Tobacco Documents Online

.., ~~~ i( (~'t' 5f' /OT "/ /( TH;1892' . pCA~NCE~R'~ RESEARi('kri46:.4e~04-4~l0~7~, Srptrmtxr ~t'9:a'61 Passive Smoking and Lung Cancer arnong J!apanes'e 1'1'omien Sum'inori Ak'iba,t' Hiroo haho, aind 1Villliarn J. Blot' F.Jd/ar,on'EQlQr Rrlra/rA~.FIDYndarlo4, HIMfA%ma;,fOran /S: A:, A',. X./, ar.dJb'ariona!'Cancrn/nri6wr„BrMrrda, k(a,)-lend'20d92./H:~/: d./' .A BSTFtAOT A cask-control stud) corsducted'in Hirosh'ima and Nagasaiki. Japan;, re.ealed rS'07 Incressed',risk of'hrng nneeraimong nonsrnohing womem whose husbands, smoled. The risks rcndcd to Increase with amount' smok'ed bl the busband, being b'ighesu rmong',wTotntn wbo worked outsiilt the home assd'w,hose hustiands were hnt3- smokers„ and'to decrease witib crossation of exposure. The fundings prot•3d'e Incentirt for further eralua- tion ofi the rteUtionship bcnween passtre smoking a'nd caiscer .mong, oonsmokers: I \rt.ODUCTIO'ti'. ` As'pan ofa case-epntroi invcsnigation o'f lung'rancer among atomic bomb surN•it-ors conducted primarily to a-a''luatt thee interactive roles of cigarettc smoking and ionizin'g radiation ('1);, d'atai ..-ere co'latct'ed on the smok'i'ng habits of ahc subjcct's spouses and'.parentsi Herein we reporti the c(rkrt of exp;osure to ;uch pa'ssive smoking, fhcusing on marricd women x•ho had ")cver sm'oked t'hemsclves. \tATFRIA'LS AND IvfETII'OD'S Since' 1951 a, cohort of 110,000 Hiiroshima and NaFasak'i atomic .omb survivors has' bocn,followtd by thr, RER'F,r'formorly ealled the ltomic Bomb C1sualty Commission (2). During,the p'eriod 1971 Io 980. 525 newJy' diagnoscd' cascs of primary lung cancer (Eighth ?c+ision ICD 162.1) were identified among cohon'mcmbcrs~: The cases .crc ascertainod from the Hiiroshima and Nagasaki Tumor and Tissuee tegistries, the RER'F'mortalit), surgical, and autopsy fltcs, and Hiro- r:irna University medic2I records. The diagnosis wa.s bascd on biopsy' r s!--~ical pathology'findings for 25%, on autopsl findings for 28':.v,,, n jIbgy for 4'Ss; and on radiolbgicol/cli'nical findings for the rr- -.'aining 43"- Since the cohort represcnts a fiQed po:pulation that'is ging oi~er time and is older than the gencraV population, the agKs at isgnosis were higher than usual for lung cancer in JSpan: the means cre 72.1 for males'and'70.2'for females; the raingcr wrre 36 to 94 for :alrs and 35 to95'for females: Controls were selected from among cohort members wiuhout lung, ancer, 2 for each cotse in Hiroshima and 3 for cach eas'c in tiagisa'kii hccontrols were indi.iduallymatched to the casca ..•ith respeet to yr f birth (± 2 lT), city of rcaidcnce ('Hirosh'ima'or Nali scx„and hethcr or not'thcy'were among the 20% of thc'eohort participating in ,e program of biennizli medical' examinations gi•en ra RERF. In {ditio,n, controls were matched'to eascs on vital status. Since most of .c cases h'ad' died„ most of the controlt were also deccasedi The. =ccasrd controls were chosen according, to the abo+-r-mr,ntiioncd .atching criteria; plurycar of'death (Y 3 ),r); rnd they w-crc' sd'caed om among,ailllnuses of death except cancer'and chronic respiratory scsst. The distribution of the'controls'serics:is as follows: alive, 13 % ; xcased from eercbroraseully discase, 2'6%; from coronary , heant dis- ~se, 13%; from other circulatory diteuc, 12'90; from acute respiratory Rrvri•tdI1OyZ/tS; nerisrdl4/24/g6; .cceptrd S/79/8'6. TAc ootts or publication of this artictr wtrr dtfnlirdln p,un,by the papmtnt pr2r charsrs. This andck mvt7 thortlorr be hcrtbD-murktd'ad.erriiror+rnt in -ord.nes ath Ia US;e, Section'.I73.4 toltltlo) indicult Ihiss f.cti "To ~hom rnavests for