Tobacco Documents Online

Comments on the Review Draft released by the Environmental Protection Agency entitled "Health Effects of Passive Smoking: Assessment of Lung Cancer in Adults and Respiratory Disorders in Children" George Feuer, Ph.D., C.Med.Sc. Departments of Clinical Biochemistry and Pharmacology University of Toronto, Toronto M5G 1L5, Canada

Since 1963 I have been working in various fields of toxicology. Between 1963 and 1968 I carried out toxicity testing on various food additives at the British Industrial Biological Research Association, Carshalton, England. Since 1968, I have been studying the effects of various chemicals and drugs on hepatic function and structure at the University of Toronto. Since 1974, I also have been studying the mechanism of chemical carcinogensis. These investigations have centered on the basis of chemically induced hepatocarcinoma and mammary carcinoma. In addition, I have been focusing since 1984 on the oncostatic role of the pineal gland in the development of malignant melanoma and breast cancer. I have a wide variety of teaching commitments, including the University of Toronto, Canadian Memorial Chiropractic College and Ontario College of Naturopathic Medicine, where I teach toxicology and pathology., including among many topics aspects of cigarette smoking and its effects on the lung and cardiovascular system. I have published more than 300 research papers and reviews, and I am the author and co-author or five books. My curriculum vitae is attached. Recently a Review Draft document was released by the United States Environmental Protection Agency on the "Health Effects of Passive Smoking: Assessment of Lung Cancer in Adults and Respiratory Disorders in Children" (1). Based in part on 1986 reports by the National Research Council (2) and - 2 -

the U.S. Surgeon General (3) as well as on subsequent epidemiologic publications, the draft report concludes that environmental tobacco smoke (ETS) causes or is associated with (a) lung cancer in adults and (b) respiratory disease and pulmonary effects in children. Specifically, the report states that "passive smoking is causally associated with lung cancer in adults and that exposure of young children to ETS from parental smoking, particularly during infancy, is associated with increased prevalence of acute lower- respiratory tract infections (bronchitis and pneumonia), respiratory symptoms of irritation (cough, sputum, wheeze), and middle ear effusions (a sign of chronic middle ear disease)" (page 1.1). Before reviewing the EPA draft report, I had already been familiar with the publications that are the bases of the EPA document in relation to ETS and lung cancer and had reached the conclusion that the data in those papers suffer from numerous inconsistencies, misreporting biases and misclassifications. In my view, the pertinent studies are both-too contradictory and too preliminary to permit any reliable conclusions to be drawn from them (4). In the present discussion, which I have undertaken at the request of The Tobacco Institute of the United States, I provide a critical analysis of some points of the draft report that further confirms my assessment that a causal relationship between ETS and lung cancer has not been 3

established. This assessment is based on (a) the lack of biological plausibility, (b) the controversial nature of the relevant publications, (c) heredity and genetic factors, (d) the presence of ETS components in other substances in our environment and (e) the presence of numerous respiratory toxicants and carcinogens in our environment and their possible health effects. (a) Biological Plausibility. This topic was addressed in the EPA document as follows: "Based on these analyses and following the U.S. EPA guidelines for carcinogen risk assessment . . ., EPA concludes that environmental tobacco smoke is a Group A (known human) carcinogen. This conclusion is based on a total weight of evidence, principally: Biological plausibility. ETS is taken up by the lungs and distributed throughout the body. The similarity of carcinogens identified in SS (sidestream smoke) and MS (mainstream smoke) along with the established causal relationship between lung cancer and smoking make it reasonable to suspect that ETS is also a lung carcinogen." (page 1.3). Uptake and distribution of ETS in the lungs, however, do not constitute evidence of a causal relationship with lung cancer. Although the lung is the entry site of ETS, ETS-containing blood also circulates in other organs. There is no credible evidence, however, that ETS is responsible for the development of cancer in any other tissues. There are several chemical carcinogens that do not act at the site of entry. Many are activated by hepatic - 4 -

metabolism (5); for instance, 7, 12-dimethylbenz(a)anthracene (DMBA) is a chemical causing mammary adenocarcinoma in rats (6). DMBA (a procarcinogen) is converted in the liver to metabolite(s) that act on ductal areas of the breast tissue resulting in cancer. DMBA metabolites neither produce cancer in the liver, which is the entry site of DMBA and major site of metabolism, nor in any other organs, other than the breast. However, inhibition of DMBA metabolism results in a loss of carcinogenicity (7). Furthermore, using another example, such as aflatoxin, which has been shown to cause liver cancer in man and experimental animals (8), the shortcomings of biological plausibility can be illustrated. Liver cancer is fairly common in certain African and Asian countries, and it probably is associated with aflatoxin contamination of food (9). This disease is rare in the United States; the incidence of hepatic cancer represents only 150-160 cases per year with an estimated aflatoxin involvement in 50-60 cases (10). It is very likely that many more people consume aflatoxin- contaminated food. Therefore, if biological plausibility is a reasonable hypothesis, one would expect a much higher occurrence of hepatic carcinoma in the U.S., which is not the case. Basically, biological plausibility is a hypothesis with limited value, as I pointed out in these examples (and many more examples can be cited). It is well-accepted by - 5 -

