Document 93jJ67yO2Rp7XJq2oLDwQbNYe

ri fa i. mm Supplementation and lx> phj|ral irsmns l 'i N e w lie n tr V M lh j . n U C, Adam k 1 .M itili. M , R o lli, 0 . Supharkarn V uni (.m g I C.jsI ih and -opliagi-jl i an mogi-n ms m o d rl liti th t<le n iiIn a im ii n i m i, .im i |i i i i | i t uve lai lo Food C h em Io v i.n l. .'4 I M I H I 1*86 Ih R e n d in g V *. L H rn atle A s Rose. I. i unti In Plessi | I' NuiniM 'O jl Min H Alni .n | mi)h iIj <mm piedipoed In rsophageal u n .<> N u li Carne 4 2'2 .'Ih . I <t 17 Si<Miri. C l> k.ughn M 1 . R n M i'l R R . R l's lu I 11. Bownwn 0 N jiln / M j t n i i u 'J N . I(N |. M . C..il.III. A. aiul H alli. C C 1 st.iiilishnn-nt jnl h.n.i i r n u lm n n i 5V 40 I antigen m m v K U liic ri hu man rsnp h ag i-jl i'|u ih elial celi ( j n n i K r , ^1 365 1?1. 1991 18 Bai >* k. M $ M anno. M . R C.onning W ! .. arai Slonr. G CUm.il g ru ih and u -iu l |ir. ,p.ig.itimi n i al c x ip h ag ra l rp iih rlia l ir li In Vtlrn iKo kvillr). 19 401 4 1 S. 1981 19 Regnici M anl Ib im n n M 1.25 l>ihydi*syviiannn I ) , t lim u la ln i i m .tirali ih r tasi irps n i rp id e n n jl iliile n n iu im n n i cu liu red human kciaiiiMK ylrs I Ull. n -nlialion. 4? 171 188. 1991 20 1lennmg. I l Mi h arl I) . C h rn g C . S trin rfl. P . H oibruok. k .. and lu s p a 5 f a lt inni egulatuin il gio w ih and d ille rm iia iio n o i muove epMieim al celi in lu ltu ie <`rll. 19- 245 254. I9 60 . 21 C o n . Y i D iav ia m l . ( m ir in o . M . Inivnsenri. I R and Stngh. P D iiie ie n iu l n i n i ni ( a 1' nn pollile.non ni litm ath. i otnm. and pjiirt*aii< u m N i r l l i l i n n i f l u l i i i N u li. Cant <. 14 149 157. 1990 22 1he. | f . Il.iu g rn . A . A u liup. I l . M iO le n d o n . I A . liu m p 8 I . and H a liti. ( < ( kmal growlh of rp iih rlia l eli In n n nurm al aduli human l>ont hi 1 am e i R e. 4 1 2294 2104 1981 21 Rree. O I I am i l i miniail. R 0 Supprrssmn il d> plana and h> |M*i|>laMa L ah n ini in igai ultured univo hladder epithelium la n ce Re 1 SBh 592 1978 24 Rrsh-i R . R n /rn I' fin -m an . Z fin- N . B ai/ilai. M . 5haul. M and Shki link. I It ic i 1>( a 1 jk iu m m in h-l liel in ih r don r p iih r lu l h>|V ipruM rr.iliin miliK r ii h N m elhyl N n m o N nnrnw iguam dinr m ali n n a lu x 1 aU n im and (ai d n `i C arn e i Rei So 1764 1767. 1990 25 ( am ili. k K . I.ii.ih v in t A ttk i*l. I A . and N rw m a ik . H I ( ali inni am i ah im genesi ni ih r naniinai> gland Ani. | Chn N u lr Sa 2Ohi 208i. 1991 2b lunhala. ( Sudo k . ami Matsushima. I ( all m m h U xidr inhibits MuiiuLilKUt.il n'jili a live I >NA synthesis l>y v mIhhti 1 hlo n d e in Ih r pylofic f liiK O M ill ia l s lo n u l i C'.w<iiKtgenests(Inml I. 10 2 1 1 5 - 2 1 )7 , 1989 27 / k e i. I F . and B ri . B N < alt mm ili'ln 1en< y and g.aslnc lesions m i l i r u l Am | Palhol S i 14 6 194 1 28 1ipkin. M . I in-dnian. t .. W h m w ik . 5 . 1 . and N rw m a rk , H I . C oiom c rp iih r lu l 1 r ll p i.J ile ia litm in respond*' and n o m rsp n n d eis in supple m rn l.il ilii'iary rah mm C am Res . 49 248 254. 1989 29 I ipkin. M . a m i N rw m a ik . H I III- l n! ad d ed d ie la iy calcium nn c n k im r epithelial <idl p m lile ijlM m m subiccis al high rn k (o familial ro k m k cancel N tn g l | M r d . I l l 1)8 1 1JB4, 1985. 10. H o ren . P . T u rm an . I . fane. N . W av. Y . am i Ron. C O ia l <alt turn vu p p in t*! iik 1ra ve d i n Ial e p ilh rli.il pn Jilri.iiu in n l p r iu in i al rnk (or iok ue cial la m e r Gul. W 650 655. (9H`i ) l (a m . |. I . laiarw vki. R . Allow. F 1 . lu ira m l, | . lu k . G . D . and M aium dar. A P N S u p filrm rn la l all m m vu||Ui'ivrN tik>mc mu oval o in ilh m e d e i a ih o v y la ir a liviiy in rk le ily patient with adertnm aiouv p o lyp i. C a iN i'i Rev S I: 3 4 1 6 - ) 4 | 9 . 1991. ) 2 Lkx X . |a ini>y. K I . lo n g . B B . Oavitlson N O . Silim . M . O . and Braviluv T A k /j mutations m I 2 d in irih y lh y d a /m r mdu ed nlom r lum nrt: r f iw l <>1 vupplem enlal d ie laiy l a k m m a nil vitamin D deli iency. Cancel Re.. S> 4 ) 0 5 4 109 1991 ) l N r w n u ik , >1 I . . W a ig o v u h . M | . and B iu ie . W . R C olon r a n r r r and dietary Ml. phovphale and la k iu m - a hypnihrMS | Nad Cancel. Im i . 7 2 1 )2 1 125. (98 4 ) 4 id r ls lr tn . ('. S.. Ih o m p M m . 5 M .a n d D a v - K I A lie ie d m liacel lular ca k iu m egolalinn m hum an m k u iM a l cam 1*1% and m ' no rm al' a d M ir n im u t o u C am e R ev. 51 4492 44 94 . 1991 )S h . | . Ia > k u . P R l i. C. . Bkg. W . Yu. Y . Irv h o w . A. G . Sun. Y . Yang. C'. 5 . Yang. Q . la n giea I A . Zheng. S , ( jr r n w a l l. P . and Cahill. 1 In li iv en im n vludir in lin v ia n. China: an update I N u ll. C'.inwlh f a m r r 1 199 206. 1986 ) 6 W an g I 1 ) . I ipkin. M 0'*- 5 I Yang. C. R . Yang. C S . and N rw nv.uk. I I I la lie lm g Hide and (.ilieliikg d iclidiuiK in o f ell in esophageal rp u h rU u m id individuals al int eased ok lu esophageal 1 am e i in Huivian. China ( aK ri R e s . So 2651 2651. 1990 UJ .* .' Ht lMUIM\lh U\ I'HU |Hili'(iiiiilii|(y. Rim ill'! & P irtrn lU f Meeting Report Epidemiology of Ovarian Cancer A p p j u h e b R. P a te l' a n d G. Iris O b r a im I ionujr.il PiogMniN Hi..in h. 11(Ii iiimiI>>k> "1 Hiosi.tlivlu I'iukoiii. Otvivmn ol ( .iim<f In.lojty N.ilnm.il I ,imci InUiloli-. Bi-i Im-v I j . M -k >Lim i A workshop on the epidem iology o l ovarian canc er was held at the* National Institutes o f I lealth, Belhesda. M a ry land, on Septem ber 10, I `PM. rh e goals o f the w orkshop were to address: (a) reproduc live fac tors; (b) exogenous horm one use; (c l d ie t, life style, and health practices; (d) tumors of lo w malignant potential and nonepithelial germ cell ovarian tum ors; (e) prospective studies; ( f ) d e te rm i nants o f genetic susceptibility; and (g) directions for future research in ovarian cancer. Session 1: Reproductive Factors Alice W hittem ore (Stanford University School of M edi cine. Stanford. C.'A) reporte d prelim inary results o f the; analysis o f com bin ed data trom 1 3 U.S. case-control studies on ovarian cancer ( II Ih e objective was to assess the influence on ovarian cancer risk o f events affecting ovulation, such as lim in g o f m e iia rrh e and menopause, lim ing and outcom e o f pregnane les, duration o f breast feeding, and duration ol O C J use. These data were examined in light o f tw o cu rre n tly prevailing hypotheses about ovarian c ancer risk I he- ` o vu la tio n * hypothesis was proposed by Falhalla in l `)7 l (2). It stales that re pealed m inor trauma to the e p illie li.il surface o f the ovary, caused by inc essaiil ovulation, is a m ajor risk f.u tor for ovarian cancer, f lic "gon adotrop in" hypothesis ( t). proposed by Sladel a few years later, suggests that e x posure o l the* ovarian ep ith e liu m to high levels o f circ u lating pituita ry gonadotropins enhances ovarian cancer risk. H ie com bined data showed strong (rends o f de creasing ovarian canc er risk w ith inc leasing parity, dura lion o f breast-feeding, and duration o f ( X use. Failed pregnancies l i e , spontaneous a lio riio n , ectopic preg nancy, and miscarriage) were* also proiec live, but less so than fn ll-le im pregnane ies. There was a weak trend o f 1 <incer risk w ith age at men.irc he. no clear trend w ith age at menopause, and a greater risk reduc lio n assoc lated with each increm ental m outh ol pregnancy than w ith each m onth of ( ) ( use. The trends o f dec ceasing ovarian c.mc er risk w ith inc ceasing parity and duration o f ( )( use are consistent w ith both the ovulation and gonadotropin hypotheses.*1 Kr. n o i l I/*'.' 1 In w in >i)l h -(|i i 'M s In r.-|..... 1 'In il ili I h e .nl(li-\M*cl 't i n - .iM h.-m .nions iim -al .h i- CM m .il o iili.n <-|ili\c. I H ki.lllli.ni-. I t i. 1>1i^l.*^*1. litio Kk\ l.-l..ll>l- lis i Itil.-iiu/iiiK A m o n ili ol breast leerling i> 01 m o ie iiio n tlis a llei delivery was lo und to lie less prole live than a m onth w ith in the liis l (> m onths of delivery. This supports the ovulation hypothesis be.ms' the longer a w om en breast-feeds. the less idle live l.u la lio n is in suppressing nvulalion. The risk reflue lio n assoc laled w ith breast -iced ing ccmllic is w ith the gonadotropin hypothesis: in lac tat ing wom en, follie le siim ulatiog horm one levels rem ain elevale! un til the retu rn ol ovarian estrogeni Inn lion. The gonadotropin hypothesis predicts that boast teed ing w o uld me rease the risk o f ovarian canc er The relalionship betw een in le rlilily and in re.ised ovarian cancer risk was als evaluated. Am ong nulliparous wom en, ovarian canci>r risk did not vary liy m arital status or gravidity. There was a weak asso ia lio n w ith duration o f longest attem pt at pregnant y. total d ilu tio n o f unprolec le d intere nurse U d ore piegnan y. or his lo iy o f clinic ally diagnosed in le ililily not .ilirib u le d to le itility problem s o f the m .iliv The n -k was elevated among w om en whose in fe rtility was attrib uted to ina<le<|iiale ovulation and am ong w om en w ho had used (Tlilily drugs. fp ith n lia l tum ors of low malignant potential, lie (|u*nlly called borderline tumors, have various leatuies o f m.ilignant epith elial ovarian cancers, but they do not invade the* ovarian stroma. W om en w ith these tum ors are usually younger w hen diagnosed arid have a better prognosis than w om en w ith malignant Illin o is ( irai on traceptives an* generally less protective against border line tum ors than against malignant invasive tumors. David Kos' (Am eni ,111 I lealth In u n d a tio n , Valhalla, NY) nuli aled that tw o w idely disiussi*! hypotheses ol ovarian earner 1 ausatim i (2 . I) may relate io .d ie i.n y di* Is on endogenous horm ones. Ito ih hypotheses ,ue com patible w ith the know n in le rn a iio n .il and 1eg10n.1l associations betw een ovarian [.m ie i risk and dietary tat icm sum ption. Although such .isso 1.1Iions do not im ply causality, the increase in incidence o l ov.iu.in, breast, and prostate cancers as fat consum ption has tin teased in |apan is particularly striking The plasma estradiol and esim ili* levels ol I *<111 prem enopausal and postm enopausal w om en w*ie lo und to decrease when they were s w ilih e d lim it a typical high-fat ( l r) <10% total calories lu m i la ti diet to one provid ing only 20% 01 less o| calorics lu m i tal Su ite lung from a lo w Itlie i to a high lih e r diet had a m iiiiI.ii ellec I A lo w tal diet also 4 aused some redo lio n 111 m id ln le a l phase plasma 1 11 levels. Kos* slated that there are several studies ol w h ich epidem iologists should he patii nl.irly aware w hen think in g n h o u l ovarian 1 an 0 1 . I or i*x.imple. .1 study published by l.loyd et .if. (41 m e iils o n iiin ia lio u and elahoiaiion. This was a .is* c o n im i study 1 examine the ette 1 ol nutritimal lai tors on n ie iis lm .il Inn lio n and hone den Nily II) ciliegiai' w om en. Mire* groups, m all lie d w ith Mi-fling Rrprl: I pirlrmNil.ig* nl O tJ lijn Cil)i ir>|>ec i lo age. height. .m il weight w r it- studied: seden lary women w ith tegular menses. athletic w om en w ith regular c and 1 small group o i a th le tic w o m en who h.iil he m e oligomeric x r lie ic . Average age ai menarc he was grealer in norm al ( M l yeais) and nligom cnorrheic I H . I years! alhleles ro m p a ie d w ith the sedentary wom en (12 2 years). Aveiage hone dertMly was low er in the oligomeiionheic alhleles t ornpared w ith normal ath letes and sedentary wom en. Dietary iiher intake was significantly higher in the nligom enorrheic w om en com- pared w ith the other tw o groups ot wom en. It was to m luded that increased dietary iiher intake was asso tia ie d w ith menstrual dyslunrtnm in t ollegiate alhleles. I'lasma hormones wen* not measured in this study. In another study (5t. menstrual irregularities were reported among vegetarian w om en com pared to om nivorous wom en whose diets were somewhat higher in fat hut nun h lower m Iiher I lie sigmlic am e ol these findings in the (o m e si oi ovarian ra n te r risk needs further <onsideration Pirke c l .1/ ((>) studied groups o l prem enopausal . women w ho wore pl.i< ed on weightrodne mg diets, one group receiving a vegetarian diet and the other a m ixed diet. W eight losses w ere similar in h o lh groups, hut after f weeks the w om en eating vegetarian diets exhibited a shortening o l the menstrual ry t le and some liecam e anovulatory. These <hanges were assoc ialed w ith a sig nificant reduction in the height ol the ovulatory plasma I It peak and in the luteal phase LH com entrations. M idluieal plasma estradiol levels w ere also reduced in wom en on vegetarian diets. I till el .1/ (7) cauied out .1 dietary in terventio n study on 4 volunteer nurses Two w om en w h o were switched troina mixed diet to a vegetarian diet reported shortening o l the menstrual cycle by 1 2 days, and their serum LH (leak occurred earlier in the menstrual cycle and was o f reduc ed amplitude. fro m these data, it seems that la i and Iiher iniluenc e the menstrual cyc le, the hypu thalam opiluilary regulation ol ovulation, and the levels of i irculating estrogens. M oreover, it appears that a low lat/high fiber diet may result in fewer ovulatory cycles, w h ic h could favorably iniloem e ovarian t am er usk. PCs ol the I ami f sene* at cum ulate rapidly in the preovulaiorv follu nl.tr fluid and reai h .1 peak at ovulation, the stimulus being provided by gonadotropins. Both l.H and estradiol at te le ra le PC production, w ith PG f c o n tributing to vasodilation and progesterone production, and PCM., stimulating contraction o f the m yoid te lls involved in the extrusion o l the ooc ylt* frillic It*. Inhib ition ol PC synthesis prevents rupture o f the follicles, and suppression ol ovulation ta n he induced in women by treatment w ith a PC synthesis in hib ito r, indom ethacin. It is not yet know n w hether long term treatm ent w ith drugs su< h as indnm eih.it in reduces ovarian cancer risk. Session 2: Exogenous Horm ones M alcolm Pike (University of Southern California. I.os An gelos. CAI began his presentation by jx n n lin g out that the age-spec ifit incidence of most ilun-horm o ne-dependent epithelial ( ant ers shows a linear it * rease (on a log log plotJ. However, there is a distinct reduc tio n in the rale ot increase around age 50 (8) This protective effect 01 menopause is the most fundam ental e pidem i ologic al t.u t about ovarian cancer. Since gonadotropin levels are high in postm enopausal w om en, it seems that high levels of gonadotropins per se may not he im portant in the etiology o i later unset ovarian cancer. I ligh levels of gonadotropins may he im portant in premenopausal w om en, how ever, w hen they have .1 "substrate" lo work o il, namely ooc ytes. Pregnane ies have a grealer protective* effect (ex pressed as rc-dut lio n in risk per m onth) than breast feeding or ( K use*, but die latter tw o variables are subject lo m uch greater measurement error. 1hree aspec ts o f the protective effec is o f O f use need further research. First, how does this effect change w ith tim e and w ith inc leas ing age since stopping use/ The Cancer ami Steroid Hormone* Study data, for example, slop al age 54. Sec ond. W h itlc n io re s metaanalysis suggests dial the use of <)Cs for longer than 6 years confers no additional protec tion. This is contrary to other data (9, 10), and further inform ation on long term use* is needed either to refute this finding or to try lo explain it in terms of, for example, the tim e to retu rn o f ovulation after cessation o f O C use Third, d o m odern low -dose pills have* the same protec tive effect on ovarian cancer risk <i* older higher-dose pills i The answers to these* questions .ire very im portant, since they have a profound c ite d on the risk-benefit equation for O f use. The apparent lack of effec t ol age al menopause on ovarian cancer risk described by W h ille m o re and c o l leagues is likely to shed light on the etiology o f ovarian canc er. W om en w h o have had their last menstrual period after age 51 have u n d oubtedly ovulated m ore lim es than w om en whose? last m enstrual period was before* age 45. This observation may be* evidenc e against the 'incessant ovulation' hypothesis. Prcgnam y. variables related to m enstruation, and O C use cannot explain th e international differences in ovarian cancer rates, it is c o m civab lt* that diet may explain part o f the geographic variation in ovarian cancer rales. Noel Weiss (University o f W ashington. Seattle, WA) reported that although estrogens used for horm one re placement therapy may reduce gonadotropins lo a level between pre- and postmenopausal values, and hence perhaps reduce risk of ovarian c.nicer, no evidence is available lo support a decreased risk ol ovarian cancer associated w ith ih r use o f postnienopaus.il hormones. W h itle n io re and 1 olleagueS have evaluated the relation ship o l m enopausal estrogen* to the risk o f epithelial ovarian cancer using data Iron five hospital-based and five population-based case-control studies.` No clear trends in risk w ere observed w ith inc reasing duration o f estrogen use 01 w ith inc reasing lim e since last estrogen use after adjustm ent lor d uration o l use. O verall, m e n o pausal estrogen use was unrelated lo risk lo r epithelial ovarian earners o f the end om e trioid cell type. Weiss observed a slightly increase'll risk in his study o f menop.msal estrogen use and ovarian i .nicer 1isk (11). Carlo (a Vecchia (istilulo di Kuerche Kirmacologic he. M ario Negri. Milan. Italy) reported results o f the mela.malysis of three hospital-ha secI <ase-t o iilro l studies ' A s W h 1 llr M K 1 ir r l.il ( fuii.u le tislu * ri-l.iliuic l o nv.iti.iii .mi u\k M j Iikimmi nrh. ti.il i.iik i- r s versus inciiii|Mr>dl I'slin gi ii u se. su lm n iled lot |ml il.t .ill. >*1. f J iu rr (pirirmuitufty. RionuiLi f s & Prevrnlion o i ovarian c .iiu p r cond u cte d iii ( ir c c io , Italy, and the UniW'd Kingdom (12 14). in .isu'ssm vnf o f the roles o f | rity and age .11 firsl b irth in parous wom en, hoth fat lors had a weak, nnnsigniticaril influence on r a n te r risk. W hen partly and age al lirst h irlh w ere considered sep arately. there was an inverse trend in risk w ilh increasing number o f hirths. H ow ever, <1 significant trend em erged only from the British study and was largely reslrit le d lo women w ho reported lour or m ore births. Com pared to women w h o lio re ih*ir firsl c hilrl al age 24 or less, the increase in risk for Ihose w h o bore ih e ir firsl (h ilrl at age 15 or m ore (RR 1.4) was ot borderline statistical signifi cance. Ih e poo led analysis also indicated that abortions conier lim ited p iu h x lin n o f 10 40% against ovarian can cer in women w ho reported tw o or more spontaneous nr induced abortions. Unlike the W h itle m o re analysis, there w.is no e v i dence o f an assoc'.itton lie tw e n n nvari.cn cam ei and age at m enarchc. and there was a 1 onsisienl trend o t in creased risk w ilh late age al menopause. Ih e strength ot the assoc iation was relatively weak, w ith .1 relative risk o f less than 2 w ith m enopause after age* r>2 com pared w ith earlier menopause. The fleet o f age at menopause seemed lo Ire long lasting and to increase w ith age at diagnosis o f ovarian cancer. Compared w ith never users, the RR lo r <)( users was 0.6. The RR estimates tor use nl OCs w ere even lower in w om en reporting their lust use before age 2r*. The risk o f disease der leased w ith l l i r ilm a tio n o l use, being 0.7 in w o m e n reporting <)( use lor less than 2 years and 0.4 lor O C use for 5 years or m ore. Ih e protection persisted even after discontinuing O C use. the RR lacing 0.5 in w o m en reporting ih e ir Iasi O C use 15 years or m ore before diagnosis o f ovarian cancer. The protective effect o f OC use on ovarian cancer emerged consistently in all age and parity strata. Session 3: D iet, li f e Style, and H ealth Practices James Marshall (Stale U niversity of N e w York. Buffalo, NY) reported on prelim inary findings Irom an ongoing case-control study of diet and ovarian cancer. There was a slight excess o f caloric intake am ong cases relative lo controls. C rude dietary fiber appeared to exert some* protec live eflec I ro t.il vitam in A intake* had a marginally significant protective eflecI. w h ich seems to be d u e lo -carotene, not retinol. There was no relationship be tween olacsity or alcohol intake* and ovarian cancer. Daniel Cram er (Brigham and W om en's llo s p ii.il, Boston, M A) pointed out that the* " incessant ovulation" hypothesis (or the etiology ot ovarian cancer does not have supporting animal m odels and does not fully explain usk fac lots sue h as early m enopause, radiation, in fertility, and heredity, or the protective role of pregnancy and oral contraceptive use. In 1`MH. C.ardner (15) proposed that ovarian (.n ice r was caused by hypergonadotropic hypogonadism, /.*.. high secretion of gonadotropins due In ovarian failure or lack o f lecclhuc k co n tro l on the pituitary. The th e o ry was based o n .1m111.il experiments, whic h dem onstrated that if you caused m m yte death by gonadal irradiation or use o f one yte toxins such as poly cyc lie hydro arlions, thereby raising gonadotropin levels, ovarian c an er resulted C ongenital defic ienc y o f not yles (germ cells) also increased ovarian c.nicer risk. I hat go nadolropin stim ulation was a necessary com ponent was in ferred from the a b ility o t p ituita ry a b la lio n in d c lii m o. y o l g o n a d o tro p in releasing h orm on e to b lo c k tu m o r de ve lo p m e n t Relevant e o f anim al m odels to h um an rlis e a s e jn o tle ls has been q ue stion e d, since a nim al tum ors are p rim a rily strom al, w hereas hum an lu m o is are pn m a n ly e p ith e lia l. ( ra n te r a n d W elc h 116), Im w e v e i, p m p ost'd that a stim ulus that causes s tm in a l Illin o is m ro d en ts m ay also prom oter d ilfe re n l typos o l o van an tum ors in w o m e n A c c o rd in g to then th e o ry , th e first stage o f tu m o r genesis in volves fo rm a tio n o l in c lu s io n c ysls (islands o f o varian surf.u e m e s o ih e liu m ) b y e n tra p m e n t o f o v a ria n sorlac e e p ith e liu m in to th e o v a n a n stro m a . In th e sec o n d stage*, d ifle r e n lia tin n , p r o lile ia iio n . and eventual m alignant transform ation of the epiihc*liuni lin in g th e nu fu sio n c ysts o t ur d u e e ith i*r in d u e l 1 s lim illa tio n b y g c m n d n lm p in s 01 to m t li ie i l s tim u la tio n b y S teroids indue eel b y g o il.id o lio |H n s to . id ililn in . C i. m m e l . 1/ ( I / ) s p in (||,I. 111.11 I,|li in . i\ ,|si 1-i 111 tin 11 iii.il it . i, | a m i bee o in e me 01 j 101.16mI in to th e 1111 lir - ii m . \ -.6. 1 1.11 liib u lin g lo th e u s k o l o v .iu .in 1 ,1111 i *i . u lm b *. 1 . 1-I->11 w it h so m e (*p id i* m io lo g ii .I el.ii.i, I x p e n iiie iil. il a n d c linn -el s im ile s l u t e M'p t u ii- ii .an assoc a lu m o l galac lo s e o iis u n ip tio n a m i iiie i.ib o le s n i w ilh h y p e rg o n .itlo t 1o jtic h y p o g o n a d is m lu a c . is e e o n lio l study o f ovarian cancel, ( ramei er .1/ ( ill) ........ .. dial w o m e n w it h o v a r ia n c a n c e r c o n s u in e a l d a ir y p io t lin is w i l l ) .1 h i g h e r c o i l l e i H c*t p n - l i y d n i l y / e e l | . m | i *. ( v u g a i il a n d f o lia g e 1 lie e s e ) anal liaal low-............. i- iili. ilio ii* . o l g.ilaa li.sa* I plio-.pll.ale in nh Ili.lir.li'l.i'.i- .ill a'li.'i III" llial aolive'll', gal.11 lease* lo glm o'.e. liiao all. I amin >1 a an* n Risk tor ovarian aam -i was lel.ili'il lo lln* i.ilm ol l.n Iu m * consumption to linnslerase aa livily. ( axes bail a mean lactose consunipticm:iransferase activity ratio ol I 17 compared wilh 0.9ft lor controls. Ilier* was a highly significant trend for increasing ovarian canor risk wadi increasing lac lose* consumption:! r.inxfer.ixe ac livity ratio Biological specimens arc* needed lo further investi gate ihe possible interactions ot various exposures with bloc hemic al 01 mola'C ulai gametic lac lens, sue b as (cans (erase* activity. (.Tamer reconimended expansion ol Ja milial ovarian cant er clinics, which will permit Ihe ida*n rific .tlion and lormalion of pedigrec*s lor linkage analysis, identification of phenotypic markers for ovaiian cancer. <ollea lio n ol biologic al spec mens, and iili*n tifi( ahoo oi iiiarkc*rs that prececli* tum or cli*velo|>nii`iil. Such clinics erne`m irage developm ent o f c Imic al str.itegies lo i prn n .iiy and secondary disease prevention. M ils u r u M<ri (K u rcim e U n iv e rs ity , k u r o in e ( ily . |a p an ) report**! th e results o t tw o m o l.(a na lyse s cat p u b iisb e d case* c o n tro l stu d ie s can Bavarian Q ua e r. I n b a l st**r ili/a lic a ii w as In line ! to In* signilia a n lly assen ial**al w ith re d u c e d risk cal o v a ria n <am e r w h e n n u llip a ro u s w a n n e ii w e re e x t l u d i`( l fro m Ih e a nalysis S ince s u la le ililily is (e la te d t o la oth an in t le a se d risk o f av.iri.in ana a n d a d e c re a s e d lic q u c m y at tu b a l s le iili/a t in n !* iabsa*iveal re la tio n sh ip c o u ld In* in d ire ct. A m o n g A sian w o m e n , in due e d aba art icaii had a slig h t p io te c liv e c ite 1 It has b e e n h y p o th e s iz e d th a t a p o te n tia l <a n no g eim agent e nters the p eriton ea l cavity llucaugli ih e l.illn p i.in lu b e . irritate s the pelvic p e r iliiiii'iiiii. piaacliues pro lile ra tican . .m d , w ith a d d itio n a l f.ic lo i> . re su lts in the d e v e lo p m e n t o l c a n c e r ll this h y p o llu *s i> is c a a iie il. lig a tio n caf th e lailcapian tu b e m a y p io le a t against o v a n a n aa m e r b y p re v e n tin g th e c ana a-r maleic m g cm p io in o h n g agents Iron) e n te iin g the p e rilo n e .il c.ivily M rrtin i Krptur f ^irirmiolofi ni Ow im r Crim Session 4: Borderline and N n nepilheiiai Tumors la A ii'iu c M cGowan (George W ashington University, Washington. IK ) emphasized llu* m r d for obtaining a more detailed <lum al history, greater involvem ent o f a gvnetologic pathologist. .md closer m lc i.u (io n o f a pa biologist w ith an operating surgeon in studies ol ovarian t am er i le stall'd (hat primary peritoneal c a moma ex hibits many symptoms similar to those o l prim ary epithe lull ovarian t am er and that iherefoio there is a possibility ot misdiagnosis C urrently, gynei ologual pathologists preler the term "low malignant potential tumor" over "borderline tum or." W om en w ith low malignant potential tumors are usually younger w hen diagnosed and have a better prognosis than wom en w ith malignant tumors. Analysis o f c o m bined data from nine case-control studies o f ovarian canter in the U nited Stales revealed that the* risk profile lor tumors o f low malignant potential was sim ilar to that tor malignant ovarian tumors w ith tw o e x c e p tio n s / The risk tor low malignant potential tum ors was less clearly reduced among w om en w h o had used OCs. There was clearly an elevated risk o l low malignant potential tumors among w om en w ith a history ol in fertility. This was the greatest dtiierem e in risk iac tors between low malignant potential tum ors and invasive care iiioina. Carolyn W e s llio ll (Columbia University, N ew York. NY) stated that ovarian tumors ol low malignant potential are almost eniuely serous and m ucinous epithelial tu mors. W hen diagnosed, they lend to be localized to the ovary, w h ich accounts lor their good prognosis. There are very lew data regarding their incidence Frequent misc lassitic ation of these tumors reduces the value of data irom clinical series. Benign ovarian neoplasms can be of epithelial, stromal, or germ cell origin. Their occurrence is not roc orded by tum or registries, and misc lassification o f the histological diagnosis is relatively com m on. The tumors o f germ cell origin, the teratomas, are most com m on and have* a ummoclal age distribution w ith a peak near age TO, (lie shape of this age inc idem e curve resembles that o f lestic ular tumors The benign epithelial tum ors occur somewhat less frequently than teratomas, are m ore sub let t to diagnostic misc lassification. and occur about equally am ong w om en from their teens through the 70s. No study has found any evidence that OCs protect against either the teratomas or the benign epithelial tumors, nulliparity and infertility may increase the risk o f these tumors. The stromal tumors are rare and occur in postmenopausal w om en There have not been any re ported epidemiologic al studies of the stromal tumors. Session 5: Prospective Studies Graham C o ld it/ (Charming laboratory, Boston, M A) re ported prelim inary findings from an ongoing prospective study of over 121.000 registered U.S nurses w h o were aged to to 55 w hen recruited in 1976 The original aim o f the study was to look at exogenous and endogenous hormones and tin* risk ol breast, uterine, and ovarian cancers The questionnaire obtained details on the fo l lowing roprodcii five lac tors age at menace he, age at first 1K I i.itfis el .1/ ( hdc.ii ifiisiK > [-Idling ii. iv.111.111 c<m<r f risk Lpiihelirfl 1 in t'i\ nt Iihv 1n.1l1KrL.111i |i<>ii-niial submlied lo* puhiu dliun. birth; parity; weight and height; menopause, including type o l menopause; il postmenopausal, use o l replace ment hormone's; C)C use, including details o f interval o f user, but not details o l ac lu al preparation used. Data were colleded-cm ~drfioronr rype5~ot conirac eqirion. in cluding tubal ligation and vasectomy. The cohort was follow ed w ith a biennial questionnaire, vvhic h allow ed the w om en to update their exposure information. Two hundred forty-seven ovarian cancers w ere re ported in the nurses' cohort by the end o f 1988. There was an inverse association, adjusted (or parity, w ith the use of OCs. W o m e n w ith five or m ore years o f O C use had a relative risk o f 0.6 for ovarian cancer. There was a decreasing risk w ith increasing parity. There' was no dea r relationship o l risk w ith age at first birth, even after adjustment for parity, lo n g duration of postmenopausal horm one use appeared to increase nsk slightly. The relative risk assoc iatc'd w ith tubal ligation, adjusted for age and parity, was 0.17. A food frequency questionnaire was administered to the cohort in I960. There was no suggestion o f in creasing risk w ith increasing saturated fat intake am ong the cases diagnosed during 8 years o f follo w -u p . For lactose intake, only the lo p q u in tile had a slight excess o f case's. To elate, there is no relaiionshij) w ith alcohol intake, lig a ie lte smoking, oi various o ther nutrients e xam in ed. Session 6: Fam ily H isto ry and G enetic Events Bruc e Ponder (University o f Cambridge, Cambridge, Eng land) review ed prelim inary data from a large populationbased study carried out using public health records in the United Kingdom. Information on cancer m ortality by site in first-degree relatives (2106 parents and 1949 sib lings) ot 1183 index case's diagnosed w ith ovarian cancer before age 60 (w ho were under age 17 in 1939) was obtained. There was a substantially increased relative risk o f death from ovarian cancer in first degree relatives o f cases. The risk appeared to he greatest w h en the index case was diagnosed before age 50. D espite the substan tial relative risks, the absolute risks are still small, w ith a l in 40 chance o f death from ovarian cancer by age 70 in a sister or m other o f an index case diagnosed below age 50. How ever, there was heterogeneity in the data; w om en w ith tw o affected relatives were at very substan tial risk. In these' families, segregation analysis was c o n sistent w ith the inheritance o f a single autosom al d o m i nant gene w ith high penetrant e. This hypothesis can only be proved by demonstration o f linkage lielw een ovarian cancer and a know n genetic marker, w hich ap pears to he a marker on the long arm o f c hrom osom e 17. Ponder suggested the establishment o f registries of w om en w ith ovarian cam ers at different ages, sibling pairs, and pairs o f closely related w om en w ith ovarian and breast earner, so that studies can be in itialed to determ ine the* risk attributable to specific mutations. These families w ill be useful for pathology studies, since so little is know n about the early stages o f ovarian cancer d eve lojim ent. There is a need for studies to assess the proli'c live elfer t ol prophylactic oophorectom y for w om en al high risk for ovarian cam er. Ponder concluded his remarks by briefly describing a U.K. National Study recently in itialed to id entify all families w ith tw o or m ore ovarian cancers. Camrr tpidrmiiilii|(y, Binmarkets A Prevrnlii f Irn ry I yne h (C m ighton l Jnivnrsity. Omaha, NE) c*m phasizird the* need for ih o establishm ent o f fam ilial cant er registries and ih e rolle< lio n o f in form ation on cancer of all anatomic sites am ong fam ily m emhers. He also recammendcdthenpst3l)lishnTencois|)e( im enixanks'lorstore rapidly frozen tum or, norm al ovarian tissue, and DM A for distribution to interested invostigalois. Session 7: Research D irections David S< hotte nfe ld (University o f M ichigan. Ann Arbor, MI) summarized three key areas for future research on ovarian cancer: 1 Family registry resourc es should he further developed to address linkage markers o f susceptibility, assess the inlerae lions o f in herited risk w ith environm ental ex posures, and assess the value o f possibly protective interventions such as the use o f <om hin a lio n OCs or other c. hem opreventive agents. 2. Epidem iological studies should further address the relationship o f variations in reproductive patterns to ovarian cancer, the potential risks associated w ith endogenous gonadotropin levels, and the use o f e x ogenous horm ones, such as estrogen horm one re placem ent therapy and fe rtility p rom oting agents. { S tu d io of ovarian cancer should use precise patho logic classification Reference* 1 Whiiirmoft*. A S.ecaf (hjrjrienslus olaiing in ovarian (nt er nsk. CulUbor.ilivr analysis oCtwelve U S t um* i onlml Unities. Am. I E|>*r1rmwil . m press. 1992 l F.uhalla. M f Inci-ssani nvul.ilioii .< I.11 loi in ovarian neoplasia laneel, / 16V 1971 1. Sladel. 8. V. the t-liology end prevention ul ovarian o ik p i Ain. ] O tH lH G y n e c o l. 121 7 7 2 -7 7 4 . 1975 4. Ikiyd. I.. Buchanan. | R., Blnrei. s Waldm.ui. C. J.. Myeis. C. and toid. 8 ( IntefieUnonships n( dn-l. athleiir activity. menstrual status, and bom- densiiy in roHegule Mimi-n Am. I Clin. Nnlr. 46 681 684. 1987. S IV . 1 Is,-II. A H . Ha.ll.nl.unn-w. M J . 1ro te in e 1 A . Al,,,. In 1 P . N e lle lm k. * 1 d 1loyd 1 M eilslfii.il ill feri' es , tin- I ll V g el.10.ill and noniveg.-ia.i .in d iels A ir1 I C lm N u ir S 1 87*1 8 8 ri. 1191 6 Pilke . X M . Si h<w.-igei. 111 . 1aesste. K . D ii kllau l. H . Shsv gel, M . and W a ei h ile i. M D ieting Illlli.e m e> ih e llle .1siru.1l < 1-. S vers,,, ,SImvegel, Ifl.li 1 (Ile i 1fill M en i . 41. m it i emu. 1981. 7 1lilt. P . < tl.HI. P . t i i Ih 'I11. 1 . W ylifter. 1 . and Kmno. X Mk -I and . 'U d o ri.. 11- lel.ile f l f . l lin e i ( . UH f-r (I'iiila 1. 1 l l p o IH2G. I `I7 7 ti Pike. M . ( Age fetali-d lai In is in am ers o l Ihe liiit-.i|S|. 0v.n y and enfiarne m u m . | < hi on l i e ... 4 0 (Seppi 2| r.9S (,9s 19ii ; 9. Ih e 1( am e i .illMl SCemid H o in io m ' Siurty o l llie ( iih is ol1 disease Co n lm l and ih e N .i Ik m u I Ii i MiIi i Ii - if ( liik l I li-allh .m il I lu iu .u i I le ve lu p im -nl. I In- ii-diK lion in nsk o l o v a iu n i.i m e t .ism 1.1I1 1 I w iili u i.it contra ejrtive use. N . 1'ngl J M i-il., IM i'G S O 6SS. l'IH 7 10 1he W l <C) ( o lLiliof.lllvi- Study III N eoplasia and Sli-iiil(l C 011IM 1 e p lives CiJiltir-li.it <van.m ,1111 er .m d <o nilnm -d oiat 1u ftll.lt r |itiv r s ln( | i|iu te n iiu l.. / S t8 S4S. 1*189. I I . W en s. N S.. I you. I I . K iis tifu n iiiflliy . S . Ilieli-M , S I . I ill. ( M . and (Ailing. ). k. N m if onir.i e ptive estrogen use .1111I llie 1 iiife u i e ol ovarian <.inei. |. N atl ( a i e i lns| . 68 4 r, 98 . 1987 12 Negri, t ., fla m e s * In. S , I/m u m . A.. Hi m ill, M . I. i V e d lu .i c . P .if.i//im , ( , Hei.il, V.. Hoyle. I'., and Tn< lio|Milos. I) Pooled analysis of I t u m p f.m 1 Hs*-cof Unit sludies I Ke| iff id live I.M in is .m d nsk ol fjn lh i-lia l ovation .iih er In i. |. ( am e i. 49. Si> Sl>. 1991 I t . f r.tiH esc hi, S . t iiV ei f Ilia. ( . H onlli. M . l/iin n u . A . N eg n I . P .ii,i//in i. f , trie Iio | m>u Ids, I and Rcial. V . Pooled analysis m I t imp.in l.ise t o illfi.l studies o l ovao.in 1 am e l. It Age at nien.il lie and .if m en m j m iim - Ini I ........... . -19 S7 M l. 1991 14 I I.IIH I-M In. S . I 'm . //III,. I . N eg n . I . tinIII. M . I .1 V e i lll.l. C Uf-f.il. V . I /i m oil. A . and 11m I io | lotilos. 1J P o o k -d .....lysis.I 1 I <i| x-.tii 1 a sf -i o n lm l sludies o l <-jiiliielial ovanan am er. Ill O ra l o n lia i e p live use Ini \ C an c el. 4 9 :6 1 6S. 1991 15. Gard n er. W . V H o rm o n al imlsalanc i-s in luninngem-si% c . im e iM e s . 8- 197-41 I. 1948. 16. C r.im ei. I) . W . and W e ll li, W K I V ien u m a n ls o l ov.10.1n .im i-i risk II Infi-rcnc es fi-garrlm g (Hilliognm-sis | N all. C .iim er lost . 7 l 717 7 2 1. 1981 17 (ra n te r. |> W . W i-liti. W k . hr oily. K t ., and W o |i 1. iiow ski. < A . O varian i .1111 er am t la i i . Cane er (Phila I. Sc |7 7 171. 1*187 18 C ra m e r. ( ) W ., H a rlo w B I . W itk -ll. W ( . W e ll h . W K . H ell. 1. A . Sr u tly . R. I . , N fl W . ( . a n d Kna|p. R < io ta lo s e 1 n its iiifip lio o a n d n ie la b u lis m in re la tio n 10 ih e risk o l o v a ria n c a m e i I a m e l. 2 C.(. 71. 1989. t BEST AVAILABLE COPY / M .ic iiiiIIrh Press I ul . I `7X`7 Dietary factors and epithelial ovarian cancer \ i . u ' O n S i m 1 V u 1 .1 ( i m i '. J in n M in Y u a n '. R .G . 7 ic g lc r ! & L A . U r in lo n 2 " s h .m v h tn < .in . > h i 'U lu li'. I h /'.it l i n r n i o l P fin h in h > io i t . 2 'IHi W i n R iu u l. S h a n f ilia l 2 (7 0 (7 .0 . i'f o p U - 's R e p u b l i c o f C h in a .n i./ I m i>'m u <ii.ll I f h h n im lm n H itiiiili. \ , U m ilili ( a t i n ' i I n s l i l n t r . I w c t i i t M V P l a z a N o r t h . R o o m 4 4 .1 . l i c l h c s i l a . \ m ;.%/ / s i Niiiiinan D k I.iiv data lim n .i population hasod case-control study ol 172 epithelial ovarian cancer cases .Hid I coniiols were analyse! A significant (/'< C M M l lose response relationship was found between intake I lat 11out .iinina! M im a s and i i 'k ol ovarian cancer, but plant lat was not associated. Although the clfecl ol .niiiii.il i.n was Loiitm iihled b ) education, an adiusted kids ralio o f I X pci sisled lor those in the upper t|it.iilik \<nip.iicl to the lower ipi.irtile ol consumption \ P for trend = l) O j) A lte r adjustment lo r anim al fat intake c .iIo m Ik and p io tem in take had inium i.il clfecls on risk T o ta l vegetables were found to be somewhat p io icvin c. bio the nievli.oiisiii o f action vv.is uucle.ir W eight, height and relative weight (weight,height ' ) were not tel.iled to risk ol iv.oi.oi e.llKCi S u b s ta n tia l c v u Ic ik c m d k . iie s ili.it d ie t is .t m .i |o i I n c l i t i in tlu ' i use o l s.iin the m ost in ip o ii.n it a m i p te v a le n i 1111111.11 t . i i u e i s i's p t v ia l lv c a n c e ls if t l u d ig e s t iv e l i n c i a m t ............ lpeol. io o k t is (W d ii.m is \ \ \ e i s b o i g e i. IbXf) \ 11 in w K v i n c i u >*i lie i.o v la i o t h e a e t io lo g y il o v a r i . i o u n ci li.o K e o s im '-v s Hy I In s*mc c p u lc m u 'h 'p ic a l s im ile s i \ i in si i <i i ,` A I 11 I ' l ' s . ( i . o r n i i l o / . P I S I R o s e e t a i . |* s |. i ,i \ e s i l i l i . i .il l't x ' i I v |v i lin e ili, il stu d ies b ave s ii.m i l I h a i hei.M V la t Is Ic l.llC il to C Ilih lg C Ih M ls llO IIIIO IIC Ic 'c ts p i. s k iin g a p la u s ib le me lu iiis o i lo i (h e :iss k la lio ti i \k ill 11 A M a M a h o n . |'J X 4 /*| -\ p o p u l a lio n - h a s e il ca.sc- lo n t to l s|olv in s h a n g h a i w in c h o b ta in e d e s te n s iv e ih c i.n y i n t a k e d a t a . v lle ic d t h e ><pp*i i n n i l y u> s i t n ly d ie c l l e d s id li. I.n v 1 .0 . e a l t ' i i . s a n d o d i c i l i u t i ic iit s o i l t h e n s k I >. I I i l l i . i l k C l M a lc lia ls and in elhods I he S li.m g h .il p o p iila tio ii-b .is e d e.ineet ic g is tix e n a b le d rapivi i>!*i 11111 a i i o i i o f i t \ . " i . u c .u n c i p a i ie o t s lo r t h is e a s e - c o i t l i o l i m is 11 1' 11111 u't I 11 i s io d v w e ie .il l w o in ,ii agct i s 7t > y c a t s -.U li . . i n . n i .i ir i n c w lv li.ig l 'si't l in lir e M u n g i m i m h . i i i a le a Iv n v ic n I N ip ie n ilv t I ' t s i ,o td h i J u n e l `W > N\nien w i i l i b o i d e i lin e i \ i ' t i i i i t o i i i s a m i u < i i 'p e i n i . i t i e i i i t o u l e n t s v eie e s lin l e d ........... s tu d y < I 2 \ S e l ig ib le c a s e s . 2 2 9 t S S 1 " . , | i i l . 111n*t a l t e r 21 i s I " Ic ic a s c il a n d e i g h t t * I " ..) t in t i n e e * a b 'c a se s w e it - A lo d c i l ( lin ts.. i I .i m i it is | > p a lh ili> g ic a l d a t a at d i.is'in 'sis. alo u t: w ith m lo im .ilio n m l te a I m et it a n d ir in il v i.- ic a b s i i .is t e l i i o i i i h o s p it a l i es o i Is N e a r ly a l l 1 i .1 o l th e . ' s w e i e h is io lo c ii .iH v e ' i i l i t m e i l . w i t h t h e i t a i - ' i m l i i I v i n c li i"nsel i l n o i i i 'l i c u b o u lt i .is o lili ! (1 I " ..) o i s la m a i e s a n i m a i n tu t2 <"..) \ to ta l o l 172 w o llte n i ' s i |t .. \ c i d i a i ' i i " s i `d w it h e p i t h e l i a l l n n i o o i s a i e t h e ' . lis o l l il is 11r . -it'.' il a t i n 1111 <1 V a . > elee te il h o n t t h e S h a n g h a i c e n c i a i p o p 'ila tio ii hv a -i m l.m l la m in ili p i cedui e to in a lili each is.. I o i v a i l i o m i n i i ne h t ' i i s i h o l i l c o m u n i ic e l a r e s id e n * '.d t in t' o l a p p i o v n n .itc lv 4.1 MNI ii n t i v u h i . i l s ) vv.is r .n u lo in l v - . t . . l h o i n t h e I J' 7 In lisciti Id c o m m it t e e s u t t h e S h a n g h a i m b h i m . i 11 >i>i vs In. 11 o n e I m i i s c l i i ' I d g i o n p t c o iis is lm g . .1 t *1** " il l l l . i l v i'. 1 N 2 " li. .it s e li o l ils ) u . sClgV t i l l S t ilt - 'V . |ll Itti t w o U .>111' il w i i l m ) l i v e v e a i X O l a g c o l l it e n u ll's .is w e ie i an l in i'. '1 tel t i o m a ll d i g M e w t u u e i i O n e s 1 '. c I .' s i l k l l l l b . l l vO l i li .1 a i u l t h e th e . an a lle n ta le I U I W M b .i I n a IV " I ,i b d . i l c i . i l . t. p h i o c k i o m y w e r e . p ia v .l V i l i . d i. i o. n x M l '`li' t i 'i l . m i l I <a o m e n a g t e e d ! .. p II II l| .ile I n b ' i t n .ib o n vv IS o lle t e d t in m i i ' l i la c c io - la . ' H it. ' -. 1 Vv n a m 1 m ic i v ie WCl ' 1 Ite si.n i|ail t 1, | - d . \ i ... i i.. M m K . v ..1 1 tiii. |. ,v - ..m l o i VuIsMe-d.I !t.o. ... i l ' \ u cus I'JXS questionnaire eovereil dem ographie characteristic:. repro ductive hisbirv. medical history, fam ilia l cancer history. |v is h m I habits, oeetipation and diet. Inloim a hon on UMial adult ei>iiM iiiipiion o f M coim iion looils in Shanghai was obtained. Study .subjects were lust askial about hove often they ale each foul (daily, weekly, m onthly, yearly, seldom or ncvci). followed by questions to dense lhe giants ol food eaten pet unit lime. The women. wh> gcnci.illy were tespoiisible hr buying and preparing ihc meals lo r their households, adpisied the v|itanlities purchased by the Iraelioti they consumed. The food com position table published by the Chinese Academy o f Medical Sciences ( I %>Xt ) was employed to con ve il these loods in to nutrients. The m ajority o f nutrient values were based on data derived horn the Shanghai atea: when this in form ation was missing, values fio m Jiangsu province, and occasionally from llcijiug. weie utilised Data were not available on saturated and iuis.it ur.ited la I in Chinese foods si it was not possible to examine these tw o variables M u ltip ly in g the reported daily consumption (in giams) ol each in iliv id n n l food by the m iln c n i content pci gium m that food, and then summing vi all loods. gcnciulcd lo i each in dividual the total daily ingestion o f each nutrient. In add ition. lol groups were burned based on dietary o r botanical sim ilatities. l oi ex ample. meats included pork, p o ik hops, spareribs. pigs' feet. s.ih*l pork, pork livei. ugati meats, heel. lamb, ehiekcu anti duck: the enietlerotis vegetables in'ltided greens, cabbage. Chinese cabbage and c atilillow er: and almm consisted ol 1Is m the nion latmly (see tab le V | lo t further details) I lie !Is la tio (OR . estimate! tcl.itive tisk) was employed in m e.isiinug the association Ivtw eett diet and ovarian eaiieet Itaseil upon (lie d is trib u tio n among the controls. i|ii.i11ill* m is weie used to ctc.tic eategorieal vaiiahles. and tlu lowest 25".. was chseu as the teleieni gionp. T o assess ami conn! lot sutuces ol eon louiuling. analyses utilised e o m litim i.il logistic icgiesstmi techniques (l.u b in . 1*7X1). I tend tests weie pci Itinned In beating ealegone.il vaiiahles as i outiiuniis in (he models Results t ases am i m m o ls w e ie w ell m .itch ed o n age. w ith the mean ag e b e in g 4X *7 ve is lo i cases a n d 4.x x y c a is lm c o n tro ls < ases te n d e d to I v b e tte r c ilu e a te d a n d h a v e h ig h e r averag e lu'U scholtl tiiio in e s th an e o n tio ls ( t a b l e I): hovvexcr. the e lic its ol iiiiiim e al lin e e sepaiale tim e iv iitu ls (ptesenl and a p p io x u n a ie lv H i a n d ?D years I v l m e d ia g n o s is ) c o tilil he e x p la in e d bv th e 'Heels ! c lu c a iim i. C ases te n d e d m m e lien ib .n i e o iu io ls (i be n u llip a io u s .m il to be sm<*kers. the m e a n n u m b e r o l live b u lb s w as 2 ( lt eases a n d 2*7 lor m ittid s ( ases m m e lic n ie n il\ tc p m ie d histories ol ovarian D I I T A R V I AC TORS v. I able I D istrib u tio n ol selected demographic charneleristics and estimates nor the trend test was statistically sigftillcani. risk faelors C arbohydrate intake, on the other hand, was associated w ith Ape (%) s' 29 * Cast's in 5 ( 'nutnils 110 a non significant reduction in risk. The potential confounding effects o f other risk factors on nutrient intake were considered, including education, income, number o f live births, ovarian cysts. sm oking, o ra l co n tra 411 4*) Ml 5') >60 174 174 ceptive and medroxyprogesterone use. lu h o s le rilis n iio i'r aiuT 29 7 32.6 I I I I ) usage. It was found that education substantially altered 25 21X a number o f effects. A lte r adjustment fo r education."trends l ilucjlion ("..) College Senior high school Junior high school Primary school No formal education in risk across quarlilcs o f consum ption remained o nly for 14.5 2.4* total fu l (/* = 0.0.1), anim al fat </* = 0.07) and anim al calorics 21) 4 14.5 ( / ' = 0.0b). Thus, those in the highest quarlilcs o f intake 24.1 33.1 showed ORs o f 17 (95% C l = 1 .0 - 3 2) and 15 (95% IX.6 23 3 0 = 0 . 9 2 9) compared to those in the lowest q ua rlilcs o f 17 4 26.2 anim al fal and animal calorics intake, respectively, fu rth e r Incnnw/month,'capita in 14X2 oncan Yuan) adjustment for income o r other risk factors did not affect 46.5 41) 2b either these point estimates or the trends in risk w ith these Nulliparity (%) Number of live births (mean) Ovanan cysts (%) Smoker (%) 1uN.sterilisation (%) Oral contraceptive usage ("-) FUI) usage <%> Mvdiovypiogeslcrone usage 20.9 2.U 12 2* 2.9- nutrients. The trend w ith total protein became non-signi ficant ( / ' = (). 12) after adjustment fo r education. In addition, 99 1.2" the association w ith carbohydrate intake disappeared after II 0 7.6 adjustment. 12.2 23.3* f.lTccls o f foods grouped hy dietary and botanical s im ila r 1)4 n r ity are shown in fable I II . H igh intakes o f meat, red meat, 7.0 15.7* and dairy products and eggs were associated w ith elevated 14.0 3.5* risks. In contrast, intake o f vegetables and legumes apivared Comparisons o f cases to controls hy i lest: */' ' 0 ItS; */<() 01 to reduce risk, although neither trend was statistically signi ficant. Selected subgroups o f vegetables, e.g. yellow-orange, dark-green and cruciferous vegetables, were not associated cists .iiul usage o l` medroxyprogesterone or oral contracep w ith reduced risk. Again, education exerted m ajor c o n fo u n d tives (although lltc excess was exclusively lo r shori-lcrm use ing influences, w ith associations lo r meal, red meal and I o f the p ill), while controls had higher rates o f tubo- dairy product intake no longer remaining significant after I sterilisation and use o f intrauterine devices (IU I)s ). appropriate adjustment. However, the decreasing risks w ith The rclatiottship between various nutrients and risk o f vegetable and legume intake persisted after adjustm ent fo r I ovarian cancer is presented in Table II High protein and fat education and animal fal. although the trend tests were not ; intake were significantly related to an increased risk o f significant. 1 ovarian cancer. Compared to the lowest quartilc. the highest Since anim al fa l and total calories were correlated, and quarlilcs were associated w ith O Rs o f 1.7 (95% C l - 0 . 9 3.4) both appeared to increase risk, attem pts were made to j and 2.3 (95% CT = 12 4.4) lo r protein and fat. respectively. id entify the m ajor determ inant. Hy cross-tabulating these tw o I I here was a remarkable difference between the effects o f fat variables, it appeared that only fat intake exerted an effect ' from plant and anim al sources, w ith the ORs o f I lie highest on risk. The ORs o f anim al fal adjusted for calories and i|ti;irtile being O.X (95" C T =0.4 1.4) fo r plant fat and 2.2 education were 1.4. 1.9 and 1.7 for the second, th ird and (95% 0 = 1.2 4.2) fo r anim al fal. ORs increased w ith high fo u rth quarlilcs. while the corresponding O K s o f calories animal protein, w ith plant protein having no apparent effect. adjusted lo r anim al la t and education were l.l), 1.2 and O.K. APhough those in the third q u n rlilc o f c a lo rific intake were Attem pts to disentangle the ell'ecis o f anim al fa l and anim al at increased risk, the trend in risk across all quartiles was calorics were unsuccessful because o f the high corre la tio n o f not statistically significant. However, there was a significant these measures (Spearman's correlation cocfticicni. r = 0.97). trend ( / ' < 0 0 | ) in risk when calorics from animal food alone Sim ilar analyses to disentangle the efleels o f to ta l protein i were considered. H igh rib o fla vin consum ption was also and anim al fat also revealed that anim al fal intake was m ore (elated to a slightly increased risk, hut neither the point im portant than protein intake. A fte r adjustment fo r anim al Table II Risk n f o varian cancer hy selected nuirienis ( nut, OR OR tiiliu.xlcd for film mini! c, V : /.vr f o r Ifa id 0 . / ' m ine V, V; Protein P la n t Annual 1 at Plant A n im al Calorics P la n t A n im al Carbohydrate C rude libre Vitamin A Carotene Ascorbic acid t< T llu a m in (11,1 Rihollavin ( I t ,) Retinol Ml I I IX 1.7* (1.03 1 0 OX O.K U.9 OKU 1.0 0 4 1 7 1 6 0.03 1.0 0.9 1 0 0.9 2.0* 2 3* < 0.01 I.U O.K 0.51 1 0 1.4 2.1* 2_2* < 0 01 1.0 1.0 1.0 OX 16 OX 1.3 0.7 0.29 0.22 1.0 1 0 2.1* 2.2` < 0 .0 1 1 0 0.5' u x 0.5* 0.04 11) 1.5 12 II 0.7X 1 II l.l OX 1 0 0 66 10 1 1 04 10 0 70 1 0 0 9 0.7 0 9 0 54 1 0 0.7 09 11 0.59 1 0 11.9 1.4 1.4 0.14 M l 1 2 1.2 1 4 0 36 1.0 1.0 1 X 1.4 1 0 0 4 l.l 12 1.0 o.x 1 5 12 1.0 l . l 1 X 1.9 1.0 0.4 1 1I O.X Ml i 4 19 17 1.0 1.2 1 6 12 M l M l 1.2 0.9 1.0 0.4 1.9* 1.5 1 0 U 6 I.U U6 M l 14 14 11 1.0 1 0 U.X 0 9 1 0 1 3 1.0 11 1.0 1 0 0 7 0.4 1.0 OX 1.2 1.3 1 0 1.0 l 5 1.0 1.0 12 1 0 1.0 ( ) , . lowest 25% : ( ) , . low er middle 25-;. : Q ,. highcir m iddle (J , , highest 25% ; * / ' < 0.U5: *7' < 0 01 7*0/ for Ira n i l ` voltI f 0.1 2 0 4X 0 36 0 03 o .5 x 0 07 0 40 0.99 006 0 39 0 91 0 71 0.97 0.5X 0 30 0 64 0.77 '<4 M U ' I I I M i l . I 1able 111 Risk o f ovarian earner bv selcilicit food groups ( 'nule OK OK </;%Irti /tir flint titilli1 / i ll for 7in/ for fremi tremi Q, (7, 3, Qx I' nilur Mv.ii 1u u X 1 t 1a II us Uni ....... 111 (19 i: 1 x 101 .......... 1II l ' MX 1 < 0 62 1. )i 1(1 Il / li / i J (l XI li m i .niil -i 1II 10 1l 14 Il 1) 1fifs 1II 10 1^ 1 1 0 39 Vi.vl.ihlc 11) (1 7 0 s OX 0 45 M.uk gn-vii Vegetables 1II 1 1 1(1 1 1 0 7<i Vi llmv >'l il ilv vegetables 1u i : 11 n 5S l 1Ikllll-'lls vegetable 1II II x 1o 10 (1XI Villini) 1II 1II 16 1 1 u x\ 1li-iinus 1II 119 II >' tl.K 0 19 10 l) 6 0 9 1 3 o 33 ` . ......... 11Sill .ill's 10 (1 V n x o 6 0 16 Vegetable ,'iK 1o 1 1 u II 0 9 0 52 10 0 6 l.l 1 2 0.30 10 0 X 10 14 (1 19 1.1 119 0 X 1 1 0 7X 1II II 7 II 7 Il 9 0 70 in 1o 1 1 U 4 0 9X 1.0 1.0 0 9 l.l 0 62 10 0.7 0 5 o x 045 1.0 1.2 1 1 1 i 0.53 10 1 1 10 0.92 1.0 I.U l.l 12 0.55 10 l. l 0 6 14 0 54 10 o x 0 4* OK 0 22 ft) 0 4 06 0 9 0.68 10 0.6 10 U.7 0.61 1.0 I I 0.0 0.9 n.5x <,, |,llll-sl 7'"'.. 1).. lower middle ... <3,. higher umilile Q.. highcsl 25%; W 0 05 1ahi IV Risk l.l .IV anali cancer by relative weighl ( Vo.-S ( iiiitit'h ( rutlr OK A tinnirti OK Rei. K ill- MClglil IM l'Iglll lll'lg lll'l b s ii i<3,> Is s? 2'i X? u y . i ' s i : : n 03,1 \" 0 IU . IS 41 1 0 I.U 56 41 1.6 2.1 is 42 1.0 1.2 46 44 1.3 1.6 I >U is .u l|u s k \i li vii iK til loll .im i .iiinn.il Ini intake. O , l.nxoi i ) ; . I.m c r m iddle ?Stn. Q t. higher m iddle 2 5 V . 0 . highest 25% 95% a | (nr titl/usU'il O R) 1.0 4 2 06 26 OX 3.3 I,ii ;im l e d u c a tio n . th e O R s asso ciated u t ili p ro te in in ta k e w c ic re d u c e d t o (i I 4 a n d 1.0 lo r the second. th ird and i o n i i l i ip i. u i lie s <*l c o n s u m p t i o n I lie < ) R s l o r p r o g r e s s iv e q u a litie s o l a n im a l f.il m l.ik e a lle i ad tiis lm e n l lo r p io le iu m l.ik e a n d e d u c a li* m i w e re 1.4. I 7 a n d 1.4 I iW e I V pi.exei.its th e a s s o e ia tio n b e tw e e n e la tiv e w e ig h t (w e ig h t lis ie h i l a n d o v .-u ia n Carteer N e ith e r ilie O R s fo r th os e w ith n u le .i'. il h n .lv m ass in n th e (te n d tests w e re M e in lii.iiii A d ju stm en t Id i ed u e a iio n an d a n im a l la i in tak e d id not s u b s ta n tia l!) a lle i th e ic la lim is ln p o l n sk to relativ e w e ig h t O t lie i a m i l i o p o m c iile m e a s u re s , s iie h as lic i|*lit. .iv e ia iie a d u lt vveieln . m a x im u m a d u lt w e ig h t o r w e ig h t/ h e i g h t 1 * vi e i e u n i e l a i e d t o ris k . O n ly i*ne ease a n d no c o n tro ls le p o ile d re g u la r a lc o h o l s o iis iu n p iio n . Iin iiim e th e a b ility to Im ih c r assess th is e x p o - sine as an a V b o ln g u al la it o r I tisinssio n liile in .iiio n .il eo m p .u iso n s m d k jte (h a i o v a iia n can cer in c i d ern e i.ite s c o n e late W illi |v i e a p n a la i a v a ila b ility (A r m sir. ne A D o ll. M ' s. R o se e t t i l . 19 X 0 ) fh e in c rea se d iik id e iis e o l o s a iia n e a iu e i a m o n g la p .m ese m ig ra n ts lo the I S \ has been m w c .l as lu ilh e i s n p p o it lo t an a c tio ln g ic a l io le .a d ie t.u v ).ii illa e iis /e l A K m d ia la . I9 6 K ) H o w e v e r, s ik h d c s iiip iiv e -in d ie s . v m icc i n ed w ith to ta l p o p u la iio n >hai ai lei i .1k s i.Kin i ili.m illusi- o| individuals. d<> not I 'l'ii ii n |11 poli-m 1.11 i o n |o n iid e is s m li .is p.intv and socioei on, nine si.i his I o llo u -n p studies p io v a le so m e s u p p o rt lo r a n a s s o c ia tio n ol . >v,n i.iii i .mcei nsk w ill) lai mi. ike. although the icsiilts -lie lin i 11Misivicia! \ lo llo u up siuilv id Seventh Dav \dv emisi in.un '1 w h>mii an- .V.. ILiei' vceciaii.ms. showed .llld.lldlscd m Mluli.lV ia ti.. I'M "V ai ian eaikei ! about 0 (< li'lllp .llC tl In ib i- genci.il ! a lii" i ma population lP hillips i ll . in s ili R,;p,n im e the piclim in .ii v lesiilts lo i a 20-yc;tr follow -up study am ong I6,I9(J white Seventh D ay A dven tists. Snowdon (19X5) found that women who consumed high am ounts o f eggs or fried food wore al a three-fold eveess risk. It was suggested that the use o f fat in the process o f liv in g , especially animal fat. was more im portant than the eggs (Rose & Boyar, 19X5). However, M orm ons in Utah showed a standardised incidence ratio o f 1.7 fo r ovarian cancer compared to the (IS population, although their ilic i is not unusually low in anim al lut ( l.)o n ct a!.. 19X0). Kinlcn (19x2) also lound no reduction in risk among nuns who cither com pletely or partially abstained from meat as com pared to other nulliparous women. C ase-control studies arc somewhat more provocative. In one case-control study in the United States, cases were found to consume more whole m ilk, butler, animal fat and satur ated fat and less skimmed m ilk , m argarine, vegetable fat and m isaluratcd fat (Cramer at ut.. 19X4). Supporting these findings. I .a Vecchia ci ni. (I9X.7) found significantly elevated risks among Italian women vsho reported frequent con sum ption ol meat, ham and lals. especially butter. In add ition, low risks were associated w ith consum ption o f fish, giecu vegetables, cario ls and wholemeal bread or pasta. In another US .studs no elcei ol fat was found but a protective e lic it o f vitam in A intake among women aged 30 49 years was noted (Myers r/ < il. 19X3) However, all o f Ihc previous studies included only a lim ited n in n iv i of food items, and no data about p ortion si/e were available Ihc present study is the lifs i to obtain sufficiently detailed dietary data to allow for an assessment o f ovarian cancer iixk in relation to such nutrients as fat. protein, calm ics. vitam ins, etc Ih c broad range o f nutrient intake was an obvious asset (see I able V II fo r details), l or example, between the 25lh and 75th |ci ceiltiles ol intake there were appioxim aiely HMl. 3lK> and 22*". increases for total fat. animal lat and ascorbic acid, respectively. Ih c study in d i cated that high anim al fat intake was significantly related to the risk o f ovarian cancer, an effect that was not found for I ) l i : I A K Y I A i T O R S 95 fal from pLin( sources. This effect was nor explained by calorific intake, relative weight or non-dietary risk factors although adjustment tor education resulted in a d im in u tio n o f the observed risks. This could represent true confounding by true lifestyle risk factors or a systematic overadjustment. Thc-cITcci-of-calories.. which._on]y_exisLcd_lor calorics front animal sources, appeared to be explained by high anim al fat intake. In addition, total protein intake was not related to risk o f ovarian cancer alter adjustment fo r education anil animal fat The mechanism whereby anim al fa l m ight increase the risk o f ovarian cancer is not clear. A n cflcct o f diet through a hormonal mechanism is consistent w ith findings o f oestrogen receptors in epithelial ovarian tum ours (F rib e rg el a t, 1978; Moll c l at., 1979; Cialli el at.. 19X1). It has been suggested that a diet high in anim al fats can produce cxtragcnital oestrogen via gut bacteria ( H ill el at.. 1971) and that oestrogen hioavailability is altered in vegetarian women (Arm strong el at., 19X1, G o ld in <7 a t . 19X1). On the other hand, it has been suggested that diet may operate directly, since anim al fal could contain carcinogenic contam inants, such as polycyclic hydrocarbons, which arc recognised carci nogens o f the ovary in certain anim al species (Cram er et at.. 1984). A n alteration o f the im m une response o f the host by dietary factors is yet another plausible mechanism (C arroll. 19X1). O ur study did not show a protective effect o f vitam in A. front either plant (carotene) o r anim al (re tin o l) sources, and failed to support Byers's previous finding (Byers c l at., 1983) However, a slight protective effect was observed w ith high intake o f to ta l vegetables, an effect consistent w ith that reported by La Vccchia el at. (1987). The mechanism o f a possible protective effect is not clear, although vitam in C has been suggested to be involved in m aintaining the in teg rity o f the in tercellular m atrix, enhancing the immune response, prom oting tum our encapsulation and preventing oxidative degradation (W ille tt & M ac M ahon. l9X4/>) Nevertheless, the possibility that our observed association arose solely by chance cannot be eliminated Interpretations other than causal relationships should be considered because o f the lim itations o f case-control studies in assessing the effects o f diet. However, results w ith food frequency questionnaires have been found to be reproducible and correlate w ith intake determined by m ore detailed dietary methods (Block. 1982; W illett & M acM ahon. I984</). In addition, the estimate o f consum ption am ong o u r controls for most nutrients appeared quite reasonable when compared to data from a representative Shanghai pop ulation (5-day weighed foiKi records from 2,000 people, conducted in Shanghai, 1981) (Table V ) We attem pted to m inim ise recall bias by asking about usual adult diet; however, i f we were in fact obtaining in fo rm a tio n on diet affected by disease status, it would be djfTicult jo know -how such a bias w ould operate. In addition, hoth the interviewer and study subjects were" generally unaware o f the hypotheses about diet and ovarian o n c e r. A lth ough n u tritio n a l status m ight have influenced the survival lim e o f cases, this should not have a llix ic d our results since only 8.1% o f eligible cases were excluded because o f death. Finally, o f concern was the possibility o f residual confounding, particularly whether the anim at fat association merely reflected as yet undeterm ined lifestyle patterns. The fact (hat the association persisted after adjust ment fo r education and income lent support to a true effect, but some caution in interpretation is in order. The calculation from this population-based study o f an attributable risk fo r animal fal intake (adjusted fo r education and ascorbic acid intake) indicates that, i f real, the re la tio n ship could explain as much as 34% o f the incidence o f ovarian cancer in China. High fat intake thus m ight partially explain the different incidence o f ovarian cancer between women in China and Western countries, as well as the increasing incidence among Chinese im m igrants to Am erica (Waterhouse c l at.. 1982), although the findings arc not conclusive. Further well-designed studies o f diet on ovarian cancer risk arc obviously warranted. Table V Comparison o f mean intake o f nutrients between controls in the present study and m representative sample o f Shanghai residents Conirotx o f present .<fifth- Sample o f Shun#hoi resilientf Calories (kcal) Protein (g) Fal (g) Carbohydrate (g) Crude fibre (gl R clinol (RF.) C arotene (RF.) Thiam in (mg) Riboflavin (mg) Ascorbic acid (m g) 2.365 2 70.1 6X.I 367.9 4.1 129 5 542.7 1.0 0.7 1214 2.437.0 74 9 65.5 3X2.3 6.7 156 3 633.3 2.1 1.0 117 9 ` Based on 5-day food records in volving 7.959 person days, conducted in September 19X2 in Shanghai. Hoth sexes com bined. Tftblc V I Food items included in questionnaire 1 Rice 2. NomJIcs 3. Steamed buns 4 Pork (ta t and lean) 5. Pork (l.it only) 6 Poik (lean only) 7. Pork chops X. Pork sparcribs 9 Pigs1 feel 10. Sa hcil pork I I . Pork liver 12 Other pork organ meals I* Beef 14. I.m iih 15 Chicken 16 Duck 17 Pork and cereal sausage* IX Fggs 19. Vegetable oil (rapcsccd. soya bean, sesai ne) 20. Lard 21 Fresh fish 22. I resh shrimp 23. Fresh crab 24 Y ellow eel 25 Sailed fish and dried fish 26. Cow s' milk 27 Soya bean milk 2X. Powdered cows' milk 29. Cakes, biscuits, pastries 30. Candy* 31. Sugar* 32. lee m ilk bars* 33. lee cream* 34. Soya hcan 35. Textured Miya products 36. W heal gluten 37. D iie d beans o r |>cas 38. Peanuts 19. A ll vegetables* 40. Greens 41. Spinach 42. Chinese chives 43. Cabbage 44 Chinese cabbage 45 C au liflo w er 46 Celerv 47 Bcaiisprouls 4X. Aubergine 49 W ild rice stem 50. Snow peas 51 String beans, asparagus beans 52. Lettuce 53. Chinese w jx g o u rd 54. Cucumbers 55. C arrots 56. W h ite radish 57 Mushrooms 5X. Sweet green, teil p cp|vrs 59. Tom atoes 60. Apples 61. Pea rs 62. Oranges, tangerines 63. W atermelon *N o l included in analysis (items 17. 