rrprinu'andlro'rrespondtncr ftom outsidF the United itrs thoutd. Ee adllressed, t Dt•p.ftment of' Ep.idemiolorfy and. St'atittiot;. .distion ERrcu Researoh Foundarion, S•2' ItijipamaCLk, Iltdosh`ima 730,.. pan (5. A.): a'nd from the United Statcs; at Epidemiology and Bfostatfui'es, otram4 Witional Cancerlhstitutt, L&mdo. Buildins 3'C16, Btthesdl, MD, disea.se, 9%; from digestive discase; 8°e; from accide'nls, 6%;',and from othcricauscs, 14%„ Interviews w•cre sought during, 1'983' with all cues and'eontrolsi or their next of kin, w•ho li,ed in Hiroshima and Nagasaki. The interview. crs were aware thit the studyconeerncd lung canccr, but thq' were not, told of the casc-control sua'tus of' thc study subjects. A structurcdl questionnaire Was us:cd to obtain historSes of ciguet¢e' smoking andl demographic, modieal, oecupational, and other factors. llf thcindividual: was marrncd; inquiry w•u made about the smoking stat'us of: thes'pou'sc„ including the average number of cigarettes smoked per day, age started smoking; and; for those W'ho stiopped, the age ofee'ssationofsmoking. Using this in'formation, together with Ihe numbers ofi yr the husban'd and wift li.+ed'togcther, an index o',f exposure to:tlie spouse's'smoking w•as ealculated. In ad'dition, a single question was asked regarding whether the subjvct's mothon and/or f'athor smoked when the subject was living at'home as rchild. OR were calculated as measures of the association between lung cancer and pu'sivcsmoking and othcr factors ('3I): Estim'aacs oftiheOR, and'corrMponding significance tests, were o;btiined by alcontiitional' logistic regression analysis for matchcd data ('4): Test's for trend uscd consecutive integers for Ie+xls of the ordercd'eatiegorics. Beeause there were, a yriori hypotheses that passive smoking might incrcasc lung nncerrisk, aJlisignificancc tcsts for passive smoking cfflrcts'.were one- sidcd+ with 90' C1 used' for intcnal estimatcs of the OR. Because interest foeused on s'pousc smoking patrtorns: elim;inatrd from thc, anal)•ses'+YCre the one eavc and,6 controls amon,,malca andithe 4 euK,s susd' 7 eontrols' among' females who wcrc never mi Among the married indit•iduaJr, almost aJl had been marricd'to onl)', one'spousc. Among those wi'th more th3n one spousc, inform'ation was aN ailab)c onlj for the mos't'rccent: Also oxcludc,d from cachltable were individuals witih missing data for thc t-ariabl'c bcing studicd. R ES ULTS. 1'nteeicx•s N•crcobtained for 428 cases and 9'57 controls,, respertii+•cly„ 8'I m and 82% of the eligible cases and controls. The two p'rim'ary rcas'ons for nonres;ponso werc the refusal of next of kin to answer questions about their deceased relatives and the decision not to attempt to IOcatic nexbofkin fors'ubj'cct's N•ho had m'o*•cd'out of]-llirosh'imalor N'agasa'kii The distribution of informants is gis•en in Table I, indica!ti'ng that the informa tion for most of the subjrcts N•as provided by'ncxi of kin. The type of'respondent, ho..•cvcr, w•as sivnilar for cases and con'trols.. Table 2 shows the lung,canczr OiR' according to the s!mokinig status (smoh'cr t•rrsus never smoked) of the subjects and thcitri spouses. In both s'e,xes the,re was an increased Iwng cancer risk associated with d6rect'sm'oking. As indicaitcd, almost all ('93ro), of the mailc lung cancer cases N cre smokrrs„but onlh• a minority (38%) of the wo'micrn with lung can!ccr in this population were reported to have ever smoked: Altihough not shown, tlrt OR inci with the numb'ers o'f cigarettcs usualQy smokcdi per day during adulthood for both men and N•o,men: Among mahcs who smoked 1 to 9, 110 to 19. 20 to 29„and 30+ cigarettes per day, the OiR were 1.7, 1.8, 3.4„and 9:7„respcctir,•rly (P'for trend <0:01). Among females who smokcd 1 to 9, 110 to 19, and 20+ cigarettes per day,,thc OR w crc 1.9, 2.0, and 4.9 (P for trond''< 0,0i1): 7lable 1'shos•s that among female nonsmokers marricd to smok'ers, therc was an clrratr'd risk for lung cancer (OR a I S"90:"t f"Ii - I 0{n_ 7 S' P () l171 _• Ir~ L, ••t • r

