Tobacco Documents Online

PItEVENTIVE wEUICIr+{ 13, 680-69G ('1984) Cancer Mortality in Nonsmoking Women witK Smoking Husbands Based on a Large-Scale Cohort'Study in Japan' TAKESF~IiRAY91MA Narional Cancer Crnrrr Rrsrarrh' /nsrinlrr.`I-! Tswkiji S-cbnmr. Chaolu: T.L~u ll}t. lupun Mortality of 91.540 nonsmoking wives was studied in relation to the smoking habits of their husbands by means of a cohort study in Japan. Durina 16 years of foVloW-up. :00 deaths fiomJun` cancer took ptace.,The relative risks of lung cancer in these nonsmoking wives were I.00. 1.36. 1.4?. 1.58. and t.91 when husbands were nonsmokers. ex-smokers. or daily smokers of 1-14. 15-19. or 20 or mort: ciprettes daily.nspectively. Corresponding relative risks for stomach cancer were 1.00. 1.16. I.oO. t.00; and' 1.01. respecuvely Spec- ificity of association and internal consistencies were observed. Among cancers of each site. a similar tendency toward risk ekvation in nonsmoking wives with smoking husbands was observed for nasal sinus unctr- brain tumors. and cancer of all sites besides lung cancer. tn interpretin` these results. the significance of proximity in exposure to stdestream smoke in Japanese homes was stressed., c asw A&.kisw rn.... tm. INTRODUCTION The possible health hazard due to passive smoking was evaluated by the ob- servation of mortality in nonsmoking wives with smoking husbands. As reported previously (6), nonsmoking wives of heavy smokers had a significantly elevated' risk of lung cancer. Results of our large-scale cohort study reported here not only confirm the results of the previous report, but also reveal additional evidence of the health consequences of passive smoking by pointing out excess deaths due to cancer of other selected sites. MATERIALS AND METHODS A prospective cohort study on the health consequences of cigarette smoking has been in progress in Japan since the fall of 1965. In total. 265,113 adults (122.261 men and 142.857 womenl ages 40 years and above. 94.8t7c of the census population in the study area in 29 Health Center Districts in Japan. participated. They were interviewed from October I to December 31. 1965. and have been tracked by establishing a record linkage system between the risk factor records and death certificates. The 16-year follow-up results of this census-populatiQn-based cohort study were used as the materials for the study: RESULTS In a large-scale cohort study carried out im Japan from 1966 to 1981. non- smoking wives with smoking husbands were found to carry a significantly ele- I Presented at the Symposium "Medical F4npectivea on Passive Smokin`.-" April 9-12. 1964. Vienna. Austria. 690 0091F7at5tlit S3.00 , Cup.nWlil i1N+ rq A.:.,M~ Prr... tn. AM n{~~. ,I rtpuJ,'wiwn,n .w+ r,winrs.r...J .