29. It), 31. 12 and 33 were rarely eaten, and item 39 was repetitive). Food items included in food groups: m eal. 4 16: red m eal. 4 I t . 13 14; poultry, 15 16: fish. 21 25; dairy and eggs. IX. 26. 2X; eggs. IX. vegetables. 41 46. 4X 5K; dark gicen vegetables. 40 42. yellow orange vegetables, 55; cruciferous vegetables. 40. 4 * 45: allium . 42: legumes. 34. 15. 37. 3K. 47. 50. 51; fu n is. 60 63. complex carh o lu d i.ilc s . I t. 36. vegetable oils. 19. \ I \ I M I I S. 1alile S II 2*lh alni 7Mli pen entiles tgdav ') lor daily lilia le o f selci lerci nuirienis and food groups among controls \ it/rii'/il s 75% Food groups 25% 75% l*ii)fi*in Plain protein 49 0 3(> X XI 1 52.0 Meat Red meal 364 29.9 106 4 K3.9 1 al Pialli l.ll A nim al lai ( ..UlCs Plani calm ics A iiiiii.iI i.dories ( .if bnlivdl.llc ( `link- libre Vii.im m A (Rl-'l C.u.'lcne (R l ) Kelim >1 ( R I ) 1In .m ini r H | l (m g) R ib o llam i ( It .) (mg) W inhii acni (mg) 39 4 237 13 0 1.901 6 1.6179 1*X 3 313 X 25 349 4 2X3.3 4X 0X 04 6X.2 77 0 3X.I 46.6 2.674 9 2.155 5 529 5 416 5 5.0 X63 5 65(1.0 20X 4 11 09 144.X lis li D airy and eggs I S Vegetables Dark-green vegetables Yellow-orange vegetables Cruciferous vegetables Fruits Com plex carbohydrales Vegetable oils Legumes Album 15.1 8.2 X.2 241.9 70.2 00 9X.4 22.8 363.1 13 2 46 3 0.0 SX 4 55 9 32 9 528 0 174.2 0.5 234.0 164.4 513 2 24.7 95 8 2.2 R ffm iK 'o \ K \ l M I < i V . U K IIK O W N J I) . t I \ H k I . ( I i A 4 o th e rs ( I 9X I ). Du-I j ik I ic p io d ih liv c horm ones A study c l' vegetarian and iic iiM 'iV l.iii.n i 'liiu iin p .iic .il w o m en J.N < 7 6 7 . ?fi| \K M M K n M i. H a n o i l . H (I9 7 A ) f n v iiiin m v n l.il factors and i . llliC I lllv l.tc ilii -MI.I lllo l|.i|l|V III t h llc f c lll lo lllllllc 's . SI n il \ | K l l.ll i i l i ' i o i i i ! d ii'I i i i p i . n i i t i - In i / i IS . <d 7 I II i a I. I . . l l `JK.1 \ liM l'U 4)1 v a lid a tio n s i ll illi' l.iT ) assessment n iv lh .u K m J //;./, 1/../ 115.492 H I I K-, I \ I \ H M I \ I I I t . H M I A M S . V II I I I IN ( A: S W A N S O N . \ l t l ' i N i ) \ . , i x i . ' i i i c i l siu d s <il iln .-l.li\ a m i n o m h c ia iy fa c to rs in is.11i.in i .ills ci /. 7 1 . i*XI \n K O M k k l l 'i v l i N in ii.il l.il\ am i y a m c i C m in 'i Res. 41. ii. -l-i I M IN I s i \< \ l > l \ | \ " I M I D I ! \ l S( I I NC I S ( I9 S I | h .iiu i C o m p . f s n.-ii I i IV * 'p ic i H e a lth Publishing ( c lli*ijiii|* I II W I I H OVA M i l l I I w. k . I l l ' l l U IN S ll N . i ; I I . u i l l I I I . W A s i I M i< l l l ' v i| D ie l .iis .ifiifii.il i.ii in ii* l.iin mi i c o v a ria n m itcei t iA ()/ /,/ (rM tiii'l. 6.). S O I k l l l l k i i 1 1 . k i l l \ M H K. S 1*1 K S K .N I I ' A K O R s l l N . H ( l ' ` S | O n i c i c p l . n \ In r .. O i.> |V ii\ <I .) ,n u l .n u ll o g cils <I >| I T ) ill jn m i.n i i*iid i< nii 11i.il .u e m -u ii.i a m i n s .m .m Illin o is .1 in l i / n / c l c. i 5?. 2M t . M I I M n . i . l i i W S S I i 1)1 . N il O i l I I I . ( i \ ts o ilic rs (1 9 X 1 ) M k m i 11i i t v i * o l m u ltip le s te ro id h o iiii.> iic lc v c p to is in dtscaseirc c .n ni ricopia ' i n In u iia ii vai) ( o n . i t . 4 7 . 1297 .n i d i n If k \ l d i m i d i n / . I I i m ) I k . I 1 . S W | I N S O N . 1. . w \ l< |( v v i M l ,V V.1 )K II At I l M (1 9 X 1 ) J lie v i o f d ie t o n e vi io n .11 o l i - m .< |-e ri' ill p ie .n id p i.N iiiic in ip a iiN .il w o m e n t ,4>l, . 1 /V. v 41 . * 7 T1 If \ l N S / I 1 w .v k 1 K IM \K A M r 1`k iX i S tu d ie s o l Jap an ese u n c i .m In 1 M >*11.1lit V I ll - I ll ca n ce l a n d o d ic i tlise;iNV\ a m o n g 1 ip .m e n,- in ih e t m ie d S ia le n , / \ ( /. 411. 4 3 I H M M i (.' > \i< I > l* .v W l l l l W l s K i n e I **711 ( i n i h a c i c i i a .ill.l .n il- li'i'S i .1I I I i"l o f i li o liro.isl ii/rli . I ii. 172 l i o n . J A . ( A n n o . T A .. K L I l.v . K M . (ik liC N W A l.D . P A C IlO R O S f . S. (1 9 7 9 ). L s lro g o n a n d p ro g e s tin b in d in g in cytoso ls o f o v a r ia n a d c n iv a r o n o n iu s . O h u e i. C vnecol., S i , SO. k lN I I N. I J (19X2). M e a l and la i co n su nip lio n and cancer m orlul- i i y A study o f siricl religious orders in llrila m . /a in rv /. i. 946. I \ VI ( ( IM A . ( . D H A R I I. a . N i c k l . I. A 5 ill hers (19 X 7). D ieta ry facto rs an d the risk c l e p ith e lia l o v a ria n can cer. J N C t. 79 , 663 I i'H iN . J (19X1). A co m p u ter p ro g ram lo r lire analysis o f m atched ease-control siudics. Com put. R w ineil. Res . 14. U K . I ) O N . j I . ( i A R D N I k . J W . A W l- S f . I ) W (19X11). Cancer risk and life-style: Cancer am ong M orm ons from 19 67 19 7 S. In C w u c r / .ii/i'/m 1 in D c lh ia f Populations. Cairns. A .. I.y o n , J.I.. Jk SloliiicL. M . (ed.s) p. 9 3 . Ila iih u ry Report N o . 4. C old Spring Harbor i.aboralory. IM III I IP S. R I. . ( i A K I - I N K I i l . . 1... K U Z M A . I W . H lil S O N . W .L .. L O T /-. I A B R IN , H (19X0). M o rta lity am ong C alifo rn ia Seventh-M ay A d v e n tis ts lo r selected cancer sites. J N C I. 6 5 , 1097. R d S li. 1)1 A M O Y A k . A I* <19 X5 ). D iet am i o v arian cancer. J A M A . 254. 2553. R o S I , D P .. H O Y A K . A .I* A W Y N D L 'K . 1: L (1 9 X 6 ). International comparisons o f m ortality rales for cancer o f the breast, ovary, prostate and colon and per m p iin loud consum ption. C ancer. 58. 2363 Sn o w d o n . I ) A (19X 5). D iet and o va ria n cancer. JA M A . 2 M , 356. W V I I k M O l'S l . J . M l 'IK . C . S IIA N M C U A K a T N A M . K A P O W I:| I.. J < |9X 2). C om er Im n h in e in liv e C niitnnnis. V o lu m e IV . IA R C , Scientific P u b licatio n N o 42. I.y o n w i t I 1-11. W ( ' & M a< M A I IO N . H ( |9 X 4 n ). D ie t a n d c a n ce l ov v i v ie w ( liis t o f iw i i p a i ls ) <V I'.ngl J M e t! . 3 1 0 , 6 3 3 . w i l l l . l l . w.c & M a> M A I I O N . I) (1 9 X 4 6 ). D ie t a n d c a n c e r o i ci view (second o f tw o parts) N hunt. J. M ed.. 31 0, 697. ail an veil l l AMS. G M A W t lS IH 'k id R. J l l (19X 6). F o o d and cancer C au s e an d cflc clV S nrg C lin N o rth b n . 6 6 , X7 3. Kl J ( <111, er ( 59, 97 99 io The M acm illan Press LM .. 19N9 Zoladex: Endocrine and therapeutic effects in post-menopausal breast cancer A L . H a rris ', J. C a rm ic h a e l1, B .M .J. C a n lw c ll' & M . D o w s c ll2 'University Department o f C lin ica l Oncology, Regional Radiotherapy Centre. Newcastle General Hospital, Newcastle upon Tyne NF.4 hlfF. and 1Endocrine Department. Chelsea Hospital fo r Women. Dorehousc Street. London SW J 6 L T , UK. Summary T h e endi*crinc and therapeutic cITccls o f the L I I R I l agonist Zoladcx have been assessed in 28 post-m enopausal women w ith advanced hreasl cancer, f ourteen had responded to previous horm one therapy and 14 had no previous hormone therapy. There were tw o partial responses and tw o patients w ith stable disease for m ore than 6 months in the form er group, and one partial response and tw o w ith stable disease for m ore than 6 months in the latter group. To xicity was m inim al. A ll responses occurred in soft tissue. Six out o f seven patients w ho received tam oxifen after progression o f disease on Zoladcx showed a response. Peripheral estradiol levels were measured, and they fell after I m onth from 33 pm ol I 1 ( 2 0 . s.d.) to 2 2 p n io ll 1 ( M , s.d.) ( / * < 0.U05). Responders and non-responders showed sim ilar changes in estradiol. O cstronc levels did not change significantly. These results suggest that Zoladcx acts indirectly via changes in peripheral hormones, rather than directly on L I I R I l receptors on the tum our. Recently lu lcinising horm one releasing horm one (L H R N ) agonists have been shown to have direct in h ib ito ry cITccls on human breast cancer cell lines in vitro and low ulTinity binding sites were demonstrated (M ille r et a t . 19X5; Blankenslcin et n l. 19X5). L I I R I l binding sites have also been demonstrated in p rim ary breast tum ours (Eidnc el al., 1985). Therefore, L I I R I l agonists such us Zoladcx (Cioserelin) have been assessed lo r potential efficacy in post menopausal breast cancer, on the assumption that a therapeutic effect may indicate a direct antitu m o u r effect (Plowman et a t. 19X6; W axin nii et al.. 19X5). whereas in pre menopausal women the m ajor effect would be via ovarian suppression (N icholson el a l.. 19X5) However, other endocrine effects may iHXiir in post-menopausal women. In the post menopausal wom an, adrenal and ovarian androgens arc the m ain source o f estrogens (Judd et al., 1982, 1974; G ru d in et al.. 1973). They arc converted peripherally hy aromutasc to ocstronc and ocstradiol. Inhibitors o f aromalasc (e g n m ino glutcih im idc) arc effective endocrine therapies in the treatment o f post menopausal breast cancer (H a rris et al., I9X3<) and occasionally pre menopausal breast cancer (D c/w oda et al., 19X7, W ander et al.. 19X6). In the latter case, it has been suggested that intratum our conversion o f androgens to oestrogens may be im portant and direct in h ib itio n may be im portant, w ithout any detectable lo wering o f peripheral oestrogens (M ille r et al.. 19X2; lic/w o d a el al., 19X7). Hydrocortisone alone can suppress adrenal androgen production (H arris el al., 19X4) and hence lower peripheral oestrogen levels (H arris et al., 19X4). We have recently shown that the androgens produced by the post-menopausal ovary arc under pitu ita ry I SH and I II control (Dow sctl et at. J9XX) Zoladcx may therefore have indirect endocrine effects on post-menopausal breast cancer. We have evaluated in this study the therapeutic effects o f Zoladcx and the correlation o f response w ith peripheral endocrine changes in post-menopausal breast cancer patients. Twelve patients had received Tam oxifen and three had a complete response, seven had a partial response and one stable disease F.ighl had been given am inoglutthim ide and lo u r had a p a rtial response. N o patient had received chemotherapy. O ther pre-trcaim cni characteristics are shown in Table 1. Patients received m onthly subcutaneous Zoladcx. 3.6 mg. after in filtra tio n o f the local site w ith lignocainc Response was assessed hy U IC C criteria (H ayw ard et al.. 1977). A fte r Zoladcx failure, patients who had received no previous endocrine therapy were treated w ith Tam oxifen 20 mg daily when appropriate. O cstronc and oestradiol were measured hy radio -im m uno assays which we have previously described (H a rris et a t. |9X3u; D ow scit et a i , 19X7). These analyses have been specifically developed for the analysis o f post-menopausal plasma oestrogen levels and have sensitivity lim its o f 30 and 3 pm ol I 1 respectively. Results Responses to endocrine therapy Three patients showed p a rtial responses fo r durations o f 21 weeks and 141 weeks (continuing) and one patient died after 6 weeks from other causes, f o u r patients showed stable disease for more than 6 m onths (range 29 66 weeks). A ll responses were in soft tissues. Twenty-one patients had progressive disease. Previous hormone therapy did not affect the likelihood o f response to Zoladcx. O f the 14 patients w ith previous endocrine therapy, tw o showed partial responses and tw o stable disease, while o f the 14 w ith no previous endocrine treatment, one showed a p a rtial response and tw o stable disease. T o x ic ity was m inim al, w ith no com plaints o f problems Materials and methods Twenty-eight post-menopausal patients w ith locally advanced r progressive hreasl cancer were studied fou rte e n had no previous endocrine therapy and 14 had previously shown either stable disease lo r more than 6 m onths o r partial response to other endocrine therapies. (.Mirespondence A . L. H u n ts Received X July 19XK; and in revised lo rn ). I t August I9XX T aM 1 Pro-1real meni characteristics Menu Median Raille Age (years) 'lim e from I.M P (years) Disease-free interval (weeks) Pro-treatment weight (kg) lim e from first relapse to M arl o f Zoladcx (weeks) 67 7 12 K 21 2 j 3 19 103 45 0 471 6? 65 42 102 73 2 ft SOX Previous endocrine therapy. 14 Sites o f disease: S o il tissue 23. lung 1. liver 1. nodes 9. bone 7. V Akticancer Research h 417 434(ivsn) Ovarian Cancer (Review) Etiology, Diagnosis, Prognosis, Surgery, Radiotherapy, Chemotherapy and Endocrine Therapy B E R E N D J. S L O T M A N and B. R A M A N A T H R A O Department o f Endocrinology, Academisch Ziekenhuis Vrije Universiteit, Amsterdam, The Netherlands A ttract. The increased incidence o f ovarian cancer tn women whose ovulations were not suppressed by pregnancy o r oral contraceptives and the increase in the incidence o f the disease wi the onset o f climacterium support the hypothesis that ovarian cancer is an endocrine-related disease. A n im al ex perimental results further support this contention. However, Identification o f a population at risk, prevention, and early detection is difficult. A t present, tumor markers are useful in monitoring the disease, though cannot be used fo r screening. Hast o f the diagnosed cases are o f advanced stages. Besides mging, tumor histology, and residual tumor load are o f prognostic importance. The need fo r accurate in itia l surgical staging with optim al Conor cytoreduction and the importance o f second-look wrgery to confirm the response to therapy are emphasized. The precise role f o r radiotherapy in the treatment o f ovarian cancer is s till to be established. Presently, response rates o f W% are achieved using cispladn based combination che motherapy. In spite o f this, long term survival has not im proved. Endocrine therapy has hitherto been used empirically and mainly as a last resort in the treatment o f advanced ovarian cancer. Response rates o f about 10% have been reported. Information on tumor predisposition related to hormonal control should be a key parameter in selecting the appropriate therapy. Tum or analysis has shown that androgen receptors predominate in ovarian cancer, compared to estrogen and progestin receptors. A clinical trial based on endocrine para meters is warranted. Contents Introduction Etiology Endocrine factors O ther factors Diagnosis Symptoms Tum or markers Preoperative evaluation Prognostic factors Stage H is to lo g y Tum or load after surgery Surgery Prim ary surgery Second-look laparotomy Second-look laparoscopy Secondary debulking surgery Radiotherapy Chemotherapy Single agent treatment Combination chemotherapy Second-line chemotherapy Intraperitoneal chemotherapy Immunotherapy Prediction o f response to chemotherapy O ther approaches Endocrine therapy Steroidal and anti-steroidal agents Steroid receptors Conclusions References Introduction Correspondence to: Dr B R. Rao, Dept o4 Endocrinology, Academisch Ziekenhuis V n je Universiteit. P.O. B o i 7057,1007 MB Amsterdam, The Netherlands. Key Words: Ovarian cancer, etiology, diagnosis, prognosis, airgery, radiotherapy, chemotherapy, endocrine therapy. Cancer o f the ovary is the leading cause o f death among women w ith gynecologic malignancies (1). In spite o f the improvement in the treatment o f several forms o f cancer in the last decades, the survival rates for patients w ith ovarian cancer are still poor. O nly 5 to 15% o f the patients w ith advanced disease (F IG O stage 111 and IV : Table l) survive 5 years or longer (2). The 5-year survival rates fo r other stages vary between 45% for stage II to 72% for stage la (2). Since most o f the patients are in stage I I I or IV at the tim e of A n t k a n c e r R e s e a r c h &: 417-434 (i% K ) diagnosis, (he overall survival rale for ovarian cancer is low. Ovarian neoplasms can originate from the surface epithe lium. ihe germ cells and the ovarian siroma (3) (Table II). The vast majority o f the malignant ovarian tumors are o f the common epithelial type, and include serous, mucinous, en dometrioid. dear-cell and undifferentiated adenocarcinomas, as well as the rare malignant Brenner tumor. These tumors are derived from the surface epithelium of the ovary, which is the adult equivalent of the mesothelium of the embryonic gonad (Mullerian epithelium) There is a dose relationship between epithelial tumors and the Mullerian epithelium. Serous tumors resemble the epithelium o f the Fallopian tube, endometrioid and clear cell tumors the endometrium and mucinous tumors the endocervical epithelium The papillary serous cystadenocarcinoma is the most common type of malignant epithelial ovarian tumor, constituting about 40 (4) to 68% (5). The malignant mucinous tumors constitute 3 (6) to 21% (5), whereas 5 (5) to 20% (4) o f the malignant epithelial tumors are o f the endometrioid type, and 5 (7) to 10% (4) of the clear cell type O ther epithelial tumors are less common The spread of epithelial ovarian cancer occurs mainly by continuity to the peritoneum and paracolic gutters, omen tum. diaphragm, and serosa o f the liver. The proxim ity of omentum, terminal ileum, cecum and sigmoid colon to the pelvis makes these organs frequent sites o f metastatic im plantation Obstruction of the lymphatics o f the diaphragm by tumor cells may lead to the formation o f ascites Further more, direct growth o f tumor to the surrounding tissues and lymphatic dissemination may take place. Lymphatic drainage from the ovary is mainly to the paraaortal nodes, but occasionally also to the pelvic or inguinal nodes. However, extra-abdominal dissemination by the lymphatic route is slow Haematological spread is usually late and infrequent ( 8 ). Etiology Understanding the epidemiology and etiology o f ovarian cancer may contribute to prevention and earlier diagnosis However, in spite of extensive research on endocrinological, environmental, and genetic factors, the causal analysis of ovarian cancer still escapes comprehension Endocrine factors. Chronic administration o f estrogens, progestins and androgens has been shown to result in ovarian cancer in animal studies (9-11). High levels of gonadotropins have also been correlated with ovarian neoplasms in animal experiments (12). Epidemiologic studies also suggest that ovarian cancer might be an endocrine-related tumor. Am ong women with ovarian cancer, an increased incidence o f nulliparity and a lower mean number of pregnancies were detected (13-17). Several authors noted a decreased incidence o f ovarian cancer after the use of oral contraceptives (18-21). Casag- randeefu/ (15) found an increasing risk in ovarian cancer; the ovanan cancer incidence rising steadily with the total time it ; a woman's life during which her ovulations were not suppres- j sed by pregnancy, lactation or the use of oral contraceptive*. ( It has been suggested that each ovulation causes a minor trauma o f the ovarian surface, which cumulatively could contribute to ovarian cancer (22). During pregnancy and during the use o f oral contraceptive agents, ovulation is suppressed and the levels o f gonadotro pins arc low. The high levels of gonadotropins in women in the early post-menopause may play a role in the development o f ovarian neoplasms (23, 24). This is supported by the dramatically increased incidence o f ovarian cancer in women above the age o f 45 years when gonadotropins reach high ^ levels (23). In this connection, high levels of androstenedione, the precursor o f estrogen, in the postmenopausal years may also have a role in the development of ovarian < cancer. In some studies estrogen replacement did not seem to > effect the incidence o f ovarian cancer (25, 26). However, others reported that more women with endometrioid cancer of the ovary were found among those who used estrogens during menopause (27) A two to three - fold excess of ovarian carcinoma was found in woman who used conjugated estrogens and diethylstilbcstrol (DES) (28). In animal stu dies, exposure to DES in utero could induce cysiadenocarrinoma o f the ovary (29). To date, no relationship has been demonstrated between intra-uterine DES exposure and ovirian cancer in humans. In a long term follow-up study of women who used medroxy-progestcrone acetate for con traception, no increased risk of ovarian carcinoma vat observed (30). Other factors. The incidence o f ovarian cancer is higher io most industrialized countries, except Japan. Dietary or other environmental factors may be responsible for this phe nomenon, since in Japanese women who emigrated to the U.S. an increasing incidence o f ovanan neoplasms is seen (26, 31). In the U S. white women have a significantly greater risk of developing ovanan malignancy than black women (26). However, the results o f studies o f dietary factors in rela tionship to ovarian cancer are contradictory. A n increasing risk for ovanan cance r with increasing consumption o f animal fat has been reported (32). Carcinogenic substances in fat are presumed to promote tumor growth directly, or via the ability to increase estrogen production by bacteria in the gut (32). O ther investigators could not confirm this observation (33). Coffee drinking, smoking, and alcohol consumption do not seem to increase the risk o f ovarian cancer (24, 32, 33). There is controversy about a link that has been suggested between ovarian cancer and talc (34,35). Cancer o f the ovary has also been linked to asymptomatic mumps infection (36). It has been suggested that the infection might lead to premature depletion of oocytes, resulting in higher levels o f gonadotro pins (36). m* Slolman and Rao: Ovarian Cancer (Review) Table I. Stagei o f primary carctnoma o f the ovary, according to the Imemanonal Federation o f Gynecology and Obstetrics (IV74). Tabic II. Summary o f the classification o f malignant ovarian tumors, according to W. H. O. (3). Sup 1 Suge U () <ii> Suge lb <> () Stage ic sur n Suge Ila Stage Mb Stage lie Growth limited to the ovaries. Growth limited to one ovary. No tumor on the external surface: capsule intact. Tumor present on the external surface; and/or capj.uk ruptured. Growth limited to both ovaries; no aschctNo tumor on the external surface; taps.uk intact. Tumor present on the external surface; and/or capsule ruptured. Tumor at either Stage la or lb. but wnh obvious ascites present or positive peritoneal washings. Growth involving one or both ovaries with pelvic extension. Extension and/or mciastasct to the uterus and/or fallopian tubes. Extension to other pelvic tissues. Tumor at cither Stage Ha or lib , but with obvious ascites present or positive peritoneal washings. | || III IV V VI V II V II V III Common "epithelial" tumors. A. Serous tumor*. B. Mucinous lumori C. Endometrioid tumors. D. Clear cell (mcsoncphtoid) tumors. E. Brenner tumors. F. Mixed epithelial tumors. G . Undifferentiated carcinoma. II. Unclassified tumors. Sex cord stromal tumors. l.ipoid cell tumors. Germ cell tumors. Gnntidohlastoma. Soft tissue tumors not specific to the ovary. Unclassified tumors. Secondary (metastatic) tumors. Tumoi-likc conditions. s u r in s u r iv Spedai category Growth involving one or both ovaries with intrapentoncal mciastasct outside the pelvis and/or positive retroperitoneal nodes. Tumor limited to the true pelvis, with histological ly confirmed extension to small bowel or omentum. Growth involving one or both ovaries, with disUne mtastases or pleural effusion rs present, with positive cytology or metastasis to liver paren chyma. Unexplored cases which are thought to be ovarian carcinoma. In an attempt to derive a common lin k between different explanations o f the etiology o f ovarian cancer, Cramer and Welch (24) described a model in which inclusion cysts are formed in the ovary by incessant ovulation or by foreign body stimulation. Stimulation by gonadotropins o r estrogens may lead to differentiation, proliferation and malignant trans formation o f the inclusion cysts. The factor responsible for the final step, i.e. malignant transformation, remains unknown (24). Epithelial cancer o f the ovary can also be fam ilial (37, 38). The hereditary pattern was found to be consistent w ith an autosomal dominant gene w ith variable penetration (37). Some genetic syndromes are also associated w ith ovarian cancer. The occurrence o f tumors in dysgenetic ovaries is well known (39). la patients with Peutz-Jcghers syndrome more sexcord stromal tum or* are ten, and in patients w ith m ultiple basal cell carcinomas, ovarian fibromas are seen more fre quently (31). Ovarian cancer has also been related to the ABO-system (40). In a study o f 1930 patients versus 24120 controls, more women with ovarian cancer were seen with blood group A than expected (40). The reason for this is unknown. Diagnosis Symptoms. Early ovarian cancer has an asymptomatic charac ter. For this reason, ovarian cancer is frequently called a silent killer. Complaints arise when there is expansion o f the tum or, adherence to the surrounding tissue and form ation o f asciies The presenting symptoms arc usually o f short duration and most frequently consist o f pain and abdominal distension (41-43). Less often, abnormal vaginal bleeding is noted or an abdominal mass is found (42,43). Extensive abdominal disten sion can cause dyspnea. Asymptomatic prolapse o f the uterus or vagina can become symptomatic. Tum or deposits on the small bowel can change its m o tility (44). Ascites or large omental mtastases can lead to dyspepsia and decreased food intake. The weight loss caused by decreased food intake may not be apparent due to abdominal distension by ascites and tumor volume. Tum or markers. Tum or markers might be useful for the diagnosis o f ovarian cancer and for m onitoring the disease status during and after treatment. Elevated plasma levels o f carcinoembryonic antigen (C E A ) are noted in 46% o f patients w ith ovarian cancer (45). Elevated levels o f C E A have also been noted in patients w ith carcinoma o f the cervix (53% ) and endometrium (37% ), patients w ith benign gynecologic dis eases (18% ), and in healthy volunteers (11% ) (45). In patients w ith ovarian cancer, the plasma levels o f C E A are related to stage o f disease and tum or D N A content (45). A relationship w ith the histological type has also been described (46). In patients w ith CEA-positive tumors, determined by im m uno histochemical staining, plasma C E A levels could be used in the follow-up (46), especially in cases o f m inim al tum or lesions 4m A N I I C A N t 't K R E S EA R C H & 417 434 (IW K) after .surgery (47). Furthermore, the use o f radiolabeled anti bodies against (T .A can he helpful in the detection of prim ary and metastatic ovarian tumors (48. 49) Tumor-associated antigens are another group o f tumor markers. They are glycoproteins that are present on the cell surface of cancer cells, but not on the surface o f normal cells. Radioimmunoassays, using polyclonal reagents, have been developed for the detection o f ovarian caretnoma-associated antigen (O C A A ) (50) and ovarian carcinoma antigen (O C A ) (51) O C A A and O C A could be demonstrated in the serum of approximately 70% of patients with ovarian cancer, but the value of O C A A and O C A in the surveillance o f patients with ovarian cancer was disappointing (50. 52). However, scrum levels of NB/70K. isolated from O C A , correlated well with stage of disease and the residual tumor burden in ovarian cancer patients (53). Bast and co-workers developed a monoclonal antibody (OCI25) against the antigen C A I2 5 . and reported elevated levels o f CA I25 in 82% of patients with ovarian cancer (54). Although determination of CA 125 is not useful as a screening method for ovarian cancer. CA125 has been found to be an useful parameter in monitoring the disease (54-58). In patients with recurrent or progressive ovarian cancer, increased CA125 levels were seen in more than 90% o f the cases, whereas in only 34% o f these cases increased levels of C E A were noted (56). Rising levels o f CA125 may indicate tumor progression, but decrease of the serum levels to normal ranges does not always indicate complete tumor eradication (54, 57. 58). A number of other monoclonal antibodies that react with epithelial ovarian cancer have been investigated (52). The monoclonal antibody OV-TL3. which in contrast to 0 0 2 5 recognizes a common antigen on most ovarian carcinomas, including the mucinous, may be of value for rapid diagnosis of ovarian cancer (59). This is yet to be demonstrated. A t present, tumor markers are not specific enough for the diagnosis of epithelial ovarian cancer. However, presently available markers can be useful in the surveillance o f patients during and after treatment. The use o f radiolabeled antibodies against tumor-associated antigens may become valuable in the detection of metastase* (60-62). In addition, tumor markers may be used as targets for antibody-directed treatment (63). Preopcrativt evaluation. Among the various well known pro cedures. ultrasonography recently, has been used for the diagnosis o f ovarian cancer (64). In general, finding a cystic lesion without septac suggests a benign tumor, while tumors with multiple internal echoes or solid parts must be considered as malignant (65,66). Furthermore, ascites can be detected by ultrasound. In addition, computerized tomography can detea upper abdominal lesions and retroperitoneal lymph nodes from the pelvis to the diaphragm (67). Nuclear magnetic resonance imaging and computerized tomography appear to give similar results in patients with gynecological tumors (68). Lymphangiography has advantages over other methods of assessing retroperitoneal nodes. The nodes do not have to be enlarged and metastases as small as 3 mm can be recognized (67). The overall accuracy o f lymphangiography using strict criteria is generally reported to be in (he range of 80 to 90% (67). Evaluation o f lymphangiography has shown that nodes which were positive at lymphangiography were also positive when surgically biopsied and was false-negative in 19% (9/41) o f the cases (69). In doubtful cases, lymphangiographic find ings may be confirmed by percutaneous transperitoneal fiat needle biopsy (70). Prognostic factors The prognosis of patients with cancer of the ovary dependsoa several more or less independent factors, namely stage, hist logy and amount o f residual tumor after surgical resection. Furthermore, the age and the condition o f the patient arc of importance. Stage. U n til the standardization o f classification o f ovariu cancers by the FIG O (Table I). various dassificalions, based on operability of (he tumor and on anatomic extent of disease, have been used (71). It is dear that because of this variety of classifications, comparison o f results from different studies was often impossible. The clinical stage of tumor growth at the lime o f diagnosis generally considered the best prognostic indicator (72). The increasing accuracy of staging o f ovarian cancer led to i> creased survival rales o f both patients w ith low-stage cancer and patients with advanced cancer (73). The latter can be explained by the fact that w ith more accurate staging, higher stages not only include patients with Obvious advanced d> ease, but also patients with minimal abdominal lesions, result ing in a better prognosis for the whole group of patients. This inform ation must be taken into account in comparing current survival rates for different stages with previously reported rales. Stage 111 ovarian cancer (Table I) includes patients with abdominal metastases varying from microscopic lesions to large tumor deposits. In a recent revision o f the FIGO staging system (Table III) , stage III disease is divided into stage II1A w ith microscopic seeding to the peritoneal surface, stage 1I1B with pentoneal implants not exceeding 2 cm in diameter, and stage 1IIC with larger abdominal implants and/or positive inguinal or retroperitoneal nodes (74). . Peroperative rupture o f (he tumor in stage 1 disease doa not seem to be o f great influence on the prognosis (75, 76). However, decreased 5-year survival rate of patients with stage I cancer and peroperative ruptured cysts was rcporled in another study (77). Histology. The histological type o f the tumor seems to be leas im portant in prognosis. In general, mucinous, endometrioid and clear-cell tumors have a better prognosis than other epithelial tumors (78). When tumors are analysed for histolo gical type. after correction for stage and/or grade, the prognos- mucinous borderline rumors has been reported < ) The T *b k 111 S u g e i o f p rim a ry carcinom a o f the o\-ary. th` UCyca, survival ra.es were 83% and 73% . r e s ^ u v e ly (87). Ir J l- i f b,Mna ,,w7< Because o l ihis favourable prognosis, tumors of low polcnlt Growth limited to ibe ovane* G io . l h lim .lc J lo o u v .ty . no M c ilo N o >umor on the Hem al ,uilcc. c.piulc moon Growth limited to both o v .n e t no m e n . No tomo, on the e tle in .l .u it.tx : p i t . l t mucl Tumor et.het S t,, = l.o t lh .h u t wrlh rumor on the .u rl.ee o( .me o. both o v .n e t, o . w.th capsule ruptured, or with .K ite , prevent conum in, m ill,- malignancy should be regarded as a specific entity In the dc.crm.nal.on o i his.olog.cal type and grade of ovarian tumors, a stgn.f.cam v a r ia b ly hai bccn repo.tcd (91-93) Discrepancy was seen particularly in the d.scnm na lio n between borde,1,ne and well differentiated rumon. 9 92) and between undifferenlialed and serous lumors (92 . II has been ahown lh al differences between d tllercn l Plho g- n .n t cell, ot with poottve peritoneal m h tn p isls in grading ovarian lum ots also reflcci differences m P 8 Growth involving one ot both ovnrie. with pelvic eiicntion. E ric norm .n d /o , m c ..,.a c . to Ibe u le ru . .n d io . nosis (93) A part from an inlerobserver variation, an tnlrao ,, e r variation was noted (93). For these reasons rhe assess ment o f more objective and rcpr.iduc.ble techniques is re . Stage Mb Stage He 1 . ** fallopian wibev * icm ton in olhci pelvic li o c t Too.. .1 either S.a,c I I . or lib . hot . . . h tumor on ibe surface of one oi bolh ovarms; or with capsuled) ruptured, or with ..c . U t preacnl conUining malignani Us or wiih positive peritoneal washings Tumor involvin, o n . er both o v .n c . with pen rune,I impl.nl ouinde ibe pclvt, .nd/or ponitve rcuop cn io n ca l or inguinal nodes quired. Quantitative morphometric methods have ^ ' wn a aood correlation w ilh prognosis of ovarian cancer (9 ). L o m e trv is especially im ponanl in the differenlialion of borderline and malignant tumors and in identifying those borderline lumors which are likely lo pursue an unusual malignani course (94). Flow cytome.r.c and cytopho.ometnc analysis of D N A levels in the lum or have been reported lo be o f prognostic value (95-97). iii. 5," C , , U Stuc lllb H Histologically confirm ed m icrmcopic ceding o i ibdomrn.t periioaeal vutl.ee. Node. ne,.t,ve. H o n d o ,i l l i eonlitmed ttnpU o t, peritoneal surfaces, none cscecdmg 2 cm in d,a- Tum or load a/rer surgery. In pat.en.s with advanced ovarian cancer the sue o f residual lum or lesions after treatment is the " " " im portant prognos.tc factor (79 . 83 . 98-IU1). An tm mctcr. Nodes negative. Abdominal implants > 2 cm in diameter and/or positive rciroperiioncsl or inguinal nodes. Growth involving one , both ovaries, wllh d.s,.n i meiasiases Parenchymal liver metastasis equals Siage IV . creased complete response rate to chemotherapy, ' "firm ed ,, second-look laparotomy, was observed tn pu n u w ith residual lestons < 3 emcompared to those w.th residual tumors > 3 cm (98). In another study the survival rales o f P,ien' s metastases smaller lhan 1.6 cm after surgery were found to be significantly h.gher than those of pat.ents wt.h large, mlas- " ^ ^ r v i v a l rate for patterns wtrh large abdominal m e t tases surgically reduced below a certain k m il. was found to be K value loses is importance (78. 79). Thus mucinous adeno- A n n e a l fo th/surviva, rate o f p .,. e n w ^ below that lim it without any treatment (99.101). However beneficial effect could be demonstrated m patterns w iih in i tal ^ IX ^ U o M h c lu ^ r U ^ ^ v ^ n *J metastases Urge, than 10 cm (101) This is cons.slenl w ith the hypothesis that spontaneous mutation in large lumors w ill have already resulted tn cell clones rests,an. to system treat- . item factor " f " ^ j^ 'a -- } p r o m t s e s ^ c i.O y o . low stage ovan.n musi be patd ,o . group o f lu m ort wiih me^ a " suit Ol all these observatrons. patterns w ith advanced ovarian cancer are now d.vidcd into ihose w.th m inim al iesidual disease (largest residual lesion < 2 cm) and tlm w tl bulky residual disease (largest restdual lesson > 2 cm) (103). Surgery o f the ovarian tumors are o f rhe borderime type f w M n 11 to Of the cases, the tumor I. diagnosed," .rage HI or IV (87-89). Borderline tu m o ra h .v c . " W c p r o g - aosis, regardless of spread o f tumor (86,90). A 5-ye ^ T o f 91% for serous borderline tum o r. and o f 81% for lient w ith ovanan cancer, n F* nd |()f |he A nticancer Research h 417-a.vi o <mmi masses In case o f complication* such as bowel obstruction, intervention laparotomy may be necessary. Primary surgery. A (para)median abdominal incision is made, extending from the symphysis pubis to above the umbilicus. After opening the peritoneum, ascites is aspirated for cytolo gic examination. In the absence o f ascites, peritoneal washing is performed. Then, careful inspection and palpation of the intraabdominal organs and peritoneal sufaecs is earned out. Suspicious lesions are biopsied. and the pelvic and paraaortic nodes are palpated. The next step is removal of the prim ary tumor, along with a total hysterectomy and bilateral-salpingooophorectomy. In case of extension o f the tumor to the bowel, bladder or cul-de-sac, the retroperitoneal approach may be of benefit, since deep infiltration of ovarian carcinoma in the peritoneum and other stmetures is unusual (104). In some cases, resection of bowel with reanastomosis or colostomy may be indicated The greater omentum is removed below us attachment to the transverse colon (infracolic omentectomy). In cases o f gross involvement o f the omentum, a total omen tectomy should be performed (105) When feasible, an appen dectomy should be carried out If no metastases are seen or palpated, biopsies are taken from the peritoneum o f the cul-de-sac. the vcsico-utcrine pouch, the parietal peritoneum below the arcuate line, the pelvic walls, the paracolic gutters, the sigmoid colon, the hepatic and splenic flexures, and the right diaphragm (lOti). Tumor adhesions are also biopsied. Clinically suspicious pelvic or paraaortal lymph nodes are removed (100). Clinically negative paraaortic nodes are sam pled in the area near the insertion of the ovarian veins (100). The surgical findings and treatment should be accurately documented Inoperable tumor masses may be marked with clips. Such a procedure may be helpful in subsequent tumor identification for radiation treatment, computerized tomogra phy and during second-look procedures. The nght hemi-diaphragm is susceptible to implantation o f tumor (105-108). Metastases to diaphragm in patients with presumed stage 1 or (1 disease who underwent laparoscopy within one month after exploratory laparotomy, were observed in up to 44% of the cases (105-108). Spread o f tumor via the lymphatics is im portant in ovarian cancer (8). In presumed stage I ovarian cancer, paraaortal lymph node involvement is reported in 10 to 18%, and pelvic node involvement is reported in about 9% of the cases (107, 109). The frequent involvement of paraaortic and pelvic lymph nodes in apparently low stage disease supports the view that extirpation of paraaortic and pelvic nodes should be a part of any staging laparotomy for ovarian cancer. Since the size of residual tumor masses after treatment is found to be the most important prognostic factor in patients with advanced ovarian cancer (82.83.98-101), the importance of cytoreductive surgical treatment cannot be overstated. Debulking of the tumor may achieve as much as 99.9 per cent tumor volume reduction (99), and restore gastrointestinal function and improve the patient's nutritional status(44). The residual small lesions appear to be more vulnerable tod- 9 motherapy and radiotherapy since they are better vascularixsi ' and oxygenated and contain a higher fraction of actively 1 proliferating cells (110) Resection of tumor bulk may be easy to perform in patients with minimally invasive growiag tumors. In other cases, debulking with "maximum effort" necessary. No difference in survival was found between pa- i tients in whom debulking, leaving masses less than twoca, - was readily accomplished, compared to those who had "m u- ' imum effort" debulking (100). In a study of 70 patients, optim al cytoreductive surgery could be achieved in 87% of the cases (111). In 20% o f the patients bowel resection waa . required (111). The m orbidity and m ortality o f patients wm ' found to be acceptable (111, 112). In some patients, however, J cytoreductive surgery is not feasible, because of the patients' poor general condition, or due to the site o f metastases. Sometimes debulking surgery may become easier after a few successful chemotherapeutic cycles. ^ In young women with a desire and a potential for child- , bearing, with a w ell differentiated, unilateral, encapsulated, nonadherent, common epithelial ovarian carcinoma, the liter* us and contralateral ovary may be preserved if metastases ate excluded (73. 105, 113). Suspicious areas on the preserved ovary should always be biopsied. Blind biopsy of the preserved ovary has also been recommended, especially in cases o( serous tumors, since these tumors are frequently bilateral (73). If there is any doubt, a conservative approach should be chosen. If necessary, extensive surgical treatment a n be performed after detailed histological examination of the re sected tissues (73). Close patient follow-up is mandatory. A fte r completion o f childhearing, resection o f the residual internal genitalia should be reconsidered (73). This concept supported by the study o f Munnell et a l consisting of 190cases o f stage la ovarian cancer (114). They concluded that there was no statistically significant difference in 5 year survival rates between patients in whom unilateral salpingo-oophorectomy was performed, compared to those in whom the reproductive organs were removed (114). Second-look laparotomy. Second-look surgery is defined as an exploratory laparotomy to assess the cancer status in patients who are clinically free of disease following a planned course of chemotherapy (115). The term "second-look operation" h also used to refer to restaging operations and secondaiy cytoreductive surgery. Second-look laparotomy follows the same principles as the * staging laparotomy. Ascites or peritoneal washing fluid k examined. Remnants of ovaries, uterus, or omentum art removed. Residual tumor masses are removed as far u possi ble. Sites, known to have residual cancer at the initial lapar otomy are biopsied. M ultiple biopsies are obtained from the peritoneum of the bladder, the cul-de-sac, the pelvic walls, the paracolic spaces, and from the diaphragm. Enlarged pelvic and paraaortal lymph nodes are removed. I f there are no enlarged nodes, sampling can be performed. However, if A Slot man and Rao: Ovarian Cancer (Review) ' Thk IV . Frequencies o f p enm en! disrate at second - /cm* laparotopy in ; pmtma with advanced ovarian cancer with clinical complete response. Atffcx Dwpfcl (116) On 017) fa m a (118) Webb (119) Capri*o<l (120) .Ibecfcio (121) m itili (122) atrek (123) I*iftc r (l2 4 ) lobcm (125) |U|u (126) fWbbi (127) Yf 1986 1981 1986 1982 1985 1984 1984 1984 19R5 1982 1982 1983 Number Percentage 7/24 9/20 9/19 32/59 161/246 13/20 37/55 38/56 7/10 19/24 53/65 26/30 29 45 47 54 65 65 67 68 70 79 R2 87 biopsies or peritoneal washings already were positive, lymphadenectomy is not required. The diagnostic value of second-look laparotomy in patients with a clinically complete response has been justified in several studies. Persistent disease, diagnosed by second-look lapar otomy, has been observed in 29 to 87% o f patients with advanced ovarian cancer who were clinically free o f disease (Table IV ) (116-127). This clearly demonstrates that secondlook surgery is useful in m onitoring the response to treatment. Negative findings at second-look laparotomy arc not a defin i tive sign o f tumor eradication. In patients with moderately and poorly differentiated tumors with either negative o r microsco pically positive findings at second-look surgery, a recurrence rate o f approximately 50% was reported after a follow-up period o f at least 4.5 years (128). In another study, consisting of patients with advanced disease and a negative second-look laparotomy, 24% recurred w ithin 5 to 32 months after lapar otomy (129). Second-look laparoscopy. Attem pts have been made to use laparoscopy in the m onitoring o f the disease status (105, 106, 130-133). A m ajor advantage o f laparoscopy is the possibility of performing repetitive procedures w ith minimal m orbidity and discomfort to the patient. In some studies, laparotomy was performed in case o f negative findings at second-look laparo scopy. A tumor was found in 20 to 55% o f the cases (106,130, 132, 134). These data indicate that laparoscopy is o f lim ited n lu e in the m onitoring o f ovarian cancer. However, perform ing laparoscopy prior to second-look laparotomy may be valuable, since the finding o f an unresedable tumor mass or diffuse intraperitonea! spread may obviate the need for lapar otomy (130, 132, 133). Secondary debulking surgery. The benefit o f cytoreductive surgery at second-look laparotomy is a m atter o f dispute. The primary unresolved question is whether the association be tween bulky disease and poor prognosis is merely due to the presence o f a great tumor load or to the fact (hat bulky disease is associated with more aggressive tumors w ith decreased sensitivity to chemotherapy. O nly in the form er situation might secondary cytoreductive surgery be beneficial. In some studies, secondary debulking operations were found to be o f value (135-137). In the study o f Schwartz and Smith (136), survival after second-look surgery varied directly with the amount o f tumor found at second-look laparotomy and the size o f residual masses after operation. In patients with residual tumor masses o f more than 2 cm in diameter, the 2-year survival rate was 9% , compared to a significantly higher response percentage o f 47.5 when all residual tum or was removed at second-look operation (136). In another study, an average survival o f 44 months in patients w ith residual tum or masses o f 2 cm or less was reported, while in patients with larger residual lesions the average survival was 4 to 7 months (137). However, in most o f these studies, nocisplatin-containing chemotherapeutic regimens were used. In studies o f pa tients who had received cisplatin-containing chemotherapy, no beneficial effect o f secondary cytoreductive surgery could be demonstrated (126, 130. 139). On the basis o f the above cited studies, it can be concluded that surgical resection o f bulky tumor should be performed early in the course o f ovarian cancer to be effective. Radiotherapy Radiotherapy has been used as a primary treatment for nonresectablc ovarian cancer, as an adjuvant to surgery, for palliation in advanced cases, and for the treatm ent o f recurrent disease. However, (he role of radiotherapy in the treatment of ovarian cancer has been the subject o f considerable con troversy, and is yet to be defined (72, 73, 140). In the Gynecologic Oncology G roup study o f adjuvant treatment in stage 1 ovarian cancer, no beneficial effect of pelvic irradiation could be demonstrated (141). In a study of 54 patients with stage Ia2 disease, Dembo el a lconcluded that postoperative pelvic irradiation was an inappropriate treat m ent. because relapses occurred throughout (he peritoneal cavity (142). Since ovarian cancer is considered to be a disease o f the whole peritoneal cavity, intraperitoneal adm i nistration o f chromic phosphate ( U P) and colloidal gold ( l9N A u ) has been used in patients with stage I disease (143). The colloids are taken up by the lymphatics o f the diaphragm, thus following a well recognized pattern o f metastascs. H ow ever, no significant differences, in relapse or survival were found between patients treated with intraperitoneal MP and patients treated with melphalan (144). A t two-years, the disease-free survival was about 80% in both groups (144). in a prospective study o f patients with stage lb , 11, and 111 ovarian cancer with small or no macroscopic tum or residuum after surgery, the effect o f postoperative radiotherapy amd/or chemotherapy was evaluated. It was concluded that abdominopelvk irradiation, using the moving strip method w ithout diaphragmatic shielding, significantly improved patient survi val rate and long-term control o f occult upper abdominal A N I K A N l LR R fcS fcA K O l S 417 4.U (IW K ) disease in 2510 30% more patients lhan did pelvic irradiation alone or together with chlorambucil treatment (142. 145). In patients with stage II or III with large residual disease, cures were rarely obtained with radiotherapy. However, no careful initial staging was performed in this study, the extent of eyloreductive surgery and the amount o f residual tumor masses were not reported, and chemotherapy was not op ti mally administered. In another prospective and randomized study, whole abdo minal irradiation, using a moving strip method with liver and kidney shielding, was compared to melphalan treatment in patients with stage I III tumors with no residual tumor or residual masses levs lhan 2 cm (146) In both groups, the same 5-year survival rales were found, being 45 to 65 per cent (146). However, the morbidity in the group o f irradiated patients was much higher. Hacker et at (147) found whole abdominal radiation to be useful in patients who responded to primary chemotherapy and had minimal residual disease at second-look. The use of radiotherapy after chemotherapy, however, is often limited by the fact (hat a high percentage o f patients is unable to tolerate this treatment (147. 148). O ther investigators could not demonstrate a positive effect of salvage treatment with whole-abdominal irradiation (149. ISO). Reviewing the literature, it can be concluded that wholeabdominal irradiation is more effective than irradiation lim i ted to the pelvis, and that the beneficial effect depends on the amount of residual tumor. Although whole-abdominal ra diotherapy may be curative in selected patients, the role o f radiotherapy in the treatment o f ovarian cancer is limited and is still to be defined. Chemotherapy .ViVig/e divi/ treatment. Alkylating agents were the first group o f drugs widely used in ovarian cancer Most series have reported response rates o f 40% to 60/i (72). For patients with advanced disease, median survival varied from 3.5 to 21 months (151). Melphalan. chlorambucil, cyclophosphamide and thiotepu have shown comparable response rales (72, 152. 153). Special care should be taken with these agents, because of the increased incidence of acute nonlymphocytic leukemia after long term use (154). The response rate of doxorubicin in untreated patients was found to be similar to that of alkylating agents (155). Hexarncthylmclamine is also effective in ovarian cancer (156. 157) A response rale o f 31% has been reported in previously untreated patients with advanced disease. Although this is less lhan the rates which can be achieved with alkylating agents, hexamethylmelamine can be useful in com bination with alkylaiing agents, since they have different mechanisms of action and no cross-resisiancc (157). Cis-diammincdichloroplatinum ( II) (cisplatin) is the most active non-alkylating drug in ovarian cancer. Response rales as high as 50% (158) and 67** (159) have been recorded after in itial treatment with cisplatin in advanced stages of ovahaa 0 carcinoma. In previously treated patients, response rates of 33% and 52% were achieved with doses o f 30 mg/m2and 100 mg/ni2 respectively (160). Analogues o f cisplatin with reduced toxicity have beca developed. Carboplatin (JM8) and iproplatin (JM9) hast shown sim ilar response rates to those o f cisplatin. with reduced neuro-, nephro-, oto- and gastrointestinal toxicity, but an increased myelotoxicity (161-163). Newer drugs with activity against ovarian cancer further include ifosfamkk, dihydroxybusulfan. prcdnimustinc. galucticol, and mitomyda (163). Antimetabolites such as 5-fluorouracil and methotre xate are shown to be o f little value as single agent treatment . ovarian cancer (151). Combination chemotherapy. The use of combinations of drag) with different toxicities and mechanisms o f action could theoretically lead to better results o f treatment. Only drugs that are effective as a single agent should be used, and ideally. ' at their optimum dose and time interval (152). Using non-cisplutin-bascd combination chemotherapy, i response rate o f 47% and a median survival of 14 month vm recorded (152). This is comparable to those obtained with monotherapy with alkylating agents. Only a few prospective randomized studies showed superiority o f combination che motherapy without cisplatin over melphalan (164, 165). I n i study in which a therapeutic regimen consisting of hexameibyF melamine, cyclophosphamide, methotrexate and 5-fluorouncil (Hexa-CA F) was compared with melphalan treatment,! significantly higher overall response rate (75 vs. 54%) with more complete responses (33 vs. 16%). longer median survival (29 vs. 17 months), but also an increased toxicity was seen ia the group o f patients treated with Hexa-CAF (165). Other workers failed to confirm the therapeutic superiority o f HexaC A F (166. 167) Several cisplatin - containing regimens have been studied ia prospective randomized (rials. The combination of dsplatia and adriamycin was associated with a clinical response rate d 80% (168) This was significantly higher lhan in patients treated with cisplatin alone, or with a combination o f thiotepi and methotrexate (168) In a comparison o f a combination o f cisplatin and cyc lophosphamide (CP) with cyclophosphamide alone in patients with advanced disease, the 2-year survival o f patients treated with the combination regimen was 62% , compared to 19% for the patients treated with cyclophosphamide alone (169). The use o f a combination regimen containing cisplatin. chlorambucil and adriamycin led to sim ilar results, as did the combination without adriamycin (170). Another three-drug regimen, consisting o f cyclophosphamide, adriamycin and cisplatin (C A P ), has been used by several workers, and an overall response rate o f approximately 80% was recorded (171-173). In a prospective randomized tria l, the CAP-regimen was compared with the CP-regimcn (173). Objective re sponse rates were similar for the two regimens (81% for CAP h i Slotman and Rao: Ovarian Cancer (R e v ie w ) w76% for CP), but a higher percentage o f survival complete response was recorded in the CAP-group (62 vs 39%) (173). This advantage was especially observed in patients with small residual disease ($0% vs 15%). However, survival was not statistically better fo r the CAP-group, since in some the tumor eventually recurred (173). Using a combination o f cyclophosphamide, hexamethylme lamine, adnamycin and cisplatin (C H A P ) in patients with advanced ovarian cancer, an overall response o f 92% was recorded, w ith a complete response rate o f 42% , and a median survival o f 24 months ( 174). However, in patients w ith tumor lesions larger than 2 cm, response was comparable to that of single agent treatment (174). In another study, an overall response rate o f 79% was reported using C H A P , w ith a median survival o f 31 months, compared w ith 50% response and 20 months' median survival in patients treated with HexaCAF (175). High response rates after the use o f C H A P were also reported by others (98,176,177). In the group o f patients treated with C H A P , earlier and more severe neurotoxicity was noted. A clinical tria l to compare the effectiveness o f a com bination o f cyclophosphamide, hexamethylmelamine, adriamycin and carboplatin (C H A C ) is in progress (179). Second-line chemotherapy. Second-line or salvage chemother apy is used after failure o f in itial chemotherapy. The best results are obtained when lim ited prior chemotherapy is given. To date, the highest rates o f clinical response are achieved using cisplatin alone or in combination treatment as a secondline treatm ent. Prelim inary results o f high-dose cisplatin admi nistration, with hypertonic saline hydration to reduce nephrotoeixity, showed a response rate o f 35% in patients who had relapsed on standard dose cisplatin (180). in another study, the use o f high-dose cisplatin in 31 patients who did not have prior cisplatin treatment resulted in a response rate o f 55% (181). O f nine patients who had already been exposed to lower doses o f cisplatin, only 2 (22% ) responded (181). W ith some combination chemotherapeutic regimens, re latively high response rates (40 - 60% ) were recorded in patients who received previous chemotherapy ( 182 - 190), but most responses were only partial and o f short duration (4-6 months). Using a combination o f hexamethylmelamine and cisplatin, a clinical response rate o f 55% was reported (182). With adriamycin and cisplatin a response rate o f 42% was recorded (183). The CAP-regimen showed rates o f 30% (184) and 48% (185). The combination o f hexamethylmelamine, adriamycin and cisplatin led to response rates o f 40% (185) and 58% (186). In the trial o f the Southwest Oncology Group, a response rate o f 48% was achieved w ith a combination of adriamycin, 5-fluorouracil, hexamethylmelamine and risplatin (187). Using the C H A P regimen, response rales o f approx imately 50% were recorded (188, 189). A fte r schedule mod ification and intensification o f the CHAP-regimen. a response rate o f 72% was achieved (190). Resistance to chemotherapeutic agents is a m ajor problem in the treatment o f patients with ovarian cancer. Recently, experimental model systems were developed in which mechanisms o f resistance could be investigated (191). in some resistant human ovarian cancer cell lines a decreased accu mulation o f adriamycin was observed, which could be reversed by a calcium channel blocker (192). This has led to a clinical (rial w ith adriamycin plus verapamil in refractory patients (193) . Resistance to melphalan and cisplatin is found to be associated w ith increased levels of glutathione (191). The use o f a specific in hibito r o f glutathione may be useful (191). Intraperitoneal chemotherapy. Intraperitoneal (i.p .) delivery o f chemotherapeutic agents offers the possibility o f exposing the abdominal cavity to much higher concentrations than those which can be achieved by intravenous administration (194) . This difference is caused by the fact that peritoneal clearance is generally slow in comparison to the total-body drug clearance. Drugs should be administered in a large volume (usually 2 liters) to make distribution through the whole intraperitoneal cavity possible (194 - 1%). The i.p. distribution can be visualized using computerized tomography or radio-isotopes. The distribution may be lim ited by adhe sions following surgery and the presence o f m ultiple tumor masses. The effectiveness o f i.p. treatment further depends on the size o f the tumor lesions. Furthermore, the effectiveness depends on the ratio o f plasma clearance to peritoneal per m eability o f the drug A high ratio is required for successsful treatment (194). The ideal i.p. drug should have a large peritoneal to plasma concentration range to be used, no local peritoneal toxicity and no cross-resistance w ith agents used for systemic treatment (195). In pharmacokinetic studies, a marked pharmacological advantage has been demonstrated for 5-fluorouracil. adriamycin, methotrexate, melphalan. cis platin, carboplatin and some other drugs (196). For i.p. delivery o f chemotherapeutic agents a Tenckhoff catheter has commonly been used. Complications, including bowel perforation, infection, hemorrhage, abdominal pain, and catheter leakage, blockage, and displacement are re ported in 67% o f the cases (197). W ith a single-use Tenckhoff catheter, the occurrence o f complications may be reduced (197). Another system used for i.p. administration o f drugs is the totally im plantable Pon-A-C ath system (198). The m ajor problem in i.p. treatment - the form ation o f a fibrous sheath around the catheter - occurs in both systems (199). Cisplatin is currently the drug o f choice in i.p. treatment (199). In general, no severe intra-abdom inal complications of this drug are seen. In two studies o f patients with minimal residual disease, i.p. cisplatin treatment resulted in surgically complete response in approximately one-third o f the patients evaluated (200, 201). In i.p. cisplatin treatment, thiosulfate may be used to decrease nephrotoxicity (194). Promising results using i.p. treatment have been reported in patients with small residual disease and in patients w ith malig nant ascites. Probably i.p. treatment may also be of benefit in patients with a negative second-look laparotomy and in pa tients w ith early stage disease who are at high risk for relapse. 425 ! A N IK AN O -K RESEARCH H; 417-434 (I9H8) Controlled clinical trials should define iho advantages o f i p. treatment in ovarian cancer. Immunotherapy. The addition o f an immunoputentiating agent should increase the capacity of tumor rejection via non-specific cellular activation and increase the recognition of and response to specific tumor antigens (203). In this aspect, Corynehacierium pur nun. BCG. and ovarian cancer anti serum have been used in addition to chemotherapy in the treatment o f patients with advanced ovarian cancer with prom ising results (2t)3 - 20b). Recently, the potential of other hiologicals. such as interferons (207 , 208) and inicrlcukin-2 (209) has been investigated. Surgically proven responses were recorded in 5 of II patients (45% ) with relapsing ovarian cancer using i p. intcrfeion-o2h (207). The responding pa tients all had residual tumors less than 5 mm, whereas none of the patients with larger residual masses responded Hredictum o f response to chemotherapy. Attempts have been made to predict the sensitivity of tumor cells to drugs in in vitro experiments. In this aspect, the effectiveness of therapeutic agents to those cells which are responsible for growth o f the tumor by cell proliferation, i t*, the clonogenic cell population, is of special importance. The tumor stem-cell assay, first described by Hamburger and Salmon (210), may be used for predicting response o f ovarian cancer to chemotherapy. In (his technique, suspensionsofmonodispersed tumorcells are incu bated with a number o f drugs. A fte r washing and incubation in a soft-agar medium for 2-3 weeks, the number o f colonies is counted and compared with colony formation in the absence of drugs. Inadequate growth of tumor cells has been a major problem using this technique (211. 