Tobacco Documents Online

CHAPTER 4: ENVIRONMENTAL TOBACCO SMOKE AND CANCER fl CARR J. SMITH, PH.D. SENIOR SCIENTIST L

ENVIRONMENTAL TOBACCO SMOKE: A COMPENDIUM OF TECHNICAL INFORMATION Section/Page/Line Specific Comments ETS and Cancer/Paae 41/Line 14 The phrase 'in the general population" needs to be clarified.- There are significant differences in histologic distribution between males and females (1)- ETS and Cancer/Page 41/Line 17 The sentence 'Bronchioloalveolar cell carcinoma represents about 5 percent of the remaining lung cancers' is incorrect. Five percent of ten percent is 0.5%. This cancer represents from 1.1 to 9.0% of lung cancers (2). ETS and Cancer/Page 41/Line 24 - The statement, 'Most of the early evidence indicated that tobacco smoke was a potent respiratory carcinogen" deserves comment. Toba,:co smoke has not been shown to induce lung cancer in experimental animals. 0D OQ ~ ~ N ~ 1 p

ENVIRONMENTAL TOBACCO SMOKE: A COMPENDItM OF TECHNICAL INFORMAT[ON Section/Page/Line Specific Conwents ETS and Cancer/Paae 41/Line 41-42 The statement "Agents other than tobacco smoke may also cause lung cancer, and cases occur in lifelong nonsmokers" implies that lung cancer is unusual in nonsmokers. M. A. Schneiderman and D. L. Davis, National Academy of Sciences, and D. K. Wagener, National Center for Health Statistics (3) recently made the following comment: "If these estimates are accurate, only colon and prostate cancers in men (mortality rates, 21.4 and 23.5 per 100,000, respectively) and colon and breast cancers in women (15.7 and 27.3 per 100,000, re;:pectively) had higher mortality rates than did non-smoking-attributable lung cancer for both blacks and whites.` Further, blacks might be at higher risk than whites for non-smoking-_ attributable lung cancer, the computed residual "rate" being 67% higher in black men than that in white men and 16% higher in black women than that in white women. As computed herein, non-smoking-attributable lung cancer is one of the most common causes of cancer mortality.... " 2

ENYIRONMENTAL TOBACCO SMOKE: A COMPENDIUM OF TECHNICAL INFORMAT[ON Section/Page/Line Specific Comwents ETS and Cancer/Page 41LLine 48 Line 48 states that "many other occupational agents are suspect respiratory carcinogens'. The list of occupational exposures not mentioned that have been associated with increased lung cancer risk includes: Welding (4) Painting (5) Pesticides (6) Diesel Exhaust (7) Working in foundries (8-10) Ceramic and pottery workers (11-12) Rubber workers (13) Herbicides (14) t r. 0, 3 L.

ENVIRONNENTAL TOBACCO SNOKE: A CONPENDIUN OF TECHNICAL INFORMAT[ON ` Section/Page/Line Specific Coments ETS and Cancer/Pa4e 42/Line 2 The statement "a clear pattern of genetic susceptibility to lung cancer has not been demonstrated" is incorrect. Tokuhata and Lilienfeld (15,16) first observed the familial aggregation of lung cancer in 1963. In the absence of smoking, they 1'ound.the relative lung cancer mortality rate among first-degree relatives of lung cancer patients to be 2.7 times higher than that among relatives of controls. 0oi et al. (17) reported a 2.4-fold greater.risk among relatives of probands after controlling for the effects of age, sex, and environmental factors„ Samet et al. (18) have shown a 5.3-fold increase in adjusted risk in individuals with a parental history of lung cancer. A recent study by NcDuffie et al. (19) also demonstrated a two-fold increase in lung cancer risk fur relatives of probands in a Canadian population. Further evidence for a potential familial link was provided by Joishy-et al. (20) in 1977. They reported the occurrence of alveola'r cell carcinoma in a pair of male identical twins. The cancers in these twins presented with strikingly similar ages of onset, histopathologic features, and sites of metastasis. After adjustment for age, sex, race, and smoking, the first-degree relatives of lung-cancer patients and of patients with chronic obstructive pulmonary disease have been shown by Cohen et al. (21) to have significantly higher rates of impaired forced expiration than first-degree relatives of patients with non-pulmonary disease. Similarly, Guirgis, et al. (22) reported an increased frequency•of pulmonary diseases in the relatives of lung cancer patients. These findings suggest*that lung cancer and chronic obstructive pulmonary disease may share a common familiar component. There may be genetically determined metabolic differences between lung cancer patients and matched controls. Ayesh et al. (23) have reported differences in the metabolic pathway of debrisoquine 4-hydroxylation in a white English population. They demonstrated that lung cancer patients showed a preponderance of probable homozygous dominant extensive metabolizers (78.8%) with a few recessive poor metabolizers (1.6%) compared with matched controls (27.8% and 9.0% respectively). These authors concludE:d that the gene controlling debrisoquine 4-hydroxylation may be a genetic determinant of susceptibility to lung cancer. Caporaso et al. (24,25) have recently confirmed these observations in a group of white and black American patients. They have estimated that extensive metabolizers are al; 4.5 times the risk for developing lung cancer compared with poor metabolizern. 4

