scientists than some of the principles set forth in the IRLG or in the OSHA cancer poli- cies. It seems to me that the message I got listening to the scien- tists is that it is impossible for them to decide between them as a matter of science. If I may make one further point, I said at the beginning that I thought that a cautious policy approach was often appropriate where public health is concerned. I do not believe that at a policy level this administration has flinched from taking extremely conservative regulatory approaches, even in the case of scientific uncertainty. There is a great deal of scientific uncertainty about very, very low level effects of lead. But we, based on a very contentious scien- tific record, decided to strengthen quite considerably the lead- phasedown regulations based on that record. In addition, it is important to realize that, in the regulatory area, strict controls does not necessarily mean better health results. We know from a decade and a half of research that the very strictness of the FDA's regulation of new drugs has had a very pronounced negative effect on public health in the United States by keeping off of the market life-saving drugs for long periods of time. We know this too from the saccharin controversy, where many scientists believed that notwithstanding the fact that saccharin had been shown in some tests to be carcinogenic, there was some reason to believe that banning soda pop with saccharin in it would have adverse public health consequences. People might drink more coffee. That is not good for you. People might drink soda pop with sugar in it. That is not good for you either. PAGENO="0373" 369 It is not simply a question of strictness versus leniency and Mr. Ritter, I know, has been a long-time advocate of comparative risk assessment. One has to look at the public health effects in their to- tality, comparing a given level of control or a ban with doing 1CSS or nothing, and looking at all of the effects, not just simply one side of the ledger. Mr. FL0RI0. Mr. Ritter. Mr. RITTER. I really have no questions, but I would point out that in this OSTP policy group, there is work on risk assessment, and it has gone out for review. I think that is something that is commend- able. Here is a group with a very early and a very rough draft, that has circulated it quite widely to the scientific community and to all parties concerned. With the new National Academy of Sciences report, "Risk As- sessment in the Federal Government, Managing The Process," I am sure that some of the very extensive work that has gone into that report will find its way to the OSTP group, and that that doc- ument is probably a fluid document. I have no further questions. Mr. FL0RI0. The gentleman from New Mexico. Mr. RICHARDSON. Mr. DeMuth, today's Washington Post states that you are a contender to be the head of EPA, and in that con- nection, I would like to ask you one question. I understand that you studied with Professor Posner at the University of Chicago. As I understand it, he was a very strong advocate of economic risk de- termination in tort law. At 0MB, I understand that you are in- volved in, the area of regulatory relief and cost-benefit analysis. And my question is, If you were appointed head of EPA, where would the balance be? I think one of the messages of the recent problems we have had with EPA is that cost-benefit analysis have been a buzz word for, I think, destroying the trust of the American people in our environmental prOtection programs, if you were ap- pointed to this position, on which side you would be in your cost- benefit analysis? Mr. DEMUTH. We have been talking about caution this morning and I suppose answering your question is a good occasion for cau- tion. But with regard to public confidence in EPA, I am not aware of any of the controversies of recent weeks being related to the ap- propriate role of economic analysis in environmental regulation. I am quite proud of many of the things that have been done at EPA in the last 2 years that are explicitly a result of our attempts to make environmental regulation more efficient. We inherited a situation where effluent guidelines in industry A, would be several hundred dollars per pound removed, and in industry B, would be 15 cents per pound. In such a case we know we must be spending too much in one place or too little in another, and the administration has been very active in trying to make the effluent guidelines more efficient. The emissions trading policy, a direct result of economic re- search, which was bottled up for years in the previous administra- tion, is being applied aggressively, we can document trades in States from New Jersey to Oregon that have saved scores of mil- lions of dollars per trade and increased pollution abatement. PAGENO="0374" 370 So, I do not believe, and I am not aware of any charge, that the use of economics to try to make environmental regulation more ef- fective has led to any diminution in the respect for environmental regulation, and I certainly believe that it should not. Mr. RICHARDSON. Well, I think we disagree here. I think, if any- thing, the message of reduced appropriations for EPA enforcement has been that there has been a public perception that in this cry for deregulation, in this cry for less emphasis on redtape, that we have posed risks to the public, and that is precisely my point. And in this connection, I am concerned about another matter. I understand that Mr. Keyworth-is his title science advisor to the President? Mr. DEMUTH. Yes, sir. Mr. RICHARDSON. Now, I understand that he reports to you. Is that correct? Mr. DEMUTH. No, he reports to the President. He does not report to me. Absolutely not. Mr. RICHARDSON. As I understand it-- Mr. DEMUTH. Pardon me. He is head of the Office of Science and Technology Policy within the Executive Office of the President. Mr. RICHARDSON. Does he chair a subcommittee on cancer policy? Mr. DEMUTH. Yes, sir. He chairs a subcommittee that is develop- ing scientific principles underlying regulatory policy in the area of risk assessment for the Presidential Task Force on Regulatory Relief. The Vice President is the chairman of that, and I am simply the executive director. Mr. RICHARDSON. So in effect he does report to you. Mr. DEMUTH. No, he does not. Mr. RICHARDSON. He reports directly to the President? Mr. DEMUTH. I would say he reports to the Vice President and the members, his comembers of this task force. He is a member of the task force. I am essentially the secretariat to the group. Mr. RICHARDSON. Well, you see, this is my concern. I have seen with great emphasis the fanfare over the fact that the President has a science advisor that is not at all involved in line authority, that he has direct access to the President, and I just think-I do not see it, and I think this is part of the problem. Scientific advice to the President of the United States has to be independent, has to be objective, and I wonder if we get mired in the issue of whether regulation or deregulation is political or not, that this taints the independent advice that I think the President should get. Mr. DEMUTH. Sir, if I may say, there were questions raised earli- er this morning as to why we did not simply continue with the