Tobacco Documents Online

lNt.J.Canrer.•83.h3s-G3Vt19991lkY~ , ~ Ja8„2Q76 O 19'N Wdey-Liss. Inc. EXPOSURE TO ENVIRONMENTAL TOBACCO SMOKE AND RISK OF ADENOCARCINOMA OF THE LUNG Paoio BoaEYTw'•. Wol(gang,Wa~~z=• Fn:drik NrnEac" Mush MwcsRU•. trcrx BROSCE-kiatl.t•ELd. Crisciaa RmsFA• Va)( CoxsTU.'rrnESCti''• Lorcnzo Sraror:aro`• Ha)iru BAn.'awly,.anvu••. Su>anne LEw'•n'• Valerie GABORiE..L' and Sitrwtte B>a:wwoW' 1fnrernurtnnal Agency far Resrntrh an Cannr ffARC1 L)r.n. France '8rca+rn fnmitate firr Pn.rnHnn Research and Sorid Medicinr fB)PSA Bmmrn• Germaa,r 'lnsr(rure of £nvitrmmrnrot MedfcFne, Xarolinrka lartirate. SracYhMnr. SnYden 4lnsrirureaJCarc(nagtaesi; CannrrResean•ACeaar• Mauow: Rwssiaa Federation 'Narfaec! Reseatrh Cenr~~ fimErtrimnmenr and Nmhb (GSF), Manirk 6ennnany •Eptdemiolo8x Unit lnritnn Region. Rowe. Italy 'Nmiottol lnsrltttte af PrNit• XealtB, RttcAatest- Rommria s Venem Cancer Rtgisrrr Padua- ltal) 'DtParrnrmrofLnng Disseases._Medlcd School. Po:nan• Poland taSc)taof ofPa6fk NeatNk Utrinrrsity ofCalifanua at Rrrketq: 8erke[n: CA. US.t ttNatfmml lnvituteofNealtAaeAMedical R[ssarrh(lNSERM) Unit 351. Yllejujf Francr Wa condueted a arfe•ctanbW .tudp of adanocarc6toma of tha lung and tpw ta.nvfrantnanW mbacce tmoke (ETS) M 7 eeawwin. Wt intarvlavnd 70 easa of tdenoearei• rw.na oNAe iury and I7f peydlatlen or hospital cont.Wa. A6 wbjac4 had anehad ifew.r than !Q6 dgar.rtas 8r their . YMHasp.Evar eryeaure.to•E7S i. otn ehnpar.nta during . eftlldlroed .wa associated whh a.d.er.aaal Hak fodds .asfe (O!t) 0.a. fg%eurNidW40 irroaeval (C() 0.7-).Tjeatrd th.nnweas n wgge.tlor~r et.r.da.aaaYegtrand in ridrw)tltYtmwirtg w~u wt Maszociapd i.ph ae k~iw~isad r4k (OR .&. 957[~C( - OS-1.Q, wh(fe fhe OR d.rar .xpoaurs to ETf at the . werfplse waa~(s (fsx u o...d•o). For (wus eapmva. •wres,.an( aaad rtdc-rras e6afe..d arnwtg the AfgWy aape..d, artdeha OA arrrorrg tlwa. w(M tfi. Mg(wc dnratinn of enpomr. te Ff44orw fMapeute er.e tla. worYphca wa L!(95%C19S-iay A.Y.~.rr(drwaaesdnnt.dfoecureent prponvr to.FTS [torrt.Wrr• nnueew Our.a.arlta tenR'm peav)eua raporaa eta waak.l7ace af aduk !7•5 upes.w~. ae elYi.d ~ ot 1M ia.g- 1lhaandcenfnurding sa.rweb.uteludedasa~mnfartl»npqrened.a.aaa in.Nk Lam ehitdheod .>~wrra (ea } Ceec.r t7~it-g39, flN. e/999NSIn-lisr,lnc- . . , MATERlAL AND METHODS During 19'74•1986, we etao)led non•amoking cases of )ung adeooatcinana from 9 ceatm in 7 countties:S(ockholm CSwe- den)- Paris (France)- Bremen and Munich (Crnnany). )3dua. and Rornr: (uly). Pbznan {PoiandL Moaeov. (Ruuia) and Baeharest (i(omaaia)• 71lese suhjects-were pert of alatger-study.itned ar assasing the tate of mukers of Individual pmoepp-DiiGy to lung uacer among nott-smoknts(dara nat shown). The subjects tmm Swedee wam alsopart of a(xevious(y reported ma)ysis (NyheB et aL. 1998u1n in Sweden ard Germaay.eotrmota werc sekttM atnong aao-s(notett from the underlying (wpuluion, wtn•(e:n Padnaand Romntia, they included both subjeasiinm Brc undeAyi.; Popala- fioe and hospital Fatiaus• in the otlrcr mmes- drcy were selected mwng tqrtsmo)nng healthy iMividu.bae patiennadmit(ed to the sqnc hospitals as the cases. Conuois rae froqueacymatdtedto eaxs aa age and gendrr. Patients admh(ed to the hosp(ula for oobaao•ee(atcd diseatts were aat.msidett:d as pa(naiW controls. CYses aadcaniro(s watadmin(staed a snnduNiu4puWioovsire, which includcd detailed sections.on oe:aslooal~ stpopag. ETS expau(e during ehildhaod. ETS,eapusute dw(ng adui(hocd Rora ihe spause md at the ivorkplace, occupaatonat eaprtsuras, diet and famBy hislcry of cancca The.aection on ETS exqisute had been -usedinanearlierstudycondunedinE.urope(Boffemerd 1999) inaa uriaa mr(ninc measa<emeMs 60~ ~~ a ~ ry ga Expvsura,ro eavironmenla) (ohacco smoke (EiSt has Eeen ~ associatedwi(hanincreasedriskoflun eancerin <RihoRnal,(990A s(udiet ' Noah Atnetica• Asis and e~a~~ ~' defined as twn•srnukers (lase cases and conuds aM during . eondti[Y6"d-r(t Ernope fHactshiw theyr Gvee had snwked (ewer than 400 agaedtes nr the tqvivakart er af,_ t9971• We have reported the results or a mo)tfCCtme carearotarol studyofiungcancereonductM in 7? 