Tobacco Documents Online

radiation (approximately 0.003 to 30,000 l4Hz). These two regions of the spectrum are emphasized because they are of regulatory concern to the agency and because the preponderance of information is in these regions. The EPA stated in the review draft that the evaluation of the likelihood of human cancer risk is based on a judgment as to the overall weiaht of evidence that a carcinogenic response is causally related to specific levels or types of exposure (emphasis added). Since the establishment of causality is often difficult, the weight of the evidence approach relies on the combination of empirical observations and inferences grounded in reasonable scientific judgment. Under this approach, the evidence from human studies is considered most important, with lesser importance being attached, respectively, to chronic lifetime animal studies and to ancillary evidence, such as short-term tests of genetic toxicity, mechanistic studies, and evidence of carcinogenicity for chemical analogues to the agent under study. Accordingly, the EMF draft document considers human, animal and supporting evidence in separate chapters. As was the case with diesel emissions, EPA held a preliminary scientific workshop for the purpose of discussing EMF issues. A revised draft was prepared, incorporating the suggestions made by the panel of expert reviewers. This second draft was issued

in October, 1990, and was subsequently reviewed by a subcommittee of the SAB, January 14-16, 1991. The subcommittee plans to meet at least once more before preparing the final report. C. BACKGROUND CONCERNING THE ETS REVIEW DRAFT In the Spring of 1990, OHEA prepared a preliminary draft document dealing with health risks associated with ETS ("ETS Review Draft"). That document, among other things, stated that exposure to ETS caused up to 3,800 lung cancer deaths per year and advocated that ETS be classified as a Group A carcinogen. The document was circulated internally at EPA for review, and it was eventually "leaked" to the press in May, 1990, prior to being released to the public. The EPA also issued a press release and allowed press interviews concerning the draft report's conclusions prior to issuing the draft or making it available to members of the "regulated community" (i.e., those members of industry who would be directly affected if ETS were to become a regulated substance). A separate draft policy guide on workplace smoking ("Policy Guide Draft"), containing statements regarding the risks of ETS, based on the ETS Review Draft's conclusions, was prepared and disseminated by the EPA as a review document at the same time. Both "final" draft ETS documents were released on June 25, 1990 and ultimately assigned by EPA to the Indoor Air Quality

Committee ("IAQC") of the SAB for review. The EC of the SAB selected members of the IAQC who would review the ETS drafts.2 The IAQC held an open meeting on December 4-5, 1990, to review the EPA draft documents. Part of the purpose of the meeting was to formally receive written and oral comments from interested parties on the draft documents to aid the IAQC in its independent review process. However, due to procedural ground rules established by the IAQC Chairman at the beginning of the meeting, there was virtually no substantive debate or discourse between members of the IAQC and the public. At the end of the meeting the draft documents were, in effect, remanded to EPA for additional work. The meeting was conducted as though there was no question about the scientific foundation of the draft report and as if the basic issues in the risk assessment were already resolved.3 2. In doing so, the EC selected at least one highly visible anti- smoking advocate and several others who had previously either participated in the drafting of the report or who had reviewed and commented upon the draft report prior to its release to the public. Coincidentally, no scientists were selected by the EC who had expressed views contrary to the draft report's recommendations. This is true despite the EC's announced purpose of selecting "neutral" scientists for membership in the IAQC. 3. In addition, several scientists with critical views concerning the two draft documents were not afforded any opportunity to speak -- in contrast to the fact that nearly unlimited time was allotted at the meeting for presentation by anti-smoking organizations. - 12 -

Despite the fact that some IAQC members expressed major criticisms of the drafts at the meeting, the Chairman of the IAQC closed the meeting by stating that the panel had reached a "consensus" on the classification of ETS as a human carcinogen. This statement was affirmed by the Chairman of the IAQC to the press at a press conference called by the Chairman immediately after the meeting. As a result, he was quoted extensively in the press the next day. A. THE INABILITY OR UNWILLINGNESS OF THE EPA TO FOLLOW ITS OWN RISK ASSESSMENT GUIDELINES The authors of the ETS Review Draft clearly did not follow the EPA Guidelines for risk assessment procedures. This omission is particularly glaring when contrasted with contemporaneous EPA work in other risk assessments. Although the ETS Review Draft claims to provide a weight of the evidence analysis for ETS as outlined in the EPA Guidelines, the approach taken in the ETS Review Draft is neither comprehensive nor critical.4 4. In its April 1990 review of guideline revisions for assessment of suspected toxicants, the Environmental Health Committee of the SAB writes: "Assessing the comoleteness and aualitv of the database for a specific agent is an important part of the risk assessment process" (p.6). - 13 -

