Tobacco Documents Online

i! Gw.w f 11!PDl' L'Z 9ti2._ 9aT C MaamilLn P+,ea L W~~. :9p0 Lung cancer among women in north-east China A.H. wu-Williarns', X.D. Dai', W. BCot': Z.Y. Xu', X.W. Sun', H.P. Xiao', B.1. Stone, S.F. Yu'. Y.P. Feng', A.G. Ershow2, J. Sun', J'.F. Frau.meni Jr' & B.E. Henderson' Drpaamrnt o!' P.evrr•ttve Nedroutt. L hlvtrsuv Of Southern Caltfornta School Of Mtdfcosr. Lau Angt1eJ. CA 9003'3. USA: %ailonal Cancer Inrttrurt.,Berhtsda, .'4ID 20892. USA. 'Hcrbtn Wtdiral School. Harbur• HeJonigliarrj Province. Ptople s Republic of Chuea: anC' L.wanut; Public Hta/rh and Anu•Epidrin+a Srarion: Sht+tya+j. Lmor+uI Pravtrset. Prople s Republic of Chwa. Srann.ry A caae-control'uudy of lung cancer tnvoivtng intervwws wrth %S fernak pauenu and 959 controls in SAenyang and Harttn, two industnal cities which have among the htgtiest rstes of lung artorr in Chtna. raveakd that cipretu smoking is the main tausal factor and aaoountrd'for about JS'/% of the rurnoun among women. Although the amount smoked was low (the casa avenged eght npratea per day): the percentage of smoken among women over, age SO in theae cnua was noriy double the nauonal avenge. Air pollution from coal burning stova was implicated. as risks of lung nnac tnas.rd in proyortson to years of eipwura to 'KanB' and other heaua8 de»aa mdtgenoui to the repon. In additson. the nurrber of moL cooked by deep frying and the freqpeecy of smokiness during aooktn8 .ert aanaatrd wttb risk of lun8 oem. Mote tasn than controb reported workplacs e:pwuees to coal dux and to smoke from burnin8 fuel. Elevated nsks were obsetwed for anxlttr worken and deirmaa.d risks for te.ulc workers. Prior chromc bronc]tittu empAysaos. pneumonta. and rscent tuEem+/6.u contributed stps6unt/y to lung onar tuk as did a butory, of tuber culous and lung canoer in famtly members. Higher intake of earoteae-r>cb vqetabrs w not protective apttut lung arner in this populatton. The find,n0 were qualttatuvely, stmtlar ucroa the ma)or o<ll types of lung ueker~ naept that tht associations with smoking and prevtous huy dve.ae were stronger for squamouu oat aii nncen than for adenonrnnoma of the lun& The rate of' lung atscer among Chinese femaks is among the highest in the world. Elevated incidence. particularly of adcnocaratnotna of the lun8, has been noted for Chinese females residing in difierent 8eographx atsas, including Singapore (Law tf al.. 1976). Hong Kong (Kung tr al.. 1%4). Shanghai (Gso tt al.. 1988) and the Uhtted States (Hinds rr al'. 19814. The high rates an unusual because few Chitseae women smoke tob.eeo. Within Chitsa. elevated rates of female IunB cancer ars found in urban areas such tas Shang- hai and to rural as well as urban areas of the nortbea.stetn proviaces of Liaoruni and Heilonpian= (National Cancxr Control Office. 1980; Xu ef al:. 1986). Reasons for the geo- graphic vanauon and clustering of high rates of lun8 cancer in the northern provitxes are not known. We report here the results from cax-concrol studsa conducted in Sbenyaa= and Harbin: the two major industrial cities in Liaonins and Heilonlian; provtnces. to evahute the role of ssveraJ potet} ual risk factots. In 1985-87. investiptors from the Liaoning Provines Fublic Health and Anu-Epidemic Station and the US Natr!onal Caocer Inuttutc cooduaed a 1[rp tun8 caneer study inctud- in8 017 mila (729 taxta. 18i eontaob) and 1.073 females (518 cawes. 555 oentroll). Dusin= the satne tinse period. iar.r upton at Harbim Medical Cotiep and the Utsisrtity of Southern California cendut.-tad a txae-oontrol study focu.ad on femak lung t,aacar (446 taaea, 404 controb): Investiptors from both studia met during the piaanin8 phase of the study and adopted a tmibnd protocol to aaoertain aad nlret tsaea and controls, and a oommon qtaesoonnairs for tde iatavinr component of the study: Data on risks from taook* and ait pollutwn among mcn and women in Shenyang bavn bnn publisbed elsewhere (Xu er al.. 1989). Hereia we tepott rinks amont females attaodstsd with a variety of faciora, inaa.~aj sample sias by oeuty 80'Y6 by combining itsformacon from the two otin. Co.>eVOOdiacs A. H'. W> W iid.a. kww"d 27' M.rcia l99Q and s nvr.d form 25 1dy ll9tl. Cast arcerawrwerf We sought to enrol all newty diapnomd pnmary lung nnarrs in females in the study areas between 1985 and 1981. Uulia- in8 the cancer, repsu>ta of Harbin and Shenyang. a system of rapid can ascenainmeat was estabiisbed with the ooopera- tion of all the major hospitals serving its uea (about 35 in each aty). lo brnef, the admittin8 phyncans at each paru- cipaun8 bosptalIcomplbted a can abstract form whetsever a lung canarr was diagnosed. We tsoeived tDex abstracts on a bi-weekly basis and selected u elipbk caaem those wttti primary, incident luns cxncsrs diapc»ed among femak rsst- dents of the study ara who were a8e+d len than 70 years at the nme of diai nooa. The lung caaar diagno+is and all-typs clasaBnnon were versbed localty