Tobacco Documents Online

't Sobua*1, T., R. SuZUki*1, N. Rakayama*1, C. Inubuse*1, M. Matcuda*2, O. Doi*3, T. kori*4, K. Furuse*s, M. lukuoka*6, T. Yasumitsu*7, 0. Kuwabara*'8, M. Ichigaya*9, M. Rurata*l0, x. Ruwabara*S1, X. Nakahara*12, S. Endo*13, and s. Hattori*13: pA88SQ$ BMO]CIN4 AMONG NONSMOXZNO *OMEN AND '1'US RSLATION8HIP BETIFEEN IND80R AIR POLLDTIODt AND LVNG CJINCER INCIDENCE•-AE8DLT8 ot A!db'SrTSCEKTER ChsB COiZTROLLED STDDY. Can to Rinsho, Vol. 36, No. 3, pp.329-333. 1990. Introduction According to the 1987 population dynamics statistics eompiled my the Ministry of Health and Welfare, the age-adjustQd mortality of 1ung cancer in both men and women is the second highest (after stomach cancer) among cancer mortalities". The age-adjustad mortality in stomach cancer has been steadily declining since 1960, wwhile that of lung cancer has been rapidly increasing during the same pQriod. If this trend continues, the ranking of these mortalities will be reveraed by the year 20002) . *1 Survey Section, Osaka Prefectural Adult Disease Center. *2 Department of Internal Hadicine, osaka Prefectural Adult Disease Center. *3 Department of Surgery, Osaka Prefectural Adult Disease Center. *4 Department of Surgery, NatSonal Sanatorium xinki Central Hospital. *5 Department of Internal Medicine, Natianal Sanatorium Xinki Central Hospital. R6 Department of Internal Medicine, Osaka Prefectural Habikino mospital. *7 Department of Surgery, Osaka Prefectural Habikino Hospital. *8 Department of Znternal Medicine, National Sanatorium Toneyama Hospital. 69 Department of Respiratory Diseases, Osaka Red Cross Hospital +10 Departrent of Thoracic SurQery, Kitano Hospital. *11 Department of Respiratory Diseases, Kansai Denryoku liospital. *12 The First Department of Surgery, Osaka University School of Medicine. *13 The arganization to Eliminate Lung Cancer from Osaka.

For risk factors to expiain this rising incide-ce of lung cancer, smoking is cited first. In JApan, the incidence of smoking among men has been declining in recent years but it was reported to be 554 in 19883), a rat. considerably higher in comparison to the incidences in the western world. Thus the promotiost of an antismoking policy is considered to be necessary. hlthough the incidence of smoking has recently been increasing among young women, it was reported to be mere a f%3) in 1569, a figure much lower th~an those in western societies. To reflect this situation, the population-attributable risk of lung cancer due to smoking is ?1t in men and only 26% in vomen4). In other words, it is suggested that risk factors other than smoking exist in the development of lung cancer among women. =n the past, for risk factors other than smoking that cause lung cancer to develop, studies have been conducted with a major focus on occupational exposure. In women, however, it is necessary that studies on exposure to various elements in the everyday environment supersede the investigation of the effects of occupational factors. However the relative risk of these elements in everyday env'sronment is expected to be comparatively small. Thus it is necessary to include as many cases of lung cancer among nonsmoking women as possible for the analysis. The purpose of the present study is to elucidate the risk factors of lung cancer among nonsmoking women. Therefore the status of passive smoking and the relationship between indoor air pollution and the development of lung cancer were investigated, using the data from the case-controlled studies conducted jointly by a number of health facilities. It should be noted that the study uses data collected up to the tnd of March 14a9t thus the liqures may be different in the final section of the report which is scheduled for the coming year. 1. tubjects and Method. The "Organization to Eliminate Lung Cancer from osaka*I), with the participation of 8 major hospitals in Osaka specializing in the treatment of lung cancer, has been conducting a hospitnl.- based case-controlled study since January 1936. Among th. 2 2023512387