A significant number of relevant, published studies reporting no significant associations between ETS exposures and nonsmoker disease were not considered by the authors of the ETS Review Draft, including: (i) Twenty-five disease; (25) studies on childhood respiratory (ii) Thirty-five (35) studies on ETS exposures in realistic settings; (iii) Relevant toxicological data from animal inhalation studies and short term tests for mutagenicity; and (iv) Published criticisms of various studies. Additionally, the ETS Review Draft fails to provide an exposure assessment, a necessary condition for a risk assessment, according to the EPA Guidelines. The fact that no actual exposure data exist for anv of the epidemiologic studies is not addressed in the Draft.5 Confounders in the association between lung cancer 5. The SAB's ad hoc panel on dioxin (November 1989) criticized a~ Review Draft on Dioxin for just those oversights. The ad hoc ~ Panel urged the authors of the Draft to "distinguish between ~ excellent versus poor experimental data" (p.4). The panel CJ recommended that "the document distinguish between excellent CA versus poor experimental work and thorough studies versus(b (continued...) Q~ ~ ~ - 14 - ~

and ETS exposures were also disregarded. Confounders in the association between childhood respiratory disease and ETS exposures were only partially considered. The ETS Review Draft employed a method of aggregating study results (meta-analysis) which did not adjust for statistical significance, study design or study strength. Significantly, and of concern to the public, the EPA provided a quantitative risk for lung cancer which presumed causation. The ETS Review Draft relies exclusively on one kind of evidence, epidemiologic, to the exclusion of other disciplines; even so, of the 24 epidemiologic studies on spousal considered, 19 were = statistically significant. smoking The ETS Review Draft also ignores the guidance set forth in the Regulatory Program of the United States Government (April 1, 1990 - March 31, 1991) ("Regulatory Program") which provides in part: Current practice gives undue weight to studies that show positive relationships. Resulting risk classifications are thus conservatively 5.(...continued) cursory examination. Not all data should be given the same weight" (p.13). The SAB Report also notes that most of the epidemiological studies evaluated for dioxin "do not provide definitive data" and should be classified as "inconclusive, due in most cases to design limitations such as inadequate power and inadequate exposure assessment" (p.10). The Report concludes that "without good exposure data, the epidemiologic studies are meaningless." - 15 -

biased estimates derived from samples of similarly biased observations. Perhaps the most important current need in regulatory decision making is for carefully prepared and scientifically credible estimates of the likely risks involved. Decision makers at all levels need unbiased and impartial risk information so they can focus their attention on significant problems and avoid being distracted by minutiae. For policymakers and the public to continue to rely on risk assessment in the development of regulatory initiatives, a renewed effort must be made to separate science from policy and provide risk information that is both meaningful and reliable. (p.25) The Regulatory Program clearly states: Unfortunately, epidemiological research suffers from its own set of limitations. For example, retrospective studies often have difficulty correlating morbidity and mortality with exposure to specific substances. Exposure data are commonly lacking, incomplete, imprecise, or affected by systematic recall or selection biases. Furthermore, the risks these studies seek to detect are often very small relative to background, thus making statistically significant effects difficult to observe. When health effects are latent, correlating exposures to illness is even harder. Similarly, the Regulatory Program states: [S]esides these unavoidable methodological limitations, epidemiological studies often suffer from outright bias. Many studies employ scientifically questionable procedures aimed at demonstrative positive relationships between specific substances and human illness. Some researchers used inappropriate statistical - 16 -

r procedures to 'mine' existing databases in search of associations . . .6 No consideration of any of these aspects is apparent in the ETS Review Draft. A summary of scientific comments submitted to the public docket was subsequently drafted by the author of the ETS Review Draft. That summary clearly reveals many of the Review Draft's major weaknesses and deficiencies (see Attachment II). The EPA's simultaneous publication of the Policy Guide Draft and ETS Review Draft is clearly in violation of the principle of separation of risk assessment from risk management. While the Review Draft with its assumption that ETS is a carcinogen is in public review, the Policy Guide Draft has already accepted that assumption and sets out to provide guidelines for "implementing" smoking restrictions and for "strategies" to reduce exposure to smoking. (Policy Guide Draft, p. 2) Thus, the "conceptual distinction" of risk assessment and risk management is inextricably blurred in EPA's dealing with the subject of ETS. The ETS Review Draft is inconsistent with the contemporaneous EMF Review Draft released by the EPA in June 1990. The EMF Draft examined AQ epidemiologic studies and a_ll the available animal studies and short-term tests on the subject. It ~ 6. For a further analysis of the implications of science or O regulatory decision making, S.ee Countina on Science at EPA, N W 1990, and August 10 Science ACS Plenary Focuses on Risk , , Assessment, C&EN, September 3, 1990 (Attachment 3). ClI CO 01 - 17 - kQb W N