in each study atea by a paml' of pulmoauy spaaatiw aad parbobpsta. CaRaot adecriw Contmia .as femal" raodomly seieet.d from the pissral poyutatium of Harbin and Shenyang. Controb wsrt frs- queatlY, matcll.d by 5-year ye Roup to the e:pacud diatri- bution of caaea, wEicb was detamimd in advance uan8 the number and age ditanbution of femak htni canaer cata rrported' in the two dos ia 1983. A tbreo-stap nmpfixtlg procedure was ua.d to neiea e.eb oootroL The initial umt for randomiaaooe was tJ,e txilghbourbood oommiaas, of which tbere an about 1,l00 atrb in Harfiin add in Sasayna& Committeal wen randomly a.ian.d witD replacemtmt afbr wdghtin8 by, tb.v poQuht;oe tsa Then we raadomly chose a bou.eboid Qovp from tbe apptvsim.caty 10-25 So+r.boid Foupe witaio .aeb sniacs.d tseigEbtxtrbood oommitsw. In tsa final itallle., attson8 aD fcmaar ia ts. S-y.ar a,P osttpey. within tha boareaoid Vrsap, om was raadomly tstdas.d N ~ ~ W QW-*Xmwiplt ~ A ururxsaad. prreadd' qtaaioemin .as tand by ttaiaad •~ intes++ieovs who ooodtssad panonal mter.ies .ith the pv- .r~ txipaau ia t!>eir bc,mea or work as or ia tbe boapitall ~chaic. 7b izu~ ptbe'td iaformatioe oo demopapbi,c faaocs. ~+ atxive and patsit r. molls azposar., lifetat rmdeaaaa and nill'if oaupamoal biestirs, dia aa,d oookia8 ptacaa., parsoaal drtory of noatm.kVaac Mn8 thiraa., bsstary of tuissrcvlos

LL'NG CANCER AMONG WOMEN fN CIiINA vO (TB), and cancer in first deQee rciaures, and repro,du,rn.e facton. Qirsuons on s>nokinl tndudsd the amount and types of tob•ezo pt'oducta srtsoked. age rhen smoktnl stan_ eQ and for e=-Qnokers. age wbea smokial atopp.d. To assess passave srooks exposurt. we mksd about lifeumr msi_ dentul exposure to tobacco smoke from t:ohabuants. usc(ud- tn1 the amount and durauoo of exposure from each smokinl cohabtant. in addition, we asked' d the subjact wu exposed to passtve smoking at each work place. For each rosrdtnor in which a sub)oct lived for three or morrs years, we asked tn detail about heaunl and cooking pracvcxs, including methods for heating and cooking and typea of fuels used. Several questtons were asked about 'Kanf . bnck tieds eom- monly used in the north<astent pan of Clutta, which an lieated'etther dirsctly by a stove underneath them or by pipes connected to the cooking uo.u. To assesa dietary habits 5 years prior to intervuw, we asked subyects to estimate their frequencies of intake of 33 food items, inclUdmi stapk grains (nce. what, maue). soya bean produeu (Eean curd. ferment- ed bean paste). dried pess and beans. antmal protan sources leus. fish. shellfuh. liver. poultry, pork). ferttsensedtsa)ted foods., akoholic bevenla, and~ fresh vegetables and frwts. Also tncluded were questions on diaptosn by a phyaoan of previous each lung drseaaes. age at lung diaeise diaposts. and if hospitalisation was reqtured. Infornuuon on outcome of each pnepsancy. age at menarohe and at meoop.use was also ehated. As a qualityeontr!ol measurs. interviers were cassette•recorded for review by a field supervttwr. Sraturrcal'nsrrkodr The data were edited. coded. keypunehed and submitted to computsnssd ranp and conastency checks. The staustxal analyses Wess baaed' on muluvaroate techniques for ase- control' data I Breslbw ! Day. 19b0). Uncotrdittonal lopttsc regression analysa wers used to esumate summary relauve risks (RRs) of lung cancer aasoaatod with vanoua factors while adjusting for other facsore. RRa were ealtulkted for all lung cancer comboed and for spanfic csU tyya. We pteaent results for squamous cslliand oat.'small cell cancers combined because we had too few oatvsmaU adl ancers to conduct separate analysis and beausr these two oeit types of lung canarr ant mors strongly associated wtth smoking than adenoamnoma of the lung (Lubin & Blot: 19P). Our analysu for adenoarsanottu of the lung did not ieclude Wys cell canan. There were too few larp oatl assoess for in- clusion by cell type. In the analyss ittdudinl all atE*asv, the regression models contained tertae fot age (hsa than 50. S0-S9. 60-69 years), eduntioe (no formal ydtxationm pn- mary or secondztry scbooL high sc~ool and hiiher), t®oinnB ( non-smoksr. wmok,ed 1-19 dsuetss pQ day nad 1-29 y(an. 1-19 opnretea per day and 30-39 yara, 1-19 aprrttea per 8ay and 10 + y.ars. 20 + tapnar p.r day and f-29 yaars. 20 + cilgarertsa p: day and 30-39 y..n, 20 + ciwerus per day and 40 + y.an) and snrdy omtrs (Harbin verstas Sb,ea7ant). We aYo ooeductrd aoalyt.a restrtcted to nootmokam deleting the tmolknnl variabhr in the regesaion tssadd tmd adjusnn= only on tya, adutatio.. and centre., Rwi (h - 310) adenocaranonsaa, 21% (R - 201), squamom cctl cuminomas. 16% (r, - 117) ,o.usmall cell caranomas and the rematnder .trs IarBc odf caranotaaa, mutures of ot6a cell types or the cell type svaa tsot known (n - 66). A toul of 959 controb (404 in H'arbtd 555 io Shenyanl) wers tnterviewed: Cases (tnesn aBr 33:9 yean) and controls (mean a1e SS 4 years) wers clbsety mattlyd on a/e but cases wert less educated than controls.. Re{auve to tDcrr wnlt no formal edu,cauon. the RRs forr women with pnmaryr3umor school. high school technical school or college eduauon was 0 9. 