Tobacco Documents Online

I I I I I 11 I I I I I I I I I I I I Seminars in Surgical Oncology 9:73-79 (1993) Recent Developments in the Epidemiology of Lung Cancer GEOFFREY C. KABAL Pho From the Department ol Eoidemioiogy and Social Medicine. Albert Einstein College or Medicine, Bronx. New York Lung cancer is currently the leading cause of cancer death in the United States and also the most common tumor worldwide. Changes in the distribution of histologic types over the past two decades in the United States, as well as high rates of lung cancer in certain suboopula- tions, require explanation. While cigarette smoking and specific occu- patibnai exposures are firmly established as important risk factors for lung cancer. recent work provides evidence that other factors may play a role either as indeoenaent risk factors or as modifiers of the effect of smoking. This paper reviews the eptdemiology of lung cancer. with an emphasis on developments in the past decade. ' 1993 wiiey-Uss. Inc. KEY WORDS: lung neoplasms. smoking, environmental tobacco smoke. radon, diet, host factors INTRODUCTION The Magnitude of the Problem The United States is currently in the sixth decade of an epidemic of lung cancer. Lunz cancer is the leading cause of cancer mortality in both sexes. accounting for an estimated 146.000 deaths in the United States in 1992. or 34'% of cancer deaths in males and in femaies (1]. The American Cancer Society projects that 168.000 new cases will be diacnosed in 1992. ac- countme for 18% of new cancers in males and 12% in females [1]. Between 1930 and 1987 the age-adjusted lung cancer mortality rate increased from 4 to 74 cases per 100,000 in males and from 4 to 27 in females [1]. During the past few years. lung cancer incidence rates in males have begun to level orT, but those in females have continued to rise. It is clear from these ngures that lung cancer will remain a major public health problem for decades to come. While lung cancer incidence in males is approaching its peak in the United States. rates in developing coun- tries are increasing, presaging a globalization of the epidemic [2]. Lung cancer is already the most common tumor worldwide [2]. Differences in age-adjusted incidence rates for spe- cific histologic types of lung cancer by sez, race. and calendar time period suggest that different histologic types may have different etiologies. In white men in the United States, rates of adenocarcinoma and oat cell carcinoma increased over the period 1969-1988, while the rate o f sa, uam ous cell carcinoma decreased. In white women, all major types showed an increase [3]. Squamous cell carcinoma is still the predominant histological type among males. whereas adenocarci- noma predominates among femaies [33. Among non- smokers. the proportton of adenocarcutoma is ereater than in smokers [41,, and parttcuiarly so in females. reaching 78% in one series of lung cancer cases in nonsmoking women [5]. While cigarette smoking and specific occupational exposures have been rirmiv established as important risk factors for lung cancer, over the past decade there has been increasine recoenition that smoking and oc- cupauonal exposures may not explain ail of the varia- tion in lung cancer incidence within countnes and be- tween countnes, and that other factors may play a role either as independent risk factors or as modifiers of the edect of smoking. Some issues that remain to be eluci- dated include ( t) the high rates of lung cancer in Chi- nese women, who have a low prevalence of smoking; (2) the higher incidence of lung cancer in black Amen- Address reorint mqtssu to Gcodrev C. Kabat. Ph.D.. Albert Ein- stein College of Medinne. Deoarrment of Eoidcmiolo¢yand Sodal Belfer Bldg. Rm. 1307- 1300 Moms Park Ave.. Bronx. Medicine. NY 10461-1602. ' 7 1993 Wiiey-Liss, Inc.