IRLG approach to cancer and risk assessment begun during the Carter administration. It was exactly because this was a group that consisted entirely of people in the regulatory agencies. We wanted the President's science advisor, who was not a regula- tor, who has no regulatory responsibilities, to be in charge of this independently. The people that are working with him include people from the regulatory agencies, but also from many agencies in HHS, the National Institute of Toxicology, and so forth, who have no regulatory responsibilities at all. PAGENO="0375" 371 And the point is that the scientists in the administration who can call them as they see them without regard to the interpreta- tion of a particular statute or whatever to be involved in the con- struction of some risk assessment guidelines, and we are doing this. And our development of these guidelines is absolutely an open book. As soon as we had a reputable early draft on paper, we got it out as fast as we could, as Dr. Keyworth said, warts and all, to sci- entists on all sides of these controversies. We have given anybody who wants them all of the comments that have come in. We are going to go through another round in the Federal Register. And the whole point is to get a large consen- sus of the scientific community on these issues, apart from the pressures of day-to-day regulatory decisionmaking. Mr. FL0RI0. Would the gentleman yield? Mr. RICHARDSON. Yes, sir. Mr. FLORI0. On that point, my only apprehension is that the an- nouncement today is, the intention is to publish the policy in Janu- ary of 1984. That almost presupposes that we made up our mind that we are going to have something on line in 1984, notwithstand- ing the results of floating it out and getting criticisms or getting observations on it, which again puts into question the desire, and it is commendable that it is out there, but it almost seems that some- one has decided what the outcome is going to be, and it will be pub- lished on such and such a date. Can you relieve my apprehensions by saying that that date is not fixed, and that hopefully as a result of some of what we see as a very legitimate criticism that the agency has already received, be- cause we were provided with sothe of these statements of criticism, and they appear to be voluminous and, I think, in some respects appropriate, but I would hope that the agency, that is, the task force, would respond to those criticisms, make some changes if ap- propriate, and reserve the right to go beyond that. The product that is finally published is a product that is work- able, and hopefully accepted by the scientific community in a way that it is not accepted by the scientific community at this prelimi- nary stage. Mr. DEMUTH. Certainly our hopes may founder, but we have set this primarily as a deadline to Ourselves, to keep ourselves ener- gized and not permit this project to wither away as the IRLG project did in the previous administration. It could be that it will be much more difficult than we anticipate. We certainly do not intend to set down in a final guideline anything that is more hard and fast or absolute than a very broad scientific consensus permits. In fact, one of our difficulties with the previous efforts is that at some points they do seem to be excessively rigid. They do adopt a cookbook approach that says, in this case you will do this, and in that case you will do that, when there just is not a scientific con- sensus that is nearly so clear cut. So, we intend to be no more concrete than a considerable degree of support from the scientific community permits. Mr. FL0RI0. Thank you. Thank you very much. Mr. RICHARDSON. I just want to conclude, Mr. Chairman, by my saying, should you be appointed to this position, I just hope when a question like this is asked again that you mention the areas of PAGENO="0376" 372 health and safety of the American people. If I can stress again the message that I sense from this whole affair with EPA, it is that the people are willing to talk about regulation if it means protection of their own health and safety. That is the message I want to leave you with. Mr. FL0RI0. Thank you very much. Mr. DEMUTH. Thank you, Mr. Chairman. Mr. FL0RI0. We are pleased to have as our last witness, and I ap- preciate his patience today, Dr. John W. Hernandez, the Acting Ad- ministrator of the Environmental Protection Agency. Dr. Hernandez, welcome to the committee. STATEMENT OF JOHN W. HERNANDEZ, DEPUTY ADMINISTRATOR, ENVIRONMENTAL PROTECTION AGENCY, ACCOMPANIED BY BETTY ANDERSON, OFFICE OF HEALTH EFFECTS, OFFICE OF RESEARCH AND DEVELOPMENT, AND RUSS WEIR, ACTING DI- RECTOR, HAZARDOUS SITES Mr. HERNANDEZ. I wonder if I could have a couple of my staff join me. I will identify them. Mr. FL0RI0. Your statement, which has been provided to the committee in advance, and we appreciate that, will be made a part of the record in its entirety. You may feel free to proceed in a sum- mary fashion, and for the record, we would like you to identify your colleagues. Mr. HERNANDEZ. Thank you, Mr. Chairman. I would like to have one other join me. I have not asked him yet. He will find this a little surprise. Joe, come on up. Starting on my far right, Dr. Betty Anderson, who heads our group which is commonly known as OHEA. It is the Health Assess- ment Office, and it is in the Office of Research and Development. My testimony will touch on the history and activities of this office. It is the independent group within the agency that does health ef- fects analysis for all the other offices in the agency, although each of them has some of their own technical capabilities. Next to me is Dr. Dick Hill. Dick is a medical doctor and he is the chief scientist in the Office of Pesticides and Toxic Substances. And on my left, Dr. Joseph Cotruvo. Dr. Cotruvo is, I think, a chemist who is knowledgeable in toxicology and has been in the drinking water program for a number of years. Mr. Chairman, I would like to read about the first 9 or 10 pages of my testimony. I have made some changes, so your staff may want to pick some words or some key word changes. In the second part from the testimony which I will not read, I also have a few changes, which I will give to your staff. I will also provide these changes for the record. As I have said, I am John Hernandez, Acting Administrator of the U.S. Environmental Protection Agency. I am pleased to have the opportunity to appear before you today to discuss principles re- lating to the evaluation of carcinogenicity data and the use of these evaluations in making policy decisions involving the control of car- cinogens in the environment. Decisionmakers at regulatory agencies like EPA are faced with an almost daily need to protect public health by limiting exposure PAGENO="0377" 313 to potential human carcinogens. This is an old, not a new, perplex- ing problem that has become more complicated as we acquire more and more understanding of the mechanisms of the toxic effects of chemicals on animals and humans. It is indeed a paradox that confounds reason. The more we know, the more difficult it has become to make socially acceptable deci- sions. Let me try to frame this problem succinctly. How should or can a public policymaker integrate and evaluate