1.0. 0;! respectively (RR for finnr trend-0.9:, 95% Cl 0.1-1.0). Srnokw= hab'ru Table I shorrs the pervrntales of women by 5-year aBe group who smoked cigarettes for 6 months or longer. The preval- ena of smoking in the Ieneral popuLoon (l.e. among con- trols) vaned w+th aBe, being much higher lapproumatdy 40'i.) among women 50 or over than among women below 50 (smoking rate 24'i.). but rncreased risks were seen tn smokets at all' a1ts. For all lung canorrs combined. smoken. atpmesrsd a 2.3-fold (93'i. Cf increased nsk of lung attoer. The aR-. education- and ctry-adjusted RRs for smoking were d:2 (95% C1- 3.0-.5.9) for~ squamous cdl canar. 2_2 (95% CI I'a'=3 2)' Nor' oai snsall' cell cancers. LS (95% CI il.l-1.9) for adenoearetnonu of the lung and 2.5 (95% CI 1.9-3.3) for the 'otber' category which included thosst diapo!sed ebntally. ItrBe cell atscsrs. and thost with mixed or unknown oNl type. Most ( S7*i,) ases began smok- ing before they were 20 yean oWt compared to a0•i% of controls; the averapr age when subpcta began to smoke was 19.9 for cases and 24.0 for controls. The women were not heavy smokers. Few sub)acts (9b, wes. 4% controls) smoked 20 or mors caiarsttea per day. and the mean daily number of dlatetta smoked was 11.1, for casa usd~ 6.1 for controls. Plevertheieaa, tbere na suEf~at varuuon tn amounts smoked to show that ruta of lung canaer Itpu- ficantfy (,•<0.001) increaard .rti invaaanj numbers of aprattea smoked per day and ath mceaanl dunuon of smoking (Tabk 11). Ckar indepeadent ditv srt saea wtth each meaauts of smoking expown within cateprsr of tbe otha., with the asroaatioa stronRr for sQuttmousroat oaU csmnomas than for adenonrtaaotae- At the same lsvel of smokutB. 2- to 4-fo6d daQersaots tn the tnasatude of the nsk bec.hsa the two osii typw wes typralky oE.erv.d. las:i+r swo1"q Tabie lI7 xhos the RR• assodas.d .ith paainwe stteoks exposues. >tnt among aL subjaeta alkrt adJusttfB for Pnno°.1 smokinB and tEea among ooe-smoksrs. Eilb"tot prr ast of aL at+s and controb rsport.d Aavm1 k..d h as iirw oae of taar rs.daoo.a .ith a c»BsEitaat .`o .as a smoter. Tbaa .en no znt6aaee nts-onosra(' diA - otr is ea.r havinB Iived with a t>motar. ertwpt foe eaass>DOtasa .bo ly..d with a sqam who tootad. wb.m the riak .s t.duod (RR 0.7; 93% C1 0.0-0.!). The iosr.d r>ak uaoeiatad with a spar .!o sssottrd w.a taa o.ly i. Hartit.: 60% of non-.moAnB ao.rrolr aad M% of so.assotieB eatr in Hartiia rapor,.d trt the .pors e.r nsok•d, oompar.d to 52% of noe-smoatieg ooaroiY a.d 52% of sotasolQaK AIl interviews wete coeduestd in 19a-{7. At the elor of use rsjcrwrment. 1.049 el+ybik pati.nta A.d bas+n idtnsided by the Ffarbin and Sryayans canae regstruw. Nms-htudnd and sixty-four (91.hS) .ers int+rvir..d, 32 (#.1%) dird befon our asempt.d ocotaet, 50 (4.1%) w.e not bcserd and tluw (0.3%) t~f4s.ad to puticp.aa Forty-two p.r osat (a a 103) of the aa. ..et+ diagsioaad by tiawe b.op.y. 32% (a - 309) by cytobo- and 26% (a - 351) by, cadioion. AhSoco the paaanta*a ol patboioB- ialty and cytoiapcally oossrm.d t:ttm w.n hotc in SbenyanB than in Harbin. tly aff-typa dinnbutions w+sts stmilar. La the combined sst of catr, thers .as 44% 1laM (. tr.valss d rtal bf yfver aga IaP Wcarruip- dipig eitrw rwYs fnr t..1 cas a.naaaa .ak soktal Cr. Cwr.i ,yr (»s) • a.lte, t • a.Aat RR !1!% CI1 <!C 201 34 163 24 . 1.6 (I.(L 1a) !0-S6 203 40 1% 3! 2.7 (1.1, t,0) JS-" 232 62 ' 241 43 10 (1.a.3.0) 60-M tM i1 191 31 3.2 (11. S.0) 6] + 137 60 161 40 2.2 (t.1. 3 S)

954 A.H w'L'•w<1LLIAMS rr al TaWe R RR and 95% Cl for lunt canoer uaoaated with intenrry of aawtjna by aeUitypr Dwauow o/ lrnksrj I). ss 1. Crll7ypr CiTmrtrrt prr dov l -19 )0-Jy > 40 AU lunl cancsr 1- 19 1 3(1.0; 1 7r (1181125)6 2.6 (1.9.,3.5) /146/83) 3:2 (2.4. 4.3) 1It7+,103) ~20 1L8(09:3.6) (19d4) 3 3 (118. 6.2) (33/IS) 5,7 (2.9. 111.5): (36,11) Squamous.oat orll I-19 Z.0 (1.3. Z9) (a8rf23) 39(2:6:5'9) (5643) 4.7'(11.7.1) (64/103) ?~ 20 20107;5 4Q 16-14) 3.8(i.7.88) 1I0!IJ) I2.DId3.270) (17,11) Adenacamnoma 1- 19' 0 8(0 3; 1 3) (30r 128), 1'.7(a.l.2.s) (37183) 2:0'19.3'.3.01 (4s/)03) ;?t 201 0 8 (0.3: 2.6) (4, 14 1 3 8(1 a. 8.0) ud, is)' 2:8 (f.0: 74) (7111) '95% confidence intefvab. 'Numbers of cases.contro)s ane rn pucn[lwes: T.Y/e III RR for lung cancer atuoaated with pasa,ve unoke TNIe fV Relauve risk of lung onoer ••-^^•ted witA ynn of.use of exposure aN rr.b/rclt Yo-P,r,rwheri o.Jy Paurrr Sovice of'P¢rnrr swwkr Cnsr, Cnwr. trrwke r.ynianr r.cponre cowerolt RR' cowrroLr RR' Any cohabtnnt o0 112 lill yes Avt-8.2 Spaue no 398402 yes 558 551 Mother no 543 595 yes 413 359 Father no 484 SIS yes. +7243e Workplace no1 403 "8 yes 563 513 ?4,l7 0.e 343,515 0.7 212:271 0.9 205331 0T :98+410 10 119,192 0.9 235,332 1.0 182:2s0 1.1 197 301 11.2 22228 301 I.) 