Tobacco Documents Online

Cancer Detection ana Preti•erttion, li8(6):321-330 (1994) Women and Lung Cancer: A Comparison of Active and Passive Smokers with Nonexposed Nonsmokers G. H. Miller, Ph.D., C. P. C., F.A.I. C.,a Joseph A. Golish, M.D., F.A. C. P., F.C.C.P.,b Charles E. Cox, M.D., F.A.C.S.,c and Donna C. Chacko, M.D.d •Stud+'es on Smoking, Inc., 125 High Street, Edinboro, PA 16412; °Department of Pulmonary Disease, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44106; °8reast, Cancer Center, H. Lee Moffit Cancer Center, 12902 Magnolia Drive, Tampa, FL 33682-0179; and °Radiation Oncology Center, Palms of, Pasadena Hospital, St. Petersburg, FL 33707 A,ddross rcpnnt requests to Dr. G. H. Mi1kr. ABSTRACT: Prior to the 1920s. lung cancer was a rare disease. However, the current increase in lung cancer appears to parallel the increase in smoking for both men and women wittt a 30- to 50-year delay: National lung cancer deaths continue to rise, with over 168.000totsl deaths estimated in 1992. VS%omen are now showing higher percenta¢e increases in lung cancer than men from active smoking. ~ The data from the Ene County Study on Smoking and Health (ECSSH), a population study. were used to measure the effects of both active and passive smoking on women's lung cancer mortaliry: The three major categories of exposure (no known or minimal exposure„pa.ssive smoking exposed, and active smoking) were used in the analyses. The results from the population data in Erie County; PA, were based on 528 nonexposed nonsmoking women. 3138 exposed nonsmoking women, and 1747 smoking women, Deaths due to lung cancer as a percentage of total deaths excluding traumatic deaths were 0.2% for the nonexposed nonsmoking women. 0.9% for the exposed nonsmoking women. and 8.0% for women who smoked The data showed'that women smokers died of lung cancer ara rate 9 times greater than exposed'nonsmokers and42 times greater than nonexposed nonsmokers. KEY WORDS: life expectancy, lung cancer, mortality, passive smoking, second hand smoke,smoking, women. 1. INTRODUCTION Prior to the 1920s, lung cancer was a rare disease.' Since that time the number of cases has increased substantially. The American Cancer Society estimated that over 168,0001ung cancer deaths would be reported in the U.S. in 1992.2 The increased incidence of lung cancer cases ap- pears to parallel the number of years of smoking by both men and women. During WW I and WW II, military personnel were provided with cigarettes and smoking be- came an accepted behavior pattern. Extensive advertising by the cigarette companies glamoriz- ing cigarettes made such deep inroads into the American subconscious that over 60% of the male 0361-090XJ94/5.50 ® 1994 by CRC Press, Inc. population became cigarette smokers. Prior to WW II few women smoked; after WW II over 30% of the adult women reported themselves as smoking regularly. A number of studies in the 1940s, 1950s, and 1960s reported on the dangers of cigarette smok- ing.3 The results of these studies were summa- rized in a 1964 report prepared by Dr. Luther Terry, the Surgeon General, showing how devas- tating smoking is to health.3 The mortality rate from lung cancer for men became significantly higher than that of any other type of cancer. The early reports on lung cancer in women did not show a high mortality rate. However, the cumulative smoking years of women compared with men were very small at that time and thus 421

Women vnd'Lung Cancer did not show the same high lung cancer mortality rates as men.' Both health professionals and the public, therefore, came to believe that women, because of their low rate of lung cancer, were less susceptible to the detrimental effects of cigarette smoking than were men. However, as more and more women contin- ued to smoke and were joined by women in younger age groups, mortali'ty and morbidity rates increased dramatically for women. Lung cancer mortality rates over the last 25 years for women cigarette smokers has increased as much as 500% and~ appears to have superseded breast cancer as the prime cancer killer among women 3 Cigarette smoking, apparently, is now achieving the same devastating effect on women as it has on men when cumulative years of smoking are taken into account, as shown in Figure 1 where the percent- age of lung cancer of women is converging with that of men. The most recent research on lung cancer among women has examined active smoking and exposure of nonsmokers to carcinogens in to- bacco smoke, which has been classified as pas- sive smoking. tobacco smoke pollution, or invol- untary smoking. Before the late 1970s, passive smoking has been assumed to have few harmful effects for either men or women: The recent review of the literature on passive smoking and lung cancer can be separated into three different types of reports: (a) comparisons of active and passive smoking,b-12 (b) passive smoking only;"-2' and (c) reviews summarizing the effects of passive smoking.=5-'0 Nearly all of the reports support the conclusiom that passive smoking is responsible for lung cancer in non- smoking women. Results from the Erie County Study on Smoking and Health (ECSSH), which was designed to provide population-based data that could contribute the most representativeness to the subsamples selected for analysis, are re- ported in this article. II. METHODS Data from the ECSSH are based~on informa- tion obtained from the Northwestern Pennsylva- nia Study on Smoking and Health (NPSSH): - now renamed the ECSSH because the data includes only deceased residents of Erie County.4113 This population study was initiated in 1973 to determine 0 10 20 30 40 50 60 70 80 90 PERCENTAGE OF LUNG CANCER DEATHS FIGURE 1. Comparison of lung cancer deaths; male vs. female from 1955. 422

