Tobacco Documents Online

III THELAKCET,SEPTEMBER10.1983 595 PASSII'E SMOIXiG AND LUNG CANCER 1+EtAY0 CORREA LINDA WILLIAMS PICKLE ELIZABETH FONTHAM YOC'PIhG L11: WtLL1AM HAENSZEL IJlrparrmnu of Parholop, louuiana Sratt Urtrcrrrrt}° INcdrcal Center,Neu t7rltans, Louirtvna,' ErruironrenualEpidnrtwlvrv, BraneA, Nartona! Cancr+ Inrniurs National IruritYtts of Healrh, Brnceida, Maryland IlLnou CancrrCauncil Cl¢ago Illiroii, US9 Ssrrr,trrrary Questions about the smoking habits of part-nrs and spouses were asked in a case- control study involving 1338 lung a=r patients and 1393 comparison subjects in Louisiana, USA. Non-smokers married to heavy smoken had an increased risk of lung cancer, and so did subjects whose mothers stnoked,'I-here was no association between lung cancer risk and paternal smoking. The association with maternal smoking was found only in smokers and persisted after controlling for variables indieat ive of aaive smoking. It is not clear whether the results reflect a bio)ogical'effect associated with raaternal lsmoking or the inability to control adequately for confounding factors related to active smoking, This preliminary finding deserves further investigation. Ilntrnduction THE possibilit}{„of passive or involuntary smoking being a causative factor in Jung cancer has been investigated in several countries.' This report describes a case-control: study oflung cancer in Louisiana in which questions were asked about the smoking habits of the spouses and parents of 1338 )itng cancer patients and~ 1393 comparison subjeas (controls), Materials and Methods Current primary lung cancer cases were identified from admission and pathology records of all participating hospitals in twenty-nine Louisiana parishes (counties), which ineluded' all southern, one central (Rapides), and two northern parishes (Caddo and Bossier). Patients with bronchioalveolar carcinomas (32o.ses) are not included in the present report. All major hospitals in the study area participated except some in the eity of New Orleans where, for, logistic reuons„interviewing was deliberately limited to two large hospitals serving the medically tminsured population and two large private hospitals. For ach subject a control was randomly aelMed from patients attending the same hospital and matched by ratt; sex, andage (within 5 yean),' Patients whose main diagnosis was emphysema, chronic bronchiiis, chronic obstructive pulmonary disease, or cancer ofthe larynx, oral avity, cesop}iaguc; or bladder, were excluded from the control selection procedures. The admission diagnoses of the controls were distributed in the following categories: ardiovascvlar 15-3%; 6astroiatestinal 13%; musculoskeletall 10%; senitourinary 7-3%; ophthalmology and otorhinolaryngology 6-6%; otber ttunotus 5°-b; diabetes 5Wa; trauau 3•7aAb; penpheral vascular 3-7%; pulmonary 2-7%; cerebrovascular 2- 5%; andinfections 2%. Local professional interviewers, trtuned for this investigation and thoroughly familiarwith local culture, interviewed subjens(7646of the cases and 89% oftbe controls)or their next ofkin. The questions covered occupation, residcncy;dia, smoking anddrinking habits, health, water supply, and other relftted items. Information elicited on the smokinghabits of the spotuc or parent included type of maTCrial' stnokedl duration of smoking hsbit, and dail} amount. Questions on parental habits referred to the period "during most of your childhood". Histolbgical confuntation was obtained for 97% of the cases. Nlissing data were excluded from the tables. Standard unmatched pair methods were used toestimate relative risks. All p values are bssed on 2-sidcd X7tats. Results Spoum Smoking We idcntified non-smokers with hutg cancer and compared the smoking histories ofthcir spouses with those ofspotues of non-smoking controls. Only 10 out of 1036 male cases were non-smokers: 2 reponed' occupationa) exposure to dust (sttret-sweeper, log{vner); 1 was a steam-pipefiner; 2 lived in the immediate vicinity ofindiutrial plants (grain elevator and cement, oil'refincry); I was married to a heavy smoker; and 4 were long-time chewers oftobacco, There were 25 non~ smoking, ever-married women with lung cancer out of 302 female cases; 2 of these chewed tobacco regularly. For I female and 2 male non-smoking patients no information was available on the smoking history of the spouse. 