'Adjusted for aae. edtaouon, personal smoking and study am. 'Adjusted for ale. educauon. and study area.'P<0.03. cases in Shenyang There were no siRni6ant trends in risk wtth tntetsstty (i.e. number of ci=arrttes smoked by family members) and' duration of exposure (i.e. years of smoking by cohabitants), earspt for an increasisg risk assoei.aed with increaap0 mte-y of fBtLes''s smokins a the preaenx of the inder'siibjeR` .", There was a small excess risk associated with panive smoke exposure at the workplace. For all i sub7etts, the smokinE•ad7ssssed RR was 1.2 (95% CI 1.0-1i.4). The result was similar for noo•sttsoket+s (RR 1.1: 95% CI 0.9-1.6). There were no stpt6cant dose-respoturt trends associated with years of passive smoke exposure at work. ffrorurr md cook'uit pracricer Table IV preaents M associated with dwation of use of iCang and other heating devwea. Elevated tiiirs were oEeerved for increasing years of use of Kani (partinilarty when hatsd by stoves underneatltl: hesud btict .r.lla or Eoon (i.e. heated by pipes leading from the stoves to the wall or 6oor),, coai' stovea and coal burners. On the otber haad. decraned risks were ob+erved for increas;iaII years of uae of twnsroaa- burntng stoves and central heanag The patterns wsrs poer- al)y similar for smotas and noo-smokera, and for squamous/ oat atl otrinomas and adeaooranom:_ We aao euminad the nska associated with years when coai wood. and central beatinj served as the tnain fuel i for beatins The RRs teadad to rise .vitn inaeaazni utt of ooaJ and d.cGae .vitb inawing use of wood and central beatiej, but uone of the traoQs was sisni6eant Casea tnore often reported that their bomee bacame smoltyy during caokinl and that they tnors froqvently bad irritsud eyes during cookinE (Table V). TAere also was a agnibtaat trend in risk with inawnng numba of tnaah cookad by deep fryinE, although this taetbod of ooobzg was oot frt queotly usnd: The rrsults were stmtkr for squamous/ou oeL caaoers and adeaanranoma and for smokas and tron- smokass. Occuporiow Subjects .ere asked about aD'joba in .+f3icb they bad rvrtsd I or more y.an, sntb c.aes and controls compusd in tarma spenRc trcaun{ de.xo EzpvSwr ,~rorsi Carrrconrrols RR!9Ji Cl Kans 0 2340 1 0 1-39 3640376 1 i4 (0 t. 2 41 40-49 132t144 1 1(0:6. 2.8) s0 + 4151393 1.6 109.2 :.8) Burning Kanp 0 677'740 1.0 I-20 106/91i 1.2(09. 1.7) 2d+ 173'l122 1.3(1.1.2.01 Coal sto.s 0- 20 192R26 1 0 Z1-40 311,4qs 1.2II'.0.1.6) 41 + 233/242' 1.3 (1'.0. 1.7) Monioa) stoves 0 212:183 I.0 1- 20 367i 340 0 8 10 6. 1. 1) 21-30 2391293 0 1(0.3.0:9) 31 + I Ii/l3s' 0.8 (0.l. I.I) Heated bnck rallul{oon o 5861651 I 0 1-20 127,'9a ils (d 1. 2.1) 21 + 243121 W I'.4 ('I.I. 1.9) Coal barners 0 52!/!t3 1.0 1-20 2511202 1.2 (1.0. 116) 21 + 173/16i 1.1 (0!!. 14) . Csntral h.at 0 6021J73 1.0 1'-20 213/201) 1.0 (0a. 1:3). 21 + 139/110 0.8 /0.6. 1.0). 'Adjusted for ap, aduutwn. persoaal smoking aod study ares. Tib V Relrtrve ruk of hay csinwr arooat.d with fe.qoency of dsp frytng aad in unuuoa .Ms oooksnR Ca.nrrawrnewr RlC r41% Cll Deep fry (nmn per awet!) 0 324M03 1.0 1 326/360 1.2 0.0,' L1) 2 170/107 2.1 (t.3. 2.1). 3+ 121181 1.9(1.4, 2:7) Eye irr>:nt(om nevw,nnly 617/'7n 1.0 fomatios 21!/I63 1.6 (f.2. 1:Q, fnqeiaa t9/S6 1.9 (1-3. 2:6) 'Adjuend for ap, .dmatw4 pvsoeal wooking ud atLQy ara of t1Yir empbymeat in 29 job eatepxia Moa (77%) romea bsld u ieaat oets job outade the boms, but tsoi- 6cantly inceaaed reslzs +aere ob.er.ed ooly for metal oneltia~ work (RR 1.5; 95% C1 1.0-21), .5ija a.~atty detze.aad rish was oaarev+d for tceda .ork.n (Rll 0b; 95% Cl 0.3-1.0). The ..n al.o asked a they srs expoeed to 12 .peei8c diew, tmoRs or fttma at work, with from I to 16% ttportiag on-tJw-job ezpowna to tLe 12 pollution nema Ciaa rcporvd ezpo.urt to ooal' dust (KR 1.5; 95% CI 1.1-20) and to trmote from bytntng fud (R.R 1.6; 95% CI 12-22) apillaaaiX morn ofEed

prior fautg diuau Tabk. VI lists RRs of lung cancer associated with spenfic pnor chronic lung diseases. Lung diseases that werc first diagnosed wrthln three vean of lung cancer di,agnosls (and a comparable time period for controls) were excluded from the analysis After adjusung for smoktng. history of any prior lung disease was associated with a 50•/s intrea.xd risk 195•/, CI l.=- 1.81: The excess was greatest for pneumonia (RR :.ID. An increased risk was found for bronchitis and.or emphysema. but the association was limited to sqtumous.oat cell cancers (RR 1.6) and not found for adenocarnnoma (RR 0.19), We investigated whether risk of lung cancer varied aaord- mg to the lag trme following the diagnosis of prior lung dtsease. Earlier detectton of chronic bronchrus, emphysema conveyed greater risk. Relative to those wtth no history of chronic bronchtus,emphysema. the RRs were 1.3. 1.3. and 1.7 respecttvely for conditions detected 4-10. 