Tobacco Documents Online

(i32'. "IlhlF NEW 'F:NCisAialtD JK)'U,R'NAL OF M1Ft'JIC;INE Sept. 6, 1990 LUNG CANCER AND EXPOSURE TOTOBACCO SMOKE IN THE HOUSE'HIOLD. I)IXtcHT -I'..(ntir:RICH, 1)il):5., R'ILfD.H., l'4'. t)otr(;t:.A's TitrnMl+so.: Ptt.f)., Ltas R. VARELA, M,.1)-„ YHtD,,* IaETE'R.G~REENWALI), ttiLD..,,1)k~.f :H., S~HP.R'Rl"~, G.HDROSII'', M.S~~., L'.AT.Hh'~ I UCCI, IUS., MUHARSM~ADIB- ZAMAN,.M~.D., MYRON R'. MIELAN[ED„ j%+fLD-„ MAUREEN ti:[ELV;. R.N., ANID M~ARTIN.I". MCKNEA~hFl`; M,.D... ~ Abstract Background The relation between' passive smokingiand lung cancer is of'great publicheadth'impor- tance. Some'previous studies have suggested that expo- sure to environmental i tobacco smoke in the household can cause lung cancer, but others have foundlno effecti Smoking by the spouse has been the' most corlnmonlyy used measure'of this exposure. Methods. In order to determine whether lung cancer is associated' with exposure to tobacco smoke within the household, we conducted a population+based case-con- trol study of 191 patients with histologically confirmed pri- rnary lung cancer who had never smoked' and an iequal! number of persons withoutl lung cancer who had never smoked. Lifetime residential histories including informa- tion on iexposure to environmental tobacco ismoke' were compiled and analyzed. Exposure was measured in terms of "smoker-years," determinedlby multiplying the'number of years in each residence by the number, of smokers in the household. T HE 1'972 Surgeon Ceneral'srrp(ort dealt with the health consequences of passive srnrokinG,or envi- ronmentail tobacco smoke for the first tinic.' Ln19$6the entire'repnrt'wasdevotedltotihe i'ss('ie:,iit coneludedi tdiati"involuntairy,smoking,is:a cause rnf discase includ- ing IlJng cancer in hcalthy non-smokcrs."I Mmre tlhan~ a dozen epidemiolqgic studies tiavr asscssed the relYt- tion bctween passive smtDking and IunG'cartccr."I''The'g findings have ranged from no detrctahle increase in ri5k't''I'tolarnoderate (,about ta,vofillcd)y statisticallysig- nificant iincrcase:'' A4kost studit•s ha%e Eound on(v.' small elevations in,risk, which arc frccquentdy not'sta- tiistically signific'ant.'-"~" In a meta-analysis of aCl thc available stiudics in:L'9£3fD, 1\':alc1',et al. filund'a sliglnly iiocr~eas'ed risk of lu'n.G:can'ccr as:st/ci.utctd with etnvironr mental tDtDlcco: snlnke." We undertook tlhc current stiudv' in arr attempt to clarify further the rol't+:of pa>;xivc s:nDt/king in causing lumg c.arncer. Iln tlnis raport' we diseuss t•xIDUstnrc tcD tuo bacco smoke' in the household as a pussih(h causc ml- lu'nQ cancer among nolDSnl<,kons. M. ETH O~.DS l\Sr c:nnduclyd a Ir,pulannn-b:-d, iinli. rdi:;:llbm.rlchrd. c:rsr- cnntrul studb in Ncw Yi:rk:tiuur/i,nni I'17i' tn: 1'1133:.'I~hr catsnsscr,tn drawn frnrn crcrwStanditrd \lu•trnpo,ht.rn Stntl.ctiral .\hn:rs From the:De(tartmant of tpilem:oiogy, and. Rubli..ll:elth. 1'a!e t+nversn). Sch,x,l of \tudirinr;.Neu-Huvcn. (',inn. (l).r.J.Ii the lr! o6 Suunccm Ma.me. f'unland iW'ID.T..I: the.liyruluruun Cnun:nl , t\l-::,,. Met::cu:(*ity. (1..V R.)..IhcN:rGomaIC.anccrln.wulr:.N'.a,hinEtrne.l).(~_tl'Go:,Ihc4cwYiurk. St:ite Ikprtmentod Ilcalth, AthanytS.C.. c. r u thc Sh+:m.h;oncrinc \t8nwriali Invtilu(t.VOw'i,r.kIIM.H!7., MiR'A1.I.anJlAlh'an_v.S1 d:a.a'.tiollcg'. A7han).. ,V.Y. (M.K . M.F M.I, Adlfrc., rcprmt rrqac.t,n+ 1)iE: n.h ,o the Ihp:rnt~ mcnn of 1 plc:n-litgvenJ Nuhlnc Health. YAlc I n:-r nc,Suh,"+ta( Mu.bdinr: rXl (Zdlcge. Sr. OLEI'll ItIS);,Ncw Havcn. Cluaillt; Surrpx)ncd in part ti)'.y frantA tl"()l cA4^Itll Jlkl R{1!I (',11_0SKl.frnnt thcNatiunal In+btu:cxuflFBc:Jrhi •Ik.ca.oJ. Results. Househ'odd, exposure to 25 or more smoker- years during childhood andiadolescience doubled the risk of lung cancer (odds ratio; 