212). In 90 to 99% of the cases where drug resistance was predicted, no clinical responses were seen (213, 214). In approximately 40% of the cases in which drug sensitivity o f the tumor cells was predicted, it was found to be false-positive (213). A t this moment, donogenic cell assay directed treat ment is still an experimental approach. Its value has to be established in prospective clinical trials. The tumor stem-cell assay may further be used for the screening and identification of new anticancer agents (215). Additionally, it may be em ployed in the detection of the presence of viable tumor cells, for example in (he peritoneal fluid. Non-clonogcnic methods for studying drug sensitivity in clude the subcutaneous implantation o f ovarian tumor tissue in nude mice, a congenitally athymic mouse mutant (216), and the subrenal capsule assay (217, 218). In the subrenal capsule assay, tumor tissue is transplanted under the renal capsule, a site which appears to enhance growth of xenografts by provid ing a rich vascular bed, where it proliferates during a short period. A fte r six days tumor growth is determined. Tabic V. Treatment with progestim in advanced ovarian (oncer. Invctiigiiinr Jolie (223) Varg ,224) Ward ,225) Tim oihy (22b) M;dkuian (227) Kauifm.in (22H) M ilk a * ..n (229) Bcrgqvisl (230)* Slayton (231) Mangioni (232) Mangioni (232) A -bo (233) Rendimi (234) Trope (235) Hamcrlynck (23b) tendoni (237) Wecih (23H) Cicislcr (239) Sikic (240) Agoni A A B B C D D D D r> i) D n D D D F E E Rouic; i.m. i. m . i m. i.m P `> p.o. p.. i.m. im im p.o p.o. i.m. i m. i.m i. m . po po. p.o Rcipomc rlc 1/10 1/ 6 3/15 1/ 7 21 9 1/1 1/19 31 4 IV |9 5/33 IV30 1/27 18/33 1/25 1/41 9/67 0/ 2 10/22 4/47 (10 %) (17 %) (20 %) CM * ) (% ) (% ) (5% ) (75 %) (0 *1 (15 %> ( 0) ( 4 %> ($5 %) (4% ) ( 2 %) (7 *i ( 0%) (45 %) (9 *) Total 53/360 < U * > 1 A 17-o-hydroxypiogciofom: - 17-n-capmaic. - 17-u-hydfoxy l9-norprogcicronc - 17-n-caproaic. C b. 17-a-diaicthyl b-dohydrnprogcsicronc. D medrox yprogesterone acciate. E - mcgctirol telile * i.m.: intramuurular p.o.: per o*. 1 A ll (ages. result from the experimental studies with M ullerian inhibiting substance (M IS ), that induces regresssion o f the Mullerian duct in the male mammalian embryo. It has been demons trated (hat MIS inhibits the growth o f human ovarian cancer cell lines in the soft-agar assay (221) and in nude mice (222). Further studies on these approaches are indicated. Endocrine therapy The fact that the ovary is not only the main source o f eslrogea and progesterone but also a target organ for these and other hormones suggests that epithelial ovarian cancer may be an endocrine-related tumor. Investigations to determine the role o f hormones in ovarian cancer include empirical treatment w ith hormonal agents in patients with advanced disease, stu dies o f steroid receptors in ovarian cancer and animal experi ments. Other approaches. Other interesting approaches in the man agement of ovarian cancer are the use o f antifibrinolytic agents (219) and the application of platinum as a potentiator for radiotherapy (220). New experimental approaches may also Steroidal and anti-steroidal agents. Estrogens were thought to be possible promoters o f ovarian carcinoma and estrogen counter-acting progestins have therefore been tested in the treatment o f ovarian cancer (223 - 240) (Table V ). In most of A2h Slotman and Rao: O varian Cancer (R eview ) V I. Treatment! with progestin in combination with other agents in ovarian cancer. h w tig n o r Agcnu1 Response rale f M (22)) Prog. Estrogen M e . (241) Prog + Estrogen Freedman (242) Prog Estrogen Hyeia (243) Prog. Tam oiifcn Moteen (244) Prog t Tam oiifcn Gmhnc (245) Prog + Endoian fc w m t (2*>!` Prog Melphalan Prospettive controlled Mudici M m (24*) M PA CAF (32 patients) CAF W (2<7) M P A Cyclophosphamide (71 patients) Cyctophuspharrude CF Untine (24) M PA*CAP (U patients) CAP (V K 2/11 9/65 1/ 1 (V|7 41/54 280) ( 0%) ( 18%) 14%) (l i% ) ( 0%) ( 76%) ( *5% ) ( 59%) ( 55%) ( % ) ( 58%) ( 42%) ( 0%) ( 50%) > |M P A Mcdroxyprogcsierooc accinte. CAF Cyclophosphamide, Adriamydn, and 5-Fluorouracil. i CF Cyclophosphamide. 5-Fluorouradl. CAP Cyclophoaphamidc. Adriamydn. and ds-Platin. 1 AU Mages. jt , -5 die early reports, both subjective and objective improvement kavt been reported as response. In Table V , however, only the objective responses by current standards are listed. In most studies, true objective response was recorded in only about 10 to lS % o fth c patients, with an additional 10% of patients with stabilization o f disease. It has been shown that intramuscular administration o f medroxyprogesterone acetate , (M PA) is more effective than oral administration (232). In the study o f the EO R T C Gynecological Cancer Cooperative Group o f M P A (500 mg daily i.m . for four weeks, with a maintenance dosage o f 1000 mgMPA weekly), only one partial response was recorded among 41 patients (236). The duration of this response was only 20 weeks. W ith high-dose megestrol acetate treatment, 1 complete and 3 partial responses were recorded among 47 patients with advanced disease. The sur vival fo r patients who responded to treatment was significantly longer than fo r nonresponding patients (240). In a study o f 33 patients with advanced endometrioid ovarian cancer treated with M P A , 18 responses (55% ) were recorded (234). It must be noted, however, that 79% o f these tumors were well differentiated. In the same study, 10 patients V with well differentiated stage M l endometrioid ovarian cardaomas were treated w ith M P A . Five o f them had a complete response (234). Geisler (239) recorded 6 complete and 4 1 "* partial responses among 22 patients treated w ith megestrol acetate. Survival o f patienu who achieved a complete reponse was 5 to 36 months. Patients w ith partial remissions ( survived 4 to 10 months (239). Progesttns have also been used in combination w ith other agents (225, 230, 241-248) (Table V I). Freedman er 4J (242) Tabic V tl. Trraimeni with anti -estrogens in advanced ovarian cancer. Investigator Agent Response rate Myers (243) Schwarts (249) Landoni (237) Shircy (2541) Weiner (251) Sicvin (252) Total T a m o iifc n T a m o iifc n T a m o iifc n T a m o ii fc n T a m o iifc n Tam oxifen 2J 2 1/ 13 0/ 55 (V 22 V 31 IV 22 6/145 (in o % ) ( 8%) ( 0%) ( o% ) ( 10%) ( 0% ) ( 4%) studied the effect of a sequential combination o f M P A and ethinylestradiol in 65 patients with refractory ovarian carcino mas. Nine (14% ) patients responded and 13 (20% ) had stable disease. Vascular complications were seen in three patients. In one o f the patients hemiplegia occurred. In a phase-ll study of cyclic treatment w ith M PA and tamoxifen, no objective re sponses were recorded (244). In a review of patients postoperalively treated with M P A plus melphalan, a response rate o f 85% was reported (230). However, about half o f these patients had early stage disease. G uthrie (24S), studying 54 patients treated with Gestronol and continuous oral cyclophospha mide, recorded 25 complete (46% ) and 16 partial responses (30% ). None o f these patients had received previous radio- or chemotherapy. However, these high response rates with a com bination o f progestins and alkylating agents could not be achieved in three prospective, controlled studies, in which treatment with chemotherapy plus M PA was compared with chemotherapy alone (246 - 248). The synthetic anti-estrogen tamoxifen has also been used in the treatment o f ovarian cancer (237, 242, 249-252) (Table V II) . Landoni ei al (237) recorded no responses among 55 patients w ith advanced disease treated w ith tamoxifen. However. 19 patients (35% ) bad stable disease (237). Another study also recorded no objective responses, but objective stabilization o f disease was seen in 80% o f 22 refractory patienu treated w ith 20 to 40 mg tamoxifen daily (250). In a study o f 31 patients with recurrent disease, 1 complete and 2 partial responses were seen (251). Six patients (19% ) had stable disease. Median survival was 16 months fo r responders and 7 months for nonresponders (251). Recently, the androgen fluoxymesterone was used in the treatment o f 16 ovarian cancer patients (253). N o responses were recorded (253). The LH R H agonist D -T rp -6 -L H R H (Decapeptyl) has also been used in the treatment o f ovarian cancer (254 , 255). In a study o f 10 women in whom agonist treatment was started after failure o f chemotherapy, tum or stabilization o r shrinkage was seen in 50% o f the cases (255). The rationale fo r this treatment is the dow n regulation o f gonadotropins and thus the m inimization o f the steroid out put, if any. o f the ovarian carcinoma. It is not yet known whether total down regulation o f gonadotropins can be achieved with this form o f treatment in postmenopausal women. 427 A n t ic a n c e r R e s e a r c h h. 4 |7 -4 M (|vhk) '.c Summarizing the literature, a response rate of 15% was recorded among 36(1 unsclccted patients treated w ith progestins Most responses were seen in patients with serous and endometrioid tumors. Duration of response was mostly short. Stabilization of disease was seen in approximately 10% o f the patients. Using progestins in combination with chemother apeutic agents, high response rates were recorded (76% and 85%). hut these could not be achieved in prospective control led studies. The combination of M PA with ethinylcstradiol showed response rates o f 14 to IH%. It has been postulated that simultaneous and sequential administration of a potent estrogen and progestin may provide therapeutic advantage and increase responsiveness o f the tumor to treatment (242). A response rate of only 4% was noted in the six studies of tamoxifen treatment Stabilization of disease was seen in up to 80% o f the patients (250). It must be stated, however, that the meagre response rales using endocrine treatment may be partly due to the fact that endocrine treatment was often initiated after failure o f other therapeutic approaches, as a last resort. Since endocrine treat ment has lower toxicity than current chemotherapeutic agents and seems to improve the quality o f life in patients with advanced ovarian cancer, further investigations are necessary to determine its role in ovarian cancer Additionally, endo crine treatment deserves to be evaluated as a primary treat ment due to the increase in the information available, discus sed in the following section. S trrn u lra fpiorx. Steroids interact wiih specific receptor pro teins inside the target cell, after entering the cell probably by diffusion The complex of hormone and receptor is translo cated ii the nucleus and interacts wirh specific genomic sites on the chromatin, resulting in the stim ulation o f specific K N A and proteins In lhisclassic.il ' two-step" model, receptors are thought to be localized in the cytoplasm in an inactive form After activation by binding with the ligand, the complex is transported into the nucleus (256. 257) Recently, receptors have also been localized specifically in the nucleus (258. 259) Irrespective o f this controversy as to the precise location of the steroid receptors in the cell, it is clear that their presence in specific cells indicates that those cells may be sensitive to specific hormones Information o f the predisposition of tumors to hormonal control is an important parameter for selecting a specific endocrine treatment. To date, a number o f studies have been performed in the management o f breast cancer, based on the presence o f specific steroid receptors Additionally, estrogen (ER ) and progesterone receptors (PR) have been convincingly demonstrated in cancer of the breast and their importance as a predictor o f response to hormone treatment and as a prognos tic indicator has been extensively recorded (260. 261) Breast tumors containing EK are likely to respond to endocrine treatment, whereas tumors without ER respond less frequent ly In the presence of both ER and PR the likelihood of response is higher (262) In the last decade, several reports on the presence of steroid receptors in ovarian cancer have been published (230, 234, 263-304) (Table V III) The criteria for positivity vary arooag the series. Furthermore, differences in reported frequeodcs may be attributed to different assay methods, different rumor types, and the lim ited number o f tumors studied. If the criteria o f the authors are accepted and the overall frequencies an analysed, estrogen receptors are seen in 63%. progcstcrooe receptors in 48%. and androgen receptors in 69% of the tumors. The presence o f both ER and PR is seen in 36% of the epithelial ovarian cancers. Estrogen receptors without PR arc seen in 28% o f the cases, whereas in 25% neither ER nor PR are present The presence o f PR in the absence of ERissccaia only 11% of the tumors (Table IX ). This may be partly due to the fact that PR synthesis is an estrogen related process(305). -v Compared to ER and PR. the presence o f A R has beta investigated in only a few studies (265.269,276.277,282,294, '] 304). In the study of Kuhnel era/ (304). 85 o f 94 tumors (90%) j contained A R . In 19% o f the cases PR were seen in combiaa- ';*jj tion w ith A R and in the absence o f ER (304). suggesting that ' androgens may also induce PR synthesis. The fact that ovariaa 'i cancer is predominantly seen in postmenopausal women might ( be attributed to the predominance o f A R in ovarian rumors, } since postmenopausal ovaries produce little or no estrogen bat } continue to produce androgens. Since A R are regulated by * androgens, it is likely that accumulation o f androgens taka t place in ovarian tumors without conversion to estrogens. This androgen to estrogen conversion is partly controlled by the enzyme aromatase It has been demonstrated that turnon without aromatase activity are A R positive (306). In these tumors, androgens are not converted to estrogens and caa influence the tumor cells through the A R . In these patients treatment with anti-androgens might be o f value. A n inverse correlation between the presence o f ER and tumor growth kinetics has been demonstrated in breast cancer (307). In endometrioid cancer o f the ovary, the presence of ER and PK is reported to be a reliable criterion in the selectiono( patients for progestin treatment (234) However, further stu dies are required to determine the correlation between recep tor content and response to treatment in ovarian cancer, la these studies, not only (he presence or absence o f receptors should be compared in response to treatment, but also their concentrations. Differences in results of receptor studies may be partly attributed to the fact that within one tumor, the cell population may be heterogeneous for hormone binding (308). This might also influence endocrine treatment. Probably, only the prolif eration o f receptor containing cells can be influenced and not the proliferation o f receptor negative cells. Since endocrine treatment is most frequently used after J failure of chemothcrupy. it is o f importance to know the effect o f chemotherapeutic agents on receptor levels. It has been demonstrated that EK continues to be expressed in patients with advanced disease after chemotherapy (291. 295). If fai lure of hormonal treatment is due to low concentrations of PR, 42s Slotman and Rao: O varian Cancer (R eview ) I s i ld*c V IH . Estrogen (ER), progesterone (TR), and androgen receptor ' iAMf presence in epithelial ovarian cancer -v fcmiigaior Number ol HR patient* m. ( * ) PR A R - (%>no no. (% ) Tkpor (263) I m (264) fnherg (2*5) l * o (266) Dapuni (267) Friedman (268) Gfilli (26V) Httnrl (270) . H o i (272) ; KaupptU (272) ***(2 7 3 ) Sledman (274) f a t a u i (274) ctggvixt (230) Bcrggvni (230) Grcwnun (275) Crewman (275) CUIk (276) Hm i i Iio (277) Hoh (278) Farley (27V) KlHncI (2X1) FflcKicrcr (281) Own (2X2) Ouinn (282) Quinn (282) Rcndina (234) Sunk oki(2X3) Schwartz (284) Ford (285) kmci (286) Kauppila (2X7) Lcibach (288) Follow (28V) Spona (2VU) Teufel (2vi ) Tcutcl (291) Viemko (292) Wilkxxk* (293) Wn (294) Geyer (295) Gronrooi (2V6) Lamia (297) Richman (298) Schwaru (299) Schwanz (299) Ivcncn (300) Suiioo (301) Solion (301) Toppil* (302) Agarwal (303) Khnd (304) 8 5 4 29 57 34 8 4 16 25 21 II 4 II H 18 5 9 12 29 2 22 44 37 36 31 43 7 30 39 42 68 21 43 68 153 ISO 45 4V 32 171 21 50 52 113 V9 31 32 20 60 16 94 0 (0) 2 (40) 1 (25) 29 ( H**) 29 (50) 18 (32) 34 (100) 34 (MM) 5 (63) 5 (6.3) 2 (50) 6 (75) 1 (25) 8 (50) 3 10 (40) 10 (19) (40) 15 (71) M (38) 5 (5) 8 (73) 1 (25) 3 (38) II (61) 4 (27) 6 (67) 5 (56) 8 (89) 5 (*2) 24 (83) II (92) 1 (50) 2 d ') 10 H 5 ) 3 (14) 27 (61) 18 (41) 17 (*6 ) It (31) 35 (81) 31 4 (<7) 2 IU (32) (72) (2V) 16 (53) 15 (38) 6 (15) 35 (83) IX (43) 60 (88) 55 (81) 13 (62) 13 (62) 31 (72) 20 (47) 40 (59) 27 (40) 97 (63) 6V (46) 40 (89) 41 28 (57) 14 (91) (29) 28 (88) V (28) 10 (31) 112 (64) 84 (>9) 13 (62) 12 24 (48) 22 (57) (44) 38 (73) 58 (51) 50 16 (52) 15 (M ) (48) 22 (69) 7 (35) 27 (45) 32 6 (38) I I 52 (55) 49 (53) (69) (52) 85 (90) Total 1058 (63) 712 (48) 132 (6V) Table IX . Presence o f estrogen (ER) and progesterone receptors (PR) in combinations in epithelial ovarian cancer. Investigator Number ER PR ER -FR - ER - PR E R - P R - Moll (271) Kauppila (272) Bcrgqvin (230) G alli (276) llih n c l (280) Pflcklcrcr (281) Ouinn (282) Jonc (286) Kauppila (287) Lcihach (288) Follow (28V) Spona (290) Teufel (291) Willcock (293) Wrz (294) Gcycr (295) Lanila (297) Ivcrscn (3111) Topptla (302) Agarwal (303) Khnel (304) patients no (*) 16 3(19) 25 10 (40) 21 8 (3 8 ) 8 3(38) 15 4 (2 7 ) V 5(56) 12 3 (25) 44 H (25) 32 8 (25) 42 15 (36) 68 52 (76) 21 9 (4 3 ) 43 17(40) 68 22 (32) ISO 54 (36) 49 10 (20) 32 9 (2 8 ) 171 67 (3V) 50 14 (28) 31 H I (32) 60 20(33) 16 5 (3 1 ) 94 31 (33) no (%> 5(31) 9 (36) 7 (33) 2(25) 5(33) 1 (M> 7(58) 16(36) 8 (25) 20 (48) 8(13) 4 (19) 14(33) 18 (26) 44 (29) 18(37) 19(59) 47 (28) 10 (20) 6(19) 7(11) 1 ( 6) 21 (22) no (% ) 0 ( 0) 1 ( 4) 0< 0) 0< 0) 0 ( 0) o< n) 0 ( O) 7(16) 3(10) 3 ( 7) 3 ( 5) 4 (19) 3 ( 7) 5 ( 7) 16(11) 4 ( 8) 0 ( 0) 19(11) 8 (16) 5(16) 12 (20) 6(38) IK (19) no ( `if) 8 (50) 5(20) M 2*) 3 (38) 6 (4 0 ) 3(33) 2(17) l(23) 13(41) 4 (10) 5 ( 7) 4 (19) 9(21) 23 (34) 36 (24) 17 (14) 4(13) 38 (22) IK (36) 10(32) 21 (34) 4 (25) 24 (26) Total 1077 390 (36) 297(28) 117(11) 273(25) induction o f receptor* might increase activity (301). Tamox ifen may increase the number o f PR and decrease the number o f ER. From this point o f view, alternating treatment with tamoxifen and progestins might be useful (301 ). However, as discussed previously, in a trial o f 29 patients with advanced disease, no beneficial effect o f this combination could be demonstrated (244). In mtastass or recurrences o f ovarian carcinomas steroid receptors are rarely seen when the prim ary tumor is receptor negative (271, 278. 2K4. 289. 291. 292. 295. 299, 309) Several authors correlated the presence o f receptors with prognostic factors as histology, stage and age. No correlation between receptor content and stage o f disease could be de monstrated (266, 284, 286. 295, 300,301.310). Some workers found no correlation between the presence o f ER and histolo gic type or grade (266, 284, 286. 289, 295, 298. 310). Others reported endometrioid tumors to contain more PR, often in association with ER (234 , 268 . 285, 291. 301, 304), while serous tumors are often ER positive (282, 291). In mucinous (282 , 285, 287) and in clear cell tumors (291, 295) lower amounts o f steroid receptors were detected. Friedman <rf al (268) reported that most ER positive tumors were seen among postmenopausal women with poorly differentiated tumors, whereas PR positive tumors were associated w ith well d iffe rentiated tumors in premenopausal women. The association - - . 417-434 ( l w * , Ilficrenliiilcd tumors m r.r. r , h s founJ lh" i well u r b . i . h K R . n d i T , , ^ ^ ^ 'LU" ,ain' d ER ^ k*`" "<' (4. .w) lio w .;lev; ,;rpR^nr ,7h's1hcr tumors of postmenopausal , nnien , . , PR ,29d' 295) *" receptor presence h j \ ;.l k ` 3 ' rcPi,rted. Steroid .. ...... b e tw e e n receptor conir nr a a i '^a,Ci^ l ^c relationship wilh tumors containing high le 7s'of Pr I Z " ^ ' should be removed as fa, as posstble The use of iU' f ' y ' 'be managenten, of palienls wnh o a7 7 - r " tb v r ,ua,ere^ * - 2 ^ 2 = p s 5 s .x = 2 S ........ . . L ,7 ' " ,,7' `>,C i>' r.` r,i!n,,,c;,B,;c*,f,h^ 'ih--....... -> >n ovarian cancer Despite the fa c t treatment was admin,Sternd S S r * d mp,nci8y r,,g t,,a " 7uh7u7 " of s d e e l , ; ' 0 * W ,O X ,m M '> lu % were'reastded.* men, ha, the leas, im ou -h.AS ^ r j r p" r " * -- r.,, frctiim cni. i lowcvcr. inftirmatinn'ah. T TM '' " ' L'leclion of become useful for the selection o f th U ,t'i:`:p' or conlenI TM V Putte,.is With .trom Itase n W 'P "ale treatment progestins. an, " 7 1 1 . md s,nKa" unt,-androgen treatment might he o f valu i,, 7 h . 7 . available in form ation strongly supp orts^ncedf WOn' e" elm,cal trials to evaluate the , r r P , ' ,' " ,la" " o< and in comb,nation with 1.IIR H analoguM,andr0* ' " S i-hriTj.!diiiSma,T,,g : : " r : r us,ni chr,,ni' u n ^ , ' 7 ~ 7 ^ r ^ 7 S ' o `i:,,,:' cn' horn'on" adequate ,rea,,,ten, an. ,, ^ ^ .'"h ' 7 '" " " " f aging o v a ria n c a n te r rem ain s difficult and ,t present r r <,ues,'Ms * - ~ - s r z r z Acknowledgements cancers ,314. 3 1 5 , J Conclusions 'ra im e n t o f HR-rich gynecolt tgic" mah^'na^nci?' Ahho'ugh'th '" " * * dCJ' h fr m endocrine, envtronmen al 'h * ' ' S eV' de" cc importance in th ^ d e v e h tp m c m n t/n ^ '^ ' aC' rS " * bC < pathogenesis has si,It ,, , hc eiarifLd. '.'".P'" '" l m fo r p e ; 1: : , 'u rn " K " 10 d e v e lo p g iiid e- ....................................................... A " * r t l X s r ,, t e V ^ s uIwctsMcU^ M '"'1' C " " ' Puftdslim Fonds. The Netherlands Mauols cn A nn. d< Koct References ^ l ;/ ;i s `" 'T '* '' DW Tike ep-demtodogy <* c anM a s . ( nooi r _ , _____ _ I M .lkava. j , o p Deekc, DO and Webb ^ ,971- d ScdunO nwi; I`r a j |^ l7 5 *" wlanficaiKNiMd K-n6o5l1n-1J*M6,^.dl72 * " * * * n *7d* 7andd ccuK Vr,|| e r a W ,l-ci.*h`c*o7,.,, * * " hV c l j*1 * ' * ' * * ' TM<Vm4c*m J.tl Sloim an and Rao: O varian Cancer (R eview ) AaMnl J Ep Biol 40 1W IW . I62 I (imitimi WTJ Funb ladra oa cuperimai o van ii lonoaifcncua hoc Aim Am O n . R n I J00. 195 Montai ES: C aiciM fcw acikia a t imfcu p n Br I Caacn 12.' 41 - 41. 195 I M ia d MS and BkkM| CM: D o d o p m ii a t import a ita ral otry mtm m ipptm n op M o iplccp P.oc Sot Eap Biol Med IS: IM-179, 1944 I Co m DW . Hulclipoa OB. Welch WB. Sc.Dy R and R y u KJ n-------- ------at a ttiM C M C ti hak. I. Rcptoductivc eipctkaau w d laMly Malory J N M Caocar InM 71 711-71*. I9BJ M y D. Lilacnfcld AM. Diamond EL u d Brom ID I Aa phfeMotocac tapdy at Ihc icUdomhcp ud reproductive capcneaoc to m m i odIh* ovary. Am I Epidemici 99: 190 m. iv4 CiMftknrV r r . I xhmc EW. Rika MC. Roy S. Rom RK u d Hcmknon BE - la m p a mlatiop* and ovariaa caooer l a m i I I I7B-I7J. I9T9. HildMth NG. ReIvey IL . U Volai VA n ml Am apidtmioiOBK M-dy od epithelial atrdnoma od ihc ovary Am J Eptdemiod I4- M-4A5. I9l. Icambunkl I . CurMaaU W. Gadoanka H , Rovakki M and Wackat - RwydM B: Cam egprrod ppdy od ktpk-nak fartan mranaa cardnumaa Oyacood Oncol II l i t . I9B1. C raim to. Dtacaw Coouoi Cauteri and Su.oM Hormone Snady: Oral oonmoeptiv uh aM die nak od m in a a canora IA M A 249 I5M I W . 19*1 Cram DW. Hmchiaon GB. Walck WR. So d ; RE and Rnapp RC Faeton a (V a li| Ika aaaoaatkm odoral oontnccpaivMand ovanan canon. N Ea|i I Med 307- I0 O 1051. HO. C(Mcn ten Dtarate Coniti! Caooce and ilka NaitonaJ lotiiauu ud Child Health aad Haman DevcIopnMol Cancer and Steroid Hotmonc Slndy: The reduction la nak od ovarian n am aamciand uh orai-eoauacepuv me N Ea|i I Med 1/6 450-655. I Faialla M i Meeaaanl ovuUllon - a Indo* In ovtnaa tuhopiaaiaT l u a l M : 141. 1971 I tradei BV; Thn cdoiofy and provando# odovanu caneu Am I O b w i Oyncooi 121: 772*774. 1971. I Corner DW and Welck WR; DciciMnanU od ovarian ca tto rtak. Il lateromei (ardirti palhufrnetta I Nati Canee Inai 71: 717*721. I9B) > Annefa n |F . O'Fallon W and Borland LT: Eiogcnow ucatrogea and ovnrian canai Lancel II 869-X). 1977 > Wyndtf EL. Dodo Id and B utto HRR: Epidcmialofy od enneer ad (he avary. Cantar 21: 2 170. 1969 Wciaa NS. I yun IL . RrMmamunhy S. Dietan SE. I jR 1M and Dalam IR: Bdcmoon- moeptlv etlrofea u and th wcvncan od mali ennocr. 1 Nari Canear Im i od 95*98. I9R2. Hoextr R. Orar LA and FianmeM Ir JF SitRsoctirod (dtatkyMUhcvtrol) u d Ibe rtak od m in m ancer Lanari //: 511-514. 1977 Walkr BL: Tumore od Iemale odliprlnj od mica capoaad pceoaully lo dMhyMiitac- rrai I Nati Cantar inai 71: 111*157. I W Liana AB, Oretnapu Lcvcnaon A . Layde 7M am t Rtak odbacaM. wenne coapm. ami ovanan canoer In m a t a rccrivinj medioaypeofcauiopc Injacliona. 1A M A 249: 2909-2912. I U Lucana# CH. Eliutotyndunenrodihchmuovnry: A revtn JNata Canon Im i JJ: I4dd-I61. 1974. Croma DW . Walch W R . Hutchtaon GB. WIMc W and ScuRy RE. Diclary animai fai In reUlto# io ovanan canoer ntk Oteui Gymcod 41 U H M . 198. yen T. MarahaH 1. Oraham 8. Mcnlln C and S u *, M A caac oontroi Mudy od dmiary and nomtaeiary faeton la ovarian enaear I Nall Canoa Ina 71: 481-484. 19*1 Cramcr DW . Wdch W R. Sodly RE and Wojekeho-vl CA: Ovarian ronca and U k A cnaa-oontroi Mndy. Cane 30. 172 174. 1982. Roe FJC: Conirovany. coaaactic lata and ovarian canori Laurei II 744, 1979 Cranici DW . Wckh W R. Cacatili 5 and SeMy RE Mumpa. mcMictae. manosame. and ovanan canea Am 1 OtMct Gynoooi 47-. 1-6. 1983. Lynch HT, Albano WA. Lynch 1F. Lynch FM and CampbeB A: SuivedWact u d manafcmcni odpnlicau al 61*6 pcncnc nak lo orano# cardnoma. Otamcl Gypnooi 19 589-594. 1982. Rive MS. Mcllbn a . Twkad. Y. Naaro F. GrcenwaU F aad McPhna ME: Faaiai ovarian canea repnry Ofcnui Gyncooi 64. 195 199. 198. T r it i 1 and Bocakowahi R: Oratetene* od iitmon In dytfanelic gonadi. Canon im -1110. 1967. B|nkKota, E Btood (towp dmuiburlon In uoman rtekovaia cannar Im J Egidemidi 11: 15-17, 1984 Reni SW and McRay DO: Frimary roncar odIh o w y . Aa nilyela od149 roana. Am 1 Ottalei Gyncooi 4U H M . I960. FIver MS. Lek S and Bariuw II: Frooperanve and hMmopnarKa aulmilon la ovnrian maUgnarom Otmai Oyneeol 4f: 112-115. I9M. Reanady CR and Ootdoa H: Ovanan roncar: iha ano yen* n p trU ta od a U rta genomi hropiuL Br J Obatm Oynnecol 88: 1186-1191. 1981 FuBar AF and 0 rifiliha CT: Ovarian canrot udmeia - mrgkai Mcroctiona Oypeooi Oncol 8 1*110. 1979 ' Va# Naell Ir IR . Demataboai ES. Marno# MB. Ooy EC and Fnvhk El: Btaehamlcal marken In tha piaamn and 1untori od pariena wkh typatoo(te malitaanrim Canoa 4 t: 95-301. 1981 Va# Nagnl! Jr IR . Dematafcon ES. Ony EC. Shatkcy RM. R ayW a Fand Ooidanbcrg DM: CartinonmbtyrMc wllgen In ovario* pitMUai rjw a li mi rarrinrimai The pr* nome va te od m aro ind acri! piaamn deima/autoro Cantar 2: 35-2140. 1971. Khoo SR. W btuka S. Io n a I and hiaduy E. Frodkrivc vate oderini roroaocro btyonac aailgrn tenia in lo a g u ia l o t e e p odcairian ronca Canea 4S: 2471-2478. 1*79, , Ootatetaag DM. DeLand F. t e C m mi It e od rodmdaMtnd la t i t i * 11 10 cardnocatbryonk uligan fot tha tauem n and toroliuiion ad Marna amari by n u m i pharoronnint N End I Mad 298: I184-IM8. 1978 Van Nate* Ir JR. Rna E. Caper t m mi: Rndt e mmonorteucttei ad p riM iy and meiaauiic ovanu canon minf redktebcted anijhnthra to cnrrinnamhryoMr amigen. Canea Re* 49 502 306. 1980 BKarucharya M o d Bario* II. Ovnrian lama* a r n ic a Canea 42-. 1614-1420. 1*7. KnaufSnnd UrbachGI: TW dtvehyienl oi a dowbic-antibealy r>i9rlmmimnmaiy tot dcicamg ovarian inmof-akkuuMicd amigcti liacnun OCA in plaima Am J <)r*ici Gynccui III: 7MI 7*7. 197k. Baal KC and Rnapp RC ImmunokitH- appro*hn 10 ihc management uf tnatun cartteom*. Semin Oncol II. 264-274. IVX4 Rnuf S. Raima 1. Hclmkamp BF Harwell l.W. Bcccham I and Lord FM Monocionai annNute a p t M human uvarmn romor aaaociatcd aaliftnn N B /hK Froparaltcm and me in a rodaoiattmiinoaiany foe meauriii| NB/7DR in eerwn. Cancel Immunol lmmunroher 21: 217 225. IS*V> Bmi l l RC. Rlug T L . Si John E rt mt. A radioimmnnoaiaay m ini a ammocluoal antibody lo anuiMoe Ibc courvc od cpilbclial ovarian cancer. N Engl I Med W : IU -M n . 19*3 Batl RC. Rlug T1. Schaciri E rr ml. MotMuring human ovarian cia.t iw w with a eombinaiMM ud CA 125. CA 19-V. u d cartinoemhtyunic anligcn Am I O teet Gynecol I49-. 551 559. 19*4 Bruachi FA, Bnrhul F. Rapan C. D t Roten M, Irmn O and Riaucr f : Im p it e iu l atudy od CfcA and CAI25 m m u n a cancer (iynecui (tarotW 21 1-4. 19*5 R u b . H B. tropicrM DR. Rilpana SI. Myn MB and Hum A R.4c od CA 125 a. iwmoe malkcl in ovaiian cnirinuma ObaiM Gynecol 67: 471477. |hn. Rhoo SR. ilH iti T Webb kdi. Dickie G ). Rcanicy I I I and Mackey EV: Fiokctivc value od acnai CA 125 anligen tecta in ovanu cancer evaluated by accond-look laparotomy Eui I Cancer Clin Oncol 21 745 771. |97 Foeta 1G. Fctcn D. Van Mcgen V r* ml: Monoctnnal ambody againm t e n ovarian luttnor-ataociaicd aiuiactu J Nail Cancer Inat 76: 781-791. 19M6 Rakriown HF and Epcnctoa AA: The rota od monocionai aaibodm in tumour diagnotai. Cancer Treat Rev II. 241-252. 1*16 Grammka M. Bntlun RE. Staepheid IH rr ah A proipeanr acudy od 121-1 labeled munrxlunai ainiNvly imaging in o va iia cancer J O n O m i 4 7X1714. 19*6 Hunter RE. Dnhcny F. Griffin TW er of Ik e od Imhurn I I I lobclnd OC 111 monoclonal naibudy in (be deterion od ovnria cman Oynecoi Oncol 27: 123-117, 1987 Houghton AN and Schcinhcig DA- Monocionai anlibodiea: pocealni appikaione 10 Ih t ireaimeai at u a a i Sanaa Oncol II 165-179. 19*6 Andulf E . Svalcmua E. and A Mrdr II Ulifavonogiaphy lor early detection od ovarian carcinoma Br I O b *at Gynaecol . 12X6-12*9. Itea Metre HB. Fairant F. and Cite T DiWimiKHs ud bemga and malignaart ovarian cyma by Olraaound Ri I (lbalei (iynarcul *1 893 99. 1971 Rcdunnd CR. Mental Jr FA and Warki ID: Prcopciaifv tomography od oatagnani ovanna nenpiatmt Am I RieiMgrmi/ 117. >1-82. 19X1 I .atala JR: Ncwci dugmiauc apprnnrhct in the evalnaliun od gyneooiogic mnHgnanctaa Obaiei Gynecol Sur* ,J7: 417-44. 19*2 Bart JR. I:Jlii III . Rupecky RR rtmi. Aaveiwncni of primary gynrcuiugic maligasancia: Companion ndll 15-T in ttliw MRI wiihCT Am J Rocnigc* 141. 1249-1257. 19X4 Muaumcd R. Bandi A. B<dn G o a l . Lymphangiography In paliemi wxh ovartaa cpilbeUnl cancer. An cvaluarioa od 2X9 amacouerive u m Cancer F 1444 1449. 1977 Madnioch FR. Thomson RR and Barbaric Zl.: FticHanaoua irantpcioocwai lymph node btopty m a mean ed tmpruvt| lymphograptax itaagnuari Rarfadogy 111 647-649. 1979. Van Ordcn DF.. McAlltaicr WB. Zernc SRM aad M o d i JM: O van u candnoma The problem! ud waging and grmfeng. Am I Obaiei Gynecol 94 195-202. 1966 Tubat JS, and (iriffilht CT Managemcnl of ovaimn enretnama. Cnneiw ooocepet aad fa u n protpatt (rwo p *m ) N F.ngl J Med 294: 8IR82J and 877 182. 1975 Richnrdaoo GS. Scully RE. Niknu N and Nelson Jr JH. Common epuhchd cancel ud itac ovary ilwo pnrta) N f-ngl J Med 112 415-424 and 474-4X). 1985. lacmolionaJ Federal* od GywcciUogy and Obairirica Chnngct in dcrmlilm od danroal Waging for carcinoma od Ibc cervu and ovary Am I Obatel Oynecoi IM : 261-26. 19*7. Day ir TO. Smith IF. Diaymnri and waging od ovarian carcinoma Sctee Oncut 2 217-222. 1975 Father RT.Faker O I and Wilbaakt GD: Cancel od ihc ovary. Survival wadica baud upon operative therapy, ehemorherapy. and radrodlcrapy. Am JObwel Gynecol lo t 87B8M. 197(1. Wrbfc MJ. Decker D O X u a ay E and WUisaim T l FaevuiInOucncug wrvivai in Mage I ovarian cancer Am J Obiicl Gynecol 116 222 22. 1971. Sotbc B.FrankcmJnl Band VerciaB. Importance a t tuairitogJc grading in ihc pogm< of eprthalial ovaiinn carcinoma Obwei Gynecol 59 57A582. 1982. Malkuian Jr <iD. Mchon I II IJ . O'Brien PC and Gieeae M il: FiognuatJc Wgniflcarwc aI htaaokigic claaarflcaiton and grading of epdittachal maJIgxancWt od itac ovary. Am J Obaiei Gynecol 149. 274-284. 19*4 Baiter HRR. Sommm SC. Snyder R and Rnon TH: Hiuotogic and auetaat grading and wromal iiactnm* a tetaect h r progntea in o v a te cancer. A m I Obala Oyneool 121 791-*n. 1975 Onota RF. Garvin AJ. Coma f . Simon RM and Young RC: Advanced o v a te cancer Conelnttam ud Mwodogtc grade with nginn i 10 therapy and awevfvil Cancer 43 572*581. 19*0. R D. 11uiop TO. Spinei J. U R khc JC. Y n ^ N rod Boytea DA: Ovarian ` ualyaia od pmp v w r factor!. Obetet Oynecoi 63 264-2XJ. Signfkann R. A te F and Godterg B: Faogiaoaaic (acton In mahgnaa cpiitactel ovarte lamon. Gynnrol Oncol J3 J70-J80, 1983 FUn WR Ovarte eptltaehal n n w t od bocderiUe aaaHgnancy (rorrinontaa od low malignant potential) Hum Pacha! 8. 541*549. 1977 Ovanan Tumor Panel od itac Royal Chitage od Obmeiitean* and Gyaaccoiogtei Ovartaa epiihdiai lumown odboederite mlignaacy pattanCngical (tararea and awtetn tta ta Br J Obmcl Oynacoi 741*730. 1983 Rmnell p: Th parbaiogirol .......u n i t od ovarte noopiaama I. larteacuna to the common 'cpuhehaJ' fm oun and analyte odbewi^ pdlhchnl' inmoura. Pathology / /: 5-24. 1979 Nfcnri N. Survey odrtlnirnl hchater odpaiieaa with borderttne apMhahal tumor ad itae ovary Oyoxed Oivccd 12: 107-119. 1981 Julte CO and Wooitauff JD: Th bioiogic behavior od low gride pnpilary acioua enrianau od the ovary Obaiei Oynecoi 40 *60 667. 1972 Barohifl D. Helkr F, BritMuwekl F. Advaru H. Gallup D and Part R: EptiheUaJ A N T IC a NCER R b S K A R ril * 4I7-4J4 , K riit^ rn : 'rr, r;r * "*'*um >Jl 8 < i< ilS .O (iilM M < T h d U w nih . i, . , s % s p a a rM i H ir"~ " ~ " n> >n ihr a le p n *,**, cprtlKhjl m * an w iw koi |iJ. | hy difieren oaiteu.m.^ "* . P*" ** vanatehry Pl>"a K ll> i W o iiiK M k r k .i. " " 1C i l l I h r pfogaumc l a U 04 inda M l ll r J h , |w l ,, fc(i , * _ ,, ,. U IW iky .4 cellular d n a <unicat.<ev.rv, . .* * *v*U- AS Mi ia iictvo il M lIN ** W 7hii ivm Cancer R K.slt nN M |'|.C ,- hC|n M c , |h |( || . ... ... * ' mop* pr,.*.* U u in advanceJ . jn JB'r A * * Jr l t '" C* i r " * * P*"" ^ < * hk.*Und L P r id , C ,,,0f, ,V J ' 7 ) . IM 7 " " .f; ' 'r r .~ .,, Glcvo ( A lalun OC. B. u> H *" " " AiK.iK.YdraaiLva lira I b l mlcnwvr h. ' ' * * d ` HJKam Rr * * * * * O n ic i Gyncvri JJ | W i , * ttemo.hera,,, KcuoJ took i * * * * ,n<nl tie m u i.e cftfiecr Am I rw '" t a iii. 1 , , , , 0 ^ 2 1 . reiitfm w .-- k m / rtwm.Mlwi.py u H ,n a ti. . . J ,, 0< *np'OKiT Ki dcieraune ih , -- k . . I S tu n CK`f . J d lim M i . . . , , J . J T ' M* J,, U K" * - i m . . * " " KJ i v I77. a- * . IV, 2 T " ` r M " ' i j 'V ' S5T M r - . i l (JKK,,sI-------- iv u:rr z ^:Tk* m w - -- ...... laparwo,, . lhc . ,, ^ . * 7 ' * ' ? *!? 11 Wh,mw j MA W INC IV) ,iW4 ^ vcB < ,, ,* ,. of |h< '7 ^ ^ ,W Vu* 1 S1.. M ille r V. Caleauf a it J - t - , r * r M s ^ f C* "" r " a a : i S T ^ " " " M l H in lte kp h ,!) MM f u T " C.,n M ' 5# MOh-77|J. 1 ! A' M l-5m . |H| rT t P,,h* ,fc* " " >rflh*o*iy.O ,hrC, ' '* *u'*cry lor n cyiracduc " * * " * -- irid u u a A l A * a . M ^ l K ^ l h ^ ' 10< * ' * ,h*< HMMi m h a u l. ,,1C . ' -, - ,U l.? * ,h* *I'W* MiMMivMy ri IT inaiv.n r Sj ci>ma*'O ni wcol-/iJw25-1i-w2)., " B S T `. S r * <1" " S" < a . . . . I .. "."^n:.;rM,r 's . f_*_"__** t ' TrVCH Oneri / k _ " .C Y I Sema IJM V , IV . - ......................... - ..to V . 0 ,, *-- CSNoce5. 7J v2JN1S2WI. Siene i 5 ^ r ,< ` v* A ICV^m - r n , 1 S S S . S s r i " , * " - . ^ armeni o< cany cano* t t dm v .^," JAM A 7.Vi RlTJ-AOh. IVhi "t - ~ ----------------- - ^-PMteM fcrtuHyfegartyouaimtec*,, * <>tB ^ ,"*d 1 Inodenot o p^,. ,, in *fhcJul ore*.,. ,d ,h* v>. ,, . .. f * K * ' ^ p * wide M e u n .ic . ,, ' l - \ : % ^ ~ < & Z X Z 2 Z-Lnm,, ,, -- a. . LD. IYKA >'WWM, end m ^ U d ,!, (nie, G>ncc.d fcrA M ` M rib.hM h-.H A , . , a v ,, . ,, m ,M ,l.l l . 1 ||llIlin . l l ' r \i,< ly lu Jan.( j m ||| >n c^wc<J. f t W iH J C . T W V , in . Za i'KZ-i..nhrT, Vt - . VOm..| /.'(Suppl,:IVj" J' ;Wt ltt - - K r = : S 5 ? ' V * w s rnt4J111.ua lue Siete 11.IV U..W-V fc0 ,nJ lnncr H A H o ta a it 1 T iIT a S C l ^ i S S r ' ^ -- ^ 3 a ~ = & " S a s"Atiiuei ch<auiK((tp7 Carnati U: OM O. IM Voan, R0 Oieat.fcc,.p. , j ,.v,, , . , 2n7-2R> |V75 " * (iteci, p m and prcicM Stana j. Am J ii rrK.e. r V v n T . T j'/ " ^ .'^ ^ cJ ^ fc,W| " M**c 111A| . * * t r .4 ihc ..............n...ti Nuig Un. > ... a . . k " " ' PII. Naim.,i m K k tiin t s v i ' ' ,lx " < <>m .A /r l i a n a ism1 .. r t .p n r c r ,,,, ! ^ .t r s r ...... .. . - J i)0 . , d n ^ a n . k Hr H i Z c , M' N< M l Sapk-, j, i> ^ ... . i ` *** u Z C i S I J. mavfcnma, MM. Wheri.m J| ,, l-V -r.t. y m td.to.cd ...ti.ta ..m -7 . M a i.n .<t.K Jlurl '-' I K.acrc,.. M . M ta -ta C T C,,u,,, l y ^ ' c` ' k -k tcU -Z iy rn . v Z n X m , <7- 47"- IWJ " , -- M.kv le JJ The .......... ,,h "*'* WJ. (Ielle, PH t n T p '* * i r t w Ma*ai l.a .............. ...,,ta .tramTv. r u , ^ awg^al 'JJ Bcek JS. Ilac tc i Np. J tg tiw I D K.h r T . u n tU l /V ` M-M IVlU lapa..a.l in d tp . III %L,J M, . ^ r ^ * ' ^ ' 1 * MK: Snund look diH-ew a tiu t Orner H y n c .r n t *17 >. ' iv ^ t >hn`c<l ''Me vociaied - ilk i' Tien Rev I. W I III |v74 T V c h tm .ih t.tp , ul.wanna IM <>rccoe M il Hm-c Jr J|) , ,, i r l ... (Xvicyln leakem it aller Ihcrany nuh !**"* A " "J* AJ. Acme nntym ... s w a s - , a 5 5 5r"" ^ An evala,,,n 11IW4. IV7V `'7 1* t o T t ? ' * , l h t , ami i l . Mgr MP He.amallrylmebmta: meni .4 tiran cancri Am J OtHici Oyneori /JJ i:r:tr :y r `nunt..... ........ ' j *. m 'e7 T .L I *'"* * ... .t:L"si "r ` "*" ". r -- 7 r , ' p,nt,H t " " " " d K k k m ^ m M w ^ h ll'f l ` tv! n ,fc* ' ,rT: A ski cupialin in advanced 0 . ^ 7 7 1 ,W C jl**n AH: -- 1 Brun**U VHC. C r ^ V . I) o T C , i ' ' i '* ' n ` " 'I'a -m t faucet 11. 27n ;;i . 1 , ' ^ ncaimem ul advanced Ovaina in T.TcvmM >n777 `,'t 1 ^ ",0"7 *" 'M M K> M ckiflnt JA H tH e /u n L ^ ' 5' lW * Warrtinm1*i4n ,nhVc ipnlaVnineidv ahjitnagcmem ,,1 t^dav a.a,.v.dj .tr"u*rai amma,nMa AMmaaJmOd-Uluiuckt . . . lt ^ j . , ma s a s s s s r.jrr? ^ mclphnkiB In advanced i v u v . i . r n . ... sxteassa 2 .' IS ," "*" * 1 'iT.'SErtg*" ^ "'pkalan n atlriamyete- .M" h K ,,,> i f *ml S r .,, * ,* CM A n tl.u . d - Mvlc-.h.i.l.Jd kI.vdai- lt|x.nH.aay mrpti.*enMu W. ,, ak . a JIM!#, j#'*w , *,Ar '. J OPKrm,",'Orynceaenc*e,ra>a:l C ., lt n j I j tmJ ticr.htrr.VT Dm Ovarran cante, it.I.nt.gnc Iwav./ al icoind h>k lanar,aurnv Ot,.henna, DM Copeland U . W h t . L r j rcHawdcn .a j . #0ccd 1 M .n g ,,m ,C .b u l.,U Muhe, p ,nd BrH^.1 7 ^ 1 , " ,m ,n *" J"'*cnCa nnh no Z - V ^ * ,9m < kie .m ln td * H2V-HJJ. imi ................ ..... r ,, v ,, . i `"J Pi mU1 I - "*7', ^ 7 * , l f A M ) vtri*" end. . ! RJ. S ,,a y v n h ,, A ^ V "' '" r " k - C . ,, I , . ,, U p U U11K . *.... ,,,, ' ^ `T ^ srr-'r Vi ci -.J Sloiman a n d R a o : O v a ria n Cancer (R eview ) iaaw m > adrUapeia C--m i <7; Z222V4. m i. Dndw DC. RewitiR^TH. Mattaaiaa Jr OD. WeM MJ. JaMertaa JA and b M o a m III IV r n r i u CAtanom o L te i OjaMtoi OO a i-*K 7 . IVC R` * r" *" *u e I Barker OH and WUiahe. E Randowwd mat ctanpaiar, cM ocftm or * tik ton-dotc enptaua. rtOne--ifcwnl. ta d 1 m m carn M a a L a m tl /. 747-750. IW l I Eftfbc* C t. Plt k o n L . Suhaaaa FB and B kn iag J: Tie tim e m U< advtaoad epkhckal m n t cancel atng otpltika. adnamron. and cyiotan - iKe Indiana UaK ctiiiy ipcrlcnM O w O hutt G ra te d 10 O i MJ. |R) I WBhtmt a . Mead B. Aiw dd A. Gfeen J. M m t R and W t a d o a t M d wanaa carcinoma. la kamd oomhmattoa. Cancer dV 177 |7*J, |VU 1 Ccau PF. I r u u t M . O w n S a t ! A landumiaed trial rna ytrin a cmplann plm CTdepkmpkamidc > m u citplaita. doauevlNoa. and cyctopKotpAtmui la aaKansd evanaa r u n I Clai Oncol * V05 V7I. I M . I Vogt SE. Pagano M. Kaptaa BH GmeimakJ E. Anencan J and Beane B C a p riM Meed o>mbma'ion chtmdKcrapy toe advanced ovarian canon H if overall r rate till cwairvc poicM ^I only m v o m a amb amnN 1 2024-20JO. I W I N djl IP. Tea Bokkcl Hamm( WW Van der Barg M EL tt ft. I U*9J ( l i t t . CAF r f CHAP-5) in i. Lanari li. 59L40U. 19*4 i Bnacknee IIW . Cokto a . U ddbeig JO t i ml. C ^ t a iw r padgnoan ad palnaia artA poorly ddWrratlated ovarian cancel ItfJ: fc55~Sl. I W r R ia l S. Anieby SO. 8 m S and Faka Z. C t* montarmi C H AD ChcaKMhtrapy. accond look indacuve wrgery and boia ahdonunaJ ktmhallrra fee age III svimi caranoma Ploc A Soc Cbh Oncol 2: 144. IM I I EdmonaonJH.McGocmactGW. FWmlng rReW Comg.riaoa of cyeJophoaphamkfe pfeadaplalin m a i heaameihytmalamina. - r ' - r ^ - T * " - M *' tM l k m d triplain In csmbtaaiioa at kaluaJ chemodmnpy k e auge III rad IV ovarkM cantai Cancer T ra * R i y t 1245-12, IMS Tra Bokkcl Hataah WW. Van dea l t t | M tL Vaa Ooncrom AT. Vemockcn JB. Vaaferi C aad Route.deal I A p t o l nJy of CHAC-I C ara T ira i Rea l (Saggri. A).T7-I2. IM5 1 Orala R f. Co.ria BJ. Jacob J. Wcakey MH 0*dm a Y and Y o m * RC: High oae triplaiIn la kyperroaic ulm c Ana fe ra Med 100 IV-24. IM 4. I R a tta GH and WiHahan E Uh of fegh-doM ondKMotofenmmfe piafenm (II) (NSC-IIM7J) felkratog U tla e o f ptevM cricmothertpy fee advanced carcinoma of dm ovary. B i J O tw cl Gynaecol U 1192-119. IVRI I Vogl SE. Pagano M . Davit TE et ml; Hat i mcihylmcfeminc rad rttp itim In advaaccd oviriaa cancar arici fadre of alkyriungageot therapy, f arane Tram Rep M : U U 1290. I9B2. I Beriooa KE. Pranmatiei M W. r * T and Kennedy BJ CiadkAko f e r a ( II) and J 42 2027-200. 1*71 I Bcraath A . Andrena T. D iton R af. Long la m k fee. Adriamydn. Cfepfeianara) m CAP (Cycfepfenphama r i. AdnamyOr,. Cm plaimira) In alkylating raaknnni advanced ovarme cardm na. frac A Soc Cito Omni I; HO. IM 2. earanoma Frac Am A h Cantar Rea U : 147. IM I I Vogl SE. Pagano M . Rapita BH tt W: Coi M an ta canoa ni* hetemeih ylmelamina, cteplatmam. and donorafecriitftei (riferc al prior thartpy Obuei G ynrari 5* 4J5-h40. IM I ' Albera DS. HHpcn RO. Moon TE. Manlmbem PW and Rlvkla S: Confetaaima chemotherapy fot tlk|feiu-icM<a<d ovattai carcinoma A peclimtnary tepori o l a Starthnrai Oneuogy Groap Urti Canon T r it i Rcp J JOI JUJ. 1979 I Rana R. Harvey H . Andrew, T n J. Phaae 11Utal of cyctuphoapkaaiude. bcu fluih ylmclamine, adriamycra. and craricMorudatMineptolmnm ( II) r m h t t M i che moOerapy in advanoed cmanaa earanoma Canoa Treni Rep kJ: J07-J09, 1*79. 1 Roteknei HW. Cohen G l. Deppn 0 r i mi Tieatauni of chemutricrapy-rctriiaai dvaooed ovariin canoa M in a oomMamlon of cydopfe*(*am ira. keamntlhyhnrla mine. adntmycm. and tradammantdichkMoptaiInum (CHAP) Oynacol Onori 12: ri150-15). IM I I Bntckncr HW. Cohen CJ. Dcppc 0 mi. O vaia* canee: Scrindala modificaIkon rad fe. heaa da-plana rgimen ( O lA P II) Ptoc A a Soc Cito Oncoi I: 107. 1943. I Orate RF: Pharaaootogtc revcraal ot drug tetraanae m T in ta aten l e r i t Uncol 12: 711. IMS n TC. Youra RC. Xterrier Jr RJ aml Orate R f RcvoiaJ of by vaaparad fe human ovaran canoa Sdenca 224. 9M-M4. I Orate RF. Rugan AM . Hanukoa TC. K k tritr R and Yimng R C Varaparad pkra adaarayckn m rtfe a s r y ovaran ma car D triga of rtm ical nial on baria oI ra r a mi of --------------- . . . ----------------- -- ~ - ~ Re 2): B2: M l . 1*71 i Myen C: TW ma at Sanala Omari JJ: 275-2*4. IW4. . M art mat dm paritornai cavity Cancel a CD. Dcniino PR. IkaAt Rev JJ: 21*242. 19*4. X. Mark M A rad Cohen CJ: Carimira rmrapllra- Potemffia CE. Hovmll SB. Ma.kmaa M. and l n W E: 1 peritoneal accnaa J Clm Onori 2: i m i , IM . O n ia R: lairapeviioneal 1R tm ia kim i) In (Ha management of m Onori J2: 75-BO. IMS McVie JC. Te Bokkcl Htrinlnk WW. Aanaca E. Stav FreakJm H laiiapcmoncal chemotherapy m mtaamai prviccuon Proc Am Am O in Onool 4: 123. IM5. 201 Cohen CJ Therapculic tlaging Im: Brucine HW and Cohen CJ (o ri). Ovariin conce ne* apptoachr, iiH caiic laica Momuova. NJ. Saeber ri Mclaiyrc. IM 4. pp 17 *1 2112 Honcli SB. Pltifte < F.. Wuig WE and Oldmn RA. Iiuraprntontai caathammlntdKh loeoplaut*um .ih ayaieaMc itutnuMaie punniuo C arn i H o t i 1126- 1* 11. ivnj 20) Creatann W T. Gali SA Bterang JA ri mi. OMraonamunorhaapy tn die managrmcni ri paliviary Mtge III uvarian larirann A liyc*rp< CWokogy G rw p Uudy Conctr Ticat Rep a i )|9 )2). 1V79 204 Albert* D.N. M.tm TE. SUphenv HA ri 1 Hamhumicd Mrady ri c te n m n .a n rh ti apy fe advanoed uvatmn sar.inuraa A ptcbntnuiy repori ul a Vwihncm Oaacohigy Cmp rtudy Canea Tieni Rep 4J: 125 .1)1. IV79 205 O rd a SE. Rmcnahern N. Rtem JL. Ixthcl S. Pino Y Tue.e* J and Enmgti O The Iniegfattun ri ncn ihtiapmr and rathaiion in th taanagcmcnl cri ovaiian cantai Cantar 44 VNMkh. 1WI 206 Beai Ji RC. Rack Jh. Obaiu H rial: lairaptnnmc.l lamutauihciapy .4 buamnovaimn cuciaura* urth Cuyrararnum pravwn Carati Rei 4.1 IJ*5l4l)|. 14J 2U7 Bach JS. Ilraria NF. LichicnMaia A. ci al: laiiapcruoncal iccumbumni a mierte1un fa -vaKafe- immuraiiheiapy m v i p ili cprthehal ovari* cacci A Gyncuringic Oncufegy Group Mudy Carati Rei 41 4447-445.1. im i 204 Wclande1 CF Uic ri tnurferon fe he irtaimeni ri ovanaa canea ai a uaglc ageal and m cafertunalaon mth cyiaauuc divgi Canee Sv. fcl7 t.iv |4MT 204 Chapraan PB. Hekei T. Cabnfevc J timi Aprirne I pikd udy cri iau apriduneal ill.-l fe ovanan carati Pira Ara Soc Chn Oncoi 5: 2). Ito 2)0 llamburgei AW and Salrara Sp:. Pnearay bumitay ri huraaa luraw Meracelli i e m 97 46l-fc.l. 1977 211 Sclby P Buick RN and Tafcnuck I. A cintai appianai ol ih .huanaa turno Miai celi aaaay*. N Log) J Mcd W 129 |54 |9rt) 212 llanauakc A R. lUnorakc U and Von litri D I). Rcccal rraproverneait fe ihc human hamor efeoing aiaay tot teraaiiviiy leumg cri muncupUtaic agenu Era I Canea Clm Onori 21 402feftS. IM 7 21) A fee ill DS. Salmun SE. Chea USC r i r i. In vano cfemigctuc aaaay tot pecda.rm| rvaponac taf ovanan canea lo chcmcMhaapy Lanca II 140542. I9fe3 214 Wctond CE. lloracBey IID . and Furiar* VW fe vrao chcraotherapy mufeg ul gynccologK lurauev Baan ira plantuag Ihrapy7 Am J Orina Oyntari 147: 14*145. I*J 215 Ile nera CJ. Pclgnra O . Rirkch WJ. Dcluuync FMJ and V.myi GP -Viebte' lutnra calla m pnallrieiapy burpay tpcciHcM A plnlial applnalnm .4 human (nun cfenugcntc eli d ilu Arca Paihri Lori Med lo 7 41-M. IMS. 216 Rygaard J and Povteen CO: IteWrouanaptonuina ri a humaa moNgnaaii iwnnui ui 'niade* mie Atta Patte Mici cri.4 Scarni 77. 754-76. lite 217 Macnpoa J. Kangm L and GrOnaoua M Rapante r i ovanan canee. iu u m tead eytotoitc ageau fe rie rari,cari capante aiaay: pari 1 Oriate Gynecri M 7t* 7|). JM5 214 M4anp4a I. Sutuenal capaale aiaay at a ptadkcio r i cimieri icipuaa r i ovaiian canee. IO chemotbei ipy pari II- Ori*lei Gynecul M . 7|-7I|. im j. 219 Ograra H. SckigucM F. Taiuka N tt mi. Elicci ri annliteincriyin m e n u on human concer amie rame A r a u a i Rei 2 JW 5 0 . 1942 220 Dmpte EH. RicIuwmmI RC. O'H aia JA andCougJHmGT Carbuptetm a poacwhaira r i radutwn Ihcrapy. Canoe Tirai Rev 12 (Sappi A). 111-12. I9HS 221 Fulter Jl A f. Guy S. Badnk GP atri Donate* PR. MuUafea rahrtai.ag wbrianca tehitun ori.riiy growib r i afeunuovariucMCinoraacclllfec. JChn Erfecruul Mciab 54: 1(151 MISS. 1943 222 Donatew PR. Fulter Jl AF. Scully RE. Guy SR and Budiri GP Mullcrmn intehirmg attritiance inhfe.lt giravlh r i a humaa ovanaa carati tn aarie auce Ara Suig 194 472-4141. lvl 22) Jorici B. Fingendone Mi Ih tiraiavem .4 advaaccd mahgaaa r.a o vn ol rieeau ovary and irtcui Bi J Crarai 10 2W-22I. I9fc2. 224 Valga A and llcraiktea I ENca r i |7-alpha hydn.iypnrgciier.ma 17-a-capaoate on '" relvu audignaram turaci Gynecri J l 3l-*2. IVM 223 Waad IIW C Ptv^eu.^cii Iherapy Ira ovaiLa caiifeuraa Br J Otuiei Gyaacort A). 355 559. 1472 126 Tmurthy I Prugeiiugca ihcrapy ha m u u a caKynuraa Hi J Obatel Gynoccri IN 561 .Vb). IW2 227 Malkauan Jr G l). Dckhcl IM I Jragcaaca 1.0 araJ Wtriri MJ Kvaluafem ri 6,l7a- dimahvl ri dekyOnpcgevinone lari KcaiaacM .4 rccurrcH aad inciaaaalk (vnccuigk udigneray Ara J Giraci Gynact UH 4fc! 4nV 197 221 Rauftman KJ M in .p a x u r i advanred manca carvi*.mm Mcd Chn Nraih Ara JO M.1-RV,. IWfe 229 Malhauna Jr G l). Dehkci IX . J.agcnaca ET) and III: Med..rpugeucr onc acc.aic (ut he ncairaeru >( ni4aie and rceuncai ovante caroatra* Cantei Treal Rcp J 915-914. 1977 2)0 Bcrgtjriu A. RuliaraJci S atri Th.ricH J: A Mudy r i nrrogea airi paogetelone cyi<4 rccepu.i U iu n li.lK W in ricn.gn aad aabgnaM ovarlan lamraa aad a irvicu ri Maligna ovuiaa lamanin licalcd wrth medroty progeaterone acelate Alio Ohaiet Gynco Scorri Vagg IOI 75-gl. im i 2)1 Slayum RE. Pagano M and ficee R II rrofrrtfelheiapvferfcdvMCcd ovaiian canta A phmc II Fmirra Craapeiauve Onouhigy Group trial Canea Tram Rcp*): 995 496. MI 232 Mangioni C. Franoeacfe S. La Vcachia C and IT Iacci M High-doac raedrray- piugcucroem m ( , ^ j , g g^JT*' ** * 23) AahoK. Fedenon AG. Hald I andD amata (MPA) fe adva udy Cannar T re * Rep 46: 407-404. 1942 2)4 RerahnaGM. CterairiloC and GfevtmaJaa M. 1 fe Marma a idoanm inid caroto-- . Era J Gynaac Onori J: 241-24*. 1912. 233 Trop4 C. fohraaon J-E. Stgurdaeon R and Smonaca E: F` * onc aratale (or ih* noiama r i a h m e rl o 1*41-144), 1942 23* Haracrtynck JVTH MrakeraAP. Mangioni C ri: Prive llnWriaMdnraypropraicuMaociara wadianoBd ovarlancanoa; An EORTCOyaocrioghalCaiwar Cooparailra Oroup ktady Oyuaool Oneri 22 )I> -)I6 . I 0 237 Landonl F. Epta A. Gorga CJ, RegaDo M. Vaaacna L aad Mangioni C H rara al ireairacH m advanoed cpdhcLol ovmvoa canoa In Para*il F (ed). Aau Onuogem In oncotogy Fan. prcicnl and proapcclt Eac. Medica. Aawt.rhm pp 262 267 234 Wcclh JB Laigc duce piogcuia paJInnon: vaUabte in aohd luraot patteon Proc Am ni ** A N T U "ANKER R E SEARCH : J 17-434 (lANOi) A w l am... . f* lr.5. IVT4 gyncodogn- lurroui I InimuWipeloa Id a * and the dealian coaled charcoal amwy I (<cilci Ml IKc um <( h ig h .W iiH'grM Cancel **. 2lSJ-2lf4l. I92 tole-i*,atan, dito Vitim Owral 1} INuppUl 22. i'* ? l H lhncl R. kclean G H II. Marlin ID . Mavierv AM . MiCanncy AJ and Tuaddk L I Sik* 111 V - .k k i NA H jik x iN C n al High ** nK|C*lol acci-nlc therapy .X m i n Edrogen and paugolcrone icci pvon in lum on <d the human ovary Gyncooi Onool I f A |'fl r e II Mikl; l> Ihd Nix iKcih ilrhxnia ( Jncotugy Gfuup Semin Ornad M M SI. 192 / i*>n>i> ivxr. Ptkidciei A er nl. fn. Gvaiial tunvoren (DaJlenhaeh ltllrg O cJ| Springei Vtrlag I J..ll,-v O . 1rxv.tm.iu K-N sftu Jon*-* I A I>n..*-tnl ajktiwl pi .igx-vuifcii therapy in advanced 19*2. p 275 .).J||M.JA.VI |N tlhlU lt ( i) k \ l OVhKO.d 1ft >52 I.W. 19) ! Ouinti M A . Pcaaer P. Kimac R. Funder IW . Fiuiune D and PcppereM IU Cyioptnaolc ' liM .ln .ii k S W r | l l . l > J i ^ l l r i" l I lluntl .litad and mr>Jri>>|M|(r*iri.><K cienuJ irv x p u v i m uvanan lutiuuitv Hi J Otwiei liynaeeid m / 754-759. IMJ I|Vl>^-> ixpMhrliul. Vhava* II hluJy Cancel ilea l Hep I Saarrkotki S. Selahder K. kalliu S and Pyuyncn P. Slcrmd rceepton Ul norami and ftciipla.ltc lemak rcpt.ulu.iivc iiu v c i liynewd Obviei ln .f u H 2IM-2I2. 19*2 I M v.iv CM Mi*C (>l n d M i,.u I I Advanced o.aiian Kcvpamt I I Sehuaru FN. Livobi V A . Ilrldrclh N. Mael mky N i. Nafiohn FN and EitenleMAi anl>. *li-`gvn iJk'tupy l aO.ll * ' ?Vd |<aKI Km .gen ic-cepruiv an uvanan epalhelial caicimuiia Ohucl Gyncent J9 229-274. 15(2 < laC.dv^A A tkncKv-n k an.1 V II A Cy.ln h-iim.-naJ li.al.v Can. c> A I Fold I C, Beiek JS. I agawe 1.6 riml. I.uiogcn and progevinonr icccpion in mrmim pha.< II mal I ai I I anci ( lm O iu.'l t | ' 9|n. Iv*i7 ' G o ltiin l> 7 he llcalftwni ni advan.c.1.vviaJ.iv.nal.in-ana.d theiivalv -ilh ycvanniid rtcnpJawnv Gynccvd O m ul l i 2W Rt4 IWO i l.mcv l.A . Fduardt Cl . Fiecdaun Rs. Tan M l and (iallagci IIS: Eumgea and ami ..ainnu.nav .al cv.i.i(d>--v<<hun.>.k Ri I IMmei C.ynJweid SO 7 Xkl. IvM prugevicrruvc rtcepror inert in ptimary cpuhclial ovanan earoruutvav lm J Cancel JB: v buinnia C and H viicii O la Atniia.lv V-.-ovd Ini.nialnmal iya .p *iu in "Role ad 507-571. m i mcdi.iv.paiigcvltl.xK .dale (MPA I lacndmlinC'la'Ialrd lin o u n " Rome , Italy. >h 1 Kauppita A . Varnko P. hi.men S. Sicnbftck F and Vihku R. Clinical ugniheame al V |-l l. l <k* . I * | . p 1) ruingen and progeuift Icecpi.uv in ovanan cancel CRwjcI Gynecol ft/ 5211)26. IW 1 Jungi > . |.ci I) and 5nn I II In I'l '. r i d i n p the Itucrnalional Symp,mum on l lerbaeh S. M ilk l N. Slayton RI: ri a/ IIiuiu. uk iccepeiw in ovaiun canv i i hat Mcdi.'a.pioyrviri.-nc Acctalc It a.all. I Mel nine 70I Panrumi I . KcUcgnru A. Am Aw Cancel Re* 24 |7ft. m.V Ri.huvl.lli I Telia t una t i . vdvl I luvipta Medna. An.vl.ldam, I92. p 90 i PnU.ru k. Sehmidi Mallheven A. I li.lfmannCi. Sehwcikhan 15. KicicnbclgR.Maarl N c n iin a U. Kjikkur*. n I and kah*|ua K In H>k id atnllaOv. Inlci naiumal and GnU I I I ill-Eiifadnd and ) l l-R Vl'll hmckrtg m c-ynuokt ol normal and aeoplaam k v m p n w i Medi.itypcog.~viei.uic .'dale licneva, Nwiuciland. I l l Jaih2cvcla. human uvanan tivauc lm I Cancer JI UI-fttM. 19} 19*2. p <7 i Spun* J. Giivch E. Suirci I I and K irrri K: knrogca -- and gniagcn - receptor*fc * SvhnalK Pi-.. k .a lio g C M ae lu J y N Naliol.n * and I ivenlckJ A Tamoailcn Iheiapy ovanan careinum* Gyncod Ohnei Invcvl 16 IW -m . I9N) lm a.l>an.cd ..valan vaneo O tn l.i G>ik *.i1 tv 5N<` M . iWO Teulcl G. Geyer 11. De (iiepm o G. Fuchi A. Kkme W and Plkkarri A: Oanoga J Shucy I7R kavafl-gh l l I I. iicivhcnvon DM Fccedmah RS. Copeland IJ and loot land Progctlemnrarvpioeen in maligneo Ovanailumoecn Gcbumh u Fiaucaheflh A I A (am.uiKn iheiapy id epithelial ovalian vancef (Rniel (lyncwdne >75171. I W 7)2 740. m i I Weiner SA. Alhenv ITS. Suron FA . Daviv J and (iftiw u D Tamoailcn therapy in Vrerikku t . Kauppda A and Vihko R Cyiiwd awl ntackai rurogen and pro ^ w (eeutienl epilhclial Ialian ealviluima (iynvc.d Om.d .' 7 i h } | V IWt7 Irocpior I and 17 beiahydroi yitcnud dchydrrgcnaK aelivdy in non diKaacd inane M l : M e n M l Halve* VI. O d v c o RI.Shepheid III . t h ill, . a . l l andMeadCiM A phave in benign and malignant lumnrv id Ihc hunvan ovary Ini J Cancel J2. 41M22. HD. II vio.lv .d ...................... alia* vaneo I d 1 Cance. (lira Omid M U . I9M - Wilkocht D . Topp.la M. Hudum 0 5 . Tyler IJP. Baird PJ and Eavimcn CJ. E r l if I Kavaftagh I I tktiallnn I I and R.dvellv Wk Androgen Iheiapy in ihc HCaimenl ul and progevrei.mc icixpou* m human iwanan lumoet GyneevdOnctd 16.2*6-231. M 2 lellacloiy (parhelia! ovarian cancel Cancel Itc a l Rep 7/ 517, IV7 Wuir H . Wa*wKi E Ciluki P. Sehuli K D and Kaitei R Multiple eyloplaamk naMd I Palmar II. Ni.vdl I. Slock dale A n d Advanced ovarian carcinoma Revpimae lo Ihe hormiuK reiepriu* in benign and malignani irvruri luimm and 1ilitcaac-frac m a n agwmvi I) Tfp-h-l IIR II Cancel lie a t Rep rw | 141 I VO. I ' Tumor Diagn Thai 4 15 Hi. 19) ' kullai.de i S l.l IRI I apmivl licalmcni in i.vatMll cancer lm Symp un 'Ikum onal Gcyei ll. lenlclG . De CiregiiruiGand Pllcidciei A Steroid-hruaaun RciepeoreaheM manipula!nMI id *au*rr' . June 4.5 anJ *. Ivor. K.Uieidam. the Ncthcrlanda. p 0). CHai laIkarrmorn und ihrc klinrachc Bedeuiung OnknkgK 7 |Suppl2) 44-52. KM. v (io rik i I. Iid i 0 . ShyamalaCi. Smnh D and N.uidrv A I hunvuve reeepi.uv S tu k o M i (ironmo* M Kangat L. Macnpat I Vanharama R. Nanuncn Al. and Jnhanaaan 4 ihe inlelaciau< .4 evimgcn null Ihe uieruv Re. cm Pi>| llo im Re, 24 45-81). 196 Slcrouj recepuu and revpoma ul uvanan cancer to hotmimet an n w . Br J Ohaen 1 Jenvcn I V .S u iu liT . KanivhunaT Siumpl W l:. lungHol PW and De Sombre R A Gynaecol VI: 472^7*. |9M iwvi vrep mecharm/n l.u ihe miclaevion rvf rvtiadvul tih la i uicnat Pioc Nail Acad So Lamia M EuradnM and ptogcilcrunc fr a pior* in normal ovary and ovarian M l USA 6324*3. lledl Acta OtMtcf Gyncooi Stand 4 *97 SU). I9M I King W l and Cnceor GL Momnlonal anubudKv localue octilOfcn icoeprcu m ihe I Richman C M. Holi IA . Lorincr MA and Hcrtu AL: Pcnrwcncc and d iuhbM a d mu-lei id laigci eellv Nature .<07 745-747. |9vj earrugen receptor in advanced epilhclial ovanan Ciirewna after chemotherapy OMR I Vkclvh.>n> tk'V. I.irhcrman M t and tm rvki I. Nuclear tucaliulmn ol unoccupied Oynceid 65: 237-26). IW ) ociiiopcn icccpion Naiuic *07 7al 74V. i*AM > Schwani PE. Merino MJ. IjVoi*. VA. l-a-icnc. R. MaeL ew , N and EinenlcU K 1 M i< iuin ifc`l Sicioid hoimonc reecpiuiv in hicavi .aiwei ireaimcnt Miatcgy Rece HniopaihrilogA corrclaln* ol caiiogm and pugenin receptor prnecin In e ' Prog llo im Rev M I.U I . |vwn ovanan camnocnaa Obwci (iynccvd 66: 42*-*)J. 19*5. I Bomuni P. Juhnuui P. Anderum K n a! Primary hruiwuial therapy <d advancedNcau i Ivciwn OE. Skaarland and Uieaicr t Sicrord receptor a n n u l in hum cancer v.nh meg n u id acetate Predulive value ul cviiogcn itvcptor and progeHcroivt lumon aurvival u l paiKnir vnih ovarian carcinoma related to Mcrovd incept iecepa.it level* Semin (tnc.d 12 4* 54. ll< Oyncviu Oncui 21 65-76. | M I Mctiuo. tk i . I l . m i l / KH Pcar--nO II andSegal.. A < urrem xaiua ol evimgcn and SuiionGP. Senior MB.Sliauia IF and Mikvna II. Ediogcn and progcateron pn.gevirione levxptorv in hicavi .an.ci Canee *V iv V l 4 7 , IV77 m epnhcliaJ ovanan nvalignancic* Gyncced Oncvd 2). 176-1*2. 19*6 I Tayku RW Biuvk M ti and king RIH I he u<c cd ikvuadiul uptake and handing-me Toppda M . Tyler IPP. Fay R n a i. Sk k M rcceploo in human ovarian . liudicv in cnjumcinul and uvanan .aicuioma Potlgiad Med i 49. T. IV7J review ul lour yean mute coikcium l l J Oftuci Gynaecol W: 946-992. IW * I kiang DT and keaordy B l. tviiugen icec|Moi away m ihe dillcrcniwl dugnoaiv ad Agaiv.fl N. Rao D L . Mmgcihan K rr t i Clinical evaluation ol Mcroid r adenruiicimunm JAMA 2 Hi 1 IM IV77 ovanan ncoplmmt. lm I Gynaecol ObMd 25 1*5 149. 19*7 l hriheig H i. Kullandei S Peivijn IP and Kortlcn CB tin icecpaonloi ctiiogcnv < ti) Kuhncl R. De GraaH J. Rao RR and Slulk JG Androgen receptor prcJ and andi.igcnv iD ITTl ut human end<ameiiial eaiciruuna and ovalian lumuvri Acta human ovarian carcinoma J Stcroad Bwchcm 26 )9)-.W7. 1W7 Ohci.i t.jAec-vU Stand >7 > 1 2rd I t 7 Run BR. Wkw WCi and Alkn WM Progoicrnnc 'receptor" in laMu mere. I. i Bihi.i M l V h na ru PI. I iVidv. VA and KrvenlcU A l. L .l rugen hutdmg maeio- Characrematron and carra^bol-17-bxla augmcoiaiwn. Endocnruilogy 92: IZ29-12M. nv-.Lc'.ulrv in human ovaOjn caieuviuna I ah Invcvl 41) 241-242. 197V 1973. ' Dapcinitl Daccohi.hiel t i anJ Margienri R Sler.ud hormone ceeeptid dmrihulHin In > Kuhncl R. Dctcmarre IFM Ran BR andSlolk JG ( uriclaiiimof animaitac actlvllyuat hrcavi and genual car.oumia t i n . i i Treat Rep n l | lw>. jvTv tlciuid rcccpioo in human ovalian carcinoma Anli.aaKci Re* 6 BM-BV2. IW*. I heJntan M. I . y M. M aikovnr A. J.mca II. Rcwcr K and Ikdln.anP MtaeyirearduiSd. Bauer WC and Rao BR. Subpopulaiioo* id brxatt carcinoma defined hy (KR> andpiogcvienme rreepror* (PRl in ovartan earner - clinical and paihdogical S-ptosc fraction, morphology, and eurrgen receptor iu iu i Lab Inwar 79. 225-233. conclalHMi Clin Rev 7 Jn5A IvTv IV7P. > Cirilli S Dc Giovanni C. Galh MC n a/. Sici.Md hormone iccepcnn in normal and Rao BR and Meyer IS F.Miogen and progctiin rcccpron in normal and canon ikwm. ftciiptavtK human rvaiy < aftcer Tieai RcpftJ HIM. 1979 fa McGuire WL rr a/ (cd) Progciiciiutc receptor* in normal and Koplaatic lianan*. i HahnelR Martin IL) Mavieiv AM,Rata|CjakTandTuaddk t : Evirngen leeepron and Raven Prew. New York. 1977. pp 155 1*9 bknuJ hnurumc level, in rcrwcal eaicuvnmi and iHhei gyncontogicaJ Iunion Gyneool Kkdfman PG and Snien PK: Sea Mcrovd icccpion in gynccokgK cancel OhmR Oncol M 22n-2J. 1979 Gynecol JJ. 64*452. IVMi l Hed JA. Capeno 1A. Kelly KM . Gicenaid P and Chnioal L EaLmgen and pregona i Fnedlamler M . Fuo MS. Oumn M rr W ReUiromhrp cd Merotd rceepton and aamnr binding m cyiinola o l o*an*n adeiucaionomai Ohviei Gynecol JJ. SISSI. 1979. pkudy in epilhclial uvanan cancer Proc Am Aaa Cancer Ra* 26 27. 1965. I KaupfaU A. Itnnc O. Syrjila P and Vihkn R Euiogen (EK) and peofedin (PR) Hammerman S. Dcaaopuuku RI. Raphael C and Rou ): Gonadotropic hormnm , leceproit ta henign and malignant ovaran lum on Canec Tieai Rap 41 11*7. 1979. binding to human ovanan tnmon Hum Palhol 12 8*6-190, 19*1 I llnne 0 . Kaoppda A. Syryali P and Vihko R Cumparium nl cym ol cRiogcn and Raiamcmi H. Kauppda A . Ronnbcrg L. S<lander K and Pyekynen P LH (HOG) pnvgcvun ietcp.or alaiui in malignant and benign lumuet and [mor-like k u o n i of revcpior in benign tod nulignani lumon o il he iwary. Acta OtoMctGyncooi Scand)p^ . human ovary, lm I Canec 2) I7S 179 IVMi W l 8) 86. IW I l Stedman KE Mtu-irc GF. and Morgan RT tw iogcn irecpiui piuleuu in diverv human i Ei k m i G. Brack C. Siwrm R. Ball P andOberhcuicr F: G N R II binding-aiicaInhmm lumort Aich Srg H i: 2*4-24. |9w) epilhclial ovarian cainnoma 1. Symp on` Hormonal manipulaiion oi cWKcr". lam . I C itanan SkT. Satan RA. Weed Ji JC and McCaiiy Ir KS Ovarian carcinoma, 4.5 and 6. IW 6. Ronaidam. The Neiherlaruft. p 1*3 htviohvgu and clinical c.irrelalnm of cyinplavavic cinngrn and piogeNciuftc binding. i I lochherg RB. MacLuaky N I. Chambcra J. Encnfcld AJ. NahoJnv F and SdkwarV f t Gynecol Oncol 12 JI9-J27. I9l Concconiuon ol |l6e-'"l|ldo-eUindiu4 in human ovarian lumon In nvo and eamM- I (.alii Mt IN GiovanmC . N n id k lliG /fW The iu.vuiicncc of muhiple vicroid toimouic rion with cMtogcn rcccprur conicol Srcrcudt 46: 775-71*. ]96. rcecplorv m diKatc-lice and nanptwiK human ovaiy Cancel 47: 1297.15U2. I9l. I lircm a MA. Fkdl JA and Orerne GL Munockwal anubudy reoogniiiun ol nnddph Ilanuliiin IC . Davici PandCinllnht K Androgen and nc*irogcn binding in eyluanli id lurma oi cungm receptor lagged with | 1],l|MeihOay-iodovinyl ctuaihol In ovnrha human ivvai.an lumvorv I Fwhicf W 42I-4M. I>ml carcvnoma* J d m EmhAimol Mciah 61. *12-417. IW ). llo li JA. I yule CR. Iuna.1 M A. Sicrn SD. Picia ML and I Icchu Al . biiiogcn recepuu and pcioaidavc aeimiy in rpRhclul ovarian caieiooma J Nall Cancel liu i 67: 17-7IN. ISMI | Farley A l O Bncn T Mftyei D and faylnl CR The deieaion ol cvirugen icccpion in Received January 14, 1988 Accepted March 31,1988 An t ic a n c e r R e s e a r c h 5-0 2 (1988) Radioimmunolocalisation of Bladder Tumors Xenotransplanted in Nude Mice Y . h ANSSON1, S. Pa U U E '. H . B E N-ASS A1, U. R U D B E R G 1. A . KARLSSON* and P. P E R E M A N N ' `Department of I m n t u n o l o g , . 12 M S tK kM m Sweden -P" -s,i2 82s'xkhotm` Sw< convenient accessibility in the amount, and homogeneities We have previously reported on Ike derivation of required. Although Ihe technique of in vivo localiwuon u mouse monoclonal antibodies (Mobs). idenlifyint several ceU becoming a clinically uftd option for the diagnoul of certain surface anngens selectively associated wiA TM lumon, thu it ill far from the case for mo lum on. One type bladder (TCC) IM l Three of these Mabs (4ES, of malignancy th u has noi been extensively investigated in this t f 4) have now been assessed for their ability to bcahse TCC- respect i. cancer of .he urinary bladder However, recen, tumor xenograft, in nude mice. The btodumbutton of' l- devdiopmems in defining a number of antigens associated with tabeled intact Mabs as well a, the corresponding Fab and Ibis disease (tor review see 5) have permitted such studies, and F(ab-), fragments from two of them were investigated m anim a few reports on the localisation of this tumor type in the nude als carrying TCC lumon or antigen negaave control tumorv. mouse model have been recently published f(v*). Using direct measurements of excised tissues, all three anti Using the same model, we have now investigated the in vivo bodies were seen to accumulate specifically in the TCC tum n - behavior and lumor targeting capacity of three of our anti- ` giving tumor to normal tissue ratio, of between 3 and 20 TC C antibodies (4E8, S M H ad 8F4). The target antigens for depending on the Mab used and the time after mjecnon. Anti these Mabs have all been defined ascelt surface antigens w iih . body fragments were generally more effiaent in their to a liM - very selective expression on bladder tumor tissue (1 4 ). Loca don, mainly due to a dramatic reduerton in the Mood back- lisation was performed with both intact M .b as well as mUpound as compared to Intact Ig. One of the antibodies. 