SESQUI NEW YORK UNIVERSITY MEDICAL CENTER A private university in the public service Institute of Environmental Medicine 550 FIRST AVENUE, NEW YORK, N.Y. 10016 AREA 212 340-5280 ANTHONY I LANZA RESEARCH LABORATORIES AT UNIVERSITY VALLEY LONG MEADOW ROAD, STERLING FOREST, TUXEDO, N.Y. 10987 MAIL AND TELEPHONE ADDRESS: 550 FIRST AVENUE, NEW YORK, N.Y. 10016 February 7, 1983 Dr. Charles M. Benbrook Staff Director House of Representatives Committee on Agriculture Room 1301, Longworth House Office Building Dear Dr. Benbrook: Thank you for your letter of January 6 and your memorandum of me. In reply to the questions you have raised, I do not feel competent to comment on whether there have been changes in EPA's pesticide regulatory policies. I can assure you, however, that I know of no scientific evidence to justify a major departure from the approaches to carcinogen risk assessment that have been endorsed by the IARC (eg, IARC Monograph Suppl. 4, 1982) and the IRLG (eg, IRLG, 1979). The possible existence of different classes of carcinogenic chemicals (such as initiators and promoters, genotoxic and non-genotoxic agents, etc.) may ultimately call for differing approaches in risk assessment, but our knowledge of the action of such substances is still too limited, in my opinion, to enable confident distinctions to be drawn for purposes of regulatory decision-making. Thank you for the privilege of commenting on this matter. If there is anything I can do to be of further help to you, please don't hesitate to call on me. js Sincerely, Сини е ция Arthur C. Upton, M.D. WILLIAM M. UPHOLT PH.D. 525 EAST INDIAN SPRING DRIVE SILVER SPRING, MARYLAND 20901 301-588-5389 March 22, 1983 Dr. Charles M. Benbrook Staff Director Subcommittee on Departmental Operations, Research and Foreign Agriculture U.S. House of Representatives, Washington, DC 20515 Dear Dr. Benbrook: I am glad I finally got a chance to see your report on EPA Pesticide Regulatory Programs. I am sorry I did not receive it in time to get my comments to you before the recent hearings, but I am not sure they would have been pertinent to those hearings anyway. Your report is very impressive in its clear, easily understood description of the present laws under which EPA regulates pesticides and how they are currently interpreted and implemented. The changes since I retired at the end of 1977 are not too great, but very significant. I am concerned by most of the same developments as are you. I think the matter of "cut and paste" has been over emphasized, at least in the press. Were that the most serious problem, I would not be too agitated. I am more concerned with Roy Albert's suggested change in cancer policy. But even that is but the latest evidence of what I consider a very basic myth that underlies regulatory policy, and has ever since about the end of World War II. To explain the nature of the myth, I think I should first outline what I believe to be the basic concept of regulation. Specifically, it is that "safe" can be defined, as regards regulation of hazards, only as "of acceptable risk" and that the acceptability of any risk depends upon the benefits and costs of reducing the risk further (or in view of alternative available risks). But, what is acceptable depends upon the nature of the hazard (as well as costs and benefits) and the perception of the seriousness of the risks in the minds of the concerned public. Thus, the regulatory decisions must be made by a politically sensitive administrator who is sensitive to the perception of the public, as well as understanding of the costs, benefits, and risks involved. Finally, from a practical standpoint, the only regulatory actions available to the regulator are designed Dr. Benbrook March 22, 1983 Page 2 to reduce exposure. In the case of pesticides, reduced exposure can be achieved by a wide variety of mechanisms such as selection of the toxicant, modification of formulation, method of application, dosage, timing of application, restrictions on uses, methods of storage, transport and disposal, and, of course, prohibition of use by untrained operators or even prohibition of manufacture. These methods of reducing exposure create the alternative regulations that must be considered by the regulatory decision-maker. In order to estimate the risk involved in any particular regulation, the responsible official (and hopefully the concerned public) must recognize the principle that no substance is so toxic that it cannot be used "safely" just as no substance is so "safe" that it cannot be used dangerously. In the early 1950's, regulatory officials generally did not recognize these principles and as a result they viewed their responsibility as assuring a "safe" food supply and that only "safe" pesticides be used. On the other hand, toxicologists who recognized the above principles found it easier to identify for the administrator a level that they felt was of negligible risk, or some other "judgemental" term and, to label it as "safe" as though it represented an absolute zero risk. Their method of doing this in terms of tolerance for pesticides as required under Sec. 408 of FFDCA, was to misuse a dosage-response characteristic of most pesticides. Specifically, most toxic chemicals produce more than one type of effect in living tissue but usually one predominates at higher dosages. Thus, as an example, even one molecule of cholinesterase inhibitor will inhibit a molecule of cholinesterase, but it will be undetected by the most sensitive analytical methods available. Larger dosages will produce a level of inhibition that is detectable by suitable methods but will not be observable clinically. Higher dosages produce some signs or symptoms that are trivial. As the dosage goes still higher, the signs or symptoms change markedly and at something over 80% inhibition, death may follow very promptly. With some materials, like vitamins or fluorides, very small dosages may produce desirable effects, but at higher dosages the effects become