2.07; 95 percent confidence interval, 1.16 to 3.68). Approximateljr 15' percent of the control subjects who had never smoked reported this level ofexposure. Household exposure of lesslhan25 smoker- years during~childhood and adolescence did!notlincrease the risk of, lung cancer. Exposure to a spouse's smoking~, which constituted lessthan onethind of total, householdi exposure on average, was not associated with an in- crease in risk. Conclusions- The possibility of recall bias and other methodologic problems may influence the resudts of case- control studies of environmental tobacco smoke. Nonethe- less, ouriindings regarding exposure during early life sug- gest that approximately 17'percent of lung cancers among nonsmokers can be attributed to high levels of'exposure to cigarette srnoke during childhood and adolescence. (N Engl J Med 1990; 323:632'-6.). ('Buff'alo; RncLrstrr, Svrtcusr, t;tica,Rhmr::\9bam-Srh'onrctadp- Trny, BinKhamt,tnt and gratrr New York. escludiqq the h.e bor- oughs nfl Nr c 1"ork'', (:ity), This s;a:Graphic arra compriscs 23'rnun- ties;.w•.itluapprosimatcls 12:idiaGnnsticor:trcattncnt fiicilitii•s:.:rnda popula~.tiimba~c•uf nra.rln~ 10 ntillinn por,ple. A special systrm f+rthe rapid.ascnrtainrnrrnt nf~ casrsof IunG cancer wasrstaltlisticci in thcsc. I2:i:fAc.ilitirs:sr, that patientscould bridrntifird and rnrnllyd assntmaf4er d:agnn.iF n pussihlc: A'll ncw cases of lungr,ancrr(rliaqnnscd clinicallir-,h:st duGt(..rih•, nr hnth) ~rrcc rrGula:rlvidcntilird, ar ohr partinip:uiint; hnspiLtl.. I he Nrw 11nrk St:rtr C:utrrr Rrejstrv, s.'as: ch'eckrdlrnut{nrlv:tn irkntif-•r anv.casrathatmiGlitliaor been missedl by thc hrapitai-bnsrcl reporrtin>; ssstrmj . Lnf,rmari,.n nn snwkin~,s,Au:initiallpuhtaincd frrnn tlie palients' mrdli.+al rncr,rds ,\II' Ihr c.tsr p:rtii.uta rrpnrtrd as hanine ncs-or smoked or asfnrmrrs srnvkecs.urwhnsrsmokin,t; historv was un. k'natsan ccrrr rr.nt::clyd bs trlrphnnr; ;it+d du ir smnking statcts was cnnfirmr& 'll;, hc-iucludrd as.t"c,rsc" in d:r studv. a patienctiadl,n rrsidc in,thc ';-t,ntutt% arra- hc hrtsccen NI ,u:d HQ!'vears of atr;, ne.-cr hacr -r„k<~cl mnrrthan Illllicit;:urtrrs (imnstnnkrrs)~nr~ ha%~c. smc>krd,arsr-mc tirnr but nnt hactsmnkrct morr.lhan It10eiQarrnrsin thr 10 vrars hrlr,rr diar;panis (furmrr ..mnkrrs); and base hern Givrtt a dia¢nnsis ull printarN lung canrrr brtnvrcn.f uh• I, 1Pfi'_); and IDcccmhr.r:il. 1981. nhat wa:s cunhrunrtl on rrr±rauninatinn nfi thrpntha Nryical +prcimrns.:rud t linical rtcurd>: \lidrs-.nr binck'ssrnf 1is:- suns a r• annilul,lribn adl bur liur.of lha. casr pauents. All matrriials wcrc rrsii-rd lis intrsti~atnn wlro worrblindrd wilh rrsprcctnatie patirnt'sinitial:diagnn.is, sntnking '.historc,.:utd other risk Gtctars. litJrrvirws s.rrx' crmduc lyd'witlt 7fi'prrrcnt ull the rligihkpatirntae nr their cLosca ec::iLtldr rr.l.:;ivvs.orrfrir•nds,(sunrnRatt's):.. C:rmu<d .n/qrrt, svrir inrliciclualln•m:uchadtuthc patients atrd t,arr srlcrtrd hv zrrrrnioZ thr lilrs of thr \rw \'inrk State Ihpart- mrnt nf' \Intnr \'rhiclrs. Fhis sourrrnf controls wascnnsidrrrd appropriarr sincr• in.vns pupuLuuun-based and provided most of the in/i+rmatiiuninrarsserv tnprfnrrn~.thc malrtiinQ; A list of pntcntic+l unntonl suhjrrts tim rach c:rsr ptticut s.as srlnclcd rnn tlir, hasis nfl ::g<- (iai t liin Il., r a-z..and lcnuntcrd rrsidrncr. pntruti:d cnn- trol subirris c.r rr cnntectnd liv telrphunr: Thr first eligible suhirct crhuccas !i.und ur.malclithrcasrpati(•nt in trnmsofsmnkine histor oc~, (tnn.mtu kcrn:°. G inn:rr smnkr:r ) end who aQrrrd in participate s,•aar.nrnllnd.in :lr,.:udt. :\nUadtl:uunal rcmatc'h:nr;carilrhlb cc+nsidr'rrd .W . thr:i:nr r.frln:ant,dlhruu:a -- rhrispruf' intrroir.v -i.r., dirrct' intrrn'ir.v sciih tlieprtia+nr::c cuntrul xuLjrr uvrrsus,intrn+ravsviih asurrng,urrnspnndrntA lKhrn a surrne:ucrasr pa.tii•nt had tn br intcnirsord, tcr;d.n iutt•rvirwrtf a surrnt(:uc' lorhisor ht'r marhodl