Tab1e 1~ frmrnrarr dinr;Awioa of nrpndrnrr St> of ttudy cubjret ('7G) . M ak Femak ltespondest Case Control Cax Controll Self' 6 7 16 19 sponse SI 41' 12 11 C411d' 23 22' 33 33 PJauthtcr-Iblaw 11 12 111 1'7 Others 110 110 271 19 Tota1 9: 100 ]01) 100 100 . rr 264' 595 : 164 3.62 Tabk 2 Od+it ruiaeJwlrnl curcrr arcordint,to smckiny uarrat oftArsnbXct rnd. Aro/Arrspo>lac Sex of tubjtet Subjtct smokrr Spouse rrnoker Case Control OR" 90% Cl' Ma1e Nob No, ' 16' 1ID'J'. l.0' Yes 3' 9 1.1 (0.5.5.6) Yrs No 190 36E' 3.4 (2.1.5'•SD Yes SI 96 4.2 (2.4, 7.3) Femak No, No, 21 !2' 1ltY Yes 73 1ds 1i5 (J.O, 15) Yn No li 14 212' (0.9, 5:1). Yes 50 36' )i6 (2:1„6.1) "Odds mtio and 90'.. Cl from matoh'ed anrllzu. ' Indi.idual reported emr to ba.e unokcd citirettez ' Referent oteaory. Tabk 3' Odttt raribs/orbrnir rae+Crr omonr noumokinr wo+nrrn ornordint to ,lufband's rtrai dailf- coiuvrnption of rirorrnra No. of cfgartate3's busband usuallry smoked/dar ayr ontrol R"' 0% Cl• 0 21 L2 1.01 1-19 29 90 1.3 (0.7, 2.3) 20-29' 2'2 Sa 1.3 (O.b; 2:bp 30+ )2 23 2.1 (0:7; 2!Sj "Odds ratio and 90% C1 from matclicd analysit- (r fon trrnd - 0.06) - Table 4 Odd1 rotios Jor bn j, eurrrr mont w+ona~,Unt .omex orroi to ArsDbnd'a di oftmo,tinr.t citnrrrrro whilr n+arrud. Yr busDand trooked eit.ncttes: Case Conuol OR' 90'7VCl' 0 21 32 1.0 1I-19 20 30 211 (t.0;,4.3) 2tD-39 29 111 1.5' (t):1,,2.7). 40+ 22' 59 1.3' (0;7„2.5) ' Odds ratio and 9" 'C7 from matchcd analnit. cient datai for detailed anaJyscs of passir,e smoking patterns were aNailable only'for female nonsmokers. The data for nonsmoking w•o'mle'niare e2lteg,orized in T4ble 3 accordinlg to the number of' cigarottes the husbandl usua111y smoked per day during adulthood. There was an increasing, lung,caneer risk Mithiincreasing a'maunt smoked per day by the husband, with the OR slightly exceeding, 2-fold for women w-hose husbands were heavy smokers. No monotone trend'of increasing risk associatedlwith increasing duration of cxposure to husband's smoking was found (Table 4)j Risks according to time of exposure are examined in Table S'. The odds ralios were lower among `ex-passive smok'ers" than a~m'ong womeni who had been cxposed to their husbands' smoking x•ithin thc, past 10 yr. The rcd'ut:tion in risk with cess3tion olexposurc remained aftor adjus'ting for the amount of cigarettks'smoke,d per day by, thr spouse. Tabk 5 Oddr nriot/orlrnr z.ncrr.monr noruntoklnt womzw .ccordinj 10' rrrrnty o/raposrrr ro A'trtD.wi tmoA'Vel T1me of erpoturc Cau Control OR' 90x' Cl• None 21: 12 ).0 N1ot nposad'within 1110 yrs 3111 37 , 13 (0.9, 2.4) Fipoxd'within last 110 yr 40 ts 1.11 ( L0, 3.2) (J for trend - O:OSD ' Odds rt.tio and 910% C1 from matchrd analysis. a TTiese,'ca-passi.t tmokrrs' arc tbrnr rhosc b'usbi QuhlsmokicZ 10 c: ttwre yr pnorto the diarnosit o(lunj tsooer(or 10 ot'mone yr primr to the datt ofukrtioa for conuols)'.orahosc,who+•crc not living .viih their buutivod3 becausc of separationr, dirorce, or hu dutb 10 or a+orr yr prinr to the dtasnosit. Table 6 Oddi rniiot o/)rnr rmwrrr amont norvnwRiny,.wmrw i.ororrfinp to nArir otrwj!ntion and tbri AxsAandi' unoAiar sraticr Oavpition of subjpcrt Husb.nd't tmokint tutui ' Canr Control ORt 90f.' C1' Houxwifr` Ncmr