reported that the relative risks from the various studies averaged 3.0; the authors concluded that those reported risks were not compelling enough to establish EMF as a GrouR B carcinogen. The ETS Review Draft, on the other hand, looked at only 2A epidemiologic studies which reported half the increase in risk (1.41) noted in the studies on EMF, and also failed to consider pertinent animal and exposure data. The Draft nevertheless concluded that ETS J& a Grouo A ("known human") carcinogen! The unwillingness or inability of EPA to uniformly and objectively follow the EPA Guidelines is dramatically illustrated by a visual comparison of the three draft reports issued by the EPA in 1990, within 90 days of one another. Table 1 which follows is a brief comparison of the EMF and ETS reports as to adherence to the EPA's Guidelines. Table 2 is a comparison of the ETS and Diesel reports. These graphic comparisons clearly confirm that the EPA is not uniformly following the EPA Guidelines and is at best conducting regulation by random procedure. - 18 -

TABLE 1 Diesel Emissions 1. Exposure Measurement "A major difficulty with the epidemiological studies regarding exposure to diesel engine emissions was the meastuement of the actual diesel exhaust exposure. Most studies either compared persons in job categories with pnesumably some exposure to diesel exhaust to either standard populations (presumably no exposure to diesel exhaust) or with men in other job categories from industries with little or no potential for diesel exhaust exposure. A few studies have included measurements of diesel fumes, but there is no standard method for the measuremertt." 2. Strength and Consistency of Association . Animal Studies 4. EPA Draft Recommendadons TABLE 2 "Evidence for the potential carcinogenicity of diesel exhaust in humans is limited; however, a few recent studies have indicated a small but significant increased risk of hmF cancer in occupational exposed workers.' A statistically significant risk of 2.6 was reported for miners and heavy equipment operators. "Studies employing rats and an adequate experi- mental design were nearly allpositive in demon- strating diesel exhaust-inducxd increases in tumorigenicity: ' Group 111: Probable human carcinogen EMF (Electromagnetic Fields) 1. Strength of Association 'The association between cancer occurrence and exposure to either EFL or RF fields is not strong enough to constitute a proven causal relation- ship, largely because the relative risks In the publtshed reports have seldom exceeded 3.0 in both childhood residential exposure and in occupational situations:' 2 Consistency of Association and "The consistently repeated pattern of lymphoma, Statistical Significance 3. Confounding Factors 4. Animal Studies 5. EPA Draft Recommendation leukemia, nervous system cancer and lympheuna in childhood studies and the tuling out of several confounding exposure factors in the Savim et al. (1988) study argue in favor of a causal link between these tumor types in children and exposure to ELF magnetic or electric fields. However, the fact that the odds ratios are small and In many cases not statistically significant Indicates that the association may not be strong and therefore argues against a causal relatlon- Ship " "No other agents (confounding factors) have been identified to explain this association." "Both animal and in vitro studies are needed to discover the relevant exposure factors and their interaction and to gain some understanding of the mechanisms of action." No classification ETS (Environmental Tobacco Smoke) These studies compare individuals with higher ETS exposures to those with lower exposures. All of them have made observations on never-smoking married women. Those married to a smoker are assumed to be at higher exposures than those not married to a smoker. These two groups are referred to as "exposed" and "unexposed." respectively, following the established terminology. 18 out of 23 studies are not statistically significant. The EPA calculated the relative risk for ETS (via mets-analysis) to be 1.28. Animal studies attempting to elicit lung cancer as a result of ETS exposure have failed to demonstrate any significant increase in cancer. Group A Known human carcinogen ETS (Environmental Tobacco Smoke) All studies used in the meta analysis calculation have relative risks less than 3.0. The EPA calculated the relative risk for ETS (via meta- analysis) to be 128. 18 of 23 studies are not statistically significant. A number of independent risk factors (confounding factors) for lung cancer have been identified (occupation, cooking oils & diet).1'hese factors were not addressed in the Draft. Animal studies attempting to elicit lung cancer as a result of ETS exposure have failed. Group A: Known Human Carcinogen 19-20