adverse. Clearly, it is important to know what type of effect a particular toxicant produces in humans. Hopefully, it will produce similar effects in properly selected experimental animals. Thus, in general, experimental animals can provide useful information on the adverse effects that can be expected in humans and their relationship to dosage. This phenomenon permitted toxicologists to construct dosage response curves and at lower dosages no adverse effects would be detected. This, of course, is the basis for the presumed "no effect level." It is clearly improperly labeled for there are always some effects on cells at lower levels. "No adverse effect level" was a better term but even that was inaccurate since lower levels ordinarily produce an effect that may be adverse cumulatively. Thus, the term "no observed effect level" is better. Unfortunately, there is a tendency to change that accurate term to a completely unacceptable term, "No observable effect level." No toxicologist should defend that term because what is observable depends upon the techniques used in searching for an effect. Thus, "NOEL" should be used only for Dr. Benbrook March 22, 1983 Page 3 "No observed effect using the number of animals and the analytical or clinical techniques described." I apologize for belaboring this apparently trivial point but it has become the key to some of the most important problems you identified in your report. From "no observable effect level" it is reasonable to assume that there is a threshold below which no effect occurs. Such a threshold may in fact exist but I once heard a highly respected pathologist say that he believed that with adequate electromicroscopy it will eventually be possible to visualize the effect of any single molecule on a cell. I am in no position to argue the validity of that opinion, but I am confident that to assume the presence of a threshold dosage is useless unless it can be determined either experimentally or theoretically. (There are special cases such as fluoride in drinking water that have a determinable level that is optimal and thus could be considered the equivalent of a threshold.) I am sure you will see that if there is no demonstrable threshold, the muchdiscussed difference between carcinogens and other toxicants becomes of little importance in terms of extrapolation models for dosage/response determinations. Of course the fact often emphasized by Dr. Saffiotti and others that a single hit may initiate a self-generating growth that eventually may kill the individual is of tremendous importance in terms of the acceptability of any particular level of risk. Thus, it is a very important factor in risk/benefit analysis even though it does not validate the NOEL safety factor as a scientifically sound method for other risk/benefit analyses. This, of course, is why I am disturbed by Dr. Albert's suggestion to use it for epigenetic carcinogens. It is no more valid for them than it is for other toxicants. Perhaps this will explain why I object to current methods of setting tolerances even though I have nothing better to suggest. What we need is better models for dosage/response of all toxicants as well as all carcinogens. If we can develop such models (and I suspect several will be needed to cover all pesticides) then we will have a much sounder basis for quantifying estimates of risk associated with each alternative regulatory option under consideration. Also, we need much better methods of estimating exposure. Those used for oral exposure in establishing tolerance are reasonably sophisticated, but estimates of exposure by dermal and respiratory routes are still very poor. I believe EPA should require exposure estimates of comparable quality to toxicological assessments now required. Then, perhaps, we will also be able to present better (more quantified) descriptions of both risk and benefit (or costs, depending upon the chosen terminology) to the administrative decision-maker. Perhaps by that time our social scientists will also be better prepared to explain public perception of acceptable levels of various risks to the administrator as well. Meanwhile, I am also concerned about another aspect of our regulation. Is not our real objective to protect humans and their environment from excessive adverse effects? If that is true, we should also be searching for a reliable method of Dr. Benbrook March 22, 1983 Page 4 assessing our success or failure. I have no ideas as to how to do that for our environment, but we should be able to set up a national, long-term epidemiological surveillance system that would detect changes in our health status attributable to pesticides. In 1967, the Pestcide Program of the PHS initiated such a program but it died in about 1972, due to a number of factors, but at least in part because it was short of sufficient epidemiological competence and resources. It is a massive undertaking and I believe it would require special legislative authorization to assure very long term commitment and multi-agency support to assure competency. A couple of years before I retired, I proposed this to the then Administrator of EPA and he invited the heads of NIH, CDC, and OSHA to join us in a discussion of the possibility. Much to my disappointment, the chief of EPA Research and the Director of the National Institute of Environmental Health each took the attitude that there was no need of inter-agency participation--either group said they could do it themselves given the necessary funds. The representative of the Surgeon General told me privately that we could never get those two to collaborate and that it was impracticable without them. I agreed, so I never called a second meeting of the group. But I still think it is the only reliable method of assessing the effectiveness of the regulatory programs. I apologize for the length of this letter, but appreciate the opportunity to express my concepts of needed changes in the EPA. Sincerely, Million shipholt William M. Upholt WMU/s1m |