6ti~eancemres araoua(ofmbaceo.fmmcigara.ciguillosapipe.l7(iscarrespalls (Boffew er at, 1996)• In dat .swdy-.vhich is Mc tasges+ such so about J rigarene a daydueioQ I year. Noa~aniokets ieclmded in the investignrion eondueeed- in F~ T rocc daected a mndcrate `m- ~y were classified according to ever eaptuure (d ETS tmm crease in risk ot lung wncer from ever expnwtre ur'ETS ftom the rhe modter a thc fathet• duriwg childhood (age 0-(g) atnl ever spoarse Jrela((ve tisk (RR) 1.16. 95% cnnhtlence interva) (C() K~° R°R during adulthood to E[S from the sp9use. Ef5 at the 'YarkPlace nr bdh- Further E7S Caposnrc vatiables ffia( we 0.9-7-614t and at t)e avskplacf•(&R 1.17,95%C)0,94-1.35)with dexNnped include: L chlldhand: ever exposure to fa(Mrs ana a suggesitonotan incrcasittg lteriA in risl with ine.easing dursrim rnatlrcr•s smoke. u><aldararion of e.pasure and""duration of of exposure to either source of ETS or a cotnbination of the t+w.' exposure weighted for dro typt of smdcer (rnoUxr- i: Cather. O75k - Additional important tesuhs of aar study included a higher risk tor ?spoase: cumulxite eapnsure (number of cigarettes smoked per aqtumous•ceA and sma(1-ceN carcinou(a than for adenocarnihnma. day by the spouse in the Presence dlhe inde: xnbjrctmultiplied by a decrea.e in ri.dc with increasing rirne s(nce slapping ETS thenumberofyearaa(eaposure).duratianvfeapowre(tmmberof . expourre, andlack of tuhg cancer risk assax9ated with I:TS daily hours of expowre trw0iplied by thenuiisber of years of exposure during chi)dhand: •)7K RR fur any chiidttaad exposurt esposure): 3- worteptace: duration of expasune (numbx oT daily was 0.79 (93'A Cl 0•6i-O.96) with a suggestion of decreasing risk with inereasiogdnrarionoft:pasute. YVe conducted aa addil(ona/ cnse-conrrol srudy among non- •Conespondmcro ta' Ihdt o( Fn.ironrtsma( Cancer Ep{0eniobgy smokfngtungad~nopreiaomaeasesandconuolctoassaswhe(her feretnaiaulApeneyferRaean9retrCanrec /SOcounA)brrt•7homac . out results caold he reproduced iaan indeperaknt pqwla(ian and F-69M Lyun• F.anor. Fa.: +33-(-T73ga42. E-mmT: nw(anaetan-fr .. . topiovule additional evidence nn the association between Fi7S . expOSUreatdiungadenOCardnomainEOtnpeanpopulaGons .. ' Rettived?3Apri(1999;Revi.ed2glane1999

636 9()f FG1TA LTAL tA6LS i- xElt:elEDCHARA[TERISTKS DFTItE iTt'DV w/PULATlns [~ faqnA. Nr9n~ Mi11a~ n Age Rmt91(years) 18-54 11 15.7 48 27.0 53-b4 30 28.6 63 35.4 65-74 31 44.3 42 23.6 75+ 8 11.4 25 14.0 Gender Men 4 5.7 41 23.0 Womm 66 94-3 137 77.0 Cannrrv SweAen 16 22.9 25 14.0 Gennany II 15.7 67 37.6 France 4 5.7 li 6.2 italy 9 12.9 31 17.4 Romania 9 12.9 14 7.9 Russia /6 229 20 I12 Poland 5 7.1 10 5.6 T{u1E n-DDDS RATIUt OF Ll 80 ADEMOCARCmOn1A iEn'.t El(POSLRETO ENVIRDN~ MEWALTDaA(YC15.NOKE OF AARFMS DIRI`/G CNn.inf00p ,.Ir*~ M.xs N01P~ Wtawda nR %'ifl Un<xpnred 32 55 1.0 Ref. Everexposed' 33 110 0.6 0.3-1.2 Exposed to father's 33 101 0.7 0.3-1.3 smoke Exposedtonwthefs 3 18 0.4 0,09-1 B smoke Dvrotiun of espowrr (.etg ted yea/tl I-10 8 2 .0 .4-2.4 10.1+ to so 0.3 0.1-0.9 Trmd 0-06 OR: odds ratio adjuswd for age, gender and cenoe; Cl: con6dence intervai; R<f~ rcfercxe casegnry. 