Vol. 323 No. 10 THiE'NEW ENGLAND JOURNAL OF'MED'ICI''NE 633 control, even when the control subject was availkbk and willing to be interviewed.. Further information on the methods used in the study is availabteelsewhere.'y Data werc'.colleeted for 339case-eontrolipairs. Of these, 2421 pairs were formcr, smokers and 1i97 pairs had never smokedl Separating the residual dfccts of direct smoking from those of passive smoking amongiormer smokers;involLes,more complex analytic and inter- pretational issues than does an examination of the effects of passive smoking in thosrwho have never smoked. '17iis report is therefore limited to persons who never smoked. Six nf the 197 pairs who hadl never smokedlwere',mismatched in terms of the type of interview (direct vs:surrogate) and have therefore been excluded. Thus„the, analyses reported here were based on 191 matched case-control pairs. A total of 129 pairs were interviewed directly;,and surrogates were interviewadlfor 62- All information was collected during a face-to-face interview with use of a precoded questionnairo Case patients and control subjects were interviewed in exactly the same fashion, and except for items concerning the clinical aspects oflthe current medical condition, both,groups answered the same questions. Information about smoking',in the household was collected sepa- rately for each residencein which the,subjeet had lived forone,ycar or more, up to a maximum of'12 residences:,The numberof "smok- er-years"'of exposure was calculated by multiplying the numbenof years the subject'lived in each residenee by the number of smokers (includinG thespouse) in th'at residence.'Che products for all resi- dences'were summed. Smoking,by thc spouse was also recorded separately from that by other household!membcrs in a subsequent sectinn of the q}testion- naire. The informaunn consisted of the number of years the spouse had smoked'whilc living with the case patient or control subject and the numbernf cigarrttrs smoked perday,. Smker-vears of exposure from the spouse's smoking were calculatcdlin the same manner as for the entire h'ouselioldl Pack-years ol' exposure from the spouse were calculated by multiplying thcnumberof packs smoked perdary- by the numtxrnf years that the spouse s'mnked while.living with the suhjecr If! the subject had been married to more than one'smoker, then the numbers of smokcr-years andipaek-ycars nfl exposure for all spouses wore'summcd. 'I'hc qurstiontnire a lso ~inciudrd scetiims nn natHasurc tn envirtrn- mentaLtobaeco smoke in the workplace and inso€ial settings out- side the home. Ttieformat for these'qucstions differed from that used tu eollect'data on exposure in the housrhol8i The sctmmarv results of this analysis are presented here;: detailed findings are aM1•ailable' elsewhere.19 Statistical trchniqpcs:appropriate for the analysis of individ'ually ntatched rasr--ountrul studies were usrd." Forclitrity of presenta- tion, percentages were tabulated for case patients and cnntrol'sub- jocts separately,,raahcr than for matched pairs. However, odds ra- tios were computed on the basis of the matched' pairs. The orntdltiunallogistic-regression model was used in the multivariate analyscs." Comparisons of the effects of exposures that occurred during diffl•,remt'periods of the subjects' lives were based on evalua- tion of differences-in the magnitude of appropriate logistio-regres- siun cnctticicnts. For statistical testing of Ihese differences, we used the varianne-cavariance matrix frum the logistic-regression analyses. RESULTS Smoking by spouses contu-ibuted a large propor- tion of lifetime exposure to environmental tobacco smoke but'w.