2 female patients' husbands were smokers but the amount and' duration:was unknown. Table I distributa the non-smoking, ever-married men and svomen according to total lifetime pack-years smoked by their spouses at the time of the interview. The relative risk of lung cancer is raised~when the spouse is a heavy smoker. Similar tabulations for smoking subjects did not show an increased risk associated with smoking of spouses, except for light smoking men, (less than 20 pack-years), who had a relative risk of 1- 5 when married to heavy smokers (41 pack- years or more). Case-control comparisons basedion current daily, number of cigarettes smoked by the spouse yielded almost identical findings, including relative risk estimates, with those presented in table 1. The apparent passive exposure effect was present in women over and under 60 years of age, although ttnalli numbers made the subgroup findings not statistically significant. Analyses limitedto cases and controls interviewed in person indiated that systematic bias in personal versus nextof kin responses can be ruled out as a potential explanation for the findings. The same conclusion was reached, when relative risks were race adjusted. Inclusion of bronchioalveolar carcinomas resulted in slightly lower odds ratios: males I- 69; females I- 77; both sexes I • 75. Parrnrs' Sinoking Xabirr Smoking habits ofthe parents strongly influenced smoking habits in offspring (;able I1). He•avy smokers were more likely than the other patients to have had smoking parents. The smoking histories of the parents in our series were associated with each other. There were 201 spouse pairs of smokers, compared with 7 36 expected if the status of each parent was TAatIE I-NOtt•iMOf.IIiG,1tYEa-MAltlIIED LLRIG GNCEa CASES AYD CbArLROts AND LwrTlMF OONSL'safrror: oFCJGARFTTPS t(YTHF]R ..if01'SES, CApreftesIIn10keCt tT'ipOlife (psck-yean) Nonc t-40 >,i/1 Maler C.ases 6 2 I 0 C l 15 1 ontro s 4 2( p Odds ratio 1•0 2• 0 Fnsaln Cyes 8 5 9 Controls 72 38 23 Odds rnuo 1.0 tL 3.52• 2. 07 JRth KlfS Odd1 rnuo (adtusted foraa) I•0 •18 Il• •PCO.o5

596 TADLE I1-CIGAAETTE USE OF CONTKOL SUa]EGTS tY SMSOKING CJtTECARY.Or THE1.R PAREhTS Father aarole'er Mel tier ewier No No Maln Noninwkers 18% s2a, 8% 27% Ea,mrwkers 25% 26% 24% 26% Current mwk'cn 56~6 42% . 68% 47% Taal'numbcr 475 484 79 880 F+n.di, N6n-anoicea 37% 6m 19~6 S6'ro Es-anokm 20% 14% tsa 17% C.urrtm smoRors 43% 20% 566. 27% Taal number 130 154 H 250 independently distributed. Classification of the status of one member, particularly the mother, indirectly conveys informatiozl on the atatus of the marital partner. When the smoking stanu of each parenr is classified separately the relative risks of lung cancer for persons (both sexes, smokers and non-smoken) with a positive paternal and': maternal i history of smoking are 1• 04 and 1' • 66, rapenively (table In). Scrutiny of the data shows that the increased risk associated with maternal smoking is significant in smoking males (odds ratio 1• 4) but not signifiant in smoking females (odds ratio 1• 2): No significant increases in risk were found in non.smoken )rut small numbers preclude adequate analysis (there was only one non-smoking lung cancer patient whose mother was a srnaken): To n:mos,e the confounding etTen of the other parent, .re comidtsred eactt subset of cases and controls for which only one of the parents smoked. The respective relative risks, controlled for spouse-smoking status, for positive paternal and maternal histories ofsmoking were 0,95 and 1-47, respectively. Thus, smoking status of the mother increases the relative risk of lung cancery but smoking status of the father does not. The effect of maternal smoking did not seem to be dose related; our questionnaire did not cover this point extensively beatue we doubted