retrospectively the smoking habits of all deceased residents (t-oth male and female) of Erie County, PA. All information was based on telephone in- terviews with close relatives of the decedents. The standardized interview form was devel- oped cooperatively by the director of the study,, members of the local health associations, and the Pennsylvania Department of Health. The inter- view forms were designed so that the important data could be obtained in a minimum amount of time so as not to burden unnecessarily the rela- tives being interviewed. Information was collected on the decedent's smoking history - type of tobacco used, amount used, lifetime duration of smoking, amount of inhalation (none to deep), percentage of tobacco product smoked (quartiles), attempts to quit smoking, age at which the dece- dent started smoking, occupation, other possible moderate or high-pollution exposures, amount of exercise. height and weight, cause of death and diseases prior to death, including length of time for these diseases. In order to obtain information relating to pas- sive smoking, additional questions were incorpo- rated into the questionnaire concerning the smok- ing histories of the spouses and parents of the deceased (amount of their smoking, whether or not they stopped smoking, age at death, year of death, diseases and cause of death) and whether the deceased had additional exposure in the house- hold or other locations. In addition, information was obtained on other sources of exposure out- side of the household such as bingo halls and card games where smoking is permitted. This informa- tion was used to establish estimates of length of exposure from all sources of tobacco smoke from spouses. parents, others in the household, or from other environments. This detailed type of data have not been available in most studies on active or passive smoking and provides important infor- mation from detailed population data that can be used to demonstrate the detrimental impact of these forms of tobacco exposure. Death notices and obituaries from the Erie County newspapers for the years 1972 to 1976 and 1979 to 1984 provided the names of almost every person who died in Erie County along with the names of their closest surviving relatives. Lists of all the deceased provided by the Pennsylvania Department of Health made it possible to locate Cancer Detecrion and Prevention all those decedents who were not declared resi- dents of Erie County, who died from accidental causes, or who were under 30 years of age. In order to provide the best estimates of the full impact of both active and passive smoking, all deaths listed as due to accident, homicide, or suicide, all decedents under age 30, and all non- residents of Erie County were omitted from the analysis. Eliminating these members in the above categories reduced the total population size by approximately 23%. In order to obtainl complete information on the smoking histories of the decedents, as such detailed informationi is not available in hospital or other medical records, telephone numbers were sought for up to three surviving relatives for each decedent listed in the death notices and obituaries in the Erie County newspapers. In Pennsylvania, these listings provided the nearly complete popu- lation of the deceased for all counties for each year. Telephone numbers were obtained for 85% of the surviving, relatives by research assistants making use of telephone books for the different years, which were supplied by the General Tele- phone company (GTE). The director of the study and interviewers who were trained by the director called the decedent's relatives up to a maximum of six times for their responses and were able to contact 90% of those called. Interviewers did not know the cause of death while gathering the data. The inter- viewers asked the relatives of the deceased the cause of death of the deceased and any prior illness. Although the cause of death obtained by the interviewer was listed on the interview form, only the designated cause of death based on the Intemational Classification of Diseases (ICD) codes were used in the final coding and analysis. The ICD codes for all the deceased were provided by the Pennsylvania Department of Health for each year that was used in the analysis. A com- parison of the accuracy of the off cial ICD code and the reported diseases by the close relatives showed high reliability in the range of 85 to 90%. The additional diseases prior to death were re- corded with the designated ICD code. Of the rela- tives contacted, 95% cooperated in the study.. Three major categories of exposure were des- ignated for the study. The fust was the nonexposed category (NX) that contained only those women 423