-ts not the ch'ief source of exposure. Table I shows the amount of' exposure to environme.ntall tobacco smoke (expressed in smoker-years) during', childlhood and adolescence, during adulthood, andi from the spouse for thc 191 control svbjFcts who hadd never smokedl 'I"here' were only'small differences lie- tweeit men and women. The'spouse contributed about 30 percent of the lifietime smoker-years of exposure; the correlatiion' coefFcients for exposure from the spouse atid lifetimr expnst'Ire were 0.37, for men and Tablet'. Distribution of Smoker-Years of Exposure to Environ- mental Tobacco in the Housetiold.' c.Tt(:OtM Or. Exp-lE. AtEM wOMEM Li(eti.me smoker-years. mean 2'S D. 46.6:±53:7' 52.7'_42:.99 Smoker-years durmg, chddhuxd and adblescencet Mean ±SD 15.4+_ 20:6 16.1± 16.2' Percent of lifetime exposure 33;1 30.6 CornelYtion with lifetime exposuo: • Smokcr-years from'spouse 0'92 0.61 Mean !SD: 13.0±17:0 16.2"_16:7' Percent ofilifetime exposure 28:0 30.7 Correlation with lifetimeexposme: 0i37 0:51'. Smoker-yeaosduringadulthoodfromsources.otherthan spouse Mwni:tSD Percerubf lifetime exposure Correl.Nonnwith lifetime exp,osure Iq;.1-±39.0 38.9 . 0:91 20.5±29:91 38:9 043 •B1sed lom 43'.mate tutd 1 46 ,femate rontrol subjects ho had inever snokedd mese,.tlun tOD, cigaueaes. Infomutionion smoker-ycurs~of caposun;m ftonsqhespaae is fer 45 nute nnd, 140: femalecoatrat subjeets:.Smuker-years wcrt calEUlated Iby tnuhiptying the number of yeass Ne wbjett lived.in a meidence bythe nwnberW smokersin:the hnusehold.l tlkssnhan,21 years.nfisRe: 0.51 for women. Exposure during childhood and ado- l'escence (<2Il years of age) contributed a similar per- cenitage of the lifetime smok'er years but correlated more closely with lifetime exposure (correlationcoeffi- cien't, 0:92'for men and 0.61 for women)'. The,average lifetime exposure was 46.6 smoker-years for -men and 52.7 smoker-years for women. IDuring adulthood, household exposure from sources other tihan the spouse was somewhat greater nhan from the spouse. Table 2 shows the odds ratios for developing lung cancer in'relation to the degree of exposure to tobacco smoke in the householdl for the 191 nonsmoking case- conurol pairs. The data are stratiifledlby levels of expo- sure (measured in smoker-years) and by the periods of life when the exposure occurred. Exposures during childhood and adolescence were defined as exposures that occurred when the subjects were less than 21 years of age. Exposures during adiul't~hood include all household exposures from1 21 years of age to the time of diagnosis; Alt'houQh the odds ratios fiori the higher exposure categories are somew han higher than those for the lower categories; no clear dose-response rela- tion is evident, andl most of the 95 percent confidence intervals include 11.0,. For exposu'res in childhood and adolescence,,the highest level'of exposure is associated with the greatest elevation in risk, and the 95' percenu confidence initerval',excludes,the nulfvalue (odds ratio;, 2.07;,95 percent confidence interval, 11.16 to 3!68),,For the 129 cast•-conoroli pairs who were interviewed di- rectlv, the odds ratio for persons with 25 or more smoker-years of exposure in childhood and adoles- cence was 2.311 1 (95 percent cotrfidence interval„ 1. 116 to 4:61),. With smoker-years during chiildlhoodl and adoles- eenceandlsm'oker-year~sduu-ing adulthood treatedlascontinuous variahles amd includ'ed simultlane!nusly in a logistic model, each inrrcment of five smoker-years of exposure duriniG childhood and adolescence was found to inereasc the risk of Iun,t; cancc.r by 615 percent (95 percent continenc.eintervad„0.1 to 13L2). OYa tihe other hand, cach adcLition,tl five snac/kf-r-yc.urs ctf cx'posurc.