11-20, and', 21 + years before lung cancer diagnosrs. On the other hand: the RRs were higher for more recent diagnoses of pneumonia and TB. The RRs were 2.7. I:S and 1.9 respectively for pneumonia. and 2.8. 1. 1. and 1.2 for TB fint detected 4-10, I1-_0 and 2l + years prior to lung cancer diagnosts. The elevated risk associated with TB diagnosed 4- 10 years prior to lung cancer was signrficant: it was observed for both squamoust oat cell cancers and adenocarcinoma of the litng. and' among non-smokers as well as smokers. fineih hurors of TB and cancer We observed a significant 60•/. (95'r. Cl 1.2- 2.1)iincreased risk associated with TB in a household member, with similar risks for squamous,oat cell cancers and adeaocatrrtnoma. The familial associati!at was seea in smoiten aad twnr smoken. and remained unchanged after, aadjusting for personal' history of TB. The risk assoctated with family his- tory of TB increased with decreasing age when the index subject was first exposed: After adjusuog for smoking. exposures at age <21. 21-30 and >30 conferred risks of 1.7. 1.5 and 1.2 when compared' to those with no household TB exposttre.. Family history of lung cancer in first degree relativea. reported by 4.5•i% of the cases. was associated with a sipti- ftcant 80'/. (95'/. CI 1.1-3.0) increased risk. There was little difference in risk by cell type or smakint status. The risk of lung cancer was somewhat higher among those with a family history of other cancers (RR l.A:,95Y, Ci 1.0-2.0). with the exasa nsk being higher for adsnocarnnoma IRR 1L8) than for Iqusmousroat ceU carsaers (RR 1.1). .4fenarrual and nyrodrcri.e Jacrorr Table VIt presenu rists of lung anoer by various menstrual and reproductive factort Tbe:e were little or no association with age at rnettarche. panty. hysterectomy: spontaneous aborrtion. pregnancy rmultutB in difficult 4bour. and use of oral contraceptives. Thers was a signibtant 50'/, t93'/. CI 1.2-1.81 increased risk associated with history'of ausearria.ge, and caxs tended to hsve a later a=e at natural menopause although the trend was not smootlrl LUNG CANCER AMONG N'OMEN'fN CMINA 995 TrW VII Relauve nsks of lung ancer associated with ntenstrual and reproductive factors Casrr'cowrroL RR' '9J'/• Cli Ar au menarche I6• 194 192 10 16- 17, 427 412 11(01. 14) I4-IS _95:276 1 1 I0,8: 1L4) <14 5564 09(06. 1;4) Number of children <3 193 :05 1 0 3-4 119 300 ~ 1 1 109. 1 5i 5-6 •'327" 10108:,141 7+ 169 174 1 0 t0 7: 1 3) Aye at natural menopauu <45 77 112 10 43-49 373 303 1.7 (1.2. '_.4) 50- 54 :71327 11 (09.1l1 5s+ 31 .18 Cr(1.0:32) Postuve history o( Hystereetoeny 16,36 1 0 t06: 1 6) Miwrria1e 112.126 1 5 0 2. 141 Sponc abortion 239-211 1 1 109. 141 DiAScuh labour 16 61 1 ,3,10 9. 1,91 Oral contraceptive v 64 09105. 1.21 •Adlusted for age. eduonon. personal smorint and studr area, D+rrotr /acrors The diet of the sub)ects was domttsated' by staplt gratns Imedian intake among controls - 1.095 limes per yearl: fresh vegetables 11.199 times per yearl. fermented salted foods 1730 times per year). and soya bean products (365 ttmes per year). Less frequent was consumpuon of antmal protein sourm (231i times per year), fresh fruits (52 times per year): and peass and beans (~12, times per year). Risks of lung canctr in relation to dietary intake are shown in Table Vlll. Higher frequeesaa of intake of vetetabks. either thosc rich or low tnn carotene content were not stgntfsantly protective agarnsr lung eanar. The three foods with the htglmt carotene con- tent in thts study population were drsed hot raf' peppers (16.9 mg of carotene per 100 g). dark leafy greeru (2.7 mg of carotene per I00 g), and carrots (2.0 mg of carotene per 100 g). Carrots and dried hot red pepprrs werrt consumed Itsa often by aus compared to controls, but t!>eae ttema were not frequently consumed (man intake among controls was 41.4 and 70.0 times per yeu respactuvely). On the other haadd eases had slrghtly htgher inukes of the mors commonly consumad dark lafy IIeena (ave;rsge intake amorsj controls was 163.5 times per year): Casra rsporud higher frequencies of intake of aaimal pro- tein and fresth fruits. Few womett ( l2`/. casa versua 8% controls) drank alcohol moee than otsa a year, but they showad a significant smokinj-adjusted 30% increased risk of lung cansxr cotapued to thos who did not dntsk at aU. Hov+ever,, thea was tw cfctv tresd with itsRSasi.ng alcobol consumption. There were no apprsaabk dlRetsnces in die- ury patterns foc squamouuoat ceil tzoeen verstr adeaocar- ei.noma, nor for smoksrs veram ews-smokas. ~..~ TaW MU Retati.e rnk for, htnte nacer a..on.ud with ptnioaa Itmg danna W ' All7irW .Sqr.wwrlwr AAn.ararrwrmise r ~ Caant controtr RR' r 9!x C!1 M Rr 1P RR' ~ Poptive ststory ol` chronic braaclnu 210/137 1.4 (1.2. 1.11) 79 1.6' 46 0.9 W andlor etnpayresa 66/3! 2-t (1.3. 3.3) 23 2.T IS 1.6 leal Praunwilas tuberrulo.u 10343 ' 1.3 (0.9. 1.7) 33 1.2 33 1.1 ' 0 .Aajtaud for aya..dursood perao°al eooittt6 aed acuQy,utat RiNIbv olc.na ath facwr.'95X con6deats imrmt• esctu0ts I.O.