patients newly admitted with a diagnosis of primary lung cancer, 458 men and 193 women were registered by the end of March 1989. only 23 out of thes• 658 men (3.5%) were lifetime no~smokers. Among the 193 female patients, however, 120 (62%) were nonsaokers. The present analysis was limited to these nonsmoking female patients. Of the lung cancers in these nonsmoking women, 78% war• classified as adanocarcinoma. For control, $19 nonsmoking female patients were selected from those newly admitted to the same hospitals with a diagnosis other than lung cancer. Neoplastic diseases were the predominant diagnosis (8s4) of these patients. Breast cancer (240 cases) was the most common, followed by stomach cancer (63 cases). For the analysis, the ages criteria at the time of admission was set from 40 to 79 years for both the lung cancer patients and control. The information on smoking and indoor air pollution was obtained from a questionnaire, filled out by the patients, which was distributed and collected at the time ot admission. Any questionabl• responses concerning the present analysis were excluded from tabulation. Medical information such as histological type of cancer was obtained from attending physicians. The age-adjuated odds ratio and 95% eontidence level wera computed by the Mantel-Haenszel metbod using the PROC FREQ of SA56). Factors with an estimated value of the odds ratio over 1.5 or below 0.47 waas selected and subjected to a logistic regression analysis using PROC LOGIST of SASn. 2. Results Table 3 shows an ag.-adjustsd odds ratio where the population was divided into those with smoking husbands and other mQmbers of the lamily who were smokers. The tabla was prepared to find ths effects of passive smoking or lung cancer in adults (prior to hospital admission). The odds ratio when the husbands wers smokers was almost 1 (0.94), while smoking by other membQrs N in the household raised it to 1.4s, indicating a slight increase ~ in risk. Most of the other smoking members in a household were r .? children. ~ To find tha effects of passive smoking during early t~ ~ 3 ~

childhood, the age-adjusted odds ratios were computed when fathers, mothers, or other members of the household were smokers. Trie results are shown in Table 2. When fathers were smokers, the odds ratio of passive smoking was 0.60, with a significant reduction in risk. When mothers were smokers, the odds ratio was computed to be 1.71. The risk rose in this instance but the change was not significant. The odds ratio for smoking by other members of the families was computed to be 3.13. Table 3 shows the effects of using room beaters (which may be a eause of indoor air pollution) on the development of lung cancer as an age-adjusted odds ratio. The ratio was computed separately for each age category when the heater was in use. For heating facilities which may be the cause of indoor air pollution the following were included: unvanted stoves using gasoline, gas, ooal, charcoal briquettes, or wood; or a brazier, clay charcoal stove, or foot warmer which burns charcoal or briquettes. Air conditioners, stoves with vents, electric stoves, and electric foot warmars were excludQd from the study. Odds ratios were computed in relation to the use of these sources of pollution at each age level whsn the heaters were used. As shown in Table 3, the ratios were near I for each age level, showing no significant relationship. The effects of using foot warmers (burning charcoal, small briquettes, or round briquettes for the source of heat when sleepinq) on the incidence of lung cancer were studied and expressed as an age-adjusted odds ratio_in Table 4. These ratios, shown by age level, were near 1 for all ages, indicating no significant relationship. Wext, age-adjusted odds ratios were computed for the effects of using straw or wood for cooking on the development of lung cancer (Table 3). Again, the ratios were computed for each qroup. The ratio was 1.33 when the fuel was used at age 15 and 1.90 when used at age 30, with the latter showing a statistical significance. All the patients who had been usir,g straw or wood ~ for cooking at age 30 had also used the same fuel at age 15. ~+ Therefore the odds ratio computed at age 30 was interpreted to ~ 4 ra co d