Women and Lung Cancer who had no known reported exposure to •obacco smoke. Although it is unlikely that anyone could have avoided all tobacco exposure during their lifetime, the nonexposed category was established to classify decedents who had no known exposure or at most a very limited exposure to tobacco smoke. The second category, the passive smoking exposed group (EX), includes those nonsmoking women who were reported to have had moderate (20 years) to lifetime tobacco smoke exposure, that is, women ~ nonsmokers with sources of expo- sure to tobacco smoke such as a smoking spouse, additional smokers in the home, smokers in the workplace, and exposure in other environments. Inasmuch as only two of the nonsmoking women who contracted lung cancer had less than 20 years exposure, 20 years exposure was selected as the minimal exposure to passive smoking for these analysis. The specific category of 20 years was selected for those deceased who were exposed by one or more parents since a pilot project estab- lished 20 years as the best approximation for childhood exposure. The 20 years was then: added to the years of exposure by the spouse to obtain the total years of exposure for those exposed dur- ing childhood. If there was no childhood expo- sure. then only the exposure by the spouse was considered. In the present study no further break- down of the categories of 20 years to lifetime exposure was completed'as the ana]ysis was made prior to the completion of the 20-year study - 1972 to 1991. The third category, the active smoking cat- egory (SM), was defined as women who smoked more than 20 packs of cigarettes during their lifetime. Data were gathered for approximately 55% of the deaths among Erie County women for the months completed in the years 1972 to 1976 and 1979 to 1984. A total of 5413 interviews were obtained from relatives of the deceased women: 3666 nonsmoking women - 528 NX, 3138 EX - and 1747 Smoking women. Cause of death as reported on the death certificate, supplied by the Pennsylvania Department of Health listings, was used for the analysis of the data following the assumption that the physicians listing of cause of death would be more accurate than that of the relative's designations. The decedents were placed into one of the three different exposure categories for coding according to the information obtained from their smoking history as noted above (NX, EX, SM). The population data were analyzed to deterrnine the effects of both active and passive smoking as wellias to determine the effects on those who have had little to no exposure during their lifetime. This report deals with comparisons of these three levels of exposure of the population data to esti- mate the incidence of lung cancer caused by the different exposure levels. In addition, the retrospective case-control tech- nique, as embodied in the National Cancer Institute's newest software, EPI INFO (Version 5), was used to determine levels of significance. Lung cancer deaths (cases) were computed as a percent of total deaths (controls) for all three categories for the analysis. Statistical analysis utilizing odds ra- tios (OR) and standard mortality ratios (SMR) were computedL III. RESULTS The results of the analysis of the population data from the three groups of exposure levels is presented in Figure 2. One of 528 deaths (0.2%) among women with no known exposure to to- bacco smoke (NX category) was due to lung can- cer. Of 3138 deaths, 27 (0.9%) among nonsmok- ing women who were exposed to some sources of tobacco smoke (EX category) were reported as due to lung cancer. Among women with a known smoking habit (SM category), which included former smokers and late-starting smokers, 140 of the 1747 deaths (8.096) were due to lung cancer. A comparison of the population of lung can- cer deaths obtained in the Erie County Study on Smoking and Health with the actual, population of lung cancer deaths as a percentage of total deaths for women in Erie County reported for the years 1972 to 1976 and 1979 to 1984 is found in Fig- ure 3. Using lung cancer as the cases and non-lung cancer as the controls the three categories of ex- posure levels were analyzed using the chi-square test to determine any signif cant trends. A highly significant chi-square value of 207 (p <001) was computed for the 3 x 2 table (refer to Table I). The 424

Cancer Detecrior+ and Prerenrlort NONEXPOSED EXPOSED SMOKING LEVELS OF EXPOSURE FIGURE 2. Lung cancer: nonexposed nonsmokers with exposed nonsmokers and active smokers. FiGURE 3. Erie County lung cancer data: Erie County data vs. Pennsylvania Department of Health idata.