Mi A H w'U.WILLIAMS rnad' TaYY V1[I RcdauW nsk of luni oncer a»oaated with dinary faaon l.wk. I twrs prrDirrmv jorta yWy/ f'au/ronrrol RR 19JSf C11 Supk Vain < 1095 308, .166 I.o 1095-11 46 352; 396 01107. ,1.1) > 1146 :90.290 0.901. ,1_2) Peas and ticaru <1 256,241 I.0 <-1S 221 244 0.9(0.7;,1.2) 16-52 31913!14 1 110e. 1..) > 52' 1601152 t.o (0.7;,1-3) Soya bcan products <1i53 2]2217 1.0 153-363 204r266 0.7,(0.5.,09) 366-a15 265%2S0 0.9 (0.7;,1.2) >sIS 23B/219 1.0(0.9 .,1.3) Antmalprotein <109 t56=23J 1.0 109-230 239/236 1.6(1.2. 2.11 231-4t2 235i237 1.60.2. 2.1) >6s2 336i2N1 2.3 (1.7; 3.0) Femxnted,salted < 366 234,173 1.0 foods 366-625 IR9/I5A 1.2 (0.9, 1.6) 626-990 3291306 1.2 (0.9. I.S) >990 2Ut219 0.9(0.7a 1.2) Veyrtablee' IoW in <366 2St,25'1' 1.0 carotenr content 366-547 256,237' 1.0 (0.l, 1.3) 5al-730 2a1;240 1.0(0 .a. 1.3) t ;0 731 19l:210 0:1 (0.6. LI,) Ve{rutikt high in <731 201 221 1.0 carotene content 731 i- 1095 355:331 I.1 /Q9. LO 1096 -1 460 195 / 197 1.0 10.1. 1.3) ;0 1461 205,201 0.9(0 7, 1.21 Fresh fruits < 19 203/232 1.0 19-52 2091249 L0(o.1. 1.3) 53-132 256r231 14( 1.0. 1,1) > 132 2t{/240 1.50.2. 2.0) Alcobol tieverqss 0 649/706 L0 Ii-12 110/98 1.3(09. 1!7) 13- 52 s3 v76 I.o {0:7. 1.5) > 12 116,75 ) 1.300. 11) 'Adjusmd for age. educaoon, personaJ smotunl and uudy' arn. '(nclVtde rhite potato. p.k r.ea potato. white .eletabin. yelbw and peen lourdi. `Indudea ultsd vegetatiim dart r.es potato.,ydlos Iroeti squasli dart peen lefy, pxn,, yello. .nd light p.en IaJy ve{eubto, curou, red peppen, dnad hot red peppen, p.en ps, tomatoes. MtJtirar•iau artolysis The factors found to have a sipii5cant effect on risk of lung caneen in univariste analysis wers evaluated simWtaneoual)r in muluvarute unconditional lopstic rsQesdon analytia In addition to smokin{. the following varnabb bad a aiRsiRwst effea on nak of lung caacsr (r<Q05) and t>sey enut.d tJ>e vepesaaaa model in the otdet tn abowra: deep-frying. eye imution pxumottia. Sotneboid tubarulota& burning Kan{, self•rcported oavptionai expsaus to burning fuei; puore smoking from any bottaehoid tnemb.r and heatad brick ..W Aoor. Dlniani.o TEi><,popuiatioapu.d nae-oonvol audy conducted 'm two lr* npa6ess C3iaws tatie. rereaNd that at iwrt 33% of the lunt&.ooen ataon{ women ean be explainid by tiptettr smokin{. Altbou{h this attribuubie riak ia lbw oompared to Car~•••n femak populatiom (Lubin t liot. 1964). it is hither, than else-bere in China (Chan .t d., 1979•, 4'sao rt d 19sE). maialy bsnuae of a bi{her prsvaleact of smoking women in tliiu repon. Smokin{ ratss among women ovec ap 50 - oearty doubk tboa fosmd in SSanghai or nationally in China (Gao rr d:, l9'Ia, Wen{ rr d.. 19i7). Furthermom women in Harbtn and Shenya,n{ aarted to smoks at a rnlativey young a{e. As cotnpar.C to women in Shan{hai where 19•r, of female smokra in the Paeraj poyulauoe bc'pn smoking at tyc 19 or youn{er., apProximate}y 40% stan.d at tJsn ap in nortDern CAin& Htno., even teou{! atnouau smoked wras b. (.v.npn{ aighi eipratyl par day among the axs). smokin{ contributa to the elevated rates of' lung "caacer ataoa{„nonl"n,,,,CEianc._womea. It also appan to-sooount^ for tlse~bi~er peroeteta{e (44%) of squatoowJolt3,;eU,ca,ts<ea in our study verszu 12% and 35%. esspectlvely;'fe Shtsnjhjj'aad Hon{ Koti{ (Gao er al:. 1988; Kun{ rr al.. 1984). The tslativify low mean daily number of ci{amtes smoked by thne wotnen may explain tbs lower relative risks of lung cancer among Chinese compared to Caucastan smokers. We obaerved no overatl association between lung canoer risk and passive smoking. Our rssulu vanetl by wurss of puaive smoke expxura, however, with non-smokin{ tises reportin{ less exposure from spouw (but only in Harbin), morr exposure from fatlien, and similar exposure from mothen when compared to norrsmokin{ controls. Despte the large sia of our uudy, we were unable to danfy the ma{nitude of nsks due to puavt smoking. recognised aa a catue of-tunr nttoer around,abe world (Sur{eon Creoersl. 19E6):,Parhaq in this study population the effects of enviroo- mentah tob.oco;'Imoke-.vaa 'oliarvr.d'* the=;t•atbeti:bsavy plpOSW!'a'"' W'~'POU4tiDp',, ftomy;;OCaI'burnin{ ICan{,'.: other indoor beatins soun~e4,..nd higA levels of net{libourtiood afr pollution (Xu er a/.. I9S9). Poltution from coal burning sams likely to contribute to north<astern China's elevated lung canoer rates. Risks in- creased with incrrssin{ yean of use of burning Kan{ and heated brick walis/Eoors, and we observed weaker but stmilar trends with use of coal stoves and coal burnen. Levels of air pollution have been reported to be high in both Harbm and Shenyan{, with both indoor and outdoor winterutne benso- pym-?