Tobacco Documents Online

-'STATEMENT Irving I Kessler, M.D., Dr. P.H. Department of Epidemiology & Preventive Medicine University of.Maryland School.of Medicine- Baltimore, Maryland I am a physician and epidemioloclist who has been active in cancer research for many years. My academic affiliations are with the University of M<<ryland School of Medicine, where I have been Professor of F:pidemiology and Preventive Medicine since 1978 and Chaifm<<n of the Department until 1988. Previously, I served as Professor of Epidemiology at the Johns Hopkins University. My profe!ssional expertise is in research on disease of humans, health regulation, medical administration and clinical education. My curriculum vitae is attached. I have been asked to comment on "Environmental Tobacco Smoke and Cancer," Chapter 4 in "F:nvironmental Tobacco Smoke: A Compendium of Technical Information" recently issued in draft form by the U.S. Environmental Protection Agency. Cancer in general, and lung cancer in particular, are relatively prevalent conditions which provoke concern among scientists, public health officials, health regulators, the corporate community and the general public. All share an earnest desire that the causes of these diseases be elucidated, that effective treatments or preventive measures be developed, and that individual risks be substantially reduced or eliminated. However, the achievement of these objectives can only follow the acquisition of the requisite

i etiological knowledge and the subsequent formulation of clinical and public policies taking the ne!w knowledge into account. Regulatory.actions-.initiated on any.basis other than that of the hard facts, however well intentioned, must inevitably fail. - -At a time like the present, when good health is widely regarded as a right rather than a privilege, activism by individuals or political bodies responding to public pressures is not uncommon. It is fair to say that activist efforts sometimes even succeed in changincf established government policies, typically diverting f'unds and public attention from one disease or condition to another of more current interest. AIDS, asbestos-related disease, DES, saccharine, the Dalkon Shield, and the lat:est neoplasm befalling a prominent public figure are among examples that may be cited. In recent years, environmental tobacco smoke has attained similar notoriety as a controversial issue raised to an intensely high level of public concern through activist pressure, media attention, and the subsequent reactions of governmental agencies and other public bodies. The EPA document is described in its preface as "intended to be useful for a diverse audience including: decision makers such as labor and management officials con-cerned with workplace exposures, public health officials and corporate•medical directors who are concerned with making health policy recommendations, educators, industrial hygienists and safety officers, ETS researchers, indoor air

pollution investigators and legislators who are considering legislation to restrict smoking in workplaces, restaurants and public access.buildings . . . . It is hoped that the technical information in this document, written by experts in the field, will provide information necessary to allow the public, government agencies, and the building industry, to make well-informed choices regarding exposure to ETS." Implicit in this statement is that the facts concerning the relationship between envircnmental tobacco smoke and cancer have been fully established and that public policy can now be modified to take these facts into account. Accordingly, the statement is addressed tc scientists, educators, public health officials, corporations and legislators, whose cooperation would be required to implement changes in health policy regarding environmental tobacco smoke and cancer. In reality, definitive or even reasonably convincing evidence on the biological relationship between passive exposure to environmental tobacco smoke ard cancer has not been assembled. There are no conclusive facts for the biomedical scientist to consider, for the public health official to regulate, for Corporate America to acknowledge or for the general public to understand. In no way can evidence of the quality presently available enable interested parties "to make we11-;informed choices regarding exposure to ETS". In fact, the author of Chapter 4 freely admits to the limitations of the evidence, including-substantial deficiencies in-most of

the published studies. The reader is therefore quite unprepared for the author's ultimate agreement with the conclusions of the three consensus confere!nces of.the World Health Organization, the National Research Council, and the Office of the U.S. Surgeon General, conclusions which ignore the scientific evidence in a rush to consensus. The general public is, by and large, ignorant of the methods of epidemiologic study, particularly the manner in which these differ from the experiments of basic medical scientists. In the laboratory, the reactions of biologically identical animal clones exposed and unexposed to a test - treatment are compared. The results serve as the basis on which to accept or reject the scientific hypothesis in question. By way of contrast, the epidem'.ologist studies heterogeneous human subjects, facing almost insurmountable ethical and judgmental problems in exposing them to the test treatment and other essentially unavoidab:.e methodological difficulties: At the completion of the si:udy, the laboratory scientist can attribute the experimental outcome to the test treatment without undue concern for methodological bias or confounding by the subject's gender, prior disease history, behavior, diet, environment, methods of selection, motivation or extent of cooperation. On the other hand, the epidemiologist is usually left with these problems incompletely resolved and thus confronts .influential roadblocks to a full understanding of the biological significance of the investigation. The basic problem.-of the

laboratory scientist is generalizing the-.experimental...- conclusions from the test animals to man. The fundamental prQblem of the epidemiologist -is deriving:-scientifically."- justified conclusions on the biological phenomena that are being studied. The "significance" alluded to throughout the chapter is statistical rather than biological, anc the critical distinction between the two should not be ignored. It can be affected by the number of cases studied ar.d the appropriateness of the statistical test us.ed as well as by the underlying biological reality under investigation. Most of the cited studies are deficient in respect to the - representativeness of the cases and controls utilized. None has resolved the basic problem of defininci ETS and measuring it quantitatively in human subjects who m<<y be exposed at home, at work, and at leisure throughout t:heir lifetime. Most fail to show a dose-response relationship between ETS and cancer, and odds'ratios are not often greater than 2.0, i.e., well within the range of biological insignificance as understood by experienced epidemiologists, Some of the studies do not take into account the histopathological distinctions of lung neoplasms, each of which may be subject to a specific etiology. Response rates anong the human subjects, an important desideratum for evaluating all epidemiological studies, are poorly documented in a number of the studies.