4E8. body fragments and was measured etiher directly on excised was also employed for external imaging wah gamma camera tissue or by immunoscimigraphy The biodistnbution of In scintigraphy using ,,iln o rlstla , tracers. Excellent visualisation and 1MI-conjugaied antibodies was compared M A c am or sties could be obtained bodi wiA Fab fragments and intact antibody wtAin 12-24 horns after fiction. As expected, background radioactivity was significantly f-v^wtih fragments than wtih whole molecules. " `In appeared in all instances to be superior to Ihe Znjugaic, In conclusion, Ac present study indlextia that Ac three ano TCC antibodies may become useful for Ihe in vivo Materials and Methods C M U no Three c u M u h e d T C C II t o T24. TCCSuP X H U M V f l) were t o . A i.li.c n n e u - I W - * " ' line L S lM T (9) nnd Ihe m elm om n line. 15*5 ( 10) diagnosis of bladder cancer in man. M onoclow U m i M w . Three d llle re n l m on octo n n n lib o d i . * t _ Immunolocaliiaiion ha* gradually reached a age where il h u begun to fulfill *ome of Ihe ixpedahon. aaoaated with it. O l ^ importance in .h i. progreu b n been the deriwuon of S g e tt 1 HEX ( * l | 0 2 * > - " i l " 1 d U ' '" 1 T T C * ^ * . i W e T t o r e J e d -n re w . H W l r 1 > h ~ *" < " " d n * " > d =p re v^.o vly b0y v , t 1i ' ) --1 ^a d d it^io n . . m o v e M . h ( 1 , 0A1U) .u b o d re . - culture w p c m n la n li by n ilin ity e h m m .io u .p h y 0 Pm rem ... monoclonal anubodie.. not only lor their ^ e p h n r o r e (P to m m e i. Floe C h e m to h . U p p v ta . Svreden) ! defining new and anlable target um gen. but a l for their r^rmmondence to: Dr. \ . Hanuon. Deptrtmonl <9 Immun 5 ^ k b o t a . S-106 n Stockbolm. S l o . Key Words: R . d i o . m m u n o l o c l i H t io n . human blO dcr cancer, monoclonal nntibodic., nude mice. A f - A M . h w ere o b u in e d b y iDCubilioc fo r 12 h r * i t ? r '~ * ' u * J r a o ,) p e p s'. (W o tth in g lo n B m chem kals. F te eh old. I ) * " " " L t J S T e r e d l i n e ( p H 4 f t ) ( , 2) . - n r e d |, e . v e in d ia l v ie d *ifi pho|*h*te buffered line (P S ) fP '1 ^ S S 4TM. r.o *r""s i m ocp u iie d oo * FPLC gel column (Superwe 12. Phnnc.*). ^Monovalent F .b I,. m e n u .e r e otonmed by ,ncub.uon 1 3 b n .1 0 2 7 8 -6 9 1 5 (9 4 )0 0 1 0 5 -7 Fd Chetv. Toxic. Vol. 32. No. 12. pp. 1173-1184. 1994 C opyright 1994 Elsevier Science Ltd Printed in Great B ritain. A ll rights reserved 0278-6915-94 S26.00 + 0.00 Review Section BIOLOGICAL EFFECTS OF COSMETIC TALC A. P. Wehner Biomedia! and Environmental Consultants, Inc., 312 Saint Street, Richland, WA 99352, USA (Accepted 16 June 1994) Summary--A review of the literature reveals two primary issues: (I) a weak, but not causal, association of hygienic use of cosmetic talc and ovarian cancer; (2) lung changes in animals exposed to talc aerosol concentrations that resulted in lung overload. The evidentiary weight of the most significant of the epidemiological and laboratory studies and their biological significance for human risk assessment are briefly discussed. Publications describing granulomatous lesions attributed to talc on surgical gloves, and consequences of accidental inhalation of baby powder by infants are also reviewed. The literature reviewed does not provide any convincing evidence that pure cosmetic talc, when used as intended, presents a health risk to the human consumer. Introduction Hildick-Smith (1976 and 1977) and Lord (1978) have reviewed the older literature relating to talc. HildickSmith concluded that the normal use of cosmeticgrade talc does not present a health hazard. In his review of biological effects of talc in laboratory animals. Lord found the mineral to be fibrogenic but he observed that the fibrotic response was a function of the administered dose and `that there are levels of exposure that are tolerable'. In none of the studies reviewed was there any indication of neoplasia. A computer search of the literature of the last 15 years provided well over 700 titles/abstracts on various aspects of talc and talc issues. Of these. 137 w;ere selected for review as they describe epidemiolog ical. clinical and relevant animal and in vitro studies. Papers on exposures to industrial talc, often contain ing more toxic impurities, such as asbestos fibres, were excluded as irrelevant to the safety of cosmetic talc when used as intended. Studies of lesions fol lowing intravenous injection of talc by drug addicts were also excluded because of unintended use. The term `cosmetic talc" in this review refers to talc of high purity as produced by today's responsible manufacturers. A review of the selected titles reveals two primary issues, namely (I) hygienic use of cosmetic talc and ovarian cancer risk, and (2) health effects of inhaled cosmetic talc. In addition, a number of papers Abbreviations: B A L = bronchoalveolar lavage fluid: C l = confidence interval; N TP = National Toxicology Pro gram; OR = odds ratio; RR = relative risk; SCE = sister chrom atid exchange: UDS = unscheduled D N A syn thesis describe granulomatous lesions attributed to talc on surgical gloves, and consequences of accidental inhalation of baby powder by infants. This review focuses on publications dealing with these and related issues. Hygienic use of talc and ovarian cancer risk The Harlow paper Publication of a papier by Harlow et al. (1992) revived concern that chronic talc use in genital hygiene might be associated with an increased risk of epithelial ovarian cancer. That concern was orig inally raised 21 years earlier in a paper by Henderson et al. (1971) on the same subject. Harlow et al. (1992) provided the most detailed results to date on the relationship between pierineal talc exposure and ovarian cancer. The study is the only one to date designed specifically to address this issue. It is among the largest studies (greatest statistical piower) and one of the few to use closely age-matched neighbourhood controls. The authors interviewed 235 white women diag nosed with epithelial ovarian cancer between 1984 and 1987 in Boston and 239 population-based matched controls. Approximately 50% of cases and 40% of controls used talc on the pierineum. under garments, sanitary napkins or diaphragms. This resulted in a 1.5 odds ratio (OR) for ovarian cancer [95% confidence interval (Cl) 1.0-2.1]. Perineal exposure to talc resulted in significantly elevated risks in the subgroups who applied it directly as a body powder (OR 1.7. 95% Cl 1.1-2.7). on a daily basis (OR 1.8. 95% Cl 1.1-3.0). and for more than 10 years (OR 1.6, 95% Cl 1.0--2.7). The greatest risk was 1173 A P Wl HM K observed in women with more than 10.000 talc and transferase a genetic risk factor. Whittemore applications during years when they were ovulating <7 al. (1988) observed in a wcll-dcsigncd study (188 and had an intact genital tract (OR 2.8. 95% Cl cases. 539 controls) that the ovarian cancer risk in 1.4-5.4); however, this applied to only 14% of the women who had consumed coflcc for more than 40 women with ovarian cancer. years was 3.4 times that of women who had never The authors acknowledge the inherent limitations regularly consumed coffee. This relationship is more of epidemiological studies. In their statistical analyses robust (P =0.01) than the marginal significance of they adjusted for several confounding variables and the relationship with perineal talc use [relative risk conceded the existence of additional but intangible (R R )= 1.40, / = 0.06], variables. In-person interviews included enquiries A causal relationship between talc and ovarian into dietary history but it is not clear whether, cancer requires that talc particles, administered to the for example, lactose consumption and coffee con perineum or to the vagina, can translocate to the sumption were determined and adjusted for; both ovaries. Whether or not inanimate particles without substances have been associated with a significant locomotion of their own can do this without assist increase in the risk of ovarian cancer (Cramer et al., ance has been the subject of a number of investi 1989; Whittemore et al., 1988). Except for these two gations in humans and animals. Results have been cases, no diligent efforts appear to have been made inconsistent and ambiguous. For a discussion of to identify other confounding variables that could relevant findings published before 1986, reference is account for increased incidence of ovarian cancer, made to Wehner et al. (1986). such as vulvovaginal diseases (e.g. yeast infections Henderson et al. (1986) observed talc particles and trichomoniasis) and obesity. McGowan et al. in the ovaries of all eight ex-breeder rats after (1979) reported a higher incidence of rubella infection intrauterine deposition of a talc suspension. Follow between the ages of 12 and 18 years in ovarian cancer ing intravaginal deposition of talc suspension, talc patients than in controls, but there was a lower particles were found in the ovaries of only two of frequency of mumps (West, 1966) and of pneumonia six ex-breeder rats. Retrograde flow of menstrual and influenza (Joly et al., 1974). products into the peritoneal cavity through the Harlow et al. (1992) mention in one sentence that oviducts is well known; therefore, translocation to the they used meta-analysis in the statistical evaluation ovaries of talc particles deposited in the uterus of their epidemiological data. Meta-analysis requires in 250 n I saline, perhaps aided by the hydrostatic rigid criteria that must be strictly observed to yield pressure of the saline solution during and after meaningful results (Chalmers et al.. 1989; Yusuf deposition, is not surprising. Yet in two of six rats et al.. 1985). Without details on the meta-analytical the inanimate talc particles managed to cross the procedures used by the authors it is not possible cervical barrier and reach the ovaries, perhaps also to determine whether the use of meta-analysis aided by hydrostatic pressure during deposition. was appropriate, and if so, whether it was used Henderson et al. (1986) express a guarded opinion on appropriately. the association of talc with cancer, stating that "car The sample population in the Harlow study is cinogenic activity of talc has not been established limited in numbers and restricted to white women although its ubiquitous presence in the environment in the Boston area. With an overall adjusted OR of and its elemental similarity to asbestos has brought it 1.5 (Cl 1.0-2.1) the results are statistically barely under suspicion". Elemental similarity, of course, significant. Extrapolating from results in that small, does not necessarily mean similarity in biological restrictive group to the general population will effects. It is well known that differences in physical require caution. properties, such as shape and surface characteristics, The authors conclude that their "data support the between elementally similar substances are respon concept that a life-time pattern of perineal talc use sible for significant differences in their biological may increase the risk for epithelial ovarian cancers", effects. The pharmacological action of certain ele but they do not claim to have established a cause- mentally and structurally identical agents (isomers) effect relationship. depends on whether they are dextrorotatory or Related literature levorolatory. A remarkably large number of talc particles were Harlow et al. (1992) cite a number of literature found by Henderson el al. (1979) in human ovarian references on talc translocation and putative associ tissue, namely from 6900 to 55.100 particles/g wet ation between hygienic use of talc and ovarian cancer. weight tissue in three normal ovaries, from 17,400 to These references contribute to the ambiguity of the 24.300 in three cystic ovaries, and from 6400 to database as some of them suggest translocation 24.300 in three ovarian adenocarcinomas. In another or support an association whereas others do not; reference (Henderson et al.. 1978), 2 x I0J particles/ again, others suggest alternative explanations for mm-' are reported from oxygen incineration studies of greater-than-average incidences of ovarian cancer. human ovarian tissue. For example, Cramer et al. (1989) suggested that Talc particles in human and animal tissues can be lactose consumption might be a dietary risk factor identified by X-ray fluorescence and X-ray diffraction Biological effects of cosmetic talc 1175 (Wehncr et al.. 1977c) and by neutron activation of talc before animal exposure and subsequent y-ray analysis of animal tissues for at least two suitable radionuclides (Wehner el al., 1986 and 1977b). The neutron activation technique eliminates contami nation problems during sample collection and processing. Talc particles found, for example, on ovarian tissue might be contaminants deposited during sample collection and processing. For talc particles in ovarian tissue, contamination during sample collec tion and processing can be ruled out. Wehner et a i (1985 and 1986) investigated talc translocation in Cynomolgus monkeys (Macaco fascicularis). The female of this species resembles the human female closer than any other animal model in the physiological and anatomical parameters of inter est. The authors deposited a suspension of neutronactivated talc in the posterior fornix of the vagina on 30 consecutive work days (45 calendar days), that is, through at least one menstrual cycle (Wehner et al., 1986). 2 days after the 30th talc application the animals were killed. The following samples under went y-ray analysis: vagina with the cervix of the uterus; uterus; the entire oviduct in three sections; ovaries; and peritoneal lavage fluid. Only the vaginawith-cervix samples--the site of deposition--con tained varying quantities of talc. No talc was detected in any of the other samples. The detection limit of the neutron activation/y-ray analysis technique under the described experimental conditions was approxi mately 0.5 ftg talc (Wehner et a i. 1986) or about 1/230,000-1/245,000 of the estimated talc deposition in the posterior vaginal fornix (Wehner et at., 1985). Harlow and Weiss (1989) interviewed 116 patients with borderline ovarian tumours on their use of hygienic powders and found `no appreciably altered risk' in the use of baby powder or cornstarch. How ever, the unspecified smaller number among those women who use deodorizing powders alone or in combination with other talc-containing powders had a 2.8-fold higher risk than 158 women without perineal exposure to powder. In the light of this newly discovered confounding variable, Harlow and Weiss recommend that the specific type(s) of powder used should be identified in future studies on hygienic powder use and ovarian tumours. Cramer et al. (1982) observed a statistically signifi cant (P < 0.003) relationship between epithelial ovarian cancer and talc used for dusting the perineum or sanitary napkins in 215 women. However, Cramer et al. (1982) found no relationship between ovarian cancer and talc exposure from dusting condoms or diaphragms, even though talc, in the latter appli cations, is deposited close to the cervical os. Hartge et al. (1983) made a similar observation from their epidemiological study. Their data indicate that the use of talc on a diaphragm did not appear to increase risk and that there was no overall association between talc use and risk of ovarian cancer although a small group of women who specifically reported 'genital use' of talc showed an unspecified excess relative risk. Mostafa et al. (1985) observed in 175 grossly normal, surgically removed human ovaries a 9% incidence of magnesium silicate granulomas and an additional 9% incidence of histological changes very similar to these granulomas. Booth et al. (1989) reported an association between infertility as well as late onset of menopause and increased risk of ovarian cancer. In their opinion, the evidence linking talc with ovarian cancer was contro versial, and they stated that more studies are needed to clarify this issue. The association between `fibre' exposure and ovarian cancer has been described by Rosenblatt et al. (1992). Cases were ascertained between 1981 and 1985. The authors define fibre exposure "as exposure to asbestos, talc (which may contain asbestos), and fiberglass". The authors observed elevated, but statistically not significant, risks with the use of condoms, powdered diaphragms and genital bath-talc. The risk for use of talc on sani tary napkins was significantly greater than unity (R R = 4.8; 95% Cl = 1.3-18.0). The authors con clude that "The results of our study and others suggest that genital fiber exposure may be associ ated with an adverse effect . . . but further study is needed to determine if this relationship is causal in nature." Tzonou et al. (1993) examined the use of anal gesics, tranquillizers and perineal talc application as risk factors for ovarian cancer in 189 cases and 200 controls. Talc use was determined on a no/yes basis without attempts to quantify application. The authors adjusted for a number of confounding vari ables, among them age. years of schooling, body weight, age at menarche, menopausal status, parity (nulliparous/parous, parous age at first birth, smoking (non-smoker/ever smoker), consumption of alcohol (glasses/day) and coffee (cups/day). There was a marginally significant inverse association with an apparent dose-response trend between frequency of analgesics use and ovarian cancer, and a highly significant dose-dependent positive relation between hair dyeing and ovarian cancer. As to talc, the authors state, "although the number of talc users is in general small and the respective confidence interval fairly large, there is clearly no evidence of an increased risk associated with perineal application of talc." Chen et al. (1992) investigated risk factors for epithelial ovarian cancer in Beijing in 112 cases and 224 age-matched community controls. Among a number of other risk factors, the authors reported in their abstract an elevated risk in women with a history of long-term (> 3 months) application of talc-containing dusting powder to the lower abdo men and perineum (RR 3.9. 95% Cl 0.9-10.63). 1176 A P Wl MSI K Examination of the full paper reveals that these figures arc based on seven (!) eases and five (!) controls. Kupryjanezyk (1989) described multiple talc gran ulomas. inclusion cysts and adhesions in close prox imity to an adenomatoid tumour of the left ovary of a 41-year-old patient. She suggests that talc crystals, as well as repair and inflammatory processes, should be taken into account as initiating factors in the development of ovarian adenomatoid tumours in susceptible patients. In their paper on the aetiology of ovarian cancer, Baylis et al. (1986) conclude, `i t certainly cannot be said at present that talc causes ovarian cancer." However, because the cause of ovarian cancer is unknown, and because of the "unexpected and un warranted" presence of talc in ovarian cancer tissue and talc's "chemical similarity with asbestos", the authors are pursuing further investigations. In their review paper on the epidemiology of ovarian cancer, Greene et al. (1984) summarize the role of talc and asbestos as follows: "while the above observations [reviewer's comment: referring to some of the publications reviewed here] are provoca tive, a conclusive role for talc or asbestos or both in the genesis of human ovarian cancer has yet to be demonstrated in either cohort or case-control studies." There are other papers in which an association between hygienic talc application and ovarian cancer is mentioned. Most are reviews, referring to the literature reviewed here. None provide new infor mation on the talc/ovarian cancer issue. Hamilton et al. (1984) injected 100/ri saline con taining 10 mg talc into ovaries of rats and observed no evidence of neoplasia. In their review, Longo and Young (1979) point out that "epidemiological data could be interpreted as showing that the risk of developing cancer from an occupational talc exposure was due to contaminating asbestos." This could equally apply to chronic ex posure from hygienic use of talc. The authors support this view by subsequently slating " . . . consumer talc products marketed before 1973 were variably contaminated by asbestos", and. later. " . . . data collected on populations exposed before 1976 may reflect the hazard of contaminating asbestos rather than talc. Unfortunately, adherence to the revised Cosmetic, Toiletry and Fragrance Association guide lines is voluntary . . . so that, even now [1979: year added], commercial talcs are not certain to be asbestos-free." Cralley et al. (1968) found fibre con tents ranging from 8 to 30% in cosmetic talc products available at the time of their investigation. In considering mechanisms that might be involved in ovarian carcinogenesis. Venter (1981) points to the extremely high concentrations of gonadotrophins and potent steroids in the follicular fluid that is released monthly into the pelvic cavity by the rupture of the ovarian follicles. The concentration of these chemicals correlates with the mitotic and biosynthetic activities of granulosa cells, oT which approximately 50 million accumulate in a follicle during the follicu lar phase with about 6.5 ml of antral fluid before the follicle is transformed into a corpus luteum. Given the fact that oestrogens cause proliferation of certain cells. Venter proposes the hypothesis that the antral fluid could act as an ovarian cancer promoter. Dietl et al. (1986) emphasize that the ovarian surface epithelium is a dynamic tissue with distinct morpho logical differentiations: it may proliferate inwards and form crypts and inclusion cysts or it may develop superficial papillary excrescences. In addition, con stant metaplastic changes may take place in various parts of the mullerian epithelium. These growth processes appear to be influenced by endogenous and exogenous factors. It is conceivable that these factors, in combination with the physiological, biochemical and morphological characteristics of the ovarian tissue, can induce neoplastic lesions in the ovarian surface epithelium. The repeated breaks in the epithelium that occur during ovulation apparently increase the risk of developing neoplasia. Biological effects o f inhaled talc The literature search revealed that publications on biological effects of inhaled cosmetic talc arc sparse and basically fall into two categories: (I) findings in animal studies: (2) accidental inhalation of large quantities of talc by infants and small children. Talc pneumoconioses in adults Feigin (1989) distinguished between three forms of pulmonary disease caused by talc inhalation, namely (I) talcosilicosis. (2) talcoasbestosis and (3) pure talcosis. Feigin writes: "Talcosilicosis is produced by exposure to talc usually from Italy but also from California associated with silica and other nonasbestiform minerals___ The clinical and radio graphic manifestations resemble those of silicosis. .. . The only documented difference between silicosis and talcosilicosis is the presence of talc, as demon strated histologically. No other differences have been described. The radiographic changes are indistin guishable from those of silicosis," Talcoasbestosis is caused by inhalation of talc containing asbestiform fibres, such as tremolite and anthophyllite. as mined, for example, in upper New York state. On pure talcosis. Feigin writes: "Symptoms and pulmonary function test results consistent with restrictive pul monary disease are well documented in pure talcosis; airway obstruction may also occur. The clinical form of the disease has most often been documented in miners and other workers involved in the obtaining and processing of pure talc. It has also been docu mented in people exposed to cosmetics, but only when the exposure was very heavy and prolonged" This reviewer could find only two literature refer ences on talcosis in consumers. Wells et al. (1979) Biological effects of cosmelic talc 1177 describe a case of talcosis due to talc abuse, initially suspected to be tuberculosis, in a 41-year-old house wife. On direct questioning the patient admitted to very heavy (at least once a day) use of talc powder over her whole body in an unventilated room for many years. The other reference is to a paper by Tukiainen et al. (1984) who describe two cases. One of them involved an elderly female smoker of 20 years with a history of several operations. When she presented years later with non-productive cough, dyspnoea, tachycardia and low-grade fever, talcosis was considered a possibility on account of her 10-year use of talcum powder two or three times a day in an unventilated room. The authors eventually diagnosed chronic sarcoidosis with talc deposition in the lungs. The other case was an elderly female smoker of 30 years who had been occupationally exposed to industrial talc from 1958 to 1968. Scatarige and Stitik (1988) reviewed the liter ature on the induction of thoracic malignancies in inorganic dust pneumoconiosis. They concluded that "talc pneumoconiosis does not appear to be associated with an increased risk of lung carcinoma or mesothelioma, and animal studies have failed to convincingly demonstrate the carcinogenicity of talc." Accidental inhalation o f large quantities o f talc by infants A number of reports describe consequences of accidental inhalation of large quantities of baby talc powder by infants. Although the cases do not consti tute talc use as intended, they are nevertheless included in this review for the sake of completeness as they present a form of--albeit accidental--talc inhalation. Brouilletle and Weber (1978) describe the case of a prematurely born 1-month-old girl, presented at a hospital in cardiac arrest. She was covered with talc powder which had been poured into her mouth and nose by her 3-year-old brother. Following appropri ate treatment of her severe pneumonia, she was discharged after 12 days of hospitalization without apparent consequences. The authors refer in their paper to at least 24 previous cases of massive talc powder inhalation by infants. Most of the children were older than 6 months, and those old enough to play with the powder container were considered at risk. The mortality in these cases was 20%. Mofenson el al. (1981) point out the potential hazard of careless use of baby powder. They reviewed the experience of the Poison Control Center at the Nassau County Medical Center. New York, with baby powder inhalation in children less than 5 years of age. in whom most of the incidents occur. Of approximately 4300 calls in a 6-month period. 40 concerned baby powder inhalation. Symptoms included cough in 14 children, dyspnoea in five, sneezing in five, vomiting in six. and cyanosis in one. McCormick et al. (1982) describe the hazards associated with diaper (nappy) changing, based on statistics from the Massachusetts Poison Control System. Of 138 cases of exposure to various 'poisons' during diaper change in a 3-month period, powders accounted for 47%. Symptoms such as coughing, wheezing and shortness of breath were described as mild, occurring most often with powders. Motomatsu et al. (1979) report the clinical case of two baby girls who died following accidental inhalation of baby powder and unsuccessful treat ment. "In order to investigate the harm of baby powder," the authors placed eight mice in a box, the bottom of which was covered with baby powder which was then " blown up" with compressed air. Neither aerosol characterization nor other measure ments or experimental data are provided by the authors. Four mice, removed from exposure after 30 and 60 min. "recovered completely" . Two other mice, removed after 90 min exposure, died within 6 hours and the last two mice died after 2 hours of exposure. Histopathological findings include haemorrhage, oedema and desquamation of bronchial epithelium. Pfenninger and D'Apuzzo (1977) describe two cases of powder inhalation. One was a 7.5-month-old girl who had inhaled Fissan baby powder containing talc, zinc oxide and other unspecified substances. The patient responded only slowly to treatment and required 19 days of hospitalization but eventually recovered fully. The second case involved a 13month-old boy who had inhaled Merfen powder containing talc and borate of phenylmercury. The patient responded well to treatment and recovered completely within 4 days. De La Rocha and Brown (1989) report a case of baby powder inhalation followed by adult respirat ory distress syndrome in a 16-month-old girl who recovered fully after 20 days of hospitalization. Gutermuth et al. (1980) describe accidental talc inhalation and treatment in an 11-month-old female infant and in a 21-month-old boy. The first patient recovered after a 16-day hospital stay, the second one after a 13-day hospital stay. Mussi et al. (1979) report a fatal case of powder inhalation in a 14-monih-old girl. After a character istic fairly asymptomatic initial period--in this case 12 hours--the girl did not respond to treatment and died 41 hours after the accident. Swanson-Bicarman et al. (1991) report the case of a 10-month-old boy who was hospitalized for 16 days following massive talc inhalation. The patient had recovered completely by the time of his dis charge. Butenandt et al. (1981) treated a 9-month-old male infant who had accidentally inhaled several grams of Pcnanten powder which contained 96% talc. Prompt bronchial lavage plus appropriate treatment resulted in an asymptomatic status after only 4 days of hospitalization. 1 178 A P W fhner Cotton and Davidson (1985) describe the ease of a prematurely born 4-monih-old male infant with a tracheotomy tube in place to maintain airway patency. During diaper change at home, baby powder was spilled accidentally and apparently blocked the infant's airway. At arrival in the hospital the boy was in cardiac arrest. He was resuscitated but died the following day. Pairaudeau et al. (1991) report respiratory arrest in a 12-weck-old boy, following a talc spill on his face during diaper change. He recovered during several days of hospitalization and treatment. Cruthirds et al. (1977) diagnosed progressive diffuse pulmonary fibrosis (talc pneumoconiosis) in a 10-year-old girl who had inhaled a considerable quantity of baby powder at 2 years of age at which time hospitalization was not thought indicated. Articles by Rumack (1982), Hayden and Sproul (1984), Wagner and Hindi-Alexander (1984) and Hollinger (1990) are mainly brief reviews which do not contribute new scientific information above and beyond the papers reviewed in this report. Animal studies Wehner et al. (1977c) exposed hamsters to a respir able talc aerosol concentration of approximately 8 mg/m3 for 3, 30, or 150 minutes/day, 5 days/week for 30 days, or for 30 or 150 minutes/day either until they died naturally or for a maximum of 300 days. The hamsters received cumulative exposures rang ing from about 15 to more than 6000mg/hr/m3. Estimates based on a pulmonary deposition and clearance study with neutron-activated talc (Wehner et al., 1977b) indicate that 0.05-6 n% talc, depending on the length of exposure, was deposited in the hamster lungs at each exposure. Estimates based on infant-dusting experiments (J. N. Sivertson, per sonal communication, 1976) show that the weekly hamster exposures, expressed in mg/hr/m3, exceeded the average weekly infant exposures by some 30 to 1700 times, depending on the hamster exposure group. At death, the lungs, trachea, larynx, liver, one kidney, stomach, uterus, one ovary, or one testis, and all tissues showing gross lesions were collected for histopathological examination. The talc exposures did not affect body weight, survival or the type, incidence or degree of histopathological changes in the exposed groups compared with sham-exposed controls. Wehner et al. ( 1977b) determined pulmonary depo sition, translocation and clearance of inhaled talc in hamsters by a single 2-hour nose-only exposure to neutron-activated talc and subsequent serial killing. Lungs, liver, kidneys, ovaries, skinned carcass and 2-day and 4-day excreta were subjected to y-ray analysis. The isotope " Co was used to estimate talc quantities in the samples; the isotope "S c was used to check the validity of " Co as a tracer for talc. From 20 to 80/ig talc (approx. 6-8% of the quantity inhaled) was deposited in the deep lung with a biological half-life of 7-10 days. Alveolar clearance was essentially complete 4 months after exposure. No translocation of talc to liver, kidneys, ovaries or other parts of the body was found. To validate the interpretation of the pulmonary deposition, translocation and clearance data (Wehner et a!., 1977b), Wilkerson et al. (1977) investigated whether radionuclides leached from the neutronactivated talc in serum and in dilute hydrochloric acid. Leaching was negligible in both liquids, but somewhat higher in dilute hydrochloric acid than in serum. A gavage study with neutron-activated talc showed that talc absorption in the gastro-intestinal tract of hamsters is also negligible (Wehner et al., 1977a). In 1992, the National Toxicology Program (NTP, 1992) prepared for public review and comment a draft technical report on a comprehensive chronic inhalation study conducted at the Lovelace Biomedi cal and Environmental Research Institute. The study, designed to investigate toxicology and carcinogenesis of talc in rats and mice, followed the standard NTP experimental protocol for chronic inhalation studies. F344/N rats and B6C3F, mice were exposed for 6 hours/day. 5 days/week to an intended talc aerosol concentration of 0, 6 or 18 mg/m3. In rats, the exposures resulted in impaired respiratory function; increased lung weights; inflammatory, reparative and proliferative processes in the lungs; hyperplasia of alveolar epithelium; interstitial fibrosis; accumulation of macrophages in lymphoid tissue and regional lymph nodes; and occasionally squamous metaplasia. Incidence and severity of these changes generally were a function of dose. The incidences of alveolar/ bronchiolar adenomas and carcinomas were signifi cantly higher in female rats (but not in males) of the 18 mg/m3 exposure group than in the controls. A significantly increased incidence of phaeochromocytomas of the adrenal medulla in talc-exposed rats of both sexes cannot be explained, as there is no known mechanism by which talc particles deposited in the lungs can affect the adrenal medulla, with the possible exception of a stress-related effect owing to a high pulmonary particle load. In mice, the talc exposures produced chronic inflammation and macrophage accumulation in the lungs, but no hyper plasia, metaplasia or interstitial fibrosis, and no pulmonary neoplasms. As a relatively recent, comprehensive study that yielded interesting--even puzzling--results, the Lovelace study deserves special comment. Although it was generally well conducted, the study has flaws that can interfere with the interpretation of its results. Inclusion of negative and positive dust con trol groups would have allowed unambiguous determination of relative toxicity/carcinogenicity of inhaled talc. As it is, the question remains whether the observed pulmonary lesions and other changes in the talc-exposed rodents of the Lovelace study are Biological effects of cosmetic talc 1179 talc-specific or a non-specific foreign-body (dust) reaction that is to be expected as a consequence of inhalation exposures at concentrations that result in lung overload. The investigators were unable to maintain target aerosol concentrations for the 18 mg/m' rat exposure group during 19 of the 113-122 weeks of exposure. For 7 of these weeks the rats were exposed to approximately twice the intended aerosol concen tration. Even the two intended exposure concen trations led to an impairment of lung clearance mechanisms; both of them, therefore, meet the cri teria for a maximum tolerated dose or maximum functionally tolerated dose. On the basis of present knowledge and standards for conducting chronic inhalation studies to investigate carcinogenicity, the chosen talc aerosol concentrations in the Lovelace study were too high. The carcinogenic response observed in female rats of the high-dose exposure group might therefore be attributable to a secondary effect of carcinogenesis, based on a high particle load in the lung (lung overload condition). The Lovelace study can be considered irrelevant for assessing the pulmonary oncogenicity of inhaled talc in humans and the authors of the NTP draft report do not imply any such relevance. Talc aerosol doses received by users of cosmetic talc are several orders of magnitude lower (Aylott et al., 1979; Russell et al.. 1979) with no danger of reaching a lung overload condition. The increased incidence of phaeochromocytomas in male and female rats at the high talc exposure concentration remains a perplexing phenomenon that needs to be independently confirmed. Its relevance for humans is questionable, last but not least, again because of the excessive amount of talc ac cumulating in the rat lungs and the possibility of subsequent stress-related effect. Background inci dences of phaeochromocytomas in control rats were already rather high, and no significant increase was observed in the low exposure groups, although these groups also clearly showed symptoms of lung particle overload with impairment of alveolar macrophage clearance function. The benign pulmonary lesions in the talc-exposed animals generally were those typically observed in inhalation exposures of laboratory rodents to high concentrations of a variety of dusts. There was a significant incidence of bronchiolar/alveolar ade nomas and carcinomas in the female rat group ex posed to 18mg/m\ but not in the males, and not in mice of either sex. The biological significance of this observation remains uncertain. Pickreil et al. (1989) investigated the relationship between the inhalation exposure concentration of talc and the resulting lung burdens and histological lesions. Rats were exposed to 0. 2.3. 4.3 and 17 mg talc/m' for 6 hours/day. 5 days, week, for 4 weeks. Lung burdens were 0. 0.07, 0.17 and 0.72 mg lalcg lung, respectively. Mice were exposed to 0. 2.2. 5.7 or 20.4 mg talc/m'. which resulted in lung burdens of 0, 0.10. 0.29 and 1.0 mg talc/g lung. Histological changes consisted of a modest increase in talc-con taining free macrophages within alveolar spaces in both rat and mouse groups exposed to the highest concentration of talc. Wagner et al. (1977) exposed groups of rats to mean respirable concentrations of 11 mg SFA chrysotile or Italian talc/m \ 7.5 hours/day, 5 days/week, for 3, 6 or 12 months, respectively. Some rats were killed 10 days after termination of exposures, others after one year. Minimal to slight fibrosis as a function of exposure duration occurred in the dust-exposed rats. In vitro studies Beck et al. (1987) investigated the toxicity of quartz- and asbestos-free talc and of granite (12% quartz), collected from worksites, in hamsters that received the dusts by intratracheal instillation. Dose-response (0.15, 0.75 and 3.75 m g/100 g body weight) and time-course (1-14 days) studies were conducted in bronchoalveolar lavage fluid (BAL). One day after exposure, both talc and granite caused elevated enzyme levels, pulmonary oedema, and increased cell numbers in BAL. Macrophage phago cytosis was inhibited. Response levels were either between 'nontoxic' iron oxide and toxic alpha-quartz or comparable with alpha-quartz. The response to granite dust decreased fairly rapidly as a function of time, but talc exposure resulted in longer elevated enzyme levels and decreased macrophage phagocyto sis. The authors conclude that, given similar mass deposition in the lungs, talc causes more lung injury than granite. Endo-Capron et al. (1993) studied genotoxicity of three talc samples in rat pleural mesothelial cells, using genotoxicity assays for unscheduled DNA synthesis (UDS) and sister chromatid exchanges (SCEs). Attapulgite and anatase served as negative controls, chrysotile and crocidolite as positive con trols. The positive asbestos controls enhanced UDS or SCEs in treated cultures compared with untreated control cultures, but the talc samples and the negative controls did not. Davies et al. (1983) tested the cytotoxicity of seven specimens of high-purity talc dust in mouse perito neal macrophages. All samples consistently showed moderate macrophage cytotoxicity, suggesting that they would be fibrogenic in vivo. Talc granulomas from powdered surgical gloves Talc powder is fibrogenic when administered by various routes to many species of animals (Lord. 1978). When introduced into open wounds it may induce talc granulomas. This phenomenon is used therapeutically in pleurodesis. the deliberate pro duction of adhesions between the parietal and visceral pleura by (usually surgical) deposition of talc or kaolin to treat recurrent pneumothorax. Chappell 1180 A. P. Wehner et al. (1979) published a survey of the long-term effects of talc and kaolin pleurodesis. There was no increased incidence of lung cancer and no case of mesothelioma in 199 traceable patients who underwent pleurodesis 14 to 40 years previously. Weissberg and Kaufman (1986) successfully used talc for pleurodesis in the treatment of resistant empyema in five patients who completely recovered from empyema with no undesirable side-effects. However, in the vast majority of cases, talc granulomas are undesirable effects of wound contamination with talc, usually from surgical gloves. Sparrow and Hallam (1991) implicate talc powder in the aetiology of an appreciable number of umbili cal granulomas excised from infants and young children. In view of these cases and with reference to the sometimes disastrous consequences of accidental inhalation of baby powder by infants, the authors strongly discourage the routine use of talc powder in the care of infants. Healey and McDonald (1977) observed a talc granuloma in the right hemiscrotum of a 3-year-old boy who had undergone a right hydrocelectomy 3 months previously. The authors state that the interval for granuloma formation following contamination is extremely variable, ranging from 2 weeks to 45 years, with an average of 2 months after surgery. The extent of granuloma formation depends mainly on the dose and antigenicity of the contaminant and on host response. Pratt et al. (1985) consistently found birefringent particles, which they identified as talc, in the sub serosal stroma of hernia sacs. Cellular response was remarkably mild, perhaps owing to the small particle size (about 10 fim) compared with about 50 ^m of particles in talc granulomas. The authors further hypothesize that the particles were ingested with medication or food and reached the peritoneal cavity by migration through the intact intestinal wall. Al-Sheikhli (1978) observed talc powder in granu lomas of the vocal cords of a 16-year-old girl and suspects as the cause accidental talc contamination of an endotracheal tube with which she was intubated 4 months previously. Simsek et al. (1992) report a case of severe obstruc tion of the urinary tract due to a talc granuloma in a 70-year-old male patient 7 years after a suprapubic transvesical prostatectomy. Felling and Evans (1986) experimentally tested in rats long-term peritoneal tissue response to mouldrelease agents and lubricant powder used on surgical gloves. Iniraperiioneal implantation of talc produced significantly more adhesions and more severe, persist ent, granulomatous reaction than starch powder and calcium carbonate. Sheikh et al. (1984) describe tissue reactions to talc and Keoflo (low cross-linked cornstarch) which they tested as contaminants on the surface of sur gical sutures or as pellets implanted in the abdomi nal muscle of rats. Keoflo caused a strong acute inflammatory reaction which, together with the starch, had essentially disappeared by the fourth week, leaving minimal lesions and scar formation. By contrast, the implanted talc initially induced only mild to moderate acute inflammation followed by chronic inflammatory response, and granuloma for mation by the third day. The results suggest that low cross-linked cornstarch is a bioabsorbable substance, whereas talc is not. The authors therefore conclude that low cross-linked cornstarch "is a safe material for use as surgical glove powder", a conclusion not supported by findings of cornstarch-induced lesions reported independently by several other investigators (see below). Kaiser et al. (1982) reported similar results with intraperitoneal tests in rats. Talc may contaminate surgical gloves as a mouldrelease agent during the manufacturing process or it may be deliberately added as a lubricant for easier donning. In a letter to the editor, Henderson and Griffiths (1979) state: "The large number of talc particles observed on certain American-produced surgeons' gloves, on occasions in excess of 6 x 10' particles/cm2, would suggest that it is this material that is being employed as the releasing dusting powder in the molding process." Tolbert and Brown (1980) and Khan et al. (1983) point out that removal of talc particles from gloves is difficult using rec ommended washing and wiping procedures, and that a shedding hazard might exist by mechanical dislodg ing of the particles during surgery. Rather than reviewing all publications of the last 15 years on talc granulomas, it is the limited objective of this section to show that talc granulomas following surgery can and do occur at various locations and that they can be induced experimentally in animal models. For relatively recent comprehensive reviews of various aspects of glove-related issues, reference is made to the papers by Ellis (1990) and Beck (1992), Fay and Sullivan (1992), White (1992) and Witmeyer (1992). Parenthetically, replacement of talc with corn starch powder as a glove lubricant has resulted in starch granulomas (Wilson and Garach. 1981). starch peritonitis (Ellis, 1990; Loup et al.. 1979; Urdiales Cabal et al., 1989) and intraperitoneal adhesions in rats (Kamfler et al.. 1992). With the trend since the 1980s towards powder-free gloves gaining momen tum, complications from wound contamination with surgical glove powders may become rarer but not eliminated. To avoid allergic reactions (including anaphylactic shock) in sensitive individuals exposed to water-extractable latex proteins on medical gloves (Beezhold, 1992), vinyl, nitrile, neoprene and co polymer gloves are now available: generally, these gloves are powdered (Witmeyer. 1992). Ingested talc Under conditions of normal use, talc can be ingested in one of two ways. First, when used as intended as dusting powder, very small amounts of Biological effects of cosmetic talc I 181 talc dust may be inhaled. That portion of the talc particles which is deposited on the ciliated part of the respiratory tract will be transported cranially by the mucociliary escalator mechanism and then swallowed. There are no reports in the literature describing biological effects of these minute quantities of ingested talc. Secondly, talc accounts for the bulk of filler materials in tablets. Appreciable amounts of talc can be ingested with chronic heavy consumption of pills. Anani et al. (1987) describe the case of a 46-yearold male who presented with abdominal pain. Fur ther examination and a right hemicolectomy showed marked fibrosis of the intestinal wall in which birefringent particles with energy-dispersive spectra typi cal of those for talc were found. The anamnesis revealed that, at the age of 27 years, the patient was treated over a period of 28 months for pulmonary tuberculosis with tablets containing talc (183 g talc per 2670 g total tablet ingestion). The authors specu late that the tablets ingested during this antitubercu losis therapy were the source of the talc found in the intestinal fibrosis. If this assumption is correct, the potential consequences of daily multiple pill use, as can be the case with vitamin and mineral sup plements, should be more closely examined. Moder ate pill consumption does not appear to present a risk. In their review of pharmaceutical excipients, Golightly et al. (1988) state, "Ingestion of talc is very unlikely to be toxic." The Joint Expert Committee of Food Additives of the Food and Agriculture Organisation of the United Nations and the World Health Organization stated in its report on food additives that talc was not muta genic in litro or in vivo and allocated an accept able daily intake " not specified" classification to food-grade (i.e. free of asbestiform particles) talc (FAO/WHO, 1987). Discussion The literature reviewed reflects several areas of concern regarding biological effects of cosmetic talc use. Hygienic talc use and ovarian cancer The presence of large numbers of talc particles in normal and diseased ovarian tissue seems indis putable although it is difficult to imagine how inani mate particles without locomotion of their own can breach the formidable cervical barrier and migrate `upstream' against the ciliary beat of the fallopian epithelium to the ovaries. Several investigators have reported an association between hygienic use of talc and ovarian cancer; none claims to have established a causal relationship. The epidemiological evidence linking hygienic talc use with an increased risk of ovarian cancer generally is weak and sometimes inconsistent: confounding variables were often ignored; the reported increased risk ratios, in most cases less than 2. are barely statistically significant, and epidemiological studies are not sensitive enough to estimate risk ratios less than 2. Talc use might be a causally unrelated marker for confounders associated with increased ovarian cancer risk such as. for example, vulvovaginal disease or obesity. There appears to be a consensus of opinion that more and better designed studies are needed before valid scien tific judgement can be passed on whether or not there exists a causal relationship between hygienic talc use and increased ovarian cancer risk. An interesting question remains. Talc is a recog nized fibrogen. If there are sufficient numbers of talc particles in ovarian tissues long enough to cause cancer, where is the ovarian fibrosis which one would expect to develop long before cancer occurs? The only references to ovarian granulomas are those by Mostafa et al. (1985) and Kupryjanczyk (1989). Talc inhalation In contrast to individuals occupationally exposed to industrial talc, talc pneumoconiosis from personal use of cosmetic talc appears to be extremely rare, occurring only following chronic abuse. When talc, a fibrogenic substance, is chronically deposited at doses sufficiently high to overwhelm the bronchopulmon ary clearance mechanism, a fibrotic/ granulomatous tissue response can and will occur, as observed in humans and animals. Animal studies suggest that natural defence mechanisms, such as macrophages and mucociliary clearance, can cope with exposure to talc concentrations considerably exceeding estimated infant exposures, without lesion development (Wehner et al.. 1977c). If doses in animal experiments are increased substantially so as to result in pulmonary overload, the results can no longer be considered relevant for human risk assessment. The Lovelace study (NTP. 1992) is a case in point. Paracelsus recognized some 450 years ago. "dosis solum venenum facit" (only the dose makes a poison). Lee et al (1985) confirmed this dramatically by inducing a significant incidence of squamous cell carcinoma in lungs of female CD rats, exposed for up to 104 weeks to 250 mg titanium dioxide/m3. a substance long considered by many to be biologically inert and frequently used as negative dust control. Should titanium dioxide therefore be considered a carcinogen in rats--and in humans? There is no evidence in the literature to suggest that cosmetic talc, inhaled under conditions of normal use. can cause cancer in the human respiratory system. Accidental inhalation of baby powder by infants presents a largely avoidable problem if physicians, nurses, parents and other individuals handling the infants are made aware of the potential danger and observe appropriate precautions, such as keeping the powder can out of reach of children. Consider ing the very large number of dustings administered daily to babies, these incidences fortunately are very rare. 1182 A . P. W e h n e r Other studies In vitro studies have shown that talc is fibrogenic and not genotoxic. The number of papers reporting talc granulomas caused by powdered surgical gloves indicates a certain concern. Calls for talc-free surgical gloves have been voiced. Cornstarch powder does not appear to be a suitable substitute as it, too, causes lesions. The use of talc on surgical gloves presents a special problem that does not affect the typical consumer of cosmetic talc. Ingestion of food-grade talc is unlikely to be toxic. Conclusion There is no conclusive evidence in the literature reviewed to indicate that cosmetic talc, when used as intended, presents a health hazard. Acknowledgements--Johnson & Johnson Consumer Prod ucts, Inc. granted permission to publish this manuscript which is based on a critical literature review previously prepared for Johnson & Johnson. The Cosmetic, Toiletry and Fragrance Association (CTFA) authorized the inclusion of excerpts in this review from BEC's 1992 technical report to CTFA. written by E. L. Alpen. A. J. Gross, G. Oberdorster, F. J. C. Roe, R. K. Ross and A. P. Wehner. REFERENCES Al-Sheikhli A. R. J. (1978) Bilateral post-intubation talc granulomata of the vocal cords. Journal o f Laryngology and Otology 92, 735. Aylott R. !.. Byrne G. A.. Middleton J. D. and Roberts M. E. (1979) Normal use levels of respiratory cosmetic talc: preliminary study. International Journal o f Cosmetic Science I, 177. Anani P. A., Ribaux C. and Gardiol D. (1987) Unusual intestinal talcosis. American Journal o f Surgical Pathology 11, 890. Baylis M. S., Henderson W. J., Pierrepoint C. G. and Griffiths K. (1986) The aetiology of ovarian cancer. In Gynaecological Oncology. Edited by C. P. Morrow and G. E. Smart, pp. 157-000. Springer-Verlag, Berlin. BEC (1992) Critiques o f N TP Technical Report on the Toxicology and Carcinogenesis Studies o f Talc in F344/N Rats and B6C3F, Mice and Perineal Exposure to Talc and Ovarian Cancer. BEC Technical Report to CTFA. Richland, Washington, September. Beck B. D., Feldman J. A., Brain J. D et at. (1987) The pulmonary toxicity of talc and granite dust as estimated from an in vivo hamster bioassay. Toxicology and Applied Pharmacology 87, 222. Beck W. C. (1992) Issues related to surgical gloves. Biomedical Instrumentation and Technology 26, 225. Beezhold D. H. (1992) Latex allergy. Biomedical Instru mentation and Technology 26, 238. Booth M., Beral V. and Smith P, (1989) Risk factors for ovarian cancer: a case-control study. British Journal o f Cancer 60, 592. Brouillette F. and Weber M. L. (1978) Massive aspiration of talcum powder by an infant. Canadian Medical Associ ation Journal 119, 354. Butenandt I,, Dumling S. and Hager C. (1981) Akzidentelle Babypuder-Aspiration. Monatsschrift f r Kinderheilkunde 129, 605. Chalmers T. C., Hewett P,, Reitman D. and Sacks H. S. (1989) Selection and evaluation of empirical research in technology assessment. International Journal o f Tech nology Assessment and Health Care 5, 521. Chappell A. G., Johnson A.. Charles J. et al. (1979) A survey of the long-term effects of talc and kaolin pleurodesis. British Journal o f Diseases o f the Chest 73, 285. Chen Y., Wu P. C , Lang J. Vi., Ge W. J,, Hartge P. and Brinton L. A. (1992) Risk factors for epithelial ovarian cancer in Beijing, China. International Journal o f Epidemi ology 21, 23. Cotton W. H. and Davidson P. J. (1985) Aspiration of baby powder. New England Journal o f Medicine 313, 1662. Cralley L. J., Key M. M,, Groth D. H,, Lainhart W. S. and Ligo R. M. (1968) Fibrous and mineral content of cosmetic talcum products. American Industrial Hygiene Association Journal 29, 350. Cramer D. W,, Harlow B. L,, Willett W. C.. Welch W R,, Bell D. A., Scully R. E.. Ng W. G. and Knapp R. C. (1989) Galactose consumption and metabolism in relation to the risk of ovarian cancer. Lancet 2, 66. Cramer D. W,, Welch W. R., Scully R. E. and Wojciechow ski C. A. (1982) Ovarian cancer and talc. A case-control study. Cancer 50, 372. Cruthirds T. P., Cole F. H. and Paul R. N. (1977) Pulmon ary talcosis as a result of massive aspiration of baby powder. Southern Medical Journal 70, 626. Davies R.. Skidmore J. W,, Griffiths D. M. and Moncrieff C. B. (1983) Cytotoxicity of talc for macrophages in vitro. Food and Chemical Toxicology 21, 201. De La Rocha S. R. and Brown M. A. (1989) Normal pulmonary function after baby powder inhalation causing adult respiratory distress syndrome. Pediatric Emergency Care 5, 43. Dietl J., Buchholz F. and Stoll P. (1986) Das ovanclle Deckepithel und seine histogenetische Beziehung zum Ovarialkarzinom. Geburtshilfe und Frauenheilkunde 46, 561. Ellis H. (1990) The hazards of surgical glove dusting powder. Surgery, Gynecology and Obstetrics 171, 521. Endo-Capron S., Renicr A.. Janson X.. Kheuang L. and Jaurand M. C. (1993) In vitro response of rat pleural mesothelial cells to talc samples in genotoxicity assays (sister chromatid exchanges and DNA repair). Toxicology in Vitro 7, 7-14. FAO/WHO (1987) Evaluation o f Certain Food Additives and Contaminants. Thirtieth Report o f the Joint FAO /W H O Expert Committee on Food Additives. WHO Technical Report Series 751. WHO, Geneva. Fay M. F. and Sullivan R. W. (1992) Changing require ments for glove selection and hand protection. Biomedical Instrumentation and Technology 26, 227. Feigin D. S. (1989) Misconceptions regarding the pathogen icity of silicas and silicates. Journal o f Thoracic Imaging 4, 68. Golightly L. K., Smolinske S. S.. Bennett M. L.. Sutherland E. W., Ill and Rumack B. H. (1988) Pharmaceutical excipients. Adverse effects associated with 'inactive' ingre dients in drug products (Part 11). Medical Toxicology and Adverse Drug Experience 3, 209. Greene M. H,, Clark J. W. and Blayney D. W. (1984) The epidemiology of ovarian cancer. Seminars in Oncology 11, 209. Gutermuth M,, Schirg E.. Steinbacher D. and Weiss H. (1980) Akzidentelle Aspiration von Babypuder Intensivmediitn 2, 83. Hamilton T. C,, Fox H., Buckley C. H.. Henderson W. J. and Griffiths K. (1984) Effects of talc on the rat ovary British Journal o f Experimental Pathology 65, 10!. Harlow B. L.. Cramer D. W,, Bell D. A. and Welch W. R. (1992) Perineal exposure to talc and ovarian cancer risk. Obstetrics and Gynecology 80, 19. Harlow B. L. and Weiss N. S. (1989) A case-control study of borderline ovarian tumors: the influence of perineal Biological effects of cosmetic talc 1183 exposure to talc. American Journal of Epidemiology 130, 390. Hartge P.. Hoover R,, Lesher L. P. and McGowan L. 1983) Talc and ovarian cancer. Journal of ihe American Medical Association 250, 1844. Hayden G. F. and Sproul G. T. (1984) Baby powder use in infant skin care. Parental knowledge and determinants of powder usage. Clinical Pediatrics 23, 163. Healey G. B. and McDonald D. F. (1977) Talc granuloma presenting as a testicular mass. Journal of Urology 118, 122. Henderson W. J. and Griffiths K. (1979) Talc on surgeons' gloves Getter). British Journal of Surgery 66, 599. Henderson W. J., Hamilton T. C., Baylis M. S.. Pierrepoint C. G. and Griffiths K. (1986) The demonstration of the migration of talc from the vagina and posterior uterus to the ovary in the rat. Environmental Research 40, 247. Henderson W. J., Hamilton T. C. and Griffiths K. (1979) Talc in normal and malignant ovarian tissue. Lancet I, 499. Henderson W. J., Joslin C. A. F., Turnbull A. C. and Griffiths K. (1971) Talc and carcinoma of the ovary and cervix. Journal of Obstetrics and Gynaecology of the British Commonwealth 78, 266. Henderson W. J,, Melville-Jones C., Wilson D. W. and Griffiths K. (1978) Oxygen incineration and electron microscope X-ray microanalysis of mineral particles in biological tissues. Journal of Histochemistry and Cytochemistry 26, 1087. Hildick-Smith G. Y. (1976) The biology of talc. British Journal of Industrial Health 33, 217. Hildick-Smith G. Y. (1977) Talc--recent epidemiological studies. In Inhaled Panicles IV. Edited by W. H. Walton, pp. 655-665. Pergamon, Oxford. Hollinger M. A. (1990) Pulmonary toxicity of inhaled and intravenous talc. Toxicology Letters 52, 121. Joly D. J.. Lilienfeld A. M.. Diamond E. L. and Brass I. D. J. (1974) An epidemiologic study of the relationship of reproductive experience to cancer of the ovary. American Journal of Epidemiology 99, 190. Kaiser W.. Otten G.. Kohnlein H. E and Kitzinger H. (1982) Unspezifische Gewebsreaktionen durch Hand schuh-Puder. Fortschriite der Medizin 100, 1213. Kamffer W. J.. Jooste E. V.. Nel J. T. and de Wet J. 1. (1992) Surgical glove powder and intraperitoneal adhesion for mation. South African Medical Journal 81, 158 Khan M. A.. Brown J. L.. Logan K. V. and Hayes R. I. (1983) Suture contamination by surface powders on surgi cal gloves. Archives of Surgery 118, 738. Kupryjanczyk J. (1989) Adenomatoid tumour of the ovary and uterus in the same patient. Zentralhlatt fur allgemeine Pathologie und pathologische Anatomic 135, 437. Lee K. P.. Trochimowicz H. J. and Reinhardt C F. (1985) Pulmonary response of rats exposed to titanium dioxide by inhalation for two years. Toxicology und Applied Pharmacology 79, 179. Longo D. L. and Young R, C. (1979) Cosmetic talc and ovarian cancer. Lancet ii, 349, Lord G. H. (1978) The biological effects of talc in the experimental animal: a literature review. Food and Cosmetics Toxicology 16, 51-57. Loup P.. Besson A.. Critsotakis J. and Mosimann R. (1979) Pentonite a amidon. Helvetica chirurgica acta 45, 733. McCormick M. A.. Lacoulure P. G.. Gaudreault P. and Lovejoy F. H. (1982) Hazards associated with diaper changing. Journal of the American Medical Association 248, 2159. McGowan L.. Parent L.. Lednar W. and Norris H. J (1979) The woman at risk of developing ovarian cancer Gyneco logic Oncology 7, 325. Mofenson H. C., Greensher J.. DiTomasso A. and Okun S. (1981) Baby powder--a hazard! Pediatrics 68, 265. Mostafa S. A. M.. Bargeron C. B.. Flower R. W.. Rosen- shein N. B., Parmley T. H and Woodruff J. D. (1985) Foreign body granulomas in normal ovaries. Obstetrics und Gynecology 66, 701. Motomatsu K.. Adachi H. and Uno T. (1979) Two infant deaths after inhaling baby powder. Chest 75, 448. Mussi N. C.. Borghetti V. L. M.. Mussi M. M. et at. (1979) Aspiracao aguda de talco. Journal of Pediatrics 47, 233. National Toxicology Program (1992) NTP Technical Report on the Toxicology and Carcinogenesis Studies of Tate in F344IN Rats arid B6C3F, Mice. NIH Publication No. 92-3152. National Institutes of Health, Bethesda. MD. Pairaudeau P. W.. Wilson R. G.. Hall M. A. and Milne M. (1991) Inhalation of baby powder: an unappreciated hazard. British Medical Journal 302, 1200. Pelting D. and Evans J. G. (1986) Long-term peritoneal tissue response in rats to mould-release agents and lubri cant powder used on surgeons' gloves. Food and Chemical Toxicology 24, 425-430. Pfenninger J. and D'Appuzzo V. (1977) Powder aspiration in children. Report of two cases. Archives of Disease in Childhood 52, 157. Pickrell J. A.. Snipes M. B.. Benson J. M. et al. (1989) Talc deposition and effects after 20 days of repeated inhalation exposure of rats and mice to talc. Environmental Research 49, 233. Pratt P. C.. George M. H.. Maslin J. P. and Roggli V. L. (1985) Crystalline foreign particulate material in hernia sacs. Human Pathology 16, 1141. Rosenblatt K. A.. Szklo M. and Rosenshein N. B. (1992) Mineral fiber exposure and the development of ovarian cancer. Gynecologic Oncology 45. 20. Rumack B. H. (1982) Diapers and poisons (editorial). Journal o f the American Medical Association 248, 2164. Russell R. S.. Merz R. D.. Sherman W. T. and Sivertson J. N. (1979) The determination of respirable particles in talcum powder. Food and Cosmetics Toxicology 17, 117-122. Scatangc J. C. and Stitik F. P. (1988) Induction of thoracic malignancy in inorganic dust pneumoconiosis. Journal of Thoracic Imaging 3, 67. Sheikh K. M. A.. Dugal K.. Relfson M.. Gignac S. and Rowden G. (1984) An experimental histopathologic study of surgical glove powders. Archives of Surgery 119, 215 Simsek F . Tiirkeri L., Ilker Y.. Kiillii S. and Akdas A. (1992) Severe obstruction of the urinary tract due to talcum powder granuloma after surgery. International Urology and Nephrology 24, 31 Sparrow S A and Hallam L. A (1991) Talc granulomas (letter). British Medical Journal 303. 