SYMPOSIUM: MEDICAL PERSPECTIVES ON PASSIVE SMOKING 681 vated risk of lung cancer (n - 200): nasal sinus cancer (n = 28), brain tumors (n - 34). and cancer of all sites /n =''7051. Lung Cancer A total of 429 deaths from lung cancer in women was recorded during the 16 years of follow-up (1966-1981): Of these deaths. 303 occurred among nonsmokers and 200 among 91.340 nonsmoking married women whose husbands' smoking habits were known. Tbe standardized mortality ratios (SMRs) of lung cancer in nonsmoking women were 1.00. 1.36. 1.42. 1.58. and 1.91 when husbands were.nonsmokers,-cx- smokers, or daily smokers of 1=t4; 15-19, or,?Q:ortnore'eilp-'rtttes per,day. ... ~fpecttvNya(one-tait P, value; _~ 'Table~~) A -tt~milaawdose-tesponse relationship was observed by age and occupation of'the' ~iusband (Table 2): This tendency is in sharp contrast with that' of stomach cancer, where no re- lationship at all exists between the risk in nonsmoking wives and the amount of smoking by the husband (Tables 3 and 4. Figs. I and 2). Similar trends vf lung cancer risk elevation in nonsmoking women with~ the increase in the extent of the husband's smoking were observed in each time period of observation, in each age group. both by age of husbands and by age of wives: in each occupational group. and in most areas under observation (internal con- sistency) (Fig. 3) (7). No other characteristics of husbands or wives themselves were found to elevate the risk of lung cancer in their nonsmoking partners (7) (Table 5). ,:Nonsmoking husbands; with., stttoking,wives also showed- an.,elevated-risk- of lung eancer, the SMP.s being 1.00. 2.14. and 2.31 m nonsmoking wives, wives smoking 1-19 cigarattes, and wives,smoking 20 or more cigarettes daily. respec- tively'(P = 0.0177): This observation also strengthens the evidence listed above (Table 6). TA1qLE I Ll'MG CANCER MORTALRYIT..wOME*h' AOE GROL7 AMD,lYHCSRA•JDS' SMOKIPiG HARIT. (PATIENT HERSELF A NONTMOKERY Husbcnd s snwtm{ habt Husband's RR troup Woeunotcr Easmolc'cr 1-1f4d.y IS-19+day 3D•rd+y Ta+I 40-119 4 6:'9 1 1.111 0 9.6211 . 6 5.1.48 16 10.764 35 3-.017 l0-59 !0 7.791 3 1.9L, 20 9.66! { 4.052 24 ' 9.t:0 65 33,5) 60-69' It 7.t:0 11 2.697 29 7.243 9 2.513 23 4.651 09 24..•1R 70-79 - s 755 2 3+! 2 612 1 103 ti 226 Il 2.046 Taat 37 21.S9s 17 6:12 V 26.144 24 11" 64 Z+.c61 I0D 91 sQ ' 2 3! ' 0t 7t 2 Tbc .reighted porm 1 00 . t>< 36 / 11 . t ..~ ., 1 . t 9K euimtte of wu ratio . . ~O.RS . . . 01 ~196 1.34 Mantet ei<ten wn md test-0..ed 90% chi 2:915 eo.5eenct Fu+mtb ~, i3<2.02 One-uil ` 1.04 I v.luc 0.00176 Mantel-Hrenaeliciti 1. 0655 I.0?90 3.0?9S Oec-uil t value ~ 0.1389 0.0337 0.001. - Prospccttve stud>. 1966- 199 I. Jrpan.