'S cases and 8 controls bad missing infonnatiaa=l0 esposed cases and 21 exposed controls had missing IMOnnation. hours of exposure multiplied by the nulflber of year8 of exposure and a subjective inden of smokiness of the wotkpiace); 4. spouse and workplace combitled: dundnn of exposure (sum of the indices of duration described above wititout the weighting by workplace smokiness) and tilne since stopping exposure,from both sources - For quantiwive adulr ESS egposure vmiabks, exposed subjects were classified In 3 gtoups comp.iang appmxidutelY the bottom 7556. the next 15!F aeW We top 105% of the cmtlrnB. This approach was basnd on the sesults of IS urinary qi(inine-sudy showing am taisclassification af~ETg exposure is greater in the 3 lowesu qunGlaa of the dWrlMriian than in detop quartlle (9edler ef at.. 1992). For duradon of exposure to ETS during ehildhaod, only 2 eategories wen: used, with the culpolim a the median of the distribution among cantrols We retained in the anslysis only histologically conRrmed cases of lung a&nocaci4atla;caus Gaaified as mued adenocascinoma oralsodser65atobgieal.type.weseexciuded. . -'.. We.fined multiv.rhae logistic fegression models to estimue the udds ratims (ORs) of long edenacaseinwna froln exposure to ETS from the dilferem satsces and thcir 959: confidence intervals. All the regression models included tesms for'age.110-Year groups). ~ gender and cenue. hs addition. wve gfteA models including the - potential confounders: asbw.tepidenees edlseatitm uld exposure to occupationat cascinogens. We tested the signiftcanoe of the lineu trend in risk across increasing alegor'ss of quantitative ETS exposure variables by fi1Gng an ordinal varia6le. We conducted additional anaiyses. restricting the sfudy population to women and separating young and oW subjexts with a cwpoinl at the age of 6.5. RFSU7.TS A total of 70 Ilon-smoking adenocacinoma cases and 178 non-smoking controls wese included ia 1he study. Their distribution according to age. gender. and centre is shown in Table 1- Cases (mean age 64.4 years) were Cider than controls 160.9 yeanr. 5.7% of cases and 23.0% of wntrols were men. Afl countries contributed u least 15 subjects to the study. One•htedred and seven controls were healthy individuals. and 71 wete hospital patients trearcd for a variely of diseases. including orflapaedic tondilions (N = 17). non-toh.eco-relaoedma/ignantneoptasms(Na i71.acWerespirs- (ory conditions (N = 14). artMosis (N s 5). digestive diseases (N = 4) and benign neoplasms (N = 4). Exposure to ETS during childhood was reponed by 33 tases and I 10 controls. resulting in an OR of 0.6 (95•k Cl 0.3-1 .) (Table 11). More subjects reported exposure to father's smoke than to modier's smoke, and the decrease in risk was more prmlalncedL although am s[aPstically significant because of small numben for the laner type of exposum. Three cases and 9 controls reported exposure from TAaLE DI-ODO5 MAT1fMOFLU7lG AIIFIiDCARL910MA FRUM FXPnSURE ln FI/YIXII%MFMALRInKCOSWOKE FROS1 7115 stOUSE Numuee ofmss etueie. W.c.acl. Da 95A a Unexpnsed 37 103 1.0 ReG Everexposed 33 75 1.0 0.5-1.8 Dam/ion oje+pasumr, (6arra/ ~ d4r X rpet) 1-120 19 44 0.9 0.4-2.0 121-758 6 11 0.9 0.3-3.0 259r b 7 1.8 0.3-12 Trend 0.8 Camu(mive upumn' /parFwean) 0.1-IAO 20 48 ' 0.9. 0.4-1.8 14.1-26.0 5 12 0.9 0.2-2.7 26.1+ 6 8 1.5 0.4-5.9 Trend 0.9 OR: odds ratio adjusted for age. gender and centre: Cl: eon8dence interval: Ret: rekrence case0ory. . '4 expused tases and 13 eAposed oonuds had mitsfng informs- lion.-%2 exposed cases and 7 exposed controls had missing inforsna- don. both parents (OR 0.& 95% Cl 0.2-3.6). These was a decrease in the risk of lung adenocarcinoma according to duration of exposure to ETS during childhood after weighting this variable according to type of smoker. Results were similar for the conesponding unweighred variable (nor ihown). The reduced risk of lung edenocarcinomarwas present among women and men, while the trend in decreasing risk was clearly present Only among subjects bclowage65. . Table atl presents the sesulis according to ETS exposure fsorn the ~ spouse. There was no overall Im:rease in the risk of lung adenocar- cinoma among the subjects expnsed to spousal ETS (OR I.O. 95`.f CI A non-signi6cant increased risk however, was ptesent in the highest cmegary of duration and cumulmive exposure. these two variables being correlated (Pearson correlation coe(flciem amang conools 0.81. p< 0.011. The tesults of the analysi* resnined to women were similar to those based on the whole study population, since there were no male cases andonly 6 male controls exposed (OR of evcr exposure amalg women 1.0. 