6793. Swanson-Biearman B.. Cobaugh D. J., Dean B. S. and Krenzelok E. P. (1991) Massive talc aspiration: it still occurs. Veterinary and Human Toxicology 33, 383. Tolbert T. W and Brown J L. (1980) Surface powders on surgical gloves. Archives of Surgery 115, 729. Tzonou A.. Polychronopoulou A.. Hsieh C.. Rebelakos A.. Karakatsani A. and Trichopoulos D. (1993) Hair dyes, analgesics, tranquilizers and perineal talc application as risk factors for ovarian cancer. International Journal of Cuncer 55, 408. Tukiainen P.. Nickels J.. Taskinen E. et al. (1984). Pulmon ary granulomatous reaction: talc pneumoconiosis or chronic sarcoidosis? British Journal ofIndustrial Medicine 41, 81. Urdiales Cabal G.. Lamamie de Clairac E.. Sampedro Nuno A and Diaz-Faes Cervero M. (1989) Peritonitis granulo matosa por almidon. A proposito de un caso. Revista Espaola tie las Enfermedudes del Aparuto Digestivo IS, 411 Venter P F (1981) Ovarian epithelian cancer and chemical carcinogenesis. Gynecologic Oncology 12, 281 Wagner J C.. Berry G.. Cooke T J.. Hill R. J.. Pooley F D and Skidmore J W (1977) Animal experiments 1184 A. P. W lH N I R with talc. In Inhaled Particles. Vol. IV. Part 2. Edited by W. H. Walton and B, McGovern, pp. 647 654. Pergamon. Oxford. Wagner T. J. and Hindi-Alexander M. (1984) Hazards of baby powder? Pediatric Nursing 10, 124. Wehner A. P.. Hall A. S.. Weller R. E.. Lcpel E. A. and Schirmer R. E. (1985) Do particles translocate from the vagina to the oviducts and beyond? Food and Chemical Toxicology 23, 367-372. Wehner A. P.. Tanner T. M. and Buschbom R. L. (1977a) Absorption of ingested talc by hamsters. Food and Cosmetics Toxicology 15, 453-455. Wehner A. P., Weller R. E. and Lepcl E. A. (1986) On talc translocation from the vagina to the oviducts and beyond. Food and Chemical Toxicology 24, 329-338. Wehner A. P.. Wilkerson C. L,, Cannon W. C., Buschbom R. L. and Tanner T. M. (1977b) Pulmonary deposition, translocation and clearance of inhaled neutron-activated talc in hamsters. Food and Cosmetics Toxicology 15, 213-224. Wehner A. P., Zwicker G. M.. Cannon W. C., Watson C. R. and Carlton W. W. (1977c) Inhalation of talc baby powder by hamsters. Food and Cosmetics Toxicology 15, 121-129. Weissberg D. and Kaufman M. (1986) The use o f talc for pleurodesis in the treatment of resistant empyema. Annals o f Thoracic Surgery 41, 143. Wells I. P.. Dubbins P. A and Whimstcr W F. (1979) Pulmonary disease caused by the inhalation of cosmetic talcum powder. British Journal o f Radiology 52, 586. West R. O. (1966) Epidemiologic study of malignancies of the ovaries. Cancer 19, 1001. White N. D. (1992) Understanding latex. Biomedical Instru mentation and Technology 26, 232. Whittemore A. S.. Wu M. L,, Paifenbarger R. S.. Jr. Sarles D. L.. Kampert J. B., Grosser S., Jung D. L., Ballon S. and Hendrickson M. (1988) Personal and environmental characteristics related to epithelial ovarian cancer. II. Exposure to talcum powder, tobacco, alcohol, and coffee. American Journal o f Epidemiology 128, 1228. Wilkerson C. L., Wehner A. P. and Rancitelli L. A. (1977) Leaching of radionuclides from neutron-activated talc in serum and in dilute hydrochloric acid. Food and Cosmetics Toxicology 15, 589-593. Wilson D. F. and Garach V. (1981) Surgical glove starch granuloma. Ora! Surgery, Oral Medicine. Oral Pathology 51, 342. Witmeyer R. J. (1992) Trends in glove product and process development. Biomedical Instrumentation and Technology 26, 235. Yusuf S,, Peto S. R.. Lewis J., Collins R. and Sleight P. (1985) Beta blockade during and after myocardial infarc tion: an overview of the randomized trials. Journal o f Progressive Cardiovascular Disease 27, 335. Am erican Journal of Epidem iology C o o y ritjn i C> 19 9 2 b y lir e J o h n s H o p k in s U n iv e rs iiy S o l tool o l H y g ie n e a n d P u t lie H e a lth All n g n is reserved ur*ai V o l 136. No 10 Printed m U S * Characteristics Relating to Ovarian Cancer Risk: Collaborative Analysis of 12 US Case-Control Studies IV. The Pathogenesis of Epithelial Ovarian Cancer Alice S. Whittemore.' Robin Harris.' Jacqueline Itnyre.' and the Collaborative Ovarian Cancer Group' /i/| Two hypotheses have been proposed to explain the reduced risk of epithelial ovarian I cancer associated with pregnancy and oral contraceptive use. The first states that some sequelae of ovulation increase the likelihood of malignancy and that pregnancies and oral contraceptives protect by suppressing ovulation. The second hypothesis states !that circulating levels of pituitary gonadotropins increase the risk of malignancy and that pregnancies and oral contraceptives protect by suppressing secretion of these hormones. The authors evaluate the two hypotheses in light of combined data from 12 United States case-control studies of epithelial ovarian cancer in white women con ducted from 1956 to 1986. While a number of observations support both hypotheses, there are exceptions. Differential risk reduction associated with pregnancy and oral contraceptive use (pregnancy being the more effective in young women and the less effective in older women) conflicts with the first "ovulation" hypothesis, while reduced risk associated with breast feeding and absence of altered risk associated with estrogen replacement therapy conflicts with the second "gonadotropin" hypothesis. Several findings would not have been predicted by either hypothesis, e g., only weak trends relate cancer risk to age at menarche. and, among older women, no clear trends relate risk to age at menopause Odds ratio attenuation due to errors in reporting personal characteristics may be responsible for some of these inconsistencies. Multidisciplinary research is needed to clarity the etiologic roles of ovulation and gonadotropin stimulation, both of which may enhance carcinogenesis in the ovarian epithelium. Am J Epidemiol 1992;136:1212-20. gonadotropins; ovarian neoplasms; ovulation; reproduction Pregnancy and oral contraceptive use are ovarian cancer. These associations appear to associated with reduced risk of epithelial reflect more than a correlation between low Received lor publication August 21. 1991. and in Imal torin July 23. 1992. Abbreviation FSH. follicle stimulating hormone. ' Division of Epidemiology. Department ol Health Re search and Policy, Stanford University School of Medicine. Stanford. CA. 2 Members of the Collaborative Ovarian Cancer Group Dr John T Casagrande, Department of Preventive Medi cine, University ol Southern California, Los Angeles. CA. Dr Daniel Cramer. Department ol Obstetrics and Gyne cology. Bngham and Women's HospilAl, Boston, MA; D r Patricia Hartge. Environmental Epidemiology Branch, Na lionai Cancer Institute. Bethesda, Dr Jennifer L Kelsey. Division of Epidemiology, Department of Health Research and Policy. Stanford University School of Modi erne, Stanford. CA; Dr. Manon Lee, Department of Epide miology. University of California. San Francisco. San FranciS(J0. CA. Dr Nancy C Lee. Women's Health and Fertility Branch. Division of Reproductive Health. Cent for Disease Control. Atlanta. GA; Dr. Joseph L Lyon. Department of Family and Community Medicine. The U> versify of Utah Medical Center. Salt Lake City. UT; & James R Marshall. Department of Social and Prevent** Medicine. State University of New York at Buffalo School of Medicine. Buffalo. NY. Dr. Larry McGowan, Division d Gynecologic Oncology. Department of Obstetrics and Gy necology. George Washington University Medical Cent* Washington, DC; Dr. Philip C. Nasca. New York Stale Department of Health. Bureau ol Cancer Epidemiology. School of Public Health. Department of Epidemiology. Albany, NY. Dr. Ralph S Paffenbarger. Jr.. Division oI Epidemiology. Department of Hoalth Research and Potay. Stanford University School of Medicine. Stanford. CA, Dr Lynn Rosenberg. Slone Epidemiology Unit. School of Put> lie Health. Boston University School of Medicine. BrooMne MA; and Dr. Noel S Weiss. Deparlment of Epidemiology 194 0 parity and some form of infertility that pre disposes to the disease. Two hypotheses pro ' pose direct protective effects of these factors. Fathalla (1, 2) hypothesized that ovarian carcinogenesis involves some mechanical se quelae of ovulation, such as trauma or mi totic stimuli to the ovarian epithelium. This `ovulation" hypothesis suggests that preg nancy and oral contraceptive use protect against ovarian cancer by inhibiting ovula tion. Alternatively, Gardner (3) and Stadel (4) hypothesized that exposure of the ovar ian epithelium to persistently high circulat ing levels of pituitary gonadotropins in creases the likelihood of malignancy. This `gonadotropin" hypothesis suggests that pregnancy and oral contraceptive use pro tea against ovarian cancer by inhibiting pi tuitary secretion of gonadotropins. Animal experiments provide evidence to support both hypotheses. In domestic fowl, stimulating egg production induces ovarian adenomas (5), and in rodents, anatomic al terations that result in increased gonadotro pinsecretion enhance ovarian tumorigenesis (6). However, because these experimentally induced tumors are either uncommon or nonexistent in humans, their relevance to the epithelial cancers that comprise the ma jority of human malignancies is unclear. Here we evaluate the hypotheses in light of combined data from 12 United States casecontrol studies of epithelial ovarian cancer in white women conducted from 1956 to 1986 and described in the previous two pa pers. PREGNANCIES Pregnancy induces both anovulation and suppression of pituitary gonadotropins (7, School of Public Health and Com m unity Medicine. Univer*ty ol Washington. Seattle. WA Project Consultant: Of Genrose D. Copley. Extramural Programs. Division ol Can car Etiology. National Cancer Institute, Bethesda, MD Reprint requests to Dr. Alice S. Whittemore. Stanford University School ot Medicine. Department of Health Re March and Policy. HRP Modular no 2, Stanford CA 94305-5092 Supported by National Cancer institute grants CA 0689. CA 47427. and CA 47448 The authors are graletul to Drs Sally Glaser. Esther M. John, Pamela Horn Ross, and Carolyn Westal tor helplul comments on earlier versions ol the manuscript 8). Thus, both hypotheses predict that preg nancies reduce ovarian cancer risk. In sup port of this prediction, each additional term pregnancy (even among the highly parous) is associated with reduced risk of both in vasive cancer (9, table 4) and cancers of low malignant potential (10, table I). The data suggest that each additional pregnancy after the first confers the same percent reduction in risk of invasive cancer, estimated to be about 14 percent. For the population-based studies, this reduction is smaller (p < 0.001 ) than the 40 percent reduction found for the first term pregnancy. In terms of the two hypotheses then, the population-based data suggest that some factor in addition to a period of anovulation or gonadotropin suppression distinguishes uniparous women from nulliparous women. Both the ovulation and the gonadotropin hypotheses predict that failed pregnancies protect against ovarian cancer, perhaps to a lesser degree than do term pregnancies. The data for both invasive and borderline can cers support this prediction. Failed pregnan cies are associated with reduced risk among the parous (9, table 5 and 10, table I), al though no clear effect of gravidity is evident among the nulliparous. The risk reduction per pregnancy is smaller in magnitude for failed pregnancies than for term pregnan cies. However, data on gestational length for each pregnancy, available from a subset of the studies, suggest that the decreased pro tection associated with a failed pregnancy is due to its shorter length: Among the gravid, odds ratios per month of pregnancy do not depend on pregnancy outcome. In conclu sion then, findings from the combined data concerning the relation of pregnancies to ovarian cancer risk support both hypotheses. Both hypotheses, and especially the go nadotropin hypothesis, also receive support from the observed increased risk of both invasive and borderline ovarian cancer as sociated with use of fertility drugs (9, table 3 and 10, table 1). Such drugs stimulate ovulation by increasing follicular-phase lev els of follicle-stimulating hormone (l-SII). Clues to pathogenesis arc provided by the complications of fertility drugs, which in clude multiple gestations (supportive of the 1214 Whittemo. al ovulation hypothesis) and ovarian enlarge ment from FSH hyperstimulation (support ive of the gonadotropin hypothesis) (II, 12). BREAST FEEDING Breast feeding induces partial inhibition of ovulation in lactating women ( 13. I4). Thus,the ovulation hypothesis predicts that breast feeding reduces ovarian cancer risk. Breast feeding also induces increased secre tion of FSH and reduced secretion of lutein izing hormone (15). In lactating women, FSH rises after childbirth to normal follicu lar phase values and remains elevated until the return of ovarian estrogenic function (16). Thus, the gonadotropin hypothesis pre dicts that breast feeding, with its concomi tant elevated FSH levels, increases risk. Yet the data for both invasive cancers (9, table 7) and cancers of low malignant potential (I0, table I) show reduced risk associated with breast feeding. So the data on breast feeding, while consistent with the ovulation hypothesis, conflict with the gonadotropin hypothesis. It should be noted however, that lactation is also associated with ovarian re fractoriness to FSH stimulation (15), and this characteristic may be responsible for its apparent protective effect. The ellectivcness of lactation in suppress ing ovulation is strongest during the first few months after delivery and wanes thereafter (15). The ovulation hypothesis thus predicts that a month of lactation shortly after deliv ery reduces a woman's ovarian cancer risk more than does a month of subsequent lac tation. Data available front six of the studies support this prediction. AGE AT MENARCHE AND AGE AT MENOPAUSE Studies in the United States ( 17) and Finarid (18) suggest that the later menarche iccurs the longer it takes to establish regular ivulatory cycles. Thus, the ovulation hv'Othesis predicts that women wjfto began tenstruating Irefore their teens IffiMe higher varian cancer risk than do women with iter menarche because the former stalled vulating earlier. By contrast, cessation of icnstrualiqgt at menopause is a poor sur rogate for cessation of ovulation, since o#latory cycles occur sporadically during the peri menopausal years. The ovulation hy pothesis therefore predicts that late age a menopause is weakly associated with in creased ovarian cancer risk. In agreemert with the prediction for menarche, the dm show trends of increased risk associated with early menarche, although the trends air weak (9, table 8). However, in disagreement with the prediction for menopause, the data show no clear trends with later cessation of menstruation (9. table 8 and 10, table 3l Among older women, those with late men opause have no altered risk of invasive caocer in the hospital-based studies and slightly decreased risk in the population-based studies. The slight decrease in risk in the population data supports, instead, the gonadotropin hypothesis because late menopause postpones exposure to elevated gonadotropin levels that accompany the menopause (19). This absence of increasing risk with increasing age at menopause among older women conflicts with the less rapid rise with age of ovarian cancer inci dence rates after age 55 years. Such deceler ation is seen even in countries where hyster ectomy is uncommon and therefore unlikely to bias postmenopausal rates downward by removing women from the population at risk (20, 21 ). Imprecision in estimates of age at first and last menses could explain at least some of the discrepancies. Consistent with this inter pretation are the stronger trends for men arche and menopause in young women compared with older women, who are less able to recall accurately menstrual events in the distant past (22-24). Indeed, the total age span within which women undergo men opause is relatively narrow, so reporting of menopausal age with large error (23) may misclassify women from one extreme of age at menopause to the other. EXOGENOUS ESTROGENS Estrogen-containing oral contraceptives suppress ovulation and reduce pituitary se cretion of gonadotropins (25). Thus, both ovulation and gonadotropin hypotheses pro- P a llloye i ic b ii u i u v ii ii o u i m did that oral contraceptive use reduces ovaron cancer risk. This prediction is supported by trends of decreasing risk with increasing duration of oral contraceptive use, even among the highly parous, as seen in refer ence1?, table 9 and reference 10, table 2. Conjugated estrogens also suppress pitui tary hormone levels in postmenopausal omen, although not to premenopausal lev; els (25-29). However, the data show no as sociation between estrogen replacement therapy and risk for ovarian cancer, regard less of tumor behavior (9, table 10 and 10, Bble 2). The failure of estrogen replacement therapy to reduce risk argues against the jonadotropin hypothesis, unless gonadotro pin levels in estrogen users are still above a threshold needed to facilitate carcinogenesis. PELVIC SURGERIES Tubal ligation and hysterectomy with ovarian conservation may impair ovarian function, possibly by compromising blood supply to the ovaries (30-35). Thus, they should protect against ovarian cancer, ac cording to the ovulation hypothesis. More over, hysterectomy during the reproductive years should protect more than hysterec tomy after the menopause. These predic tions are confirmed by the observed reduced risks associated with both types of surgery (9, table 11and 10, table 3), although several sources of bias (9) must be considered in interpreting the findings. In contrast, the risk reductions conflict with an increased risk predicted by the gonadotropin hypothesis, unless any estrogen deficiency after surgery (34, 36) fails to elevate gonadotropin levels above a threshold needed for carcinogenesis. TOTAL YEARS OF ANOVULATION We next evaluate predictions of the two hypotheses for associations between risk of epithelial ovarian cancer and periods of an ovulation and for the variation in strength of these associations with factors such as reference age (i.e., age at diagnosis or inter view), source of anovulation, age at anovu lation, and adiposity. Here we use the term "anovulation" to mean suppression of men ses due to pregnancy, oral contraceptive use, delayed menarche, or early menopause. In this section, we restrict attention to invasive cancers (9), whose numbers were large enough to permit evaluation of such inter actions. Reference age Pike (37) hypothesized that ovarian can cer incidence rates / are proportional to ovarian epithelial "tissue age" /, measured in units of cell mitoses: 1( 1) = </. where c is a constant. Ovulation induces a transient increase in mitotic activity in the ovarian epithelium (1); therefore, a year of anovulation reduces ovarian tissue age by the number / of mitoses avoided. Pike's hypothesis then predicts that each year of anovulation reduces the ovarian cancer in cidence rate by a fixed (absolute) amount: 1 (1 )- 1 ( 1 - /,,) = cl - lit - i,,) = cl,,. Since a woman's tissue age increases as she ages, the prediction implies that the propor tional risk reduction associated with a year of anovulation wanes with age: [/(/) - 1(1 - /,,)]//(/) = c ljc t 0 as l increases. This implication is supported by the attenuated odds ratios among older women (reference age, greater than 55 years) compared with younger women, shown in table I. This table presents odds ratios among younger and older women according to total years of ovulation, estimated by subtracting age at menarche from age at last menstrual period and then further deducting total lime pregnant in term pregnancies (as suming 9 months per pregnancy), breast feeding, or taking oral contraceptives. Women hysterectomized before the meno pause were assumed to continue ovulating until age 55 years, fable I shows strong and statistically significant trends of increasing risk with increasing years of ovulation for younger, but not older women. The trend in older women showed interstudy heteroge- 1216 Whiltemofc al wl i-'a ll t u y c i ic b ib u i u v a iia n o d iK u 11 TABLE 1. Odds ratios (OR) for invasive epithelial ovarian cancer according to estimated years of ovulation, by reference age Reference age (years) Years ot ovulation Cases No. % <55f Controls No. % OR 95% Cl B P value Cases No % 55$ Controls No % OR 95% Cl P valu <25 25-29 30-34 >35 Overall trend 208 30 2.099 47 1.0 131 19 804 18 1.8* 198 29 962 22 2.6* 145 21 595 13 2 9 * 1 4-2.5 1 9-3.6 2 0-4.1 24 5 90 6 1.0 24 10 160 11 1.3 121 23 363 25 1 2 318 62 826 58 1.5 0.72-2.3 0.72-2.1 0.87-2.4 per year p< 0.001. 1.081 < 0 001 1 .0 2 # 0.10 t Based on studies 2. 6. 8. 9, 11. and 12 (see part I. table 1 (50)). $ Based on studies 2. 6. 8. 9. and 12 Adjusted for age and study ACl. confidence interval 1 Test of OR homogeneity across studies: = 5.8. p = 0 38 * Test of OR homogeneity across studies: x2 = 11 9. p = 0.04. TABLE 2. Percent reduction of Invaeive epithelial ovarien cancer risk per year ot anovulation according to turce, by reference age Reference age (years) Source of anovulation <55 % reduction SEt 55 % reduction SE Delayed:) m enarche Term pregnancy Breast feeding) Oral contraceptive u se l Early m enop ause# 2.1 27 9 ** 8.5 8 .0 ** -1 .0 2.4 4.3 9.2 1.7 1.7 0.8 12.2 10.3 2 0 .4 * 0.4 3.0 5.2 87 7.9 1.0 i *p < 0.01; * * p < 0.001 t SE. standard error, f After age 10 years. (Among the parous, adjusted for age. study, and oral contraceptive use. and assuming 9 months per term pregnancy | Among women who had breast-fed. adjusted for age. study, parity, and oral contraceptive use. 1Among users, adjusted for age. study, and parity. # Before reference age for those aged <55 years and before age 55 years for those aged >55. Women with artifidal menopause Mr excluded neily (p = 0.04); data from one study showed a stronger trend than did data from the others. Similar trends were obtained when women were classified by the comple mentary years of anovulation (relative to a maximum 45 years of ovulation for a nulligravid woman with menarche and meno pause at ages 10 and 55 years, respectively, who never used oral contraceptives). In additional age-specific analyses (not shown), measures of risk associated with pregnancy showed strong and consistent trends of attenuation with increasing refer ence age. Specifically, the percent risk reduc tion associated with parity relative to nulli parity declined steadily, ranging from 28 percent for women aged less than 40 years to only 1 percent for women aged 70 or more. Indeed, there was little difference in risk by parity among women aged 60 years or more. Similarly, the risk reduction among the parous associated with each term preg nancy declined with age. These trends of decreasing percent risk reduction with in creasing age were present ( p < 0.01) for both hospital-and population-based studies. Such attenuation supports Pike's hypothesis. It also is consistent with a transie^ protective effect of gonadotropin sup p rt& in during pregnancy. However, it contradicts the con jecture that pregnancy induces changes that continue to reduce ovarian cancer risk throughout life. If the conjecture were true, then the percent risk reduction per preg nancy would remain constant or increase, rather than decrease, with age. The percent risk reduction associated with a year of delayed menarche also tended to wane with increasing reference age, but the trends are more readily explained by chance. Risk reductions associated with breast feed ing did not vary by reference age, while those associated with oral contraceptive use showed a nonsignificant increase with age Source of anovulation According to the ovulation hypothesis, ovarian cancer risk depends on a woman's reproductive and menstrual history only as a function of her total years of anovulation, rather than as a function of the individual components from the various sources. Table 2 compares estimates of the percent risk reduction per year of anovulation due to delayed menarche, term pregnancy among the parous, breast feeding among those who had breast-fed, oral contraceptive use among users, and early menopause among those who were naturally postmenopausal or pre menopausal. Among younger women, the magnitude of risk reduction per year of term pregnancy exceeds that associated with the other sources. Among older women, how ever, pregnancy is less protective than is oral contraceptive use. Reporting accuracy is probably greater for time spent pregnant than for time from all other anovulatory sources given in table 2. Since random recall error attenuates regression coefficient esti mates toward zero, such error could explain some of the observed differences. In developed countries, breast feeding suppresses ovulation less effectively than does pregnancy; some 40-75 percent of lactaring women menstruate while nursing (13). The findings in table 2 are thus consis tent with the ovulation hypothesis, which predicts that a month of breast feeding is associated with a smaller reduction in ovar iancancer risk than is a month of pregnancy. Although the differences in risk reduction between pregnancy and breast feeding do not achieve statistical significance, those be tween pregnancy and oral contraceptive use arc significant ( p < 0.01) among younger, but not older women. This difference in risk reduction in younger women, if not due to reporting errors in duration of oral contra ceptive use (38) or other bias, conflicts with the ovulation hypothesis, which predicts equal protection per year of pregnancy and oral contraceptive use, since these condi tions are equally effective in suppressing ovulation. The difference is more consistent with the gonadotropin hypothesis because the low potency oral contraceptives may be less effective than pregnancy in inhibiting pituitary secretion of gonadotropins (7, 39 41). The relatively large risk reduction as sociated with oral contraceptive use among older women has a large standard error be cause few older women had used oral con traceptives. Nevertheless, it suggests that the early high-potency formulations used by these women (42, 43) may be especially protective. Such extra protection also would argue against the ovulation hypothesis since all formulations suppress ovulation at simi lar rates (44), but would support the gonad otropin hypothesis since the high-dose for mulations may have been particularly effec tive in reducing pituitary gonadotropins (39-41). Alternatively, the large risk reduc tion accompanying previous oral contracep tive use in older women suggests that the biologic effects of oral contraceptives may increase with time. Like the previous expla nation. such a latency effect would conflict with the ovulation hypothesis, but not nec essarily with the gonadotropin hypothesis. The data in table 2 suggest that for all women, neither a year of delayed menarche nor a year of early menopause is associated with the same risk reduction as is a year of pregnancy, breast feeding, or oral contracep tive use. Among younger women, breast feeding and oral contraceptive use appear equally effective in reducing ovarian cancer risk, albeit less effective than pregnancy. 1218 Whittemore Age at anovulation The probability that a menstrual cycle is ovulatory varies with age, being highest in the age range 25-40 years and lowest at either end of the reproductive years (45). Thus, the ovulation hypothesis predicts less protection per year of menses suppression due to delayed menarche and early meno pause than per year of such suppression at the height of reproductive life. The differ ences in table 2 support this prediction. Fur ther. since women with late first birth tend to complete their childbearing later than do women wiih early first birth, the ovulation hypothesis predicts less protection per preg nancy to women with late first birth than to women with early first birth. This second prediction received weak support from the hospital-based data, but not from the popu lation-based data. Future studies should ob tain the timing of all pregnancies and epi sodes of lactation and oral contraceptive use. as well as information on menstrual cycle length, in order to compare ovarian cancer risk reduction per unit time at the height of the reproductive years with that at either extreme. Adiposity Extreme obesity before the menopause is associated with increased incidence of ano vulatory cycles (46) and lower circulating levels of pituitary gonadotropins (47). Thus, both hypotheses predict that pregnancy, lac tation, and oral contraceptive use are more protective to the lean than to the obese, for whom they prevent fewer ovulations and less exposure to FSFI. Only the data for breast feeding support this prediction; the risk re ductions associated with pregnancy and oral contraceptive use did not vary with adipos ity. However, the reduction associated with breast feeding was con fined to women whose "usual" value ofQuetelei index was less than 35 (p < 0.05 for both hospital-based and population-based studies). / SUMMARY The data relating ovarian cancer risk to pregnancies, oral contraceptive use. and use of fertility drugs support both ovulation and gonadotropin hypotheses. Yet there arc in consistencies; The observed protective effect of delayed menarche is weaker than pre dicted by either hypothesis, and the lack of clear trends with age at menopause, if not due to recall error or other bias, argues against both hypotheses. The greater risk reduction associated with pregnancy than with oral contraceptive use argues against the ovulation, but not the gonadotropin, hypothesis. Some of these inconsistencies may be due to differential measurement er ror among the sources, since pregnancies probably are recalled more accurately. Eval uation of this issue in cohort studies is of interest in light of evidence (2 1, 22) that contemporaneous report of menopausal sta tus and estrogen use may be more accurate than recall of these events later in life. The findings for breast feeding, estrogen replacement therapy, and pelvic surgeries are consistent with the ovulation hypothesis, but conflict with the gonadotropin hypoth esis. However, the endocrine profiles asso ciated with these characteristics and their contributions to ovarian carcinogenesis are poorly understood. Pending further research to clarify these issues, the findings cannot be interpreted as strong evidence against the gonadotropin hypothesis. Both hypotheses may be valid, with each explaining some fraction of all epithelial ovarian cancers. Histology-specific analysis, although not feasible for these data, may help to distinguish the relative contributions of each. Research involving both epidemi ology and the basic sciences will be needed. For example, ovarian cancer has a strong familial component (48), and work is under way to determine genetic loci that predispose to the disease. Markers of genetic suscepti bility will permit refined analyses that may elucidate mechanisms. Clinical research will also be useful. For example, epithelial ovar ian cancers have been observed in dysgenetic gonads that lack ova and are. therefore, inca pable of ovulating (49). This observation demonstrates that ovulation is not necessary for epithelial carcinogenesis. However, dys genetic gonads are exposed to elevated cir- t ail ioyui ui Uvufkali 0 % id. 19 culating levels of pituitary gonadotropins and contain binding sites to gonadotropins. Cancer Res 1978;38:4021-4. 18. A p le r D , V ih k o R. E a rly m enarchc, a risk factor for breast cancer, indicates early onset o f ovulatory cycles. J C lin E n d o crin o l M e ta b 1 9 8 3 ;5 7 :8 2 -6 . 19. C o o ke I D , A n d e rto n K J , L e m o n E. H o rm o n e p a t terns al the clim acteric. Poslgr M e d J 1976; 52(suppl 6): 12 - 6 . REFERENCES 20. A d a tn i H - O , Bergstrom R , Persson I, el al. T h e incidence o f o v a ria n cancer in Sw eden, 19 6 0 -1 9 8 4 . A m J Epidem iol 1990;132:446-52. 1. Faihalla M F . Incessant o v u la lio n -a factor in ova r 21. M o h le J, W h itte m o re AS, Pike M . el al. G o n a d o ian neoplasia? (L etter). Lancet 1971 ;2 :163. tro phins a n d o va ria n cancer risk. (L e tte r). J N u ll 2. Faihalla M F . Factors in the causation and in c i Cancer Inst 198 5;75:178-9. dence o f ovarian cancer. Obstet Gynecol Surv 22. Bean JA, Leeper JD , W allace RB . el al. Variations 1972;27:751-68. in reporting o f m enstrual histories. A m J E p id em io l 3. G a rd n e r W U . Tum o rig en es is in transplanted irra 1979;109:181-5. diated and nonirradiated ovaries. J N a ll Cancer 23. Paganini-H ill A, Ross R K . R eliability o f recall o f Inst 1 9 6 1;2 6:82 9-5 4. drug usage and other h ealth-related in fo rm a tio n . 4. Stadel B V . T h e etiology and p reve ntion o f o varian A m J Epidem iol 1982; 116:114-22. cancer. A m J O bslel G ynecol 197 5;1 23:7 72-4. 24. C o lditz G A , Stam pfer M J. W illett W C . et al R e 5. W ilson JE. A d eno ca rc in o m as in hens kept in producibility and validity o f self-reported m eno a constant environm ent. Poult Sci 1958:37: pausal status in a prospective cohort study. A m J 1253-66. Epidem iol 1987;126:319-25. 6. M c G o w a n L . O avis R H . In trasplen ic o va ria n grafts 25. L a u rilz e n C. O n endocrine elTects o f oral c o n tra in Syrian hamsters and peritoneal fluid cellular ceptives. Acta Endocrinol Suppl 1968.124: d istribution. Proc Soc E x p Biol M e d 19 7 0 ; 134: 87-ICK). 507-9. 26. V a rm a T R , Everard l>. Hole D . Effect o f natural 7. Parlow A F . D a an e T A , D ig n a m W J . O n the c o n estrogen on the scrum level o f follicle-stim ulating centration o f radioim m unoassayable FS H circulat horm one (F S H ). estradiol and eslronc in post ing in blood throughout hum an pregnancy. J C lin m enopausal w o m e n and its effect on e n d o m e triu m Endocrinol M etab 1970:31:213-14. Acta O bslel G ynecol Scand 1985.64:105-9. ( 8. Jeppsson S, R a n n e v ik G , T h o re ll JI. P itu ita ry gon- 27. U lia n W H , K a tz M , Davey D A , el al. Effect o f j ado lrophin secretion d u rin g the first weeks o f preg- prem enopausal castration an d in crem en tal dosages r nancy. Acta Endocrinol 1977;85:177-88. o f conjugated equine estrogens on plasm a follicle- 5 9. W h itte m o re A S , H a rris R , lln y r e J, el al. C harac- s tim u la tin g h o rm o n e , lu te in iz in g h o rm o n e , and es ; lerislics relating to ovarian cancer risk. C ollabora- tradiol. A m J Obstet G ynecol 1978:132:297-302. live analysis o f 12 U S case-control studies. II. In 28. SchilT I T h e effects o f conjugated estrogens on vasive ep ithelial o varian cancers in w h ite w om en. gonadotropins. Ferlil Steril 1980;33:333-4 A m J Epidem iol 1992;136:1184-1203. 29. Larsson-Cohn U . Johansson E D . Kagedal B, el al 10. H arris R . W h itte m o re AS , lln y re J, et al. C h arac teristics relating to ova rian cancer risk. C o lla b o ra Serum F S H . L H and ocslrone levels in postm eno pausal patients on oestrogen therapy Br J O bslel tive analysis o f 12 U S case-control studies. I II . Epithelial tum ors o f low m alignant potential in Gynaecol 1977;85:367-72. 30. Beavis E L G . Brow n JB. S m ith M A . O varian func while w om en. A m J Epidem iol 1992:136: tion after hysterectomy w ith conservation o f the 1204-11. ovaries in pre-m enopausal wom en. J Obstet G y 11. Davis O K . R a v n ik a r V A . In d u c tio n o f o vu latio n naecol Br C onintonw 1969.76:969-78. with clom iph ene citrate. In: Barbiri R L . S c h ilT I, 31. DcStcfano F, Perlm an JA . Peterson H B . et al. eds. R ep ro d u ctive end ocrine therapeutics. N e w j; York: A la n R . Liss. 198 8:1 -24. L o n g -term risk o f m enstrual disturbances alte r tubal sterilization. A m J Obslel G ynecol 1985; 12. Y eh J, R a v n ik a r V A . In d u c tio n o f o v u la tio n w ith 152:835-41. ! hum an L H -F S H and hum an FS H . In: Barbiri R L . 32 Ellsworth I.R , Allen H H , Nisker JA. O varian func SchilT I, eds R ep ro d u ctive en d ocrine therapeutics. tion afte r radical hysterectom y for stage IB c a rc i New Y o rk: A lan R . Liss. 1 9 8 8 :2 5 -4 9 . nom a o f the cervix. A m J Obslel G ynecol 1983; 13. Perez A , V e la P, M asn ick G S , et al. First o vu latio n 145:185-8. after childbirth: the effect o f breast-feeding. A m J 33. N e il J R , H a m m o n d G T , N o ble A D . et al Late Obstet G ynecol 1972;114:1041-7. com plications o f sterilisation by laparoscopy and 14. G ra y R . Eslam i S, C a m p b e ll O . R e tu rn o f fe n ility tubal ligation A controlled study, la n c e t 1975:2: during lactatio n m o n ito re d by u rin ary steroid as 6 9 9 -7 0 0 . says. (A b stract). A m J E p id e m io l 19 8 7 ; 126:761 . 34. C a llanach J. Oestrogen deficiency after tubal liga [ 15. Reyes F I, W in g e r JS D . F a im a n C. P itu ita ry -o v a ria n relationships during the puerprium . A m J Obstet Gynecol 1972;114:589-94. tion. Lancet 1985;1:847-9. 35. Radwanska E. Headley S K , Dntowski P Evalua tion o f ovarian function after tubal sterilization J { 16. Zarate A , Canales E. T h e reproductive age: the Reprod M ed 1982:27:376-84. I hum an ovary during pregnancy and puerperium . In: Serra G B , ed. T h e ovary. N ew Y o rk: Raven Press. 1 9 8 3 :2 5 7 -7 1 . 36. Corson SL. Levinson C'J. B a t/c r F R , el al. H o r m one levels follow ing sterilization and hysterec tomy. J Reprod M ed 1981:26:363-9. 37. Pike M C . Age-related factors in cancers o f the 17. W allace R B , S h erm a n B M , Bean JA , et al. M e n - breast, ovary and end om etrium . J C hronic Dis strual cycle patterns and breast cancer risk factors. 122 win; e ei ai. 1987:40 (Suppl. 2):59S-69S. 38. C oulter A Vcsscy M. M cPherson, et al. The ability o f women to recall their oral contraceptive histo nes Conttaception 1986;33:127-37 39 D ericks-Tan JSE, Krog W, Aktories k . el al. Dosedependent inhibition by oral contraceptives o f the pituitary to release l.H and FSH in response to stimulation with LH-RH + . Contraception 1976; 14:171-81. 40. Scott JZ. K leuky OA. Brenner PF, et al. C o m p a r ison ol the effects o f contraceptive steroid form u lations containing two doses o f estrogen on pitui tary function. Fcrnl Steril 1978;30:141-5. 41 SpcUacy W N. Kalra PS. Buhi W C. et al Pituitary and ovarian responsiveness to a graded gonadotro pin releasing factor stim ulation test in women using a low-estrogen or a regular type o f oral contracep tive Ant J Obstet Gynecol 1980;137:109-15. 42. Piper JM . Kennedy DL Oral contraceptives in the U nited States: trends in content and potency. Int J Epidemiol 1987;16:215-21. 43 Van de C arr SW. K ennedy DL. Rosa FW . et al. Relationship of oral contraceptive estrogen dose to age Am J Epidemiol 1983;117:153-9. 44 (oldzieher JW . dc la Pena A. C henault CB. et al C om parative studies o f the elhynyl estrogens 4 in oral contraceptives. II. Antiovulatory potency Am J Obstet Gynecol 1975;122:619-24. 45. M elcalfM G . Incidence o f ovulation from th e n arche to the m enopause: observations of 622 N n Zealand w om en. N Z M ed J 1983:96:645-8. 46. G lass AR. B urm an K D. D ahm s WT, et al End crine function in h um an obesity. Metabolism 1981. 30:89-104. 47. Sherm an BM. K orenm an SG. Measurement scrum LH , FSH. estradiol and progesterone in d orders o f (he h um an m enstrual cycle: the made q u ale luteal phase J Clin Endocrinol Mclab 19(4. 39:145-9. 48. Schildkraul JM . Thom pson WD. Familial ovmn cancer: a population-based case-control study. Aa J Epidemiol 1988:128:456-66. 49. Miller DS, Teng NN, Ballon SC. Epithelial ovaiia carcin o m a in patients with in te rs disorders: tie role o f pituitary gonadotropins in ovarian tumors genesis. Gynecol O ncol 1986:24:299-308. 50. W hittem ore AS. H arris R. Itnyre J. et al. Clianc- teristics relating to ovarian cancer risk. Collabora tive analysis o f 12 US case-control studies. I. Mak- ods. A m J Epidem iol 19 9 2 :13 6:1175--83. American Journal o( Eptdemlotogy cCopyright 1992 Oy The Johns Hopkins University School ol Hygiene a n d P u b lic H e a lth AH rights reserved Vat 136. No to P r im e d in U S A if. j* Increased Risk of Breast Cancer with Alcohol Consumption in ;. Postmenopausal Women ,'i. Susan M. Gapstur, John D. Potter, Thomas A. Sellers, and Aaron R. Folsom The association between breast cancer Incidence and alcohol consumption among postmenopausal women was examined in the Iowa Women's Health Study. In January 1986, a cohort of 41,837 postmenopausal women, aged 55-69 years, completed a questionnaire that included alcohol intake and other information. Through December 1989, 493 incident breast cancer cases were identified. Age-adjusted relative risks of consumption of less than 1.5.1.5-4.9. 5.0-14.9, and 15.0 g or more of alcohol per day compared with abstention were 1.08, 1.10. 1.08. and 1.28, respectively (p for trend = 0.11 ). After controlling for age, body mass index, age at first livebirth, age at menarche, and family history of breast cancer, the relative risks were 1.18, 1.20, 1.25, and 1.46 (p for trend = 0.04) Multivariate modeling, using Cox proportional hazards regression, revealed a significant multiplicative interaction between alcohol intake and noncontra ceptive estrogen use. The relative risks of breast cancer associated with average daily alcohol consumption of 5.0-14.9 and 15.0 g or more were 1.88 (95% confidence interval 1.30-2.72) and 1.83 (95% confidence interval 1.18-2.85), respectively, among everusers of estrogen; no association between alcohol and breast cancer was observed among never-users of estrogen. Am J Epidemiol 1992; 136:1221 -31. alcohol drinking; breast neoplasms; cohort studies; estrogen replacement therapy . ]< Breast cancer is the most com m on cancer T, among w om en in the United States and is a .- i major public health concern. It is estimated f/ that one in nine w om en will develop breast ? cancer in their lifetimes ( l) . Thus, identify(.< ing those risk factors am enable to ntodifii' cation would be useful for im plem enting Received lor publication February 21. 1992, and in final tom May 22, 1992 Abbreviation: Cl. confidence interval. ( From the Division of Epidemiology. University of Min neeota School of Public Health, Minneapolis. MN . Reprint requests lo Dr. John D Potter. Division on (Epidemiology. University of Minnesota. School ol Public MeiHh. 1300 South Second Street, Suite 300. Minneapolis. 1*55454-1015 Supported by grant R01-CA 39742 to Dr. Aaron R Folsom from the US National Cancer Institute Susan M. Gapstur was supported by National institutes of Health tm n g grant T32 CA099607 to Dr John D Potter The authors thank Dr Susan Kaye, Dr. Robert Wallace, i Or. terry Kushi. Kathleen McKeen. and Ching Ping Hong i tortheir important contributions They also thank Dr Walter let! and Laura Sam pson for permission to use the frtorverd food frequency questionnaire primary prevention strategies. In 1977, W illiams and Horm (2) reported results from the Third National Cancer Survey, a population-based cross-sectional study o f in cident cancers, suggesting for the first time that alcohol consum ption may be related to breast cancer. This report lead to an exten sive examination o f the alcohol-breast can cer relation. In a recent review, Hiatt (3) sum m arized most o f the studies that exam ined the asso ciation between breast cancer and alcohol consum ption. He noted that 11 o f 18 casecontrol studies (2, 4 -20), five o f the six cohort studies (21-26), and one meta analysis (27) have shown a positive associa tion. Three o f four case-control studies (28 31) and one meta-analysis (32) published between 1989 and 1991 also reported a |>sitive association. In general, the relative risks and odds ratios for developing breast cancer ranged from 1.4 to 2.0 for w om en who a!> 1221 vN J i ) I K N A I - K l' ID E M lO L O tiY ti I9H8 byT n e 1John H opkins U niversity Srhool o f Hygiene and Public Health s r- served Vwl. IJH. s.. * `rtnU'dinl`> t tSONAL AND ENVIRONMENTAL CHARACTERISTICS RELATED TO EPITHELIAL OVARIAN CANCER E PK O D U C TIV E AND M EN ST R U A L E V E N T S A N D O R A L C O N T R A C E PT IV E USE M ARION I.. W U ,' A I.IC E S. W H IT T B M O K E ,' R A L PH S. P A F F E N B A R O E R , J r ..1 DOR1EN L. SA K L E S,' JA M E S B. K A M P E R T ,' S T E L L A G R O S S E R ,' D E X T E R L. J U N G ,' SA M U EL B A LLO N ,1 M IC H A E L H E N D R IC K S O N ,' a n d JANET MOHLE BOETANP Wu, M. L. (Stanford U. School of Medicine, Dept, of Health Research and Policy, Stanford, CA 94305-5092), A. S. Whittemore, R. S. Paffenbarger, Jr., D. L. Series, J. B. Kampert, S. Grosser, D. L. Jung, S. Ballon, M. Hendrickson, and J. MohleBoetani. Personal and environmental characteristics related to epithelial ovarian cancer. I. Reproductive and menstrual events and oral contraceptive use. A m J Epidemiol 1988;128:1216-27. In two case-control studies conducted in the six-county San Francisco Bay Area, 111 women diagnosed with epithelial ovarian carcinoma in 1974-1977 and 188 women diagnosed in 1983-1985 were interviewed concerning their menstrual, sexual, and reproductive histories. For comparison, interviews were conducted with 752 control women admitted to the same hospitals within six months of the cases; for cases diagnosed in the later period, interviews were also conducted with an additional 259 population-based controls selected by random digit dialing. Controls were matched to cases by age and race. Qualitative and quantitative findings were similar for the two studies. In the combined data, cases were more likely than their matched controls to have been nuliiparous, to have undergone menarche at an early age, and to have refrained from using oral contraceptives. Menopause occurred slightly later for cases than for controls, but the differences were not statistically significant. Neither age at first term pregnancy (20 or more weeks gestation) nor number of term pregnancies was predictive of ovarian cancer risk. The protection afforded by oral contraceptive use was independent of parity, and it increased with Increasing duration of use. There were no trends in risk with time since last oral contraceptive use or with time since first use, after adjustment for duration of use. These observations suggest that oral contraceptive use decreases risk for ovarian cancer, rather than merely indicates fertility, which may itself decrease risk of developing the disease. The authors combined reproductive characteristics and oral contraceptive use to estimate a woman's total duration of ovulation, which was positively associated with ovarian cancer risk (p < 0.001 for trend). These observations support the concept that the greater the duration of ovulation or accompanying endocrinologic phenomena, the greater a woman's risk for ovarian cancer. contraceptives, oral; menarche; menopause; ovarian neoplasms; ovulation; parity i from several studies suggest that protect against epithelial tumors of the ovary (1-16). T w o hypotheses have been ved for publication M ay 29,1987, and BWjkal ril 11, 1988. artm ent of H ealth Research and Policy, Stan iversity School of Medicine, Stanford, CA. ` 444 High S treet, Palo Alto, CA. ' D ep artm en t of Pathology, Stanford University Schirol of M edicine, Stanford, CA. R ep rin t req u ests to Dr. M arion L. Wu, Stanford 1218 OVARIAN CA NC E R AND KKI'KOUUCTI VE EVEN TS I a -i . proposed as possible explanations for these protective effects. The first states th at childbearing and oral contraceptive use protect by suppressing ovulation (2, 17, 18). According to this hypothesis, each ovula tion increases risk of neoplastic change, possibly by traumatizing the ovarian epi thelium or by exposing it to endogenous promoting agents, such as estrogen-rich follicular fluid or high serum levels of pi tuitary gonadotropins. The second hypothesis states that parity and oral contraceptive use are not causally related to ovarian cancer but th a t the three conditions are common correlates of subfertility. According to this hypothesis, women who have difficulty conceiving are at increased ovarian cancer risk and also are less likely to be parous or to use oral contraceptives. The first part of this hy pothesis is supported by some studies show ing that women with ovarian cancer are more likely than control women to report difficulty in conceiving (11) or to have es chewed contraceptives during marriage ( 12). Here we present data from two casecontrol studies concerning the effects on epithelial ovarian cancer risk of childbear ing, oral contraceptive use, and other re productive and menstrual factors. The first study, conducted from 1974 to 1977, will be called the 1970s study; the second study, conducted from 1983 to 1985, will be called the 1980s study. M a t e r ia l s a n d m e t h o d s 1970s study Cases for the 1970s study were white English- or Spanish-speaking residents of northern California who were 20 to 85 years of age at diagnosis. These included 111 women with histologically verified epithe lial ovarian cancer, newly diagnosed from University School o f M edicine, D e p artm en t o f H e a lth Research and Policy, H R P B uilding, Stanford, C A 94305-5092. T h is work waa supported by N 1 H G ra n t C A 35007. T h e authors th an k D r. G erald M . S hefren for his help in the study. 1974 through 1977 at one of .34 hospitals in the five counties included in the San Francisco-Oakland Bay Area Surveillance and End Results (SEEK) reporting system (19). Cases with a history of surgical meno pause were excluded. Controls in the 1970s study were while women hospitalized in one of the 34 adm it ting hospitals for the cases. Control pa tients were matched to cases by age (within five-year groups), hospital of admission, and year of admission. Women were ex cluded if they had undergone surgical meno pause or if they had been admitted or treated for psychiatric, obstetric, gyneco logic, or malignant conditions. A total of 472 control patients participated. Further details of the study design can he found elsewhere (20, 21). 1980s study Cases in the 1980s study were residents of northern California aged 18 to 74 years who were diagnosed with epithelial ovarian cuncer or with a borderline epithelial ovar ian tumor between January 1083 and De cember 1985 at one of the seven hospitals in Santa Clara County or at the University of California, San Francisco, Medical Cen ter. Slides of coses' surgical specimens were reviewed and classified by one of us (M. H.). Of the 317 eligible cases identified during the study period, eight (3 ]>er cent) were not interviewed because their physi cians denied permission, 30 (9 per cent) declined to participate, and 44 (14 per cent) were too ill to participate or died before they could be contacted. The remaining 235 women (74 per cent ) constituted the cases for this study. The tumors of 47 (20 per cent) of these women were classified a s borderline; these tumors were excluded from this analysis, leaving 188 cases for the 1980s study. Two sets of control women were selected for the 1980s study. The first set consisted of women hospitalized at one of the eight admitting hospitals for the cases. These women were matched to the cases by age (five-year groups), race (white, black, other), hospital, and dale of admission wii e t a i . thin 3 months). We attempted to loe two controls (one medical admission, : surgical admission) for each case; howr, only one control was available for ne nonwhite women. Women were exded if both their ovaries had been re: ved or if they had been admitted for /chiatric, obstetric, gynecologic, or ma; mnt conditions. Of 524 eligible women, (9 per cent) were not interviewed beise their physicians denied permission, ) (19 per cent) declined to participate, I 23 (4 per cent) were too ill or died prior being contacted. The remaining 354 men (68 per cent) constituted the set of ipital controls for the 1980s study, /enty-four of these women were matched cases with borderline tumors and are luded from this analysis, leaving 280 spital controls. The second set of controls consisted of men who were selected from the general >ulation by random digit dialing tele>ne contacts, with an attempt to reach >controls per case. These women were tched to cases by age (within five-year ups), race (white, black, other), and teleine area code and prefix. Women known have a previous bilateral salpingoihorectomy were excluded. We conted a total of 2,917 households. Of these, 25 were business numbers, were not in vice, gave no answer after 10 attempts, were households containing no eligihle men. An additional 228 households re ed to provide information about potenI eligibility of their members. Of 464 iible women contacted by telephone, 408 per cent) consented to receive a letter ilaining the study, and 329 (71 per cent) eed to participate. Seventy of these were iched to cases with borderline tumors Iare excluded from these analyses, yielda balance of 259 population controls. interviews II cases and controls participated in ictured personal interviews cwducted their homes by trained interviewers. Most case interviews were conducted within three months of diagnosis. The questionnaire used for the 1980s study was similar to the 1970s questionnaire but in cluded more questions. Both instruments were designed to assess menstrual, repro ductive, medical, and family history prior to a reference date (one year before diag nosis for cases; one year before interview for controls). In addition, the 1980s ques tionnaire included questions about expo sures to environmental agents such as talc, asbestos, coffee, tobacco, alcohol, and ra diation. To facilitate recall of oral contra ceptive use, interviewers used photographs of all oral contraceptives marketed in the United States prior to 1983. To facilitate recall of reproductive events, interviewers provided subjects with a chart organized by calendar year. Data analysis Conditional logistic regression for matched case-control studies (22) was used to estimate odds ratios (hereafter called relative risks) and to test for trend. Anal yses were conducted by means of the pro grams EGRET (23) and RISK (24). Ap proximate 95 per cent confidence limits were obtained by the Wald method (22). All p values are two-tailed. Differences in relative risks for the two components were tested by likelihood ratio methods. Under the null hypothesis of no difference, the likelihood ratio statistic (defined as minus twice the difference in log likelihood lietween regressions with and without a com mon relative risk) has approximately a chisquared distribution (22). R esults Demographic and reproductive charac teristics of study participants are shown in tahle 1. Cases in both studies were more likely to be nulliparous than were their matched controls. However, cases and con trols did not differ in average age at first term pregnancy (20 or more weeks gesta tion) or in average number of abortions. & OVAKIAN CANC ER AND KEI'KODIICTIVK KVICNTS 'V a m .h 1 C h a ra c te ristic s o f s tu d y fX irlic ip a n ts, S u n F rancisco H a y A rea , I if 7-1 I if 71 a n d lUH'i I US;> CliaracieriMlic 19708 part ta p in ila ( -MHtiH (n 111) ( 'ont roi U i = 472) 11)80* p tirlirip n iilH ( 'ilrtfri (fl " 188) Pimi rul > 1liiapital ( = 280) I 'lip u liil im (/i = 2i!0 Age (m ean years) Race (% w hite) O ra l co n ceptive use (% ever use) Parity (% nulliparous) Term pregnancies* (m ean no.) Age at m enarche (m ean year) Age at first lerm pregnancy* (m ean yeurs) Abortions (m ean no.) "55.4 100 22.5 27.9 1.8 12.8 24.8 0.4 50.8 100 21.6 19.7 2.1 13.0 24.2 0.5 52.8 94.7 45.7 20.7 2.2 12 5 22.2 0.6 52.4 97.1 50.0 17.1 2.5 12.7 22.1 0.6 52.0 94.2 57.9 10.0 25 12.8 22.5 0.6 20 or more weeks gestation. A prior history of oral contraceptive use was less prevalent among participants in the 1970s study than among those in the 1980s study. This difference reflects tem poral trends of increasing oral contracep tive use in the United States. Since women in the 1970s study were interviewed at an earlier date and were older at interview than were those in the later study, a greater fraction of their reproductive years oc curred before the introduction of oral con traceptives in the early 1960s. Characteristics considered below were similar for the two sets of 1980s controls. Therefore, we report only results obtained by comparing 1980s cases with the com bined 1980s set of hospital and population controls. Table 2 gives the distribution of cases and controls and relative risks by number of term pregnancies. Parous women were at lower risk than were nulliparous women in both the 1970s and 1980s studies, al though the difference achieved statistical significance only for the 1980s data. The relative risk associated with ever having a term pregnancy was 0.64 for the combined 1970s and 1980s data (p < 0.001). Both data sets also indicated decreased risk with increasing number of term pregnancies among the parous women, although in nei ther study was the decrease statistically significant. According to the logistic func tion fit to the combined data, each term pregnancy conferred a 7 per cent reduction in risk (p = 0.18). Table 2 also gives relative risks by age at first term pregnancy. The table shows no trend in risk with age at first lerm preg nancy, regardless of the decade of data col lection. This luck of trend was also evident in both data sets after adjustment for num ber of term pregnancies. Similarly, neither data set showed any trend in risk with age a t . last term pregnancy, either before or after adjustment for numiter of term preg nancies. The effects of age at mcnarehc and of menopausal status are examined in table 8. Ovarian cancer risk tends to decrease with increasing age at menarche. 'Phis trend was statistically significant in the 1980s data (p = 0.04) and in the combined data (p = 0.03). Overall, each year of delayed men arche beyond uge 12 years conferred a 9 per cent reduction in risk. By contrast, menopausal stains und age at natural menopause had no effect on risk, regardless of the decade of interview. Women who had undergone surgical or nat ural menopause were not at altered risk in comparison to premenopausal women of the same age. Among postmenopausal women, older age at nat ural menopause was associated with slightly increased risk, but the trend was not statistically significant. T able 2 _O uarum cancer risk, by num ber of term pregnancies and by age at first term pregnancy, San Francisco B ay Area. 1974-1977 a n d 1983-19S5 No. o f term pregnancies^ None O ne or more 1 2 3+ ^ * " Unspecified Cases (%) 1970s participants Controls RKt (%> 9 5% C It Cases n (%) 1980s participants Control? -- (%) RR 95% Cl Cases (%) A ll p a rtic ip a n ts Controls n (Sir RR 95(7 Cl 31 (28) 93 ( 20) 1.00 39 (21) 74 (14) 1.00 80 (72) 379 (80) 0.67 0.40-1.11 149 (79) 465 (86) 0.61 0 .39-0.9 6 18 (16) 71 (15) 0 .80 0 .4 0 -1 .5 8 22 ( 12) 65 (12) 0.64 0 .3 4 -1 .2 0 31 (28) 136 (29 ) 0.67 0 . 3 7 - 1.2 1 51 (27) 163 (30) 0.62 0 .3 7 -1 .0 3 31 (28) 172 (36) 0.5 8 0 .3 1 -1 .0 8 76 (40) 236 (44) 0.59 0 .3 6 -0 .9 8 0 (0) 0 (0 ) 0 ( 0) 1 (0) 70 (23) 167 (1 7 ) 1.00 229 (77 ) 844 (83) 0.64* 40 (13) 136 (13) 0.7] 82 (27) 299 (30) 0.64* 107 (36) 408 (40) 0.60* 0 (0) 1 (0 ) 0 .4 5 -0 .8 S 0 .4 5 -1 .1 2 0 .4 4 -0 .9 5 0 .4 0 -0 .8 8 O verall trend per preg nancy (if parous) 0.88 0.72-1.08 0.96 0.84-1.08 0.93 0.84-1.04 Age (years) at first term pregnancy! N u llip a ro u s <20 20-24 25-29 30+ U n s p e c ifie d 31 (28) 93 (2 0 ) 1.00 9 (8) 66 (14 ) 0.39 0.16-0.92 33 (30) 146 (31) 0.70 0.38-1.28 20 (18) 103 (22) 0.65 0.33-1.2 5 18 (16) 60 (13 ) 0.99 0.49-2.0 0 0 (0) 4 (1) 39 (21) 74 30 (16) 99 68 (36) 189 31 (16) 115 18 (10) 60 2 ( 1) 2 (14) (18) (35 ) (21) (11) (0) 1.00 0.61 0.72 0.52 0.56 0 .3 4 -1 .0 9 0 .4 4 -1 .1 9 0 .3 0 -0 .9 2 0 .2 8 -1 .1 2 70 (23) 3_9_ (13) 101 (34 ) 51 (17) 36 (12) 2 ( 1) 167 (17) 1.00 1. 6. .5 (1 6 ) 0.55 335 (33 ) 0.73 218 (22) 0.57 120 (12) 0.73 6 ( 1) 0 .3 4 -0 .8 8 0 .4 9 -1 .0 7 0 .3 7 -0 .8 8 0 .4 4 -1 .1 9 O verall trend per 5 years (among the parous) *p< 0.001. t R R , relative risk; C l, confidence interval. ! 20 or more weeks gestation 1.20 0.92-1.5 5 0.95 0.78-1.15 1.03 0.88-1.2 0 O O VAR IA N O ANOKIt A N D U K I'ltO D O O T lV K KVKNTS T able 3 O varian cancer risk, by age at m enarche a n d by m enopausal status, S a n Francisco B a y Area, 1974-1977 and 1983-1985 1970s participants Cases Controls ---------------------------------- RR * n (%> n 1%) 95% C l* 1980s participants Cases Controls ------------------- ----------------- n (%) n (%) RR 95% Cl All participants Cases Controls ------------------------------------- RR n (%) n (%) 95% Cl Age (years) at menarche <12 12-14 15+ Unspecifjed 18 (16) 85 (18) 1.00 79 (71) 303 (64) 1.19 0.6 7 -2 .1 2 12 (11) 75 (16) 0.81 0 .3 6 -1 .8 4 2 (2) 9 (2) 48 (26) 111 (21) 1.00 66 (22) 196 (19) 120 (64) 347 (64) 0.79 0.53-1.17 199 (67) 650 (64) 20 (11) 79 (15) 0.56 0.31--1.03 32 (11) 154 (15) 0 (0) 2 (0) 2 (0) 11 (1) 1.00 0.90 0.63 0 .6 5 -1 .2 5 0 .3 9 -1 .0 2 O verall trend per year 0.94 0.82-1.0 7 0.89 0.81--0.99 0.91 0.84-0.9 9 M enopausal status Prem en o p au sa l Surgical m enopauset Age (years) at natural menopause <45 45-49 50-*U n s p e c ifie d 37 (33) 139 (29) 1.00 0 (0) 0 (0) 63 (34) 196 (36) 1.00 100 (33) 335 (33) 1.00 41 (22) 154 (29) 0.93 0 .5 4 -1 .6 0 41 (14) 154 (15) 0.8 9 0.54--1.47 74 (67) 333 (71) 1.16 0 .5 0 -2 .6 9 7 (6) 46 (10) 0.71 0.2 3 -2 .2 2 28 (25) 97 (21) 1.40 0.5 9 -3 .3 3 33 (30) 149 (32) 1.09 0.4 2 -2 .7 9 6 (5) 41 (9) 84 (45) 189 (35) 1.47 0.80-2.69 158 (53) 522 (52) 11 (6) 22 (4) 1.55 0 .54-4.4 4 18 (6) 68 (7) 21 (111 60 (11) 1.03 0 .4 2 -2 .5 4 49 (16) 157 (16) 51 (27) 103 (191 1.67 0 .6 7 -4 .1 8 84 (28) 252 (25) 1 (1) 4 U) 7 (2! 45 (4) 1.36 1.06 1.26 1.41 0 .8 3 --2 .2 2 0.50--2.27 0.67--2.34 0.73--2.70 O verall trend per 5 years (am ong those n a tu rally menopausa' ! 1.05 0 .7 6 -1 .4 6 * R R . relative risk; C l, confidence interval. + W om en w ith surgical menopause were excluded in the 1970s study. 1.22 0.87-1.71 1.13 0.89--1.43 1'22 VVll K T A l. ie;e were no clear or consistent puti of risk with any of several menstrual >rs examined, which included history nenorrhea, irregular menstrual cycles, midcycle pain. ihle 4 shows the effects of oral contraive use on ovarian cancer risk, after stment for number of term pregnanThe unadjusted relative risks and p es were similar to the adjusted ones. >ite the different frequency distribution al contraceptive use in the two decades, ive risks from the two studies were lar, with a combined risk reduction of er cent associated with prior oral eon;ptive use. Risk decreased with increasluration of use. For the combined data, ive risks ranged from 0.97 among .en who had used oral contraceptives >ne year or less to 0.40 among users of e than three years. The overall reducin risk per year of use among ever s was 12 per cent (p < 0.001). lie protection afforded by oral contraive use was greatest among women who used them for more than one year. The ive risk did not vary with time since oral contraceptive use or with time e first use, after adjustment for total ition of use. ible 5 gives relative risks by 3 x 3 = 9 categories of parity and duration of contraceptive use. The data suggest the protective effects of oral contraive use do not wane with increasing ty. The relative risks in table 5 were luately described by a multiplicative ive risk function (goodness of fit xi = p = 0.21). According to this function, ty status has no effect on risk ratios in ligher two durations of oral contracepuse relative to the lowest duration, ises and controls were similar with ret to noncontraceptive estrogen use durboth the 1970s and the 1980s. For the hined data, the relative risk associated . a history of any noncontracept/e esen use was 0.89 (p = 0.46). fyfrther- 6<P ^Ottl T uCO) o i ^t i^ io o o o oCTo> cm --r -o -r l/ i CD T f CM 8 r oi-> o-o oV dooo CD o> CD cCUmO) crvdi -- owrr>S ct--on<o.