* concentrations eu:edin{ standards for atiea in the United States by more than 60-fold (Dai er ol.' p:nonal communtcauon: Xu er d:, 1989). Coal burtun{; especially uac of a lonl' smoky eaal. has also been implicated in the EO lung nrtrer rates reported among women in Xuan Wei County in southern China (Mumford' rr al:. 19i7). 71m etfeccts of oertain workplace exposures on lung cinoer resemble those reported in Shanghai (Lenn et d.., 19d7., 1988), including a decrt:ased risk seen in textik workers. The excaa risk among women employ.d in metal smdtin{ u consistenr with the three-fold incneased risk amonj men exposed to inorpnic arsenic in copper smeltin{ in Shenyan{ (Xu rrd:. 1989) lod the United Statea (Lubin et d:, 19i1). The occupuional flndinp wil!' be presented in more detil' in a septnte report. Our 6ndinp that ears +ras more 6kely to cook food by deep frying and to more frequently report eye irTitation when they cooked aro oonsatent with the inaaaed naks auocsatad with exposure to cooking oil fumes in S6an{hai: (Gso er d., 1917). The association in Sbaaglui .u Itronpa for us of rapaeed cookin{ oil, but fe..romen in Harbin or Sbenyao{ used thia type of oiL att{{esani that vapon from t+evert<I types of cooking oib may be Gnksd to mcieaad riat. Coe- denaatn of both tapesead aad sroya baats eookin{ oJ voisula have beea found to be muta{enic (Qu et a1:, 1966). Further shoR-term testing of several typr of cookits{ oib u uadta- way to help ident3fy the r..ponsibk consti'tveau and p^vrid. Ita& for adldition.l study. Ceruin lung diaeaue may 6.ve an aetioiopic rds io lite{ caacsr devebpmeot (fuso er d:, 1967; , Wu er d.. 19{S~ Stei an anooatsoa is of parvcvlar itaportaeor ist Csina. Aes the pre.aknoe of chronic !tm{ ditmur is high. Indwd we found that 3 5% of the caas and 24% of the oontrols rtsportd ad prior esronic lua{ dia..an. Like oth.rs, we foued aa excsn risk of nquamwr/o.r eeli tsnart of the hm{, but sot adeaocaxkn.nma. in a.rocaanoe with rhronic bcme"itia/ empbyt+ema. Our Ondin{ of a rpa5oa inaoaard ritt uaoaarod with noent dia{eo~s of TD (is 6-10 y.an pios to luni cancsr)' ia coe.asmt .ith r.ruha from Si'na{hrai (2aens er d:. 19N). Our resulta an atpqorti•.W of a f.milial tendency in lung caaesrs (Colyen n d_, 1977; Ooi ir d., 19b6ab;, Skilji•ud er d.. 1997; Wu rr d:. 19911} Shared eawontn.oraJ'ctpasuns, familial'ajpsptioa of amaloaR babttat andJOr {eaeoc ptndatr poawe may bt+ moportaeL The prrssnp of aaw 6avin{

affected first.degrce (Ittuly members was small (4'i.). Rezcttt case-controllstudtes imGreat Britain (}kyesh elal:, 19g4'):and the United States (Caporaso er al,. 1,989): however. suggest that genetic traits may tnlluertee susceptibility in a stzeabk portion of cases. These tnvesugauons revealed SipttfscanUy increased risks of lung cancer associated with the geneuqlly controlled abt7izy to extensively metaboUse the drug debnso- qutne= a trait aftectml 5<'i, of the control population studted in the United States. We found'no strong suppon for a rolc of hormonal factors for lung cancer ovenll or spectfically for adenocarr:inoma The cases did tend to experience menopause at later a=es- but the trend' in risk with age at menopause was not smooth. Hlstory, of prolonged labour or hysterectomy, which had been suspected as risk factors for adenocarnnoma because of the potential for trauma-associated lung embolism. occurred more frequently among our cases. but the excess risks were twt stgntficanr since relatively few women were aBtxtnl. RJsk of lung cancer w•as recrntll reported to be increased'among Chinese women with short menstrual cycle length (Gao er al:, 1988). but thts variable was not assessed in the current study. In other countries the risk of' lung ancer is generally reduced among those with higher dietary intake of carotenoids iZtegler. 1989); but our findings are ksa clear. Cases had slightly higher rather than lower intake of dark green leafy vegetables. the most commonly eonsumed rich source of carotene. !Noreover, in our analysis using a com- bined index of all vegetables rich in earotene, high ftequen- Refeeas AYESH. R.. IDLE. J!. RITCHIE. 1 C. CROTHERS. MJ. ! HETZEL M R. (198Y): Metabolic oatdatton phenotypes as markers for suasrsp- ttbJtty to lung cancer. Varrrr. 312. 169. lRESLOW. N E ! DAY. V E. (1980); Sraurncd MsrAodr ir. Canrer Rrsra.ch rM Analixu of Casr-Conrrof Srr.die:. /AAC: Lyon. CA!'ORASO: N E.. FALK'. R T. ISSAQ. HJ ! 5 others (19l91. Lung cancer risk. occupational exposure, debnsoqutttt: metabolic phenotype Cancrr Rra . 49. 