634 THE NEN', F,NC[:AND f O!UIRNAL O,f INtIF.151KaN t: Sept. 6, 1!1!x) Table 2. Relation lof iSmoker-Years of Exposure to Envirtonmental I Tobacco Smoke to the Risk of Lung Cancer among Persons Who Never Smokedlfvlore than 100 Cigarettes.' Mo. o.~ cNa~ Swoc[n-Y[wu ~~. P.newn QoMrRou Ooos R.m (9Sa~:CI) CRiildhood land adolesceneet' - 0 57 (29.8) 68I(35.6) 1-24 82(42.9) 941(49.2). I.O9 (0_68- I L73:) =25 52(29.2) 2905.2) 2.07 (1-16-3.68) Adulthood 0. 44 (23.0) 39 (20.4). 1-24 37(19.4) 48I(25.t). 0.64 (Q.34-1121) 25-49 46 (24.1) 50(26.2) . 0:81 (0-43-1145') 50-74 36'(18.9) 32'(16.8), 1.00140 .52-I l93) a75 28(it4.7) 22'(It.5). 1_11 (936-2:20) Liretime 0 32(16.8) 331(17.3)'. 1-24 20(ilo.5) . 27'(I4.1). 0.7&(Q.36-1:67) 25-49 35'(18.3) 4ti'(24.t), 0.80 1(Q:43-I .50) 50-74 44 (23.0) 40(20.9) , 1.19 (Q:63-2127J 75-99 33 (it7.3) 21 (11.0). 1.80 1(Q.SD-3:90) a1!00: 27 (n4.1) 24102.6) , L13 (Q-56-2:28) •Bised inn 191 mMOhed i!ase-cpueL paia. Cl idenotascanfide-interval. Odls niios aiesAow.n for a penan with AWexpoaon: spcci6ed as compane4 with a Peasun with.u expxun: (00 smokr•ycan):.B6nuud mundiha. pcrceou`R Jo nd,always iohl 100: f clca tl!un ,21 yers of'a=e. during adulthood were estimatedl to have virtualhy no effect on risk (95 percent confidence interval, -3.3' to +2.8 percent),. The difference in the magnitudrof the effect between exposure during childhood and adolt•s, cence and exposure during adulthood did no:t achieve statistical significance (P = 0J2). On the basis of the distribution of exposure levels during childhood and ad'olescence among the control subjects and the magnitude of the effect of early exposure, we estimate thart approximately 17 percent of all l!ung cancers in nonsmokers.can be attributed to exposure to pas- sive smoke in the household during childhood and adolescence. On the basis of the odds ratios for the 129 case-crontrofl pairs who: were interv-iewed' di- rectdy, approximately 19 percent of lung cancer in nonsmokers appears to be attributabic to exposure to environmental cigarette smoke in.childhood and adolescence: Since smoking by the spouse has been the most'; commonly reported measure of exposure to environ- mental tobacco smoke, in previous studies, we exam- inedl exposure ftom: the spouse separately, although, exposure to environmental tobacco smoke from thee spouse is also included in the results shown in Table 2., The odds ratios !br exposure frequently differed ac- cording to the type of interview, especially for thee da~ta onexposure:trn a spouse's smoking. Table 3 there- fore shows the results of the analyses of exposuree to environmentall tobacco smoke from the spouse, separately for subjects interviewed directly and those, for whom surrogates were intiervietved. Theodds ratio: for the development: of lung cancer for those whoo ever had a spouse who smuked, as compared withi those who did not,: was 0.93 (95 percent confidence intervaJ„0.55'to 1.57) for those interviewed diirectllv: llnn terms of smoker-years of exposure to the spouse's smoke„the results show little effect, with an odds ratio of 1.07 ' for 25 or more smokcr-years of exposurc (!95~ percent confidence inl' 0.59 to 1.97')L Estimates based on pack-years of exposure from the spouse were similar to those based on smuker-years: For both measures, there was little evidence of a trend accord- ing to amount of exposure among, those who were exposed. All analyses were repeated for only the case-control pairs farr whom we had complete and internally consis- tenrt: da~ta for all residences and marriages. Any pair was dropped from ahesc',analyses if data were incom- pltte or missing for eitiher the case patient or the con~ troll subject, leaving 1'13' pairs of' nonsmokers., Our purpose was to ensure that our conclusions were not dependent on the particular methods we adopted to handle inconsistencies or rnissing,ittms in the data,seti The findings were similar to those for the entire group of 191 pairs. The oddk ratio:for exposu!re to 25 or more smoker-years in cl)iildhoodl and adolescence was 2.59 ('95' percent confidence intierval!, 1.22 to 5.49),. E'xposurein.tfie workplace was nncasuredby record- ing the number of smokers who worked' with each study subject dnriing his or her lifetime and tlie amount of time the subjects spetnti working with these smokers. These exposures were compared for case patients and control subjects. Estimating the odds ratio as a continuous variable for an equivalentv dilherentiall of 1150 person-years of