li=h2 6 11 20 34 1.V 0.9 (0.4; 2.1) Hcavy 1'3 35 a.S (0.7;3ll) Wbite eollir' Never 7 23 1.0 (0:4i 2.R) Lijht 9 20 1.7 (0.7;,4:1). Hcavy al 16 1.6 (0:6i 4.1) Bluc colltrf Ncn-or 6 21 1.1 (0:4, 2.9) 1;i=1it-- " S 22' 0..5 (0.2, 1 .5) Hiu•y, 7 6 10.4 (1;6; 66:7) "LitAt, husband unoked kir tban 20 eigarcnrr/d11; heavy, husband smokedl 20 ormore ci;areucs/dary., t Odda rotio.nd 90 : Cl from machod an.lrsir: rHousewife defuned at womm who w.s employed outsidc tlie bomc lor no more tStun 10 yr. ~ Rcforcncc otrpor7. •Offuce and ulef ..orhzrs& fE'acludeiS c.ses andi3:4 convoli ..ho.•arr farmtn, 4803 As shown in Table 6', tho risk or lu'ng cancer tcnded to increase in rclartion tlo exposure to the husband's tobacco smoke for each or house++ir:'e's, whi'te collar;,and blue collar workers.']lse h ighest' odds ratio occurredl for wome'n who had blue col0ar'jo'bs and'. were mafricd to men who smokcd one or more packsWcigx- rettes:per dl3y„bull the numbers involved were sma.d)i The odds ratios from the rnanchedllogislic regression analyses presented in Tables 2', to 6 are gcnierall9•, similar to unadjusted odds ratios that ean be calculated from ohe, cross-products or the numbers ofl exposed anduncxpo's'edl eascsarvd'i controls, indic2:ting that cnnfounding in unadjusted analyses by'age, city, vit'ali startus, and yr of death (the matching factors) is not substantial. We also assessed wshether the asscxiations with passive smoking were consi'stent across the various stra'la de_ fined by the matchlinlg factors. The nlumbers oCsvbjcct's in several of the caiegories bccame: quite small with this fine a aoss-elassifieation„ but the trends with husbands' smo,kiing tended to bc seen throughoul, witlhino striong differcnces by age group or by city of residencc. The trend's wt:re alls0 a!Dparcnt for eachi t)'po of informanit (self, husband, child, and oth€r); in pantilcu)'ari the elt:>•atc'd risk for heavy rclatih'e to nonexposure to husbandfr'smoking,u•as.de1t:ctc;d'when dartaiwcrrreported bg', the husbands or subjects themselves. Radiation exposure was also examined as a potential confoundcr and effect modifier. No significant inAucnce of radiatio'mdose:on the passive smok- ing association was dctecled, alahough, the trends Mith passive smoking seemed strongrr among, the unexposed. Inform'ation, on the histohogical types or lung cancer was unaw'ailablir for 43e.?;: of the cases who were diagnosed only on radinlpgical or clinica) evidence. We conducted separa'tr anal- ySes among those,wiilhiand without a pathological confirmation of lung ca!ncer and found increasrd risks asstxiatrd %&itih passivc smo'king for both groups.,Thc Oi2' among nonsmol.'ing M,omcn

rnarried to smokers was 1.4 for the cases a'nd their matched eAntratls brith a histolo'gically confurmed'diagnosis, and 1.6 for' ' those with r clinicah/radiological diagnosis. Among women withia histological diagnosis, adenocarcinoma was the predom- inant cell type„but'the distribution of histological types varied by smoking status (Table 7). The percentage of squamous and small cell carcinoma was much higher among smokers than nonsmokers, Although based on s!mall'numbcn, there were al5o more squamous and small cell canccrs among nonsmoking females whose husbands smokedi Responses to the question on p'arcntali smoking while the subject was a child were provided for only rwo-thirds of the subjects. Among these the mothers of the svbjccts were reported to be smokers for 13% of the cases and 17 % of the controls, and the fdthers, for 67% of the cases and 66% o of the controls. Hence there was no overall increascd risk associatedi