Tobacco Documents Online

J. Exp. CIin. Cancer Res_ 3J;3~ 1994 r- Is passive smoking an _ added risk factor for tungg cancer in Chinese women? LC. Koo ` Ph.D.. J.H-C. Ho = M.D., D.Sc.. F.R.C.P.. F.R.C.R... D. Saw' F.R.C.P.A., M.R.R.C. Path. 1 Depmsment of Coen+wunirK Msdicins. Univers+iy of Hong JCons. HoRs Kon[ • At. & H.D. lturirute of Radiology and Oncoloty. Qusat, Elizabeth Hospiial. Hong Kong s At. & H.D: lnstfruu of Pmlto/osy. Queen FlFzabeth H'ospiual. Hons Kong 2Q0 female lung cancer p.tients and 200 healthy district controls were leter viewed' to identify and quantify the various wurces of passive smoking among Chinese females in Hong Kong. For the eveFSmokers. passive exposure from external sources did not appear to add to their risk. For the neveranokers.. qualitative assessmente (smoke exposure categories. age when passive exposure etarted), and' quantitative assessmenn (hours. years. iatetssiry) showed no ai- ptificant differences between the date for patients and controls. Moteover. higher relative eitks were oot associated with higher levels of passive smoking foe the ever or never-smokers. T1+us. our findin=s would' seem to indicate c6at passive tmokini, as an isolated factor. did not have an influence on female lung cancer iaddence in Hong Kong. Recently, there has been renewed di- scwssion on the possible effects of passive smoking on lung cancer risk (5, 8, 19). In previous studies, on the possibility of increased risk of lung cancer among wi- ves/husbancis from their smoking spou- ses, the data (2. 9, 12, 16) were only based on whether the spouse smoked (yes/no) with no further qualifications on whether the smoker actually smoked in the pre- sence of the subject and for how long. Where a quantification a was done (5. 17, 18), it was based on the current spou- se's smoking habits. It is well known that the carcinogenetic process of internal soo- lid' c,ancers usually begins 20 or more years- beiore diagnosis when there might have been no exposure from the current sour- ce. Furthe.-more. little account was taken of changes in sisolCing habits or tnarria- ge, or the possroaility of exposure from the work environment. Some of these pro- blems were raised by Hammond and Se- likoff (11) but they have yet to be ad- dressed by epidemiological studies to date. Chinese females in Hong Kong have an average annual age-standardized inciden- ce rate of 24.1/100.000 for lung cancer (13). This is among the highest rates for women in the world. In order to more directly assess the possible role of passive smoking in lung cancer development, a retrospective study of 200 female lung cancer patients and 200 healthy district controls was begun in 1981. Hong Kong., with an average urban density of 23,0M inhabitants per square kilometer. an& 8 tas of average living space per person. is one of the most densely populated areas in the world. It is. therefore, an appro- priate place to test the passive smoking aetioiogical hypothesis. Patients and methods ReceiMed fianuaty 14. 1934. • To whom teQuesta for reprints should' be fent.. The 200 lung cancer parienu studied were from the wards or out-paeent depariments of l hospitals 27 7

Koo LC. a ai. in Hoag Kong. Patients were interviewed as they became available. Eisfit ptusible subiecu were not baterviewed because they were not ntfficiently alert p0 answer our quest;ona. Another 13 had to be sscluded after interviews had been completedi when later evidence and checking revealed that their lung wmoeus were secondaries and not primaries. Med(- tai ttcorda and radioQayha .vete reviewed by T'.1H: C.H» and pathology specimens were verifie& by DS. with her colleagues. 