14 Kabat can males in the United States: (3) etiologic factors for after quitting smoking; (6) the correlation between adenocarcinoma other than smoking: and (4) risk fac- prevalence of smoking and lung cancer mortality rates tors for lune cancer in lifetime nonsmokers. Recently in successive birth cohorts of men and women in the attention has been drawn to a number of new poten- United States: and (7) the induction of tumors in ex- tial risk factors for lung cancer. including: passive perimental animals following exposure to tobacco smoking. domestic radon exposure. diet. body mass smoke. index. alcohol consumption. reproductive factors and Following Krevbere's ciassincation (15], for a long exposure to exogenous hormones. and host suscepti- time it was generally accepted that smoking was only biiity. In addition. changes over the past 3-4 decades associated with squamous and oat cell carcinomas and in the type of cigarettes smoked in the United States not with adenocarcinoma. However. studies carried have been adduced as a possible explanation of the out in large case-control series indicate that. although observed increase in adenocarcinoma of the lung [6,7]. the magnitude of the association with smoking is This review does not aim to be exhaustive but rather smaller in the case of adenocarcinoma. nevertheless a to provide an overview of the epidemiology of lung dose-response relationship exists for this cell type as cancer with an emphasis on developments over the well [16,17]. A recent analysis of 87 cases of the rare past decade. Several topics. such as occupational ex- bronchioloalveolar carcinoma has shown a consistent posures and air pollution. are given less space than association with smoking [18]. Large cell carcinoma their importance warrants. and the reader is referred ais6 appears to beasso.ciated with smoking [17]. to comorehenstve reviews. Studies examining the use of filter versus nontilter Although most of the discussion below deals with cigarettes and cigarettes of reduced tarr nicotine yieids specinc exposures as independent risk factors. interac- generally indicate that there is a modest reduction in tions between various factors isuch as smoking. occu- the odds ratio for lung cancer associated with smoking pation. and diet) may be important in determining an these "less haiardous cigarettes." on the order of 20- individual's risk of lung cancer [8]. Classical examples 30% [l9]. In an effort to explain the higher lung cancer of interaction between risk factors for fun¢ cancer are incidence rate in black American males compared to the greatly enhanced effect of exposure to radon as whites. recent work has focused on differences in well as to asbestos in smokers compared to nonsmok- smoking patterns between blacks and whites (20] and ers [9,10]. . on the effect of mentholated versus nonmentholated cigarettes on lung cancer risk (21]. The latter study SMOKING showed no increase in the odds ratio for lung cancer in Since the publication of the first epidemioio¢ic stud- smokers of mentholated relative to smokers of non- ies linking cigarette smoking with lung cancer in 1950. mentholated cigarettes. the association has been connrmed in epidemioioeic Various estimates are available for the proportion studies carned out in manv countnes and has been of lung cancers titat are due to smoking-that is. the further buttressed by animal evidence of the car- proportion of lung cancers that would be eliminated if cmoeenicitv of tobacco smoke ( I f-13]. Tobacco is the smoking were totally eliminated. Estimates of the pro- most exhaustively studied human carcinogen. and the portion of lung cancer attributable to cigarette smok- evidence for a causal association is overwheimine. In ing in various develooed countries range from 83% to fact, one could say that the association of cigarette 94% in males and from 57% to 80% in females (12]. smoking with lunz cancer provides a model for the Smoking cessation among current smokers and pre- associauon of an environmental risk factor with a vention of smoking initiation starting in school-aee chronic disease. This model is explicitly formulated in children offer the best prospects for reducing the inci- Sir Bradford Hill's criteria for judging the causality of dence of lung cancer. Between 1965 and 1987, the an association [14]. Evidence for a causal association proportion of current smokers in the United States of smoking with lung cancer includes (1) the rarity of declined from 50?