exposure gave an odds rauio of 0.91 (95 percent confidence interval, 0.80: to I.04), indicating no evidence of ani adverse e(Tect of environmental tobacco smoke in the work- place. Our assessmentu of smoking in social' settings used an un'testcd', semiq,uainnitative index in which the case patient or crnnnroli subject used a score of 0 through 12 to indicate his or her regular exposure to tobacco smoke in social settings dluring,each decadee of hfe: Cumulative lifetime reported scores ramged' from nearly 0 to more than 70. The odds raaio for am increase of 20 un, the cumulative score: was 0.59 (:951 percent confidence interval, 0.43' to 0.8'd): O,uranalysis of exposure in social settings with use of this indexshowed a statistically significant inverse as- stnciation between environmental tobacco smoke and lung cancer. Dascusstotv We found a statistically significant adverse effect of relarivel%. high levels of exposure to cnvironmcntal to- bacco sn)nkeduring thetarly decades of life (up to the age of 2'1). For tdiose who were exposed to 25 or more smoker-years during their firsc two decades of life, the risk of lung cancer doubled. This amount of exposure is equin•alent to living, with more than one smoker throughout childhood and adolescence - a high but not uncommon level! of exposure. An exposure, ofi this level w•asreportrd fbr approximately IS percent of the controllQroup. By contrast, we found no adverse clfect of expnsure to environmcntall tobacco smoke during ad,ulthood, incl!udiing cxposure to a spouse who

VoL 323 No. 10. THE NEW ENGLANDJOURNAL OF'MEDIiC11:E 635 Table 3. Relation of Spouse?s Smoking to the Risk of Lung Can- cer among Persons Who Never Smoked More than 100 Ciga- rettes, According to Type of Interview,• VNl1A.LE' TY'E or INTERViEw SURROGATE nI ~rn+io~(93%~.11 Ever tad a spouse who smoked No - Ya 0s93 (0:S5-I'.57) 0A4 (0:19-1.02) Smoker-years of exposure fmm spouse . 0 1-24 0.78 {0l41-1.SID) 0.33 (0• 1 II -1.05)' s2,S 1.07 (0.59-1.97) 0.33 (0.12'-0.95) , Park'-ya.rs o( exposure from spouse 0 1-24 0.71 (0.37-1.35)' 0.16' (0.04-0.62) 25-49 0!98 (0.47-2.05) 0i68 (0.18-2160) a50: 1.10 (0:47-256) 0;20 (0i03- t .22) •BneA w 129cne-eoohol ip.in uneerviewMdueayaod 59 pain f« wlnm wlmEalimweee .. imerrirwed. Th[e,ofuhe 191 ippGy~weie~exehded'hec.weforune nsmbenufithe~pair Yle/e ~wn '. miasina in(omwiua atww whethrr the'E.Eject,h.d in spomse whu smehed: D.meu vynkrl yeaniof Kapoane weac,.r.ilatik fu'129 cee-<onud p.ias with dirccl iMCrnewE md.36 pain widl sweqpt.iMerview.: Dtla iua ip.A+yeanOf eapwurc.wese w.ilabk(v 122 pain with ~ dlfM'iMerYKw{ NQ f.1 p/ill'wp111fWTo~ae'~160CT'icwss eIN1e11G1esicoRfNdeOCt',IIIYfY11. Odds ruies'~ue.slnwn fa. ~perwn'wiN idlec.poswe specifiada.nxnpnN'wdh iE iprnun with mu espoaarr to 1a ilpouse whol amnkcd.. smoked. Although problems of recall and other poten- tial biases may have influenced the results, our data suggest that exposure inlearly life maybe a limibed but important contributor to the' risk of lung cancer in nonsmokers. A'previous study withia small number of subjects found little evidence of an elevated risk of lung,cancer among nonsmokers whose parents had smoked.la'Childreni of parents who smoke have been shown to beespecial'ly susceptible to respiratory prob- lems that occur soon after expos'ure to environmental tobacco smoke.2 This: type of susceptibility might initiate changes that eventually lead to lung cancer when the exposed children become adults, but we know of nospecific mecha~nismitliat'would''explain our findings. We found no adirerse effects of exposure to tobaccoo smoke in the workplace, alithougli we did not have, enough information about the level of exposure in tlae workplace to assess the precision of our measure- ments. The apparent protective effect of exposure in social settings is difficult to explain. lluring, the course of this study, regulations ini New l''ork began to restirict'smoking in the workplace and in social'set- tings such as restaurants. We did noranticipate this development' and cannot estimate fiow much thc awareness of these new restmictions might have af- ketedl the responses of the study subjects or thcir surrogates. Evidence is clearly mounting, tihat tobacco smoke inhaled passively by nonsmokers is potentiadly car- cinogenic. In a recent stt'tdy, Maclure ct ad." found'd elevated, levels of carcinogens in the blood of passive smokers. Levels of hemoglobin adducts of 4-aminobi- phenyl'',andladdiucts of 3-aminobiphenyl were