ENVIRONMENTAL TOBACCO SNOKE: A CONPENDItN OF TECHNICAL INFORMATION V Section/Page/Line Specific Co=ents ETS and Cancer/Page 42/Line 2 (Continued) Kellermann has reported that the inducibility of aryl hydrocarbon hydroxylase (AHH) in human lymphocytes is correlated with an increased risk of the occurrence of bronchogenic carcinoma (26). The use of lymphocyte AHH induction as an indicator of genetically controlled metabolic activation of chemical carcinogens in the lungs is based on the assumption that induction is similarly regulated in different tissues. Dosaka and colleagues (27) have studied the inducibility of sister chromatid exchanges (SCEs) by benzo(a)pyrene in cultured peripheral blood lymphocytes from 15 untreated lung cancer patients and 25 healthy persons including 11 high - and 14 low - cancer-risk individuals as classified by family history. After exposure to benzo(a)pyrene, the lymphocytes of lung cancer patients and high-risk individuals exhibited significantly increased SCE rates than those of low-risk persons. These authors have suggested that lymphocytes of high- risk individuals may be more susceptible to benzo(a)pyrene induced DNA damage and therefore to cancer. The spontaneous transformation of human skin fibroblasts from three lung cancer patients was observed by Azzarone et al. (28) in 1976. Rudiger et al. (29) have also reported spontaneously increased SCE r<<tes in the fibroblasts from lung cancer patients. Several studies have shown differences in HLA frequenc':es between lung cancer patients and the normal population. Ford, Newman, and Mackintosh (30) found a high relative risk of being HLA-BW22-positive and having lung cancer. Markman et al. (31) found individuals with the HLA-BW44 (HLA-Ba2) allele to be significantly over-represented in a patient population with small cell carcinoma of the lung compared to a control population (52% versus 25%, N.001). Mottironi and Banks (32) showed that AW33 had a significant difference in antigen frequency between a lung cancer population and controls. However, none of these correlations has been strong ernwgh to use HLA typing to screen patients for susceptibility to lung cancer. Heighway et al. (33) investigated the influence of polymorphic variants of the human c-Ha-ras gene on predisposition to lung cancer. They found an abnormal allele distribution of the c-Ha-ras gene in individuals with non-small cell carcinoma of the lung compared to both control and small cell carcinoma of the lung patients. These authors have suggested that this difference may represent a"degree of genetic predisposition" to non-small cell carcinoma of the lung. Another way to study the question of genetic predisposition to lung cancer 5

ENVIRONMENTAL TOBACCO SMOKE: A COMPEHDIUN OF TECHNICAL INFORMATION I U Section/Page/Line Specific Comments ETS and Cancer/Page 42/Line 2 (Continued) involves the identification and isolation of oncogenes i:apable of inducing malignant transformation. Yokota et al. (34) showed that the c-erbB-2 gene was amplified in 5 of 63 adenocarcinomas and in one of -the 38 other types of tumors studied. These authors have suggested that the arotein products of the amplified c-erbB-2 gene may play a role in the evolution of adenocarcinomas. Whang-Peng et al. (35) have reported a nonrandom chromosomal abnormality (a common deletion of the short arm of chromosome #3) in human small cell lung cancer. The expression of proteins encoded by ras oncogenes was examined by Tanaka et al. (36) in several types of lung cancer. PrDteins related to ras genes were detected in squamous cell carcinoma, small cell carcinoma, and adenocarcinoma. Individual differences in oncogene activation as a response to environmental insult may be an important determinant of susceptibility to lung cancer. The National Cancer Institute currently has a major research program on genetic typing and lung cancer. NCI Director Samuel Broder has formed a committee to bring this gene research to clinical practice as quickly as possible (37,38). In summary, the evidence for the existence of a genetic predisposition for the development of lung cancer comes from several sources: family histories, case studies, epidemiology, biochemical measurements and geretic studies. ETS and Cancer/Page 42/Line 6 The statement "Animal and human studies suggest that law consumption of Vitamin A or its precursor, beta-carotene, may also increase lung cancer risk" downplays a significant body of literature on the assoc:iation between diet and lung cancer risk. Zeigler (39) has recently written a literature review on the association between low levels of serum or plasma beta-carotene and lung cancer risk. Several of these dietary studies have reported relativc risks for lung cancer much larger than those reported in ETS studies. For example, Menkes et al. (40) reported that "A strong inverse association between serum beta-carotene and the risk of squamous-cell carcinoma of the lung wa:: observed (relative adds, 4.30: 95 percent confidence limits, 1.38 and 13.41)". Dietary factors other than beta-carotene levels have also been associated with lung cancer risk. Irv 1987, Wynder et al. (41) reportei that in an international comparison study "calories from dietary fat were highly 6 ~