^_o1 (37) (39) (17) (13) (8) (7) (9) (17) (3) (2) 188 (63) HH OH C^4 CCMO OO t--tI Or- cOo aO cd oTil CCDO CCMO dCM> dodo 8 I- OCO (COD Cf*O d ^od J 2 S^ o> CM Odi (oi' Oo -n CM 102 (54) 146) 86 (46) 42 (22) 16 (9) 20 (11) 8 (4) (23) 102 (22) (8) 39 (8) (7) 19 (4) (7) 38 (8) (0) 6 (1) rC-O (cm' cTmf cTor tC-O CO cTif CM dodo 8 t ' CtM' C<ND COO OO HO ao tClOMO) CO oo o 86 370 2 E O D5 C--M ^ ^ <*3D Z U3 Overall trend per 12 months of use (among users) 0.96 0.86-1.08 0.86* 0.79-0.93 0.88* 0.83-0.94 p < 0.001. R R . r * l a l i w rin k , a d j u s t e d fo r n u m lie r o f l e r m p r r p i a n c m <20 o r m o r* w n lu> lo m m u o n I in r a i r c c m * 0 . 1 - 2 . a n d 3 + ; C l . coniM lenc* i n te r v a l . O V A R IA N C A N O iH A N D KK R K O D IU IIV K KVKNT.S 5 more, there was iiu d e a r trend with dura tion of use. 5 cm io O CO <o? n7 i<o? o d dCO T -< 8 d3 To investigate Ilie hypothesis lli.it the protective effects of late age at menardie, parity, and oral contraceptive use are due to their common suppression of ovulation, we examined the relation between risk and estimated durat ion of ovulation, defined as the time between menarche and last men strual period, minus nine months for each term pregnancy and three months for other pregnancies, and minus the totalduration 3 2 of oral contraceptive use. Women who had undergone hysterectomy without bilateral oophorectomy and were under age -If) years o .A were assumed to be ovulating; those aged 45-55 years (rt = 75) were excluded from the analysis, and those older than 55 years d d 1C> CO i were assigned the median menopausal age of non hysterectomized women, specific for A case and control status. We excluded 5 breastfeeding from the calculation of ovu a latory years because its effectiveness in 3 J 3 h ^ co suppressing ovulat ion is variable, depend ing on the mother's nutritional status, the extent of supplemental feeding to the in & fant, and the intensity of suckling (25-27). 1 3 .A Its effectiveness for women in developed countries is weaker than that of pregnancy or oral contraceptive use; some 40-75 per cent of lactating women menstruate while nursing (25, 27). sa 8 I " I -- CO CO a Table 6 relates ovarian cancer risk to sfac estimated duration of ovulation. Data from both st udies show trends of increasing risk c with increasing duration of ovulation, with a stronger trend in the l!M0s study than in 8. -*r 2 t c 8 "3 3qj 8: MU ** 4> c B o 2a z<o; e <=> C 5g. 2oa 3* vQ.JNgK the earlier one. Compared with a relative risk of unity for those wlm had been ovu lating for less than 25 years, women in the combined studies who had ovulated for an estimated 25- 29 years had a risk of 1.58, those who had been ovulating for 50 51 years had a risk of 2.48, and those who had been ovulating longer had a risk of 5.20. According to the logistic function lit to the ; 5 combined data, each additional five years of ovulation conferred a 50 per cent in crease in risk ([> < 0.001). These findings 224 W U E T A l. O Q -<f ' o> i> < S3 I CM CO could be confounded by age despite the matching within five-year age groups. How ever, the results were unaffected when age was added to the regressions as a continu ous variable. x X 8$ D is c u s s io n The above results should be interpreted cautiously for several reasons including the ' studies' failure to interview all eligible ovar S5 CM CM ian cancer patients and a completely ran C~~ ' -H -H CM I dom sample of eligible controls, potential pitfalls in combining the two studies and the two control groups in the second study, 2s o tp ai co d> co &> and the possibility of confounding by unmeasured variables. Nevertheless, several conclusions seem appropriate. In particu lar, the present data do not support the 1 hypothesis that an association of ovarian cancer with infrequent oral contraceptive use is a consequence of their common cor relation with subfertility. This hypothesis predicts that oral contraceptive use confers no benefit on women of high parity and 3* 9 2 o CO OQ tO Q -- CO CM < 5 CM that long episodes of use confer little ben efit above th at of short episodes. These predictions are inconsistent with the pres C<Oo UO5 CM h o CM V UO ent data which show oral contraceptive use to reduce risk independently of number of So odd previoua term pregnancies, with increased 8h b- <{ 0 0 UO C o h^ reduction attached to increasing duration of use. s Oo S hh Most studies support the present find - CM CM CO -- *3 ings that nulliparous women are at elevated t S risk of ovarian cancer (1-4, 6-9, 11-13, 28- c 31). Similarly, the lack of association noted with age at first term pregnancy agrees with other data (2, 3, 13, 28, 32-34). Exceptions I are provided by data from two studies (9, j o--r ;os 3ft) showing that women whose first preg- ( -id 5 nancy occurred after the age of 25 years had two to three times the risk of those .2 .O whose first pregnancy occurred earlier. 9 ^ A S3 81 | VQ. Oi 0$0 At least 12 studies support the current finding that oral contraceptive use protects against ovarian cancer (1, 2, 4, 5, 7, 8, 11, 13-16, 36). The strong trend noted here of 5# increasing protection with increasing years of use also was found in a recent study & O VAK IA N CANCEK AN D K E P K O P U tT I VK EVENTS 1225 involving 492 cases and 4,228 controls (4) which found similar protective effects for each of the four major histologic subtypes and for each of several specific oral contra ceptive formulations. Other studies (4, 14) also support the constancy of relative risk with respect to time since last oral contraceptive use among women who have used oral contra ceptives for a fixed duration. Such a con stant relative risk model predicts th at for mer users of oral contraceptives experience the same per cent reduction in ovarian can cer incidence rates throughout their life time. However, the long-term effects of oral contraceptive use cannot yet be tested in women over 60 years of age, because too few have ever used oral contraceptives. Therefore, one cannot yet exclude the pos sibility that the protective effect of oral contraceptive use wanes long after use ceases. The present association between early age at menarche and increased risk is not supported by other studies, which have found cases and controls to be similar in their ages at menarche (8, 11, 13). Cases and controls also appear similar in their ages at menopause, according to data here and elsewhere (1, 11, 13). The failure of case-control studies to find a strong rela tion between ovarian cancer and age at menopause contrasts with the behavior of national incidence rates for the disease, which increase more slowly with age after the fifth decade (37). This drop in the rate of increase at the time of menopause sug gests a protective effect for termination of ovulation. The effect may be missed in case-control studies because they measure termination of menstrual bleeding rather than ovulation. Anovulatory cycles are so common at the end of reproductive life that perimenopausal menstrual bleeding is a poor indicator of ovulation. Late menarche, pregnancy, and oral con traceptive use decrease a woman's total number of ovulations. Therefore, the pro tective effects of these characteristics pro vide indirect support for the hypothesis of Fathalla (17) that ovulation increases risk for ovarian cancer, possibly by subjecting the ovarian epithelium to increased cellular proliferation, steroid-rich follicular fluid, or elevated levels of pituitary gonadotropins. Results from four other case-control stud ies that have examined this issue (2, 7, 8, 30) agree with the present finding that a woman's estimated duration of ovulation is a strong indicator of increased ovarian can cer risk. One study (30) found substantially larger magnitudes of diminished risk from each of the separate characteristics that reduce ovulatory years than would lie ex pected solely on the basis of their inhibition of ovulation. For example, the reduction in risk per year of oral contraceptive use was roughly four times the reduction per year of early menopause. However, these find ings do not provide strong evidence against the hypothesis that pregnancy and oral contraceptive use protect by suppressing ovulation. As noted above, a decrease of one year in age at menopause need not represent a decrease of one ovulatory year, since ovulation is sporadic in the perimeno pausal years. Furthermore, in estimating the magnitudes of diminished risk expected according to the ovulatory suppression hy pothesis, the authors assumed that the log arithm of ovarian cancer incidence rates increases in proportion to total duration of ovulation. However, it can be shown thul if cancer risk depends only on duration of ovulation and the regression model relating risk to duration of ovulation is misspecified, the estimated regression coefficients cor responding to ovulatory years lost from the separate characteristics need not have equal magnitudes. Stadel (38) and Crumer and Welch (39) have suggested that high circulating levels of pituitary gonadotropins may play a role in ovarian cancer. If so, late menarche, pregnancy, and oral contraceptive use may protect against the disease by reducing total exposure to gonadotropins (40, 41). How ever, in agreement with other investigators >6 W U E T A L . 7,3, 39, 42, 43), we found no statistically nificant associations between ovarian icer risk and use of noncontraceptive rogens, a practice which also reduces tiadotropin exposure, although not to ^menopausal levels (40). T his lack of asiation suggests th at the protective ef ts of oral contraceptive use may he m e lted by m echanism s o th er than its role reducing serum levels of gonadotropins, rther work is needed to resolve th is issue, se num bers in the presen t study were too all to investigate the finding of W eiss et (43) th a t noncontraceptive estrogen use Teases risk for ovarian tum ors of the dometrioid type. In sum m ary, th e p re s e n t o bserv atio n s )port the concept th at oral contraceptive *decreases risk for ovarian cancer, ra th e r m merely indicates fertility, which may If be associated w ith decreased risk, rthermore, they suggest th at the greater 'Oman's d u ratio n of ovulation (or o f connitant endocrinologic phenom ena), the ater her risk of developing ovarian can- R efekences Annegers JP, S trom H , Decker D O , et al. O varian cancer: incidence and case-control study. Cancer 1979;43:723-9. Casagrande IT, Louie E W , Pike M C , el al. "I n cessant ovulation" and ovarian cancer. Lancet 1979;2:170-3. C ram er D W , H utchison O B , W elch O R , et al. D e term in an t* of ovarian cancer risk. I. Reproduc tive experiences and fam ily history. J N C I 1983;71:711-16. T h e Cancer and Steroid Horm one Study o f the Centers for Disease C o n tro l and the N a tio n a l Institute of C hild H ealth and Hum an Develop ment. T h e reduction in risk o f ovarian cancer associated w ith oral con trace p tive use. N Eng! J M ed 1987;316:650-5. C ram er D W , H utchison O H , W elch W U , et al. Factors affecting the association o f oral con tra ceptives and ovarian cancer. N Engl J M ed 1982;307:1047-51. Demopoulos R I, Seltzer V , D u hin N , el al. T h e association o f p urity and m arital status w ith tlie development o f ovarian carcinoma: clinical im p li cations. Ohsiet Oynecol 1979,54:150-5. Pnmresehi S, l,n Vecchin (', H elm ric h S # \ et al. Risk factors lo r epithelial ovarian c n n c e /iji Italy. Am J Epidemiol 1982;115:714-19. P. \ H ildreth N O , Kelsey J L , L iV o lsi V A , H * a l. An epidemiologic study o f ep ith elial carcinom u o f the ovary. A m 1 E p id e m io l 1 9 8 l; l 14:398-405. 9. J o ly D J , L ilie n fe ld A M , D ia m o n d E L , et al. An epidem iologic study o f the relationship of repro ductive exp erience to can cer o f the ovary. Am -1 Epidem iol 1974;99:190-209. 10. La V e cch ia 0 , Franceschi S, O allu s O , et al In cessant o v u la tio n a n d o va ria n cancer: a critical approach. In t J Epidem iol 1983;12:161-4. 11. M c G o w a n L , P a re n t L , l-e d n a r W , et al. The w om an a l riak for d evelop ing ova rian cancer. Gy necol O n co l 1979;7:325 -44. 12. N asca P C , (ire e n w a ld P, O h o ro st S, et al. An epidem iolo gic case-cont rol study o f ovarian cancer and rep ro d u ctive factors. A m J Epidemiol 1984,119:705-13. 13. N ew houae M L , P earson R M , F u lle rto n J M , et al. A case co n tro l study o f carcinom a of the ovary. H r.I Prev Soc M ed 1977;31:148-53. 14. Rosenberg L , S h a p iro S, Slone D . el al. Epithelial ovarian cancer and com bination oral contracep tives. J A M A 1982;247:3210-12. 15. W eiss N S , L yo n J L , L i f f J M , et al. Incidence of o va rian can cer in relatio n to th e use o f oral con traceptives. In t J C ancer 1981;28:669-71. 16. W ille t t W C , B a in C , H e n n eke n s C H , e t al. Oral coni racep tives an d riak o f ov a ria n cancer. Cancer 1981;48:1684-7. 17. F a th a lla M F . Incessant o v u la tio n -- a factor in o va rian neoplasia? (L e tte r ). Lan cet 1971;2:163. 18. F a th a lla M F . Fac to rs in th e causation and inci dence o f ovarian cancer. O hstet Gynecol Sun* 1972;27:751-68. 19. H o rm e J W , A s ire A J, Y o ung J L , et al., eds. S E E K program . C ancer incidence and m ortality in the U n ile d States, 1973-81. Bethesda, M l): National C a n cer In s titu te , 1985. ( N 1 H pub lication no. 851837). 20. Fasal E , P a ffe n b a rg e r R S J r . O ra l coni raceptivea as relate d to cancer and benign lesions o f the breast. J N C I 1975;55:767-73. 21. Paffenbarger R S Jr, K am pert JB , Chang H-G. Characteristics th a t predict risk o f breast cancer before and a fte r the menopause. A m J Epidemiol 1980;112:258-68. 22. B reslow N E , D a y N E . S ta tis tic a l m ethods in can cer research. V o l 1. T h e analysis o f case-control studies. ( IA R C Hcientific pub lication no. 32). Lyon: IA R C , 1980. 23. M a u rita e n R. E G R E T softw are program . Seattle, W A : Statistics and Epidem iology Research Cor poration, 1986. 24. T h o m a s D C . G e n e ra ! re la tiv e risk models for m atched case-contro l an d fa ilu re tim e analysis. Biom etrics 1981;37:673-8. 25. Perez A , V e la P, M a s n ic k G S , et al. First ovulation a fte r c h ild b irth : the effe ct o f hreast-feeding. Am A O hstet Gynecol 1972;114:1041-7. 26. G ray R, Kslam i S, Cam pbell O. Return of fertility during lactation m onitored by uninary steroid as says. (A b s tra c t). A m J E p id e m io l 1987;!26:761. 27. S p e ro if L , (Hass H H , Knse W G , eds. C linical gynecologic endocrino log y und in fe rtility . 3rd ed. B altim ore: W illia m s & W ilk in s , 1984:249 -50. 28. M ille r A B , B arclay T H C , C hoi N W , et al. A study o f cancer, p a rity and age at first pregnancy. J Chronic Dis 1980;33:595-605. OVAKIAN CANCER AND KKPKODl ICT1VK EVENTS li-i 29. B e ra l V , F rase r P , C h ilv e rs C. Does pregn ancy protect against ovarian cancer? la n c e t 1978;1: 1083-6. 30. Risch H A , W eiss N S , I.y o n J L , et al. E v e n ts o f reproductive life and the incidence o f epithelial ovarian cancer. A m J Epidem iol 1983;117:128-39. 31. B yers T , M a r s h a ll ), G ra h a m S, e t al. A casecontrol study o f dietary and nondietary factors in ovarian cancer. J N C I 1983;71:681-6. 32. W y n d e r E l ., D o do H , B a rb e r H R K . E p id em io lo g y o f cancer o f the ovary. Cancer 1969;23:352-70. 33. V o ig t L F , H a rlo w B L , W eiss N S . T h e in flu en ce o f age at firs t c h ild b irth and p a rity on ovarian cancer risk. A m J E p idem iol 1986;124:490-1. 34. lu s h e r I . , M c G o w a n L , H a rtg e P , et al. Age at firs t b irth and risk o f ep ith elial o varian cancer. J N C I 1985;74:1361-2. 35. I a V e cch ia C , D e c a rli A, Franceschi S, et al. Age at first b irth and the risk o f epithelial ovarian cancer. J N C I 1984;73:663-6. 36. Tseonou A , D a y N E , T ric h o p o u lo s I ) , et al. T h e epidem iology o f ovarian cancer in Greece: a case- co n tro l study. E u r .1 C u iu e r C lin O ncol 1984; 20:1045 52. 37. M o h le J. W h itle m o rc AS. P ike M , et al. t ionndo tro p h in s and ovarian cancer rink. (L e tte r). J N C I 1985,75:178 9. 38. Stadel B V . T h e etiology and prevention o f ovarian cancer. A m J O h s le l G ynecol 1975:123:772 4. 39. C ram er D W , W elch W R . D eterm inant of ovarian can cer risk. 11. Inferences reg ard in g pathogenesis. J N C I 198.^71:717 21. 40. S c h ifl 1. T ile effects o f co n jiig a led estrogens on g o n adotrophins. F e r lil S te ril I9 8 0 ;3 3 :3 3 3 4. 41. Parlow A F, Daane T A . Dignam W J. O n the con cen tratio n o f radioiniiiiunoassayable F S II circu lating in blood throughout hum an pregnancy. J C lin E n d o c rin o l M e ta !) 1 9 7 0 ;3 I:2 1 3 14. 42. Hoover It, G ray L A Sr, Frniim em J F Jr. S|itim es trol (d ie th y ls!ilh es lm l) and the risk o f o varian cancer, laincet 1977; 1:533 4. 43. W eias N S , Lyon J L . K ris h im in n rth y S , et id. N o n contraceptive estrogen use and the occurrence of o v a ria n cancer. J N C I 1982,8:93 8. A ( J j r-t 'S B r. J. L a n c e r 11989). 60. 592-598 The Macmillan Press Ltd.. I9H9 ,' Risk factors for ovarian cancer: a case-control study M. Booth1, V. Beral3& P. Smith3 `Epidemiological Monitoring Unit and `Tropical Epidemiology Unit. Department of Epidemiology & Population Sciences. London School of Hygiene and Tropical Medicine. Keppel Street tCower Street). London WCIE 7HT, UK: and1Imperial Cancer Research Fund. Epidemiology <5 Clinical Trials Unit, Radcliffe Infirmary. Oxford 0X2 6HE. UK. Summary A hospital-based case-control study of ovarian cancer was conducted in London and Oxford between October 1978 and February 1983. Menstrual characteristics, reproductive and contraceptive hisiorv and history of exposure to various environmental factors were compared between 235 women with his tologically diagnosed epithelial ovarian cancer and 451 controls. High gravidity, hysterectomy, female sterilisa tion and oral contraceptive use were associated with a reduced risk of ovarian cancer. Infertility and late aee at menopause were associated with an increase in risk. While these factors were related, they were each found to be independently associated with ovarian cancer risk after adjusting for the effect of the other factors. While results from recent case-control studies have con sistently shown that multiparity and oral contraceptive use are associated with a reduced risk of ovarian cancer, the association of the cancer with other reproductive, hormonal and related factors such as age at menopause, history of hysterectomy or use of oestrogen replacement therapy is less clear. We have conducted a hospital-based case-control study in London and Oxford which was designed to inves tigate the independent contributions of reproductive history and contraceptive use to ovarian cancer risk. In particular, it was planned to attempt to segregate out the effect on risk of infertility from that of voluntary limitation of family size. The association between ovarian cancer and other possible aetiological agents was also examined. Subjects and methods Between October 1978 and February 1983 five interviewers identified and questioned women with a diagnosis of ovarian cancer and women selected as controls at 13 hospitals in London and two in Oxford. A standard questionnaire was used to obtain information on reproductive and menstrual history and on exposure to various substances such as exogenous oestrogens. cigarettes and talc. A month by month record was made of the specific contraceptive methods used by each woman between the ages of 16 and 45 years, or. if under 45 years, up to the time of diagnosis (cases) or inter view (controls). The methods were classified as sheaths, diaphragms, intrauterine devices, oral contraceptives or 'other methods' (spermicides, rhythm and coitus interruptus). Women who reported using a contraceptive diaphragm were asked if they had stored it in talc. Also recorded were months during which a woman was not using contraception due to sexual abstinence, pregnancy, menopause or because she or her partner had been sterilised. The other months when a woman reported using no method of contraception although sexually active have been classified as months of 'unprotected intercourse'. The total duration of use of each contraceptive method, of any contraceptive method, of unp rotected intercourse and of pregnancy were computed for each woman. The study was confined to women aged less than 65 years whose diagnosis of ovarian cancer had been made within two years of interview. A total of 280 cases were interviewed and pathological specimens were histologically classified by Pro fessor C. Hudson and Dr M. Curling from St Bartholomews Hospital A total of 235 women with epithelial ovarian cancer were included in the analyses. For these women, the tumour type was described as serous in 101 {430o) cases, mucinous in 38 (15%) cases, endometrioid in 52 (22%) cases C orrespondence: M Sooth. R eceived 3 F e b ru a ry 1989: jn d in revised form 9 M ay 1989 and clear cell in 12 (5%) cases. Mixed and undifferentiated types of epithelial tumours accounted for the remaining 32 (14%) cases. Excluded from the analyses were nine women with a non-epithelial ovarian neoplasm. 11 with a primarytumour in an unknown site outside the ovary. 21 with a primary tumour in an unknown site although one consistent with an ovarian origin, one with a benign tumour and three for whom pathology material could not be obtained. For each case it was planned to select two age-matched controls from women being treated in the same hospital. Women with bilateral oophorectomy were excluded from the control group as were women admitted with conditions that have been related to reproductive history or oral contracep tive use (all circulatory and gynaecological diseases, gallblad der and thyroid diseases, rheumatoid arthritis, malignant disease of the breast, uterus and bladder, and melanoma). It proved logistically impossible to select two age-matched controls for each case from the same hospital and it was decided merely to ensure that the age distribution of the controls was approximately the same.as that of the cases. For 63 cases recruited from a London hospital where only cancer patients are treated, controls were selected from other London hospitals. For these reasons, the data were analysed using an unmatched approach with adjustments being made to relative risk estimates for age and socio-economic status. A total of 451 controls have been included in the analyses. The admission diagnoses fpr these patients were gastrointes tinal disease (105). bone or joint disease (70). respiratory disease (39). renal or other urinary disease 135). neurological disease (30). fractures or other injuries (28). skin or sub cutaneous tissue disease (17). malignant neoplasms of the digestive organs (15) and bone or skin (2). benign neoplasms of the digestive organs (4) respiratory system (4) and other sites (8) and various other conditions and symptoms (94). This final category included patients with haemorrhoids (15) and those with symptoms relating to the respiratory system (10). gastrointestinal tract (20) and urinary system (10). Maximum likelihood estimates of relative risk (RR) together with their 95% confidence interval (95% Cl) and tests for trend where appropriate were computed by multiple logistic regression techniques (Breslow & Day. 1980) using the GLIM statistical package (Baker & Nelder, 19'8). All relative risks have been adjusted for age in 5-vear strata (20-24. 25-29. . . . 60-64) and for social class in six categories (I.II.Ill non-manual. Ill manual. IV and V). Age of the cases was taken as age at diagnosis of ovarian cancer and of the controls as age at interview. Social class was based on occupation (Office of Population Censuses and Surveys. I9"0i using husband's occupation for ever married women and own occupation for those who had never married Other relative risk adjustments and tests for trend have been made with the exposures as continuous variables. When the data were examined by place of interview i London or Oxford), there were no notable differences in the risk estimates . ' associated with the major variables of interest. The relative 1 risks have not. therefore, been stratified by place of interview - The terms nulligravid and gravid have been used to denote, respectively, women who have never knowingly conceived and women who have had at least one pregnancy. Parity has been defined as number of live and still births. Table II Relative risks for ovarian cancer associated with pregnanes htstorv f a r able G raviditv*' G ravid N ulligravid C ases C ontrols R R 95% C /. ;"6 376 59 74 1.0* 1.7 ( 1 .1 - 2 .6 ) Results The age distributions of the cases and controls are shown in Table 1. The average age of the cases was slightly higher than that of the controls. There was an excess of cases in social classes 1. II. and III non-manual (58%) as compared to controls (43%) (P = 0.05) and. because of this, all relative risks have been adjusted for social class as well as age. Table II shows the relative risks for ovarian cancer associated with various aspects of pregnancy history. Nulligravid women had a higher risk of ovarian cancer than gravid women (RR = 1.7. 95% Cl 1.1-2.6). The relative risks were elevated both in nulligravid women who had been sexually active and in those who had not, although significantly so only for the sexually active. Among those who had ever been pregnant, the relative risks decreased as the number of pregnancies increased. (y: (trend) = 4.3. P <0.05). Similarly, among parous women, the higher the parity the lower the relative risks (x: (trend) = 3.9. P < 0.05). After adjusting for parity, the relative risks associated with successive numbers of incomplete pregnancies (spontaneous and induced abortions) also decreased although the trend was not statistically significant (X2 (trend) = 0.5). Women having their first preg nancy after the age of 35 years had a significantly higher risk of ovarian cancer than women with a first pregnancy before the age of 20 years. Their risk was also higher than that for nulligravid women. There was. however, no marked nor significant trend of increasing risk the later the age at first pregnancy (X: (trend) = 1.0). Analyses by age at first livebirth gave similar findings. After adjustment for number of livebir ths. women who had breastfed for more than two years in total had over three limes the risk of ovarian cancer com pared to women who had never breastfed (/^ < 0.05) but overall, there was no significant trend the longer the duration of lactation. Analyses of infertility and subfertility as risk factors for ovarian cancer were restricted to the 213 (91%) cases and 240 (93%) controls who reported that they had ever been sexually active. Among these women, 30 (14%) with ovarian cancer and 34 (8%) controls reported, when so questioned, that they had had problems in becoming pregnant and. of these. 16 cases and 12 controls had never conceived. Analysis of the data on contraceptive use suggested that there were other women who might have been infertile or subfertile. Although sexually active, they had used contraception infrequently or not at all and had had few or no pregnancies. For all women who had ever been sexually active, the risk of ovarian cancer increased with increasing duration of unprotected intercourse after adjustment for gravidity (X; (trend) = 10.2. P <0.01). The effect was most marked among nulligravid women who reported more than 10 years of unprotected intercourse. Their risk was over six times that of nulligravid women who reported less than three months of unprotected intercourse (Table III). Among gravid women, those reporting over 10 years of unprotected intercourse had a higher risk than other gravid women. There was no significant trend in risk associated with the duration of use of any Table I Age distribution and average age o f eases and controls Age !years i 20-34 35 - 44 4 5 -5 4 55-64 T o tal Average age (years) Cuses 1 1 13 (5 .5 )' 27 (11.5) 87 (37.0) 108 (46.0) 235 52.4 C ontrols ' " 33 (7.31 75 116.6) 156 (34.6) 187 141.5) 451 51 4 N ulligravid and ever sexuallv active N ulligravid and never sexually active 44 *>*> 30 1.9 ( 1. 1- 3 . 1) 1.5 (0 .8 - 2 .6 ) N um ber of pregnancies 0 1 3 4 ^5 59 4 1.0* 43 71 0.8 ( 0 .4 - 1.3) 63 107 0.7 ( 0 .4 - 1.1) 37 98 0.5 (0 .3 -0 .8 ) 13 41 0.4 ( 0 .2 - 0 .8 ) 20 59 0.4 (0 .2 -0 .8 ) / fo r tren d = 4.3 P < 0 .0 if (gravid wom en only) E stim ated red u ctio n in relative risk associated w ith each pregnancy 0.86 (0 .7 8 -0 .9 4 ) P a ritv b 0 1 2 3 4 >5 66 87 1.0" 48 84 0.7 (0 .4 -1 .2 ) 61 127 0.6 ( 0 .4 - 1.0) 40 88 0.6 ( 0 .3 - 1.0) 12 30 0.5 (0 .2 - 1 .0 ) 8 34 0.3 (0 .1 - 0 .7 ) / - fo r trend = 3.9 P < 0 .0 5 (parous wom en only) E stim ated reduction in relative risk associated w ith each birth 0.84 (0 .7 5 -0 .9 4 ) No. o f incom plete preg n a n c ie s 1" 0 1 *> >3 185 330 39 83 7 IS 4 19 /' fo r tre n d 1.0* 0.9 0.7 0.6 == 0.5 (0 .6 - 1.4) (0 .3 -1 .8 ) (0 .2 - 1.7) E stim ated red u ctio n in relative 0.92 risk associated w ith each incom plete pregnancy (0 .7 5 -1 .1 3 ) Age at first pregnancy (vears)d 15-19 20-24 25-29 30-34 2s 35 N ulligravid 26 65 1.0* 73 182 0.9 1 0.5- 1.5) 49 96 1.2 ( 0 .7 - 2 .2 ) 17 29 1.2 (0 .8 - 2 .7 ) 9 4 4.1 ( 1 .1 -1 5 .1 ) 59 74 2.0 (1 .1 -3 .7 ) fo r tre n d = 1.0 (g rav id w o m e n only ) M onths of lactation' None Si 6 7- 12 13- 18 19-24 iS 25 44 107 1.0* 66 124 1.3 29 SO 0.9 13 29 1.2 5 7 2.1 i : 15 3.4 Z-' fo r tre n d := 1.8 (0 .8 -2 .2 ) (0 .5 - 1.6) (0 .5 -2 .5 ) (0.~ -6.' ) <1 . 1 - IO.81 A ll relative risks a d ju s te d lo r age a n d social class. 'Reference category. "D ata missing for I control. `Relative risks adjusted for parity. J D a la m issing for 2 cases an d I co n tro l. ''W om en w ith livebirths only. R elative risks ad ju sted for n u m b e r o f live births. contraception (X: (trend) = 1.2) although sexually active nulligruvtd women who had never used any method of contraception had about twice the risk of ovarian cancer compared to all other sexually active women (Table IV ). Of the specific methods of contraception studied, ever having used oral contraception and having been sterilised were associated with a statistically significantly reduced risk of ovarian cancer, while no method was associated with a significantly elevated risk (Table V). As only three cases had been sterilised it was not possible to assess whether age at sterilisation influenced the risk of ovarian cancer Table VI shows detailed analvses of the relative risks associated with oral 5 9 4 M . B O O T H et al. ' T a b le III R e lative risks fo r o v a ria n can cer associated w ith d u ra tio n o f unprotected intercourse by g ravid ity Cases Controls RR Mulligravid women D u ratio n o f un- p ro te c te d in te rc o u rs e (m o n th s )1' <3 4 -6 0 12 6 1 -1 2 0 > 120 4 1 20 * / fo r tre n d = 26 1.0" 6 1.5 4 0.7 8 6.5 11.2 P < 0 .0 0 1 Gravid women D u ratio n o f un- p ro te c te d in te rc o u rs e (m o n th s )1' <3 4 -6 0 78' 176 6 1 -1 2 0 > 120 51 113 10 27 37 60 =X: fo r tre n d 2.6 1.1 11 l.l 1.6 (95% CU (0 .4 -6 .5 ) (0 .1 -7 .8 ) (2 .1 -2 0 .4 ) (0 .5 -2 .4 ) (0 .5 -2 .6 ) (0 .4 -3 .2 ) (0 .7 -4 .0 ) S exually active w o m en o n ly. R e la tiv e risks ad ju sted fo r age and social class. 'R e fe re n c e catego ry. T i m e w hen sexually active and a t risk o f pregnancy but using no con tracep tio n . contraceptive use. The risks decreased as duration of use increased although, among those who had ever used such contraceptives, the trend was not significant (X: (trend) = 1.2) Whatever their age at first use, women who used oral contracep tives had a lower risk of ovarian cancer than those who had never used them, the risk being lowest in those who had first used oral contraceptives under the age of 25 years. The risk of developing ovarian cancer did not increase as time since discontinuing use increased. Women who had stopped using oral contraceptives more than ten years previously had a statistically significant reduced risk of 0.3 compared to women who had never used them. Women both under the age and over the age of 40 years had a reduced risk of ovarian cancer associated with oral contraceptive use, but the reduction was greater in the younger women. Gravid and nulligravid women who had used oral contraceptives had a reduced risk of ovarian cancer. Table VII shows the relative risks associated with age at menarche and age at natural menopause. There was no trend in risk with age at menarche (X: (trend) ==0.03), In contrast, risk increased the later the age at natural menopause (X; (trend) = 7.1, P <0.01). Women having their menopause at the age of 50 years or later had nearly three times the risk of women who were menopausal before the age of 45 years. The risks and trend associated with age at menopause were similar irrespective of whether they were adjusted for age in five year or one year strata. Table IV R e la tiv e risks fo r o v a ria n c a n c e r ass o c ia te d w ith d u r a tio n o f use o f c o n tra c e p tio n b y g ra v id ity Cases Controls Xulligravid women D u ra tio n o f use o f con- tra c c p tio n N e v e r used 15 < 10 vears 14 1 0 -2 0 years 6 > 20 vears 2 r for trend = 0.9 10 21 9 4 Gravid women D u ra tio n o f use o f con tra c e p tio n N e v e r used 32 < 1 0 years 25 1 0 - 2 0 years 55 > 20 years 64 X: fo r tre n d = 0 . 3 47 56 147 126 RR (95% Cl) 1.0* 0.5 0.5 0.4 (0 .1 -1 .7 ) (0 .1 -2 .5 ) (0 .1 -3 .2 ) 0.4 (0 .2 -1 .1 ) 0.3 (0 .1 -1 .0 ) 0.4 (0 .1 -1 .2 ) 0.5 (0 .2 -1 .5 ) S e xu ally active w o m e n o n ly . R e la tiv e risks ad ju sted fo r age. social class and d u ra tio n o f u np ro tected intercourse. 'R e fe re n c e category. Table V R e la tiv e risks fo r o v a ria n can cer associated w ith the use o f different m ethods o f contraception Method of contraception Cases Controls RR (95% Cl) Sheath N e v e r used E ver used 108 205 105 215 1.0' 1.1 ( 0 . 8 - 1.7) D iaphragm N e v e r used Ever used 178 329 35 91 1.0' 0.7 (0 .4 -1 .1 ) Intrau terin e device N e v e r used E v e r used O ral N e v e r used co n tra c e p tio n Ever used P artner w ith vasectom y No Yes 201 12 178 35 203 10 585 506 1 14 404 16 1.0' 0.8 10' 0.5 1.0' 2.1 <0.4-1 7) (0 9 - 4 9) o o Fem ale s te rilis a tio n No Yes 210 3" 5 3 45 1.0' 0.2 (0.1 0.6 ) O ther m ethods' Never Used Ever used 156 292 57 128 10' 1.1 ( 0 . 7 - 1.7) S exually active w om en o n ly. R e lative risks adjusted fo r age. social class, gravidity and d u ra tio n o f u n p ro tec ted intercourse. 'R e fe re n c e category "Use o f sperm icides, rh yth m o r coitus interrup tus. Table V I R e lative risks fo r o v a ria n cancer associated w ith o ral c o n tr a c e p tiv e ( O C ) use D u ra tio n o f O C use (years) N e v e r used <5 5 -1 0 > 10 C ases C on trols RR (95% . C l) 178 306 1.0' 24 70 0.6 (0 .3 -1 .0 ) 10 29 0 .6 ( 0 .2 1.4) 1 15 0.1 (0 .0 1 1.0) fo r tre n d w ith in users = 1.2 A g e a t first O C use (years) N e v e r used <25 2 5 -2 9 3 0 -3 4 $5 35 178 306 1.0' 6 39 0.1 (0 .0 4 6 17 0 .6 ( 0 .2 11 27 0 .7 ( 0 .3 12 31 0.7 10.4 = PC fo r tre n d w ith in users 5.9. < 0 . 0 5 0.5) 2.0) 1.6) 1.5) T im e since d is c o n tin u in g O C use (years) N e v e r used 178 306 1.0' C u rr e n t users 6 19 0 .5 < 5 12 24 0 .8 5 -1 0 > 10 9 25 0.8 8 46 0.3 /} =fo r tre n d w ith in users 2.6 (0 .2 1.5) ( 0 4 1.9) (0 .3 - 1.9) (0.1 0 .7 ) A ge (years) <40 O C use N ever used Ever used 11 1 1 1.0' 9 35 0.2 (0.1 0 .9 ) ^40 O C use N e ver used Ever used 167 295 1.0' 26 79 0.7 (0 .4 1.2) G ra vid ity G ra vid wom en* O C use N e v e r used Ever used 149 280 1.0' 27 96 0.5 (0.3 0.9) N u llig ravid women O C use N ever used Ever used 29 26 1 0 ' s 18 0.5 (0 .0 5 2 .8 ) Sexually active w om en o n ly. R e lative risks adjusted fo r age. social class, gravidity an il d u ra tio n o f un p ro tected intercourse 'R e fe re n c e categ o ry . "R ela tive risks ad ju sted fo r g ra v id ity . Ta Ay lye Arr.; \ c t \ f r RISK FACTORS FOR O V A R IA N CAN C ER 595 Table VII R elative risks for o v a ria n can cer asso ciated w ith age at m enarche and age at natural m enopause C ases C ontrols R R i9 5 % a Age at m enarche (vears)" > 14 12-13 < 12 97 89 46 X: fo r tre n d = 0.03 197 185 66 1.0* 0.9 (0 .6 - 1.3i 1.3 ( 0 . 8 - : . ! ) Age at natural m enopause (vearsk <45 4 5 -4 9 >50 10 34 47 77 84 99 X : fo r tre n d = 7.1 P < 0.01 1.0* 2.0 (0 .9 -4 .-) 2.5 (1 .1 -5 .8 ) R elative risks adjusted fo r age and social class. 'R eferen ce category. bD a ta o n a g e a t m e n a rc h e m issin g fo r 3 cases a n d 3 c o n tro ls . 'D a t a o n age at m enopause missing for 2 cases and 1 control. Women who reported hysterectomy, with or without unilateral oophorectomy, had a much reduced risk (Table VIII). Since there were only 10 women with ovarian cancer who had had a hysterectomy it was not possible to assess the effect of age at hysterectomy on ovarian cancer risk. Total duration of ovulation was estimated as the months from menarche to diagnosis (cases) or interview (controls), or to menopause, whichever came first, minus the total months of anovulation due to pregnancy and oral contraceptive use. Women who reported a hysterectomy were excluded from these analyses as it was unknown if or when they had stopped ovulating. For all women combined, there was a strong trend of increasing risk the longer the duration of ovulation (X: (trend) = 17.8. P <0.001) (Table IX). In separate analyses by menopausal status, there was no significant effect of duration of ovulation after adjustment for the `anovulatory' factors used to estimate that exposure, namely, months of pregnancy and oral contraceptive use and age at menopause for post-menopausal women and months of pregnancy and oral contraceptive use for premenopausal women. Duration of ovulation is very sensitive to age but the risks and trends were virtually unaffected when adjusted for age in one year rather than five year strata. Five (2%) cases and 29 (6%) controls reported having taken hormone pills as a pregnancy test and five (2%) cases and 13 (3%) controls had been given hormones to prevent miscarriage. For all post-menopausal women, there was a small but non- significantlv increased risk of ovarian cancer associated with ever having recieved hormone replacement therapy (Table X). The excess was confined to women who had reported a hysterectomy who had an 11-fold risk. The cases did not report more severe menopausal symptoms. Among the hormone treated women with ovarian cancer. 23% had endometrioid or clear cell tumours compared to 38% in the untreated women. The reproductive and related factors found to be statistically significantly related to ovarian cancer risk (gravidity, duration of unprotected intercourse, use of oral contraception, having been sterilised, age at natural menopause and having had a hysterectomy) are not independent and we also computed the relative risks associated with each factor after adjusting for the others (Table XI). As sterilisation is often a consequence of high parity, the risks associated with gravidity were not adjusted for sterilisation as this was considered to be overadjustment. In this study. 40% of the sterilised women had five or more children compared with 9% of the unsterilised women. Each of the variables remained statistically significantly related to ovarian cancer risk, suggesting that each may be independently associated with the risk of developing ovarian cancer. There was no significant difference between the percentage of cases (53%) and controls (57%) who had ever smoked cigaret tes. No cases or controls reported having worked with asbestos. No cases but three controls reported a radiation-induced menopause. Women who reported using talc more than once a week or daily had higher risks of ovarian cancer than women who reported less frequent use (Table XII). Although the relative risk of 2.0 associated with weekly use was statistically significant Table VIII R elative risks for o v a n a n can cer associated w ith rep o rted history o f hysterectom y and or unilateral oophorectom y C ases C ontrols R R 195% C l) R eported w om b intact R eported unilateral oophorectom y by no hysterectom y R eported hysterectom y but conserved ovaries R eported hysterectom y and unilateral oophorectom y 220 370 1.0" 5 9 0.9 (0 .4 -2 .1 ) 8 62 0.2 (0.1 -0 .4 ) ri 10 0.4 < 0 .1 -1 .1 ) R elative risks adjusted for age an d social class 'R eferen ce category Table IX R elativ e risks fo r o v a ria n c a n c e r a ss o c ia te d w ith d u ra tio n o f o v u la tio n D uration o f ovulation t y ea rs) C ases C o n i rots R elative risks adjusted for age a n d so cia l class Relative risks adjusted for age. social class and duration o f anovulation R elative risks adjusted for age. social class. duration o f anovulation and age at menopause All w o m en " <30 30-34 35-39 >40 59 163 1.0' 73 106 2.0 68 92 2.0 19 13 4.3 X : fo r tren d = 17.8 P < 0 .0 0 1 Post-m enopausal w om en <30 30-34 35-39 >40 14 53 1.0' TO' TO' 54 81 2.4 2.1 0.9 58 72 2.4 1.9 0.6 14 10 5.0 4.0 0.7 X ; for tre n d = 12.3 P < 0 .0 0 1 1.1 P < 0 .0 1 0.5 Prem enopausal women <30 30 - 34 35-39 >40 45 1 10 TO' TO' 19 25 1.5 1.1 10 20 1.3 0.9 53 3.2 1.9 X for irend = 4 4 P < 0.05 0.6 Women reporting a hysterectomy excluded: 'Reference category. "Data missing for 6 cases and 4 controls. Total months o f pregnancy and oral contraceptive use. rI ri ri O Table X Relative risks for ovarian cancer associaied with the use o f hormone replacement therapy lo r menopausal symptoms Cases Controls RR (95% Ch A ll post-menopausal women Use o f hormone replacement therapy No Yes 249 34 44 1.0' 1.5 (0.9-2.6) Women reporting hysterectomy Use o f hormone replacement therapv No Yes 5 62 1.0' 5 10 10.9 (1.7-69.0) Post-menopausal women other than those reporting hysterectomy Use o f hormone replacement therapy No Yes 177 187 29 34 1.0' 1.2 (0.7-2.3) Post-menopausal women only. Relative risks adjusted for age and social class. 'Reference category. Table X I Relative risks associaied with the factors found to be significantly related to ovarian cancer Factor RR Graviditv'-" 0 1 *> 3 4 >5 1.0' 0.8 0.8 0.6 0.4 0.5 Unprotected intercourse (months)3 < 3 1.0' 4-60 1.3 61 -120 1.1 > 120 1.9 Oral contraceptive use* Never used 1.0' ^ 5 years 0.6 6 -1 0 vears 0.6 > 10 years 0.1 Ever sterilized3 0.2 195% Cl) (0.4 - 1.5) (0.4 - 14) (0.3 - 1.1) (0.2 - 1.0) (0.2 - 1.0) (0.8 - 2.0) (0.5 - 2.5) (1.2 - 3.2) (0.3 - 1.1) (0.3 - 1.4) (0.02- M) (0.05- 0.6)* test for trend 11 d.f.) 6.5 P <0.05 7.8 P < 0 .05 4.6 P <0.05 Age at natural menopause3 < 4 5 vears 45-49 vears (p. 50 years 1.0' 1.9 (0.8 - 4.5) 2.6 (1.1 - 6.1) 8.2 P <0.01 Ever reported hysterectomy ' (0 1 - 0.5)* The relative risks associated with each factor have been adjusted for age. social class and all the other factors :n the table. 'Reference category. 'Sexually active women only. 'Relative risks not adjusted for sterilization, see text for details. JVVomen reporting natural menopause only mP <0.001. Table X II Relative risks for ovarian cancer associated with reported frequency o f talc use in the genital area Cases C ontrol> R R < Cl R eported frequency o f talc use'- Never Rarely M onthlv W eeklv Daily 6 rs h Ib I ld X- fo r tre n d Mr ').9 (0.3-2.41 10.3 1 X) :.o (1.3 -3 .4 , .3 (0.8 - 1.9) .-..so. P = 0.05 R elativ e ribki. a d ju ste d lo r age a n d so cial .:a s s . 'R e fe re n c e c ategory . "D a ta m issing fo r IS (X%> eases a n d 17 id '1 ;. c o n tro ls as q u e stio n s on laic use in tro d u ced three m o n th s a fte r stud'- regun (P = 0.007), there was no consistent trend of increasing risk with increasing frequency of talc use (X' (trend) = 3 .so P = 0.05). There was no significant difference between the percentages of cases (86%) and controls (81%) who had used and kept their diaphragm in talc. Discussion As in most previously reported studies (Booth & Beral. 1985) we found that nulligravid women had an increased risk of ovarian cancer and that risk decreased as the number of pregnancies increased. We also found that the greater the number of incomplete pregnancies the lower the risk, although the trend was not significant. Most other studies have not investigated if women of low gravidity have an increased risk of ovarian cancer because of reduced fertility or because of voluntary limitation of family size, although Jolv ei ul. (1974), McGowan ei al. (1979) and Nasca et al. (1984) found a higher risk in women who had tried to conceive but had failed. Our findings also suggest that infer tility is a risk factor for ovarian cancer. Women who had not conceived but had been sexually active for more than 10 years without using contraception had about six times the risk of all other women. Approximately half these women had undergone investigations for infertility. For only five cases and one control was the cause of their infertility deter mined. Thus, it is not possible to assess whether this high risk group had normal or impaired ovarian function. Subfertility may also be associated with ovarian cancer. Gravid women reporting over 10 years of unprotected intercourse had a 50% higher risk than other gravid women, but this increase was not statistically significant. Age at first pregnancy was not found to be associated with ovarian cancer risk although women having a first pregnancy after the age of 35 years had a higher risk compared to women having a first pregnancy at earlier ages and to nullig ravid women. Since subfertility might be a risk factor for ovarian cancer, the relative risks associated with age at first pregnancy were also adjusted for duration of unprotected intercourse. The raised risk for women with a first pregnancy after 35 years persisted (RR = 3.9. 95% Cl 1.1- 14.2). Results from other studies regarding the risk for women having a first child at relatively older ages are inconclusive, some finding no association (Newhouse et al.. 1977; Casagrande ei al.. 1979: Cramer et al.. 1983; Lesher ei a!.. 1985). others an increased risk (Joly et al.. 1974; McGowan ei al.. 1979: Hildreth ei al.. 1981; Franceschi et at.. 1982). Only Frances chi et al. (1982) found the increased risk to be statistically significant and independent of parity. Overall, there was no association between use of any con traception and ovarian cancer. Of the specific methods studied, female sterilisation and use of oral contraception were associated with a significant reduction in risk and no method was associated with a significant increase in risk. The associations remained after adjusting for gravidity and dur ation ol unprotected intercourse, the measure used to indicate infertility. While few studies have examined the association between female sterilisation and ovarian cancer, the relation between oral contraceptives and ovarian cancer has been demonstrated in many studies (Booth & Beral. 1985). Like others, we demonstrated that the longer oral contraceptives had been used, the lower the risk. Our findings also suggested that the earlier the age at first use the lower the risk and that the protective effect of oral contraceptives persists after their use is stopped. It has been suggested th a t in h ib itio n o f o v u la tio n , as induced by pregnancy and o ra l co ntraceptives, is the fa c to r w h ich protects against o va ria n cancer (F a th a lla . 1971). I f so. p o s tp a rtu m a n o v u la tio n associated w ith la c ta tio n m ig h t also be expected to be p ro te ctive . We fo u n d no evidence that the longer a wom an had breastfed the lower her risk o f ova ria n cancer indeed, the highest risk was found in those who had breastfed longest. Results fro m o th er studies are contradic- 'K RISK FACTORS FOR O V A R IA N CANCER "isk .80. ihe "ed o> of of he k. es in t\ :h o r- Jt 0 e n e k lory (Cramer et al.. 1983; Mori el al., 1984; Risch el al.. I9g3 : Cancer and Steroid Hormone Study, 1987), Our finding that age at menarche was not associated with rjsk is consistent with results from most other studies (Casasrande el al.. 1979; McGowan el al.. 1979; Hildreth el al.. ]9S1; Franceschi el al.. 1982). Age at natural menopause, however, was strongly related to risk. While Hildreth el al. (1981). Franceschi el al. (1982) and Tzonou el al. (1984) also demonstrated that the later the age at menopause the greater the risk, other studies have found no association (West. 1966; N,'ewhouse et al.. 1977; Annegers et al.. 1979; McGowan el al.. 1979; Cramer el al.. 1983). A lower frequency of hysterectomy, of unilateral oophorec tomy. or of both among cases compared to controls has also been reported from several other studies (Wynder et al., 1969; Joly el al.. 1974; Annegers el al., 1979; McGowan el al.. 1979; Franceschi el al., 1982; Cramer el al.. 1983). As these studies were case-control in design, there may have been some misclassification of controls who, rather than having a hysterectomy with ovarian conservation, .actually had a hysterectomy with bilateral oophorectomy. Another explanation for the findings might be that if at hysterectomy a woman's ovaries look diseased, it is likely that they are removed. If the diseased ovaries were precancerous. those women who might otherwise have developed ovarian cancer do not. Following hysterectomy with ovarian conservation, reduced ovarian function or ovarian failure occurs in a pro portion of women, due possibly to the blood supply to the ovaries being compromised (Beavis el al., 1969; Ellsworth el al.. 1983). Female sterilisation was also associated with a low risk of ovarian cancer. Neil el al. (1975) have suggested that the menstrual disturbance that many women experience after sterilisation may reflect changed ovarian function due to damage to the vascular supply to the ovaries. If both hysterectomy and female sterilisation can indirectly affect ovarian function then both procedures could also influence the risk of ovarian cancer. Recent investigators have shown that the longer a woman ovulates the greater her risk of ovarian cancer (Casagrande et al.. 1979; Hildreth el al.. 1981; Franceschi et al.. 1982; Wuc/ a!.. 1988). We also found that risk increased the longer the duration of ovulation. Duration of ovulation is. however, highly cor related with the 'anovulatory' factors used to estimate tinexposure. In an attempt to determine whether duration o:' ovulation had an effect over and above that expressed b\ is relation with these factors, the risks and trends were adjusted idduration of anovulation due to pregnancy and oral contracep tive use and. where appropriate, for age at menopause. Tinsignificance of the effect disappeared. We conclude that u is no; possible to determine from these data whether it is the above factors which inhibit ovulation that prevent ovarian cancer whether repeated ovulations promote it. or whether a combina tion of both is acting. Our finding of no overall relationship between hormone replacement therapy and ovarian cancer supports those of other investigators(Newhousee/a/.. 1977; Hildreth el al.. 1981; Weiss et al.. 1982). An increased risk associated with the therapv was found among women who reported hysterectomy, but the finding was based on very few cases and may have been due to chance. We did not find an increased risk associated with oestrogen therapy for any particular tumour types as suggested by Cramer el al. (1981) and Weiss et al. (1982). The evidence linking talc with ovarian cancer is controversial (Anonymous, 1977; Roe. 1979; Longo & Young 1979; Cranu-i et al.. 1982; Hartge et al.. 1983). In this study, women who reported talc use in the genital area more than once a week or daily had higher risks of ovarian cancer than women who used talc less frequently. The women were not asked how long the\ had been using talc. It is possible that because of their symptoms or disease-related pelvic examinations, the frequency of current talc use by the cases may not have reflected their frequence of past use. Since these and other results (Cramer el al.. I9S2; Hartge el al.. 1983) are insufficient to reject an association, further work is needed on the relation between genital use of tale and ovarian cancer. We would like to thank the Imperial Cancer Research bund lur financing the study. Professor Sir Richard Doll for helpful advice. Professor Christopher Hudson and D r Marigold Curling for review,nc the histology. D r Eve W iltshaw. the consultants, nurses and other M a il o f the participating hospitals for their co-operation and support o f the study. Ann Bateman. Kate Rodriques. G illian Saunders and Rosalie Thomson for their skilful interviewing, and Nina Saroi for typing tliemanuscript. Margaret Booth is funded by a grant from the Medical Research Council. References A N N EG ER S. J F.. STROM . H.. D ECKER. D.G.. DOCKERTY. M B & O 'FA LLO N . W M ( 1979). Ovarian cancer: incidence and case control study. C a n te r . 43, 723. ANONYMOUS (1977). Cosmetic talc powder. Editorial. L a n c e t, i. I34S. BAKER. R J. k NELD ER. J A. (1978). T h e G U M S y s te m R e le a se J. Royal Statistical Society; Oxford BEAVIS. E L G . B R O W N . J B. k SM ITH. M.A. (1969). Ovarian function after hysterectomy with conservation o f the ovaries in pre menopausal women. J. O b s ic i. G y n a e c o l. B r. C n lth . 76, 969. BOOTH. M. & BE RA L. V 11985). The epidemiology o f ovarian cancer. In O v a ria n C a n c e r . Hudson. C.N (ed). p. 22. O xford University Press; Oxford. BRESLOW. N.E & DAY. N.E. (1980). S ta t i s t i c a l M e th o d s in C a n c e r R e s e a r c h . I 'o l. I. T h e A n a ly s is o f C a s e - c o n t r o l S tu d ie s . IA R C Scientific Publications No. 32. IAR C : Lyon. Ca n c e r a n d s t e r o id h o r m o n e st u d s' o f t h e c e n t r e s fo r d isea se c o n t r o l a n d t h e n a t io n a l in s t it u t e o f c h il d h e a l t h A N D HUMAN DEVELOPM ENT (1987). T h e re d u c tio n in risk o f o v a ria n c a n c e r a sso c ia te d w ith o ra l-c o n ira c c p tiv c use. A. Engl. J M e d .. 316, 650. CA SA GRANDE. J.T.. PIKE. M C . ROSS. R K . LOUIE. E.W . ROY. S & HENDERSON. B E (1979], Incessant o v u latio n ' and ovarian cancer L a n c e t, ii. 170. CRAM ER D W .. DEVESA. S.S. & W E L C H . W .R. (1981). T re n d s in the in cid en ce o f e n d o m e trio id an d c le a r cell c an c e rs o f the o v a ry in the U n ite d S tates. A m . J E p id e m io l.. 114, 201. CRAM ER. D W W E L C H . W R .SCULLY. R E & WOJCIECHOW SKI. C A (1982). O varian cancer and talc: a case-control stu d s. C ancer. 50, 372. CRAM ER. D W . H U TCH ISO N . G B W ELCH. W.R . SCULLY, ILL X RYAN. K.J (1983). D e te rm in a n ts o f o v a ria n c an c e r risk I. R ep. re d u c tiv e ex p erien ces a n d Tamils h isto rv . J . S a i l C a n c e r l i n t 71 711. ELLSWORTH. L.R.. ALLEN. H.H. k NTSKER. J.A. ( I9 S 3 |. O v a ria n fu n ctio n a fte r rad ical h y sterec to m y fo r S tage IB carcinoma u | cervix. A m . J . O h s te t. G y n e c o l.. 145. 185 FATHALLA. M.F (1971). Incessant o v u latio n - a facto r in ov arian n eoplasia? Lancet, ii. 163. FRANCESCHI. S.. LA YECCHIA. C.. HELM RICH. S.P . M ANGIONI. ( X TOGNONI. G. (1982). Risk factors for epithelial o v arian cancel in Italy. A m . J. E p idem iol.. 115, 714. HARTGE. P.. HOOVER. R . LESHER. L.P k M cGOW AN. L (1983). In k an d ov arian cancer, J. A m . M e d . A sso c.. 250. 1844 ' HILDRETH. N.G.. KELSEY. J.L.. LIVOLSI. V A k 5 others 11981, An e p id e m io lo g ic stud> o f ep ith e lia l c a rc in o m a o f the o v a r \. Am ./ E pidem iol.. 114, 398. JOLY. D.J.. LILIENFELD. A M.. DIAM OND. E.L. k BROSS. I D.J I 1974, An epidem iologic study o f the relationship of reproductive ex p erien ce to c an c e r o f th e o v a ry -tin. J. E p id e m io l. 99. 1 9(, LESHER. L . McG OW AN. L . HARTGE. P. k HOOVER. R. (1 9 8 5 , Age .,i first birth and risk o f epithelial ovarian cancer. J S a il ( o n e r h r.t 74. 1361. LO N G O . D.L. k Y O U N G . R.C. (1979), C o s m c lic la ic and o v a ria n cancc; L a n c e t. ii. 349. Mi GOW'AN. L . PARENT. L . LEDNAR W X NORRIS. II.J 11 9 7 9 , i , lt. w om an al risk for developing ovarian can cer G \n e t a t 7 325. MORI. M.. KIVOSAWA. H. 4 MIYAKE. H. (1984). C a s e - c o n tr o l s tu d y o f ovarian cancer in Japan. C ancer. 53. 2746. NASCA. P C,. G R EEN W A LD . P C H O R O ST .S . RICHART. R .. 4 C A P U T O . T. 1 1984). A n e p id e m io lo g ic c a s e - c o n tr o l s tu d y o f o v a r i a n c a n c e r a n d r e p r o d u c tiv e fa ctors . A m . J . E p id e m io l.. 119, 705. NEIL. J R . H A M M O N D . G.T . NOBLE. A.D.. RUSHTON. L. 4 LETCH W O R TH . A T. |1975). L a te c o m p l i c a t i o n s o f ste ril is ati on by laparoscopy a n d tubal ligation. L a n cet, ii, 699. NEWHOL'SE. M L.. PEARSON. R.M.. FUL LE RTON . J.M.. BOESON. E A M 4 SHANNON. H.S. (1977). A case c o n tr o l s t u d y o f c a r c i n o m a o f the o v a r y . B r. J P r e r. S o c ia l M e d .. 31, 148. OFFICE OF POPULATION CENSUSES AND SURVEYS. (1970). C lassifications o f O ccupations. H.MSO: London. RISCH. H.A.. WEISS. N S.. LYON. J.L.. D A LING. J.R.. & L1FF. J.M. (1983). Events o f reproductive life a n d the incidence o f epithelial ovarian c an cer. A m . J . E p id e m io l.. 117. 128. ROE. F.J.C. (1979). C o n trove rsy, cosmetic talc a n d o v a ria n cancer L a n c e t, ii, 744. TZONOU. A.. D A Y . N.E.. TRICHOPOULOS. D 4 4 others (1984). The e p id e m io lo g y o f o v a r ia n c a n c e r in Gre ece : a c a s e -c o n tro l siud v. p ur J . C a n ce r C lin. O n c o l.. 20. 1045. WEISS. N S.. LYON. J.L.. KR1SH.NAMURTHY. S.. DIETERT. S.E.. LIFF J.M. 4 DALING. J.R. (1982). N o n c o n t r a c e p t i v e e str o g e n use a n d the oc cu r an c e o f o v a r ia n cancer. J . S a i l C a n ce r In s t.. 68. 95. WEST, R.O. ( I966). E p idem io logic s tu d v o f m a lia n a n c ie s o f th e ovaries C a n ce r. 19, 100! WU. N.L.. WHITTEMORE. A.S.. P.AFFENBARGER. R.S. 4 7 others (1988). Personal and environmental characteristics related to epithelial o v a r i a n c anc e r. 1. R e p r o d u c t iv e a n d m e n s t r u a l events a n d oral contraceptive use. A m . J . E p id em io l.. 128. 1216. W YNDER. E.L.. DODO. H 4 BARBER. H.R.K (1969). Ep id em io lo g y of cancer o f the ovary. C ancer. 23, 352. REGULATORY TOXICOLOGY AND PHARMACOLOGY 2 1 , 2 1 1 - 2 1 5 (1 9 9 5 ) Talc: Consumer Uses and Health Perspectives1 B eth esd a , M a ry la n d , J a n u a r y 3 1 -F e b r u a ry 1, 1994 Co-Sponsored by: The International Society of Regulatory Toxicology & Pharmacology and The United States Food and Drug Administration R apporteur C. Jelleff C arr Professor E m eritus, D epartm ent of Pharmacology and E xperim ental Therapeutics, School of M edicine, U niversity of M aryland, Baltimore, M aryland 21201 Received October 1,1994 W orkshop Participants G a b r ie l a A d a m - R o d w e l l , C olgate-Palm olive, Piscataway, N ew Jersey S e a n A LTEK LU S E, U .S. Food & Drug A dm inistration, C F S A N -- Epidem iology Branch, W ashing ton, DC W il l ia m H . A s h t o n , J o h n so n & Jo h n so n C onsum er Products, Inc., S killm a n , N ew Jersey H a r l a n d A U S T IN , E m o ry U niversity School o f Public H ealth, A tlanta, Georgia C A TH E R IN E J. B a il e y , U.S. Food & D rug A dm inistration, W ashington, DC JO HN E . B a il e y , U.S. Food & Drug A dm inistration, W ashington, DC DO UG B a k e r , Luzenac A m erica, Englewood, Colorado A L I B a s a r a N, T h e G lidden Com pany, Strongsville, Ohio D o n n a M . B e a c h , Cosm air, Inc., Clark, N ew Jersey G a r y B o o r m a n , N IE H S , Research Triangle Park, N o rth Carolina D a n ie l W . B r ig g s , T h e P rocter & G am ble Co., C in cin n a ti, Ohio RO BERT B r O NAUGH, U.S. Food & D rug A dm inistration, W ashington, DC M . D a n B r o w e r , H oechst Celanese Corp., P ortsm outh, Virginia A r n o l d L . B r o w n , U niversity o f W isconsin M edical School, M adison, W isconsin C . JE LLEFF C a r r , Scientific Inform ation Assn., Columbia, M aryland M IC H A E L C h u d k o w s k i, Jo h n so n & Jo h n so n C onsum er Products, Inc., S killm a n , N ew Jersey C H R IS T O P H E R R . E . C o g g in s , R. J . R eyn o ld s Tobacco Co., W in sto n -S a le m , N o r th Carolina C H A R L O T T E H . C o p p e r t H IT E , M inerals Technologies, Inc., N ew York, N ew York J a m e s D . C r a p o , D uke U niversity, D urham , N o rth Carolina E d D a v e n p o r t , W hittaker, Clark & Daniels, Inc., S o u th Plainfield, N ew Jersey C H R ISTO PH ER D a v is , M em orial S lo a n -K etterin g Cancer Center, N ew York, N ew York PA TR IC IA D e l a n e y , U .S. Food & D rug A dm in istra tio n , Rockville, M aryland M ic h a e l D e p a , S ta te of M ichigan Dept, o f N a tural Resources, Lansing, M ichigan JO H N D o u l l , T h e U niversity of K ansas M edical Center, K ansas City, K ansas W il l ia m E . D r e s s l e r , Clairol, Inc., S ta m fo rd , C onnecticut K e v in E . D r is c o l l , T h e P rocter & G am ble Co., C in cin n a ti, Ohio JOSEPH E v a l l , E sanu, K atsky, K orins & Siger, N ew York, N ew York K E N N E T H J. F a l c i, U.S. Food & Drug A dm inistration, W ashington, DC K a r e n F a s s U L IO TIS, E stee L auder Com panies, M elville, N ew York TH O M A S F a z io , U.S. Food & Drug A dm inistration, W ashington, DC D E N IS E A . F e e , Jones, Day, Reavis & Pogue, W ashington, DC K a y FITZG ER A LD , Pfizer, Inc., N ew York, N ew York JOSEPH L . FLEISS, Columbia U niversity School of Public H ealth. N ew York, N ew York 1Presented, in part, at the International Society of Regulatory Toxicology and Pharmacology/U.S. FDA W orkshop on Talc: Consumer Uses and Health Perspectives, NIH, Bethesda, MD, January 31-February 1, 1994. 