682 T. HIRAYAMA TABLE 2 LUNG CANCER MORT~ALITY IN WOMEV !Y~ AGE'GROLta.JY~ OCCLMT/ON.AVD6Y HLSaAVDS* SMOKI•+G HADIT /PATIEvT. H.ERSELF ANO!+SLIOKERC" Husband's smoking habit Husband's Husband's occupation age group Nonsmoker Exsmoker C-191day 7.0+ /day Tolall AQicultural 40~49 I 2.30' 6 5.941 9 3.636 16 1::079 +voriter 50-59 4 3',a97 16 6112 9 3,314 29 13:RI3 60-69 13 4.064 33 6.945 10; 2.152 56 13,051 70- 3 323 1 446 01 09 4 858 Totul 21 10.406 56 20.044 2d 9.3911 105 39.8&01 OUier 40-49 3 3.r.7 9 9.093 7 7.126 19 19.908 30-59 6 4.294 15 4.930 15 6.306 36 19.430 60-69 3 3.0.16 15 ]:596 13' 2.499 33 11.133 70- 2 432 4 619 1 137' 7 1.188 Total 16 11.489 43 24.140 36 16.070 95 51.699 The Weijhtcd point 1 00 1.94 41 1 I 93< 2.74 Mantel extension estimate of rate ratio . . < 1.03 1.35 chl 3.145 and'test-based 907c confidence limits One-tail' P value 0.00083 Mantel-Haensul chi 1.766 3.053 One-tail P value 0.03705 0.001111 • Prospective swdy. 1966-1981. Japan. TABLE 3 STOMACH CANCER MOATiKLITY. 1-4 WOMEN'./Y AGE GROUtr AND OY'HUSaArDS' SMOKING HARIT (.PITIEYT HERSELF A NOVSMOKER!•.' Huabatd's tanoking habit Husband's age group Nonsmoker E%.moker 1-14..day tc-19'day '-0.. day . Tou{ 40-49 50-59 . 60-69 70-79 Tout 31 60 121 7 219 6:»"9 7.791 7.IJ0 755 21.895 1. 14 ?0 4 RO 1.:55 I-92: 2.68:7 3a! 6.212 44 R.6:1 02 9.Mtl' 10v 7:a? 11 612 :46 26.1." 23 5.1!R 36 4:0f_ 40 I:513 I 105 In0 CI.R:A 48 10.761 77 9.1120 78 4.631 6 ::6 :09 25.461 I!d 3_.0]' 269 33.233 396 24.214 :9 2.0r6 854 ' ' 91.V0 ~117 1 " 1.19 he weighted point 0.9J 1._._ ~O.tln 1.0n< 0.61 1.01< Oai6 >t/antel exlenbion 1,00 estimete of rate ratio and ust-based chi -0.:f0 OneaaJ 90% eonfdence limns P value 0.3'9J~0 Mantel-HaeRSZel chi' 1.059 -0.016 -0:0J3 0A91 1 Oee-wl P value - 0.14sta0 0.4936. 0 4M"41 0.46375 ' hospecuva atudy-1066-1901.Japan.

SYMPOSIUM: MEDICAL PERSPECTIVES ON PASSIVE SMOKING 683' TABLE 4 STOVICH GAVCER \1(1tTAWTS~ tS K'UME% ~ al~~ ABE GROLY. /1~.OCCCMTIOr~,.AKI)~l1~ HL~SYw~OS ~ S-MIKfsG H'ABtT~~IPATtENiT HERSELF A.N0%540AkRr` Husband's smokin` habit I Husband-s Husband's occupation a8e group Nbnsrnoker Easmoker 1-19,day 20-/day Total ABriculturali worker Total Other otal' 40-49 %0-~9 60-69 70 - 40-49 S0-S9 60-69 70- 121 37 77 3 130 18 23 44 4_, 89 :.30' y a97 4.014 323 110.406 3•7.7 4_'4t 3•0?6 432 1't.489 41 5.941 % 6.912 116 6.1145 13 446 226 20.044 38 9.093 76 8-830 V 3.."K 3 619 200 .4.Id0 25 3.636 37 3.514 I 43 2.152 3 89 108 9.341 21 7.128 40 6.306 35 2.499 3 137 101 15.070 79 130 236 19 464 79 139 16' 10 390 12.079 13.123 13.081 856 39.94 1 19.948 19.430 11.133 1.188 51.699 h Th i d i i 1 ~ 1 Oa < , 1 05 ~ Mantel extension e we te g po nt est mate ' . - 0.89 . 0.89 chi 0.234 of rate ratio and test-based 90nc confidence limits One-tail P value 0.40749 Mantel-Haenszel chi, One-tail P value 0.:98 0.486 0.38285 0.31?48 • Prospective uudy. 1966-1981. Japan. 7 S- - a- . .I :et3:1 * F.14. t.[.i. ... •. .s.... [[ •........r. Fte. I. Relative rick, of lung cancer and anmach cancer in v1..W nonsmokins wives by husbands- smokmf hybu, IProspecuve Sludy. IW,h-19u1. Jrpan.1