with parental smoking, nor was there any significant increase after strati'fying,by smoking staitus of the subject. A'mong,maJe smok,- ers, the OIR' for lung cancer associated Mith maternal smo'king, was 1.11. DI'-SCUSSI ON The results from this casr-control study suggest that there mi be a moderate exIcess in lung cancer risk' associatkd' with p ive smoking. The odds ratios for lung cancer among non- smoking women tended to increase with, amount smoked by their husbands, a trend seen among houscK'iNes a-s well as women who worked outside the home. The highest odds ratios among,nonsmokers were for women who worikcd in blue collar jobs whose husbands were heavy smokers, women presumably with the highest exposure to environmental tobacco smoke. There was little association with parental smoking or with e'x- passive smoking, suggesting that cessation of' exposure may lower risk. The findings are generally consistent with results ofa na'tional cohort study of mortality among Japanese women (S): and of several epidemiological investiga!tions conducoed eJsewhcre in the world (6-8). Updated follow-up for the period 1966 to 198'1 of the study conductedl among an adult' population sclectcd from multiple areas throughout Japan, excluding Hiroshima and Nagasaki„ showed a gradient in mort'a'lity .vit'h amounn smoked by the husband (9). The increase in risk reached 90% among those whoscfiusbands smoked 20 or morc eigarc'ttes per da a figpre in line with the 2-fold cxccss for 30 or more cih_.ettcs per day of smokers in our study. The similarity in results, despite difiTcrent methodological a!pproachcs, suggests that the association between lung cancer and passive smoking is not an artifact of recall bias which can affect retrospective studies. Furthermore, we were unable to identify any.'strong confounding factors, including radiation exposure, that may have accounted for the passive smoking association. Iit is noteworthy that a recent surycy i:n Kyo',Io, Jypan, found significantlytJe••ated levels or cotinine, the major metabo'lite of nicotine, in the early morning urine of nonsmokers who lived in hnuseholds witih smokers or workcd'in officcs/factorics with T.b'le 7 . Itrrrnratr.Autororkdd'urriDvrion o/Irnrrrr.crnamonr/rrnalts accrorCiAtto tArdl end IAtir Asu6andj' s,no.6inr,,vmsu Ccll ry,yr (G) Sau.mous or Adenncalcinoms SvKjec1 Husbandd small crlll or lurtcttl unokt:r smoltr canctr cannn - No Mo 0 100 Ycs Ii6 84 Yrs Y1t 4: - smokers (10). The cotinine concentrations among,nonsmol' living with 2-pack-a-ddy smokers were roughly equivatcn= the eotinine levels'of'smokers of less than 3 cigarettes pert Precise estimates of'the lung, cancer risk associated w.ith level of smoking are not available, since not many smokce few cigarettcs per day. However, 3' well-known prosFe: stnd'ies of mortaJity among, simokers [the American C:: Socictys!tudy involving nearly I milllion volunteers ('11)i,t`: yr fo'llow-up oI' 2S0,0b0: Uniittd, St-Litos veterans (12)„ an.: 2@-3r follow-up of 34,000 British doctors (13')) found rri :' risks of lung,cancorof 4.6, 4'.8'I and 7.8 among 1 to 9, 1;: and 1' to 14' eigarrtte•pcr-day smokers, respectively. L:-- interpolnition bc'tweemthrsr values and the baae-line level c; for nonsmokers would'yieldiestimated relative risks for 1 n cigarette-per-day smokers of nearly 2-fold, about the same cr• of increase obscrvcd fior 'hcaVy"passive smokers in this s:-j Hence, if the Kyoto results (10) are applicable elsewhcrc,' .: if urinary cotinine levrls'reflkct levels ofexposure to thec-~rc ogcnic substances in tobacco smoke, then the observed mc; tude of the increased lung cancer risk among, passive sm: '`.