9Vhere nexsaary. additional dia- psostic procedures were requested oo compleu the data. Patients were matched with an equal number of healthy controls by age strstification (s S rears) in sach district (n a 34). and by aociotconomic status. Controls were iaterviewed at their homes wiahin a few weeks after their matched padenta bad been idrntified. Two fsatait research aa:iatants, fluent in Chinese and English. eonducted the interviesva usinj a tape r'corder and asmi-ceructured' questionnaire. Utilizing interview tahniques from the social sciences. ape- eiaily those tslated no the gathering of life huto ria, the interviewers were trained to probe for details and elaboration of facts. Data were obtai- ned on the changes in tsridence patterns since birth (where Gved. how long. bow many totether. what type of houainf. how many rooms); occvpationali history (where .rorked. what done, krel, of pollu- don. how long): active smoking (type of tobacco. method of smoking. and amount wrrently smoked and at 10. 20. 30. <0. 50 or atore years ago. and inhalation practices): passive amoituu (from whom. what type of tobacco produot. amount they proba- bly smoked per day. amotmc of time of etpoaure. when stopped or chanstod): personal and family hi- stories (ace at ntarriate, divorse. separation. and/or widowhood: number of children. occupations of pa- rents and' s pouse)t etc. The taped interviews were aanscribed' and then ehecked' by L.C•KL for points that had bem left out, or for inconsistencies. s.j. comparin2 passive smoking ezpo.ura with reaiden- tial and ntarriaae historics. Where necessary. sub- jects were recontacted' for further infotmation. The mean age of the panmts was 61.8 years (S.D. 10.0) and that for the controit was 60.6 years (S.D. 9.6). itesuIts Hrsrological' distribution The histological distribution according to WHO 2nd Edition (20) and basis of diagnosis of the patiencs are shown in Tabie I. The predominant cell type was adeno- carcinoma, forming 34.5% of the total sam- ple, or 38% of those with histological ty- ping. However, when the frequencies of squamous plus small cell types are com- pared with adenocarcinoma plus large cells, the resulting, Kryberg ratio (6) of 1.16 still showed a preponderance of the former group of tumours. This Gow rela- tive frequency of adenocarcinomas in Hong Kong Chinese females was also found by Chan and Maci.ennan (3). Smoke esposure caregories From our interviews, three major regu- lar sources of tobacco smoke were identi- Table (• Cd/ ryas and basv of diaywsis t:1S e.*. fR- f..1L «11 Amo- dar'ems Lrp ull ttisad 4rctaatd Yec)ac- cltGd Tocal ... (s) ab.. (t) r.. (S) a.. (z) .e.. (z). ... (z) Ib. (s)~ ... CZ) i:neutso.eeac aieya7' 2~ (17.0) 1~ ( 7.0~ 1(~.!) 4 ( 2.0) 3 ( 1.3) - • t •.s) 63 (32.3) y.occkae ls ( 7.0) 7( 1.3) 13 (li.!) 1( 1.3) 1( 1.3) 1( O.SY I( 0.3) 62 (31.0) t."We ..aa 1(~.3) 4 ( 2.01 t0 ( 5.0) 2 C 1.0) - - • 23 (12.,3) rl.xral - 1( 0.3) 3( 2.s) 2 ( 1.0) - - 1 ( 0.3) . ( 4.3) _-cm cycolbay - ( 3.3) .7 I1 (!.3) 12 ( i.9) 1( 7.3) - 1( 0:S) 32 (61..]) aaliolsaical i 7 ( 3.3) • 0 3. 3) clinacal l.cal S6 (2a.0) 37 (1a.3) ao (34.3) ll ( 3.3) 7( 3.3) L( 0.3) 1! ( 1.S) 200 (100 ) ' leclLdn tr:uc:brunchfal bivpar. 278'

Passive sasokinQ risk in Chinese women' fied: In addition to ever-smokers (S). there ve smoking at home (H), workplace (W), were those who had cohabiting relatives or both (HW) had RRs only marginally smoking in their presence at home (H). or those daily exposed at their workplace Table tr - Relarfva `is/a (RR) for drller+ru esaosilre for a number of years (W). In Fig. I three intersecting circles have been drawn to shown seven possible categories and one isoIated circle (N) representing those who had never been exposed to any of these regular sources. Passive exposure is de- noted by the shaded area, and includes sidestream smoke from home or work- p~- To see whether this qualitative method of assessment would discriminate higher risk groups, all patients and controls we- re fitted into each of these 8 different smoke exposure categories and the odds ratios were calculated (Table TI). If those claiming none (N) represent the standard with a relative risk (RR) of 1.00. smokers with no other source of exposure (S) or `"°~ ""`°`' fasi.aca CaaceL Q s ~ it s,,' ~ s= ~: o:;~ SR s= " ='s` sv : 7 1 ~ 3 .f/ i.3r ~ zs ~ ~ s. " ' 22 `0 z'0° tocai 200 zso JbM°"ar' °at's°ry ra°se° °iaa°L ': i: ~`. a~ ~~ li :T i:=i~ a u ao i. oo multiple sources (SH, SW. SHW) had RRs UML 200 200 ranging from 2S6 to 5.45. whereas non- smokers who were only exposed to passi• • D s aaW, o s 0.414 Passive exposure at home i No exposure clasaed f;=. t- Sew,Ee uposun cvraSwiat. 279