684 T. HIRAYAMA ft'iDol:'s Ow 111cuD 51, • 5 9 1. . .I ...... ~i i ~~'. .. ~ i ~ cO . ' ` ^~ ' .... 3 ' .3. . •. M.YM ,-4 V.1-,WN ,... • N li. Fto. 2. Mortality ratios for lung cancer and stomach cancer in eansmolung wives by husbands' smokin{ habiis.,IProspective Study. 1966-1481. Japan.l Naso!' Sinus Cancer A significant risk elevation of cancer of para nasal' sinuses in nonsmoking wives was observed according to the amount that husbands smoked, the SMRs being L00. 1.67, 2.02, and 2:SS' wheeshu'sba(tid*were nonsmokers' or.. smorters+'of 10'= 14, 1S 19;''or 20, or more cigarettes daily, respeOvely (P - 0:02482) (Table 7). No°otherrisk factors studied were identified as significantly altering the risk of nasal sim>rs cancer in women.. CY'~ 70! :f ~ nL'LD?'Cf~. P »:" .S SC•SS .~K.. w.1~x Sl • . ..1-1 lYDCC:OCtlon Of fVS00rCtS' 21 t 1 ~ 1F D !r ...• •.1,.... !Y ntrlo: of :lstrs,;IlK' ~.. ~~ i ` OK-ftT.:Craa:. Fto: 3. Mortality ratbs for lung cancer in nonsmokinjl wives by huabands• smokins habits. IPro- speetive Study. 19d6-1981.Japan./

SYMPOSIUM: MEDICAL PERSPECTIVES ON PASSIVE SMOKING TABLE 5 Ll'NG CAVCER MORTALIT. 1% NOV5VO/avG WO>.tE-4: RATtO aY SELECTEDRtSK F.ACTtMS' 685 Mortality rauo (relative riskt C'hi-sqprre values Husband's characteristics Smokin8 /20 ci8. -Idey1 1.911 9:1B Drinking 1.06 0.04 Population density: 600 1 1 600, 1.10 0.30 Women's chancteristics Occupation: A=riculture/others 0.95 0.17 Number of children: 0-3%4-9 1.09 0.48' Drinkin8: -/- 1.02 0.01 Meat: Daily.'others 1.12 0.09 Green.yelloN vepetable: Dail) others 0.68 0.93 Soybean paste soup: Daily/others 1.08 0.29 • Prospective study. 1966-1981. Japan. t Brain Tumors~ The risk of brain tumor was also observed to increase with an increase in the extent of husband's'. smoking habits, tfti#isk for:nonsmokiny women beinQ_1.00:' _ 3.03. 6.25, and 432,when•liusba'nd5;:wem°nonsmokers oc smokers of 1-1I. 15- ,~, .~,... ,~. _, 19, or 20 or more cla,arette~ly,~respecttvely (P° ~ 0.00376) (Table 8). Cancer oJA!'1 Sites- ln the case of cancers of all sites, a sirificant elevation in risk was observed. _ _.~._ the SMRs for nonsmoking women being.l::D0.1 12 "ind 1,23 when husbands were nonsmokers, ex-smokers. or smokers of 1-19 or 20 or more cigarettes daily. TABLE 6 LVNG CANCER.M.ORTALtTY1V. NoN9MORING HU9l+ANtK atWIV£S' SMOKWG HAl11T' t I b d's H Wife•s smokint habit us an age group Nonsmoker t-19/day 20+Jday Total 40-59 24 10.741 1 321 1 184 26 1l'46 60- 33 9.538 3 276 2 229 38 9.043 Total 57 1!9.279 *_ 397 -3 413 64 20?89 The vrei8hted point estimate 1 00 IK4•65 2 31 ~ 5.94 2 o( rate ratio and test.based . . 