signifi- cantly elevated in subjects with confirmed exposure. The validity' of this finding was supported by addi- tional' evidence tha~t showed a sharp decline in tihe levels of adducts among smokers who quit.2t At present, infurmartion on past exposure to envi- ronmental tobacco smoke can be obtained only by interview. The available biologic markers, sulch~as co- tinine, cannot be used to confirm exposure that oc- curred years or decades earlier. The'use of interviews to obtain al lifetime history of exposure to passive smoking,requires that the q'uestionnaire be structured and'the interview techniques be standardized so that all subjects are interviewed in the same way. We tookk steps to ensure',such standardizationi Two:recent re- ports may lead to improvedi ways to measure lifetime exposure to environmental tobacco smoke by' means of interviews.rt~` ' In one of these studies,, which at- tempted to evaluatetihe reliability of interview data'by' repeat interviews, information on, exposure dntring, childhood was found to be very reliablr." It was necessary to use surrogate, respondents for about one third of the interviews, usually because the patients were too ill to, be interviewed. To minimize potential bias, surrogates were adso interviewed fbr the matched control subject's, and'scparate estimates were calculated for respondents interviewed directly and surrogate respondents. We used equal care in all types of'interviews and in all subject areas covered in, the', interviews; however, the data we obtained in inter- views witlh, surrogartes still differed somewhat from those obtained in direct intierviews. l'naccurate report- inq' of exposure tends to bias odds ratios toward the null value unless al systematic bias is present. Data from, surrogaterespfmdentsare likclyto introduce random error, because of the surrogate's lack of de- tailed knowledge of the subject's exposure. On the other hand, it is'', possible tliat' the surrogates for pa- tients with,lung cancer might tend to underreport the ex'posure contributed by their, own smoking to a great- er ex~tent than surrogatrs ftor control subjects.,S'uch al dilterence could mask a''n actual increase in risk or reverse the direction of the association. The findings~ shown in "1'able 3 indicate that the use of data from surrogates may hav.e lcd',to an undcrestiina(iion of the elfect of exposure leom, the spouse. Alrchnugih' our.re- sults for exposure due'to smoking by the spouse differ from those of earlier sutdies,'" our f ndings regarding other types of household exposure support, the conclu- sion that exposure to! environmental tobaccfu smokee can cause lung,cancer. Akiba et al.,"' Dltlagar et, aV.,`"and Carfinkel" have reported elevations in risk of 30 percent, 50' per- cent, attd 110 perccn't„respcctively, asstociated with ex- posure to a spouse's smoking; none of these increases were statistically significant. With the exception of Chan et al.tr and Koo et al." in HonG' Kong, these and most other investigators have reportcd point estimates that'suggest an increased risk for those exposed. Thee duration of exposure, as measurcdi by the number of years the spouse smoked while livinl the sub- ject„ did not have a stacistically', signifcant': elrect in our data Two studies tfiat, used the same measure

636~ THE NF.l4r' ENGLAIVU Jp'URNAI: OF PvtIED1ICINft of exposure als'o failed& to exclude the null value-N9 Garfinkel et al:,'using a,d'ifferent measure for duration of exposure (husband's smoking in the last 5 and 25 years), found one significant association among t'ihelarge number examined. Exposure due to smoking by the spouse, expressedl in; terims of pack-yea'rs while the spouse was' living with the subject, was fbund not to be~signiificantily associated with lung can- cer. Using a comparable measure of exposure, Tricho- poulos et al.s reported relatively large increasesin risk (greater than t'wofold).. Perhaps our data do not show that'smoking by the spouse increased the risk by itself because smoking by the spouse, made up only about' one' third of the subjects' lifetime exposure to environmental tobacco smoke. It is also possiblee that physical circumstances and differences in study areas, the' size of residences, ven'tilation„ and other important physical aspects of the living'conditions; as well as soeial' habits that affect exposure within the family, will'ineed to be measured and analyzed before the:diilerences in findings among the studies can be reconciled. . The evidence