- i'n, Japan seems not' grcatily out of line with w.hat' migh; expected based on their exposure to enn•ironmental to.`'•ac smoke. It should be noted that' the risk ranios for lung cancer eieited with directsmoking (as shown in Table 2) were 1oMe: this casc-control study than typicallyy found in case-contvol n. cohort investigations in other counitmes ('14). The lower C among smokers in part a.mses from ourselection, in ordc: minimiize respondent',b,ias; of controls matched to cases on Nii status„whiich led, to the inclusion of some controls who die:'' smoking-relalkdldiseascs. However, lung cancer risk ndos S_ enally similar to those in this study were also reported prospective study of Japanesr adults (9): Bccause or the Ic•• relative risks of lung cancer among smokers in Juipan, di::c rnccs in the OR'betwce'n direct and passive sm'okc'rs'arc noc high as in western countries. Indecd1 we found OiR' for "'i ' , passive smokers to be nearly equal those for µvrncn who ~, c: reportedito be light smokers themselves. While such similzrih was uncxpocted,, characteristics such as the size and styic • residcnrtial' units might result in a higher cnnironmental-t- d'trect',tobacco smokc exposure ratio in Japan (and thus lic'ss < a difference in' OIR, for lung,cancer bc'twcen passive and di: c: smokers). This in fact is suggested by the comparison of t'` couiniine analyses behwec'n Japan and Grcat! Briuaini (10; i: where the ratio of cotinine,lcvels in p,assivic compared to dircc smokers was considerably higher in Japan. Our findiing t'`.=, lung cancer risk among nonsmokers may be Ikss,closel7• relh"c tio duration of exposure to tobacco sr.:oke„the majordctcrw•~' nanit oflung eancerrisk:among,smoktrs (13);,than to inticn,i!an&reccncy of exposvrc,also may be notrworthy;,Suchia duf?cr ence might contributc to a higher ratio inJapan of lung cainca: risks in passive comp'arcdi to direct smokers, since the currc:• p,revalence or smoking, is higher in Japan than in ciither Grc a• Britain or the United'Stabes, but the mi temporal increa_:~: in smoking bespn later (9, 16).. The, present s'tudly did not replicate the finding of a c:,~tc- control!study in Louisiana whic'his'howed'a higher risk rmor•:. malk smokers whcse mothershad'smokcd (7). Although we dir: find hiigher percentages of smokers among both cases a!r.t: controls and among both menian'd'women whose parents ha.: been smokers, there was no clevation i!n the OR among,smokir•t~ 4806 ' Thcrt it somt qutstion about Ihnir atnrrwliubiliry, sinct cotininr Ic•r6+ vi hts.7pasai.ey vmokrrqin Kyoto wtrt about ont-sc•rnthh dit /tstJsir rnorstesmoltrn,.in eontrnt.to about onr.riliicth in a rncent fflritii,h stud) (IS'.I ln,botkstudics; To-r•rr, tliaurinar~ lt.oll.incrrxstd inproponionto.rs~tim3trc I+a~ssi.t smokinr.t.pos,urr, ~ itOt "' !~Eir:,(a re t~+° tr~.t~iil

PASSIVE S'h1o}:JNG AND LUNG CA'rJCER' AMONG JAPA'NESE wOMF.N' Jalpanese men or, women associa'ted with ma'tcrn'al'or patcrnall smoking. However„it was often difficult for the respondents to provide informa'tion on parent'all srno:king, andl data on this exposure were missing for about one-third of the subjects. One of the con'cenns in this study Was the adcqpacy of data provided by' surrogate respondents. O'nl?, a minority of the patients could be interviewed directly bccau'se of the often fatad outcome of'lung ean'cer and the ncedlto includc cascs d'iagnoscd as early as 19711' in order to assembl'c sufficient numbers of subjects for analysis. The distriburtilon of respondent types was comparable betweenicascs and controls so that response bias is unlikely, but the possibility of poor quality information for'both easesand controls exist'ed. We could evaluate this possibility, however, since many or the cases and controls had provided inforirla'tion on their smoking habits in routine RERF surveys conducted' in the 1960s when all study subjects were aliive ('ll,. 