% to 31.7% in men and from 31.9% lung cancer in lifetime nonsmokers: (2) the large ma2- to 26.8% in women [11-22]. However, the reduction in nitude of the association. generally a 10-fold increased smoking prevaience has been greatest among the more risk for current smokers relative to never-smokers: (3) educated. particularly in men. and as a result smoking a dose-response relationship between amount is becoming increasingly a habit associated with lower smoked and the relative risk of lung cancer. which can socioeconomic status. Extensive health promotion re- exceed 40-fold in heavv smokers: (4) the fact that the search has focused on designing effective strategies to relative risk increases with duration of smoking and help smokers autt and to deveiop the "life skills" and earlier age of starting smoking; (5) the progressive self-esteem that enable children and young adults to reduction in the relative risk with increasing years resist taking up smoking [23]. 2081783280 I I I I I i 'I I ~ I I

I I ' I I I ! I I I I Z I I Epidemiology of Lung Cancer 75 OCCUPATION had a significantly increased risk of lung cancer rela- Studies of occupational groups have shown that tive to the nonsmoking wives of nonsmokin¢ hus- occupational exposure to a number of agents is as- bands [30.31]. Since then over 30 studies of the issue of sociated with lune cancer. These include asbestos. environmental tobacco smoke (ETS) and lung cancer radon. poiycyclic aromatic hydrocarbons. mustard have appeared. A metanalysis carried out by the Na- gas, chloromethyl ethers, chromium. nickei, inorganic tional Research Council [32] in 1986 found that the arsenic. and vinyl chloride [24,25]. Other agents sus- summary relative risk for lung cancer of nonsmoking pected of being lung carcinogens are: acrylonitrile- omen whose husbands smoked, relative to nonsmok- exposures encountered by rubber workers. beryflium. ing women whose husbands were nonsmokers, was ferric oxide dust- lead- and cadmium (241. 134 (95% confidence interval 1.18-1.53). The identification of synergism between occupa- Epidemioiogic studies of passive smoking are con- tional exposure (radon. asbestos) and cigarette smok- fronted by a number of challenges: the rarity of lune ing (9,10] has important implications for both an un- cancer occurring in never-smokers: the greater dilu- derstanding of the biological mechanisms of lung tion of ETS compared to smoke inhaled by the active carcino¢enesis and for prevention. smoker, the difficulty of obtaining accurate exposure The proportion of cancer attributable to occupa- information by means of self-reports: the lack of a tional exposures has been a subiect of controversy biological marker for long-term exnosure: misclassiii- [26], and the limited availability oi accurate exoosure cation of smokers as nonsmokers: and the possibility of coniounding by other risk factors. including diet or. data on ootennally exposed workers makes any esn- mate of this proportion highly uncertain. Within these in places like China, exposure to cooking fumes [33]. limitations. an educated estimate of the proportion oi There is inconsistency among the existing studies as to lun¢ cancers attributable to occupational exnosure is the presenceiabsence of an association; the histologic l5% of male and 5% of female lun¢ cancer cases [271. type of lune cancer for which an association is ob- served: the sex in which an association is observed: AIR POLLUTION and the timing of exposure (i.e.. childhood versus adulthood) [34,35]. It has lon¢ been suspected that exposure to environ- In spite of these problems. given the chemical com- mental (as opposed to occupational) air pollution position of ETS and what is known about the effects might contribute to excess lung cancer incidence. of active smoking, it is biologically plausible that However, studies of air pollution and lung cancer arc heavy ETS exposure over long pertods, and perhaps complicated by the fact that air pollution is a compiex particularly in those exposed in childhood. can in- mixture that varies from place to place and over time: crease the risk of lung cancer. The largest study to date by the overwhelming effect of cigarette smoking; and ot lun¢ cancer in nonsmoking women indicates that bv movement of subjects both within and between women whose husbands smoked had an increased rel- different cities. The so-called "urban factor." !hat is. ative risk for lun_g cancer (odds ratio = 1.3; 95%con- the 1.5- to 2.0-fold greater lung cancer incidence in fidence interval 1.0-1.7) and for adenocarcinoma cities compared to rural areas. can be largely explained lodds ratio = 1.5: 95% confidence interval 1.I-2.0) by cigarette smoking and occupational exposure. Nev- after adjustment for socioeconomic variables [4]. A ertheless, studies of populations exposed to point siznincant trend in the odds ratio for adenocarcinoma sources of pollution, such as nonferrous smelters, sug- was seen with