Kool..C.aaL greater than 1.00 (range 0,91-159): When smokers and those only passively exposed were grouped (fable IIB), the RR of acti- ve smokers was 3.23. and ttlat for the pas- sive smokers a non-significant 1.24. Quantification of passive smoking Our detailed interviews allowed us to estimate the amounts of passive smoking from various places in terms of hours or years (Table III). Smokers as a group had more exposure to passive smoking from others than the never-smokers. De- pending on the unit of ineasurernent, whe- ther hours or years, we found that among the smokers, the patients had more hours of exposure, but the controls had more years. Among the never-smokers, the con- trols actually had more hours or years than the patients, but these differences were mini.nsalL Hours per year was used as a measure of intensity of passive expo- sure. Overall, there was no significant difference in exposure levels between pa- tients and controis, whether they were smokers or never-smokers. Tabie III - Avera=e aecumulcriw pauivr tobacco anolce esposlve by place. rs s..k.r• t.r ,i...:-MIor.: 7! 32 M +7. }aci.au ra"1. }act..u .insasL I.: Re.. 1surs 22.773 21.317 1l.711 18.32~ t.ar. 2s.,7 31.3 23.7 23.~ IZ. Yssiqlac. lraa {.352 1.067 2.321 1.«t Y.ars , 3.1 1.7 2.0 1.2 III. Teu: ao%ac. Ibw. 26.703 23.383 17.592 , 20.737 ar. T. 30.0 32.6 U.4 ' 26.) lswrs,f.ar tS6.. 717.3 $77.3 7i2.• * ExchUded We/Y 33 pmirraamul1J NntTolf WIth- ow paaave eraouve. Since about 90% of the total amounts of passive smoking came from the home. Table IV shows the average contripution from each~ cohabiting relative who smoked in the presence of the subject. Only direct exposure was counted. Husbands who smoked, but did not expose their wives to passive smoking for various reasons. such as living overseas, on travelling jobs. etc., were not inciuded in the estimations. From the Chinese cultural practice of ha- ving extended family members living to- gether, the female could be exposed to her parents' cigarettes or pipe when young,, to her husband's and in-laws' tobacco du- ring marital life, and to her children's ci- garettes when old. Although in terms of hours/person, parents were found to be a heavy source of sidestream smoke, only a minority of patients or controls were so exposed. The most frequent source was that from the husband. About 2/3 of the total hours of tobac- co exposure were calculated from our data to be from the husband's cigarettes. Both cases and controls had an average of about 20.0M hours of passive smoking from their homes, so that no significant difference in exposure levels was found between them. Table IV -$ouree of pauiw esposure w home. A..raye cucal amsuacs ir S O.SK1 141 faeloncs 1" swcrela 1b. • fnas/*.rses sa. {nas/vsrs.e 113 11.123 111 19.314 U 27,f1t 21 .17.7ti 7 28.137 ~ 1.175 246 3.6046 33 0.170 1 13.333 13 4.33e 21.026 20.672 280

Unlike the Louisiana study (5). we found no association of an increase in disk of lung cancer among current smo- kers, ex-smokers, or never-smokers and material or paternal (yes/no) smoking habits. Smoking history and histology Among the ever-smokers, there was a predominaace of squamous and small cell types of lung tumours, whereas the op- posite pattern of a predominance of ade- nocarcinomas was found for those passi- vely exposed and the N category (Table V). There was no significant difference in cell type distribution between the pas- sively exposed women and those with no regular exposure. The predominance of adenocarcinomas in the never-smoked women as a group, regardless of their passive smoking history, has been repor- ted elsewhere (1, 4. 10). ~ Table V- Smoking kisrory and' hiscolo=y. Cr11 " 1.attag Yi..wry s.w•wKe * 7.e11 frll A/.e.caresae.a • tosp Ca11 t..r smsk.d µZ (61HS) 36% (34l!!) taeeiw e.ating 42% (2S/S!) !ti (7N'!1) im. 37S ( Z[1f) f3i (Y211f) Risk among never-smokers We have earlier shown that the average total amount of hours or years of passive smoking among the. never-smokers was not significantly different between pa- tients and controls. We also did not find a higher RR among patients with passive exposure levels of > 35,000 hours (3 hours 12 min./day x 30 years) than those with Iower exposures (Table VI). Puiive smoking risk in Chinese women? Table VI - RR o1' luni cancer an.oRg aeve.