211 027 3-2300/95 $6.00 C o p y rig h t 1995 by Academ ic Press, Inc. A ll rights o f reproduction in any form reserved. TALC: CON SUM ER USES AND H EA LTH PER SPEC TIV ES A N N E W o l v e n -G a r r e t t , A. M . W olven, Inc., A tla n ta , Georgia R O N A LD J. W U L F , Carter-W allace, Inc., Cranbury, N ew Jersey A n n G . W Y L IE , U n iversity o f M aryland, College P ark, M a ryla n d E R N S T L . W y NDER, Am erican H ealth Foundation, N ew York, N ew York JEFFREY Y o u RICK, U.S. Food & D rug A dm inistration, W ashington, DC B e v e r l y Z a k a r ia n , Cancer P a tien ts A ction Alliance, Inc., Brooklyn, N e w York W il l ia m Z a v a d o s k i, U .S. Cosm etics Corporation, Dayville, Connecticut R IC H A R D Z a z EN SK I, Luzenac Am erica, Inc., T hree Forks, M ontana 213 EXECUTIVE SUMMARY T h is issu e o f th e jo u rn a l is la rg e ly d e d ic a te d to re p o rt o n a J a n u a r y 3 1 -F e b r u a r y 1, 1 9 9 4 w o r k s h o p o n ta lc , o r g a n iz e d u n d e r jo in t s p o n so r sh ip o f th e U .S . F o o d a n d D r u g A d m in is t r a tio n ( F D A ) , th e C o sm e tic s, T o ile trie s, a n d F ra g ra n c e s A sso c ia tio n (C T F A ), a n d th e In te rn a tio n a l S o c ie ty o f R e g u la to ry T o x ic o lo g y a n d P h a rm a c o l o g y ( IS R T P ) . A lth o u g h n o t a ll p a p e rs g iv e n a t th e m e e t in g w ere m a d e a v a ila b le fo r p u b lic a tio n , th is o ffe rs a g e n e ra l o v e rv ie w o f th e s u b sta n c e o f th e p re se n ta tio n s a n d d isc u ssio n s. T h e w o rk sh o p w a s to p ro v id e a fo ru m fo r a n u p d a te d d isc u ssio n o f th e o rig in s, m a n u fa c tu re , c h a ra c te riz a tio n , to x ic o lo g y , a n d e p id e m io lo g y o f ta lc . P r in c ip a l fo c u s o f th e m e e tin g w a s o n th e la te st to x ic o lo g ic a n d e p id e m io lo g ic stu d ie s, a s th e y re fle c t o n th e sa fe u se s o f ta lc in c o n su m e r p ro d u c ts. T h e c h a ra c te ristic s o f c o sm e tic g ra d e ta lc , th e h is to r y o f its u se s in a v a rie ty o f p ro d u c ts, a n d th e c u rre n t q u a lity -c o n tro l m e a su re s to e n su re sa fe ty w e re fo llo w e d b y a re v ie w o f th e re g u la to ry h is to ry o f ta lc a n d b y a n a p p ra isa l o f re c e n t N a tio n a l T o x ic o lo g y P r o g r a m ( N T P ) stu d ie s o f c h ro n ic p u lm o n a ry e x p o su re o f r o d e n ts to ta lc . O f sp e c ia l in te re st w a s th e re le v a n c e o f th e se stu d ie s to h u m a n risk a sse ssm e n t, in v ie w o f re p o rte d te c h n ic a l p ro b le m s o f lu n g o v e rlo a d fo r th e se b io a ssa y s, a n d o f th e s ta n d in g c o n c e rn s a b o u t th e p h y sio lo g ic , a n a to m ic , g e n e tic , a n d o th e r d iffe re n c e s o f ro d e n ts a n d m a n . E x p e r i m e n ta l d a ta w e re th e n e v a lu a te d a g a in s t th e c o n tr a s tin g e v id e n c e e m e rg in g fro m e p id e m io lo g ic stu d ie s o f h u m a n ta lc e x p o su re . A c ritic a l p a n e l o f in v ite d e x p e rts a n d sp e a k e rs c o m p le te d e a c h se ssio n . F a c u lty a n d p a n e lis ts in c lu d e d ta lc m in e ra lo g is ts , g e o lo g ists, to x ic o lo g ists, e p id e m io lo g ists, p a th o lo g ists, fo o d /d r u g /c o s m e tic /m e d i c a l d e v ic e m a n u fa c tu r e rs, q u a lifie d re g u la to ry s p e c ia l ists, a n d c o n su m e r re p re se n ta tiv e s. In b rie f o p e n in g re m a rk s, D r. J o h n E . B a ile y ( F D A ) re m in d e d p a rtic ip a n ts th a t th e w o rk sh o p w a s n o t p a rt o f a fo rm a l ru le -m a k in g p ro c e ss a n d c o u ld n o t b e a fo ru m to re a ch a c o n se n su s fo r re g u la to ry d e c isio n s. E m p h a s is w a s to be o n a free in te ra c tio n b e tw e e n p a r tic ip a n ts to e x p lo re a n d p o ss ib ly re a c h so m e c o n c lu s io n o n th e v a lid ity a n d sig n ific a n c e o f th e e x is tin g k n o w le d g e re g a rd in g th e s a fe ty o f c o sm e tic ta lc . A first a n d e sse n tia l p re se n ta tio n b y D r. G e ttin g s ( C T F A ) a d d re sse d c o n c lu siv e ly th e n a tu re o f c o sm e tic ta lc , th e sp e c ific fo rm o f ta lc u n d e r s tu d y n o w re c o g n iz e d in th e 1990 U .S . P h a rm a c o p o e ia , a n d o th e r o ffic ia l s t a n d a rd s. U n iq u e p h y sic a l a n d c h e m ic a l p ro p e rtie s-- in e rt n e ss, h y g ro sc o p ic ity , se lf-a g g re g a tio n , lu b ric a tio n , ta c tile s e n s a tio n , tr a n slu c e n c y , a n d c o lo r -- m a k e p o w d e re d ta lc d e sira b le , u se fu l, a n d v ir tu a lly in d is p e n sa b le in c o s m e tic a p p lic a tio n s . C o sm e tic ta lc is a n e g lig ib le fra c tio n o f th e n e a rly 5 0 ,0 0 0 to n s m in e d fo r in d u s tria l u se in th e U n ite d S ta te s a n d to d a y strin g e n t sa fe ty a n d q u a lity c o n tr o l m e a su re s e n su re th e a b se n c e o f a sb e sto s fib e rs fo rm e rly c o n sid e re d a s a p o te n tia l h a z a rd , a lb e it n o t a d e fin e d one. D r. G ilb e r ts o n ( F D A ) re v ie w e d th e h a r m o n iz a tio n o f in te rn a tio n a l sta n d a r d s a n d re g u la tio n s fo r c o sm e tic ta lc a n d its c o n s u m e r a p p lic a tio n s . In th e ir jo in t e v a lu a tio n , ta lc h a s p ro v e n to be a m o n g th e sa fe st o f a ll c o n su m e r p rodu cts. In a d d re ssin g fu n d a m e n ta l a sp e c ts o f in h a la tio n to x i c o lo g y , D r. O b e rd rste r (R o c h e s te r U n iv e r s ity ) n o te d th a t a n y in h a le d d u st-- a n d m o st su b sta n c e s fo r th a t m a tte r -- m a y c a u se in fla m m a t io n a n d c e ll p ro life r a tio n in th e lu n g , p o ss ib ly le a d in g to tu m o r fo rm a tio n if th e c h a lle n g e is su ffic ie n tly str o n g a n d p e rsiste n t to o v e r c o m e th e n a tu ra l d e fe n se s o f th e a n im a ls u n d e r test. T h is c o n d itio n im p lie s th e p re se n ce o f a n o -e ffe ct th re s h o ld a t le v e ls b e lo w w h ic h n a tu ra l d e fe n se s re m a in fu n c tio n a l to p re v e n t tu m o r fo rm a tio n . H e n o te d th a t th e N T P in h a la t io n b io a s sa y s o f ta lc in v a ria b ly c re a te d lu n g o v e rlo a d s, th u s m a k in g th e in te rp re ta tio n o f re su lts q u ite p ro b le m a tic . In h is o v e ra ll e v a lu a tio n o f th e to x i c o lo g ic lite ra tu re o n ta lc , th e re is n o re a so n fo r c o n c e rn fo r lo w -le v e l u s e s o f c o s m e tic ta lc . A s a sta ff sc ie n tist o f th e N a tio n a l T o x ic o lo g y P r o g ra m , D r. B o o r m a n e m p h a siz e d th a t N T P p ro to c o ls c a ll fo r m a x im u m to le ra te d d o se s ( M T D s ) in a n a tte m p t to id e n t ify a n y h a z a r d s. N e g a t iv e fin d in g s m a y re c e iv e lit tle o r n o fu rth e r a tte n tio n , b u t p o s itiv e o n e s c a ll fo r d e ta ile d m e c h a n istic stu d ie s to d e te rm in e th e ir re le v a n c y fo r h u m a n h e a lth . In a d e ta ile d d e sc rip tio n o f th e ta lc in h a la tio n b io a s sa y s a t N T P , th e ir p ro to c o l, a n d p a t h o l o g y fin d in g s, h e c o n c u rre d w ith e v id e n c e o f d o se o v e r lo a d s in m o st o f th e a n im a ls th a t e n d e d u p d e v e lo p in g tu m o rs. H e e x p la in e d th e la c k o f a n o v a ria n e ffe ct o f life tim e e x p o su re in F 3 4 4 /1 4 ra ts a n d B 6 C 3 F 1 m ice . In h is su m m a ry re p o rt he n o te s th e m a n y facto rs th a t c o m p li c a te in te rp re ta tio n o f th e se ro d e n t stu d ie s. T h e re su lts o f th e N T P in h a la tio n b io a s sa y s o n ta lc w ere first e v a lu a te d a s c u s to m a ry in 1992 b y th e T e c h n i c a l R e p o rts R e v ie w S u b c o m m itte e o f th e N T P B o a r d o f S c ie n tific C o u n se lo rs. D r. G o o d m a n (M ic h ig a n S ta te TALC: CON SUM ER USES AND H EA LTH PER SPEC TIV ES 215 o f e p id e m io lo g ic stu d ie s m o re s tra ig h tfo rw a r d a n d d i re ct? U n fo r t u n a t e ly -- D r . G o r i su g g e ste d -- e p id e m io lo g y p re se n ts a n e w se t o f im p e d im e n ts. T h e e p id e m io lo g y w e a re fa c in g d o e s n o t o ffe r th e sa m e d ire c t c a u se a n d e ffe ct a ss o c ia tio n s ty p ic a l o f in fe c tio u s d ise a se s. H u m a n c a n ce rs a re m u ltifa c to ria l d ise a se s a risin g fro m a c o m b in a tio n o f sim u lta n e o u s e x p o su re s to m a n y p o te n tia l e tio lo g ic d e te rm in a n ts. T o e x tric a te th e s ig n ific a n c e o f a n y o n e o f th e se fa c to rs fro m th e in te g ra te d e ffe cts o f a ll o th e rs is a c h a lle n g in g ta sk . If o n e a d d s th e te c h n ic a l p ro b le m s in th e e x e c u tio n o f th e se stu d ie s, w e s o o n fin d th a t th e e p id e m io lo g ic fo g is ju st a s d iffic u lt to p e n e tra te a s th e o n e g e n e ra te d b y a n im a l b io a ssa y s. K e n n e th R o t h m a n (1 9 8 6 ), a le a d in g th e o re tic ia n o f A m e ric a n e p id e m i o lo g y , h a s re v ie w e d e x te n siv e ly th e se d iffic u ltie s o f in te rp re ta tio n o f c a u s a l in fe r e n c e s a n d w rite s: Despite philosophic injunctions concerning inductive infer ence, criteria have commonly been used to make such inferences. The justification offered has been that the exigencies of public health problems demand action and that despite imperfect knowl edge causal inferences must be made. T h is d e fin itio n is w id e ly a c c e p te d b y m a in stre a m e p i d e m io lo g ists to d ay . O n th is b a sis h o w e v e r-- ju st a s w e c o u ld a sk w h e th e r e x tra p o la tio n fro m a n im a l b io a s sa y s q u a lifie s a s a sc ie n tific e x e rc ise -- w e a re ju stifie d in a s k in g th e sa m e q u e stio n o f h u m a n e p id e m io lo g y . In a g e n e ra l o v e rv ie w o f o v a r ia n c a n c e r e p id e m io lo g y , D r. A u s tin (E m o r y U n iv e rsity ) n o te d th e m a n y fa c to rs th a t m a y c o n fo u n d th e p u ta tiv e a ss o c ia tio n o f p e rin e a l ta lc e x p o su re a n d o v a ria n c a n c e r. F o llo w in g a p re se n ta tio n b y D r. B ro w n (U n iv e rsity o f W isc o n sin ), th e d isc u s s io n m a d e it c le a r th a t a v a ila b le h is to lo g ic a n d p h y s io lo g ic stu d ie s p ro v id e n o b a s is to c o n c lu d e th a t ta lc c a n m ig r a te to th e o v a rie s fro m th e p e rin e a l re g io n . D r. H a r tg e (N a tio n a l C a n c e r In stitu te ) a n d D r. H a r lo w (H a r v a r d U n iv e r s ity ) p re se n te d a re v ie w o f e p id e m i o lo g ic s tu d ie s -- in c lu d in g th e ir o w n o r ig in a l stu d ie s-- p e rta in in g to p e rin e a l ta lc e x p o su re a n d o v a ria n c a n c e r risk . T h e stu d ie s re v ie w e d b ro u g h t to lig h t th e m a n y in te rp re tiv e d iffic u ltie s o f e p id e m io lo g y a s a n o b se rv a tio n a l sc ie n c e a n d a re d e ta ile d in th e p a p e rs b y D r s . H a r tge a n d H a r lo w a p p e a rin g in th is issu e o f th e jo u rn a l. F r o m th e u n iq u e p e rsp e c tiv e o f h is lo n g a n d d is t in g u ish e d e x p e rie n c e in th e e p id e m io lo g y o f m u ltifa c to ria l d ise a se s, D r. W y n d e r (A m e ric a n H e a lth F o u n d a tio n ) re v ie w e d th e h isto ry o f th e e v o lu tio n a ry p ro c e sse s th a t p ro v id e d n a tu ra l d e fe n se sy s te m s a g a in s t d ise ase . H e stre sse d th e sp e c ific p ro b le m s in th e e p id e m io lo g y o f w e a k a sso c ia tio n s: b ia se s o f re sp o n d e n ts a n d in v e s tig a to rs, k n o w n a n d u n k n o w n c o n fo u n d in g facto rs, a n d the irre sistib le u rg e to in te rp re t re su lts a s if o n ly a re d u c e d se t o f v a ria b le s o f in te re st w a s o p e ra n t, w ith o u t a c k n o w le d g in g a n d c o n tr o llin g fo r a m o re c o m p le x m u lti fa c to r ia l re a lity . F o llo w in g th e m a n y iss u e s ra ise d b y a ll p re se n te rs, th e e n su in g d isc u ssio n g e n e ra lly a gre e d th a t w h ile so m e w e a k a ss o c ia tio n b e tw e e n ta lc e x p o su re a n d o v a r ia n tu m o rs h a s b e e n re p o rte d , it w a s n o t su ffic ie n t w a rn in g fo r con ce rn . A fin a l p a n e l in c lu d e d m o st sp e a k e rs a n d o th e r e x p e rts a n d w a s a b le to re a c h a n u n a n im o u s a sse ssm e n t o f th e w o rk sh o p . In re g a rd to th e N T P ta lc b io a s sa y in ro d e n ts, it fo u n d th a t b e c a u se o f th e e x tre m e d o se s a n d th e u n re a listic p a r tic le siz e s o f th e ta lc e m p lo y e d , b e c a u se o f th e n e g a tiv e re su lts in m ic e a n d m a le ra ts, b e c a u se o f th e la c k o f tu m o r e x c e ss a t th e lo w d o se s, a n d b e c a u se o f th e c le a r b io c h e m ic a l a n d c y to lo g ic a l m a rk e r s o f e x c e ssiv e to x ic ity in fe m a le ra ts, th e p o sitiv e ta lc b io a ss a y re su lts in fe m a le F 3 4 4 /N ra ts a re th e lik e ly e x p e r im e n ta l a rtifa c t a n d n o n sp e c ific g e n e ric re sp o n se o f d u st o v e rlo a d o f th e lu n g s a n d n o t a re fle c tio n o f a d ire c t a c tiv it y o f ta lc . G iv e n th e g r o s s d iffe re n c e s o f ro d e n t a n d h u m a n lu n g s, th e lu n g c le a ra n c e c a p a b ilitie s o f h u m a n s, a n d th e p o ssib le c o n d itio n s o f c u sto m a ry h u m a n e x p o su re s, th e N T P b io a s sa y re su lts in F 3 4 4 /N fe m a le ra ts c a n n o t b e c o n sid e re d a s re le v a n t p re d ic to rs o f h u m a n risk . In re g a rd to th e p ro p o se d a ss o c ia tio n o f ta lc e x p o su re a n d o v a ria n c an c e r, th e p a n e l fo u n d th a t e p id e m io lo g ic d a ta a re c o n flic tin g a n d re m a in e q u iv o c a l. A lth o u g h it is th e o re tic a lly p o ssib le th a t ta lc c o u ld re a c h th e o v a rie s, th e a c tu a l a c c e ss to o r th e p re se n ce o f ta lc in o v a ria n tissu e is n o t d o c u m e n te d . D ie t, p a rity , c o n tra c e p tiv e u se, o v u la to ry fre q u e n c y , fa m ilia l p re d isp o sitio n , a ge to m e n a rc h e a n d m e n o p a u se , a n d o th e r fa c to rs a sso c ia te s tr o n g ly a n d p la u s ib ly w ith o v a r ia n c a n c e r in c id e n c e . T h e s e p o ssib le c o n fo u n d e rs a n d c o n tr o l se le c tio n b ia se s, p u b lic a tio n b ia se s, in te rv ie w e r a n d in te rv ie w e e b ia se s, a n d o th e r fa c to rs m a y w e ll e x p la in th e c o n flic tin g re su lts th a t h a v e a p p e a re d in th e lite ra tu re . T h e p o s s ib ilit y o f a n a ss o c ia tio n o f ta lc e x p o su re a n d o v a ria n c a n c e r is a n im p o r ta n t h y p o th e sis o f p o te n tia l p u b lic h e a lth im p o rta n c e . H o w e v e r, th is a sso c ia tio n re m a in s a re se a rch h y p o th e sis w h o se v e rific a tio n o r fa ls i fic a tio n n e e d s a d d itio n a l stu d y . E x p e rim e n ta l stu d ie s m a y b e n e e d e d to d e te rm in e w h e th e r ta lc c o u ld a c c e ss th e o v a rie s u n d e r fie ld c o n d itio n s a n d , if so , w h e th e r it c o u ld b e e m b e d d e d in o v a ria n tissu e a n d p ro d u c e p a th o lo g ic e ffe cts. F o r e p id e m io lo g y , fu rth e r re fin e m e n ts m a y b e p o ssib le in th e se le c tio n a n d c h a ra c te riz a tio n o f c o n tro l su b je c ts a n d in th e a c c o u n tin g o f p o ssib le c o n fo u n d e rs a n d b ia se s. H o w e v e r, e p id e m io lo g ic stu d ie s h a v e p r o v id e d w e a k a n d c o n flic tin g risk s ig n a ls fo r th is a sso c ia tio n , a n d it is u n lik e ly th a t fu rth e r stu d ie s m a y p ro v e a d e q u a te to ra ise c o n c e rn a t a le v e l su ffic ie n t to w a r ra n t re g u la to ry o r p u b lic h e a lth m e a su re s. REFERENCE Rothman, K. J. (1986). Modern Epidemiology, p. 17. Little, Brown & Co., Boston. ITJ Ovarian Cancer and Talc A Case-Control Study DANIEL W. CRAMER. i WILLIAM R. WELCH, MD. ROBERT E. SCULLY, MD.1 AND CAROL A. WOJCIECHOWSKI, RN* Opportunities for genital exposure to talc were assessed in 215 white females with epithelial ovarian cancers and in 215 control women from the general population matched by age, race, and residence. Ninety-two (42.8%) cases regularly used talc either as a dusting powder on the perineum or on sanitary napkins compared with 61 (28.4%) controls. Adjusted for parity and menopausal status, this difference yielded a relative risk of 1.92 (P < 0.003) for ovarian cancer associated with these practices. Women who had regularly engaged in both practices had an adjusted relative risk of-3.28 (P < 0.001) compared; to women with neither exposure. This provides some support for an association between talc and ovarian cancer hypothesized because of the similarity of ovarian cancer to mesotheliomas and the chemical relation of talc to asbestos, a known cause of mesotheliomas. The authors also investigated opportunities for potential talc exposure from rubber products such as condoms or diaphragms or from pelvic surgery. :-.. No significant differences were noted between cases and controls in these exposures, although the intensity of talc exposure from these sources was likely affected by variables not assessed in this study. C a n c e r 50:372-376, 1982. Th e p o s s i b i l it y t h a t o v a r i a n c a n c e r m a y b e c a u s e d b y e x p o su re to c e rta in h y d ro u s m a g n e siu m s ili c a te s su c h a s ta lc a n d a sb e sto s h a s b ee n ra ise d b y se v e r a l r e s e a r c h e r s . 1" 3 T h e l a c k o f e p i d e m i o l o g i c s t u d i e s r e g a r d in g th is h y p o th e sis p ro m p te d u s to in v e stig a te ta lc e x p o su re in a c a se -c o n tro l s tu d y o f o v a r ia n c a n c e r. From the Departments of 'O bstetrics, tGynecology, and Pathology, Boston Hospital for Women, Division of the Brigham and Women's Hospital, the tDepartm ent of Epidemiology, Harvard School of Public Health and the *Department of Pathology, Massa chusetts General Hospital, Harvard Medical School, Boston, Mas sachusetts. Supported by G rant Number 5-RO l CA24209, awarded by the National Institutes of Health, DHEW. Address for reprints: Dr. Cramer, D epartm ent of Obstetrics and Gynecology, Brigham and Women's Hospital, Boston, MA 02115. This study could not have occurred without the generous partici pation of many clinicians and institutions in the greater Boston area including: Dr. Emanuel Friedman of the Beth Israel Hospital, Drs. Robert Knapp and Thomas Griffiths of the Brigham and Women's Hospital and Sidney Farber Cancer Institute, Dr. A rthur Hassett of the Brockton Hospital, Dr. Joel Rankin of the Framingham Union Hospital, Dr. Edward Copenhaver of the Lahey Clinic Foundation, Dr. James Nelson of the Massachusetts General Hospital, Dr. Clem ent Yahia of the New England Deaconess Hospital, Dr. Lalita Gandbhir of the Pondville Hospital, Dr. James Whelton of Saint Elizabeth's Hospital, Dr. Stephen Alpert of the Salem Hospital, Dr. Richard H unter of the University of Massachusetts Medical School. The su perb clerical and technical assistance of Ms. Eileen McManus, Ms. Sally Cassells, and Ms. Christine Peters is also gratefully acknowl edged. Accepted for publication December 29, 1981. M e th o d s T h e case s stu d ie d w e re w o m e n w ith o v a ria n ca n c e r, d ia g n o s e d b e tw e e n N o v e m b e r 1 9 7 8 a r d n.-pU ::.,L-. 1981 a n d id e n tifie d th r o u g h th e p a th o lo g y lo g s o r tu m o r b o a r d s o f tw e lv e p a r tic ip a tin g h o sp ita ls in th e G r e a te r B o sto n are a. T h e stu d y w a s re stric te d to E n g lis h -sp e a k in g re sid e n ts o f M a s s a c h u s e t t s r a n g in g in a g e fro m 18 to 8 0 y e a rs. D u r in g th e s tu d y p e rio d , 2 9 7 e lig ib le c a se s w e re id e n tifie d . P h y s ic ia n s d e n ie d p e rm issio n to c o n ta c t th e ir p a tie n ts in 13 in sta n c e s. F o u rte e n p a tie n ts d e c lin e d to p a rtic ip a te , a n d 14 o th e r p a tie n ts h a d d ie d o r m o v e d b e fo re th e y c o u ld be c o n ta c te d . F o r e a c h o f th e 2 5 6 in te rv ie w e d case s, slid e s o f th e su rg ic a l sp e c im e n s w e re re v ie w e d b y tw o a u th o rs ( W . R . W . o r R . E . S ) . E ig h te e n c a se s w e re e x c lu d e d a s n o n o v a ria n p rim a rie s. E a c h o v a ria n tu m o r w a s c la ss i fie d a c c o r d in g to th e H is t o lo g ic a l C la s s if ic a t io n o f O v a ria n T u m o rs o f th e W o r ld H e a lth O rg a n iz a tio n .4 T h e p re se n t a n a ly s is w a s re stric te d to 2 1 5 w h ite w o m e n w ith e p ith e lia l c a n c e rs, in c lu d in g 3 9 w ith tu m o rs o f b o rd e rlin e m a lig n a n c y a n d th e ir m a tc h e d c o n tro ls. C o n tr o l c a se s w e re id e n tifie d th ro u g h th e M a s s a c h u se tts T o w n B o o k s, a n n u a l p u b lic a tio n s th a t list re sid e n ts b y n a m e , a g e , a n d a d d re ss. C o n t r o ls w e re se le cte d r a n d o m ly fro m th o se w o m e n w h o m a tc h e d case s b y p re c in c t o f re sid e n c e , ra c e , a n d a g e w ith in tw o y e a rs. A d d it io n a lly , it w a s r e q u ir e d t h a t a s u b je c t b e e x c lu d e d 0008-543X /8 2/0715/0372 50.75 American Cancer Society 7 '7 "> ,1 ' ' >. y No. 2 O varian C a a n d T alc C ra m e r et ai. 373 i i a s a c o n tro l if sh e h a d h a d a b ila te ra l sa lp in g o -o o p h o - \ ! re c to m y , b u t su b je c ts w e re n o t e x c lu d e d b e c a u se o f p rio r h y ste re c to m y o r o th e r ty p e s o f p e lv ic o p e ra tio n s. T a b l e I . Characteristics of Cases a n d Control* Case* (Total = 215) Control* (Total = 215) W o m e n w h o h a d h a d p e lv ic o p e ra tio n s w e re g e n e r a lly c o n fid e n t in th e ir k n o w le d g e o f w h e th e r th e ir o v a rie s Characteristic No. % No. % h a d b e e n re m o v e d , b u t th e n a tu re o f th e o p e ra tio n s Educational level c o u ld n o t b e v e rifie d b y h o sp ita l re c o rd s in e a c h in sta n c e . W o m e n w h o se sta te m e n ts c o u ld n o t b e v e rifie d (completed college) Religion (Roman 48 22.3 49 22.8 w e re in c lu d e d o r e x c lu d e d o n th e b a sis o f th e ir re c o l Catholic) 126 58.6 128 59.5 le c tio n o f th e s u rg e ry . T h e 2 1 5 c o n tr o ls in th is, s t u d y M arital status w ere e v e n tu a lly o b ta in e d fro m a to ta l o f 4 7 5 p o te n tia l (never married) Nulliparous 46 21.4 78 36.3 24 11.2 39 18.1 c o n tro ls id e n tifie d th ro u g h th e T o w n B o o k s. F ift y -s ix Menopausal status V ( 1 2 % ) o f t h e t o t a l c o u l d n o t b e r e a c h e d b e c a u s e t h e y ( postmenopausal* ) 137 63.7 129 60.0 h a d m o v e d , d ie d , o r h a d d isc o n n e c te d o r u n liste d p h o n e s. T w e n ty -n in e (6 % ) o f th e to ta l w ere in e lig ib le b e c a u se o f a h isto ry o f b ila te ra l sa lp in g o -o o p h o re c to m y , * Postmenopausal at time of diagnosis for cases or for interview for Controls. vy w h ile 2 0 ( 4 % ) w ere o f th e w ro n g a g e o r ra c e o r d id n o t 5 3 .5 (1 .0 ) y e ars. T a b le 1 sh o w s o th e r c h a ra c te ristic s o f sp e a k E n g lis h . O f th e to ta l p o te n tia l c o n tro ls, 155 su b je c ts. C o n tr o ls w e re c o m p a r a b le to c a se s in e d u c a ( 3 3 % ) r e fu s e d to p a rtic ip a te . I f th e 2 1 5 c a se s a re c h a r tio n a l le v e l a n d re lig io n . C a s e s a n d c o n tr o ls d iffe re d a c te riz e d a s to e a se o f m a tc h in g , 121 (5 6 % ) c a se s w e re s ig n ific a n tly in m a r ita l s ta tu s a n d p a r it y w ith p a r ity m a tc h e d w ith n o re fu sa ls, 5 8 (2 7 % ) w e re m a tc h e d a fte r b e in g th e m o re im p o rta n t d is c r im in a to r b e tw e e n th e m . o n e re fu sa l, a n d 3 6 (1 7 % ) w e re m a tc h e d o n ly a fte r tw o S ix ty -fo u r p e rcen t o f th e case s w ere p o stm e n o p a u sa l at o r m o re re fu sa ls. th e tim e o f d ia g n o sis, w h e re a s 6 0 % o f c o n tro ls w e re In te rv ie w s w e re c o n d u c te d p e rso n a lly to a sse ss a p o stm e n o p a u sa l. O f th e se , 15 c a se s a n d 2 0 c o n tro ls h a d n u m b e r o f fa c to rs fro m th e m e n stru a l a n d re p ro d u c tiv e h a d a n a rtific ia l m e n o p a u se . P a r ity a n d m e n o p a u sa l h isto ry , m e d ic a l a n d fa m ily h isto ry , a n d e n v iro n m e n ta l s ta tu s w e re c o n sid e re d im p o r ta n t p o te n tia l c o n fo u n d e rs e x p o su re s. T h is re p o rt w ill d e a l o n ly w ith th e re su lts in th is a n a ly s is a n d w e re a d ju ste d fo r a s d e sc rib e d o f se v e ra l q u e stio n s re la te d to p o te n tia l o r d e fin ite ta lc above. e x p o su re b y w a y o f c o n tra c e p tiv e p ra c tic e s, o p e ra tio n s, R e la tiv e risk s a sso c ia te d w ith p o te n tia l ta lc e x p o su re o r p e rin e a l h y gie n e . S u b je c ts w e re stra tifie d b y p o te n tia l fro m c o n ta m in a tio n o n ru b b e r p ro d u c ts a re e x p lo re d c o n fo u n d e rs d e sc rib e d b e lo w , a n d a d ju ste d re la tiv e risk s in T a b le 2. A lt h o u g h s u r g ic a l g lo v e s o f re c e n t v in ta g e a sso c ia te d w ith th e se e x p o su re s w e re c a lc u la te d b y th e a re d u ste d w ith sta rc h , ta lc c o n ta m in a tio n m a y still b e M a n te l-H a e n sz e l p ro ce d u re a s a d a p te d b y R o th m a n fo u n d .7 T h u s , a h isto ry o f p e lv ic o p e ra tio n s (a p p e n d e c a n d B o ic e .5 T o a c c o m m o d a te o th e r c o n fo u n d e rs a s w e ll to m y , c e sa re a n se ctio n , h y ste re c to m y , a n d o th e r o p e r a s th e m a tc h e d d e sig n in th e d a ta c o lle c tio n , lo g is tic a tio n s o n in te rn a l g e n ita l o r g a n s o th e r th a n b ila te ra l a n a ly sis fo r m a tc h e d d a ta a s d e sc rib e d b y B re slo w et sa lp in g o -o o p h o re c to m y ) w a s d e te rm in e d in c a se s a n d al.b w a s a ls o e m p lo y e d . c o n tro ls. E x c lu d in g o p e ra tio n s a sso c ia te d w ith th e d i R e su lts a g n o sis o r tre a tm e n t o f th e o v a ria n c a n c e r a m o n g th e c a se s, n o e x c e ss in th e o c c u rre n c e o f p e lv ic o p e ra tio n s T h e a v e ra g e a ge (a n d sta n d a rd e rro r o f th e m e a n , w a s n o te d . T h e g re a te st o p p o rtu n ity fo r ta lc e x p o su re S E M ) fo r c a se s w a s 5 3 .2 (1 .0 ) y e a rs a n d fo r c o n tro ls, fro m s u rg e ry o c c u rre d b e fo re 1 9 5 0 , w h e n ta lc w a s th e T a b l e 2. Relative Risks (R R ) for Common Epithelial Ovarian Cancers Associated with Potential Talc Exposure from Contamination on Rubber Products Case* Controls Exposure Total No. (%) with exposure Total No. (%) with exposure Crude RR Adjusted RR* 95% Confidence limit* Pelvic surgery 215 78 (36.3) 215 75 (34.9) 1.06 1.17 Pelvic surgery (prior to 1950) 215 51 (23.7) 215 48 (22.3) 1.08 1.12 Use o f condom st 169 19 (11.2) 191 30 (15.7) 0.68 0.77 U se of diaphragm t 169 37 (21.9) 191 35 (18.3) 1.24 1.19 (0.76-1.79) (0.69-1.82) (0.41-1.44) (0.69-2.05) * Adjusted for parity (nulliparous, parous) and menopausal status (pre- and postmenopausal). t Restricted to subjects who had ever been married, 374 Cancer J u l y 1 5 1982 v0i. so T able 3. R elative Risks (R R ) Associated with Using T alc for S torage Am ong Diaphragm Users* by D uration o f Use o f D iaphragm Cases C o n tro ls Duration of diaphragm use Total No. (%) who used talc on diaphragm Total No. (%) who used talc on diaphragm Crude RR Adjusted RRt 95% Confidence lim its Total diaphragm use less th an five years Total diaphragm use five o r m ore years All users 13 27 40 6 (46.2) 16 (59.3) 22 (55.0) 21 19 40 8 (38.1) 1.39 1.82 11 (57.9) 1.06 1.23 19 (47.5) 1.35 1.56 (0.42-8.00) (0.36-4.17) (0.62-3.88) Includes all women who used diaphragm regardless of m arital status. t Adjusted for parity and menopausal status, p r e d o m in a n tly u se d d u s tin g p o w d e r fo r s u r g ic a l g lo v e s. H o w e v e r , n o s ig n ific a n t e x c e ss o f p e lv ic o p e ra tio n s p rio r to 1950 w as ob se rve d fo r cases. T h e p a tie n ts (c a se s ) w h o , a t so m e tim e , h a d b e e n m a rrie d , c h o se c o n d o m s le ss fre q u e n tly a n d d ia p h r a g m s m o re fre q u e n tly fo r c o n tra c e p tio n th a n th e c o n tro l g ro u p , b u t n e ith e r d iffe re n c e w a s sta tistic a lly s ig n ifi c a n t. C o n d o m u se is n o t n e c e ssa rily a sso c ia te d w ith ta lc e x p o su re . N o t a ll b ra n d s o f c o n d o m s a re d u ste d w ith ta lc , a n d lu b r ic a n ts c o u ld a ffe c t th e s h e d d in g o f ta lc fro m th e c o n d o m . U n fo rtu n a te ly , d e ta ils o n sp e c ific b r a n d s o f c o n d o m s w e re n o t o b ta in e d . S im ila r ly , ta lc e x p o su re is n o t a n e c e ssa ry c o n se q u e n c e o f d ia p h r a g m u se . W e in q u ire d sp e c ific a lly a b o u t th e p ra c tic e o f d u s t in g th e d ia p h r a g m w ith ta lc fo r sto ra g e a fte r u se (T a b le 3 ). A m o n g a ll su b je c ts w h o h a d u se d a d ia p h ra g m , th e re w a s n o sig n ific a n t e x ce ss o f c ase s w h o re g u la rly sto re d th e ir d ia p h r a g m u s in g ta lc , n o r w a s a n y g re a te r risk a sso c ia te d w ith th is p ra c tic e o b se rv e d a m o n g w o m e n w h o h a d u se d th e d ia p h r a g m fo r lo n g e r d u ra tio n s. B e fo re th e risk fro m th is e x p o su re c a n b e a d e q u a te ly a s se sse d , g r e a te r d e ta il is n e e d e d in c lu d in g fr e q u e n c y o f u se a n d w h e th e r th e p o w d e r w a s w a sh e d o ff p rio r to u se . F u r th e r m o re , c o n tr a c e p tiv e je llie s u se d w ith th e d ia p h r a g m c o u ld a ffe c t th e tr a n sp o rt o f ta lc in th e g e n ita l tract. \ H y g ie n ic p ra c tic e s in v o lv in g ta lc w e re a ls o stu d ie d . S p e c ific a lly , w e in q u ire d w h e th e r w o m e n h a d re g u la r ly u se d ta lc a s a d u stin g p o w d e r o n th e p e rin e u m o r re g u la r ly d u ste d sa n ita ry n a p k in s w ith ta lc (T a b le 4 ). N in e ty -tw o (4 2 .8 % ) o f th e c a se s h a d ta lc e x p o su re b y e ith e r o r b o th o f th e se ro u te s c o m p a re d w ith 61 (2 8 .4 % ) o f th e c o n tro ls. T h e a d ju ste d re la tiv e r is k w a s 1.92 (P < 0 .0 0 3 ) w ith 9 5 % c o n fid e n c e lim it s o f 1 .2 7 -2 .8 9 c o m p a re d to su b je c ts w h o h a d n e ith e r e x p o su re . S ix ty (2 7 .9 % ) c a se s a n d 4 8 (2 2 .3 % ) c o n tro ls h a d e ith e r u se d ta lc fo r d u s tin g o r o n n a p k in s b u t n o t b o th . T h is d if fe re n c e y ie ld e d a n a d ju ste d re la tiv e r is k o f 1 .5 5 , w h ic h w a s o f b o rd e rlin e sig n ific a n c e (P = 0 .0 6 ). T h e g re a te st r isk o c c u rre d in w o m e n w h o h a d b o th e x p o s u r e s (u se o n th e p e rin e u m a n d o n n a p k in s) c o m p a re d to w o m e n w h o h a d n e ith e r e x p o su re . T h ir ty -tw o (1 4 . 9 % ) o f c a se s w e re in th is c a te g o r y c o m p a re d w ith 13 ( 6 . 0 % ) c o n tro ls, fo r a n a d ju ste d re la tiv e risk o f 3 .2 8 (P < .0 0 1 ) ,, r .- - " * > c o n fid e n c e lim its o f 1 .6 8 -6 .4 2 . T h e h is to lo g ic c h a r a c te ristic s o f tu m o rs d e v e lo p in g in w o m e n w ith p e rin e a l e x p o su re to ta lc d id n o t d iffe r sig n ific a n tly fro m th o se in w o m e n w ith o u t p e rin e a l e x p o su re to ta lc (T a b le 5 ). In a d d itio n , th e p ro p o rtio n o f c a se s w ith tu m o rs o f b o r d e rlin e m a lig n a n c y w a s id e n tic a l a m o n g th o se w ith a n d w ith o u t p e rin e a l e x p o su re to ta lc . T w e n t y -t w o ( 1 8 % ) o f 123 c a se s w ith o u t th e e x p o su re h a d tu m o rs o f b o r- T able 4. Relative Risks (R R ) for Common Epithelial O varian C ancers A ssociated with Talc Exposure in Perineal H ygiene Types of perineal exposure No perineal exposure Any perineal exposure As dusting powder but not on napkins On napkins but not as dusting powder Both on napkins an< as dusting powder Cases (Total = 215) Controls (Total = 2 1 5 ) Crude rr Adjusted RR* 95% confidence lim its 123 (57.2% ) 154 (71.6% ) 1 -- -- 92 (42.8%) 61 (28.4% ) 1.89 1.92 (1.27-2.89) * Adjusted for parity and menopausal status. 43 (20.0% ) 17 (7.9%) 34(15.8% ) ' 1.58 14 (6.5% ) 1.52 1.55 (_0_._9_8-2.47) 32 (14.9%) 13 (6.0% ) 3.08 3.28 (1.68-6.42) No. 2 O v a r i a n C a a n d T a l c C ram er et al. 375 d e rlin e m a lig n a n c y c o m p a re d to 17 ( 1 8 % ) o f 9 2 w ith th e ta lc e x p o su re . Discussion T h e a rg u m e n t lin k in g ta lc a n d o v a ria n c a n c e r in c lu d e s fo u r e le m e n ts: th e c h e m ic a l r e la tio n s h ip b e tw e e n ta lc a n d a sb e sto s, a sb e sto s a s a c a u s e o f p le u ra l a n d p e rito n e a l m e so th e lio m a s, th e p o ssib le re la tio n b e tw e e n e p ith e lia l o v a ria n c a n c e rs a n d m e so th e lio m a s, a n d th e a b ility o f ta lc to e n te r th e p e lv ic c a v ity . T h e m in e ra l ta lc is a sp e c ific h y d ro u s m a g n e s iu m s ilic a te c h e m ic a lly re la te d to se v e ra l a sb e sto s g ro u p m in e r a ls a n d o c c u r r in g in n a tu re w ith th e m . G e n e ric " t a lc " is s e ld o m p u re a n d m a y b e c o n ta m in a te d w ith a sb e sto s, p a r tic u la r ly in p o w d e r s f o r m u l a t e d p r i o r t o 1 9 7 6 . 1,9 E p id e m io lo g ic stu d ie s h a v e c le a r ly lin k e d lu n g c a n c e r a n d p le u r a l a n d p e rito n e a l m e s o th e lio m a s w ith a sb e sto s e x p o s u r e . 10 A n e x c e s s o f s i m i l a r p u l m o n a r y l e s i o n s h a s b e e n re p o rte d in ta lc w o rk e rs a n d se e m s to be c o rre la te d w ith in c a m o u n t o f a sb e sto s c o n ta m in a tio n in th e ta lc d e p o sits w o r k e d ." G r a h a m a n d G r a h a m 1 w e re a b le to in d u c e o v a r ia n n e o p la sm s in g u in e a p ig s w ith a sb e sto s a n d su g g e ste d th a t o v a ria n c a n c e r c o u ld b e re la te d to a sb e sto s e x p o su re , n o tin g th e s im ila r ity b e tw e e n m e so th e lio m a s a n d o v a ria n can ce rs. P a r m le y a n d W o o d r u f f 2 fu rth e r e m p h a siz e d th is s im ila r it y a n d p o p u la r iz e d th e p e lv ic c o n ta m in a tio n th e o ry , w h ic h p ro p o se d th a t e n v iro n m e n ta l c a rc in o g e n s m ig h t e n te r th e p e lv ic c a v it y v ia th e g e n ita l tra c t. Y e a r s e a r lie r it h a d b e e n o b se rv e d th a t in e rt c a rb o n p a rtic le s p la c e d in th e v a g in a im m e d ia te ly p rio r to h y ste re c to m y c o u ld b e re co v e re d fr o m th e fa llo p ia n tu b e s ." A lt h o u g h g re e te d w ith sk e p tic is m , th e fin d in g o f ta lc p a r tic le s e m b e d d e d in n o r m a l a n d a b n o r m a l o v a rie s s u g g e s ts th a t ta lc is a su b sta n c e th a t c a n e n te r th e p e lv ic c a v ity v ia th e v a g in a .2 A lt h o u g h n o c o n se n su s c o n c e rn in g th e risk s o f ta lc h a s e m e r g e d f r o m l e t t e r s , e d i t o r i a l a n d a r t i c l e s , 3,14" 16 p a r tic ip a n ts in th e d isc u ssio n h a v e a g r e e d u p o n th e n e e d fo r e p id e m io lo g ic stu d ie s o f o v a r ia n c a n c e r a n d ta lc e x p o su re , in th is c a se -c o n tro l s tu d y o f o v a ria n c a n c e r o ( th e e p ith e lia l v a rie ty , w e in v e stig a te d se v e ra l so u rc e s o f p o te n tia l ta lc e x p o su re . A m o n g th e se , th e o n ly s ig n ific a n t fin d in g w a s a n a sso c ia tio n b e tw e e n o v a ria n c a n c e r a n d h y g ie n ic p ra c tic e s in v o lv in g th e u se o f ta lc o n th e p e rin e u m . It is e sp e c ia lly n o t a b le th a t w o m e n w h o r e g u la r ly h a d b o th d u ste d th e ir p e rin e u m w ith ta lc a n d h a d u se d it o n s a n it a r y n a p k in s h a d m o r e t h a n a t h r e e fo ld in c re a se in risk c o m p a re d to w o m e n w ith n e ith e r e x p o su re . S e v e r a l p o te n tia l b ia se s m u st b e c o n sid e re d in in te rp re tin g th is a sso c ia tio n . T h e o b s e r v a t io n b y W y n d e r et a l.'1 th a t m e n s tr u a l c h a ra c te ristic s m a y d iffe r b etw e e n w o m e n w ith o v a ria n c a n c e r a n d c o n tro ls m ig h t su g g e st th a t su c h d iffe re n c e s m : iv r o n f o u n d t h e :i r v c r > tin n h p t n m l m . i l n m n f T a b l e 5. C haracteristics of O varian C a n ce r in Women with a n d without Perineal Exposure to Talc No perineal use of talc Any perineal use of talc No. (%) No. (%) Serous Mucinous Endometrioid and clear ceil Other and undifferentiated Total 66 (53.7) 16 (13.0) 32 (26.0) 9 (7.3) 123 (1 0 0 ) 45 (48.9) 14 (15.2) 24 (26.1) 9 (9.8) 92(100) ta lc a n d o v a ria n c a n c e r. W e fo u n d th a t m e n stru a l c h a r a c te ristic s o f c a se s a n d c o n tr o ls w e re v ir tu a lly id e n tic a l in th is stu d y . F ifty -th r e e (2 4 . 7 % ) c a se s c o m p la in e d o f m od erate o r severe d y sm e n o rrh e a co m p are d to 56 (2 6 .0 % ) c o n tro ls. T w e n ty -fiv e (1 1 .6 % ) c a se s c o m p la in e d o f ir r e g u la r p e rio d s c o m p a re d to 3 2 (1 4 .9 % ) c o n tro ls. T h e a v e ra g e n u m b e r s (a n d S E M ) o f d a y s o f flo w a n d c y c le le n g th w e re , re sp e c tiv e ly , 4 .9 (0 .1 ) a n d 2 8 .9 (0 .3 ) d a y s fo r c a se s a n d 4 .9 (0 .1 ) a n d 2 9 .6 (0 .3 ) d a y s fo r c o n tro ls. S in c e e n try o f ta lc in to th e p e lv ic c a v ity is p re v e n te d b y h y ste re c to m y o r tu b a l lig a tio n , it m ig h t a lso b e a r g u e d th a t th e in c lu sio n o f su b je c ts w ith p e lv ic su rg e ry in th e a n a ly s is m a y o b v ia te a n y a s s o c ia tio n b e tw e e n ta lc a n d o v a ria n c an c e r. It sh o u ld be n o te d th a t su c h su rg e ry g e n e ra lly o c c u rre d n e a r th e e n d o f re p ro d u c tiv e life fo r b o th c a se s a n d c o n tro ls, p ro b a b ly a fte r m o st sig n ific a n t ta lc e x p o su re h a d a lre a d y o c c u rre d . T h e e x c lu sio n o f su c h su b je cts fro m the a n a ly sis d id n o t su b sta n tia lly a lte r th e o b se rv e d a sso c ia tio n s. F o r e x a m p le , th e a d ju ste d re la tiv e r is k fo r th e u se o f ta lc b o th o n th e p e r in e u m a n d s a n ita r y n a p k in s w a s 2 .7 9 (P < 0 .0 0 3 ) in th e g ro u p w ith o u t p e lv ic s u r g e r y c o m p a re d to 3 .2 8 o b se rv e d fo r th e e n tire g ro u p . In te rm s o f o th e r c o n fo u n d e rs, th e a sso c ia tio n p e r siste d a fte r a d ju stm e n t fo r m e n o p a u s a l sta tu s a n d p a r ity. W e a ls o a p p lie d m u lt iv a r ia t e lo g is t ic re g re ssio n fo r p a ire d o b se rv a tio n s.6 T h e m a x im u m lik e lih o o d e stim a te o f re la tiv e risk a ss o c ia te d w ith a n y p e rin e a l u se o f ta lc w a s 1.61 (P = 0 .0 3 ) w it h 9 5 % c o n fid e n c e lim it s o f 1 .0 4 2 .4 9 a fte r sim u lta n e o u s a d ju stm e n t fo r re lig io n , m a rita l sta tu s, e d u c a tio n a l le v e l, p o n d e r a l in d e x , a g e a t m e n a rch e , e x act p a rity , o ra l c o n tra c e p tiv e o r m e n o p a u sa l h o rm o n e use, a n d sm o k in g . O u r sa m p le o f c a se s re p re se n ts m o re th a n 5 0 % o f o v a ria n c a n c e r c a s e s d ia g n o s e d in B o s to n re sid e n ts in the- s tu d y p e rio d . T h e r e fo r e , it is d iffic u lt to c o n c e iv e o f a p la u sib le b ia s in th e se le c tio n o f c a se s th a t w o u ld y ie ld th is e x c e ss u se o f ta lc . T h e r e is re a so n fo r c o n c e rn th a t th e h ig h re fu sa l ra te a m o n g th e c o n tr o ls m a y h a v e i n t r n / t i K-i.i/l e r t l n o t i/A n K i o e rvi n n r v I Ji A TV . . a. ,+ 376 Cancer July 15 1982 v0l. 50 when we restricted the analysis to the 121 cases who were matched without a control refusal, we again found a significant association between talc use and ovarian cancer. For women who had used talc both in dusting and on the perineum we found an adjusted relative risk of 2.44 (P < 0.05). Interviewer bias is also unlikely to explain the association. O f the 18 women who were initially interviewed as ovarian cancer cases but later excluded as having metastatic tumors to the ovary, only one (5.6%) had both perineal and napkin exposure as compared with 15% in cases and 6% in controls. Experimental data which might bear on the carci nogenicity of talc come primarily from models using pleural implantation of various minerals in rats.18These data suggest that carcinogenicity is dependent primarily upon the shape of the particles with long thin fibers such as those occurring in crocidolite asbestos being most carcinogenic. Talc consists primarily of plates but may contain fibers, although voluntary guidelines to lim it the content of asbestisform fibers in consumer tal cums were proposed by the cosmetics industry in 1976.1' If talc is involved in the etiology of ovarian cancer, it is not clear whether this derives from the asbestos content of talc or from the uniqueness of the ovary which might make it susceptible to carcinogenesis from both talc and other particulates. With ovulation en trapment of the surface epithelium of the ovary into the ovarian stroma occurs. If present, talc or other partic ulates might be incorporated into these inclusion cysts. Apparently implantation of foreign bodies into the lu mens of epithelial lined organs provides a favorable environment for carcinogenesis.20 Alternatively, talc might serve to stimulate entrapment of the surface ep ithelium and act in the same way that "incessant ovu lation" has been proposed as an etiologic factor for ovarian cancer.21 Given the histologic and clinical di versity of ovarian cancer, talc exposure is unlikely to be the only cause. Undoubtedly, reproductive experi ences such as pregnancies and, perhaps, oral contra ceptive use play a role in its etiology.21"23The possibility that talc exposure interacts with these variables de serves further investigation. It is hoped that this report will stimulate further study of talc exposure in relation to ovarian cancer. Animal studies would be helpful to determine whether and un der what circumstances ovarian tumors may be induced by various talc preparations. Epidemiologic studies should focus on opportunities for excessive vaginal con tamination with talc such as when it is repeatedly used in perineal dusting powders or sprays and in or on tam pons, sanitary napkins, or other products intended for in tr a v a g in a l use. M o r e p re c ise d e ta ils o n th e e x a c t n a tu re a n d fre q u e n cy o f the e x p o su re a n d the a m o u n t a n d sp e c ific b ra n d o f p o w d e r u se d a re e sse n tia l. O p p o r t u n i t i e s f o r t a l c e x p o s u r e a r e w i d e s p r e a d a n d p e r v a s i v e , 24 b u t th a t sh o u ld n o t d is c o u ra g e e p id e m io lo g ists fro m s tu d y in g th is p o te n tia lly im p o r ta n t e x p o su re in re la tio n to o v a ria n can cer. REFERENCES 1. Graham J, Graham R. Ovarian cancer and asbestos. E nviron R es 1967; 1:115-128. 2. Henderson WJ, Joslin C A F , Turnbull A C , Griffiths K .T a lc and carcinoma of the ovary and cervix. J Obsiel G ynaecol B r C om m onw 1971; 78:266-272. 3. Longo D L , Young R C Cosmetic talc and ovarian cancer. Lancet 1979; ii:349--351. 4. Serov SF, Scully R E, Sobin L H . International Histological Clas sification of Tumours, No. 9. Histological Typing o f Ovarian T u mours. Geneva, World Health Organization, 1973. 5. Rothman KJ, Boice JD. Epidemiologic analysis with a program mable calculator. N IH Publication No. 79-1649, 1979. 6. Breslow N E , Day N E , Halvorsen K T , Prentice R L , Sabai C. Estimation of multiple relative risk functions in matched casc-contro! studies. A m J E pidem iol 1978; 108:299-307. 7. Henderson WJ, Hamilton T C , Griffiths K. Talc in normal and malignant ovarian tissue. Lancet 1979; i:499. 8. Cralley LJ, Key M M , Groth D H , Lainhart W S , Ligo R M . F i brous and mineral content o f cosmetic talcum products. A m In d H y g A ssoc J 1968; 350-354. 9. Rohl A N , Langer A M , Selikoff IJ, Tordini A , Klimentidis R . Consumer talcums and powders: M ineral and chemical characteriza tion. J T oxicol Environ H ea lth 1976; 2:255-284. 10. SelikofT I J. Hammond EC (cd s). Health hazards of asbestos exposure. A nn N Y A ca d Sci. 1979; 330:1-179. - 11. Kleinfeld M , Messite J. Z aki MH. M ortality experiences among talc workers: A follow-up studv. J Occup M e d 1974; 16:345 349.' ` 12. Parmley T H , Woodruff JD. The ovarian mesot!~i:-:.-...: .'.m J O bsiel Gynecol 1974; 120:234-241. 13. ' Egli GE, Newton M . The transport of carbon particles in th human female reproductive tract. F crtil S te r il 1961; 12:151-155. 14. Anonymous. Cosmetic talc powder. Lancet 1977; i: 1348. 15. Newhouse M L . Cosmetic talc and ovarian cancer. Lancet 1979; ii:528. 16. Roe FJC. Controversy: Cosmetic talc and ovarian cancer. L a n cet-1979; ii:744. 17. Wynder EL, Dodo H , Barber H R K . Epidemiology o f cancer of-the ovary. Cancer 1969; 23:352-370. ( 18.! Stanton M F , Layard M , Tegeris A , el at. Relation of particle dimension to carcinogenicity in amphibole asbestoses and other fibrous minerals. J N a tl Cancer In stitu te 1981; 67:965-975. 19. C .T .F .A . Specification. Talc, cosmetic: Cosmetic, toiletry, and fragrance association, Inc. Issue 10-17, 1976. 20. Brand KG , Johnson K H , Buoen LC. Foreign body tumorigen- esis. C R C Crit R ev T o xico l 1976; 4(O ct):353-394. 21. Casagrande JT, Pike M C , Ross R K , Louie E W , Roy S, H en derson B E Incessant ovulation and ovarian cancer. Lancet 1979, ii: 170-172. 22. Newhouse M L , Pearson R M . Fullerton J M , Boescn E A M , Shannon US. A case control study of carcinoma of the ovary. B r J Prev So c M e d 1977; 31:148-153. 23. McGowan L, Parent L, Lednar W . Norris HJ. The woman at risk for developing ovarian cancer. Gynecol O ncol 1979; 7:325-344. 24. Blejer JP, Arlon R. Talc: A possible occupational and envi ronmental carcinogen. J O ccup M e d 1973; 15:92-97. Original Contribution !% /c '2. f e /1 3 Tubal Ligation, Hysterectomy, and Risk of Ovarian Cancer A Prospective Study Susan E. Hankinson, ScD; David J. Hunter, MB, BS; Graham A. Colditz, MB, BS; Walter C. Willett, MD; Meir J. Stampfer, MD; Bernard Rosner, PhD; Charles H. Hennekens, MD; Frank E. Speizer, MD Objective.--To assess whether tubal ligation and hysterectomy affect subse u n d e r w e n t a t u b a l s t e r i l i z a t i o n p r o c e quent risk of ovarian cancer. dure.3* S im ila rly , o v e r 4 0 0 0 0 0 h y ste r Design.--Prospective cohort study with 12 years of follow-up. . Setting.--United States, multistate. Participants.--A total of 121700 female registered nurses who were 30 to 55 years of age in 1976; the follow-up rate was 90% as of 1988. Main Outcome Measure.--Ovarian cancer of epithelial origin confirmed by medical record review. Results.--We observeda stronginverseassociationbetween tubal ligationand ovarian cancer, which persisted after adjustment for age, oral contraceptive use, parity, and other ovarian cancer risk factors (multivariate relative risk [RR], 0.33; e cto m ie s w ith th e re m o v a l o f n e ith e r o r o n ly one o v a ry a re p e rfo rm e d e ach year.6 A n in v e rse a sso c ia tio n b e tw e e n tu b a l ligatio n an d o v a ria n c an ce r risk h as been r e p o r t e d i n p a s t c a s e - c o n t r o l s t u d i e s , * '12 a lth o u g h th e m a g n itu d e o f th e o d d s ra tio s v a rie d s u b st a n tia lly (ra n g e , 0.2 to 0.9). In a s m a ll re tr o sp e c tiv e c o h o rt s t u d y , 13 a n o n s i g n i f i c a n t i n c r e a s e d r i s k o f o va ria n can cer w a s noted. A n in v e rse 95% confidence interval [Cl], 0.16 to 0.64). The association was similar when we a s s o c ia t i o n b e t w e e n h y s t e r e c t o m y a n d assessed tubal ligation statusatthe baselinequestionnaire and excluded cases in the first 4 years to eliminate any possible short-term decrease in risk due to screening of the ovaries during ligationsurgery. We noted a weaker inverse asso ciation between simple hysterectomyand ovarian cancer (RR, 0.67; 95% Cl, 0.45 to 1.00). Neithervasectomy norcondomusebya partnerwas associated with risk of ovarian cancer. Conclusions.--These data indicate that tubal ligation, and perhaps hysterec tomy, may substantially reduce risk of epithelial ovarian cancer. o va ria n can cer h a s a lso b een re p o r t e d . 12-1* '" F o r b o t h o f t h e s e p r o c e du re s, th e o b se rv e d in v e rse a sso c ia tio n m a y be du e to a b ia s re su ltin g fro m scre e n in g th e o v a rie s a t th e tim e o f s u r g e r y , r a t h e r t h a n a b i o l o g i c e f f e c t 7-8-18; th is issu e re m a in s u n re so lv e d . In th is article , w e p ro v id e th e first p ro sp e ctive a sse ssm e n t o f th e re la tio n (JAMA. 19932702813-2818) s h i p s o f t u b a l li g a t i o n a n d h y s t e r e c t o m y w ith risk o f o v a ria n can ce r d u rin g 12 O V A R IA N c an c e r is th e fo u rth le a d in g cause o f c a n c e r d e a th in A m e ric a n a ffe ct 1 % to 2 % o f w o m e n in th e ir life tim e ,2 o f w h o m fe w e r th a n 4 0 % w ill s u r y e a rs o f fo llo w -u p in a la rg e co h o rt o f re giste re d n u rse s. w om en, a cc o u n tin g fo r a p p ro x im ate ly 12 0 0 0 d e a t h s p e r y e a r .1T h is d is e a s e w ill v iv e 5 y e a rs fro m d ia gn o sis.1 T h e poor p ro g n o sis re in fo rc e s th e n e e d to fin d METHODS m o d ifiab le r is k facto rs. T h e N u rse s' H e a lth S tu d y is a p ro From the Channing Laboratory (Drs Hankinson, Hunter, Colditz. Willett. Stampfer. Rosner. and Speizer) and the Division of Preventive Medicine (Dr Hennek For editorial comment aee p 2855. sp e ctive coh ort stu d y o f 121700 m a r rie d, fem ale re g iste re d n u rse s w h o w e re 30 to 55 y e a rs o f a ge in 1976 w h e n the ens). Department of Medicine. Brigham and Women's Hospital: the Departm ent of Ambulatory Care and Pre vention (Dr Hennekens). Harvard M edical School: and w the Departments of Epidemiology (Drs Hankinson. tanter. Colditz. Willett, and Stampfer) and Nutrition (Dr W u t t ). Harvard School ot Public Health, Boston, Mass. Reprint requests to Channing Laboratory, 180 Long- wood Ave, Boston, MA 021 15 (Dr Hankinson). B o th tu b a l liga tio n a n d h y ste re cto m y are co m m o n ly p e rfo rm e d in th e U n ite d State s. F ro m 1970 to 1980 in th e U n ite d S ta te s, a p p ro x im a te ly 6J5 m illio n w o m e n betw een the a ge s o f 15 and 44 ye ars stu d y b e gan . P a rtic ip a n ts h a v e re ce ive d q u e stio n n aire s b ie n n ia lly sin ce th a t tim e to u p d ate in fo rm a tio n on e x p o su re sta tu s an d n e w ly d ia g n o se d illn e sse s. T h ro u g h J u n e 1,1988, th e fo llo w -u p ra te a m o n g liv in g p a rtic ip a n ts w a s 9 0 % . JAMA. December 15. 1993--'Vol 270, No. 23 Tubal Ligation. Hysterectomy, and Ovarian Cancer-- Hankinson et al 2813 Table 1.-Age-Standardized Prevalence (%) of Potential Ovarian Cancer Risk Factors According to Tubal Ligation and Hysterectomy Status in 1976* Rtafc Factor I !I I Yee No Yes No ParttyS 0 28 57 1 39 58 2 24 30 28 28 70 52 61 55 Age at menarche, y <11 23 23 25 22 213 51 50 49 51 OraJ contraceptive use Yes 60 54 52 49 Use for 2 5 y 17 10 13 11 S m o k in g s Never 43 44 45 43 Past 23 24 22 23 Current 33 32 33 33 Q u etelefs index, kg/m2 <21 27 28 26 26 229 10 10 11 10 'D irectly standardized to 1976 age distribution of all premenopausal (lor tubal ligation) or all premenopausal and postmenopausal women combined (for hysterectomy); numbers indicate percentages of subjects with potential risk factor. t Population includes 77 544 premenopausal women. Population includes 107 866 premenopausal and postmenopausal women. N um beis do not always add up to 100% because of missing data. Ascertainment of Exposures In 1976, p a rtic ip a n ts w e re ask e d w h e th e r th e y cu rre n tly u se d a n y m ethod o f c o n tra c e p tio n an d , if so, to sp e cify w h ich o f th e fo llo w in g m e th o d s th e y use d ; o ra l co n trace p tiv e s, rh y th m , d ia p h r a g m , co n d o m , in tra u te rin e de vice , f o a m o r j e l l y , t u b a l l i g a t i o n , o r h u s b a n d 's vasectom y. W o m en w ho had ever used o ra l co n tra c e p tiv e s w e re a sk e d to sp e cify in te rv a ls o f u se sta rtin g fro m first use a n d c o n tin u in g u p to 1976. P a rtic ip a n ts w e re a lso a sk e d if th e ir m e n stru a l p e rio d s h a d ce ase d and, if so, a t w h a t a ge th is occu rre d , fo r w h a t re a son (n atu ral m enopause or m enopause d u e to s u rg e ry o r ra d ia tio n ) and, if su r g ic a l m e n o p au se , w h e th e r th e ir o varie s h a d b e e n re m o v e d . W e h a v e a lso col le c te d d a ta o n h e ig h t, w e ig h t, sm o k in g sta tu s, a g e at m e n arch e , a n d p arity . P a r ity w a s d e fin e d a s the n u m b e r o f p re g n an cie s la stin g 6 m o n th s o r m ore. D a ta o n m e n o p a u sa l statu s, sm o k in g, and w e igh t h ave been updated e ve ry 2 years; c o n tra c e p tiv e u se w a s a sce rta in e d e v e ry 2 y e a rs u n til 1982 (b y th en use w as u n c o m m o n ) a n d p a r ity w a s a sk e d b ie n n ia lly u n til 1984. In 1982, w e ask e d p a r ticip an ts if th e y h ad e v e r co m m o n ly used talc, b a b y p o w d e r, o r d e o d o riz in g p o w d e r to a p p ly to th e ir p e rin e u m o r on sa n ita ry n ap kin s. T h e re p ro d u c ib ility a n d v a lid ity o f se lf reported m enopausal statu s has been a sse sse d a m o n g N u rse s' H e alth S tu d y p a rtic ip a n ts w h o a n sw e re d q u e stio n n a ir e s in 1976 a n d 1 9 7 8 .'9 R e p o r t in g o f age at natural m enopause, assessed am on g over 31000 postm enopausal w o m e n , w a s re p ro d u cib le (w ith in 1 y e ar) 8 2 % o f th e tim e ; re p ro d u c ib ility in c re a se d w ith d e c re a sin g tim e sin ce m enopause (ran ge, 1 to a 2 5 years). T o d e te rm in e th e v a lid ity o f se lf-re p o rte d su rg ic a l m e n o p au se , w e e x am in e d m e d i cal re c o rd s fro m a sa m p le o f in cid e n t m enopausal w om en. F o r 65 w om en w ho rep orted h a v in g had a h yste rectom y only, th e re w a s co m p le te a gre e m e n t w ith th e m e d ical record. A m o n g 66 w om en . w h o re p o rte d h y ste re c to m y p lu s exci sio n o f o n e o v a ry , th e re w a s 9 7 % a gre e m e n t w ith th e m e d ical record, a n d for 69 w om en w ho reported hysterectom y p lu s b ila te ra l o o p h o re cto m y, se lf-re p o rts w e re c o n firm e d in a ll b u t one w om an . T o d e te rm in e th e d istrib u tio n o f typ e s o f tu b a l liga tio n p ro ce d u re s p erform ed in th is co h o rt, w e re q u e ste d p e rm issio n to o b ta in m e d ical re c o rd s fro m 100 ra n d o m ly ch o se n w o m e n w h o h ad rep orted h a v in g a tu b a l liga tio n procedure. Study Population In a ll a n a ly se s, p a rtic ip a n ts w h o re p o rte d a d ia gn o sis o f can ce r (except for n o n m e la n o m a sk in can cer) o r w h o re p o rte d h a v in g one o r both o va rie s re m o v e d w e re e x clu d e d , le a v in g 107868 w o m e n in o u r b a se lin e p o p u la tio n in 1976. T h e se e x clu sio n s w e re u p d ate d e v e ry 2 ye ars. F ro m 1976 to 1988, the cohort accrued a total of 1191524 persony e a r s o f fo llo w -u p . W e lim ite d a n a ly se s o f tu b a l liga tio n to w o m e n w ho w ere p re m e n o p a u sa l in 1976 b ecau se (except fo r o ra l c o n trac e p tiv e u se ) w o m e n had been a sk e d to in d icate th e co n trace p tive m e th o d c u rre n tly use d, a n d th u s w e h ad n o in fo rm a tio n on tu b a l liga tio n a m o n g p o stm e n o p a u sa l w o m e n in 1976. T h is a d d itio n a l e x clu sio n re su lte d in the 1976 b ase lin e p o p u la tio n o f77 544 w o m e n and, fro m 1976 to 1988, a to tal o f 859 791 p e rso n -y e a rs o f fo llo w -u p . In a ll a n a ly se s, w o m e n c o n trib u te d p e rso n -tim e in each 2 -y e a r in te rv al u n til th e re p o rt of a u n ilate ra l o r b ila te ra l o o p h o re cto m y, c a n c e r, o r d e a th , o r u n til J u n e 1, 1988. Case Ascertainment In 1976 w e a sk e d p a rtic ip a n ts to in dicate if th e y h a d e v e r b e e n d ia gn o se d w ith can cer, a n d if ye s, to in d icate the typ e. O n e ach su b se q u e n t qu e stio n n aire , w e sp e c ifica lly a sk e d if o v a ria n can cer h ad b ee n d ia gn o se d in th e p re v io u s 2 ye ars. In a d d itio n , w e se arch e d th e N a tio n al D e a th In d e x to id e n tify w o m e n w h o m a y h av e d ie d o f cancer. M o rta lity fo llo w -u p h a s b e e n e stim a te d to b e 9 8 % co m ple te in th is co h o rt." F o r e ach re ported case o f o v a ria n cancer, w e re q u e ste d p e rm issio n fro m th e p articip a n t (or fro m th e n e x t o f k in for deceden ts) to o b ta in m e d ical re co rd s; th e m e dical re co rd s w e re re v ie w e d b y p h y sician s w ith o u t k n o w le d g e o f e x p o su re statu s. O n ly w o m e n co n firm e d to h a v e in cid e n t e p ith e lial o v a ria n can ce r (e ith e r b o rd e r lin e o r m a lig n a n t) o n m e d ic a l re c o rd re v ie w w e re in clu d e d a s case s in th is a n a ly sis. T h e re su lts w e re u n c h a n g e d w h e n th e v e ry sm a ll n u m b e r o f c ase s w ith b o rd e rlin e tu m o rs (n = 17) w e re e xclu de d , so w e p re se n t re su lts th a t in clu d e b o th b o rd e rlin e a n d m a lig n a n t tu m o rs. Data Analysis F o r e a ch p a rtic ip a n t, fo llo w -u p tim e, equal to the n u m b er of m on ths betw een the re tu rn o f th e 1976 q u e stio n n aire and the re tu rn o f th e 1978 q u e stio n n aire , w as a ssign e d a cc o rd in g to each co va ri ate b y its sta tu s in 1976. S im ila rly , for each su b se q u e n t 2 -y e a r in te rv al, w e a s sig n e d a d d itio n a l m o n th s o f fo llo w -u p a cc o rd in g to the u p d ate d e x p o su re s at th e b e g in n in g o f th e in te rv al. H yste recto m y statu s w as updated ev e ry 2 y e ars. B e ca u se w o m e n re p o rtin g o o p h o re cto m y w e re ce n so re d a t th e tim e o f th is re p o rt, h y ste re c to m y , a s a n e x p o su re in th is a n a ly sis, is b y d e fin itio n a sim p le h y ste re c to m y . In d e fin in g a ge at natural m enopause, w om en w ho w ere prem enopausal w hen th ey had a h ys te re c to m y h a d all su b se q u e n t p e rso n tim e a s s ig n e d to th e "u n k n o w n " c a t egory. T u b a l liga tio n an d v a se cto m y sta tu s w as u p d ate d e v e ry 2 y e a rs u n til 1982 (the la st tim e th e q u e stio n w a s asked); th e re sp o n se in 1982 w a s th e n m a in tain e d th ro u g h 1988. C o n d o m u se w as not u p d ate d , a s su b sta n tia lly few er 2814 JAMA, December 15. 1993--Voi 270. No. 23 Tubal Ligation. Hysterectomy, and Ovarian Cancer-- Hankinson el al p us 3. he i en ; 91 I yi n" of iy, . *8. m ed he re, er. .2 ia en ty i% *e-emt ts) cal .ns as. ;nt ?rre y en ith ed, )th ae, en nd re, iri- for as- up at ?v- ng me 2X- ion tge )re /s- )n - at ta- )82 d); :n - as /er l al w om en reported use at each subsequent q u e stio n n aire ; th u s, th e re sp o n se in 1976 w as ju d g e d th e b e st e stim a te o f a w o m a n 's l o n g - t e r m e x p o s u r e . In c id e n c e ra te s w e re calcu la te d fo r each c ate go ry o f an e xposu re b y d iv id in g th e n u m b e r o f o va ria n can cer cases for th a t c a te g o ry b y th e p e rso n -tim e o f fo llo w -u p . R e la tiv e risk s ( R R s ) w e re u sed a s th e m e a su re o f a sso cia tio n an d w e re c a lcu la te d a s th e ra tio o f in cid e n ce rates o f th e expose d to the unexposed. T o co n tro l sim u lta n e o u sly fo r o th e r po te n tial r is k fa c to rs fo r o v a ria n can cer, w e u s e d p r o p o r t i o n a l h a z a r d s m o d e l s . 21-22 W e c a lcu la te d 9 5 % co n fid e n ce in te rv a ls ( C I s ) . 23 RESULTS F ro m 1976 to 1988, w e do cu m ented 260 in cid e n t c ase s o f h isto lo g ica lly con firm ed o v a ria n can cer; th e se c a se s a re in clu d e d in th e h y ste re c to m y a n a ly se s. A su b se t o f 157 cases occurred am on g 77 544 w om e n w ho w ere p rem enop ausal in 1976 a n d a re in clu d e d in th e tu b a l liga tio n a n a ly se s. A t b a se lin e in 1976, 1 0818 (1 4 % ) o f 77 544 p re m e n op au sal w om en reported h a v in g h ad a tu b a l liga tio n ; fro m 1976 to 1988, 2 0 % o f th e p e rso n -y e a rs w e re a m o n g w o m e n w h o re p o rte d a tu b a l li g a tio n . A s s h o w n in T a b le 1, c o m p a r e d w ith w o m e n w h o h a d n o t h ad th is p ro cedure, th o se w h o h ad a tu b a l ligatio n w ere o f h ig h e r p a rity an d w e re m ore lik e ly to h a v e u se d o ra l co n trace p tiv e s, e sp e cia lly fo r lo n ge r d u ra tio n s. A su b sta n tia l re d u ctio n in risk o f o v a rian c a n c e r w a s n o te d a m o n g w o m e n w ho re p o rte d h a v in g a tu b a l liga tio n (a g e -a d ju ste d R R , 0.29; 9 5 % C l, 0.15 to 0.55) (T a b le 2). R e s u lt s w e r e e s se n tia lly u n chan ge d a fte r fu rth e r a d ju stm e n t for p arity , d u ra tio n o f o ra l c o n tra c e p tiv e u s e , s m o k i n g , Q u e t e l e t 's i n d e x , a g e a t m enarche, age at natu ral m enopause, and h y ste r e c to m y s ta tu s ( R R , 0.33; 9 5 % C l, 0.16 to 0.64). C o n tr o l fo r p a r ity in sin gle b irth c ate go rie s (u p to six o r m o re ) also d id n o t a lte r th e re su its. W h e n a n a ly se s w e re re p e ate d a m o n g o n ly w o m e n re p o rtin g cu rre n t co n trace p tiv e use, re su lts w ere sim ila r (age -ad ju ste d R R , 0.30). A m o n g p a ro u s w o m e n only, the a ge -ad ju ste d R R for th ose w ho re p o rte d h a v in g a tu b a l lig a tio n w a s also d e c re a se d (p a r ity = 1 o r 2: R R , 0.28 [9 5 % C l, 0.09 to 0.86]; p a r ity = 3 o r m o re : R R , 0 .35 [9 5 % C l, 0.11 to 0.68]). S im i larly, a m o n g w o m e n u n d e r g o in g tu b a l liga tio n w h o n e v e r u se d o ra l co n trac e p tive s, th e R R w a s 0.30 (9 5 % C l, 0.14 to 0.67). B e ca u se w o m e n m a y h av e had th e ir o varie s scree n e d a t the tim e o f th e ste r ilization p ro c e d u re (an d th u s w o u ld h av e the p ro c e d u re p e rfo rm e d o n ly if n o m a Table 2.--Relative Rrsk (RR) of Ovarian Cancer by Tubal Ligation Status and Use of Other Contraceptives Tubal ligation: risk period 1976-1988 No Yes Tubal ligation in 1976: risk period 1980-1988 No Yes Tubal ligation in non-talc users: risk period 1982-1988 No Yes Vasectomy: risk period 1976-19881 No Yes Condom use: risk period 1976-1988t No Yes Cases Psrson-Yssrs Ags-Ad|ustad RR (95% Confidence In te rv a l) M ultivariate RR (95% Confidence In te rv a l)* 148 690 508 1.0 1.0 9 169 283 0.29 (0.15-0.55) 0.33 (0.16-0.64) 116 477 772 1.0 1.0 4 77 481 0.23 (0.09-0.58) 0.26 (0.10-0.70) 55 211 323 1.0 1.0 3 58 482 0 22 (0.07-0.66) 0.26 (0.08-0.85) 113 560 621 37 179150 1.0 1.11 (0.76-1.61) 1.0 1.25 (0.85-1.84) 129 614 340 1.0 1.0 21 125 431 0.78(0.49-1.23) 0.76(0.48-1.21) "Controlling for age (in 5-year categories); parity ( 0 .1 .2 ,3 . >4); duration of oral contraceptive use (never, current use. duration of past use, mo: 1 to 1 1 ,1 2 to 35, 36 to 59, 2:60), age at menarcbe. y ( < 1 2 .1 2 . a 13); age at natural menopause, y (premenopausal. < 5 0 , 50 to 54, 5 5 ), hysterectomy (yes/no); smoking status (never, current, past); and Quetetet's index in kg/m2 (< 2 1 . 