whether children eould adequately quantitate their parents'' smoking history. The relative risk of lung cancer when both pmnts smoked was 1-66; there is thus no evidence of an additional contribution to risk from paternal exposure, over and above that contributed by maternal'.e:posure.. Given the enhancing efl.'ea of parental smoking on the smoking habits of the offspring, the etkcti of parental smoking on relative risk of lung cancer oould' reflect a subtle indirect association with aaive smoking by the subjea. To control for active smoking, a logistic regression aaalysis.vas done, taking into account all the active smoking variables which increase lung cancer risk: age at which ease starrted' smoking, tar content of usual brand, degree of inhalation, use of hand-rolled cigarcttes, years of smoking, mazimum amount ~ smoked; By this method of analysis the relative risk asaociated with maternal smoking was: 1-36 (p<O - 02) for both se:es and 1• 5(pC0 • 0l ) for males. No increase in risk TASLE nl'-t.trWG CJ1NQEa fASFS AND CDT7T1tO1S (IOTli IE1tP.S OOMa1NED) ACCOEDrNG TO PATY'aNAL AND MATB]tNAL sMOKING mSTORY' Fnheraanoker~, Mother aowter Ya No Yes No l unt c.ncTr S79 590. 162 1054 Centrol 615 652' 126 I 1214 Odds r.tio-auec 1-04 1• s-f Od& rrn,iordiused fon attne aawkina (b6rst utettre"ton-.ee tm )~ 0-83 •p<o•o5. tp<o-ou THE LANCET,SEPTEMBER10,19B3 was found in this model for female subjeas or for subjccts whose fathers smoked. The risk was significantly raised only in male smoken whose mothers smoked., Di.cuasioa Spousrrnwking Effect Our data strengthenrthe contention rthat havy smoking by one member of the spouse pair increases the lung cancer risk of the non-smoking partner. Heavy smoking by wives may increase the risk ofthe light smoking husband buithis finding requires further analysis and confirmation in larger series. Smoking by husbands did not affect the risk of lung cancer in women who smoked'~ (relative risk 1•03), a finding that suggests that active smoking is so powerful that it overshadows any possible additional effect from concomitant passive exposure. The proportion of lung carcinomas that were adenocarcinornas in non-smoking women was 54a1o„ eompared with 22°l'o for smoking women. The association of adenocarcinoma with smoking is weaker than for other histological types. The risk of squamous and small cell carcinomas among smokers, relative to a unit risk for non- smokersy has been reported to be 15 • 4, compared with 5• I for adenocarcinoma:` Table t may therefore reflect dilution of the relntion by' inclusion of adenocardnomas. Exclusion of adenocsrcinomas produced a significant linear trend in risk,, as found by Trichopolous er al.2 The possibility that differences in the eherniral I composition, of mainstream, (aci.e)and'sidestream (pusive)stsloke may produce differentt p7aponions of histological types of ttunours should be considered. The nitrosamine content in sidestrearn smoke is reportedi to be approximately 50 times greater than that in mainstream smoke.s The effect of the smoking habits of the spouse on lung cancer risk was first reported by Hirayama in a Japanese cohort study.' A cohort study in the United States reported a positive but not significant increase itl risk for non-smoking women married to smoking husbands.3 A ease-eontrolisrudy of non-smoking women d'ugnosed as having lung cancer in Gmce reported relative risks of approzimately 2• 5 for those married to moderate smokers and 3 for those married to heavy smokers, with a dgnificant' linear trend.2 Our numbers are small but we think,that the similarity between our findings and' those of Trichopolous et aN strengthens the suspicion that passive smoking may contribute to lung cancer risk. Pamttal SraokiV Effect As far as we know, ours is the first oasc-cantrol study of lung cancer reporting on parental smoking history. Parents' smoking behaviour influences the smoking habits of their, offspring,a' but we found that the smoking behaviour of the fatber does not influence the lung cancer risk ofhis offspring, whereas the behaviour of the mother