955F Cl 0.5-1.9). No clear pauern emesged frum the analysis based on differem age gmups. The main tesulu of the analyais or $fS expD.ule at the wnrkplace am plesenled in Table 1Y fitposure was repated by 38 cases and 97 controls, yielding an OR of 1.5 (95% Cf 0.8-3.01 without a clear indication of a dose-respotne relationship for duration of exposure. The risk of lung adeslocareinoma from

PA.CSIV E$M(1KIMG AND Ll'AG ADE\(X'ARCI~A 637 TAa1E IV -ODbS RA71US OF LUNG ADETkIfAaC1VOMA faOM ExPU6URE t0 EN V IROttM ENtAL TORACED SrrOKe AT TRE wtMKVI.\CE Nu~4er' ,~SnM aR ~ [T Unexpo•ed 31 <81 IA Ref. Everexposed' 38 ' 97 1.5 0.8-3A Dumri.in ulorpo+ure~ fLrwrr/ day X yean) 1-61.2 6 4 .9 .9-4.1 61.3-157.0 4 13 0.7 0.2-3.i 157.1+ 6 8 1.7 0.4-&6 Ttend 0.4 OR: odds ratio adjusted fm age. gender and emue: Cl: confidence imerval: Ref: refermee alegory. '1 case hid mfasing Mfommtian.-=2 exposed eases and 12 exposed coMtuls had missing informatiat. TA1G6 v-ODDS RATI0a0F u'NGAOENOCARCINWU FROM EXPOSVRE TO EHV[ROr.MENrACTORACtO~ ~ TREaPO1rs6fAlATTRE ram6n w~. N~tr{r or <wa, oii oss n Uaeaposed 20 52 1.0 Ref. Everexposed 50 126 (.2 0.6-2.5 Durarion oJesparurcr (haa.r/ day x years) . I-II1 30 83 (:L 0.6-36 112-200 g 20 1.0 0.3-3.F 201+ 9 11 1.8 03-6.2 Trend 0.5 ymn rinre )an eq»- aror 15+ 9 40 as D.2-1.5 3-14 19 36 1A 0.6-3.3 0-2 22 49 1.9 0-g-s.5 Trend 0.08 OR: odds tatin aajustrd far age, gentlrr and canne: Cl: ron6dence ialavd: Ref: reference caugory. - 13 exposed eases and 12 exposed cenaDls hed missing iafonna- tion,) expaatd cowol had missing information. workplace ETS exposure was h(gher in tuen (OR 2.6. 95% Cf 0.05-(40) than in wamen (OR 1.2. 95'k Cl 0.6-23). but this difference was not statistically stgni8cam and the risk estimate in rnen was highly unstab(e. TTe increased risk was present among subjects who were younger than 65 at the time of the study (OR 4.3. 93% Cl 1.6-131 but not among older subjects (OR 0.6. 9555 Cl 0.2-1.5: p-va(ue of ditference 0.004). ETS expowre either from the spouse or at the workplace was associated with an OR of 1.2 (95'kCI 0.6-23) (Table V). Increasing duration of exposure showed a nan-signiNcant increas- (ng trend in risk (pwalue 0.4) wit)t subjects in the category of longest duration most clearly at risk. The tt.whs, after stratification of the study population on gender and aged suggesied a higher risk in men eban in wamn, as well asanqngn younger subjects as compared with older subjects (nsults nm shown). These results ate explained by the higher risk from E7S eaposure ar the workplace among tnen and among subjects below age 65 (sec above). An effect of stopping ETS exposure is suggested by the results obtained according to lime since cessation of ETS exposure from the spouse or at the wrukp)acr. the OR of lung adenocateinoma was 1.9 (95% C1 oA-45) among subjects currently exposed bm only 0.5 (954+ CI 0?-13) among those unexposed during the last 15 years (Table V). This effect was present among women and men as well as among subjects above and belaw age 65. Adjusttmm for exposure to occupational ctrcinogens w educa- Gon did na modify the results substantially. After adjusting fot' occupn0onai cxpo.sure, the OR of ever exposure to spouse or workplace ETS exposure was 1.2 (95% Cl 0.6-2.5): that of the highest category of duration of exposure to spousal or workplace ETS was 1.8 (955b CI 03-6.2). Adjuuman far education resulted in ORs of 1.5 (9396 CI 0.7-3.3) and 1.6 495(k C10.4-6.1) for the same ETS eitpodure variables. Information on urban residence was missing for 24 ceses