2): The data in Table 8 ind'ocate very hig'h, concordance in the identi'fication of a flemale as a nonsmoker or smoker by a next o'f' kin in 1982 and by the ..•oman herself in the; 1960s. In addition to proridin'g,some confidence that the da(a protiided by surrogates are adequate; the confirmation of nonsmoking s Js by a next of kin argucs against the possibilit'yy th3t' J'apamcst..•omen ten',dltlo report'thit',m'selvcs as non'srnl)kcrs µ'hcn they actuallly smoke. The 198i2 survey revcalbd a higb'er per- ccntagc of male smokers than reponed carli'cr, but the increase -was: both for self as well, as next-of-kin ilnbcnic>vs and may reflect an alctuall increase in smoking prevalence over time. Questions about the smoking habits of spouses.vcrc not asked in the surveys in thc'11960s, so that'sclf versus surrogate repon- ing o'nI this variable cannot be assessed directly. In our s!tu,dy, hoµ•cvcr,, there were no significant difTercnccs in the passive smoking trends according to responde'nt tyqe. In particular, an increased OR was seen for nonsmoking womcn whos€:hvsbands were hcasy smokers w.hen the data M•crc reporticd' by the hus- bands themselves. Another concern in this ease{ontrol study was the reliability of the diagnoses of lung cancer. Forty-three o of the cascswcrr diagnosed solely on cli:'nical'' an'd;ror radiological csidFncc. The pereentagc' .vas high in largp pan bccause the, cohort being followed was eldcrly, and surgical' or biopsy procedures %t crc less likely , to be:performedlo',n'oldct pation'ts ThrOF? a'ssociaucd r passive smoking, however, ware similar %`'hen calculations were restnct'cdi to histologically confirmed cascs. We also ca'I- cullaticd OR after deleting 23!easos and their matched controls for ,.ho'm a'diagn'osis of possible or probable lung cancer titis made only on radiological grounds and who had sun•ivrd 5 or more yr (allll.vere in fact living as of January 198a)', since the dialgnoses for at 1t:ast some apprar to be questionable. Little change N'as nlotied. Snloking has ticcn shown tio i'nducc'adl types T.blb tl: Comhefuoe cf snwkinrrrarvs Jrum 1a1r1 i9'd '1'oasn<onrrol.fruC) and Jt£!tF asrrveyi iw 1961 to 1963 T7ir numtxrsof paired respomas.for shr, wa.infcrmnt cyrrtorirn.Krlo+ vo 58'. 6J9„t5•,u+d 92; rnpexti.elr. Sr:s of Inforsnanu 196.t-1961 currcM 1'9I8'.2~~ smoklint s.taut,(T.) subjrct in19i:1 Imokcr. 1r•cr Smoker M.k; Sclf No IS 14 Yn 0' 63' 4urrogutr ~ No 1 2 1 3 '. 1'•ts 1 ~ 74 F'~emu/r Sel(' No 1'rt I17 0~. Surnntptr~ N'ce 6t~ 3~ hts 0 of lung cancer, but its cffect' is greater for squamous and snr•ai cell carcinoma than, adc,nocarcinoma' (;1,7); WhetAer pass:°- smoking might have the same predilcction for saquamous c::r cers is not clear, but our limited histological data (Table 7) :: consistent with this notion. llt is of interest that the hlighest i for passivcsmoking has bcen'rcponedlfrom a easc-control se: : in Greece ('6; 18, 19) Mhcrc the eisrs were limited to 1.::. cancers other than adcnocarcinonia. In summary, the results of this investigation suggest :.. exposure to environmenra'l tobacco smoke may i'ncrea'sc risk of lung cancer among no'nsmokcrs.,Thc f nding's, from ,: of the tµ•o arcas of the world where the possibility of a pi:._ smoking hazardl.