-smolurs by lene(s of passive ezyosure.. 4t.gery ratieta c.aza. a ...lr. Mr 22 " 1.00 iwI 57 al 1.28 co.aa 210 1 9 16 2.02 90.96 lf.eal paeiw ~i 97 2.2a 40.69 ' S J3A00 Murs ' > JsA00 6ows It is possible that the bronchial mu- cosa is more susceptible to carcinogens before adulthood than later in life. Table ,VII summarized our data on age when passive exposure started for the never- smokers. Ther e was no significant diffe- rence between patients and controls in their ages at first exposure. In fact, there were more controls who had been exposed before the age of 20 years than their mat- ched patients. Thus our data were unable to substantiate the possibility raised by Doll and Peto (7) that e life-long exposure (including clliidhood) may have four ti- mes the effect of exposure which is 1i- mited to adult life •. Table VU - Age passive esposwr srmred lor nevrr~ .ewlcers. latl.nt. C.etzal. 11p 1a. (Z) 0.. (ZY. 0 - tf iS (23) 30 (31) 30 - 39 42 (K) So (52) sp. ! (I4) 17 (11) L.eal N !7 A..rap a" 34.6 34.3 OSa.1u Rssk for ever-smokers It is well established that not -a11 smo- kers, not even heavy ones, will develop 281

Koo LC. a alL lung cancer. To see if passive smoking adds risk to active smokets, the risks for light smokers (< 100 kg tobacco or 14 pack years) with low or no passive expo- sure (< 15:000 hours or 1.5 hours/day x 30 years) and those smoking simiIar amounts but with heavy passive smoke exposures were compared (Table YIII). The same comparison was applied also to the heavy smokers (> 100 ltg or 14 pack years). We found not only no increase but an actual decrease In the risk for both light and heavy smokers with heavy pas- sive exposure compared to those with no or low exposure. There was only an increa se in the risk related to the levels of their own cigarette consumption. This result was also found by Correa et aL (5). Table Y111 - RR• /or srrwlCtrs with mvd without passivt srposurt. T"o l.eal l.eionu Ce.er.ls St 1. 1! 1.00 3 13 0.24 Y 14 3.61 43 17 2.f7 112 63 ' S!00'ks tnbocw > 100 ki* tobaoco _< 15A00 hours ' Z 13.OOQ hours Discussioes In this retrospective study on the pos- sible influence of passive smoking on the high incidence of lung cancer in Hong Kong Chinese females„ we have attempted to identify and- quantify various sources and types. of tobacco exposure among 200 patients and 200' district controls. We have limited our data presentation to show on- ly those factors relevant to the issue of passive smoking. A more detailed descrip- tion and', discussion of active smoking as a risk factor was presented elsewhere (15). The apparent lack of an association between passive smoking and the risk of lung cancer in our study may be due to possibilities which occur because passive smoking may be only a very weak carci- t:ogen, whose effect may be concealed by other factors that play a role in a multi- factorial and multistage aetiology. Among the female never-smokers, intervening factors might cause an overshadowing or a protective effect (e.g. bronchial irrita- tion, dietary nitrosamines or beta-carote- ne). These factors in Hong Kong are li- kely to be different from those in Japan (12), US.A. (9, 16), or Greece (17. 18), and this difference may explain our different results. The possibility that the a dose-re• sponse curve resembles a logistic in sha- pe a such that a there is a dose greater than zero which produces zero response • was considered by Hammond and Seli- koff (11) and may be operating here. Certainly the lack of an increased risk for the active smokers from passive smo- king, which was also found by Correa et al. (5). Would seem to support the possi- bility that the effects of active smoking or, indeed, other factors yet to be iden- tified, greatly overshadowed the carcino- genic action of passive smoking. This, however. does not imply that pas- sive