and 39 conlrols: the adjusunenr suggested a decrmse in the risk estimates and in their precision (OR of ever exposure to spousal or workplace ETS r.l. 95% C( 0.5-2.8). OISCUSSION Our present analysis was perfomed to usess whether the results o(e krge cato-wntrd study of ETS exposure and lung cancer we had conducted (Boffetra erat.. 1998) would be repiicated In another study population and to provide add)tiwwl evidence on tlte assorimiwt between 6C5 expotrm: and lung adenoeax:inoRU in Eurapeen populuriom. YA testricted this sludy to lung adenocarci- notna, the most emnmon histoMgica( type of lung cancer among non-smoken. mainly due tu the small nnm6er nf eases in our study with other hEsrologieal types of lung cantw. The carcinogenic effacfnf tobacco smuke on edenorare(oontas has been considered weaker than on other hfstologiea) types of lung cancer. namely squamats-tell and small-cell carcinoma (fARC. 1986). Other evidence. however. 8uggests ahat the difference in risk might be smaller than previously thought f Wo- Wl(llams and Same[. 1994). A number of epidemia(ogica( studiess have reporoed results on risk of adeaocazcinotna of the lun8 following espowte to ETS. mxioiy frwa the spouse. SigoiBwnt (or borderline significant) increases inn tisk. with relative risks in the ordv of 1.3-2.1. have been reported in studies from the USA (Oarbnkel er at.. 1985; Fomham er ae.. 1994). Gttas (Kalandid( ev ai.. 1990}, Russia (Zsrfdae n at., 1999) and Hong Kong (fam nat, (987). A similar increase In risk tdbeit not s)gn)6rata, has been Teported by Stockwe0 eraL (1992) atd Brownwn er at. (1987) in the USA and by Koo ii at. (f987) in Hong Kong. Only I study. from the USA. failed to detect an fnmeased risk (Brownson er aC. 1992). In Europe, a study of Swreisfi non-smohing wranen foundan increase in the risk of kmg wncsss other than squemous- and smallcell cuciaarnes e0ereunud.t)ve ETSexposurc from the spouae of22-5 or morer pack-yms but not fur lowa expenue (Yenhagen er af.. 1987). In a muh'uzwe study fnom Wetrem EtRnpe- the OR of aderwcarcinama follomag ever expocure to spuusal ETS was tow. 1.0g (954 CI 082-1.42): the OR fol(owing woflcplaee exposure was 1.06 (93'.i: Cl 0.81-1.40). and the OR fur prolonged duntian of exposure from the spmse or in the wmkpiace was I.58 (95SF Cf 0.98-2.54) (BoBeM er a1.. 199g). In most of the stediea reporting rcsults on di(ferent histo(ogicaf types of lung cancer and ETS exposure. the increase in risk of adenocarcirwms was smatler than that of squamous-or smallsefl earcinomas (Akibder at.. 1986: Btownson rr aL 1997: Gao et at.. 1997; pershagen er aL 1987: Stockwell et at. 1992: Fontham er at. 1994t Bo6fetu er, aL /998: Zsridte a al_ 1998). In 2 studies from Hong Kong (Koo ea a1.. 1997: Lam er ol.. )987). 1 ffom dx USA (Garfinkel er aL. 1985) and I from Gteece (Kalandidi erat.. 1990). however, the risk of ndc~mcarcinoma was equal to or higher than that of other histological types. Despite the lack of statist(cat signiMcana. the rewlts on adult exposure cort8rm the presence of a small incmud risk in lung adcnocareinoma following adult exposure to BTS. In addition. the finding of a concentration o( the risk anqng the most heavily exposed subjetu is consfstem wilh a earcino@enic effect. The replication of our, pprevious finding of a prutect(ve eflect of stopping ETS exposure adds in the evidence thar this dimension of ETS exposure is an important ale tfat should be carefully examined in faturr studies conducted in populations in which important reductions in the prevalence of smokers has raken place (Nylmrg et af., I99ga). We observed a decrease in adetKUarcinoma risk following childhood ETS exposure. a small increase in risk after spousal or