0.98 . `0.90 Mantel extension 90% confidence limits chi 1.989 1 00 2 `-'S •.Z' ~ OnauD . ~ 1.19 ' P value 0.02335 btanteli-Haenszel :chi' '-• 1'"6 Ooe-tail P value 0:0177 • Prospective stud). 1966-198'1. Japan

686 T. HIRAYAMA TABLE 7 NASAL SINiS~~CAVCER MORTALITY IN WOMEN 'BS ACE GROL~P AND Il~. HLSBkSDS* SMOK-O H.iytT~ /PATIEYT HERSELF A ~ N~~ONSMOKERC' Husband's Husband', snwktnr hakn alclroup Nonsmoker Ea'•unaker I-luday I!-19.d./ :0- dal Tolst 40-19 90-39 60-69 70-79 Total 0: 6.2I9 1 7:791 A 7.1:0 0' 733 3 21.895 0 0 0 0 0 1.2«1 1.6.1 1.922 3 9.668 2.687 3 7.243 34 0 612 6.:1*- 9 -16.1" IS:ISR •:10.7h.t 1 4.032 2 9.8,:0 2 2.313 6•.65 i 0 105 0 2--6 4 10 25:461 4 K.027 7 3?.2!3 17 :.. _ u 0 ..0.6 :8 91.340 '. The weqhted point I 00 67 t=0 1 ~6 33 ~s 6.?7 estmutee of rate _ - . ~0.67 : : 0 ''~0.6e '~ 1.04 Mamel esten,wn two aed test-based ch, . 1.90 90% confdcncc One•tail'. Ymits P value 0 020. t Mamel-Haensul cM' 0.916 1.012 1.713 one-uil P value - - 0.1783 0. 0571 004336 Nare. an contpuutton. aBes 60-69 and'70-79' trere combined. ' Prospective study...1966-19Bf~ Japan. respeaMtyel'y (P =r0.00020) (Table;9) Tbis,risk elevation is influenced by the elevattd risk of lungrcPeiand cancers of oEtisi selected sites such as nasal sinus catic,cr,.,Q~itiA:tuzttor,4nd'`posst'bly also breast cancer. Risk elevation for cancer of all sites becomes nonsignificant when these cancers are excluded. No significant association was observed with other cancers suchas those of the mouth. pharynx. esophagus. stomach. colon, reetum, liver. pancreas. perito- neum. cervix, ovary, urinary bladder, skin, bone. malignant lymphoma. or leu- TABLE 8 BRAI.r.TLyOR MORTALITY Iv WOyEV.BY AGE GROL'P AV0 BY HLSBA.rOS' SMOKItiG HABIT IPATIEVT HERSELF A NONSMOKERI" Husband's aleBroup 40-t9 l0-39 60-69 70-79 Total The vsIghted poant estimate of rate ratio and test-bsed 90% eoafidsece imtts Mantel-Hatos,r.el cAi 0ne-1tul1 value Husband's smotin` habit Nunstnoker Ex•snwker 1-1.6+day 13-19day 20-d6y Total 0 6.2.'9 0 I._33 I t.A:1 6 5.08 A 10!76.t III 32.07 1 7.791 0 1.9.: r 9.ti66 t i.0i_ i 9:R20 C 33 ZI3 1 ~ 7.1.0 0 2.Nt7 5 7.241 0 2!0 A 4.631 10 24.214 1' 75c 0 liB 0 61] 0 10 01 .2n 1 I.046 3 :I.B9S 0 6.21: 10 :6.1s1: 9 11.t:B I: 25.461 34 : 9I.3+0 ~1 !B , 1943 19 ~ 1'W - 3'Ol`' 07 6"SC 2.01 a'~ ./.33 Mantel ectensoott ch 2.673 One-tatl P ahue 0.00376 1.766 2:6s6 2,317 - - 0.039.44 0.00193 0.0100.3 Autr. la t:ompuutan: a/es~ 60-69 and 70-79 trete wattiMncd':. 'Prospective study:- 1966-19fi. Japan.