Womerr and Lung Cancer TABLE I A Retrospective Case-Control Analysis of the Erie County Population Data of Women and Lung Cancer with the Comparison of Three Different Groups: Nonexposed (NX), Exposed (EX) and Smoking (SM) Group NX Controls (non-lung cancer) 527 Cases (lung cancer) 1 Total 528 EX 3111 27 3138 SM 1607 140 1747 5235 168 5413 Chi-square anaJysis (3 x 2 table) - 207„ p<0:00000001 Group Comparisons (2 x 2 tables) Groups Odds ratio (confidence Intervals) Chi-square Prob. (test) NX vs. EX 4.57 (0.67 < OR < 90.71) 1.87 0.07 (FISHER EXACT) NX vs. SM 45.91 (6.97 < OR < 886.97) 41.36 0.000001 (YATES) EX vs. SM M04 (6.51 < OR < 15.57) 171.76 0.000001 (YATES) 2 x 2 comparisons also showed significant re- sults. A comparison of the NX and EX groups showed a nearly significant chi-square value of 1.87 (p = 0.07). The comparison of the NX and SM groups produced a highly significant OR of 45.91 (p <0.0001). The comparison of the EX and SM groups also yielded another significant OR of 10.04 (p <0.0001). The results (shown in Table II) of age adjust- ment of the data based on observed death.rates in Erie County show significant SMRs for nonsmok- ing nonexposed (SMR = 17, p<0.01), passive smoking (SMR = 37, p<0:001) and active smok- ing (SMR = 170,,p <0.001). IV. DISCUSSION The data from this population study of 5-11'3 deaths among women in Erie County show some very specific trends. Few women who had no known exposure to tobacco smoke died of lung cancer: only 1 in 528 deaths. The low rate of lung cancer for the nonexposed nonsmoking women is TABLE il Age-Adjusted Data for Lung Cancer Making Use of Erie County Death Rates to Compare Nonsmoking Nonexposed Women with Both Nonsmoking Exposed Women and Active Smokers Category Expected Observed SMR Probability NX 5.8 1 17 .02 EX 45.7 17 37 .001 SM 44.6 76 170 .001 ~ Note: NX - nonsmoking nonexposed women; EX - ex- posed nonsmoking women; SM - smoking women; SMR - standard mortality ratio where the expected = 100%. 426

consistent with the results of a population study12 on- a comparison of lung cancer cases between Amish and non-Amish men and women in Lancaster County, PA. The single lung cancer case reported among the Amish in the analysis of the data from the Lancaster County Cancer Reg- istry from 1971 to 1977 was that of a male cigar smoker. Among the Lancaster County population there were 3-18 deaths from lung cancer for women out of approximately 16,000 total deaths for that time period. During this time period very few Amish smoked, and thus. they are one of the best examples of a nonsmoking popul'ation in the U.S. For the period studied there was not a single case of lung cancer among the nonsmoking Amish population. For those women in the EX group, the pro- portion of lung cancer deaths is 4.5 times higher than those in the NX group, with lung cancer reported in about 9 in 1000 deaths. This result is consistent with the conjecture that exposure to ambient levels of tobacco smoke does increase the risk of lung cancer. Approximately 8.0% of the deaths among SM women were due to lung cancer, 42 times greater than in~ NX wornen and 9 times greater than those who were exposed to passive smoking. Because of the vast differences in lung cancer mortality among the three different groups, NX, EX, and SM, of women, it can be concluded that both active and passive smokers are much more likely to die of lung cancer than those women who had no known or minimal ex- posure (Figure 2). Figure 3 shows the percentage of deaths due to lung cancer in this study (1:68 of the total 5413 decedents) is 3.1 %. The data reported from the Pennsylvania Department of Health for lung cancer for women in Erie County for .he same time period is 360 cases of lung can- cer of a total of 13,329 deaths among women, or 2.7%. Thus, the data from the completed interviews are very close to the actual lung cancer death rate for Erie County, indicatingg that the data from this population is likely to be most representative of the actual total popula- tion for that time period. The computation of the SMRs is indicated in Table II. The SMR in this instance is based on the "observed" death