increasin¢ number of pack-years of the gest that even after adjtutment for smoking and occu- husband's smoking. In the highest exposure group pation. exposure to high levels of air poilution is as- (> 80 pack-yearsl the odds ratio was 1.7 (95% confi- sociated with increased lung cancer: and analytical dence interva10.8-3.5). Other ETS exposures in adult- studies exammine lung cancer risk by urbani rural res- hood(in the household. on the job. and in social set- idence indicate that. in both smokers and nonsmokers. tings) were also associated with increased risk of lung urban residence is associated with increased lung can- cancer. but exposure in childhood was not. cer risk [28,291. Tntu. while the overall contribution of air pollution is difficult to gauge, exposure to polluted RESIDENTIAL EXPOSURE TO RADON air is likely to account for a modest percentage of lung Based on studies of underground miners. it is firmly cancer incidence (27,291. established that exposure to relatively high levels of radon and its proeeny can cause lung cancer in hu- ENVIRONMENTAL TOBACCO SMOKE mans [363. Over the past decade. the detection of In 1981. two reports were published purporting to radon and its progeny in homes has led to public show that the nonsmokine wives of smoking husbands concern that exposure to lower levels of radon typical 2081783281

76 Kabat of homes could pose a lung cancer hazard. Conditions in homes differ from those in mines, and it is possible that long-term exposure to the relatively low radon lcvels typical of dwellings may pose a greater hazard than expected based on linear extrapolation from the levels tvpical of mines [37]. In the absence of system- atic surveys of domestic radon levels in the United States. estimates of the number of persons with high exposure (defined by the U.S.E.P.A. as 4 pCi/ 1) can- not be made [37]. Epidemiological studies of residential radon expo- sure in relation to lung cancer have yielded mixeS results. Studies carried out among women in New Jer- sey [38] and China [39] detected little or no effect of domestic radon exposure, while a study from Sweden [40] provided evidence of an association as well as of an interaction between radon exnosure and cigarette smoking. These :inconsistenaes may be due to me- thodoloeic problems including: subiect mobilitv, er- rors in estimating exposure, and inadequate sample size [41]. A number of epidemioiogtc studies of domes- tic radon exposure are in progress. and it is hoped that these will help clarify some of the crucial issues. in- cluding (1) the level of radon exposure at which an increase in risk is observed. and (2) the nature of the interaction between radon exnosure and cigarette smoking (is it additive or multiplicative?). RADIATION In addition to excess lung cancer rates observed in miners exposed to alpha-radiation from radon and its oroeenv-increased risks oflune cancer have been re- ported in patients treated wnh radiation for ankylos- ing spondvlitis in the United Kingdom and in those exposed to radiation from the atomic bombs dropped on Japan [24]. DIET Epidemiologic studies of diet and lung cancer show, according to a recent, authoritative review. "a consist- ent substantial protective effect of dietary vitamin A intake from vegetable sources" [421. Some studies have found a stronger protective effect of all vegeta- bles, dark green vegetables. cruciferous vegetables. and tomatoes, than for beta-carotene specifically [43. 44], suggesting that other vegetable constituents, in- cluding other carotenoids (lutein. lycopenc) and in- doles. may be protective against lung cancer. The most widely accented mechanism underiying the apparent protective effect of vegetable consumption is the role of anti-oxidants, including beta-carotene- vitamin C, and vitamin E, in scavagine free-radicais [45]. The apparent protective effect of beta-carotene and other anti-oxidants in observational studies has led to controlled ciinical trials of vitamin A. beta-carotene- svnthetic retinoids. and vitamin E and selenium in persons at high risk of lung cancer, including smokers and asbestos-exposed workers [46.47]. Pilot studies have demonstrated feasibiiity. and full-scale interven- tion trials are in progress. Another aspect of diet which may play a role in lung carcinogenesis is that of fat intake. International cor- relation studies suggest that there is an association between fat intake by country and lung cancer inci- dence or mortality [48-50]. One study in particular [50] found that per capita supply of animal fat was stronely associated with lung cancer mortality rates. This finding appeared to be due to the interaction between cigarette smoking and animal fat consump- tion. The authors