all the varied technical and scientific information on a particular chemical that ranges from fragmentary low-level human exposure data to detailed results of metabolism and toxic effects at high doses on animals, and do this integration while taking into account the requirements and constraints of the many statutory provisions through which regulatory actions are taken. No, not a new problem, but a complex one. The manner in which EPA conducts assessments of the carcino- genic potential of environmental contaminants and the associated impact on public health has been continually evolving since EPA was established in December, 1970,. In the early 1970's, EPA had no formal internal mechanism or guidelines for evaluating potential carcinogenic agents. Rather, it relied instead primarily on independent evaluations generated by individual scientists outside of the EPA. Because these outside evaluations sometimes were based on different ap- proaches and presented conflicting results, and because of the need for greater consistency and certainty in EPA's approach to the evaluation of carcinogenicity data, an agency committee and a group of outside experts were assembled to develop a set of guide- lines for the evaluation of cancer risk by the agency. This effort resulted in the publication of interim guidelines for assessing carcinogenic risk in the Federal Register on May 25, 1976. These guidelines provided for the establishment of the Car- cinogen Assessment Group in the EPA to provide a center of in- house expertise on cancer. As an aside, Mr. Chairman, Betty Anderson-CAG, Cancer As- sessment Group, is within her organization. It has both in-house ca- pabilities and some outside consultants. The approach was adopted for a two-step decision process. One, the predominantly scientific and analytical process of defining the risk, and two, as a separate matter, the process of making a regula- tory decision. Further, the guidelines called for examining risk and benefits to the extent permitted by the applicable statute in making decisions to regulate exposures to suspecte& carcinogens. The 1976 guidelines also provided general guidance for the evaluation of scientific data to define cancer risk. Although there have been refinements as the agency gained ex- perience with the risk assessment process, the general approach de- fined in 1976 is still in use by the agency. The adopted approach called for a risk assessment process to answer two questions. How likely is the agent to be a human carcinogen? And assuming that the agent is a carcinogen, what is the magnitude of the potential public health impact? PAGENO="0378" 374 Since only rarely do we know for certain that an agent is indeed a human carcinogen, the first step involves an evaluation of all the biomedical data to determine the weight of evidence that an agent has carcinogenic potential for humans. The second step involves making rough estimates of public health impacts based on the po- tency of potential human carcinogens, the magnitude of current ex- posures, and the estimated exposures associated with various regu- latory options designed to reduce exposures. To answer the first question regarding likely carcinogenicity, the weight of the biomedical evidence is presented as ranging from the strongest evidence based on human data backed up by animal bio- assay data to merely suggestive evidence from studies showing bor- derline responses in humans or animals or positive results from short-term tests. In practice, however, most judgments on human carcinogenicity have been, must be, and in the future will probably continue to be based on primarily animal studies conducted at relatively high doses. Tumorgenic activity in animals is the signal that the agent might be a human carcinogen. However, the quality of the animal test data, the nature and magnitude of the responses, and the rela- tive merits of the studies showing positive and those showing nega- tive results are among the many factors considered in assessing the carcinogenic potential of suspected human carcinogens. The second step, that of providing some quantitative estimates of public health impacts, has been presented in terms of rough esti- mates describing the increased risk for individual population sub- groups that are exposed to the agent as well as for the total popula- tion. This second step is intended to give the regulator an idea of the potency of the suspected carcinogen, and when coupled with ex- posure information, some estimate of the magnitude of the public health problem. The potency factor is important because among the known car- cinogens, some are several million times more potent than others. There is an obvious need to take this into account in making public policy decisions, but a number of chemicals are subjected to similar analytical procedures. The decisionmaker can be presented with an array of risk projections that can be compared so that relative public health risks can be estimated. An added aspect of the quantitative risk assessment involves the extrapolation of dose-response relationships from comparatively high doses where most of the responses have been observed, usual- ly in animal bioassay studies to the much lower levels of exposure that most frequently occur in the environment. To make such extrapolations, the shape of the dose-response curve at low levels of exposure must be assumed because typically data are not generally available to describe the shape of the curve outside of the observed range, where most of the exposure occurs at relatively high doses in epidemiological studies or, as I have noted, principally in animal tests. Ideally, the selection of an extrapolation model for use in the low-level exposure range should be guided by knowledge of the mechanism by which each toxic agent acts. Unfortunately, much is still unknown about these mechanisms. In the absence of informa- tion that can lead to a conclusion on mechanisms, the common PAGENO="0379" 375 practice at EPA has been to make a conservative assumption that human health effects would follow a linear, non-threshold dose-re- sponse model at low doses. This assumption has been used because it has some biological basis and because it provides a plausible upper bound on the rela- tive individual risks associated with lifetime exposures at a particu- lar level, with the understanding that the lower bound of the risk could approach zero. Mr. FL0RI0. Dr. Hernandez, could I ask what page of your testi- mony you are on? Mr. HERNANDEZ. Seven. Mr. FL0RI0. Was this the new part of your testimony? Mr. HERNANDEZ. No, basically there are just a few line changes in the copy you have. I am going~ to about page 11 or so. And I think there are a couple of line changes in the next few lines. However, a variety of models can be used to emphasize the range of uncertainty in *these estimates. This is in line with current thinking in the scientific community, and with the advice of EPA's Science Advisory Board. The use of different mathematical models can produce widely different projections of risk from the same data base. The actual risk, however, is generally not known. In using quantitative estimates of risk, several points are impor- tant. First, the quantitative part of the risk