express the effect ot long-term exposure to this source of pollution. None of the pa.tients was still using straw or wood as a source of cooking heat. In these analyses, the odds ratios for the f ollowing 3 situations were computed to be over 1.5 or less than 0.67: smoking by father or mot,her when the patient was young and the use of straw or wood as the source of cooking heat at age 30. Using these 3 situations as variables, the odds ratio was estimated using a logistic regression model (Table 6). It was found that only the use of straw or wood as the source of cooking heat at age 30 showed a statistical significance. 3. Discussion The results of the present study suggested that the use of straw or wood as a cooking fuel in the past is a risk factor in the development of lung cancer among nonsmoking women. Gao, et a1.9a investigated women in Shanghai and reported that the use of rapeseed oi2 raises the risk of lung cancer by 40%. In the same report, the use of cooking fuel (coal, city gas, and wood) did not increase the risk. Xoo, tt a1.a' conducted a study on women residing in Honq ?Coriq and reported that among cooking fuels, the use of petroleun increases thn risk of lung cancer wbile the risk is lower when propane gas (LPa) is used. HoWever they added that these effects are relatively insignificant. MacLennan, et al.10' conducted a study on Chinese woflen in Singapore and reported that there was no difference with respect to the risk of lung cancer when wood or charcoal used as cooking fuel was compared against petroleum and gas. The subjects of these studies were all Chinese vomen. xn addition, exposure was based on whether the subjects ever used the fuel in question at all in the past or whether they are currently using it. Such criteria say not necessarily reflect past exposurs accurately. 31urthermore there is a possibility that those who were classified as *not exposed^ may actually have beer, substantially exposed. Sf these possibilities are taken into oonsideration, the results of these studies do not necessarily ~ contradict ours. ~ 2ro subjocts currently use straw or wood for cooking fuel so cNp C11 s TJ ~ 0

we could not institute a policy of primary prevention of lung cancer in relation to this practice. However the finding is considered significant in suggesting that some factor(s) closely related to our daily lives may be recogr.ized as risk factor(s) for lung cancer that appears 10 or 20 years later. No statistical significance was obtained from the effect of o~r~ maternal s:~okinq when the sub jects wexe youngo estimated odds ratia .ras hiqh (1.79) and the power .to dttee~the '"'r"ta*.5..n....r.:._... ..,_~~s..,f.::.lG..V:elliaa~i"+L .iqinflicarit difference from the cases of the present study was computed to be 56i. There=ore further studies with a larger number of subjects are considered necessary. As for the short- term effect of maternal smoking on the health of children, Tager, et ttl.11} reported that the respiratory function was depressed in children when parents were smokers and the tendency was exaggerated when the mothers wera smokers. Wu, et 0.f2~ conducted a case-controlled study on white women in Los Angeles but they failed to find a significant relationship between lung cancer and maternal smoking. In Japan, Shimizu, at al. 1a) reported an odds ratio of 1.6 for maternal smoking in relation to lung cancer in women. In the present study, the risk of lung cancer from paternal smoking was significantly reduced in a single variate analysis but the results of a multivariate analysis were not significant. Compared with the mother, the time a child spends with her father is expected to be shortt and the tendency towards a decline in risk is believed to be due to some confounding factor (such as social class). For the effect of passive smoking during adulthood, the present study focused on smoking by husbands but no significant correlation with lung cancer was established. In Japan, HirayamaW , and Akiha, et all;) found a significnat relationship between the two but Shimizu, et a1.13) did not. Dlot, et al.16) conducted a meta analysis based on epidemiologieal data throughout the world and estimated that husbands' smoking raises wivest risk for lung cancer by 304. The 952 confidence range of the odds ratio in the present study is 0.62 te 1.40 and includes 6