rates of lung cancer for the Cancer Detecnon~and'Preventton three different categories of exposure (NX. EX, SM) relative to the "observed" death rates of lung cancer in the age-adjusted population of the women in Erie County during the years under consider- ation~ without regard to smoking exposure. The comparison of the "observed" lung cancer death rates with the "expected" lung cancer death rates for these categories resulted in the following rela- tive rates: 17% for NX, 37% for EX, and 170% SM. The data show a progression of the minimal effects of nonexposure, the moderate effects of passive smoking exposure, and the greater detri- mental effects of active smoking. The NX classification in this study is an im- portant distinction from most other studies on passive smoking where both; NX and EX women are combined in the passive smoking category or the NX category is not considered in~the analysis. This distinction is necessary in determining dif- ferences that may exist between the nearly totally nonexposed and those nonsmoking women who are exposed: to other sources of passive smoking. Therefore, we recommend that studies on passive smoking make use of these three categories for classification of exposure so that a comparative analysis can be made between this study and'simi- lar studies. Another distinction between this study and others is that this study attempts to include all potential sources of exposure to passive smoking in the EX category. For this reason, data have been collected where possible on~ whether others in the household in addition to a smoking spouse were producing tobacco smoke exposure for non- smokers, such as smoking children-or other smok- ers in the household as well as other smoke-filled environments. Because there appears to be expo- sttre in the workpl.ace," working wives were placed in the EX category. Although some women are not exposed in certain environments, most women who work are exposed and therefore should be placed in the EX category. Additional substantia- tion of the increased exposure of working women to passive smoking is provided by Coughlin' and Friedman'6 This reclassification of the nonexposed non- smoker provides a much purer nonexposed group than that of classifications that include these ex- posed or potentially exposed groups in the 427

Women and Lung Cancer nonexposed group. Again for comparative pur- poses, it would be important for other studies to be sure that their nonexposed groups be as free as possible from potential tobacco smoke pollution. Population studies involve greater time and expense but such studies contribute greater repre- sentativeness of the data, It would be valuable if other whole-population studies could be con- ducted, which would allow for the comparison of population data. Because this study contained so few women exposed to hazardous pollutants, the results do not include the potential synergistic and nonsynergistic effects of exposure to pollutants other than tobacco smoke - other locations where there are factories or cities with substantial pollution. This additional exposure could increase the incidence of lung cancer. Those under age 30 years were not consid- ered in this study because the detrimental effects of active and passive smoking take many decades to produce additional diseases and earlier mortal, ity. Estimates of the exposure are based on close relatives of the deceased whenever possible in order to obtain the best estimates of exposure without having to wait for the many decades nec- essary for a completed prospective study when all of the members of the study have died. Analysis of the demographics of the popula- tion shows that 15% more working women smoked than nonsmoking women. In addition, smoking women had greater passive smoking exposure at home since 14% more of their husbands smoked than those of nonsmoking women. Therefore, smoking women received additional passive smok- ing exposure at work and at home compared with nonsmoking women, which could result in higher rates of lung cancer. A larger population base is needed to determine the approximate contribu- tion of this increased exposure