assessment should be considered in light of the qualitative evaluation that describes the strength of evidence that indicates carcinogenetic potential. Second, the uncertainties in the quantitative estimates of cancer risks arise not only from the uncertainties inherent in the extrapo- lation from high to low dose and from animals to humans, but also from the limitations in the underlying data base for the individual chemical, and from the uncertainties in the exposure estimates. Third, because of our limited understanding of carcinogenic mechanisms, and because of modeling those mechanisms, risk models cannot be relied upon to provide absolute estimates of risk, but rather, they are most useful as rough indicators of the possible extent of the potential public health problem. It is therefore necessary to take these uncertainties and the un- certainties concerning the interaction of different chemicals into account when using any estimates of cancer risk in the decision process. In many ways, this general approach has worked reasonably well, particularly for examining risks and benefits and for setting priorities. The Agency is continuing the approach of examining the whole body of scientific and technical evidence for each suspected carcinogen to decide how likely it is to be a human carcinogen and its potency. Rodent bioassay studies remain at the core of our program to detect chemicals which have carcinogenic potential. Qualitative weight of evidence assessments for carcinogenicity, coupled with quantitative risk estimates, are being used together with other rel- evant information to make decisions regarding the regulation of po- tential carcinogens in the environment. Risk estimates are generally used as guidance. Because these methods are a young science, there are shortcomings, and these are reflected in the Agency's risk assessments. These shortcomings in PAGENO="0380" 376 the Agency's approach to risk assessment stem from the limited understanding of the mechanisms of cancer in animals and humans, and our limited ability to translate from animal effects to potential human health concerns. This results in an inability to precisely define the relative public health risks as a basis for setting standards, or as a basis for identi- fying target levels of risk. For example, we recognize that the upper bound estimates of risks based on linear extrapolations may not be equally plausible for all potential carcinogens. Such estimates would be unduly conservative should an agent have a threshold or a concave dose response curve in the low-dose range. It is therefore desirable to develop scientific approaches that can describe more accurately the potential risks associated with indi- vidual chemicals. But this is difficult to do because of our limited understanding of carcinogenetic mechanisms and our almost total lack of knowledge of the interactions of many chemicals and other risk factors to which humans are simultaneously exposed. Additional problems have been associated with the use of quanti- tative risk assessment. For example, there sometimes has been a tendency to ignore the quality of the carcinogenesis evidence in the final risk assessment judgments. Since the evidence that indicates the likelihood that an agent might be a human carcinogin can range from very weak to very strong, the interest of public health protection is best served by a process that considers the strength of evidence for carcinogenicity, as well as the potency and exposure in formulating health guidance. I will now jump to page 11 and we will finish up. I had men- tioned earlier the importance of being mindful of all factors that affect the reliability of the quantitative risk estimation, wherever estimates of risk are used. This is not always done. For example, in some cases upper bound estimates of risk have been used as actual estimates of risk, and in some cases, numerical results have been presented without an expression of their uncertainties, assump- tions or limitations. Finally, there is a rapidly evolving body of information about the carcinogenic process, which may provide a better basis for assess- ing cancer risk. All the factors I have mentioned have led scientists in EPA and in other agencies and outside the Government to con- sider ways to improve risk assessment approaches. For example, EPA has considered the ways to redefine its ap- proach to risk assessment to improve guidance for programs charged with regulating carcinogens, but has not yet adopted any changes. In addition, the EPA and other agencies are focussing at- tention on research initiatives to supp.ort risk assessment. Also, as we have heard today, the White House Office of Science and Technology Policy is chairing an interagency committee to adopt principles to guide Federal regulatory agencies in the identi- fication and characterization of potential human carcinogens. The ongoing EPA efforts to refine internal risk assessment ap- proaches, the broad Federal effort chaired by the OSTP, and the contributions of other scientists and organizations, should lead to PAGENO="0381" the incorporation of the most current scientific information about cancer causation into our risk assessment process. Risk assessment techniques have evolved since the publication of the interim guidelines 6 years ago, and will continue to do so. This is a rapidly developing field, but there are still large gaps in the scientific knowledge which must be filled. In order, to make real improvements in risk assessment methodology, the expanding body of knowledge puts regulators on the cutting edge of science. This concludes my oral statement, Mr. Chairman, and I will ask, as you have indicated, that the remaining paragraphs of my testi- mony, which basically reply to many of the questions in your letter, be put in the record and I will be pleased, and my colleagues here will be pleased to answer any questions that you all may have. [Testimony resumes on p. 403.] [Mr. Hernandez' prepared statement follows:] PAGENO="0382" 378 STATEMENT OF DR. JOHN HERNANDEZ DEPUTY ADMINISTRATOR U.S. ENVIRONMENTAL PROTECTION AGENCY BEFORE THE SUBCOMMITTEE ON COMMERCE, TRANSPORTATION,.AND TOURISM COMMITTEE ON ENERGY AND COMMERCE HOUSE OF REPRESENTATIVE WASHINGTON, D.C. MARCH 17, 1983 GOOD MORNING, MR. CHAIRMAN AND MEMBERS OF THE COMMITTEE. I AM JOHN HERNANDEZ, DEPUTY ADMINISTRATOR OF THE U.S. ENVIRONMENTAL PROTECTION AGEI~CY (EPA). I AM PLEASED TO HAVE THE OPPORTUNITY TO APPEAR BEFORE YOU TODAY TO DISCUSS PRINCIPLES RELATING TO THE EVALUATION OF CARCINOGENICITY DATA AND THE USE OF THESE EVALUATIONS IN MAKING POLICY DECISIONS INVOLVING THE CONTROL OF CARCINOGENS IN THE ENVIRONMENT. DECISION MAKERS AT REGULATORY AGENCIES LIKE EPA ARE FACED WITH AN ALMOST DAILY NEED TO ACT TO PROTECT AND TO LIMIT THE PUBLIC EXPOSURE TO POTENTIAL HUMAN CARCINOGENS. THIS IS AN OLD, NOT NEW, PERPLEXING PROBLEM THAT HAS BECOME MORE COMPLICATED AS WE ACQUIRE MORE AND MORE UNDERSTANDING ON THE MECHANISMS OF THE TOXIC EFFECTS OF CHEMICALS ON ANIMALS AND HUMANS. IT IS INDEED A PARADOX THAT CONFOUNDS REASON: THE MORE WE KNOW, THE MORE DIFFICULT IT HAS BECOME TO MAKE SOCIALLY ACCEPTABLE DECISIONS. LET ME TRY TO FRAME THE PROBLEM SUCCINCTLY: HOW SHOULD A SOCIAL POLICY MAKER INTEGRATE AND