does. This difference may reflect the closer and more prolonged contact that infantss and young children have.rith their mothers than with Eheir fathers. ~ The risk of bronchitis and pneumonia is increased in i children whose mothers smoke." This.efTect is dose related; C) and is greater in the winter, strongly suggesting that passive N' smoking by the infantis causaD} related to risk ofrespirator~• (,a infection. The excess of bronchitis occurs after 6 months of CA age, suggesting that the passive immunity trsnsferred from Ii" mother to child prevents bacterial colonisation of bronchial ~ mucosa. The effect of passive smoking on bronchitis maN•, be ~, independent of the mutagenic effect of the tobacco smoke," and it is probably safe to assume that the child is exposed toW ,;A I

THELANCET,SEPTEJNBER 10, 1983 both the irritant, and the mutagenic insults carried by sidestream smoke. The observatiom that the bronchitis anributable to passive smoking occurs mostly during the first yesr of life and is independent of birth weight may reAen ~the intimacy'of mother-child contact in that period of the child's life." l Bronchitis in infants may have a long-lasting effect on the respirmtory, tract as suggested by the increase in the prevalence of cough at age 20 in subjeas who have had a respiratory'illness during the first 2 years of lifc, independent of current smoking habits. !2 Whether bronchitis is a causative factor in lung cancer is confounded by the Sct that smoking induces both cancer and bronchitis. Cohortstudta havt conduded'tbat "persons who smoke cigarettes run a higher risk of chronic bronchitis than Don-smokers and those who develop bronchitis run a higher risk of developing lung ancer".13 How maternal smoking causes lung cancer can at this arage only be a matter of speculation; By itselfy passivc smoking during childhood may not; be sufficient stimulus for carcinogenesis. However, laboratory work has shown that: transplacental exposure to carcinogens increases the carcinogenic response to post-natal exposure to the same or to a different arcinogen;`ls' benzo(a)pyrene, a mutagenic arcinogenm found in tobacco stnoke, when injected into pregnant mice, induces cancer ofthe lung and other organs of the offspring;16 tttmours develop in 33% of the offspring of pregnant hamsrerLtreated with high doses of cigarene smoke condensate;'' and small doses of arrinagens can induce tumoun in fetaltissue.1° The effects of maternal : smoking are only signifiant m males, especially the heavy smokers. Perhaps maternal smoking enhances active smoking by the offspring in subtle Ways not detecte& by conventional techniques. If our, methods for eontrolling, for active smoking are not sufficiently refined, the increase in risk associated with maternal smoking would'riot be an effect of passive smoking but one of enhancement of active smoking behavioural panerns.. This is a real possibility and we would like to encourage funher research on the subject. Conclusion The differences between the effects of passive exposure to spouse and'maternal smoking are puuling. Passive exposure to spouse smoking is mostly detected in non-smokers an& light smoking malts; taaternal passive smoking effects are seen mostly in smokers. Passive smoking from spouses is introduced in adult life and in smokers is eoncurent with~ their own active smoking. The magnitude of such an effect may be low when compared with active concomitant smoking and'it may non be detectable when both types of smoking are present. Alaternal smoking, on the other hand, exerts its influences early in life and in the absence of active smoking is probably insufficient to produce carcinogenic effects. Our findings indicate that~maternal smoking results in a slight inaease in lung cancer risk but do not indicate whether the effect is due to enhanced active smoking of the offspring or to enhanced susceptibility to lung cancer induction after the challenge of aetive smoking later in life. Our findings point to the need for more rtsearch on the subicn of passive smoking and'eancer. Since large numbers Ofcascs may be needed for adequate epidemiological analysis, muh,-institutional colliboration may be indicated. l/- ,.a<.suPtk,nedbc eontran N01{P-91023,I>CCP. Natrooal Cancea Ir-. r w,, Xa; rnmal lrusn ul a of Htahti (SoutA L.oumana) md bya aranl from of the Amer,c.n C.ncerSonety(N.orth louuuru) 597' Ca-respondrnctc sAould:br .ddremed to P. C., Ikpenmem of Patha4op. lvuuwna State Univcru.ty Medxal Center, Ne. Otle.n., 1Souuuna LA 70I;11, USA aEFYRENCES I. Husr.ou T Nm.anE+nt ~<e d bo~r ewtc+ N.