6ig ntNFkTrA 17A1_ w'orkplace exposure with a higher risk among subjccls at longest exposure. and the suggestion of a decrease in risk following cessation of ETS exposun:. All these findings parallel Ilksse of a larger investigazian we conducted in t2 Western European centres. which included subjects fronl 6 ccntres panicipating in this analysis (8dfeua re al.. 1998). Minor discrepancies can be primarily attributed to statistical instability of the resulu of the cumenl study. We interpreted our previous hndings of a decreased risk of lung cancer anqng subjects exposed to ETS during childhood as possibly due to chance lBoReua rr al., 1998). A previous study reported an increased risk following ETS exposure in chiidlqod but aat in adulthood (Janerich ar af., 1990). Most of the largest studies, howevar, failed to detect an asaociation to all lung cancen. We are aware of S prerwus studies that, in addition to the mullicentre Eeropean study mentioned above, have reported results on risk of long adenocarcirtoma following childhood EiS eapa- sure. Stockwell et aL (1992) reported a modesl. non-significlva innease in the risk of adanocacinnma following exposure to ETS ftvtn the madser but not from the father. The results of the remaining studies (Pomham er at-, 1994: Zaridre et at., 1998) a/e crmsisseensly negative- no matter which indicator of childhood ETS exposure was used (either parent or bottt fwteon being srnokers. number of smoker-yeanm etc.). In the Swedish study, subjects with at least I smoking parent had an OR of OS (95 b cr 0.1-1.9) of lung tancers other than squamous- and snrell<e3l carcinoma (Pesslsagen rr al.. 1987)- A decreased risk (OR 0.6, 959: C101-1.7) was also reported in a previous US study(Wu et al.. 1985). A possible eaplana[ion of our findings is seporning bias- i.e.. casca of lung adCnneusinoma reporting less 8equently a history of ETS exposure in childhood than enntrals. The lack of a similar biu with regard toaduk sousces of ETS exyosure- however. reduces the credibility of tltis e%planation. Similar argumema ean be used to reject the hypotheiis of selection bias due, for example. to the sekcsion of hospital conrrois For 5 casea and 13 controls. infomutiwl on childhood ETS exposure was missing (OR in this group 12. 95% CI0.3-5.0). leaving open the possibility of a smail ••missing vahtf-bias. Negative confounding by a putective facwr associated with childhood E'l'$ exposure (or a-risk factor with.ah negative association) is anotla<y possible explanation. When we repeated IMaaNysis of childhood ETS eapomre af0er adjustmcnt for urban residenGe, education and expawrc .fa oocupazional carcinogens, vx obt'aitr.d an OR of 1.0 (935E Cl 03-3A. basedon 110 subjects wiM compteteinformatien). Oespite the tack of precision of oui results- they suggest that the apparent protective effect of childhotsd ETS exposute might be due, at least in parl. to negative confounding from other risk factors of lung cancer in non-smokers, In addition, the decrease in risk was restriaed to ccntres wilh hospital controls (OR 0.4. 95'k Ci while cemres with population controis had an OR of 1.2 (95t1- Cl 0.4-3A). suggesting a possibic bias from the use of hos- pHal controls. On the other hand- we round no oarreLtion between ETS exposure during childhood and during adulthood (both from IM .pouse and at the workplace), suggest'atg a lack of reciproaal sronfounding. The lack of consistency with the results of previws studies, with the exaxption of the 2 investigations mentioned above (Wu er at.