ENYIRONMENTAL TOBACCO SMOKE: A CONPENDIIAN OF TECMiICAL INFORMATION iJ h i Section/Page/Line Specific Comments ETS and Cancer/Page 42/Line 6(Continued) significantly associated-(P<.001) with lung cancer mortality. This finding was obtained after accounting for disappearance data for tobacco (P<.001)..." In addition, Goodman et al. (42) have reported a"sign-ificant positive association of dietary cholesterol and the risk of lung cancer in men, but not in women". Therefore, dietary fat intake is a risk factor for lung cancer. ETS and Cancer/Page 42LLine 26-34 This paragraph states that epidemiologists "determine the causes of disease by studying populations" and "identify the causes of disease". Epidemiologists study "associations" of risk factors with diseases. They do not determine "causes". Sir Richard Doll (43) has commented on this inherent limitation of epidemiology, "Two methods have been commonly used - the retrospective, starting from affected patients and unaffected controls, and the prospective, starting from subjects with varying degrees of exposure to the agent under study. Both methods have their advantages and disadvantages; but neither can overcome completely the difficulty that any relationship that is found may be a secondary one and that the suspected agent may itself be associated with another which is, in fact, the true carcinogen." ETS and Cancer/Page 43/Line 22 This paragraph on the misdiagnosis of lung cancer omits the important work of McFarlane, Feinstein, and Wells (44-46). ETS and Cancer/Page 43-45LSection entitled "EDidemiolaaical Evidence on Involuntary Smokin4.and Lung Cancer" There are significant differences in lifestyle between smokers and nonsmokers. Smokers consume a diet significantly higher in saturated fat and lower in fruits and vegetables than nonsmokers (47-49). Smokers exercise significantly less than nonsmokers (50). They also sleep less than nonsmokers (51). There are also significant differences in lifestyle between the spouses of smokers and the spouses of nonsmokers. The following abstract is excerpted from Perusse et al.'s (52) paper entitled "Familial Aggregation in Physical 7

ENVIRONMENTAL TOBACCO SMOKE: A COMPENDIUM OF TECHNICAL INFORMATION :NO Section/Page/Line Specific Coo~nents ETS and Cancer/Page 43-45/Section entitled "Eaidemiological Evidence on Involuntary Smoking and Lung Cancer" (Continued) Fitness, Coronary Heart Disease Risk Factors, and Pulmmary Function Measurements": "In order to test for the presence of familiar aggregation in physical fitness and coronary heart disease risk factors, body fat, subnaximal power output, muscular strength, muscul-ar endurance, blood pressure, pulmonary functions, and several blood biochemical variables were measured in 304 nuclear families living in the Quebec city area. Analysis of variance indicated a larger between-family than within-family variation for all the variables. When all members of nuclear families were considered, intraclass correlations ranged from 0.21 to 0.34 (P < 0.01). Interclass correlations computed for various pairs of relatives revealed significant parent-child and sibling correlations for all the variables (0.14 <_ r<_ 0.55; P<_ 0.01). On the other hand, spousal correlations tended to be lower but significant (0.10 < r< 0.30; P< 0.05) for all variables except subcutaneous fat and hemoglobin concentration. These results suggest that heredity and common lifestyle shared by members of nuclear families are responsible for the familial aggregation of physical fitness, coronary heart disease risk factors, and pulnionary functions. The findings also support the notion of considering the nUclear family as a unit of intervention in the application of preventive measures aimed at the reduction of several risk factors." Sidney et al. (53) recently reported that the self-reported mean dietary intake of carotene is lower in nonsmokers exposed to ETS at home than in nonsmokers not exposed to ETS at home. These authors stated, "We know of no studies that have controlled for blood levels of caroi:enoids in the analysis of this relation...". In addition, Miyamoto et al. (!i4) has reported decreased levels of serum selenium and Vitamin E in the families of lung cancer patients. All of the studies described in this section and in the next section (ETS and cancer at other sites) suffer from the same limitation - the inability to control for the contribution of confounding variables to relative risk estimates. As previously described in this review, there are a large number of risk factors for lung cancer which exceed the National Research Council's 259'o risk increase estimate for exposure to ETS. The risk factors already discussed include occupational exposures, genetic predisposition, high saturated fat intake, and low intake of beta-carotene and other antioxidant vitamins. There are many other reported risk factors for lung cancer: tuberculosis, reproductive history in female adenocarcinoma patients, periodontal disease, and even owning pet birds. Lifestyle differences between 8

ENYIRONNENTAL TOBACCO SNOKE: A CONPENDIUM OF TECFWICAL INFORNATION Section/Page/Line Specific Comments ETS and Cancer/Page 43-45/Section entitled 'Epidemioloaical Evidence on'' Involuntary Smoking and Lung Cancer" (Continued) , smokers and nonsmokers are significant. Lifestyle differences between the spouses of smokers and the spouses of nonsmokers also appear to be significant. None of the studies associating exposure to ETS with cancer have controlled for these lifestyle differences. JJ I ii L ~ ~