concluded that cigarette smoking mav have a lesser imoact on lung cancer mortalitv in populations with a low intake of saturated fat. Other ecological studies correlating tobacco consumption and lung cancer rates between different countnes or in subpopuiauons within a country [51.52] are consistent with a modifvine effect of fat intake on the association of smoking with lung cancer. Several case-control studies provide evidence of a positive association between dietary fat/cholesterol intake and lung cancer. The first of these studies, from Hawaii, showed a relative risk of 3-5 (95%confidence interval 1.7-7.2) in males for the highest level of die- tarv cholesterol intake versus the lowest level, after adiustment for smoking and other covariates [53]. Later studies have provided generally confirmatory results [54-57.44]. In a prospective study of middle- aged American men. dietary cholesterol was as- sociated with increased lung cancer after adjustment for smoking. age, intake of beta-carotene and fat: however. the association held oniv for cholesterol from eggs. not from other sources [57]. BODY MASS INDEX Eight prospective studies and one case-control study have noted an association between leanness and lung cancer[58.59]. The association does not appear to be explained by differences in smoking habits or to weight loss due to disease. One possible explanation is that leanness may be associated with decreased levels of nutrients that are protective or with increased levels of dietary risk factors [581. Further studies are needed to determine whether the association of low body mass with lung cancer is due to the influence of factors associated with leanness or to a biological effect of leanness itself. ALCOHOL CONSUMPTION A number of reports have suggested that alcohol consumption is associated with lung cancer indepen- dent of smoking [60,61]. However. a large case-control 2081783282 ~ I I  I

I I I I I ~ I I t I I I I 1 ® I I study [62] showed clearly that although alcohol drink- ing by itself was associated with lung cancer, after adjustment for smoking, the association totally disap- peared. Results from a prospective study of men of Japanese ancestry in Hawaii [63] indicated that the association of alcohol intake with lung cancer was reduced after adjustment for smoking but still retained marginal statistical significance. Because smoking is the overwhelming risk factor for lung cancer (with relative risks reaching 20.0. 30.0. and higher in heavy smokers), and since smoking and drinking habits are highly correlated. it is crucial to adequately adjust for lifetime smoking habits before drawing conclusions about an association of alcohol with lung cancer. REPRODUCTIVE FACTORS AND EXOGENOUS HORMONES The higher prevalence of adenocarcinoma oi the iung in women compared to men suggests a possible roie of endocrine factors ( menstrual history. reproduc- tive history. use of exogenous hormonesl in the devel- opment of this type of lung cancer. Several studies have reported observations which are consistent with a role of endocrine factors: the finding of steroid receptors in some lung cancers [64]; a higher-than- expected rate of lung cancer. pantcuiarly adenocarcn- noma- among 10 + year survivors of endometnal can- cer [65]; an apparent increase in the risk of lung cancer in women receiving potent estrogens as hormone re- placement therapy [66]; and a significantly increased risk of adenocarcinoma of the lung (after adiustment tor smoking) in Chinese women with short menstrual cycles I<'_6 days) [671. [Chinese women have high rates of lung cancer tpredominantly adenocaranomai in spite of a low prevalence of smqking, potnting to the importance of factors other than smoking.) More ex- tensive investigation of the relation of endocrine fac- tors to lung cancer is needed before any conclusions can be drawn. HOST FACTORS Indirect evidence of a genetic component in the eti- ology of lung cancer comes from the facts that ( I) not all smokers who reach the age of 80 develop lung :ancer. and (2) most carcinogens require metabolic activation and this is under genetic control. Analysis of famiiies of cancer cases and controls, pedigree anal- ysis, and studies of genetic markers have been carried out in an effort to identify a genetic factor [68-721. Looking at familial clustering of lung cancer and other diseases enables one to study the interaction between genetic endowment and environmental exposures, al- though famiiial aggregation does not prove the pres- ence ot a eenettc component. since smoking habits and Epidemiology of Lung Cancer 77 other environmental factors (diet. infectious diseasesl also are known to aggregate in families. However. lung cancer tends to cluster