< a.lr~n rrE d 1~ rmcit . rudr,nl.p•n s.M.I719lI:at:r. la3-~s.andau; la6o. 2- Tndcpolau n, K.1.n" A, fpFrro. L- M.rM.6ao. t. Laa{ o.ca d p.oi.t amoWo/ l. )Ci.rn 1961,17:,I-4 1. C.rfinY<I IL T- tnndt u+lunl onrr-mn.ll,r reen/ ncn-smikn and . nwr on pa,.r md.,ry 7Na, f...n..l.,.. 190111.41111- 106lie 4. Dan HF, H.rnarrf Q'; Herrold KspeaaX npnn to tbr.Stuse" Gennal't Ad.uor3 G.mmme. - smoJ-{ and Health, lo fmaLo{ andHeItA , l'SDHEV PYbl,c HeatfhSirrwre PuE1¢nan Nb1101 iahuyton.DC. 1961, 175: S: stock SL Rrk's tb<pmi.<.nokn nm. lr.m 1980, . . 106: f. Naenl lnau- dEdeorm. Teena{r.=.&+eg Iradrn.and Mea.ennPttmuDHEQ. Natmal lls+rn.< efEAUCam, ~"~WMonDC. 1979, 259 7.lane< RR.,aenk+ FR, Kehrr MD Sinotan{leW.nu, m atacs~.~e papua.ewn . wtcepttrl #qwoerb A.1 Pr1 H41 1972, Q: 007-13 a. l/eS4 S, t>rs AM ldm .lm~m te laa.pr W asd .raeal a.oVq L.r., HrK a: s2P-u. 9 aaaaYeOrP. Rtl.rAp efaneraalsmeiaa{te.erbdnread.onalnydth<diJd'. " . to aR d f- . Au. P./ Su.! 1071- n: 521 -3'1: 10. (lndCo tA; Chi~ l. Supmur. T. C.prnx riuR-deta a orrr a aennta r.iI Mr,aw.Re 1974'., Y: ..71-a1. 1I: Celk-/. JRT, HolWMQ-Q',.Cartbi111RT..1efLeea dpt~r.< ~.ttq .w0 ~n~d pA/.Y,n,enpnrueoou.ndtiroecEitutearl7tiiWheaLlu.a+1974;u 1031-34 12.Co1ky.JRT.DouaLJt'8.14dDD Re/nraerydsemroan(adWieIntlurnnofy mly cbrteAood b.e, rvp~or3 tran ,iltean; .aoal olr,6 au plliawn .nd aal,ct 8. Md j 1973, iu11S-Hr.)..Rsuapae J- AneE.q. chroac Yr.neWeaL ~Olvq mcn.B.At.!)t973. n, n3-7s14 . vslu,o.uchSD Ganpar.nc.<d-en oerww.l o*nsateee.r 1. . Temuar L,. Mohr L':.b Tnsplecvul tans{enews IasrmtuW Actary,far Reeud or.. Cane, anemJr pub/aatwem ta4 Lron„ 1971 . 1 i-22 IS..NaplLo.NP Som<Enxra casiderann an LLt proElem . d trvnpt.rmia:: mruq~rnnuln Tmir L MoEi L', edi Tt.np/aQntsll mnnqr.non Isvnamal ABnryfa. Rmrcb r+ Camtcre aooufic pttbt,oran ao a Lrcn. iri1; 1-1T te. Ndorov. Tl'Traasptaevnul anm erleam(a)pr.aae.d. prrmt. a.drf:pe,o! M.d 1977, M: 1025-27 37. Nrdcv 1G. C~erw~at7, IN . Tamen and brperpYOe Istau tn $-n Aanarcn follo-(rrv»placm,.l andeeopueal.rreataxor.ubnaamtetmeR<mtdenu,< J C..ar Ra Ch. Owml 1979. M: N9--Sl.. 19 E.n+on RB Indi.Wuals ,rwpl.cent.lly.ap,e.ed:temeraal anot+og rer~ be.a, wacreaed aQr•er nY m adWi4&k L.rur1q0: u 123-27. ANAYHYLACTOID REACTIONS TO NEUROMUSCULAR BLOCKING AGENTS: A COMMOTILY UNDIAGNOSED CONDITION? P. R. YoLrrlcwtAN K. M. TAn.OR ]. D. WusoN Depornwcntr oJMailiciar and PA'ormacology, Ureerrriry oJilrtakltrnd'- .Sdioo4 oJM:ediaoe, AroklanQ New Zamland Satnsmary A group of 28 patients with estreme, life- threatening sensitivity to tatnmetbonium was identified and 15 were atudieddn detail by skin-testing. The femaleltnale ratio was 8/11. Sensitivity may be praent without previous exposure to suxamethonium; in 3 patients reactions occurred during the frrst exposure to anaathesia. Most patients showed one or more voss-sensitivities to~ alcuronium, tubocurarine, and gallamine. Signs of circvlatory collapse were the sole presenting feature in 50°70 of the patients. Histamine release induced by'the drug in vitro was demonstrated in some insantzs. Imti-oductioa DRUGS of the muscle-relasam group are commonly implicated in systemic reactions, sometimesaife-threatening, which occur during general~anaesthaia: Since 1977 patients at our hospital who have had severe anaatbeuc reactions have been skin-test ed to determine drug sensitivity. After tw o deaths attributed to suzamethonium~in Aucklan&in 1982, a study'ofknown sensitive patients was undenaken, initially to determine whether 'Ethycholine', the only'suzamethonium chloride available in New Zealand, differed in provoking