and its potential to increase lung,cancer in both EX and NX women. A recent report by Janerich" has attributed a doubling of lung cancer for children exposed during childhood for 25 or more smoking years. This conclusion should be checked with analysis of data from other studies that deal with exposure during childhood and adolescence. The analysis of the data from this study shows that the comparison of nonsmoking women (com- bined NX and EX categories) with the SM cat- egory yields roughly the same pattern as that of nonsmoking men with smoking men - a 10 to I lung cancer ratio for the smoker vs. the non- smoker as reported in most studies. However, when the finer distinction is made between the nonsmoking-women groups (NX and EX), we obtain a 42 to I lung cancer ratio for the smoker compared with the totally nonexposed indicating that passive smoking may be worse than reported. These data attribute a higher incidence of lung cancer to both active and passive smoking. There- fore, if smoking and other environmental pollut- ants that are causative factors are eliminated, lung cancer may once again become a rare disease.t8 ACKNOWLEDGMENTS The author would like to thank the following individuals for their assistance in the preparation of this manuscript: Dr. Marvin A. Schneiderman of the National Research Council, Washington, D.C.; Dr. A. Judson Wells of the American Lung Association, Kennett Square, PA; Dr. Thomas Novotny of the Office of Smoking and Health, Atlanta, GA; Dr. James P. Campbell of Humana Hospital, St. Petersburg, FL; Dr. Lisa A. Nemec of the Women's Medical Center, St. Petersburg, FL; and Dr. Robert H. Depue, Jr., Washington, D.C. REFERENCES 1. Breslow L. Cumberland WG. Progress and objectives in cancer control: JAMA 1988:259(17):1690-1694. 2. American Cancer Society. Cancer facts & figures. Atlanta, GA: American Cancer Society. 1990: 9. 3. U.S. Department of Health, Education, and Welfare. Smoking and health : report of the advisory committee to the Surgeon General!of the Public Health Service. Washington. D.C.: Center for Disease Control, 1964. 4. Hammond EC. Smoking in relation to the death rates of one million men and women. In: Haenzel W. ed.. Epidemiological approaches to the study of cancer and other chronic diseases. National Cancer Institute Monograph 19. Washington, D.C.: U.S. Government Printing Office, 1966: 127. 5. Loewen GM; Romano CF. Lung cancer in women. J Psychoactive Drugs 1989; 21(3):319-321. 6. Brownson RC. The epidemiology of lung cancer in metroQolitan Denver. Doctoral dissertation, Colorado 428

-.ai .rc.cuiwa ona r.Y.ventloA State University. Fort Collins. Univ Microfilm Int 25. Benhamou S. Hill C. Epidemiologie du Cancer 1985: 85-172774. Bronchique. Rev Pratic 1983; 33(25):1279-1280: 7. Gao YT, Blot WJ. Zheng W, et al. Lung cancer among 1285-1286,1289-1291. Chinese women. Int J Cancer 1987:40(5):604-b09. 26. Fraumeni JF, Jr. Blot WJ. Lung Pleura. In: Schottenfeld 8. Koo LC, Ho JH-C, Saw D. Active and passive smok- D. Fraumeni JF, Jr, eds. Cancer epidemiology and ing among female lung cancer patients and controls in Hong Kong. I Exp C1in~Cancer Res 1983; 4(2):367- prevention. Philadelphia, PA: WB Saunders Co:, 1982: 564-582. 375. 27'. Hinds MW, Stemmetntann GN, Yang HY, et al. Dif- 9. Koo LC. Ho JH-C, Saw D. Is passive smoking an ferences in lung cancer risk from among Japanese. added nsk factor for lunQ cancer in Chinese women? I Exp Clin Cancer Res 1984, 33(:3):277-283. Chinese, and Hawaiianwomen in Hawaii. Int J Can- cer 1981; 27(3);297-302. 10. Lam TH. Kung I7?v1. Wong CM. et al. Smoking, 28. Mayo P. Saha SP, Jernigan CM. Lung cancer in passive smoking and histolbgical types in lung cancer in Hong Kong Chinese women. Br J Cancer 1987; women. A twenty-five year review. J KY Med Assoc 1982: 80(5):275-277. 56(5):673-678. 29. Osann KE. The epidemiology of lung cancer in women. 11. Woolf CR. Zamel' N. The respiratory effects of regu- Doctoral dissertatiort, University of California, Ber- lar cigarette smoking in women. A five-year prospec- tive study. Chest 1980; 78(5):707-7P3. 