scientists that one piece of experience and evidence is worth more than one dozen hypotheses; therefore, we cannot argue that biological plausibility represents evidence for an ETS effect. (b) Controversial Publications: In the document an important point of the conclusion is as follows: "Consistency of response. The two completed cohort studies and sixteen of the 21 case-control studies observed a higher risk of lung cancer among the female never-smokers classified as exposed to ETS. Evaluation of the total study evidence from several perspectives leads to the conclusion that the observed association between ETS exposure and increased lung cancer occurrence is not attributable to chance" (page 1-4). The document, however, does not prove consistency of response and does not give solid and uniform evidence that ETS exposure is associated with lung cancer. Table 4-1 (pages 4-11 and 4-12) summarizes the epidemiological studies applied to calculate the overall risk. According to the data relevant to females from the 22 studies given, five, or 23%, provided a relative risk (odds ratio) < 1.0 (lack of any association between ETS and lung tumor); five, or 23%, are between 1.19 and 1.31 (very low significance); six, or 27%, are between 1.41 and 1.65 (moderate significance) and in six studies, or 27%, the relative risk is > 2.0 (representing relatively high significance) (11,12). According to the adjusted data in Table 4-2 (page 4-36), the relative risk excess from background values is not 6

different at all in six studies (29%), it ranges between 0.03 and 0.11 in five studies (24%), and in ten studies the relative risk excess is 0.2 (47%). Combining these two points, we see that almost one-third of the investigations resulted in a negative association between ETS and lung cancer. It seems, therefore, that the statement that there is a "consistency of response" (page 1-4) is not correct. Even if the overall summary of the relative risk for female non-smokers were statistically significant, in evaluating all these data we cannot ignore the negative results (28-29%). Thus a clear relationship between ETS and lung cancer has not yet been established. Very recently a Japanese group studied the relationship between indoor air pollution and lung cancer incidence among non-smoking women in a case-control study (13). This study represents a survey conducted jointly by a number of health facilities. Comparing non-smoking adult women where the husbands were smokers, the odds ratio was 0.94; when other members of the family in the household also were smokers, the odds ratio was raised to a statistically non-significant 1.45. The odds ratio for childhood exposure to ETS when the fathers were smokers was 0.60; when the mothers were smokers, it was a statistically non-significant 1.71; related to other smoking members of the family, it was a statistically non-significant 1.13. Age-adjusted odds ratios also were computed for the effects of using straw or wood 7

heating on the development of lung cancer. The odds ratio was 1.33 when these fuels were used at the age of 15 years and 1.90 when used from age 15 to 30 years, the latter showing statistical significance. This study, thus, resulted in no significant correlation between ETS exposure and lung cancer. However, indoor air pollution due to the use of wood or straw as fuel affected the odds ratio and lung cancer incidence did reach statistical significance. Considering all these publications, alternative explanatory variables should be considered for lung cancer incidence in non-smokers that might be responsible independently for the reported cancer risk. These might include genetics, hereditary factors, occupational exposure to respiratory irritants or carcinogens, possible exposure to carcinogens in the diet, various other dietary factors, differences in individual sensitivity, social and economic conditions or other specific lifestyle conditions. Further risks associated with certain occupations such as mining (14), working in petroleum refineries (15), or in the shipbuilding industry (16,17) are well established and must be considered. (c) Heredity (i) Genetic Factors The probability that heredity, environmental and common life-style factors are associated with lung cancer has been suggested for quite a while (18). Recently, using new - 8 -

methodology, biochemical techniques for risk factors and statistical analysis for the assessment of single gene actions, sha-red environment and lifestyle associations (19,20,21), some evidence has been published for the relationship between non-ETS factors and the development of pulmonary cancer. These papers provide evidence that interactions between genes, environment and life style are causally connected with the etiology of lung cancer, although the exact mechanism has not yet been verified. It was suggested a long time ago that individuals differ in their susceptibility to the effects of smoking (22,23), and the health responses to various chemicals are determined by the level of specific enzymes as related to biochemical genetics (24). The first epidemiological evidence was provided in 1961 for increased familial risk factors associated with lung cancer mortality (18). Similarly, familial aggregations_have been made for respiratory tract cancer (25) and breast and ovarian carcinoma (26). Ooi et al. (27) and Sellers et al. (28) reported an enhanced familial risk for lung cancer among the relatives of lung cancer probands after allowance was made for cigarette smoking, occupation, sex and the variable onset of the disease. These investigations have been expanded by Sellers et al. (21) using segregation analysis. For these families evidence was found for a Mendelian inheritance in the pathogenesis of lung cancer. The results of these analyses - 9 -

indicated that the data are consistent with the involvement of Mendelian codominant inheritance of a major autosomal gene that produces the onset of lung cancer at an earlier age. (ii) Biochemical Studies Differences in individual susceptibility to lung cancer (29,30) can be explained by variations in the ability to metabolize foreign compounds (31). This possibility is supported by epidemiological studies that revealed enhanced familial risks (27), family aggregations (18), and the incidence of increased alveolar carcinoma in twins (32). Cystogenetic molecular chromosomal investigations (33,34,35,36) and DNA sequence studies (37) also indicate a genetic disposition in the pathogenesis of lung cancer (38). It is well known that the metabolism of many foreign compounds, particularly drugs, is under genetic control. These include several drug oxidation reactions, hydrolysis of succinylcholine, isoniazide acetylation, spartein and debrisoquine hydroxylation (39,40,41,42,43,44). Especially well studied is the metabolism of debrisoquine, which is controlled by autosomal genes (45,46). It has been established that the ability to metabolize this drug extensively is associated with a higher host susceptibility to develop bronchial carcinoma (45,47,48). Intermediate and poor metabolizers elicit a lower risk for lung cancer. Some studies reported conflicting results (49,50) and have been criticized for the epidemiological methods used (51). - 10 -