EVALUATE ALL OF THE GREATLY PAGENO="0383" 379 VARIED TECHNICAL AND SCIENTIFIC INFORMATION. ON A PARTICULAR CHEMICAL THAT RANGES FROM FRAGMENTARY LOW LEVEL HUMAN EXPOSURE DATA TO DETAILED RESULTS OF METABOLISM AND TOXIC EFFECTS OF HIGH DOSES ON ANIMALS, TAKING INTO ACCOUNT THE REQUIREMENTS AND CONSTRA~INTS OF THE MANY STATUTORY PROVISIONS THROUGH WHICH REGULATORY ACTIONS MAY BE TAKEN? No, NOT A NEW PROBLEM BUT A MUCH MORE COMPLEX ONE. THE MANNER IN WHICH EPA CONDUCTS ASSESSMENTS OF THE CARCINOGENIC POTENTIAL OF ENVIRONMENTAL CONTAMINANTS AND THE ASSOCIATED IMPACT ON PUBLIC HEALTH HAS BEEN CONTINUALLY EVOLVING SINCE EPA WAS ESTABLISHED IN DECEMBER, 1970. IN THE EARLY 70's, THE EPA HAD NO FORMAL INTERNAL MECHANISM OR GUIDELINES FOR EVALUATING POTENTIAL CARCINOGENIC AGENTS, RATHER, IT RELIED INSTEAD PRIMARILY ON INDEPENDENT E'IALUATIONS GENERATED BY INDIVIDUAL SCIENTISTS OUTSIDE THE EPA. BECAUSE THESE OUTSIDE EVALUATIONS SOMETIMES WERE BASED ON DIFFERENT APPROACHES AND PRESENTED CONFLICTING RESULTS, AND BECAUSE OF A NEED FOR GREATER CONSISTENCY AND CERTAINTY IN EPA's APPROACH TO THE EVALUATION OF CARCINOGENICITY DATA; AN AGENCY COMMITTEE AND A GROUP OF OUTSIDE EXPERTS WERE ASSEMBLED TO DEVELOP A SET OF GUIDELINES FOR THE EVALUATION OF CANCER RISK BY THE AGENCY. THIS EFFORT RESULTED PAGENO="0384" 380 IN THE PUBLICATION OF INTERIM GUIDELINES FOR ASSESSING CARCINOGENIC RISK IN THE FEDERAL REGISTER ON MM 25, 1976. THESE GUIDELINES PROVIDED FOR THE ESTABLISHMENT OF THE CARCINOGEN ASSESSMENT GROUP IN THE EPA TO PROVIDE A CENTER OF INHOUSE EXPERTISE ON CANCER. THE APPROACH WHICH WAS ADOPTED PROVIDED FOR A TWOSTEP DECISION PROCESS: .1) THE PREDOMINANTLY SCIENTIFIC AND ANALYTIC PROCESS OF DEFINING THE RISK, AND 2) AS A SEPARATE MATTER, THE PROCESS OF MAKING A REGULATORY DECISION. FURTHER, THE GUIDELINES CALLED FOR EXAMINING RISKS AND BENEFITS, TO THE EXTENT PERMITTED BY THE APPLICABLE STATUTE, IN MAKING DECISIONS TO REGULATE EXPOSURE TO SUSPECTED CARCINOGENS* THE 1976 GUIDELINES ALSO PROVIDED GENERAL GUIDANCE FOR THE EVALUATION OF SCIENTIFIC DATA TO DEFINE CANCER RISK. ALTHOUGH THERE HAVE BEEN REFINEMENTS AS THE AGENCY GAINED EXPERIENCE WITH THE RISK ASSESSMENT PROCESS, THE GENERAL APPROACH DEFINED IN 1976 IS STILL IN USE BY THE AGENCY. THE ADOPTED APPROACH CALLED FOR A RISK ASSESSMENT PROCESS TO ANSWER TWO QUESTIONS: 1) How LIKELY IS THE AGENT TO BE A HUMAN CARCINOGEN? AND 2) ASSUMING THAT THE AGENT IS A CARCINOGEN, WHAT IS THE MAGNITUDE OF THE POTENTIAL PUBLIC HEALTH IMPACT? PAGENO="0385" ~Q1 Uc, UJ THE E(iTEET:~ LT F ii *TELCE EEL H CALC iF;OETN~ C 77 H in 22-14% O-~3%---2E PAGENO="0386" PAGENO="0387" PAGENO="0388" PAGENO="0389" 385 THIRD, BECAUSE OF OUR LIMITED UNDERSTANDING OF CARCINOGENIC MECHANISMS, AND METHODS OF MODELLING THOSE MECHANISMS, RISK MODELS CANNOT BE RELIED UPON TO PROVIDE ABSOLUTE ESTIMATES OF RISK~ RATHER, THEY ARE MOST USEFUL AS ROUGH INDICATORS OF THE EXTENT OF THE POTENTIAL PUBLIC HEALTH PROBLEM. IT IS, THEREFORE, NECESSARY TO TAKE THESE UNCERTAINTIES AND THE UNCERTAINTIES OF THE INTERACTION OF THE MANY SUBSTANCES INTO ACCOUNT AND TO KEEP THE MANY AND VARIED ASSUMPTIONS IN MIND WHEN USING ANY ESTIMATES OF CANCER RISKS IN-THE DECISION PROCESS. IN MANY WAYS THIS GENERAL APPROACH HAS WORKED REASONABLY WELL, PARTICULARLY FOR EXAMINING RISKS AND BENEFITS AND FOR SETTING PRIORITIES. THE AGENCY IS CONTINUING THE APPROACH OF EXAMINING THE WHOLE BODY OF SCIENTIFIC AND TECHNICAL EVIDENCE ON EACH SUSPECTED CARCINOGEN TO DECIDE HOW LIKELY IT IS TO BE A HUMAN CARCINOGEN AND ITS POTENCY. RODENT BIOASSAY STUDIES REMAIN AT THE CORE OF OUR PROGRAM TO DETECT CHEMICALS WHICH HAVE CARCINOGENIC POTENTIAL. ~JALITATIVE WEIGHTOF-EVIDENCE ASSESSMENTS FOR CARCINOGENICITY, COUPLED WITH QUANTITATIVE RISK ESTIMATES, ARE BEING USED, TOGETHER WITH OTHER RELEVANT INFORMATION TO MAKE DECISIONS REGARDING THE REGULATION OF POTENTIAL CARCINOGENS IN THE ENVIRONMENT. PAGENO="0390" 386 RISK ASSESSMENT METHODS ARE GENERALLY USED AS GUIDANCE' BECAUSE THESE METHODS ARE A YOUNG SCIENCE, THERE ARE SHORTCOMINGS AND THESE ARE REFLECTED IN THE AGENCY'S RISK ASSESSMENTS' THESE SHORTCOMINGS IN THE AGENCY'S APPROACH TO RISK ASSESSMENT STEM FROM THE LIMITED UNDERSTANDING OF THE MECHANISMS OF. CANCER IN ANIMALS AND HUMANS AND OUR LIMITED ABILITY TO TRANSLATE FROM ANIMAL EFFECTS TO POTENTIAL HUMAN HEALTH CONCERNS. THIS RESULTS IN AN INABILITY TO PRECISELY DEFINE THE RELATIVE PUBLIC HEALTH RISKS AS A BASIS FOR SETTING STANDARDS, OR AS A BASIS FOR IDENTIFYING TARGET LEVELS OF RISK' FOR EXAMPLE, WE RECOGNIZE THAT UPPERBOLJND ESTIMATES OF RISKS BASED ON LINEAR EXTRAPOLATIONS MAY NOT BE EQUALLY PLAUSIBLE FOR ALL POTENTIAL CARCINOGENS' SUCH ESTIMATES WOULD BE UNDULY CONSERVATIVE SHOULD AN AGENT HAVE A THRESHOLD OR A CONCAVE DOSERESPONSE CURVE IN THE LOW DOSE RANGE' IT IS, THEREFORE, DESIRABLE TO DEVELOP SCIENTIFIC APPROACHES THAT CAN DESCRIBE MORE ACCURATELY THE POTENTIAL RISK ASSOCIATED WITH INDIVIDUAL CHEMICALS) BUT THIS IS DIFFICULT TO DO BECAUSE OF OUR LIMITED UNDERSTANDING OF CARCINOGENIC MECHANISMS AND OUR ALMOST TOTAL LACK OF KNOWLEDGE OF INTERACTIONS OF THE MANY CHEMICALS AND OTHER RISK FACTORS TO WHICH HUMANS ARE SIMULTANEOUSLY EXPOSED' PAGENO="0391" 387 ADDITIONAL PROBLEMS HAVE BEEN ASSOCIATED WITH THE USE OF QUANTITATIVE RISK ASSESSMENT. FOR EXAMPLE, THERE SOMETIMES HAS BEEN A TENDENCY AT EPA TO IGNORE THE QUALITY OF THE CARCINOGENICITY EVIDENCE IN THE FINAL RISK ASS~SSMENTS JUDGEMENTS. SINCE THE EVIDENCE INDICATING THE LIKELIHOOD THAT AN AGENT MIGHT BE A HUMAN CARCINOGEN CAN RANGE FROM VERY WEAK TO VERY STRONG, THE INTEREST OF PUBLIC HEALTH PROTECTION IS BEST SERVED BY ,A PROCESS THAT CONSIDERS THE STRENGTH OF EVIDENCE FOR CARCINOGENICITY AS WELL AS POTENCY AND EXPOSURE IN FORMULATING HEALTH GUIDANCE. THERE IS ALSO THE PROBLEM OF HOW TO INCORPORATE THE FOLLOWING TYPES OF SCIENTIFIC DATA IN MAKING RISK ESTIMATES: o THE CHEMICAL AND PHYSICAL PROPERTIES OF THE CHEMICAL o INFORMATION ON PHARMACOKINETICS o INFORMATION ON HUMAN AND ANIMAL METABOLISM o ESTIMATES OF LOW-DOSE TOXICITY o GENETIC TOXICOLOGY EVALUATIONS o ANALYSIS OF RELEVANT HEALTH EFFECTS DATA FROM HUMAN AND ANIMAL EXPOSURE o ANALYSIS OF HUMAN EPIDEMIOLOGICAL EXPERIENCE PAGENO="0392" 388 o CHEMICAL PRODUCTION, DISTRIBUTION, USE AND. FINAL DISPOSITION o MAJOR SOURCES OF EMISSIONS OF POLLUTANTS o INFORMATION ON DAILY, INTERMITTENT AND LONG TERM EXPOSURE o HUMAN EXPOSURE REGIMES o ESTIMATES OF EMISSIONS AND ESTIMATES OF THE POPULATION EXPOSED TO VARIOUS CONCENTRATIONS OF THE CHEMICAL I MENTIONED EARLIER THE IMPORTANCE OF BEING MINDFUL OF ALL FACTORS THAT AFFECT THE RELIABILITY OF QUANTITATIVE RISK ESTIMATION WHENEVER ESTIMATES OF RISKS ARE USED* WE SHOULD ALWAYS TAKE PAINS TO STATE CLEARLY WHAT A QUANTITATIVE RISK ASSESSMENT MEANS' IN SOME. CASES, WE HAVE USED -- UPPERBOUND ESTIMATES OF RISKS *AS ACTUAL ESTIMATES OF RISKS AND, IN SOME CASESJ NUMBERS HAVE BEEN PRESENTED WITHOUT AN EXPRESSION OF THEIR UNCERTAINTIES AND LIMITATIONS FINALLY, THERE IS A RAPIDLY EVOLVING BODY OF INFORMATION ABOUT THE CARCINOGENIC PROCESS WHICH MAY PROVIDE A BETTER BASIS FOR ASSESSING CANCER RISK' ALL OF THE FACTORS I HAVE MENTIONED HAVE LED SCIENTISTS IN EPA, IN OTHER AGENCIES, AND OUTSIDE THE GOVERNMENT TO CONSIDER WAYS TO IMPROVE RISK ASSESSMENT APPROACHES' FOR EXAMPLE, THE