30. keley. Univ Microfilm Int 1983: 8.t-135443. Svensson C, Pershagen G, Klotninek J. Smoking in 12. Wu AH, Henderson BE, Pike MC. et al. Smoking and relation to lung cancer in women. Acta Oncol 1989: other risk factors for lung cancer in women. J Natl' 28(5)`623-629. Cancer Inst 1985: 7-t(4);747-751. 31. V utuc C. Kunze M. Zigaretten-rauchkondensat-expo- 13. Akiba S. Kato H. Blot WJ. Passive smoking and lung sition und berufsanamnese bei weiblichenlungenkreb- cancer among Japanese women. Cancer Res 1986; 46:4804-4807. 32. spatienten. Onkologie 1981; 1(S) r1 &3-187. Woodring JH, Stelling CB. Adenocarcinoma of the 14. Blot WJ. Fraumeni JF. Passive smoking and lung lung: a tumor with a changing pleomorphic character. cancer. I Natl Cancer Inst 1986: 77:993-1000. Am J Roentgenol 1983: 140(4):657--664. 15. Correa P. Pickle LW. Fonthan E. etal. Passive smok- 33. Abelin T. Current trends in the epidemiolbg,v of smok- 16. ing and ~ lung cancer. Lancet 1983: ii:595-597. Dalager NA. Pickle LW. Mason TJ. et al. The relation ing, passive smoking an& lung cancer. Schweiz Rundsch Med Prax 1989; 78(5):87-92. of passive smoking to lung cancer. Cancer Res 1986: 34. Darby SC,,Pike MC. Lung cancer and passive smok- .16: 4808-3811. ing. Predicted effects from a mathematical model'. for 17. Garfinkel L. Time trends in lung cancer mortality cigarette smoking and lung cancer. Br 1 Cancer 1988:. among nonsmokers and a note to passive smoking. J 58(6):825-831. Natl Cancer Inst 1981: 66:1061-1066. 35. Fidanza L. Franco G. Malamani'.T, Moscato G. Pas- 18. Garfinkel L, Auerbach 0. Joubert L. Involuntary sive smoking. A risk factor in the home environment. smoking and lung cancer: a case study: I Nktl Cancer G Ital Med Lav 1986; 8(5-6):233-240. Inst 1985: 75:-363-4b9: 36. Horton A W: Indoor tobacco smoke pollution. A ma- 19. Gillis CR. Hole DJ, Hawthorne VM, Boyle P. The jor risk factor for both breast and lung cancer? Cancer effect of environmental tobacco smoke in two urban 1988: 62(11):6-14. communities inrthe wesrof Scotlr{rtd. EurJ Respir Dis 1984; 65:121-126. 37. Repace JL, Lowery AH. Predicting the lung cancer risk of domestic passive smoking. Am Rev Respir Dis 20. Hirayama T. Cancer mortality in nonsmoking women 1987; 136(5):1308. with smoking husbands based on a large-scale cohort 38. Repace JL, Lowery AH. A quantitative estimate of 21. study in Japan. Prev Med 1984; 13:680-b90. Sandler DP, Everson RB, Wilcox AJ. Passive smok- nonsmokers' lung cancer risk from passive smoking. Environ Int 1985; 11:3-22. ing in adulthood and cancer risk. Am J Epiderniol 39. Tredaniel J, Hill C. Chastang C. Hirsch A. Passive 1985; 121:37-48. smoking and respiratory diseases. Current data. Rev 22. Trichopoulos D, Kalandidi A, Sparros L. Lung cancer Mal Respir 1989: 6(2):J09-120. and passive smoking: conclusion of the Greek study. 40. Vainio H. Is passive smoking increasing cancer Lancet 1983; ii`.677-678. risk? Scand J Work Environ Health 1987;,13(3):193- 23. Wald NJ. Nanchanal K, Thompson SG. Cuckle HS. 196. Does breathing other people's tobacco smoke cause 41. Miller GH. The Pennsylvania study on passive smok- lung cancer? Br Med7 1986; 293:1217-1222. ing. Ill Lung Assoc J Breathing 1978; 41(5):5-9.. 24. Wells AJ. An estimate of adult mortaliry in the United 42. Miller GH. Lung cancerr a comparison of incidence States from passive smoking. Environ Int 1988; 14:249-265. between Amish and non-Amish in Lancaster County. N J Indiana Med Assoc 1983; 76(2):121-124. C IV GO 429

I Womtn and Lwg Cancer 43. Miller GH. The impact of passive smoking: cancer deaths among nonsmoking women. Cancer Detect Prrv 1990; 14(4):497-503. 44. Miller GH. Cancer. passive smoking and nonemployed and employed wives. West1 MedI984;,140(4):632-635. 45. Coghlin J, Hammond SK. Gann PH. Development of epidemiological tools for measuring environmental tobacco smoke. Am J Epidemiol 1989; 130:689-704. 46. Friedman GD, Petitti DB; Bawol RD. Prevalence and correlates of passive smoking& krn J Public Health 1983;,73:401-405. 47. Janerich DT, Thompson WD. Varela LR. et al. Lung cancer and ezposure to tobacco smoke in the household. 1JFJ 1990; 323(90):632fi36: 48. Cassandra. P. In the 1800s anti'smoking was a burning issue. The Smithsonian 1989 July.. 430 ~