in families& even after adjustment for smoking habits [73.741. Furthermore. adenocarcinoma and alveolar cell carcinoma are more common in families with other cancers. acquired im- mune deficiencies. or heritable disorders of the lung [751. To date, studies of genetic markers of lung cancer risk. such as aryi hydrocarbon hydroxylase and de- brisoquine phenotype, have been inconclusive [76]. The findine, in four studies- that extensive metaboliz- ers of debrisoquine are at increased risk of lung cancer relative to poor metabolizers would appear to be the most promising evidence for a genetic component in lung cancer. However, no metabolic pathway linking debrisoquine metabolism and metabolism of known lung carcinogens has yet been identiued [76]. Researcn in the area of ¢enetic control of metabolic activation and detoxification or carnnogens is likely to make a major contribution to understanding the inter- action between exposure to carcinogens and host sus- cepttbility. OTI-IER RISK FACTORS A report from the Netherlands has suggested that keeping pet birds in the home may be an independent risk factor for lung cancer [77]. The adds ratio for lung cancer among keepers of pet birds was 6.7 (95% con- fldence intervat 2'-20.0) after adjustment for smok- ing and vitamin C intake. Of two more recent studies with larger sample sizes undertaken to confirm this finding. one showed a more modest association (78], while the other found no association with birdkeeptng in the home but did note a relattonship limited to keepmg pigeons outside the house_(79]. Further stud- ies which control in greater depth for lifetime smoking and which address possible sources of bias could shed light on the nature of this intnguing association. CONCLUSION Two points emerge from this brief review of the eoidemiolo¢y of lung cancer. First. although smoking, and secondarily occupation. are major established risk factors, other factors. inctudine other environmental exposures and host susceptibility are likely to play a role either as independent risk factors or synergisti- caily with smoking or occupational exposure- As noted in the introduction. the occurrence of lung can- cer in certain groups and recent changes in the distri- bution of histoiogic types represent a challenge to fur- ther research. Second. since smoking is the overwhelming risk fac- Lor for lung cancer and because of the discretionary nature of smoking, lung cancer is. to a large extent. a I

I 78 Kabat preventable disease. In view of its preventability. ef- forts aimed at reducing the prevalence of smoking in developed countries should be intensined and new initiatives under the ae¢is of the World Health Oreani- zation undertaken to reverse the expansion of tobacco use in developing countries. ACKNOWLEDGMENTS The author is grateful to Dr. Emanuela Taioli for comments on the manuscript. REFERENCES I. Boring CC. Squires TS. Tone T: Cancer Statistics. 1992. CA 42:19-38. 1992. 2. Parkin DM: Trends in lune_ cancer incidence worldwide. Chest 96:5S-8S. 1989. 3. Devesa S5. Shaw GL. Blot WI: Changing patterns of tune cancer incidence by histolomcai tvpe. Cuncer Epidemml Bi- omarkers Prev: 1:29-34. 1991. 4. Stockwell HG. Armstrone AW. Lcavenon PE: Histooa- thology of tune canccri among smokers and nonsmokers in Florida. Int J Epidcmtol 191supol 11:548-S52. 1990. 5. Fontham ETH. Correa P. Wu-WiOiams A. et at: Lung cancer in nonsmoking women: A muiucemer case-conuot stud_v. Can- ccr Epidemtol Biomarkcn Prey 1:35-43.1991. 6. Wynder EL. Goodman MT: Smoking and lung cancer: Some unresolved tssues. Epidemtol Rev 5:177-207. 1983. 7. Yang CP. Gallagher RP. Weiss NS. et al: Differences in ma- dena rates of cancers of the resotratorv tract by anatomic substte and histologtc type: An etiologic tmpiicauon. JNCI 81:1828-1831. 1989. 8. SteBman SD: Interactions between smoking and other exoo- sures. ln Hoffmann D. Harrts C(cdsl: Banburv Reoon No. 23. "Mechantsms of Tobacco Caranogenests." Cold Spnng Har- bor. New York: Cold Spnne Harbor Laboratory. 1986. 377- 393. 9. Archer VE. Wagoner JK. Lundin FE: Uranium mining and nearette smoxm_e eriects on man. 1 Occup Med 15:204-211. 1973. 10. Hammond EC. Selikoff IJ. Seidman H: Asbesms exoosure. _ ctgarette smokm¢ and deam ates Ann NY Acad Sci 330-473- 19. Wynder EL. Kabat GC: Effect of low-vicid ciearette smoking on tune cancer risk. Cancer 62:1=-1230. 1988. :0. Kabat GC. Morabta A. Wynder EL: Comparison of smoking habits of blacks and whites tn a case-control study. Am 1 Public Health 81:1483-1486. 1991. 21. Kabat GC. Heben JR: Use of mentholated cigarettes and tune cancer risk. Cancer Res 51:6510-6513. 1991. '-L Resnicow K. Kabat G, Wynder EL: Progress in decreasing ctgarette smoking. in DeVita VT. Heiman S, Rosenberg SA ledsl: "Important Advances in Oncoiogy." New York: 1.B. Lippincott. 