EPA HAS CONSIDERED WAYS PAGENO="0393" PAGENO="0394" PAGENO="0395" 3,1 - BLADDER CANCER IN HUMANS COULD BE A SPECIFIC EXISTING MUTATION PRESENT IN THE GENES OF ONLY A SMALL PROPORTION OF ALL THE POPULATION. WITH RAPID ADVANCES BEING MADE IN OUR UNDERSTANDING OF CANCER MECHANISMS, THERE IS REASON TO HOPE THAT MAJOR STRIDES CAN BE MADE IN CANCER RISK. ASSESSMENT IN THE NEAR FUTURE. *, OF COURSE, MANY SCIENTISTS SUGGEST CAUTION IN ALTERING THE CURRENT REASONABLY PRUDENT APPROACH TO CANCER RISK ASSESSMENT UNTIL SOME OF THE RECENT PROGRESS HAS FULLY UNFOLDED. AT THIS TIME, WE ARE NOT IN A POSITION TO DIFFERENTIATE BETWEEN CARCINOGENS ON THE BASIS OF MECHANISM. A. IIJMOR INITIATION AND PROMOTION - - INITIATION AND PROMOTION ARE TERMS ORIGINALLY DERIVED FROM THE TWO STAGE MODEL OF CARCINOGENESIS, AS DEMONSTRATED IN MOUSE SKIN. IN THAT MODEL, AN INITIATOR IS A SUBSTANCE WHICH CAUSES AN IRREVERSIBLE CHANGE IN THE TISSUE, THEREBY "INITIATING"THE CARCINOGENIC PROCESS. A PROMOTER IS A SUBSTANCE WHICH, WHEN APPLIED FOLLOWING AN INITIATOR, CAN ENHANCE OR "PROMOTE" THE ACTUAL EXPRESSION OF THE LATENT CARCINOGENIC TRANSFORMATION RESULTING FROM APPLICATION OF THE INITIATOR. SUBSTANCES WHICH CAN BOTH INITIATE AND PROMOTE CANCERS ARE KNOWN AS "COMPLETE CARCINOGENS." PAGENO="0396" 392 IT APPEARS THAT MANY INITIATING AGENTS AND CERTAIN OF THEIR METABOLITES HAVE BEEN SHOWN TO BE CAPABLE OF CAUSING MUTATIONS IN ANIMAL OR BACTERIAL TEST SYSTEMS AND/OR TO COVALENTLY BOND TO DWA AND OTHER CELLULAR MOLECULES, THAT SUBSTANCE IS GENERALLY CLASSIFIED AS AN INITIATOR' IT HAS BEEN PROPOSED THAT BECAUSE A SINGLE MOLECULE OF AN INITIATOR MAY THEORETICALLY BE CAPABLE OF CAUSING A GENETIC CHANGE WHICH SUBSEQUENTLY MAY BECOME CANCER, THE EXTRAPOLATION OF CARCINOGENIC RISK ASSOCIATED WITH EXPOSURE TO AN INITIATOR SHOULD BE EVALUATED BY ASSUMING THAT THE RESULTANT INCIDENCE OF CANCER IS PROPORTIONAL TO DOSE* BECAUSE INITIATORS APPEAR TO OPERATE BY DAMAGING OR ALTERING GENETIC MATERIAL, THEY ARE SOMETIMES REFERRED TO AS "GENOTOXIC AGENTS" PROMOTERS, ON THE OTHER HAND, DO NOT APPEAR TO CAUSE MUTATIONS OR COVALENTLY BOND TO DNA, BUT APPEAR INSTEAD TO ENHANCE OR AMPLIFY EXPRESSION OF THE GENETIC CHANGE CAUSED BY THE INITIATOR' SEVERAL OF SUCH AGENTS DO NOT APPEAR TO CAUSE MUTATIONS OR BIND TO DNA. ALTHOUGH MANY PRESUMED PROMOTERS ARE KNOWN TO INDUCE ENZYME ACTIVITY OR TOXICITY IN TARGET TISSUE, THE MECHANISM UNDERLYING PROMOTION IS NOT WELL UNDERSTOOD~ GIVEN THE RESULTANT UNCERTAINTY ABOUT THE RELATION BETWEEN DOSE AND PAGENO="0397" 393 RESPONSE FOR PROMOTION, SCIENTISTS DIFFER CONCERNING THE BEST APPROACH TO EVALUATION OF THE CARCINOGENIC RISK ASSOCIATED WITH EXPOSURE TO A PROMOTER. PROMOTERS BELONG TO THE CLASS OF COMPOUNDS SOMETIMES REFERRED TO AS "EPIGENETIC" AGENTS. PRACTICALLY SPEAKING, WHEN AN AGENT HAS BEEN SHOWN TO HAVE PROMOTING ACTIVITY, IT IS DIFFICULT TO BE SURE THAT IT DOES NOT ALSO HAVE SOME INITIATING ACTIVITY AS WELLJ THAT IS, THAT IT DOES NOT ALSO ACT AS A COMPLETE CARCINOGEN. IN ADDITION, SINCE PEOPLE MAY BE CONSTANTLY EXPOSED TO INITIATORS IT MAY NOT BE MEANINGFUL TO DIFFERENTIATE PROMOTION FROM COMPLETE CARCINOGENESIa INITIATION WHEN CONSIDERING ENVIRONMENTAL RISK. THIS IS THECA5E FOR 2,317,8-TCDJJ WHICH I WILL DISCUSS LATER. IT IS THOUGHT THAT SHOULD AN AGENT ACT SOLELY AS A PROMOTER, THE LINEAR EXTRAPOLATION MODEL WHICH PLACES AN UPPER BOUND ON RISKS MAY AT LEAST THEORETICALLY OVERESTIMATE THE RISK. HOWEVER, WE MUST KEEP IN MIND THAT A PROMOTER MAY ENHANCE THE EFFECT OF OTHER INITIATORS PRESENT IN THE DIET OR IN THE ENVIRONMENT. B. OXIC AND NON-GENOTOXIC A~~NJ~ WE HAVE BEEN DISCUSSING ThE POSSIBLE APPROPRIATENESS OF REGULATING AGENTS THAT DO ALTER DNA (MUTAGENIC CARCINOGENS, SOMETIMES CALLED GENOTOXIC PAGENO="0398" 394 AGENTS) DIFFERENTLY FROM AGENTS THAT DO NOT DAMAGE DNA (NON-GENOTOXIC CARCINOGENS, SOMETIMES REFERRED TO AS EPIGENETIC AGENTS). AT PRESENT, THERE APPEARS TO BE NO PERSUASIVE SCIENTIFIC BASIS FOR ADOPTING SUCH A *CLEAR REGULATORY DISTINCTIONS THIS IS AN AREA THAT IS BEING ACTIVELY STUDIED. THE CURRENT AGENCY POSITION IS TO TREAT AGENTS WHICH CLEARLY DEMONSTRATE A CARCINOGENIC EFFECT IN ANIMAL BIOASSAY STUDIES AS POTENTIAL HUMAN CARCINOGENS WHILE RECOGNIZING THAT CONVENTIONAL MODELS WHICH PLACE A PLAUSIBLE UPPER BOUND ON RISK MAY OVERESTIMATE THE RISK AT LOW DOSES FOR NONGENOTOXIC AGENTS. C. THRESHOLDS FOR POTENTIAL CARCINOGENS THE EXISTENCE OF A THRESHOLD FOR A GIVEN CARCINOGEN OR CLASS OF CARCINOGENS CAN ONLY BE DETERMINED BY UNDERSTANDING THE MECHANISM FOR THAT CARCINOGEN OR CLASS OF ACTION AT LOW DOSES. IT SEEMS REASONABLE THAT AGENTS THAT ACT THROUGH INTERFERENCE WITH HORMONAL FUNCTION, IMMUNOSUPPRESSION, OR IRRITANT ACTIVITY, E*G~, CRYSTAL FORMATION IN THE URINARY BLADDER, MAY HAVE THRESHOLDS~ HOWEVER, SUSPECTED CARCINOGENS ARE BEING EVALUATED USING EXTRAPOLATION MODELS WHICH ASSUME THE LACK OF A THRESHOLD BECAUSE OF THE ABSENCE OF ADEQUATE INFORMATION ABOUT MECHANISMS TO ESTABLISH SCIENTIFIC PAGENO="0399" 395 CONSENSUS' THE AGENCY CURRENTLY EXAMINES DATA ON VARIOUS INDIVIDUAL CHEMICALS TO CONSIDER CAREFULLY MECHANISTIC CLUES AND WE ARE WORKING WITH OTHERS TO DEVELOP SOME GUIDANCE ON THIS MATTER. 2. RODENT BIOASSAY AND SHORT TERM JJ1 ~1I~Q. TESTS WITH. RESPECT TO THE PREDICTIVE VALUE OF RODENT BIOASSAYS AND SHORT TERM IN ~IIBQ TESTS, THE AGENCY'S CURRENT POSITION IS MUCH THE SAME AS THE POSITION ADOPTED IN THE 1976 GUIDELINES. RODENT BIOASSAYS ARE USED TO IDENTIFY, AGENTS THAT ARE CARCINOGENIC IN THOSE MODELS. BY REVIEWING ALL RELEVANT BIOLOGICAL INFORMATION ON A CHEMICAL, WE THEN EVALUATE WHETHER THE CHEMICAL MAY POSE A CARCINOGENIC HAZAF~D TO HUMANS. SHORT-TERM .111 fl~Q~ TESTS PROVIDE: SUPPORTIVE EVIDENCE FOR CARCINOGENICITY; MAY PROVIDE CLUES TO MECHANISMS OF ACTION; HELP TO SET PRIORITIES FOR FURTHER TESTING OF CHEMICALS THAT ARE POORLY CHARACTERIZED; AND ARE USED IN OUR TOXIC SUBSTANCES PROGRAM TO PROVIDE PRELIMINARY INFORMATION AS TO POSSIBLE BIOLOGICAL REACTIVITY OF NEW CHEMICALS BEING EVALUATED UNDER PREMANUFACTURE NOTIFICATION. 3. EVALUATION OF CONFLICTING DATA WHERE THERE ARE CONFLICTING RESULTS, THE EVALUATION OF CARCINOGENICITY DATA REQUIRES AN ESPECIALLY CAREFUL EXAMINATION OF THE APPARENTLY POSITIVE PAGENO="0400" 396 AND THE APPARENTLY NEGATIVE STUDIES TO SEE IF EITHER THE DESIGN OR CONDUCT OF THE STUDIES COULD ACCOUNT FOR THE DIFFERENCES' THERE IS A WIDE RANGE OF SUSCEPTIBILITY AMONG VARIOUS SPECIES AND STRAINS OF BIOASSAY ANIMALS' NEGATIVE RESULTS IN ONE BIOASSAY SYSTEM MAY NOT NECESSARILY DISCOUNT THE SIGNIFICANCE OF VALID POSITIVE RESULTS IN OTHER BIOASSAYS AS INDICATORS OF POTENTIAL HUMAN CARCINOGENS' POSITIVE RESPONSES IN MULTIPLE TEST SYSTEMS, HOWEVER, DO PROVIDE STRONGER EVIDENCE THAT A CHEMICAL MAY POSE A CARCINOGENIC HAZARD' HUMAN DATA ARE HIGHLY VALUED IN THE RISK ASSESSMENT PROCESS AND WHERE AVAILABLE, ARE PREFERRED FOR RISK ASSESSMENT BECAUSE OF THEIR DIRECT PERTINENCE TO THE DEFINITION OF LIKELY HUMAN RISKS' THERE ARE GREAT DIFFICULTIES IN CONDUCTING EPIDEMIOLOGICAL STUDIES, SUCH AS FINDING AN ADEQUATE SAMPLE SIZE, AVOIDING CONFOUNDING EXPOSURE FACTORS, DEFINING THE EXPOSURES AND HEALTH EFFECTS OF INTEREST, AND CONDUCTING THE NECESSARY LONGTERM FOLLOWUP* THUS, MOST EPIDEMIOLOGIC STUDIES CAN DETECT ONLY RELATIVELY HIGH RATES OF CANCER INDUCTION' THAT IS, THE EXISTENCE OF SIGNIFICANT CANCER LEVELS MAY NOT BE DETECTED IN A POPULATION UNDER STUDY BECAUSE OF THE METHODOLOGY LIMITATIONS' APPARENTLY, NEGATIVE HUMAN STUDIES, ARE DIFFICuLT TO INTERPRET PAGENO="0401" 397 AND CANNOT NEGATE THE POSSIBILITY THAT THE AGENT CAN INDUCE HUMAN CANCERS, PARTICULARLY WHERE STRONG POSITIVE EVIDENCE IN ANIMAL BIOASSAY STUDIES HAS BEEN DEMONSTRATED. EVEN APPARENTLY NEGATIVE EPIDEMIOLOGICAL STUDIES, IF PROPERLY DESIGNED AND CONDUCTED, MAY HAVE SOME UTILITY IN PLACING A VERY APPROXIMATE UPPER LIMIT ON THE RESPONSE IN HUMANS. 