- 1985; Penhagen ee af., 1987). points toward a non-causal interpretation of our findings. If the protective effect is teal, its underlying mechanism tertlains to be elucidated. Our study sutften from a number of limitations. The use of a series of hospital•based controls may have resulted in a bias toward the null if ETS exposure was associated with sotne oftlse diseases of the coturols.'Ihe main results of Our seudy. however- held also when the 2 series of controls were analysad separaPJy. TAere was no difference in response rate between cases and conuols, reducing the likeh7wud of selection biaz. In studies of ETS and lung cancer- nen-0ifferential miselassification between cases and aawro)s of their non-smolung status ard their reported ETS exposure are 2 imponant potential sources of false positive results (Iee, 19gg). An objective meas,aanent of tixse variables was lacking in our audy, and we have no direct evidence for or against the presence of a bias. In a pnevfotss validation study based on cross inierviewsto relatives of cases and cantrols from 3 European centes, ineluding 2 ssnues participating in this study. wc found a very slrWl proportion of subjects miselassiRed according to eisher their own Isoe-smokiaf -statusrrthesmnkinghabilofthespouse(dW is.lheirownspousa,j ETS eaposwe snws), withoat evidenR of a higher proportion of miaelassiged cases as compared with controls (Nyberg er at.- 19986). The higher risk frosn workpleee ETS exposure among younger subjects than among older subjeccs suggests ehe possibility of nondiffeseatial misclassification, due to poor tecall, among older subjects. In conclusion-our results confirm in an independent population the main findings uf a larger study we have carried out a lower risk of lung. cancer aaamg subjects reporting ETS exposure during hiidhond, a smalt but siaeab)e risk after exposuse from synsat or waakplam ETS.with a suggcstion of a t/me-responso 7eluionship. and the suggession o( a decrease in risk after eessnion of ETS exposure. _ AGXNOWIEDf.EMENTS SL and F1V worked on this uudy under the tenure of Special Training Awards from the International Agency far Research on Cancer. The study was partially funded by a gram from the European Commiseion DG-Xll Contract No. EV SV-CI940t555. REFEReS1CFS I ' Ar.taA- S.. Kwro. H. nd Bwr. WJ.. Paaive sinoting and lung cancer among Japanese women. Caxer Bex. a6, 4R04-4Rn7 (/9R61. Bt_clRn-H.-Z..7o.YSW. W-andJocxu. K.-H.. Pauive smnking iaGemuny and PolaM- vrlipaiison of expnsurc lesxls-sources of.zpoxvr..-ahdity. and per~,Vtion. fjvdrniir(nys. 3, 5f19-5141 t99?t. Bolsarrn. PR and 26 urrmae Muhkenter case-comrot uudy of erposum to ensironmemal tabattv smoke and lung cancer in Europe. J- nnr. Can'er Mn.. 90, 14+K1-1450119981. Hxor.vson- RC. Auv.~u M.C.R.. Hact E.TT and Lnv. TS.. Pwsive smoking and lung cancn' In nnnxmok:ng anmen. Mlrr. J PtdJ- Hrana. 82, 15'3-1 5 3041992). Brtor1sow. RC.. Reu. IS-. Krirt_ TJ.. Ftsccsou- S. W. and PwrrL 1.A.. Risk faetors fur adenaa[einama of the lung. Am J. EMdemied.. 12S_ 25-Ls 1196-h. Fo+-rwsr. E-TH.. CaasE.. P.. Rsvwt.us, P.- Wc-Wlw.sa-A.. BL1st7.ra. PA.. Gaetneurp. R.S.. C7tw. V.W.. ALTSxxun. T., Boru- fl. Acsn.c D.FF and Lav. I.. Faviromneraal tobacco vmlkc and lung cancer in nonsmoking women: a mWt4emcr imdy. J. Anwr. Med. As.n+r-. 27/. 1752-1759 (199q.- GaG. V..T- gtar.WJ., ZMEtc. W., Fn~, A.G.. Hsv. C W_ Lsaa. LI.. ?H-NG. R-and Fawtaal:wl-l.F. Jr_ Lung cancer am. mg Chinese svomen, bu. J. Cwaet M0 aOf-6o9 t 1967s ~ Gnanaeei-l~.KTaawcr.O.aMJOfata*-L.In.dunraryxnwking snd lungcanttr. acae<ontmi stWy.l. ma. CmuvrlnxL 7R 463-id9( 179R5). H.wncsaAw.A.K--Ls..'. M.R, and WALO. NJ..Tht xzvnwlaederid~ron lung cancer and envimnmemal toluno smoke. Brir nvd. J.- 3s5, 9g0-91bt t 199/1. fARC- Tobacco smoking. /ARC Monographs n the Evaluatiun nf dn Caninogenic Risk ofCtrmicals to Hununs. Vol.3& (ntcmatimW Agency fa Research an s3ncer. Lyon t 19861. 1.rvsan:w- D.T.. Tww4non. W.D.. VAast.... LR.. GaaFnns-Aw- P.-Ctauncr. S., Tlm-C_ 2w.vAr:. M.B.. MsiAwsa. M.R.. Km.r. M, and MdCwu+. M.F.. Lung cancer and expnrue to mbaccn smoke in the Massehold N. EngL J. Med-- J2J.63 ]-6361199Ut Kat..nnuw. A- Kwrmerwrrvt- K.. Waawtzoc. N.- Btitrwi G., S.awra. R. and TuctfaPaa.os. D.. Passive s.mkirtg and diet in the etiology of Iwp cancer amung noo-snpkers. Cuarr CausrsComroL 1, 15-21 (1990i.