s SYMPOSIUM: MEDICAL PERSPECTIVES ON PASSIVE SMOKING 687 TABLE 9 ALL SITEs CAVCER MORTAUA. 1% wOMEN'.al ACiE. GROUGJIOCCl7ST1o\. ASD aY HLSt/1kNDS' S?AOKtv6 HANIT IP.ItIE%T HENSE'lfA.N0NSNOKERI" Husband's smoking habit Husband's occupation Husband's age group Nonsmoker Easmoker I-19day ?0-1dat Total Ariculturat worker Total Other . otal 40-49 30-59 60-69 70- 40-49 40-59 60-69 70- 40 96 20S 17 358 49 79 13: 17 _76 '."] 3.497 4.084' 323 10.406 3.7:7, .t_293 3106 432 11.489 9 119 5.941 ?01 6.1tI'- i7i 6.R-t5 2-1 446 715 :0:Os3 118 9:0y3 _4R' 8:11,'40 2?9 5.5" 21 619 626 ?4.1i0' 76 115 1']7 5 3'-? 10? 169' 1_9' 6 a07 3.636 21.514 2.13'_ 89 9.391 7.1I8 6.?06 2.499 137 16.070 ?35 412 705 4/ 1.396 269 496 300 44 1'.309 12.079 13.5_3 13.081 859 39.841 19.948' 19.430 11.133 1.1" S1'.699 The weiEfited point 1.00 2 1.1_ < 1.03 I°? < 1.I2 lr/antel eatension estimate of rate ratio chi 3.540 "4ti€ and test-based 90% confidence Gmits Oee-tail P value 0.00020 - Mantel-Haenszel chi 2.23. 3.6'8 One-tail P value 0_01:81 0.00014 • Prospective studt. 1966-1981., Japan. kemia. the direction of this trend being evenly distributed' to both the plus-side (risk increases with the extent of husband's smoking habit) and the minus-side (rtisk decreases with the extent of husband's smoking habit): OPSCUSSION This study confirms the correlation between lung cancer and spousal smoking reported previously. The correlation is quite specific in terms of diseases. For instance, no risk elevation at all was observed for stomach cancer. A striking internal consistency of association was also observed. The results were essen- tiallysimi3ar when observed in terms of age of husband's, age of wives, occupation of husbands, and differing periods of observation. The results are in Gne with a Greek study by Trichopoulos and others 00) and a U.S. study by Correa and' others (4) (external consistency). although they are slightly at variance with an American Cancer Society study in the United States (5) and a lxse-control' study conducted by Kabat and Wynder (8). Differences in proximity between husband and wife in daily life, room si2e, room ventilation, and' frequency of wives who work in offices in these countries are potentially influential factors in enhancing the extent of risk posed by hus- bands' smoking.

, 688 T. HIRAYAMA Histology of 21 cases of lung cancer in nonsmoking wives with smoking hus- bands was not essentially different from that of smoking women~ (adenocarci- noma. 57.1%: squamous cell carcinoma. 19.047~; and small-cell carcinoma. 4.89'r); The current results of elevated risk of nasal sinus cancer in addition to the risk of lung cancer mu strengthen the plaustbtlity.of carcinogenic hazards'of sidest ,._. .y t~N a io thr©ui 'li.~the eose as they arrin hne'`witf>"tlte°resUlti of measurements of various carcinogens in cidestiream, smoke showing them 'to' be preientiit•"higher'eoncerit6itons`ihan in rmainstrrani smoke f?. 3). These results are also compatible with,known evidence showing a possible influence of passive smoking on health including elevation of carboxyhemoglobin and nicotinoico- tinine levels in saliva, blood. and urine after exposure to passive smoking: ele- vation of hydroxyproline levels in urine (a marker of collagen destruction in lung tissue); the presence of mutagens in urine (1). small airway dysfunction, in those exposed daily to passive smoking in the workplace (11): and risk elevation for pneumonia, bronchitis. and asthma in children with smoking parent(s). When the effects of passive smoking due to husbands' smoking were compared with the effects of direct smoking in women, the results clearly indicated that the effect of passive smoking is less than one-fifth that of direct smoking. the SMRs being 1.55 and 3.81. respectively: In terms of