we report lends further supporn to the observation thart passive smoking may increase the risk of subseqpent lvng,cancer, and it suggests that it may be particularly important to protect children and adDlescen'ts from, this enNironmental I hazard. We are indE`bted to A'ndreas Nicolaou for his assistance with the computer programming used'in our analyses. R~EFERENCE51. Department of.Healtli, Edueation:.and Welfare. The health consequences of smoking: a.repost of the Surgeon General: , 1972 . WasAingtun, D C.: Gov- ernment,Piinting Office;,1972:121J5i (DHEW publication nu (HSM)72- 75.16,)', 2. Depanmem.ofHeatthi and, HumanServiees. The.heatthconsequence.e off involuntary,smoking: a:report:of the Surgeon Generall Wa~hmgton, D.C.: Government Printing O6fice, 1986: (Publication: no. DHHSI(CDC) .87. 8398. )', 3. Garfinkel L. Time trends'in lung,cancer mortalioy.anntngnonxmokersandlanote on passive.smoking. 1 Natl Cancerinst: 1981; 66:106I16. Scpu 6, 1990 4. Huayama iT.. Cancer monatity in nonsmoking women with smoking , h'us- bands based on a Jargt-scale ~cohoat isnndy in Japan. Prev Med 1994; 116W 90. 5;. TrieheapuulusD, Kalandidi A,.Spartvs L. Lungcaneerand passive smok~ing:: conclJtsian~af Greek study:.lanceti198fl; 2i677-80:. 6:. Cortu P. Pickle',LW;.Funoham E:,Din Y, Haensrtl W. Pusivesmoking,and tungeancen.Laneet 198]li,2:S95'-7i. 7: GarfinkellL„AuerbacRO;,Juuben L. Involuntarysrnoking~.and tung.cancer: rease contmi study. J Natl Cancer lhu:198fi 75:463=9. 8!, Akiba S. KatoiH, Blot..WJ. Passive',smoking.and lung cancer among Jipa- nese women..Cancer Res 1986; 46:480'4-7.. 9:, Datager,NA, Pickle LW;.Mason ~lU; et al. The rolationof passiwe smoking to tungcnrrcer: CancecRes1986; 46i4808-I I.. 10:, Kabat GC„ Wynder EL. Lung cancer in tawsmokers. Cancer 1984: 53:1214-21. 111. Sandler DP, Wilcoa AJ. Evetssrrn RB. Cumulative effects ofllifetimepas.cive smoking pa cancer risk. Lancet 1985; 1:3J 2.5. I2: SandkrDP,Evernon~RB; WikoaAJ,,BnowdcrJP:Cancerriskinadulthood frvrnvarly li'feeaprn;un:.to.parents"smuking. Am JiPublio Heattb 1985:: 7b:487-92: 13: Sandler DP;,EversoniRB- Wihcox AJ. Passivesmokinginadultlhood and cancer risk:.Am,J Epidemiol 1985; 12137-48;. 14. Pershagen~.G, Hrubec,Z,SvenssonC: Pusivesrnuking.and lung.cancer-in Swedishwomen.. Am 1 Epidemiol 1987; .125:17-24, 15. Koo.LC, HoJH, Saw D, Hb CY. Measurement of'passive smoking.and estimates of lung cancer risk amongnon-smok'ing Chinese females. Int 1 Cancer 19871 39:162-9;. 16. Humble~CG, Samet JM. PathakDR. Marriage toa'smoker and.lung cancen risk. Am 1 Public Health ~ 1987; 77:598-602. 17. CAan WC. Pung SC. Lung.cancer ininon.mok'enin Hong.Kong. In: Grund- mann E; Clemmescn 1, Muir C. edt.Geographica6 path'olugyy in canar~ epidhmiology, Ncw York: Gustav Fisher Verlag. 1982:199,202: 18. Wald N1LNanchahat K. Thompson SGi,Cuck'IeHS:..Dues.bSeathin~gothcr. people'stobaceo smoke:eause iungg cancer? BMl 1986;,293 :1217-22. 19.. VarolaLRl Assessment of theassociatiort between passive smok'ing.and lung.cancer:, (Ph.D: dissenation. New. Haven, Conn.: Yale Unibersity: 1987.). 20. BieslowNE., DayNE. Statntical methodsin~caneu.reseaach: Vol. I. The anatysis of ease,cummi studies. Lyon. France: Imernational. Agency for Research on Cancer. 1480. (IARC scientifiepublicationsno.. 3i2.). 21i- Mlclure,M, Katz,RB. Bryant MS;.SkipperPL Tanneobaum SR. Elevatedd blood levelsof carcinmgenss inpassive smok'ers- Am 1. Public Health 1989; . 79:1381-4. 22. CummingsK'M, Ma~rkello Sl. tvlahoney,MC. Marshall JR. Measurement ofhfetimeexposure.topas,ive smoke. Am1.Epidemiul1989;.130:t.22-12~ 23. Couhas.DB.. PAakc GT. SamcuJM. Questionnaire assessment of liftlimee and recenhexposurc tu envimnnuntalJobacco smoke. AmA Epi~demint 1989: 130:338?47,. 24. ChanWC, Colhoumc Mll, Fung.SC. HoiHC. Btvnehial cancer in Hbng. Kong.1976-1979. Br I.Canccr 1979; 39::182~92. 25'. Koo. LC,..Ho:JH-C, Saw D. Active smoking,and passivcsmnking.amemgw female lung.cancorpauicntsand.contmis.inHong Kung, J E:pClinCancer Res 1983; 42:367.75 Massachusetts Medical SucictV Registry on Continuing Medical Education Toobtain information,on ccmtinuinK medical cdYlt:ation courses irn.the New England,area, call: between 9:00 a.m. and 12`.(DtD noon, Munday throu);h Friday, (li',I!7)~ 893-4610 or in Massachusetts 1-$t?0-322r2:3Q3;, cwt., 1342. If writiitg„ direct correspondence to: Program Registrar, Massachusetts Medical Society, 1440 Main St:, Waltharn, NIAAf211 5 4-1 64 9. -Che booklet is tretc to MMS members; b5.0©~f(rr nonmembers. 11