14. METHODS OF RISK ESTIMATION, EXTRAPOLATION, AND ASSESSMENT EARLIER IN THIS TESTIMONY, I DISCUSSED THE USE AND LIMITATIONS OF CONVENTIONAL EXTRAPOLATION MODELS IN PROVIDING REGULATORS TO PROVIDE LEGISLATORS WITH SOME ROUGH INDICATION OF POTENCY AND POTENTIAL PUBLIC HEALTH IMPACTS. THIS IS AN AREA UNDER INTENSIVE STUDY AMONG SCIENTISTS INVOLVED IN RISK ASSESSMENT. THE AGENCY WILL CONTINUE ITS CASEBYCASE EVALUATIONS TO SUSPECT CARCINOGENS TO SEE IF THE DATA PERMIT A MORE ACCURATE DESCRIPTION OF THE DOSERESPONSE RELATIONSHIPS FOR THOSE CHEMICALS AT LOW DOSES. IN THE INTERIM, CONVENTIONAL APPROACHES MUST BE RELIED UPON IN THE MAJORITY OF INSTANCES UNTIL BETTER GUIDANCE IS AVAILABLE. 5. IMPLICATIONS OF CURRENT VIEWS ON MECHANISMS OF ACTIONS FOR THE NATURE AND LEVEL OF CONTROL FOR 2,3,7,8-TCDD. OUR CURRENT KNOWLEDGE DOES NOT PERMIT US TO DESCRIBE WITH CERTAINTY THE MECHANISM OF ACTION FOR 2,3,7,8-TCDD. BECAUSE 2,3,7,8-TCDD HAS BEEN 22-143 O-83--26 PAGENO="0402" 398 SHOWN lOBE AN EXTREMELY POTENT TUMOR PROMOTER IN THE RAT LIVER, 2,3,7,8-TCDD CLEARLY FALLS INTO THAT CLASS OF COMPOUNDS WHICH I HAVE REFERRED TO AS PROMOTERS' HOWEVER, WHILE SOME SCIENTISTS BELIEVE THAT 2,3,7,8-TCDD ACTS SOLELY AS A PROMOTER, 2,3,7,8-TCDD INDUCTION OF CANCER IN ANIMAL BIOASSAY STUDIES MAKES IT DIFFICULT TO RULE OUT 2,3,7,8-TCDD AS A COMPLETE CARCINOGEN. UNFORTUNATELY, THE AVAILABLE INFORMATION ON THE MUTAGENIC POTENTIAL OF 2,3,7,8-TCDD IS INCONCLUSIVE. THE IMPLICATIONS OF THESE DATA FOR REGULATION OF TCDD IS CONTINGENT OW THE SHAPE OF THE DOSERESPONSE CURVE AT LOW DOSE LEVELS. SOME SCIENTISTS WHO BELIEVE THAT 2~3,7,8-TCDD ACTS SOLELY AS A PROMOTER ALSO BELIEVE THAT ITS CARCINOGENIC ACTIVITY RESULTS FROM ITS TOXIC PROPERTIES AND THEREFORE THAT A THRESHOLD OF ACTION IS PLAUSIBLE. OTHER SCIENTISTS DO NOT BELIEVE THAT SUCH A THRESHOLD MECHANISM IS PLAUSIBLE, EVEN IF 2,3,7,8-TCDD DOES ACT ONLY AS A PROMOTER' THEY BELIEVE THAT OTHER MECHANISMS CAN BE PROPOSED WHICH MAKE NONTHRESHOLD, LOW DOSE EFFECTS PLAUSIBLE' IN SHORT, WE DO NOT KNOW FOR SURE' IN THE FACE OF UNCERTAINTY CONCERNING WHETHER OR NOT 2,3,7,8-TCDD IS A COMPLETE CARCINOGEN AND THE DOSE-RESPONSE RELATIONSHIP FOR ITS ACTIVITY AS A TUMOR PROMOTER, THE AGENCY HAS NOT RELIED EXCLUSIVELY ON EITHER THE PAGENO="0403" 399 RISK EXTRAPOLATION MODELS CONSIDERED MOST APPROPRIATE FOR COMPLETE CARGINOGENS OR ON THE MARGINOFSAFETY APPROACH OFTEN USED FOR TOXIC AGENTS. RATHER1 THE AGENCY HAS EVALUATED THE RISK 2,3,7,8-TCDD MAY POSE AS A CARCINOGEN BY ESTIMATING CANCER RISK USING CONVENTIONAL MODELS TO PLACE A PLAUSIBLE UPPERBOUND ON THE RISK, WHILE RECOGNIZING THAT THE RISKS COULD FALL OFF DISPROPORTIONATELY AT LOW DOSES FOR TCDD ACTING SOLELY AS A PROMOTER. 6. THE ROLE OF ECONOMIC CONSIDERATIONS IN THE CONTROL OF CARCINOGENS. A' THE TOXIC SUBSTANCES CONTROL ACT UNDER THE TOXIC SUBSTANCES CONTROL ACT (TSCA), THE AGENCY CANNOT REGULATE A CHEMICAL SUBSTANCE MERELY BECAUSE IT PRODUCES ADVERSE HEALTH EFFECTS AS A CARCINOGEN. To REGULATE A KNOWN CARCINOGEN, OR ANY OTHER HAZARDOUS SUBSTANCE UNDER TSCA, THE AGENCY MUST FIND THAT THE MANUFACTURING, PROCESSING, DISTRIBUTION IN COMMERCE, USE, OR DISPOSAL OF THE SUBSTANCE WILL PRESENT AN UNREASONABLE RISK OF INJURY TO HEALTH OR THE ENVIRONMENT. SECTION 6(c) OF TSCA GIVES US GENERAL GUIDANCE FOR MAKING SUCH A DETERMINATION. THE AGENCY MUST, CONSIDER: (1) THE EFFECTS OF THE SUBSTANCE ON HEALTH OR THE ENVIRONMENT AND THE MAGNITUDE OF EXPOSURE OF HUMAN BEINGS OR THE ENVIRONMENT TO THE SUBSTANCE, PAGENO="0404" 400 (2) ThE BENEFITS OF THE SUBSTANCE FOR VARIOUS USES AND THE AVAILABILITY OF SUBSTITUTES FOR THOSE USES, AND (3) THE REASONABLY ASCERTAINABLE ECONOMIC CONSEQUENCES OF REGULATING THE SUBSTANCE, AFTER CONSIDERING THE EFFECT OF ANY REGULATION ON THE NATIONAL ECONOMY, SMALL BUSINESS, TECHNOLOGICAL INNOVATION, THE ENVIRONMENT, AND PUBLIC HEALTH' THE ROLE OF ECONOMIC CONSIDERATIONS ENTERS THIS ANALYSIS IN THE SECOND AND THIRD ELEMENTS' IF THE COST TO SOCIETY OF THE ADVERSE EFFECTS OF THE SUBSTANCE ARE FOUND TO OUTWEIGH THE BENEFITS THE SUBSTANCE PRESENTS TO SOCIETY, REGULATION IS APPROPRIATES IN DECIDING ON THE LEVEL OF REGULATORY CONTROL, THE AGENCY CONSIDERS THE COST AND FEASIBILITY OF CONTROLS TO REDUCE EXPOSURE TO THE SUBSTANCE TO LEVELS WHERE ANY REMAINING RISK IS NOT UNREASONABLE THIS STATUTORY SCHEME MEANS THAT IN SOME CASES TOXIC SUBSTANCES MIGHT BE ALLOWED TO CONTINUE IN SOME CONTROLLED USES BECAUSE THE BENEFITS TO SOCIETY OF THOSE USES ARE HIGH' IN OTHER CASES, LESS TOXIC SUBSTANCES MIGHT BE BARRED FROM ANY USE BECAUSE THOSE USES PRESENT FEW, IF ANY, BENEFITS TO SOCIETY' - PAGENO="0405" 401 B. RESOURCE CONSERVATION AND RECOVERY ACT - SUBTITLE C OF THE RESOURCE CONSERVATION AND RECOVERY ACT, AS AMENDED (RCRA), REQUIRES EPA TO DEVELOP AND IMPLEMENT A FEDERAL PROGRAM FOR THE SAFE MANAGEMENT OF HAZARDOUS WASTES. AMONG THE WASTES DESIGNATED AS "HAZARDOUS" BY EPA, AND THEREFORE SUBJECT TO SUBTITLE C REQUIREMENTS, ARE WASTES WHICH CONTAIN MATERIALS WHICH ARE KNOWN OR SUSPECTED HUMAN OR ANIMAL CARCINOGENS' SUBTITLE COF RCRA IS SILENT ON THE ISSUE OF "HOW ECONOMIC CONSIDERATIONS ARE TO BE USED IN THE CONTROL OF CARCINOGENS" IN PARTICULAR AND SOLID WASTE IN GENERAL~ IT NEITHER PROHIBITS NOR REQUIRES THE AGENCY TO ACCOUNT FOR SUCH CONSIDERATIONS. THE AGENCY HAS TAKEN THE POSITION THAT IT MUST ESTABLISH THOSE STANDARDS NECESSARY TO ASSURE PROTECTION OF HUMAN HEALTH AND THE ENVIRONMENT. Lj5 FED. REG. 33089 (MAY 19, 1980). HOWEVER, THE AGENCY HAS ALSO TAKEN THE POSITION THAT IT WILL CONSIDER COST~EFFECTIVENESS - IN CHOOSING AMONG THE VARIOUS ALTERNATIVES THAT MIGHT MEET THE REQUIREMENTS OF THE STATUTE, AND THAT IT WILL ANALYZE THE ECONOMIC IMPACT OF ITS REGULATIONS FOR PURPOSES OF INFORMING CONGRESS OR THE PUBLIC.ABOUT THE - COST OF ITS REGULATORY PROGRAM' PAGENO="0406" 402 EPA HAS FOLLOWED THESE GENERAL GUIDELINES IN DEVELOPING ALL ITS REGULATIONS UNDER SUBTITLE C, INCLUDING THOSE APPLICABLE TO CARCINOGENIC WASTES. C' SUPERFUND THE COMPREHENSIVE ENVIRONMENTAL RESPONSE, COMPENSATION, AND LIABILITY ACT OF 1980 (SUPERFUND) AUTHORIZES THE FEDERAL GOVERNMENT TO RESPOND TO RELEASES OF HAZARDOUS SUBSTANCES INTO THE ENVIRONMENT EITHER BY TAKING DIRECT CLEANUP ACTION OR COMMENCING ENFORCEMENT ACTIONS. CARCINOGENSAS WELL AS MANY OTHER TOXIC CHEMICALS, WOULD BE INCLUDED IN SUPERFUND'S DEFINITION OF "HAZARDOUS SUBSTANCES"' SEE SECTION 101(1~). SUPERFUND CONTRASTS WITH OTHER ENVIRONMENTAL STATUTES SUCH AS RCRA OR TSCA IN THAT IT IS BASICALLY NOT A REGULATORY STATUTE IN THE TRADITIONAL SENSE' WHILE RCRA AND TSCA AUTHORIZES THE AGENCY TO PROMULGATE REGULATIONS TO CONTROL FUTURE ACTIVITIES INVOLVING CERTAIN HAZARDOUS WASTES AND TOXIC SUBSTANCES, SUPERFUND IS DIRECTED TOWARD