fASSI VE \i1fiA13C Av1i LC]Ca ADEVOCARCINt1\IA 639 Knn. LC.. Ho. 1H-C_ S.r. D, and )Pt C.Y. Meuswemnn. af{~..i.a su.,king and essianun of Iong canc<r wJ anemp n~.n.+king Chin<.e femalez. lra. l. C.wrer. 59. 1A:-169119g7/. l.w. T.H.. Kawc. LT.kt. Wow C.M_ L..w. W K.. K~vsxs. J. W.L.. Swr. b.. Nst~ C.. Saaa<vaaATf[. S_ La.a. S.Y. ttr. K.K. and Ca•%'. W.C-. Smotipusive smokiag andhistdogicai 1ypes in Iury uncet in Hong Knrg CBintic womrn. Br:+.l. Can(er. Si.b73-678 t 19117). . L¢E P.N.. ffixJnx.j/4vtkn aj mnlraA haNls and pa.uiLe snntina- u mineoJde mt?wwrre. Inc An1c Orcup. Em iron. Health Suppt. Springer. B<rRnl19Bg1. Nr9Ew.F_AGneMlus. V., Sv,µT[1+wEa. K.. $ve,ssuw. C. and PERSH~E*. G.. Environma)q) roWeeo unoke and lung CWroer in nonsmoksn--does rimt since exprurue }day asale? EpiAeatiWng.; f.3Q1-308(109ga). NnsJw. F.. Atodo. A- Bdfeua. P.. Fanes. C.. GuvlleL CA. and Pslar.e4+. G.. AEiaopnnvafidwinn sludrof smeking and environmenut sohacco smoke eaposme in nonsmokin` Iumag cancer ca.ec and canvnls. Cancer Caates CauroL t 173-182 (199gb). . P~.O.. liM~ Z and SYENSSOX. C. PaEiv fnqking and lnng cenrer in S.edish .onl[n. Aare !. Epldewud_ k2S, 17-14 (19g7)_ kuaur. E.. and )h aher.. Expoxurc nf nnnvnnkin6 n.anen In envirnnmcn- ul tMaiso xm.ke: a Ill+:..miv) coltahxative sndy. Cun.rr Causes CewenA. 1. 23T-?5. (1991i1. STOCm.nt. H.G.. C~o.....+. A.L.. Crs+". G.H.. Nuss. C.1.. Ae~ A.W.. Rvw(wM, P.A.. C~ue[oa.. E.C. and Bauss. M.R. EnWtanmdttd tnhxcn .moke udryung canser risk in nnku'wking wmten. J. na. Concer Ant_ K 1417-1411 (fD92). YrY. A.H-. H[+vexsati. S.E. Pllca. M.C. and ti•. M.C. Smdcing and wher risk hclps for Wng caneer in ~ncn. !. nat. Cancer lnsr., 74, 747-75I (19g51. We-Ww.MS. A.H. and SnxILr.1.M.. Luag cancer and ci€uene smukin;. h: l.M. Srtnet (ed./. Eyide,wiobg,v nff~.~ snneec Lung B~o1o~y 3n Health and Dise>se. Vol. T4, pp. 71-IRS. Marce) bekker. Nen Yrck (199C). Zxunm. D.. Mnaaavcrcr. D.. Z~r,.wsrA. G.. MswKOV. 2.N. and S~.. R. Exposure no environmeam9 aWxsro srndro and risk of 1u~ny eaneer in oon-smoHng womat ftwn Mas~mx. Rnssia. Ju. J. Cmnr. 75, 335-339(19Bi).