1.30 within but a clearcut conclusion could not be drawn from such an uncertain risk factor. The present study also indicated that the effect of smoking by other members of the household, rather than by husbands,.tends to be rnore significant. Shimizu, et al.u~ reported that smoking by fathers-in-law who lived in the same households has a more significant effect than that by busbands. The analysis in the present study failed to show an increase in the risk of lung cancer in relation to the use of heating equipment. For the use of various heating devices, LeunqM reported that the use of petroieum stoves raised the risk of lung cancer among womQn in Hong KonQ. In Japan, however, Shimizu1s' conducted a case-controlled study Sn Nagoya in which no increase in the risk of lung cancer was recognized in association with the use of petroleum stoves. The most serious problem inherent in the r4thodology of the present study is a large number of cancer patients (especially those with breast cancer) included in the control. 9oth the cases and oontrol wera nonsmokers and there have been no reports on the relationship between the exposure factor--the subject of the present analysis--and cancer involving other organs. Therefore inclusion of a large number of cancer patients is not considered to present a serious problem. Kowevar dominance by a single clinical entity (breast cancer in this instance) is not desirable in view of the nature of the control. The authors plan further studies using a larger number of patients and an lmproved analysis of the control. The authors express their gratitude to Dr. Eanai (Osaka hdult Disease Center) who prepared the original form of the questionnaire for the present survey under the guidance of Dr. Muir (ZARC). 7

Table 1. Odds Ratio of aassive Smoking in Adultnood (belare Hospitali2mtion) l-ura CrincerjControl 011s Rati4 (95% Confidence Level) Smoking by Husband no 56/229 1.00 yes 64/200 0.94(0.62-1.40) Smoking by other Membars no 77/384 1.00 yes 43/135 1/45(0.94-2.23) * Adjusted by age at aospital admission. Table 2. Odds Ratio of Passive 6moking during Childhood iung CancQr/Control Qdds Ratio* (95% Confidence Level) 5moking by rather no 47['144 1.00 yes 73/375 0.60(0.40-0.91) Smoking by Mother no 102/473 1.00 yes 18/46 5.13~0.95-3.10) Smoking by Other Members no 93/416 1.00 yes 25/103 1.13(0.69-1.87) a Adj'usted by age at hospital admission. N c CA ~ N 8 CA GD' CGJ

Tcble 3. Odds Ratio When Heating Equipment Which May Se The Causd of Yndoor Air Pollution Ia IIsed--Observation in Relation to the Age When the Equipment Was Used u to "na Cencer1Contxol Odde Ratio* At Age 15 (954 Confidence IAve 1 ) not used 37/150 1.00 used 83/369 0.94(0.60-1.45) At Age 30 not used 45/212 1.00 used 75/307 1.09(0.72-1.65) At Present not used 65/289 2.00 used 55/230 3.07(0.71-1.60) * Adjusted by age at admission. Table 4. Odds Ratio When root Warmers Were Used Dttring S1Qep--Observation at Each Age Level Vse o! the Igy_iament Sjuna Car+cerfC r~ .trol Odde Retio (95% Confidence Level) At Age 1s not used ?6/3Z7 1.00 used 44/192 0.97(0.64-1.47) At Age 30 not used 95/429 1.00 used 25/90 0.89(0.53-1.51) At Present not used 119/514 1.00 used 1/5 0.670.09-4.99) • Adjusted by age at hospital admission. I~7 0 w cA ~ 9 ~ CIO ' ~Ph

Table S. Odds Ratio When Struw and Wood Aro Used for Cooking Fuel the t juea Qse oi' CancerJCor.t roi Odd Rat jnnc_ , , , _ ., _, At Ag4 15 (95% Confidence Level) not used 46/252 1.00 used 74/267 1.33(0.87-2.02) At l+Qs 3 0 not used 94/469 1.00 used 261 30 1.90(1.09-3.30) At Presrnt not used 123/'819 - used 0/0 - * Adjusted by age at hospital admission. Table 6. age-Adjusted Odds Ratios for Maternal and Patarnal Smoking during Childhood; and tbe Use of Straw and Wootd as a Cooking Fue2 at Age 30, Calculated by Log3stie Regression Anal sis Fac,toe y odds Ratios (9S4 Confidence Level) MattrZal Smoking During Childhood 1.82(0.98 - 3.37) Paternal SmokinQ During ChilcShood 0.70(0.43 - 1.16) Usa o= Straw and Wood as Cooking Fuel at Age 30 1.78(1.03 - 9.10) 10