GOVERNMENT OR RESPONSIBLE PARTY CLEANUP OF HAZARDOUS SUBSTANCES RELEASED INTO THE ENVIRONMENT IN THE PAST' SUPERFUND PROVIDES THAT ECONOMIC CONSIDERATIONS ARE TAKEN INTO ACCOUNT IN SELECTING CLEANUP ACT1ONS' SECTION 104(C)(14) OF THE ACT REQUIRES THAT THE PRESIDENT SELECT A CLEANUP ACTION WHICH PROVIDES A PAGENO="0407" 4O3 "COSrEFFECTIVE RESPONSE," WHICH ALSO :BALANCES BETWEEN THE NEED TO SPEND HAZARDOUS SUBSTANCE RESPONSE TRUST FUND MONEY AT THAT SITE AGAINST AND THE AVAILABILITY OF-MONEY IN THE FUND TO RESPOND TO RELEASES AT OTHER SITES. IN ADDITION, THE NATIONAL CONTINGENCY PLAN, PUBLISHED PURSUANT TO SECTION 105 OF SUPERFUND, PROVIDES THAT THE GOVERNMENT SHALL EVALUATE THE ADEQUACY OF RESPONSIBLE PARTY CLEANUP ACTIONS USING COSTEFFECTIVE CONSIDERATIONS. SEE 40 CFR 300.68(c)~ 47 ~ ~ 31180, 31216 (JULY 16, 1982). THE FUND-BALANCING CONSIDERATIONS ARE NOT APPLICABLE TO RESPONSIBLE PARTY CLEANUP ACTIONS. Mr. FL0RT0. Thank you very much. Your last or close to your last point about EPA going forward in developing its own risk assessment policy, I wonder how that is compatible with what we have heard most of the day, which is ev- eryone advocating interagency cooperation and coordination. And of course we know that there is, an overall cancer policy assessment that is in the process of being done, and we heard from the Presi- dent's science adviser that it was early on and that what is out there already is somewhat criticized, and in some respects justifi- ably. I just wonder why EPA would in fact be implementing its own cancer assessment policy before there is an overall determination that a new policy is warranted. And in fact you are apparently in the process of making decisions, whether it be under TSCA with the formaldehyde question or the groundwater statement which is the basis on which actual decisions are being made. Why is it that you are not just maintaining the existing stand- ards, the previous standards in making these decisions, rather than going forward on your own without waiting for the ultimate policy evaluation to be completed? Mr. HERNANDEZ. Mr. Chairman, I think that so far as I under- stand the actions we have taken, they are in the same context, the same framework as we have used in the past, under the 197G ap- proach that we have. There may be some differences, and clearly there are on that issue, but my basic feeling is that we still are doing that. My understanding of what: the OSTP paper will be, at least the first piece of it, will be a very general kind of guidance, and we were looking at some other ~: approaches which perhaps might be taken as not being in conflict with that thing, that would look at PAGENO="0408" 404 how you do a different approach to risk assessment, given this gen- eral framework of how do you define a carcinogen. Mr. FL0RT0. No, we are really not talking about general versus specific. We are talking about what will hopefully be, at least by the Government, uniformly accepted procedures and methodology for making determinations on risk assessment. I would hope that we don't end up with EPA using one methodology and everybody else in the Government using a different methodology. All I am suggesting is that until there is a uniform determina- tion as to a governmental policy, it seems to me to be premature for EPA to be making decisions on a different level, whether it be in formaldehyde or ground water or TCE, and we heard some testi- mony this morning about TCE determinations, as to what are ac- ceptable exposure limits. Therefore, it seems to me that what EPA should be doing is not making changes until there are determinations of uniformity, and as we heard from the President's adviser, that is not ready to be published at this point. Mr. HERNANDEZ. I fully agree with you, Mr. Chairman. Mr. FL0RI0. Let me get to the point that we had. Who is Dr. Pal- lotta, by the way? Mr. HERNANDEZ. As I understand it, he's a consultant to the Office of Solid Waste and Emergency Response, and has been there, I would guess, 9 months. Mr. FL0RI0. Is he still there, that you know of? Mr. HERNANDEZ. I'm advised that he is no longer-as a consult- ant to-- Mr. FL0RI0. He was Mrs. Lavelle's consultant dealing with this whole question of risk assessment? Mr. HERNANDEZ. I think he advised her of toxicity of chemicals and things and approaches. Mr. FL0RI0. Do you know the circumstances under which he left? Mr. HERNANDEZ. No, sir, I don't. Mr. FL0RI0. I have, and I am sure some of the people who were here this morning can convey to you, and have conveyed to you the questions that were raised with regard to TCE and specifically at Price's landfill, and their thoughts that the initial reports that came out of the early monitoring that went on were reinterpreted so as to, under a new set of evaluating risk assessment standards, were evaluated in such a way as to allow different levels of accept- ance, and that in fact the results of that reevaluation were such that they would dictate different outcomes on cleanup standards. And the representation was made that the reevaluation would allow for additional exposure of 172-fold human exposure to TCE. I wonder if you can give us your recollection of your role, be- cause it has been represented and, I have some documents, what your role was in permitting and acquiescing in this reevaluation. [See p. 415.] Mr. HERNANDEZ. I hate to say this. I am a total blank. I remem- ber one session we had with Dr. Pallotta on TCE, but I can't re- member that it was related to Price's landfill. Mr. FL0RI0. Well, I have a document that I will be happy to share with your office that is from Mrs. Lavelle, directed to you, in which she says that "When we met, there were two specific actions PAGENO="0409" 405 that you agreed to coordinate for affected offices. To recount, they were: and the first one is the reevaluation of the health advisory on TCE as the highest priority. Second is the development of a threshold model risk assessment for nongenotoxic chemicals such as TCE. Now my only concern is that you have stated that you agree that the existing law, the 1976 approach, which does not contemplate di- visions of carcinogenic materials into the two categories, is the ex- isting policy. Yet this appears to indicate an agreement to develop a model, and then the testimony we heard today was that that model was used to reevaluate the exposure limits of TCE at Price's Landfill and wherever else. Mr. HERNANDEZ. I don't know what was used at Price's Landfill, and so I can't guess at that. Mr. FL0RI0. Do you have a recollection of an agreement to use this new approach in dealing with TCE's? Mr. HERNANDEZ. Not on any kind of specific case. As I recall, Ms. Lavelle spoke to me a couple of times about how we did risk assess- ment and who should do them, and unfortunately, she had no tech- nical support at all early on. I told her that we were looking at a variety of methods of risk assessment, one of which was: is there any kind of a basis for doing a difference between so-called epigenetic and other things that are really initiators of cancer. This is the piece that I think you have heard some about today that came from the CAG group as a draft document for other scien- tists to comment on. I think the memo that you talked about,