L, The author's discussion.of the epidemiologic literature on environmental tobacco smoke and lung cancer is reasonable and objective. He is forthright in pointing-out many deficiencies and limitations in the existing studies. His conclusion that there is a proven eticlogical relationship between lung cancer and environmental tobacco smoke therefore follows almost as a non sequitur. For example, a large paragraph is devoted to the study of Hirayama and its limitations which have been widely discussed, debated and publicized. An objective reader could not reasonably conclude that such a study contributes substantialiy to the positive evidence on the relationship between environmental tobacco smoke and lung cancer. The same applies to all of the other published studies which are flawed to one extent or another. By no means are these comments intended as, an indictment of Hirayama's study or those of the other investigators cited in this chapter. Epidemiological studies are notoriously susceptible to methodological difficulties because of their reliance on the voluntary participation of' human subjects. One must retain a modicum of understandincl for these difficulties while, at the same time, insisting upon objectivity with respect to any causal inferences that are adduced from the studies. None of the investigations cited in the chapter are capable of serving as a sulid scientific foundation fof the conclusion.that environmental tobacco smoke ik' O'f` ~ .l is etiologically related to lung cancer. If this,is true of the individual. studies, it is also--true -of: all:, of =the studies W CD

considered jointly, notwithstanding the claims of the meta-analysis. It has been noted that the.author relies substantially on the conclusions of panels commissioned by three public bodies (the World Health Organization, the National Research Council, and the U.S. Surgeon General's Office ) to buttress his evidentiary argument. The conferences which generated these reports were consensual and quasi-political rather than unequivocally scientific. This brings to mind an earlier conference sponsored by a worldwide health organization on the prevention and control of cervical cancer. For reasons having nothing to do with the scientific issue at hand, the panel members were informed that exfoliative cystological screening for ceivical cancer prevention could not be recommended in the final document. Accordingly, the members deliberated for three days on techniques for controlling or preventing cervical cancer without being able to discuss the technique which has been proven the most effective, viz. the Pap Smear. Those in attendance at the three ETS conferences understood the need to arrive at consensual recommendations aimed primarily at public health action (i.e. disease prevention) rather than at scientific elucidation of the etiological agents. The author's comment that "consensus among the thre'e groups in spite of differ'_ng-methodology, stresses the determination that involuntary smoking causes lung cancer'.' is unconvincing if only because=the panels-did

i not organize themselves primarily as scientists seeking. truths, but rather as synthesizers of existing information and supporters of public health measures in the absence of definitive scientific truths. The World Health Organization "reached its conclusion. . . largely on the basis of b'_ological plausibility". The National Research Council "based its views on a pooled analysis of the epidemiological data adjusted for . bias." The U.S. Surgeon General's conclusion "was based on the extensive information already availab:=e". These bodies were limited in their deliberations by the studies available to them, with all of their inherent diver:aity, variability, and inconsistency. None of the three consensus reports adds new information to the issue at hand. The author concedes that "the published epidemiological studies provide varying and imprecise measures of the risk, and that exposures to environmental tobacco smoke have not been characterized for large and representative population samples." Clearly risk assessments for involuntary smoking•and lung cancer are subject to substantial uncertainty. Scientists are-comfortable working in uncertainty, and their studies are designed to reduce the degree of uncertainty. On the other hand, decision-making public bodies are unaccustomed to working with uncertainty which, by their nature, they attempt to minimize in order to achieve consensus on the policy and regulatory outcomes of their daliberations. It is in this context that one-may understand-tie-implications-of

the author's statement that-"while_the results-of these new- studies will provide needed information fcr scientific purposes, the available data-and the conclusions of-the scientific community already provide a conpelling rationale for reducing involuntary exposure to environmental tobacco smoke." In other words, while it is essentially conceded that the hard biological evidence is not at hand, it is deemed prudent at the level of the lay public to conclude that indirect tobacco smoke is a cause of lung cancer. Many scientists, especially those concerned with the scientific facts rather than with serving the needs cf public health activism, would insist that prudence (or Folitics) should not play a decisive role in biomedical research. The evidence on involuntary smoking and cancer at sites other than the lung is, by the authcr's own statement, extremely weak: "Associations of involuntary smoking with cancer at diverse sites cannot be readily supported with arguments for biological plausibility." in other words, except for the lung, the author concedes that substantial evidence on associations of passive smokir.g with neoplasms does not now exist.