+as first' postulated, add to alm acicumula, : body of cvidcncc on the issue. While the total evidence is . definitive and not all studies show signifean'tl9', positive ciaaions (20-22), the results are suggestive enough to ftlrther caaluauion in )argcr, sstudies ~,he're passive smo".: cx:posurlescan be mo'reflrlly,quan'tifiled. AC}:1O\k'LEDGh1E1TS N',c thank Dr. Rotxrn Htoovcr and! Dr. Joscph Fr.umcni, helpful s'uaostions; Dr. B: J. Stone and D:r: 125, Lubin for ad•icc . computer i'ssistarscz, and Thcre" Pino and Michcdc Rlzsa for rr scripl preparatlon,, R;E F E REti'CES 1. Blot. 1\t. l.. A'kita_ S., and Y.aso, }H. lo,oix.iat.noi.tiom..nd lunr ca-...rr•ibw includintg prrliminnr7cnulu from n c,u¢<c.owoll studl, rmo-..b tamb suni+on. lic R.. Prcnticr s.nd D. Tlhompscn (rdi.), Atornr. ,. Suni•orDstn; pp: 276-248. Rhilldrlphis:SLAht, 1!9I841 2. Ekr1s.,G. W.. Kito, H, and Lknd, C. E. Studies of thc morutih of A-. s.uno•ors:.Radist. R'n., !'d:.613-6a9, 1197.I: 3,, Bhcslo-, N. E.. .nd Dsy„N:,E. TTr anul1nis of cascronuol vtud:cs. :. (dns. ALrncy, Rcs: C.nc:r) Sci! Publ:, 32: 1-:51, l!9;SO. t~ Lubin, J. Aoomputorprognm forthranaljzu of m.tchrd ciuc-<~ ' stud;rs. Comput. Biomed. Rrt'.• lI 138-143, 19311. 5. }tuspama, Ti. t`'On-s'mokint:wires ofhca.3smok,cnt:hsnc a hiphcr- lunt cancor.,a stud) from Japan. Dr. Ntkd; J,., 232: ' 183-185. 1'98:1. 6. T.nchopovlos, D.., );adandidi,.A_, parros,. .,vnd:+tacFtahon, B:. Lung r. -- and'passi.c s'mokint, Int. J. Csncrr6 1:r.1-J„1981:. ,I Corrr.• P., PKkdc; L. „ ont om4 ,E.. Lin, 1'.,.and Ilarc,ntacl, w'. P:-srno6int and Iung,cnncru:.Lincct. Ir59'5-597,.1'9ISJ. 8'. Gvfkn~kr1;,L., Aucrt»aA, 0., andiJoubrn, L. Ihroluntuy, smotint, r:,,: c-tt.r.s. cs`uconrrcl srud):.1.. Nad. Cinrxr Inst., 19i a63"69..1985:. 9~9 t.liragama, T. Ptssi.csmot~inj;andllun=eanecrc consisttacy,ei Lancm. ): 1 4:5-1i1:6', 1983: 10. hlstsukura;,Sl, Taminuto,.T., Kiuno, N., Srino,.l'., ){rmrd.z, It.. Lic`::• : M., ti.'ak.jimr, IL. and 1Nirata, Y. Efkrcta of cnsvonmcntrl tobacco s: omuninar7cotininc.rscrraion in.nonsmolrrs:r+idtincc for pA36•r t^c N: Enti. r, r.ard.. Jll: a,28-a32.19a4. :- 111. Illmmnnd„E. C..Smollint.in~rrGtion.to.thz.dcath ntcs.of one rnillin -. and:womrn: N.tl. Ca'ncor lnst. Monogr., 19: 127,20a, 1'9E6'. 1!:: Rogot. E.. .nd Mum7; J_ L Smokint andlcauws of drrth.mon: •otrrans: 16.ycars.ofobs+.cr~aion..Public )tcalthRop.., 95:.2113-2;21r'I'J. Dol1; R., snd Prt'o• R.,Ftonality in rnluion to smokint: 20 ycars obsrn :; onmuli Drritish dc+rsorn. Dr.. Mt J. l: 1ISI2'5'-1536, 1976'. IN. Surgron,Gancrat. Tlhr,hcrtconsrquencrs of 1mokinj: cancir, W'shir. rX': NparImrns ofIfralih and'1lluman Scr~icrs;.1982. ~ IIS. H'ald, N. J., Bonrham, 11• B0c1„ A., Riitchic;,C.. )Itaddo-, l:, and G. Urins:o ootininc u mulur of brcashint,othcrpeoplr'i tobacco snc Llnccd..1: 23n--27 ). 19841 1i6'. Doll1 R.. rnd Pruo.• R. Tlhr,c-auses,ofiearAcor: J.. N.d. C.:ncrrlnss.,.66: :. 17112; 1991. 1:7• Luhin, J•.N., andlBlotlw~ 1. Assessmrnt ofIuntc~wcr riskfacrc:-: Isistologic ntntnry,. 1. Nsd: Canccr lnst:,,z7: 3'13-389„1'984. 118. Tvichopoulos;.D:,l;atlndidi, A;, and Syarroa, L Lung cancer usd'p>:• smokint: concJusion of Greek studr: liannt, )i,677-d'7i11, 1933. 19. Trichopcsvlos„ID: Pssi.r s'molint and luntnncrrg tAnoct, /: 68t, 196r M GuAnicll:L. Time trrnd.in lunt,cuvttrmorulitl.smont non,smolrra a,nntr on patvitr vmkinK: 1. Nslt. Cancrr Intt., 66: lOd1-I1o56, 1'93,1 :.1i. Kc.o. L C:. I'to,,J. 18.,.and Sr., D:,Is.pasti.evmotint.ansdN}E rist rr;. fcir lung crntra in Chinrw .nrnrn: 1. Exp Clin. Cancer Sfcl.. l; 2"" : 1994 .., l..h+r: G C, and \\"lndrv, f'. t,. t!unt cancer in.nnnsmotornCanczr (I': ,: !3~ 1:1 4 .1 :.J1. I19M4 a'8(17'