smoking is innocuous, as it may con- tribute an added risk of other respiratory, and cardiovascular diseases (8, 14, 16). The possibility of other factors like diet, previous history of respiratory diseases. occupational: exposures, use of inhalants, etc.. overshadowing or inhibiting the ef- fects of passive smoking on the risk of lung cancer among never-smoked females in Hong Kong and also the roles of these factors in the carcinogenesis are being investigated, It is hoped that more direct assess- ment of passive smoking by other wor- kers in other areas can shed more light on the passive smoking controversy.. 282

Abmwl.dae-.a,rs W..rish to etprsts our Sraticude and thanks to the following iastitutions and' individuals for their help in• this ptoiecit- the Hong Kong Anti-Caaccr Soaety. aod the Research and Conference Grants Committee .nd the Medical Faculty Research Grant Fund of 1!n University of Hong Kong for financial appon: Professor CK Mok. Drs. W.K. iim. NCX. Lo. May Wei. W.C. Chan. L. Hou. S.H. 4+ow. K.W. Chan and K.T. Tham and S'uter M. Aquiiw for their advice atd' lielpc Mr. C. Cban. Ms. C. Tong and Ms. N. Ise for help In data eollecrion and awlr- tfv: and Ma. G. lsu. Mrs. T. Lam and Ma. A. Chow for aectstarial anistaoce. Refereaces I. Chan W.C.. Colbourne %t.I. Fung S.C.. Ho H. C.: Bronchial cancer in Hong Kong 19761'977: Br. J. Cancer. 39: 1i2-192, 1'979. 2. Chan W:C« Fung S.C.: Lung cancer in twn- onokert in Hons Kong. in: Grundmann E. (Ed.). Cancer Campaign. vol. 7. Cancer Epidemiology. Stuttgart/New York: Gustav Fischer VerlYe. 199- 201. 1982. 3. Chan W.C_ Macl.enaaa R.: Lung cancer in Hong Kong Chinese: mortality and hismlogical types. 19641972. Br. J, Cancer. 35: 226•231. 1976. 4. Cooper D.A.. Crane A.R-. Boucot K.R.: Pri- mary earsiaotaa of the luns in eonsntokers. Arch. Enviivtt. Health, 16: 398-4C0. 1968. !. Correa P.. Pickle LW:. Fflntfiam E.. Lin Y.. Haens:el W.c Passive smoking and lita8 cancer. l,aneet, ii: S9S-S97. 1983. 6. Doll R.. Hill A.B.. Kreyberg L: The atnifi- eance of call type in relation to the aetiolCgy of lung cancer. Br. J. Cancer. 11: 43-a. 1957. 7. Doll R.. Peto R_ The cassses of extuer: quan- dtative estimates of avoidable risks of cancer in the United States today. I. Nat. Caaces lnst.. 66: 1192-1308. 1981'. Passive sntoicinQ risk in Chinese wotaen? 8. F.ditorial: Passive mmoitins: Forest. Gasp md' facn. Lancet. i: Ss8-349. 1982. 9. Garfinkel' L_ • Time trends in Iune nnces 6or tality among noo-naokets and a note an p..- aive smoking. J. Nat~: Canccr Inaa 66:, 106i- 1C66. 1981. 10. Greea LP, Beophr P.: Cardnoma of the lung in noo-smokin8 Chinese women. West. (. Med.. 136: 291-294. 1982. 11. Hammond E.C. Selikoff I.I.: Passive smoking and Gatf cancer with eontmenn on two new papea. Fstviron. Researeh. 24: a44-tS2. 1981. 12. Hinrama T: Noo-.mokinj wives of heavy atttos kets have a higher riak of lung cancer: a ttudy from Japan. Dr. Med. 1.. 212: 183-1aS. 1981. 13. Ho J.H.C.. Chan C.L. Lau W.H;. Au GKH» Koo LC- Caace: in Hong Kong: wnu epide- miolotieal observations. Nati. Cancer Inst. Mo- nM.. 62: 47-SS. 1982. 1:4. Hugod C.. Hawking L.H.. Ascrup P.: Expo.ure of passive mrokers to tobaer.o smoke consti- toettta. lat. Arch. C+ceup. Environ. Health, 42: 21-29. 1973. 13. Koo LC:. Ho J.H.C.. Saw D.: Active and pas- sive smoking among female lung cancer pa. tienta and eontrol, in Hong Kong. J. Exper. dl Clin. Caocer Researeh. 4: 367-37 S. 1983. 16. Miller G.H.: The Pennsylvania sn+dr on pas- sive smokin8. J. of Breathing. 41: S-9. 1978. 17. Tric3topoalaa D.. 1Calandidi A.. Spasro. L. Maa Mahon B.: Lung cancer and passive smoking. Int. I. Cancer. 27: 1-4. 1981. Ii. Tsic'aopoulae D, Kalsndidi A.. Sparroi L: Lntte cancer and passive smoking: conclusion of Greek atttdr. Laneet. ii: 677-678. 1983. 19. US. Department of Health and Human Servi. as: The health consequences of smoking: Can- ea: a tsport of the Surgeon General. Washing. ton. D.C.: U.S. Public Health Service. 1982. 20. World Health OtTaniution: The World Health Organization histolotical typing of lung rumours. 2nd Edition. Am. J. Clia. Patti.. 77: 123-136. 19M 283 I