1991. 205-213. 73. Walter HL Vaughan RD. Wvnder EL: Primary prevention of cancer among children: Changes in eearetc smoking and diet after six years of interventton. JNCI 81:995-999. 1989. 24. Fraumeni JF. Blot WJ: Lung and pleura. In Schottenfeld D. Fraumeni JR (eds): "Cancer Epidemioloey and Prevenuon." Philadelphia: W.B. Saunders. 1982. 564-582. '_5. Tomatts L. Aiuo A. Wilboum J. Shuker L: Human carnno- gens so far identified. Jpn 1 Cancer Rcs 80:795-807. 1989. 36. Peto R. Schnetderman M: "Quantification of Occupational Cancer." Banbury Report No- 9. Cold Spring Harbor. New York: Cold Sprlne Harbor Laboratory. 1981. _. Doll R. Peto R: "The Causes of Cancer. Quantitative Esti- mates oi Avoidable Risks of Cancer in the United States Today." Oxford: Oxford University Press. 1981. Tomaus L: Outdoor atr pollution and tune canccr. Ann Oncol 2:265-267. 1991. =9. Persna8cn G. Simonato L: Epidemmlo0cai evidence on air pollunon and cancer. !n Tomaus L/ed): "Air Pollution and Human Cancer. European School of Oncology Monographs." Berlin: Spnnger-Verlag. 1990. 63-74. 30. Hirayama T: Non-smokmg wives of heavy smokers have a higher nsk of tung cancer. A study from Japan. Br Med J ?82:183-185. 1981. 31. Trichopoulos T. Kalandidi L. Sparros L. McMahon B: Lung cancer and passive smoking. Int J Cancer 27:1-4, 1981. National Arademv of Sdences: "Environmental Tobacco Smoke: Measuring Exposures and Assessing Health Effecu." Washtn¢ron, D.C.: Nauonai Academy Press. 1986. 33. Koo LC. Ho 1H-C. Ryiander R: Life-htstory correlatcs of environmental tobacco smoke: A study on nonsmoking Hone Kong Chinese wives with smokme versus nonsmokmg nus- bands. Soc Sci Med 26:751-760. 1988. 34. Kabat GC. Eoidcmtoiomc studies of environmental tobacco smoxe anu lunz cancer. Proceeotnes of the Toxicoto¢v Forum. 496. 1979. Winter Meeune: 187-201. 1990. 11. United States Depanmrnt of Health and Human Ser•nccs: 35. Lee PN: "EnvtronmrntaP Tobacco Smoke and Mortality." "Reductne the Health Conseouencu of Smoktne: 25 vears of Baeel: S. Kar¢e:. 1992. progtess." A Repon of the Surgeon General. (Jmted States 36. National Academy of Sciences: "Hralth Risks of Radon and Department of Health and Human Semces. Public Health Other Intemally Deposted Aloha-Emsuen." National Acad- Service. Centers for Discase Control. Center for Chronic Dis- emy oP Scenes Report: BEIR IV. Washington. D.C.: Na- ease Prevention and Health Promouon.OBtm on Smoking and ttonal Academy Press. 1988. Health. DHHS Publicauon No. (CDC) 89-841 l. 1989. 37. Harley NH. Harley JH: Potential lung cancer risk from indoor 12. International Agency for Rexareca on Caneea Tobacco Smok- radon exposura. CA 40:265-275. 1990. ing. IARC Monographs on the Evaiuauon of the Carcinogenic 38. Schoenber¢ J B. Klo¢ JB. Wilcox HB. et al: Case{ontrol study Risk of Chemicals to Humans. Vol. 38. IARC. Lyon. France. of residential radon and lung cancer among New Jerscy 1986. women. Cancer Res 50:6520-6524. 1990. 13. Loeb LA. Ernster VL. Warner KE. et al: Smakine and lung 39. Blot WJ, Xu Z-Y. Botce JD. et al: Indoor radon and tune cancer. An overview. Cancer Res 44:5940-5958. 1984. cancer in China. JNC[ 82:1025-1030. 1990. 14. Hill AB: The environment and disease: Assocauon or causa- 40. Svensson C. Pcrsnagen G. Klommck 1: Lung cancer in women tian? Proc R Soc Med 58:295-300. 1965. and type of dwelling in rclauon to radon exposure. Cancer Res 15. Kreyoerg L: Histological lung camzr types. A morphological 49:1861-1865. 1989. and biological correlation. Acta Pathol Microbtol Smnd 41. Lubin JH. Samet JM. Wein'oer¢ C: Design issues in epidemi- Suppl. 157:1-92. 1962. olo8tc studies of indoor exposure to Rn and nsk of lung can- 16. Damber LA. Larsson LG: Smokine and lung cancer with spe- cer. Health Phys 59:807-817. 1990. clal regard to type of smokme and type of cancer. A case- 42. Colditz GA. Stampter MI. Willett WC: Diet and lung can~n control study in north Sweden. Br J Cano;r 53:673-681. 1986. A review of the eptdemiologic evidence in humans. Arch Intern 17. Morabia A. Wvndcr EL: Cizarecte smokm_e and lung cancer Mcd 147:157-160, 1987. cetl typn. Cancer 68:2074-2078. 1991. 43. Le Marchand L. Yosbizawa CN. Kolonel LN, et at: Veectable 18. Mombta A. W'vnder EL: Relation of bronchtoloalveoiar ram- consumption and tune cancer nsk: A populauon-based case- noma to tobacco. Br Med J 304:541-543. 1992. control studv in Hawaii. JNCI 81:1158-1164. 1989. I It I I I I I It I 2081783284 1