attributable risk. however. the effect of passive smoking on lung cancer in women is nearly as important as that of direct smoking because the population of intrahousehold passive smokers at risk is four times greater (n - 69.645) than the population of active smokers (n - 17.366). Therefore. although the relative risk of indirect smoking is much smaller than that of direct smoking. the absolute excess deaths from lung cancer due to passive smoking may be quite important because of the large size of the exposed group-especially in countries such as Japan where the majority (nearly 70~7c) of adult men smoke. but only a minority (1S% or less) of adult women smoke. Passive smoking can be divided into (a) direct passive smoking (direct inhala- tion of sidestream smoke before being diluted by room air) and (b) indirect passive smoking (inhalation of room air polluted by sidestream smoke) according to thr extent of proximity effect. just as droplet infection is separated from~ droplet nuclei infection in acute respiratory communicable diseases: the effect of venti- lation is of limited importance in the former case. although quite significant in the latter. Smal[room size and congested living conditions in Japan (and possibly also in countries like Greece) are naturally more conducive to direct passive smoking. As described in a previous report. the age-adjusted mortality rates for lung cancer are increasing rapidly for both men and women in Japan. As only a fraction of Japanese women with lung cancer smoke cigarettes. the reasons why their mortality from 1!ung cancer nearly parallels that of men have been unclear. The current study attempts to explain at least a part of this long-standing riddle. Although the average rate of female smokers in Japan has remained fairly stable over, the past 20 years. a statistically significant increase in the mortality rate for lung cancer in nonsmoking women was observed in our long-term follow-up studyy of a Large-size population. Mortality rates per 100,000 for ages 50-59. 60-69. and 70 and above were 7.1. 17.7, and 31.01 in first 10 years of folJow=up and 9.9. 27.1. ~ N ~ C1't ~ N tA ?§4

SYMPOSIUM: MEDICAL PERSPECTIVES ON PASSIVE SMOKING 689 and 44.3 in the subsequent 6 years. respectively (P - 0.00373). This phenomenon can be interpreted as the influence of widespread exposure to passive smoking in Japan. As also emphasized in this earlier report, these observations strongly question the validity of the conventiona!' method of assessing the relative risk of developing lung cancer in smokers by comparing it with nonsmokers. This study shows that nonsmokers are definitely not a homogenous group and should be subdivided according to the extent of previous exposure to indirect or passive smoking. The observation of the effect of passive and: active smoking on lung cancer risk in men and women revealed a similar effect of both active and passive smoking on lung cancer when nonsmokers without exposure to intrahousehold passive smoking were used as the unit risk group (Fig. 4). The<ibbsecvatton ofan elevated risk of brain tumors in nonsmoking women with- tttno)~Ktisbands tis-of importance in considering the etiology of btam turttors Ca~iie °' wtitch'otiii ctirrcntknowle~ge is qu12e'lirntted~ ~speciallj~ttt>relstlon t~,# stcrwa ar`ep "ot~'orartbt~influence~of. passive srno{ring on'childhood brain tu= mors (9). tMWlt!'f OiIIK ~ R .rl~.~ w.,,t M~IT~ . -M~~• Ir1l~ ~~111~. ~11~) i~. M MiTw T7~ N N) 7~. IN at~ H N7 1Mf4171w *' ~ri ~MN~. ~`~M7f ~~iN~ M l CCwrIRKt ww..~~ !.N •.a 4.fl •.r 8 .w •.q I~tIt~K • 1. I.N t ri 7-~f~ 1-91 t-N~ • a~ 1.4 )N {H •~ wti.~ ~.«....~ r.. ~..•.. Ftc 4. Active and passive smoking and4ung cancer mortatity; Relativc risks IRRkwith 90% con- Eidence intervals. (Prospective Study. 1966-198I. Japan.)