3675. CHAN. W C. COLIILORNE. M:1.. FUNG. S.C: A HO. H.C t19791. Bronchtal i ancsr in Hong Kong 1976- 1977. Sr. J. Corn. 39, 182. COHEN. !l. DIAMOND. EL.. GIIAVES. C.G. a 6 others /1977). A common familial component in lung nnat ind ehtonte ebatrve- uve pulmonary disease. Lorcrr, IL 523. ERSHOW. AG a CHEN. W K- (1990). Chinese food ootnpotuion tables a translatton with English common namea. Latut sesentt)Sc narrra. and Pinyin romantaad tnunlitersttom. foal Can/. And. (io the press). GAO. Y T. RLOT. W Ji. ZHENG. W. t! ottten (1987): Lung oam among Chinese women. Jnr. J: Coarn. M. 604.. GAO. Y T. tfLOT. w 1. ZHENG. W.. FRAUMENL 1.R t NtiU: C.N. (1988): Lung oancer and smokuy ta Seanghn. lat: J: Epuktwrmf.. 17; 277: HINDS. N W..STEMMERMANN. G.N., YANG. H.YV a 3'3 tt<bees (1911'). DiRerenoea In lung cancer from rtsokiy amory /apanaa. Chinesc and'Ha.rauan woanen in Hawaii. Aa.,J: Cwrn. 37.297. KL'NG. V. SO. K. t LAM. T. (19H). Lung nncer in Hoty Kong Chinese: mortakty and ItntoWye typa 1973-19N2 lr. J. Cwrw. lL+. 381. taW. C H.. DAY. N'C R SHANMUGAAATTIAA/. K- (1976). httodoO ratn of npro>3c daolopnl typr of IuR{ cannr iR Srsppore Chinese diaMtt poupa, attd teeu aetiobOd stgntEtaeaa /at. J: Cawrer. 17„ 304. LURIN. 1 H. R/IAT. WJ (191it)Y Aaaaatmmt of ItrV t:we riak facton by histologic tate=ory. J. N.d Cwnrr hw-, 73. 393. MUM'fORD: I L.. HE. X.Z. CHAIMAN. R1 i 9 otAan (19871. La.B nnoer and usdoor ur pollntioa ia Xvaa MHai. l:hua Srier.. 215, 217, LLNG CANCER AMONG WOMEN JN CHINA 1167 ctes of intake dtd' not confer a significant protecuve effect. Reasota for the absence of proteeuve eCeev art not cltar. A posstbk expianauon u that thrtx-foutths of the studyy population ate vegetables htgb tn crroteae content at least twice a day so that the nearly uniformly high intake of carotette{ontatntng foods Itmtted varubthty and' hindered detectton of' an effect. Data on plasma carotene levels from this study populatton wtil be tmporunt as a more objective measure of thetr, dtetary intake. Misciasstfieation of intake also may have dampened trends We did not have inform- ation on portion size and the highest carotenecontatrung food in this population is dned hot red peppers. usually used as a condiment. In addiuon, recall of past diet may have been influertced by recent dietary tmprovements. perhaps more so among cases who may have been given preferential dietary treatment because of their illness. In summary; this iovesuptson revaled that contrary to a prion expectation in China. cigarette smokrnj is the major cause of lurtg 'ancer amoni Woaron us, oonAsast Cliim and contributes to the area's high rates of mortality from this tumor. Prevention activities should emphastse smoking cessa- tion: while additional study may help clarify the role of indoor and outdoon air pollutaon, chronic non-malignant lung disease, occupational exposures. famihall susceptibility and other factors in the aettology of lung cancer We thank Joan HoWland for preparation of the manu.enpt. NATIO;NA1 CANCER COKTROL OFFICE (1910). Vw/Nt /iunrrre of Gev:raphy .frlar of Cewrn .KorraJuy ar rAr IroP/r + Rrprb'Lr of CAwo: China Map Press. BrJma. 001. W L. ELSTON. R C. CHEN. V w_ RAILEY.WILSON. 1 E a; ROTHSCHILD. H(I9861. Incrsaa.d faeulul nsa for lung cantsr. J: Nar/ Cawrrr hur:. 76. 217 001. Wl.. ELSTON. RC. CHEN. vW. {AILEY•wILSON., 1E ! ROTHSCHILD. H(19{6)' Famthal tung nntsr.corncung an error in nkulLUon. J Nar! Catrn !nu. 77. 990. QU. Y'H . XU. G X. HUANG. F. FANG: 1 C & GAO. Y T 119g6): An Ants test on other by-pradaeta of the hsttng of eoottn{ oala. Tw.or. 4 sE. SKILLRUD.D W.OFFORD. K,t ! MILLER. R.D. (1916). Higher rnk of lung cartar in chronc obatnetive pulnsoeary dfaaase: a prQ specurr tnatched controlled uudy Awra. larmr. MeI.. 1(b, 303. SURGEON GENERAL (19{6). 7Ar HrelrA Cor{rrwrrwrn of Ikro/rr rvv Swaetwt Deparvnent of Health and Htaman Servsoa (CGCI. RuElintsas Ntesttis 17-319{. Qiowrvaret htnung OR1oe Wishiajtoe_ D.C. srENG. XZ. HONG. LG: a CHEN. D.Y (19tT7). 3rnoking pnevahncs ie Cluess ag•A 13 aad abo++e: CA.a M.L J.. tM. N6- MU. A.H.. YU. M.C.. T11OMA1 D.C.. lIKL M.C. a Htt.NDER3ON. 31.G (191i)I. Iersoaal 'aad fatruly, history of lung di.ean aa nat facton for adenasaratsoma of the hteA Carn Rsa:. 4L 72'19. XIAO. H. L XU. ZY (1995). Air polluti'os and'ttay caaor in Liaoe- ing rrvvsnts. lbopie's Republic of C3in.. IvCI' Mar.oar,. N. !3. XU. LY_ RLOT. MJ., XtAO. H.!' ! 7 otbsn (19t19): Srnokints- a+r pollution and the Irgh rsta of tiaq omr t. Shaeyan{, China. J. N.N C..crr Aur.. al. 11[p. ZHENG„N.- tK.OT. MJ. L1AQ M.L & S otlof (1987). LYy a0t: and prsoe ntbs+eubr nfta:aos al SiaoBltr. &P. J. Cwrrr. !i. 501. Z1EGLiL t_G. (Htl9). A t..rw of eq.Odolot;c ev~d.aoe thar" oroteeoar ndttcr the mt of oaaae. J Nrv.., 199. 116. ow,