I I I I I I I I I I I I I I I I i 44. Jain M. Burch JD. Howe GR. Miller AB: Dietarv factors and risk of lung eancer. Results from a casetonuol studv, Toronto. t981-85. 1nt 1 Cancxr 45:549-553. 1990. 15. Packer JE. Mahood 1S. Mora-Arellano VO. et at: Free radicals and sineiet oxygen scavengers. Biochem Biophvs Res Commun 98:901-906. 1981. 46. Malone WF: Studies evaluating antioxidants and betataro- tene as chemopreventives. Am J Clin Nutr 53:305S-3135. 1991. 47. Omenn GS. Goodman GE. KJeinman GD. et at: The role of intervention studies in asartaimne Ihe contribuuon of die- tary factors in lung cancer. Ann NY Acad Sci 534:575-583. 1988. 48. Kolonei LN. Hankin JH, Lee J. et at: Nutrient intakes in relation to cancer tnndrnce in Hawaii. Br J Cancer 44:332- 339, 1981. 49. Wynder EL- Hebert JR. Kabat GC: Association of dietary fat intake and lung cancer. JNCI 79:631-637. 1987. 50. Xie 1, Lesaqre E. Kesseloot H: The rciauonship of animal fat intake. ctgarette smoking, and lune cancer. Cancer Causes Control 2:79-83, 1991. 51. Wynder EL-Taioli E. Fujita Y: Ecolo¢ic stuav of lung cancer risk factors in US and Japan, with spcnai ref erence to smoKUtg and diet. lanan J Cancer Rcs 83:418=23. 1992. ::. Taioii E. Nicolost A. Wvnder EL: Posstbic roic oi dlet as a host actor in me euotoey oi tobaceo-mduad tune cancec An eco- logtcal study in southern and nonhem ltaly. (nt 1 Eptdemtoi 20:611-614. 1991. 53. Hinds MW. Kolonel LN- Lee J. et ai: Dietarv choiesterol and lung cancer risk among men in Hawaii. Am I Clin Nutr 37. [92-193. 1983. 54. Goodman MT. Kolonel LN, Yoshizawa CN. et al: The effect of dictarv, cholesterol and fat on the risk of lung cancer tn Hawaii. Am J Epidemtoi 128:1241-1255. 1988. 55. Byers TE. Graham S. Haughey BP. et at: Diet and lung cancer: Findings from the Western New York Diet Studv. Am J Epidemtot 125:351-363. 1987. 56. Knekt P. Seppanen R. Jarnnen R. et at: Dietary cholesterol. fatty aads. and the risk of lung cancer among men. Nutr Cancer 16:267-275, 1991. 57. Shekelle RB. Rossof AH. Stamler J: Dietarv cholesterol and incidence of lung caneer. The Western Electric Study. Am J Enidemwi 134:480-484. 1991. :3. Knekt P. Heliovaara M. Rissanen A. et at: Leanness and lung cancer nsK. Int J Cancer 49:208-213. 1991. 59. Kabat GC. Wvnder EL: Bodv mass mdcx and lung cancr nsx. Am I Epidemtoi 135:769-774. I992. 60. Pouer JD. McMichael AJ: Alcohol. beer and lung cancer-A meamnaful reladonship? tnt J Epidemml 13:'-40-242. 1984. Epidemioiogy of Lung Cancer 79 61. Stockwell HG. Matanowski GM: A case-control studv of lung cancer in painters. J Occup Med 27:125-126. 1985. 62. Williams RR. Horm JW: Association of cancer sites with to- baceo and alcohol consumouon and soaoeconomtc status of patients: Interview study from the Third National Canmr Sur- vey. JNCI 58:525-547. 1977. _ 63. Stemmerman GN. Nomura AMY. Chvou PH- et at: Prostuc- tive studv of alcohol intake and laree bowel cancer. Digesc Dis Sci 35:1414-1420, 1990. 64. Cagle PT. Modv DR. Schwartz MR: Estrogen and progester- one reoeotors in broncbogentc carnnoma. Cancer Res 50: 6632-6635, 1990. 65. Anaeeers 1F. Malkasian GD: Patterns of other neonlasia in patients with endometriaL carctnoma. Cancer 48:856-859. 1981. 66- Adami H-O. Persson 1. Hoover R. et al: Risk of cancer in women receiving hormone repiaeemrnt therapy. Int 1 Cancer 44:833-839. 1989. 67. Gao Y. Blot W. Zheng W, et at: Lung cancer among Chinese women. Int J Cancer 40:604-609. 1987. 68. Wu AH. Yu MC. Thomas DC. et al: Personal and familv history of lung disease as risk factors for adenocarctnoma o[ the lung. Cancer Res 48:7279-7284, 1988. 69. Sellers TA. Bailev-Wilson JE. Elston RC. et at: Evidence of Menaeiian mhentance in the pamo¢enes¢ oi lung cancer. J J N C I 8 2:1272 - L2 7 9. 1990. -0. McDume HH: Clustenne of cancer m families of patients with prmtarv lune cancer. 1 Clin Epidcauol 44:69-71. 1991. "I. Lynch HT. &imberlin¢ WJ. Markvska SE. ct al: Genctta and smoiung-auomatcd canccrs: A study of 485 families. Cancer 57:1640-1646. 1986- 7'-. Ayesh R. Idle JR. Ritchie JC. et al: Metabolic oxidation phenotypes as markers for susceptibility to lung cancer. Nature fLond1.312:i69-170, 1984. "3. Tokuhata GK. Lilicnfeld AM: Familial aggregation of lung can¢r in humans. 1NCi 30.289-312. 1963. 74. Ooi WL. Eiston RC. Chen VW. et at: Increased familial risk for lung cancer. JNCI 7b:217-17~ 1986. 'S. Mulvihill !J: Host factors in human tumors: An example of cco-genencs in oncology. JNCI 57:3= , 1976. 76. Law MR: Genetic predisposition to lung cancer. Br J Cancer 61:195-206. 1990. 77. Hoist PA. Kromhout D. Brand R: Pet birds as an mdepenaent nsk factor for tune cancer. Br Med J 297:1319-1321. 1988. '8. Kohlmner L. Amtmeer G. Bartolomevnk S. ct at: Pet birds as an macoenocnt nsk factor tor lune cancer. Case-control studv. Br Med J 305:986-989. 1992. 79. Gardiner AJS. Forev BA. Lee PN: Avian cxoosure and bron- choeentc carcinoma. Br Mcd 1 305:989-992. 1992. I