21 to < 2 3 , 23 to < 2 5 , 25 to < 2 9 , 2 9 ). tAssessed among women without tubal ligation in 1976. lign a n t o r p re m a lign a n t c o n d itio n w a s foun d), w e w ish e d to d e te rm in e w h e th e r d e te ctio n b ia s m ig h t acco u n t fo r th e in ve rse a sso cia tio n . W e w e re u n a b le to a sse ss o v a ria n c a n c e r r is k b y y e a rs sin ce tu b al liga tio n because, for w o m e n w ho re p o rte d a tu b a l liga tio n on th e 1976 q u e stio n n aire , w e d id n o t k n o w w hen the p ro ce d u re h ad been perform ed. T h e re fo re , w e a sse sse d th e a sso c ia tio n betw een tu b a l liga tio n sta tu s a s rep orted in 1976 a n d o v a ria n c a n c e r r isk fro m 1980 to 1988, th u s a llo w in g a m in im u m o f a 4 -y e a r la g fro m th e tim e o f th e p ro ce du re to th e b e g in n in g o f th e p e rio d at risk . O f note, w o m e n w h o re p o rte d h a v in g a tu b al liga tio n b etw e e n 1976 and 1978 had the procedure, on ave rage, 8 y e a rs a fte r th e ir la st p re gn a n c y . W o m e n re p o rtin g a tu b a l liga tio n a s o f 1976 had th e ir la st p re g n a n c y an a v e ra g e o f 10 y e ars before; thus, the actu al ave rage la g w a s p ro b a b ly a b o u t 6 y e a rs ([10 y e a rs - 8 y e a r s] + 4 -y e a r la g). T h e a g e -a d ju ste d R R w a s 0.23 (9 5 % C l, 0.09 to 0.58); m u ltiv a r ia te r e s u lts w e re sim ila r (T a b le 2). In a d d itio n , w e re v ie w e d all m e dical re c o rd s o f o v a ria n can ce r case s to d e te rm in e h o w m a n y w e re de tected a t tu b a l ligatio n . O n ly 51 o f the 157 w om en w ith o v a ria n can ce r in th e tu bal liga tio n a n a ly sis w e re 46 y e a rs o f a ge or y o u n ge r a t d ia gn o sis (an d p re su m a b ly at risk o f tu b a l liga tio n ), a n d o f th e se , 38 re co rd s in d icate d h o w th e can ce r had been detected. O v a ria n can cer w a s de tected a t tu b a l lig a tio n in ju s t on e case. T h e o b se rv a tio n th a t o v a ria n can cer in cid e n ce is re d u c e d a m o n g b o th o ra l c o n t r a c e p t i v e u s e r s 24 a n d w o m e n w i t h a tu b al liga tio n su g g e ste d th at contracepto rs in g e n e ra l m a y be a t lo w e r risk . W e th e re fo re a sse sse d th e re la tio n sh ip be tw e e n se v e ra l o th e r co n trace p tiv e m e th ods and o va ria n cancer, u sin g the sam e b ase lin e p o p u la tio n a s in th e tu b a l lig a tio n a n a ly se s, b u t e x c lu d in g th o se w h o re p o rte d a tu b a l liga tio n in 1976. N e i th e r v a se c to m y in a p a r tn e r n o r co n d o m u se (the tw o m o st co m m o n contracep tiv e m e th o d s a fte r o ra l c o n trace p tiv e u se an d tu b a l liga tio n ) w a s sig n ific a n tly a sso cia te d w ith r is k o f o v a ria n can ce r (T a b le 2). In all, 1 1 0 1 7 (1 0 % ) o f w o m e n re p o rte d h a v in g h ad a h y ste re c to m y in 1976; fro m 1976 to 1988, 1 5 % o f th e p e rso n -ye ars w ere am on g w om en w ho reported a h ys terectom y. W o m e n w ho had a h y ste r e cto m y w e re s lig h tly m o re lik e ly to be o f h igh e r p a rity an d to h av e use d oral co n trace p tiv e s fo r lo n g e r d u ra tio n th an w om en w h o d id n ot h av e a h yste re c t o m y (T a b le 1). T h e a ge -ad ju ste d R R o f o va ria n can ce r a sso cia te d w ith h y ste re c to m y w a s 0.66 (9 5 % C l, 0.45 to 0.99); th e m u lti variate R R w a s e sse n tia lly u n ch an ge d (T a b le 3). A m o n g p a r o u s w o m e n , th e age-ad ju sted R R for those w ho had a h yste recto m y w a s sim ila r (p a rity = 1 o r 2: R R , 0.61 [9 5 % C l, 0.31 to 1.21]; p a r ity = 3 o r m o re : R R , 0.77 [9 5 % C l, 0.45 to 1.31]). T h e R R o f o v a r ia n c a n c e r a s s o ciate d w ith h y ste re c to m y w a s so m e w h a t lo w e r a m o n g w o m e n w h o e v e r u se d o ral c o n tra c e p tiv e s ( R R , 0.45; 9 5 % C l, 0.20 to 1.01) th a n a m o n g n e v e r u s e r s ( R R , 0.77; 9 5 % C l, 0.48 to 1.22), b u t th e C Is w e re w id e. W e a lso a sse sse d r is k b y tim e sin ce h y ste re c to m y to d e te rm in e if th e in v e rse a sso c ia tio n w a s m a in ta in e d o v e r tim e. W e o b se rv e d a sta tistic a lly sign ifican t in v e rse a sso c ia tio n in th e 5to 9 -y e a r c a te g o ry a n d a n o n sign ific an t d e cre a se in r isk 15 o r m o re y e a rs a fte r JAMA, December 15, 1993-- Voi 270. No. 23 Tubal Ligation. Hysterectomy, and Ovarian Cancer--Hankinson et al 2815 the p ro ce du re , a lth o u gh th e n u m b e r of c ase s in e ach c a te g o ry w a s sm all. M u l tiv a ria te e stim a te s w e re sim ila r. F in a lly , w e a sse sse d th e a sso cia tio n b etw e e n h y s te re c to m y an d o v a ria n can ce r a cc o rd in g to a ge a t w h ich the p ro ce d u re w a s p er form ed. W o m e n w h o h ad a h yste rec to m y before 45 y e ars o f age w ere at a so m e w h at, but not sta tistica lly sig n ifi cant, lo w e r risk th a n w o m e n w h o h ad the p rocedu re at o r afte r 45 y e ars of age. T u b a l liga tio n an d h y ste re c to m y h ave been h y p o th e siz e d to d e cre ase o v a ria n can ce r risk b y p re v e n tin g v a g in a lly in tro d u c e d talc, a p o ssib le r is k fa c to r fo r o v a ria n cancer, fro m re a c h in g the o va r i e s . 25 W e d i d n o t f i n d a s t a t i s t i c a l l y s i g n ifica n t a sso c ia tio n b e tw e e n ta lc u se an d o va ria n cancer, h o w e ve r, because the q u e stio n w a s n o t a sk e d u n til 1982, th ere w ere re la tiv e ly few case s and th e C Is w e re w id e . T h e re fo re , w e a sse sse d the re la tio n sh ip o f tu b a l liga tio n an d h y s te re c to m y w ith o v a ria n can ce r a m o n g w om en w ho reported th at they never u se d ta lc on th e ir p e rin e u m o r on sa n i ta ry p ad s. T h e in v e rse a sso cia tio n w ith tu b a l lig a tio n w a s u n c h a n g e d ( R R , 0.26; 9 5 % C l, 0.08 to 0.85) (T a b le 2) a n d the a sso cia tio n w ith h y ste re c to m y w a s s o m e w h a t w e a k e n e d ( R R , 0.80; 9 5 % C l, 0.43 to 1.49). W e o b ta in e d m e d ical re c o rd s fo r 61 o f 100 w om en w h o h ad re p orte d a tu bal liga tio n (1 0 n o n re sp o n d e n ts, on e re fu sal, a n d 28 m e d ical re c o rd s u n a v a ila b le p ri m a rily b ecau se th e p ro ce d u re h ad been p erform ed so lo n g ago). T h e typ e s o f p ro c e d u re s w e re d istrib u te d a s fo llo w s: 33 liga tio n s (in c lu d e s P o m e ro y an d Ir v in g p ro ce d u re s, 5 4 % o f to tal), 19 c o a g u la tio n s)fu lg u ra tio n s (3 1 % ), se v e n sa l p in ge c to m ie s (1 1 % ), o n e b a n d in g (2 % ), an d one fim b rie cto m y (2 % ). COMMENT In th is la rg e p ro sp e c tiv e stu d y , w e foun d a stro n g in v e rse a sso c ia tio n be tw een tu b a l liga tio n an d o va ria n cancer. T h e a sso cia tio n re m a in e d sign ific a n t af te r c o n tro llin g fo r a n u m b e r o f p o te n tia l o v a ria n c a n c e r r isk fa c to rs an d did not a p p e a r to re su lt sim p ly fro m scre e n in g the o v a rie s a t th e tim e o f th e p rocedu re. W e a lso n o te d a n in v e rse , y e t w eake r, Table 3.--Relative Risk (RR) of Ovarian Cancer From 1976 Through 1988 by Hysterectomy Status Hysterectomy! No Yes Time since hysterectomy, y 1-4 5-9 10-14 >15 Age at hysterectomy, y <45 45 Cases 192 28 5 3 11 9 6 22 Person-Years 855 827 144 879 28 555 42 136 37 708 36 479 64 141 80 739 Age-Ad|uated RR (95% Confidence In te rv a l) 1.0 0.66 (0.45-0.99) 0.96 (0.39-2.33) 0.29 (0.10-0.85) 0.92(0.50-1 70) 0.62 (0.31-1.21) 0.51 (0.23-1.15) 0.72 (0.46-1.13) M ultivariate RR* (95% Confidence In te rv a l) 1.0 0.67(0.45-1.00) 1.03 (0.42-2.51) 0.29(0.10-0.88) 0.92 (0.50-1.69) 0.62 (0.31-1.22) 0.48 (0.21-1.08) 0.76 (0.48-1.19) Controlling for age (in 5-year categones); parity (0 ,1 .2 .3 , 24); duration of oral contraceptive use (never, current use. duration of past use. mo: t to f 1, 12 to 35. 36 to 59. 260); age at menarcrie. y (<12. 12. a 13): tubal ligation status (yes/no); smoking status (never, current, past): and Quetelefs index in kg/m* (<21. 21 to <23. 23 to <25. 25 to <29. 229). tD ata on hysterectomy status was missing tor 40 cases and 190 817 person-years. a sso c ia tio n b e tw e e n h y ste re c to m y an d o varian cancer. O u r stu d y h as se ve ral stre n gth s. Its p ro sp e ctiv e d e sig n e lim in a te s th e po te n tia l fo r se le ctio n b ias. A d d itio n a lly , b e cau se e x p o su re in fo rm a tio n is co l le cte d p rio r to d ise a se d ia g n o sis, re c a ll b ia s is e lim in ate d . W e a lso acco u n te d fo r m o st k n o w n o r su sp e cte d o varian can ce r risk fa c to rs in o u r a n a ly sis. T h e stu d y a lso h a s se v e ra l p o te n tia l lim itatio n s. S o m e p re m e n o p a u sa l w o m e n m a y not h av e re p o rte d th e ir tu b a l lig a tio n b ecau se th e y w e re n o t se x u a lly ac tiv e a t th e tim e th e q u e stio n n aire w as co m ple te d, b u t w h e n a n a ly se s w e re re peated a m o n g w om e n re p o rtin g cu rren t co n trace p tiv e u se , re su lts w e re u n changed. Because w e aske d about contracep tive u se o n ly u n til 1982, a n u m b e r of w o m e n in o u r u n e x p o se d c a te g o ry p ro b a b ly su b se q u e n tly h ad a tu b a l ligatio n . H ow e ve r, b y 1982 the n u m b e r o f new re p o rts o f tu b a l liga tio n w a s ra p id ly d e clin in g, so th a t th e sh ift in p e rso n -y e a rs fro m n o n e x p o se d to e x p o se d w o u ld be q u ite sm all a n d b e ca u se th e se m isc la ssife d w o m e n te n d to be b o th o f h ig h e r p a rity a n d o ra l c o n trac e p tiv e u se rs (T a b le 1), t h is e r r o r w o u ld te n d to b ia s o u r R R t o w a r d 1.0. W e w e re u n a b le to a sse ss fu lly o v a ria n can ce r r is k b y tim e sin ce tu b a l li g a tio n to d e te rm in e w h e th e r th e p ro te ctive e ffe ct w a n e d o v e r tim e . H o w ever, w h e n w e a sse sse d o va ria n can cer risk fro m 1980 to 1988 b y tu b a l ligatio n sta tu s in 1976, th e R R w a s u n ch an ge d . A lso , a m o n g c ase s 46 y e a rs o f a ge o r y o u n g e r a t d ia gn o sis, w e fo u n d th a t o n ly one case h ad b een dete cte d a t tu b a l li gatio n . S o m e w o m e n m a y h a v e h ad a m a lign a n c y d e te cte d d u rin g a tu b a l li ga tio n p e rfo rm e d b efore 1976 w h o m igh t o th e rw ise h av e b ee n d ia gn o se d d u rin g o u r stu d y p e rio d ; h o w e v e r, g iv e n the a b o v e re su lts, th is b ia s is a lso lik e ly to Table 4.--Studies of Tubal Ligation and Ovanan Cancer and Summary of Relative Risks (RRs) Source, y Koch et a l.13 1984 Koch et a l," 1988 Mori et al.6 1988 Boom et a l* 1989 Shu et at.101989 Whittemore et al,'2 1992 Hospital controls Population controls Hankinson et al. 1993 Summary RRl| Design Retrospective cohort Case-control Case-control Case-control Case-control Prospective cohort S iz e * 4/22 095 200/211 110/220 213/420 172/172 517/1970 766/4089 157/859 791 Covarfatsa Controlled Age, parity None Age, parity, marital status, number ot induced abortions Age. social dass. gravidity, unprotected intercourse Age. education, parity, age at menarche. ovanan cyst Age, study, parity, oral contraceptive use Age, study, parity, oral contraceptive use Age. parity, oral contraceptive use, age at menarche and menopause, smoking, hysterectomy. Quetelefs index RR (95% Confidence Interval) 2.4 (0.9-6.7)f 0.8 (0.5-1.3 )| 0.5 (0.25-1.0 )t 0.2 (0.1-0.6) 08(0.4-1.8) 0.59 (0.38-0.93) 0.87(0.62-1.2) 0.33(0.16-0.64) 0.63 (0.44-0.91) Cases/controts for case-control studies: cases/person-years for cohort studies. (Confidence intervals calculated from P value provided in article. (Crude relative nsk and 95% confidence intervals calculated using Cornfield's method from data provided in abide. Pooled analysis of eight studies: lour with hospital controls and four with population controls. Previously published findings from Whtttemore et a l' (1988) and Irwin et al4 (1991) were included in me pooled estimate. Calculated using a random effects model (see text); estimate indudes all seven previous studies plus me current study. 2816 JAMA. December 15. 1993--Voi 270. No. 23 Tubal Ligation. Hysterectomy, and Ovarian Cancer-- Hankmson et al be m odest. F in a lly , it is p o ssib le th a t w e failed to c o n tro l fo r o th e r o v a r ia n c a n c e r risk facto rs (eg, fa m ily h isto ry o f o v a rian can ce r) th a t m ig h t h a v e d isto rte d ou r fin d in gs. H o w e v e r, w e c o n tro lle d for m o st k n o w n risk fac to rs a n d th e a s so ciatio n w ith tu b a l lig a tio n w a s stro n g , th us m in im iz in g th is p o ssib ility . B e ca u se th e fallo p ia n tu b e s a re lig a te d or occlud ed a t b o th tu b a l lig a tio n a n d hysterectom y, sim ila r e ffects on o va ria n can cer r isk m ig h t b e e x p e cte d ; y e t, in our d ata the re d u ctio n in risk a p p e a re d gre a te r fo r w o m e n h a v in g a tu b a l lig a tion. T o d e te rm in e w h e th e r th is e ffe c t w as due to the d iffe re n t a g e s w h e n th e se tw o procedu res are u su a lly p e rfo rm e d , we assessed cancer risk b y age at h y s terectom y. O u r d a ta su g g e ste d a slig h tly la rge r p ro te ctive e ffe ct in th e w o m e n w ho w ere y o u n ge r at th e tim e o f h y s te re cto m y a lth o u g h o n ly six c a se s h ad su rge ry before 46 y e a rs o f age. S im ila r re su lts w e re re p o rte d in a p o o le d a n a ly s i s o f 1 2 c a s e - c o n t r o l s t u d i e s . 12 T h e e f fects o f tu b a l liga tio n a n d h y ste re c to m y on risk o f o va ria n can ce r a lso m a y v a ry ; even th e te ch n iq u e s u se d fo r p e rfo rm in g tu b al ste riliz atio n s (eg, c o a g u la tio n v s u se o f a b a n d o r r in g ) m a y v a r y in th e ir su b se q u e n t e ffe ct o n o v a ria n fu n c tion, p e rh a p s b e c a u se o f d iffe r in g d e gre e s o f tissu e d e stru ctio n .4 S e v e ra l m e c h a n ism s o f o v a ria n can cer e tio lo gy h av e b ee n p ro p o se d . F a th a lla 26 s u g g e s t e d t h a t `i n c e s s a n t o v u l a tion" in c re a se s th e r is k o f o v a ria n c a n cer p e rh ap s b y e x p o sin g th e e p ith e liu m to ste ro id -ric h fo llic u la r flu id o r b y s u b je ctin g it to in cre a se d c e llu la r p ro life ra t io n .27 I n s o m e , 2823 a l t h o u g h n o t a l l , 30 s t u d ies, tu b a l lig a tio n w a s a s s o c ia te d w ith an increased fre q u e n cy o f a b n o rm a l m e n stru al cycle s a n d th e re fo re th e p ro c e dure m ig h t also d e cre a se th e n u m b e r o f o vu latio n s. A lth o u g h w e w e re u n a b le to ad d re ss th is issu e sp e cifica lly , a g e at natural m enopause does not v a ry b y tu bal liga tio n sta tu s in o u r d ata. T u b al ligatio n a n d h y ste re c to m y m ig h t also se rv e to p re v e n t ta lc e x p o su re , a h ypo th e sized o v a ria n c a n c e r r is k fac to r .252112 H o w e v e r , w e f o u n d t h a t t u b a l ligatio n w a s a lso h ig h ly p ro te c tiv e in w o m e n w h o re p o rte d n e v e r u s in g talc. In con trast, the a sso cia tio n b e tw e e n h y s terectom y an d o varian can cer w a s som e w hat w e a k e r w h e n a sse sse d in n o n -ta lc users, a lth o u gh th e C Is o ve rlap p e d . If oth er c o n ta m in a n ts a p a rt fro m talc in fluence ris k o f o v a ria n c an c e r, tu b a l li gatio n a n d h y ste re c to m y m ig h t se rv e to prevent such exposures. H ig h circ u la tin g le v e ls o f p itu ita ry g o n ad o tro p in s also h a v e b ee n h y p o th e siz e d to p l a y a r o l e i n o v a r i a n c a n c e r . 33 H i g h go n a d o tro p in le v e ls h a v e b ee n a sso c i ated w ith stro m a l o v a ria n tu m o rs in a n i m al stu d ie s3 an d go n a d o tro p in s stim u la te th e g ro w th o f ce ll lin e s d e riv e d fro m h u m a n o v a r i a n c a n c e r s . 34 I f t u b a l l i g a t i o n i n f l u e n c e d o v a r i a n c i r c u l a t i o n , 35 p la sm a h o rm o n e le ve ls a n d o v a ria n fu n c tio n m ig h t be affected. L o w e r p re o v u la to ry lu te in iz in g h o rm o n e le v e ls h av e been re p o rte d a m o n g w o m e n w ith tu b al l i g a t i o n i n o n e s t u d y , 36 b u t n o t i n a s e c o n d s t u d y . 37 P l a s m a g o n a d o t r o p i n l e v e l s w ere un changed w hen asse sse d before a n d a f t e r t u b a l s t e r i l i z a t i o n 38; h o w e v e r , th e b lo o d sa m p le s w e re d r a w n o n ly 3 m on th s after the procedure an d th e as s a y co e fficien t o f v a ria tio n w a s 2 5 % , th u s th e a b ility to de te ct lo n g -te rm h o rm o n a l c h a n g e s w a s lim ite d . L o w e r e stro g e n 352629 a n d p r o g e s t e r o n e l e v e l s 37'40-41 h av e also been note d in w o m e n w h o had a tu b a l ste riliz atio n re la tiv e to w o m e n w h o w e re n o t ste riliz e d , a lth o u g h o th e r s h a v e n o t e d n o a s s o c i a t i o n . 3629-42-43 I n n on e o f th e se stu d ie s w e re o th e r p o s sib le d e te rm in a n ts o f h o rm o n e le ve ls, su ch a s w e ig h t o r d ie ta ry in take , co n sid e re d in the a n a ly sis. T h e d a ta a sse ssin g p o ssib le h o rm o n al chan ges after h ysterectom y are even m ore sparse. E stra d io l an d p ro g e ste r one le v e ls w e re d e c re a se d 3 to 4 w e e k s a f t e r h y s t e r e c t o m y 44 a n d a r e d u c e d o v a ria n b lo o d flo w w a s n o te d im m e d ia te ly a f t e r h y s t e r e c t o m y . 45 H o w e v e r , h y s t e r e cto m y d o e s n o t re su lt in im m e d ia te ce s sa tio n o f o v u la tio n a s e v id e n c e d b y o v u l a t o r y i n c r e a s e s i n p r o g e s t e r o n e . 46 In p re v io u s stu d ie s, in c lu d in g a re ce n t p o o le d a n a ly sis o f c a se -c o n tro l s tu d ie s , e it h e r a s t a t is t i c a ll y s ig n i f i c a n t 912 o r a n o n s i g n i f i c a n t 61011-12 i n v e r s e a s s o c i a tio n b etw e e n tu b al liga tio n a n d o v a ria n c a n c e r h a s b ee n re p o rte d (T a b le 4). In a re tro sp e ctiv e coh ort stu d y , a n o n sig n ifi can t in cre ase d risk o f o v a ria n can ce r w a s found; how ever, th e stu d y w as sm all a n d c o n t r o l f o r p a r i t y w a s l i m i t e d . 13 T o su m m a riz e th e se fin d in g s, w e calcu la te d a w e igh te d a v e ra ge o f th e R R u sin g a ra n d o m effects m od el th a t a llo w s for h e t e r o g e n e i t y a c r o s s s t u d i e s . 47 O v e r a l l , a m o n g p a st stu d ie s, a sig n ific a n t 3 1 % re d u c tio n in o v a ria n c a n c e r r is k w a s noted and, w h e n in c lu d in g th e cu rre n t stu d y , th e re d u ctio n is 3 7 % . If th e in v e rse a sso c ia tio n b e tw e e n tu b al ligatio n an d o v a ria n c an c e r sim p ly re su lte d from scre e n in g th e o v a rie s at th e tim e o f th e p ro ce du re , th e g re a te st re d u c tio n in r isk w o u ld b e e x p e c te d in the first se ve ral y e ars a fte r th e p ro ce du re . W h itte m o re et a l7 re p o rte d a g re a te r d e cre a se in R R a m o n g w o m e n w h o h ad a tu bal liga tio n w ith in th e p re vio u s 10 y e a rs than a m o n g w o m e n w h o had the procedure m ore th an 10 ye ars before, a lth o u g h th e se d iffe re n ce s w e re n o t sta tistica lly sign ific a n t. M o re re ce n tly, th e in v e rse a sso c ia tio n b e tw e e n tu b a l liga tio n an d o v a ria n can ce r w a s p re se n t a t le a st 15 y e a rs a fte r th e p ro ce du re .8 O u r d a ta a lso su g g e s t th a t th e in v e rse a sso cia tio n is u n lik e ly to be sim p ly a n effect o f scre e n in g. In a po o le d a n a ly sis, h y ste re c to m y w a s asso cia te d w ith e ith e r a sta tistic a lly s ig n ific a n t ( R R , 0.69 fo r s ix h o sp ita l-b a se d stu d ie s) o r n o n sig n ific a n t ( R R , 0.88 fo r six stu d ie s w ith p o p u la tio n c o n tro ls) re d u c t i o n i n o v a r i a n c a n c e r r i s k . 12 A l t h o u g h th e C Is w e re w id e , th e r is k a p p e a re d lo w e st a m o n g w o m e n h a v in g a h y ste r e c to m y e a rly in life, s im ila r to o u r fin d in g s. In a d d itio n , th e re d u c e d r is k a p p e a re d to b e p re se n t a t le a st 10 to 19 y e a rs a fte r h yste recto m y; o u r fin d in g s a lso ten d to su p p o rt a lo n ge r-te rm p ro te c tiv e effect. O u r d a ta in dicate a stro n g in v e rse a s so ciatio n b etw e e n tu b a l liga tio n a n d su b se q u e n t risk o f o va ria n cancer. A m o d e st p ro te ctive effect o f h y ste re c to m y on o v a ria n can ce r risk is also su g g e ste d . It m a y be a p p ro p ria te to c o n sid e r th e re du ced risk o f o varian can cer a fte r tu b al liga tio n , n o w se en in se v e ra l stu d ie s, w h e n ch o o sin g a m o n g a lte rn a tiv e m e th o d s o f co n trace p tio n . F u r th e r re se a rc h is n e e d e d to d e fin e b io lo g ic m e c h a n ism s an d to a sse ss effects o f d iffe re n t tu b al liga tio n p rocedu res. This study was supported by research g ra n t C A 40366 from the National Institutes o f H ealth. D r Hankinson was supported in p art by Research Service A w ard 5T32CA09001 from the National In s titu te s o f H e alth . D r C olditz is supported in p a rt by Am erican Cancer Society Faculty Research A w ard FR A -398. W e are indebted to th e regis tered nurses in the study fo r th e ir continuing cooperation and to M ark Schneyder, B arb ara Egan, G ary Chase, Maureen Ireland, Lisa Dunn, L o ri Egan. Karen Corsano. M A , and K ate Saun ders fo r th e ir ex p e rt and unfailing assistance. References 1. Cancer Statistics Review 973-87. B ethesda.M d : U S D ept of H ealth and Human Services, National In s titu te s o f Health; 1990:128. N a tio n al In s titu te s of H ealth publication 90-2789. 2. Schottenfeld D, Fraum eni J. Cancer E pidem i ology and Prevention. Philadelphia, Pa: W B Saun ders Co; 1982:871-880. 3. Moses V I, Rubin G L , Layde P M , Hughes JM . Tubal sterilization among women of reproductive age, United States, update for 1979-1980. MMVfR CD C S u rveill Summ. 1983;32.'9SS-14SS. 4. Huggins G R , Sondheimer SJ. Complications of female sterilization: immediate and delayed. Fertil Steril. 1984;41:337-355. 5. Bachman G A. Hysterectom y: a critical review . J Reprod Med. 1990;35:839-862. 6. M o riM , Harabuchi I, M iyake H , C asagrand eJT. Henderson B E , Ross R K . R eproductive, genetic, and dietary factors for ovarian cancer. A m J E pi demiol. 1988;128:771-777. 7. W hittem ore A S . W u M L , Paffenbarger RS, et al. Personal and environm ental characteristics re lated to epithelial ovarian cancer, I I: exposures to talcum pow der, tobacco, alcohol, and coffee. A m J Epidemiol. 1988;128:1228-1240. 8. Irw in K L , W eiss N S , L e e N C , P eterson H B . Tubal sterilization, hysterectomy and the subse quent occurrence of epithelial ovarian cancer. A m J E pidem iol. 1991;134:362-369. 9. Booth M , B eral V, S m ith P. R isk factors fo r JAMA December 15, 1903--Vol 270. No. 23 Tubal Ligation. Hysterectomy, and Ovarian Cancer--Hankinson et al 2817 ovarian cancer a case-control study. B r J Cancer. 198930:592-098. 10. Shu X O . B rinton L A , Gao Y T , Y u a n J M . Popu lation-based case-control stu d y o f ova rian cancer in Shangai. Cancer Res. 1989;49:3670-3674. 11. Koch M . Jenkins H , G aedke H . R isk factors of ovarian cancer of epithelial origin: a case-control study. Cancer Detect Prev. 1988:13:131-136. 12. W h ittem o re A S . H a rris R , Itn y r e J, and the Collaborative Ovarian Cancer Group. Characteris tics re la tin g to ovarian cancer risk: co llaborative analysis of 12 U S case-control studies, I I : invasive epithelial ovarian cancers in w hite women. A m J Epidemiol. 1992:136:1184-1203. 13. Koch M . S tarreveld A A , H ill G B , Jenkins H . The effect of tubal ligation on th e incidence of epi thelial cancer of the ovary. Cancer Detect Prev. 1984:7:241-245. 14. W y n d e r E L . Epidem iology o f cancer o f the ovary. Cancer. 1969-23:352-370. 15. A nnegers J F . S trom H , D e c k e r D G , D o c k e rty M B , O 'Fallon W M . O varian c a n c e r incidence and case-control study. Cancer. 1979:43:723-729. 16. M cGowan L . P a re n t L , L e d n e r W , N o rris HJ. The woman a t risk for developing ovarian cancer. Gynecol Oncol. 1979:7:325-344. 17. H a rtg e P, H o o ver R , M cG o w an L , L e s h e r L , Norris HJ. Menopause and ovarian cancer. Am J Epidemiol. 1988:127:990-998. 18. W eiss N S . H a rlo w B L . W h y does hysterec tomy w ithout bilateral oophorectomy influence the subsequent incidence o f ova rian cancer? A m J E p i demiol. 1986:124:856-858. 19. C olditz G A . S tam p fer M J, W ille tt W C . e t al. Reproducibility and validity of self-reported meno pausal status in a prospective cohort o f w om en. A m J Epidemiol. 1987:126:319-325. 20. Stam pfer MJ. W ille tt W C , Sp eizer F E , et al. Test of the National Death Index. A m J Epidemiol. 1984;119:837-839. 21. Cox D R . Regression models and life-tables. J R S tat Soc. 197222:187-220. 22. D 'Agostino RB . Lee M L , B elan ger A J.C upples L A , Anderson K. Knnel W B . R e latio n o f pooled logistic regression to tim e dependent Cox regres sion analysis: the F ram ingham H e a r t S tudy. Stat Med. 19903:1501-1515. 23. M iettin e n O. Estim abiiity and estim ation in casereferent studies. A m J Epidemiol. 1976:103226-235. 24. Hankinson SE. Colditz G A . H u n te r DJ, Spen cer T L , Rosner B. Stampfer MJ. A quantitative assessment of oral contraceptive use and risk of ovarian cancer. Obstet Gynecol. 199230:708-714. 25. C ram er DW . Welsh W R . Scully R E . W oj ciechowski C A . O varian cancer and talc: a casecontrol study. Cancer. 1982:50-372-376. 26. F a th a lla M F . Incessant ovulatio n-- a fa c to r in ovarian neoplasia? Lancet. 19712:163. 27. Stein K F , Allen E . A ttem pts to stim ulate pro liferation of the germinal epithelium of the ovary. A nal Rec. 194932:1-9. 28. DeStefano F , Perlm an J A , Peterson H B , Dia mond E L . Long-term risk of m enstrual disturbances a fte r tubal sterilization. A m J Obstet Gynecol. 1985; 152:835-841. 29. Rulin M C . Davidson A R , P h illib e r SG, Graves W L , Cushman L F . Changes in m enstrual symp toms among sterilized and comparison women: a prospective study. Obstet Gynecol. 1989:74:149-154. 30. DeStefano F. Huezo C M , Peterson H B , Rubin G L , Layde PM . O ry H W . M enstrual changes after tubal sterilization. Obstet Gynecol. 1983,62.-673-681. 31. E g li G E , Newton M . The transport of carbon particles in the human female reprodu ctive tract. Fertil S te ril 1961:12:151-155. 32. Henderson WJ, Ham ilton T C , G riffiths K . Talc in norma) and m alignant ovarian tissue. Lancet. 1979;I:499. 33. C ram er D W , Welch W R . D eterm inants of ova rian cancer risk, II: inferences regarding patho genesis. J S a il Cancer Inst. 1983;71:717-721. 34. Simon W E . A lbrecht M , Hansel M , D ietel M . Holzel F . Cell lines derived from human ovarian carcinomas: grow th stim ulation by gonadotropic and steroid hormones. J N atl C ancer Inst. 1983; 70:839-845. 35. Cattanach J. Oestrogen deficiency a fte r tubal ligation. Lancet. 1986:1:847-849. 36. Alvarez-Sanchez F, Segal SJ, Brache V , Adejuwon C A . Leon P, Faundes A. Pituitary-ovarian function afte r tubal ligation. Fertil Steril. 198126: 606609. 37. Radwanska E, Headley SK, Dm owski P. Evalu ation o f ovarian function afte r tubal sterilization. J Reprod Med. 198227:376684. 38. Sorensen T , Ladehoff P. Lindholm P, Q vist K. Follicular stimulating hormone, luteinizing hormone and estrogen levels before and a fte r female steril ization. Acta Obstet Gynecol Scand. 198160:559 -561. 39. Corson S L . Levinson CJ, B a tzer F R , O tis C. Horm onal levels following sterilization and hyster ectom y. J Reprod Med. 198 126:36366 9. 40. Donnez J, W auters M . Thomas K . Lu teal func tion a fte r tubal sterilization. Obstet Gynecol. 1981; 57:66-68. 41. Radwanska E . BergerG S. Hammond J. Luteal deficiency among women w ith norm al m enstrual cycles, requesting reversal o f tubal sterilization. Obstet Gynecol. 1979;54:189-192. 42. R ivera R, Gaitan JR , Ruiz R, et aL M enstrual patterns and progesterone circulating levels fol low ing d iffe re n t procedures o f tubal occlusion. Con traception. 1989;40:157-169. 43. H elm G, Sjoberg NO . Progesterone levels be fore and a fte r laparoscopic tubal ste riliza tio n using endotherm coagulation. Acta Obstet Gynecol Scand. 198362:63-66. 44. Stone SC. Dickey R P . Mickal A. The acute ef fect o f hysterectom y on ovarian function. A m J Obstet Gynecol. 1975:121:193-197. 45. Janson PO. Jansson I. The acute effect of hys te recto m y on ovarian blood flow. A m J Obstet G y necol. 1977:127:349652. 46. Doyle L L , Barclay D L . Duncan G W , K irto n H T . H um an luteal function following hysterectom y as assessed by plasma progestin. A m J Obstet G y necol. 1971;110:92-97. 47. DerSim onian R. Laird N . M eta-analysis in clini cal trials. Controlled Clin Trials. 1986;7:177-188. 2818 JAMA. December 15. 1996--Vol 270. No. 23 Tubal Ligalion. Hysterectomy, and Ovarian Cancer--Hankinson el al A m e r i c a n .Jo u r n a l o k E p i d e m i o l o g y V o l. loi). \ o , -, Copyright c 1989 by T h e Johns H o pkins Un iversity School oi Hygiene and Public H ealth A ll rights reserved P rim e d in U'.N.A. 1980-1985. Amoi 1 B rief Original Contributions an independent cent), 83 of 88 (9- A C A S E -C O N T R O L S T U D Y O F B O R D E R L IN E O V A R IA N T U M O R S : T H E IN F L U E N C E O F P E R IN E A L E X P O S U R E T O T A L C as borderline ov high degree of chose to include BERNARD L. HARLOW and NOEL S. WEISS whose tumors Through random Harlow, B. L. (Harvard Medical School, Brigham and Women's Hospital, Boston, a control group MA 02115), and N. S. Weiss. A case-control study of borderline ovarian tumors: similar to the ca? the influence of perineal exposure to talc. A m J E p id e m io l 1989;130:390-4. county of resid The authors interviewed 116 female residents of western Washington State with serous and mucinous borderline ovarian tumors diagnosed between 1980 and 1985 and questioned them on their use of hygienic powders. A sample of 158 control women from the same counties were identified through random digit dialing and were interviewed as well. Neither the perineal application of baby powder nor the perineal application of cornstarch was associated with an appre ciably altered risk of borderline ovarian tumors. However, women who used deodorizing powders alone or in combination with other talc-containing powders had 2.8 times the risk (95% confidence interval 1.1-11.7) of women who had not undergone bilate eluded from the ; the study metho> ( 10) . Reproductive, ries and informr to talc were obt. interview. An o; had perineal exposure to powder. These results suggest that future studies of women to specii ovarian tumors in relation to the application of talc-containing powders should brand name(s), c consider ascertaining the specific type(s) of powder used. perineal applica- ovarian neoplasms; talc itary napkins, a prior to diagno controls). Affirr. In light of the marked differences in agespecific incidence and patient survival be tween borderline and malignant epithelial ovarian tumors (1), we conducted a casecontrol study of borderline ovarian tumors to determine whether etiologic differences between these low-grade tumors and their malignant counterparts exist as well. As part of this study, we sought to investigate the possible etiologic role of perineal expo sure to talc. Interest in talc as a potential ovarian carcinogen has grown from reports of oc- Received for p ublication August 4. 1988. and in cupational asbestos exposure and ovarian cancer (2-4). Mineral talc, similar in chem ical composition to various asbestos min erals, is the common base for most dusting powders that women may apply to the per ineum, sanitary napkins, or diaphragms prior to storage (5). Presently, three epi demiologic studies have examined the as sociation between talc exposure and ovar ian cancer (6-8). M a teria ls and m e t h o d s The Seattle-Puget Sound Cancer Sur veillance System classifies borderline ovar ian tumors according to the World Health Organization International Classification of egorized either a containing powc izing powder, ar cum or "dusting' We were sue; views from 116 eligible) ana 15 those eligible). sponse rates ca: Since previous have reported cancer risk in rt tory and exoge: controlled for a oral contracepti means of stratit fin a l form February 28. 1989. From the Departm ent o f Epidemiology, School of Public Health and Com m unity Medicine. University o f Washington. Seattle. \VA . and the Division o f Pub Diseases for Oncology (ICD-O) (9). Female residents of three urban counties of western Washington State diagnosed as having a Women who i dusting powder lic Health Sciences, Fred H utchinson Cancer Re search Center. Seattle. \VA. Reprint requests to Dr. Bernard L. Harlow, Ob stetrics and Gynecology Epidemiology Center. H a r serous or mucinous borderline ovarian tu mor (ICD-0 codes 8.440-8.431) were iden tified from the files of this population- sanitary napkii age--had an adj developing a bo vard M edical School. Brigham and Women's Hospital. 221 Longwood Avenue. Boston. M A 02115. . The authors thank Dr. Daniel W. Cramer for his helpful advice and comments. based cancer reporting system. Included were white women aged 20-79 years whose tumors were diagnosed during the years per cent confie 1table 1). We f ciation accord! Voi. l'io. Nu. % fintee in L ' N..U I M O R S: r ton, lors: tate 980 e of digit aby . jresed lers not of uld i ovarian in chem;tos min t dusting >the per.phragms :iree epi1 the asnd ovar- :er Sur ne ovar1Health nation of Female western aving a rian tu re idenulationncluded 5 whose e years BORDERLINE OVARIAN TUMORS AND TALC EXPOSURE 391 1980-1985. Among those, tumors subject to an independent pathology review (73 per cent), 83 of 88 (94 per cent) were confirmed as borderline ovarian tumors. Given the high degree of histologic agreement, we chose to include the additional 33 cases whose tumors had not been reviewed. Through random digit dialing, we identified a control group of white women who were similar to the cases with respect to age and county of residence. Controls who had undergone bilateral oophorectomy were ex cluded from the analysis. Further details of the study methods are described elsewhere ( 10). Reproductive, sexual, and medical histo ries and information on perineal exposure to talc were obtained during an in-person interview. An open-ended question asked women to specify the type(s), but not the brand name(s), of powder they had used for perineal application after bathing, on san itary napkins, and for diaphragm storage prior to diagnosis (or a similar date for controls). Affirmative responses were cat egorized either as one or more of three talccontaining powders (baby powder, deodor izing powder, and other or unspecified tal cum or "dusting" powders) or as cornstarch. We were successful in obtaining inter views from 116 cases (68 per cent of those eligible) and 158 controls (74 per cent of those eligible). A detailed discussion of re sponse rates can be found elsewhere (10). Since previous studies (including ours) have reported an association of ovarian cancer risk in relation to reproductive his tory and exogenous female hormones, we controlled for age, parity, and the use of oral contraceptives during the analysis, by means of stratification (11). R esults Women who reported any perineal use of dusting powders--either after bathing, on sanitary napkins, or for diaphragm stor e s --had an adjusted relative risk of 1.1 for developing a borderline ovarian tumor 195 Per cent confidence interval (Cl) 0.7-2.1) Cable 1). We further examined this asso ciation according to both the specific method of exposure to dusting powders and the type of powder used. The analysis by method of use indicates that a smaller pro portion of cases than controls used talccontaining powder or cornstarch for dia phragm storage. The risk associated with the use of talc-containing powders or corn starch after bathing was 1.2 (95 per cent Cl 0.6-2.6). Women who reported any use of talc-containing powder or cornstarch on sanitary napkins had a risk about double (relative risk (RR) = 2.2. 95 per cent Cl 0.8-19.8) that of women who reported no talc use. This risk was the same for women who reported applying powder both after bathing and to sanitary napkins. No in crease in risk was present among shortand long-term diaphragm users, the risk was not modified by the use of cornstarch versus other talc-containing powders, and there was no variation in risk with increas ing number of days of use (not shown). When we compared cases and controls by the type of powder used, there was no excess risk of borderline tumors among women who applied cornstarch, baby pow der, or unspecified talcum powder alone or in combination to the perineum. However, women who applied deodorizing powders with or without baby powder (only baby powder was reported as a second powder in women who used deodorizing powders) had nearly three times the risk of developing a borderline ovarian tumor compared with women who reported no perineal use of powder (RR = 2.8, 95 per cent Cl 1.1-11.7). When we examined the type of powder used according to the method of applica tion, the excess risk due to the use of de odorizing powders was present regardless of whether it was applied after bathing or to sanitary napkins. No subjects- reported any use of deodorizing powders for dia phragm storage. D is c u s s io n Our results of perineal exposure to talc-- no association among women who applied talcum powder to diaphragms, but a modest increase in risk among women who applied 392 HARLOW AND WEISS T able 1 Perineal use of talc-containing powder and cornstarch among women with borderline ovarian tumors and their matched controls, by method of use and by type of powder used, western Washington State. 1980-1985 Cases (n = 116) Controls Crude Adjusted (n = 158) RR- RR+ 95fo Cl* No perineal exposure to powder 67 94 i.ot Any perineal exposure to powder 49 64 1.1 1.1 0.7-2.1 Method of use Diaphragm storage only Diaphragm storage only or with other methods After bathing only After bathing only or with other methods Sanitary napkins only Sanitary napkins only or with other methods After bathing and on sanitary napkins 8 11 24 34 7 14 7 21 0.5 27 0.6 30 1.1 37 1.3 4 2.5 10 2.0 4 2.5 0.5 0.2-1.4 0.5 0.2-1.3 1.2 0.6-2.6 1.3 0.8-2.7 2.2 0.8-19.8 1.9 0.9-6.9 2.2 0.9-19.8 Type of powder used Cornstarch only (no combined use) Baby powder only Baby powder only or combined use Talc, unspecified (no combined use) Deodorizing powder only Deodorizing powder only or combined use 4 18 22 13 10 14 7 0.8 31 0.8 34 0.9 19 1.0 4 3.5 7 2.8 0.8 0.2-3.8 0.8 0.4-1.9 0.9 0.5-2.0 1.0 0.4-2.4 3.5 1.2-28.7 2.8 1.1-11.7 Method and type of powder used Any powder use after bathing Any use of deodorizing powder No use of deodorizing powder Any powder use on sanitary napkins Any use of deodorizing powder No use of deodorizing powder 10 5 2.8 3.1 0.8-10.9 24 32 1.1 1.1 0.5-2.4 8 4 2.8 2.6 0.9-22.4 6 6 1.4 1.5 0.4-6.5 * RR, relative risk: Cl. confidence interval. + Adjusted for age (20-39, 40-59, or 60-79 years), parity (nulliparous or parous), and use of oral contraceptives (ever or never). t Reference group. talc-containing powders to the perineum or to sanitary napkins--are consistent with those previously reported in studies of ma lignant ovarian tumors. Cramer et al. (6) observed a 50 per cent excess risk among women who used dusting powders or who applied talc-containing powders to sanitary napkins, and a relative risk of 3.3 among women who applied both. No association was found with use of talcum powder for diaphragm storage. Hartge et al. (7) also found no excess risk among users of talc for diaphragm storage, but they did report an association with perineal application (seven cases, three controls; RR = 2.5, 95 per cent Cl 0.7-10.0). Whittemore et al. (8) reported a 40 per cent excess in risk of ovarian cancer associated with perineal ex posure only and a modest increase in risk with increasing numbers of applications per month. An association between talc use and ovarian neoplasms seems biologically plau sible. since particulates contaminating the vaginal area may migrate into the pelvic cavity and since particles of talc have been observed within ovarian tissue 12-15). It is also conceivable that the excess risk as sociated with appln neum and to san: seen in the three which inquired ab< 1 could have been du restricted to the l ders. The lack of t women who used t< diaphragms (both previous studies) ? since deodorizing used for diaphragdifferential asbest . different types of cannot be ruled manufactured co , were required to c mineral talc, but marked as cosmet . conform to theS' ! 1976, a study of : [ ders labeled as t ders, or body pov stores in New \ c 1975 reported th; tions of asbestii \ phyllite ranging * cent (4). A lth o u g h it is i of association an | powder exclusiv j uct labels baby I tain only talc V deodorizing sub from deodorizii ! and perfumed o t h e r h a n d , inc: odorizing subst l free and bondei ' asbestiform lib' W e suggest ' the results oi t 1 among women odorizing powc chance or appl: m alignant, ovi latter possibil: risk associate' containing po \n tumors and their ate. 1980-1985 ed 9571 c r 0.7-2.1 0.2- 1.4 0.2-1.3 0.6-2.6 0.8-2.7 0.8-19.8 0.9-6.9 0.9-19.8 0.2-3.8 0.4-1.9 0.5-2.0 0.4-2.4 1.2-28.7 1.1-11.7 0.8-10.9 0.5-2.4 0.9-22.4 0.4-6.5 >ral contraceptives RR = 2.5, 95 more et al. (8) ess in risk of :h perineal ex:crease in risk plications per talc use and logically plauaminating the ito the pelvic ale have been je (12-15). It xcess risk as- BORDERLINE OVARIAN TUMORS AND TALC EXPOSURE 393 ) sociated with application of talc to the per ported in previous studies of women with ineum and to sanitary napkins that was malignant ovarian tumors. In addition, be seen in the three prior studies, none of cause of refusals and other reasons for non which inquired about the type of powder, participation, we were unable to include could have been due to a strong association approximately 30 per cent of potentially restricted to the use of deodorizing pow- eligible cases and controls. Since nonpar ders. The lack of an increased risk among ticipants were similar to participants with women who used talc-containing powder on respect to certain characteristics such as diaphragms (both in our study and in the age and county of residence, we have no previous studies) supports this hypothesis, reason to believe that there was any dissim since deodorizing powder was infrequently ilarity in their use of talc-containing pow used for diaphragm storage. Furthermore, ders. differential asbestos contamination among Given the clues provided by this study different types of cosmetic talcum powders regarding the possible importance of de cannot be ruled out. Until 1975, US- odorizing powders, it would be advisable for manufactured cosmetic talcum powders future studies to elicit information on the - were required to contain at least 90 per cent brand names of talc-containing powders mineral talc, but until 1968, some products and the timing and duration of use of each marked as cosmetic talcum powders did not type of talc-containing powder. Although 1 conform to these guidelines (16, 17). In these data need replication, they raise the 1976, a study of 21 consumer talcum pow- possibility that the risk of ovarian tumors | ders labeled as baby powders, facial pow- in women who apply deodorizing powder to : ders, or body powders obtained from retail the perineum may not relate to talc per se stores in New York City between 1971 and but rather to asbestos contamination and/ 1975 reported that 10 contained concentra- or a substance or substances used specifi 1 tions of asbestiform tremolite and antho- cally for deodorization. ( phyllite ranging from 0.2 per cent to 14 per " cent (4). R eferenc es Although it is difficult to explain the lack of association among women who used baby powder exclusively, according to the prod 1. H arlow B L , Weiss NS. L o fto n S. T he epidem iol ogy o f borderline ovarian tum ors. JN C I 19S7:78: 71-4. 2. Graham J, Graham R. O varian cancer and asbes * uct labels baby powder is reported to con tos. E n viro n Res 1967;1:115-28. tain only talc and no other minerals or 3. Longo D L , Young RC. Cosmetic talc and ovarian cancer. Lancet 1979:2:349-51. deodorizing substances. The product labels 4. Blejer HP, Arlon R. Talc: a possible occupational from deodorizing powders, body powders, and perfumed dusting powders, on the other hand, indicate that they contain de and environm ental carcinogen. J Occup Med 1973;15:92-7. 5. Rohl A N , Langer A M . S e lik o ff IJ. et al. Consumer talcums and powders: m ineral and chemical char odorizing substances and a variety of other free and bonded silicas (potentially high in asbestiform fibers (18)) in addition to talc. acterization. J T o xico l E n viron Health 1976:2: 255-84. 6. Cramer D W , W elch W R. Scully RE. et al. Ovarian cancer and talc. Cancer 1932:50:372-6. We suggest caution when interpreting the results of this study. T he elevated risk among women who specifically used de 7. Hartge P, Hoover R. Lesher LP. et al. Talc and ovarian cancer. (L e tte r!. J A M A 1983:250:1844. 8. W hittem ore AS, W u M L . Paffenbarger RS. et al. Personal and environm ental characteristics re odorizing pow ders co u ld have been due to chance or applicable only to borderline, not lated to epithelial ovarian cancer. II. Exposures to talcum powder, tobacco, alcohol, and coffee. Am J Epidem iol 1989;12S:122S-40. m a l i g n a n t , o v a r i a n t u m o r s . W e b e li e v e t h e 9. W orld Health O rganization. In ternational classi latter possibility to be unlikely, since the fication o f diseases lo r oncology. Geneva: W orld - r>sk associated with the use of any talc- Health O rganization. 1976. 10. H arlow B L , Weiss NS. Roth G. et al. A case- containing powder was similar to that re control study o f borderline ovarian tumors: repro- 394 HARLOW AND WEISS ductive history and exposure to exogenous female hormones. Cancer Res 19S8;4S:5S49-52. U . M an te l N;. Haenzel W . S ta tistica l aspects o f the analvsis of data from retrospective studies of dis ease. JNCI 1959:22:719-48. 12. Eglie GE, Newton M D . T h e transport o f carbon particles in the human female reproductive tract. F e rtil S teril 1961;12:151-5. 13. Venter PF, Itu rra ld e M . M ig ra tio n o f particulate radioactive tracer from the vagina to the perito neal cavity and ovaries. S A fr Med J 1979:55:917 19. . 14. Henderson W'J, Joslin CAF. T u rn b u ll AC, et al. Talc and carcinoma o f the ovary and cervix. J Obstet Gynecol 1971:78:266-72. 15. Henderson W J, H a m ilto n T C . G riffith s K. Talc in norm al and m alignant ovarian tissue. Lancet 1979;1:499. . 16. H ild ic k -S m ith GY. The biology o f talc. B r J Ind Med 1976;33:217-29. 17. C ralley L J, Key M M , G roth D H , et al. Fibrous and m ineral content o f cosmetic talcum products. Am In d Hyg Assoc J 1968;29:350-4. 18. P aoletti L, Caiazza S, D on e lli G, et al. Evaluation by electron microscopy techniques o f asbestos contam ination in industrial, cosmetic, and phar maceutical talcs. Regul T o xico l Pharmacol 1984 4:222-35. ' \ ( [ , j : : A.UKKH.WN loi K N A I. UK Eft! (',)])> right r. 1989 by The -Nr. y right reserved A CASE G U Y R. N E W E L L . The incidence of p in the United States of lung cancer. In U and 28.500 deaths fr expected (1). Over 9 manifest prostate t aged 60 years or mo ual, occupational, other factors have 1 etiology of this canc studies have not bet fying potential risk no way to prevent basis of current km undertook this case tate cancer to exa that have been sugg ing the risk for or development of pro S u b je c ts As part of the re: adult patients at T M. D. Anderson C to complete a det risk factor questio; ously (4). Besides graphic informatio ligion of origin, ant Received for publica: lorm November 14. 198 D epartm ent o f Ca: The U niversity o f Te Center, Houston. T X . ` Department of Urn M. D. Anderson Canee: R eprint requests to I of Cancer Prevention a Texas M . D. Anderson Holcombe Boulevard. 1 Supported in part b> Petroleum Co.. Am arii: The authors thank adm inistering the r.> Dariene W om ack for e d ia b e to lo g ists a n d in m a jo r m e d ic a l ce n te rs w h e re fu n d u sco p ic e x a m in a tio n is do n e ro u tin e ly a n d c o m p e te n t ly . H o w e v e r, in th e office o f th e p r im a r y care p h y sician s, w h e re m o st d ia b e tic s in th is c o u n try re ce iv e m u c h o f th e ir care, a n n u a l e x a m in a tio n o f t h e f u n d i t h r o u g h dilated p u p ils re g re tta b ly is p e rfo rm e d in fre q u e n tly if a t a ll. G iv e n th a t c irc u m sta n ce , a n a b n o rm a l to u rn iq u e t te st re su lt d e m a n d s a co m pe ten t fu n d u sc o p ic e x a m in a tio n to ru le o u t p ro life r a tiv e re tin o p a th y , o fte n b y re fe rra l to a n o p h th a lm o lo gist. I w ish to e m p h a siz e th a t I a m n o t a d v o c a tin g th a t th e to u rn iq u e t te st re p la ce r e g u la r fu n d u sco p ic e x am in a tio n . If D rs A ab y and Zegarra have a co st-e ffe c tiv e stra te g y to e n su re a d e q u a te a n n u a l e x a m in a tio n o f th e 11 m illio n d ia b e tic s in th e U n ite d S ta te s "b y a p h y sic ia n w h o c a n re co gn ize e a rly p ro life ra tiv e d ia b e tic re tin o p a - . th y ," I w o u ld h a p p ily e n d o rse it a n d d isc a rd th e to u rn iq u e t te st; u n til th en , th e to u rn iq u e t te st w ill id e n tify n in e o f e v e ry ten p a tie n ts w ith d ia b e tic re tin o p a th y w h o n e e d to be re fe rre d to su c h a p h y sic ia n . M a n y o f th e se p a tie n ts' c o n d itio n s a re c u r re n tly u n d ia gn o se d u n til lo ss o f v isio n o c cu rs. D e cre a se in c a p illa ry fr a g ility w ith im p ro v e d d ia b e tic c o n tro l n o te d in se v e ra l p a tie n ts w a s n o t m e a n t to im p ly re gre ssio n o f d ia b e tic -re tin o p a th y. H isto lo g ic a l stu d y, h ow e ve r, m a y c o n firm th a t th e to u rn iq u e t te st d o e s a c c u ra te ly re fle ct th e p r o g r e s s io n o r re g re ssio n o f d ia b e tic d e rm a l m ic ro a n g io p a th y . A t p re se n t, th e v a sc u la r o r p la te le t a b n o rm a lity c a u s in g c a p il la ry fra g ility in d ia b e te s is u n k n o w n . I a m c u rre n tly in v o lv e d in a s tu d y c o rre la tin g th e to u rn iq u e t te st w ith flu o re sc e in re tin a l a n g io g r a p h y in th o se p a tie n ts w h o d o n o t h a v e id e n ti fia b le d ia b e tic re tin o p a th y o n o p h th a lm o sco p ic e x am in a tio n . Wu u m A. A r m o iM . MO Waalem Montana OWc HHW 1. C artw right GE: D iagnaetic Laboratory H em atology, ed 4. New York, G ru n t A S tratton Inc, 1968, p 367. Z Stobbe H, ROrup Q Zura N achw tis von K apillarschsden im Rahmen der Mikronngiopathie-Diagnootik beim Diabett* m ellitus mittria S trau v tn u ch s. Sekmeiz M*d W ocKenmhr 197*109:1308-1810. 3. Rodriguen R, Root HP: Capillary fragility and diabetic reti tit. N E ngl J Med 1948^38:391-397. T alc and O varian C ancer To the E ditor.--C r a m e r a n d c o -w o r k e rs' re ce n tly re p o rte d o b se rv in g a n a sso c ia tio n b etw e e n ta lc u se a n d r is k o f o v a ria n cancer. W e th e re fo re e x a m in e d d a ta o n ta lc u se th a t tw o o f u s (L .M . a n d L .P .L .) h a d c o lle c te d a s p a rt o f a case -co n tro l in te rv ie w stu d y Estimated Relative Risk of Ovarian Cancer, According to Reported Use of Talc No laic mentioned Any talc mantionad No diaphragm uaad Diaphragm uaad. no talc Diaphragm, with talc No body talc Soma body talc "A * ovar" Ganftai" Lags only Not gandal Unknown whara C a i*s 82 87 92 14 2S 77 54 37 7 1 8 3 C ontrola 61 100 118 11 41 84 78 57 3 0 8 10 O n gaais, sanitary napkins, or undsrwaar. E stbnatad R alathra Risk 1.0 0 .7 1.0 1.6 0 .8 1.0 0 .8 0 .7 2 .5 -- 0 .8 0 .3 98% ConManca In tan rai 0 .4 -1 . t 0 .7 -3 .7 0 .4 -1 .4 0 .5 -1 .2 0 .4 -1 .2 0 .7 -1 0 .0 ... 0.3 -2 .S 0 .1 -1 .2 o f e p ith e lia l o v a ria n can ce r co n d u cte d fro m 1974 to 1977 in th e W a sh in g to n , D C , area.1 T h e case s w ere 197 w o m e n w ith p a th o lo g ic a lly c o n firm e d p r i m a ry e p ith e lia l o v a ria n c a n c e rs tre a te d in p a rtic ip a tin g h o sp ita ls. T h e c o n tro ls w ere 197 w o m e n tre a te d a t th e sa m e h o sp ita ls fo r c o n d itio n s o th e r th a n gy n e c o lo gic , p sy c h ia tric , o r m a lig n a n t d ise ase s o r p re gn a n cy . T h e c o n tro ls w ere fre q u e n cy m a tc h e d to th e case s on age, race, a n d h o sp ita l. T h e in te rv ie w e rs a sk e d q u e stio n s a b o u t re p ro d u ctive a n d se x u a l h isto ry, m e d ic a l h isto ry , d r u g use, a n d o th e r e x p o su re s. Q u e stio n s a b o u t ta lc u se w ere ad d e d to the q u e stio n n aire a fte r the stu d y began, so 135 case s a n d 171 c o n tro ls w ere a sk e d a b o u t ta lc e xposure. T h e re p o rte d ta lc u se a m o n g c a se s a n d c o n tro ls is g iv e n in th e T a b le . W e e stim a te d th e re la tiv e r is k to ta lc u s e r s a s 0.7 (9 5 % c o n fid e n c e in te r v a l [ C I] = 0 .4 to 1.1). T h e e s tim a te w a s un affe cte d b y a d ju stm e n t fo r race, age, a n d g ra v id ity . N e ith e r w o m e n w h o u se d ta lc on th e ir d ia p h ra g m s n o r th o se w h o u sed it a s body p o w d e r se e m e d to be a t e xcess risk . W o m e n w h o u se d ta lc a s a b o dy p o w d e r w e re a sk e d h o w th e y u se d it. A m o n g th e te n w h o sp e c ific a lly m e n tio n e d u se o n sa n ita r y n a p k in s, u n d e rw e ar, o r th e g e n ita l are a, th e re la tiv e risk w a s e s t im a t e d a s 2.5, b u t th e s m a ll n u m b e r o f e x p o se d w o m e n y ie ld e d a n u n re lia b le e stim a te (9 5 % C I= 0 .7 to 10.0). O u r d a ta th u s in d icate n o o v e ra ll a sso c ia tio n b etw een ta lc use a n d risk o f o v a ria n cancer. A lth o u g h a sm a ll g ro u p o f w o m e n w h o sp e c ifica lly re p o rte d g e n ita l u se o f b o d y ta lc u m p o w d e rs sh o w e d an e xcess re la tiv e risk , use o f ta lc on a d ia p h ra g m , w h ic h w o u ld be th e c lo se st e x p o su re to th e o v a rie s, d id n o t se e m to e le v a te risk . C h a n c e , b ia s in se le ctio n o r o b se r v a tio n , o r co n fo u n d in g m a y h av e in flu e n ce d th e se e stim a te s. O n e im p o rta n t p o te n tia l b ia s to c o n sid e r in t h i s a n d C r a m e r 's s t u d y i s a d i f f e r ence b etw e e n case s a n d c o n tro ls in re c o lle c tin g o r re p o rtin g ta lc u m p o w d e r use, e sp e c ia lly in th e g e n ita l area. T a lc e xposu re w as not a m a jo r focus o f th is stu d y, a n d few d a ta are a v a il a b le to a ss e s s th e lik e lih o o d o f re ca ll b ia s. S u c h a b ia s c o u ld ste m fro m c a se s' h e ig h te n e d a w a re n e ss o r fro m th e fa c t th a t c o n tro ls w e re in te r v ie w e d in th e h o sp ita l w h ile m o st c a se s w e re in te rv ie w e d a t h om e . O n th e o th e r h an d , th e q u e stio n s a b o u t ta lc u se w ere ra th e r sim p le a n d u n a m b ig u o u s. A lso , w e n o te d th a t c a se s a n d c o n tro ls w ere e q u a lly lik e ly to re p o rt d o u c h in g . S in c e re p o rtin g o f use o f d o u ch e s m ig h t be su b je ct to the sa m e re c a ll b ia se s a s ta lc use, th is o b se rv a tio n s u g g e s ts th a t little re ca ll b ia s o p e ra te d . A n o th e r p o ssib le in te r p re ta tio n o f o u r fin d in g s o f n o a p p a r e n t e ffe ct o f u s in g ta lc o n th e d ia p h r a g m b u t so m e effect o f p e rin e a l u se o f p o w d e r is th a t ta lc itse lf d o e s n o t in c re a se risk o f o v a ria n can ce r b u t th a t p a tie n ts w ith o v a ria n can ce r h av e o r p e rce ive a g re a te r need fo r u s in g b o d y p o w d e r in .he g e n ita l are a, fo r re a so n s re la te d e ith e r tc th e b io lo g y o f t h e d is e a s e o r t o life - s t y le . V*V. a gre e w ith C ra m e r a n d c o -w o rk e rs th a t o th e r e p id e m io lo g ic d a ta w ill be u se fu l. Fatibcm Hm it m , MSc R o w k t Hoove*. MO N*ttonal C ancer institute n>ltWMll. Md Uwoa P LlSh u , MPH La m * M cOow a . MO G aorga W ashington Univarsity Me ftcat C enter W ashington. DC 1. C ram ar DVV, Welch WR, Scully RE, et al: O varian cancer and talc. Cancer 198&30:372-376. 2. McGowan 1. P aren t L, Lednar W. at aL The woman a t risk for developing ovarian cancer. G ynecol Oncol l97*7:32S-344. 1 8 4 4 JAMA, Oct 14, 1983-- V d 250, No. 14 L e tte rs Perineal Exposure to Talc and Ovarian Cancer Risk BERNARD L. HARLOW, PhD, DANIEL W. CRAMER, MD, ScD, DEBRA A. BELL, MD, AND WILLIAM R. WELCH, MD O bjective: We sought to determine whether the use of talc in genital hygiene increases the risk for epithelial ovarian cancer. M ethods: We interviewed 235 white women diagnosed with epithelial ovarian cancer between 1984-1987 at ten Boston metropolitan area hospitals and 239 populationbased controls of similar race, age, and residence. R e su lts: Overall, 49% of cases and 39% of controls re ported exposure to talc, via direct application to the per ineum or to undergarments, sanitary napkins, or dia phragms, which yielded a 1.5 odds ratio (OR) for ovarian cancer (95% confidence interval [Cl] 1.0-2.1). Among women with perineal exposure to talc, the risk was significantly elevated in the subgroups of women who applied it: 1) directly as a body powder (OR 1.7, 95% Cl 1.1-2.7), 2) on a daily basis (OR 1.8, 95% Cl 1.1-3.0), and 3) for more than 10 years (OR 1.6, 95% CJ 1.0-2.7). The greatest ovarian cancer risk associated with perineal talc use was observed in the subgroup of women estimated to have made more than 10,000 applications during years when they were ovulating and had an intact genital tract (OR 2.8, 957c Cl 1.4--5.4); however, this exposure was found in only 14% of the women with ovarian cancer. Conclusions: These data support the concept that a life time pattern of perineal talc use may increase the risk for epithelial ovarian cancer but is unlikely to be the etiology for the majority of epithelial ovarian cancers. (O b stet G y necol 1992;80:19-261 T h e th e o ry that h u m a n o v a ria n can ce r m a y be m e so th e lio m a s th at o rig in a te fro m a sb e sto s e x p o su re w a s first p r o p o s e d b y G r a h a m a n d G r a h a m .1 L a te r, P a r m le y a n d W o o d ru ff2 s u g g e s te d th a t e fflu e n c e s th a t m a y a rise fro m th e v a g in a , u te ru s, o r tu b e s m ig h t e n te r the p e lv ic c a v ity a n d in te ra c t w ith o v a ria n su rfa c e e p ith e - F rom the O bstetrics a n d G yn eco lo g y E p id em io lo g y C en ter, B righam a n d Women i H ospital, H a n ia rd M ed ica l School, B oston. M a ssa ch u setts. S u p p o rted by g ra n t R01 C A 4 2 0 0 8 fro m the N a tio n a l C ancer In stitu te liu m to in d u c e s u c h m e so th e lio m a s. L o n g o a n d Y o u n g 3 fu rth e r s p e c u la te d th a t c o sm e tic ta lc m ig h t act in th is m a n n e r b e c a u se o f its c h e m ic a l sim ila rity to a sb e sto s. A lt h o u g h th e re is little d o u b t th a t a sb e s to s m a y in d u c e m e so th e lio m a s, th e lin k b e tw e e n g e n ita l e x p o s u r e to ta lc a n d o v a r ia n c a n c e r is le s s cle ar. T h e fe w e p id e m io lo g ic stu d ie s that h a v e e x a m in e d th e a ss o c ia tio n b e tw e e n ta lc a n d o v a ria n c a n c e r re p o rte d o n ly m o d e s t e le v a tio n s o f risk , b u t d a ta o n a ll p o te n tia l s o u rc e s o f g e n ita l ta lc e x p o su re w e re lim it e d .4-7 T h e p u r p o s e o f t h is r e p o r t w a s to p r e s e n t fin d in g s fro m a n e w case -co n tro l stu d y o f o v a ria n can ce r c o n d u c te d in th e B o sto n m e tro p o lita n area, in w h ic h w e c o n sid e re d a v a rie ty o f m o d e s fo r p e rin e a l ta lc e x p o s u r e , th e fre q u e n c y a n d d u ra tio n o f u se , a n d v a rio u s re p ro d u c tiv e c h a ra c te ristic s th at m ig h t in flu e n c e th e a b ility o f ta lc to e n te r th e p e lv ic c a v ity a n d affect th e o v a rie s. Materials and Methods B e tw e e n J u ly 1984 a n d S e p te m b e r 1987, w e id e n tifie d 394 w o m e n b e tw e e n 1 8 -7 6 y e a rs o f a ge d ia g n o se d w ith b o rd e rlin e o r m a lig n a n t e p ith e lia l o v a ria n c an ce r at o n e o f te n p a rtic ip a tin g h o sp ita ls in th e B o sto n m e tro p o lita n a re a . P e r m is sio n to c o n ta c t e a ch p a tie n t w a s o b ta in e d in a d v a n c e fro m th e p h y sic ia n o f record . A n in -p e rso n in te rv ie w w a s c o n d u c te d w ith 272 (6 9 % ) o f the 3 9 4 c a s e s id e n t ifie d . T h ir t y - o n e p e rc e n t w e r e n o t in te rv ie w e d b e c a u se of p h y s ic ia n a n d /o r p a tie n t re fu sa l, p a tie n t d e a th , o r re lo c a tio n . T h e fin al s a m p le fo r a n a ly sis w a s fu rth e r re stricte d to th e 235 w h ite w o m e n c o n firm e d a s h a v in g a n e p ith e lia l o v a ria n tu m o r b a se d o n a n in d e p e n d e n t p a th o lo g y re v ie w c o n d u c te d b y tw o of the a u th o rs (D A B a n d W R W ). C o n tr o ls w e re se le cte d fro m th e M a s s a c h u s e tts T o w n B o o k s , a n n u a l p u b lic a tio n s th a t list re sid e n ts b y n a m e , a g e , a n d a d d r e s s a c c o rd in g to v o te r p re c in cts. F o r e a ch n e w o v a ria n c a n c e r c ase in te rv ie w e d , a ra n - VOL. 80, N O . 1. JULY 1992 0029-7844'92'S5.00 19 Table 1. Influence of Any Perineal Talc Exposure* on Ovarian Cancer Risk by Characteristics of Study Participants, Boston Metropolitan Area, 1984-1987 Cases Controls Total Talc exposure Total Talc exposure Crude OR 95% Cl All subjects 235 114 (48.5%) 239 94 (39.3%) 1.5 1.0-2.1 Age (y) <50 96 41 (42.7%) 101 28 (28.0%) 1.9 1.2-3.4 a 50 139 73 (52.5%) 138 66 (47.8%) 1.2 0.8-2.1 Education (y) sl2 93 51 (54.8%) 115 48 (41.7%) 1.7 1.1-3.0 >12 142 63 (44.4%) 124 46 (37.1%) 1.4 0.9-2.4 Manu! status Never married 40 14 (35.0%) 24 6 (25.0%) 1.6 0.5-5.7 Ever married 195 100 (51.3%) 215 88(40.9%) 1.5 1.0-2.7 Religion Jewish 35 21 (60.0%) 21 12 (57.1%) 1.1 0.4-3.9 N'on-Jewish 200 93 (46.5%) 218 82 (37.6%) 1.4 0.9-2.5 Weight (lb) <140 123 53 (43.1%) 125 49 (39.2%) 1.2 0.7-1.9 2140 112 61 (54.5%) 114 45 (39.2%) 1.8 1.1-3.1 Use oi OCs t,moi 23 66 31 (47.0%) 82 27 (32.9% ) 1.8 0.8-4.5 <3 or never 169 83 (49.1%) 157 67 (42.7%) 1.3 0.8-2.3 No. ol live-bom children 0 79 30 (38.0%) 43 12 (27.9%) 1.6 0.7-4.0 1 31 21 (67.7%) 27 5(18.5%) 9.2 2.9-46.2 2 40 24 (60.0%) 63 26 (41.3%) 2.1 0.9-3.3 23 85 39 (45.9%) 106 51 (48.1%) 0.9 0.6-1.6 OR = odds rane; Cl = confidence interval, OCs = oral contraceptives * Sources of perinea) talc exposure include: dusting of underwear, diaphragms, or sarutarv napkins; use by partner on his permeai area; use as a body powder d o m n u m b e r g e n e ra to r se le cte d o n e p a g e fro m th e to w n b o o k c o r r e s p o n d in g to th e c a s e 's p re c in c t o f re sid e n c e . B y w o r k in g fo r w a r d in th e to w n b o o k , w e s e le c t e d th e fir s t five f e m a le s u b j e c t s w it h in 2 y e a r s o f a g e o f the c a se a s p o te n tia l c o n tr o ls . O f th e se five , th e first su b je c t o f th e s a m e ra ce a s th e c a se w ith o u t a h isto rv o f a b ila te ra l o o p h o r e c t o m y w a s a sk e d to p a r ticip ate in tn e stu d y . O f th e 526 c o n tro ls c o n ta c te d , 239 in t e r v ie w s w e r e c o n d u c t e d (2591 c o u ld n o t b e r e a c h e d , 109c re p o rte d a h isto ry o f b ila te ra l o o p h o re c to m y , a n d 199c d e c lin e d to p a rtic ip a te ). F u rth e r d e ta ils o f th e s tu d y m e th o d s can be fo u n d e lse w h e re .* T h e in -p e rso n in te rv ie w fo c u se d o n th e fo llo w in gd e m o g ra p h ic a n d o c c u p a tio n a l h isto ry ; m e d ic a l a n d re p ro d u c tiv e h isto rie s, in c lu d in g p re g n a n c ie s, h o r m o n e s u se d , a n d g y n e c o lo g ic o p e ra tio n s; d ie ta ry h is to ry; c ig a re tte s m o k in g ; a n d h y g ie n ic p ra ctic e s. T h e h y g ie n ic p ra ctic e s in c lu d e d in fo r m a tio n r e g a r d in g the u se o f d o u c h e s, ty p e o f sa n ita ry p ro te c tio n u se d , a n d p e rin e a l e x p o s u r e to talc. Q u e r ie d s o u r c e s o f p e rin e a l ta lc e x p o s u r e in c lu d e d d u s t in g o f u n d e r w e a r , sa n ita ry n a p k in s , a n d d ia p h r a g m s ; e x p o s u r e v ia h u s b a n d 's u se o f talc; a n d m o re d ire c t e x p o s u r e to th e p e r in e u m a s a b o d y p o w d e r. N o re lia b le in fo r m a tio n o n ta lc e x p o su re d u r in g in fa n cy ' w ith d ia p e r in g c o u ld be o b ta in e d , a n d w o m e n u s in g ta lc a s a b o d y p o w d e r o n a re a s o th e r th a n th e p e rin e u m w e re c o n sid e re d n o n e x p o se d . F o r each e x p o su re , w e in q u ire d a b o u t b ra n d s u se d , a ge at first u se , to ta l y e a r s o f u se , a n d fre q u e n c y o f u s e p e r m o n th , to e n a b le u s to e stim a te th e to tal life tim e n u m b e r o f a p p lic a tio n s fro m a ll so u rc e s o f e x p o su re . D iffe re n c e s b e tw e e n c a se s a n d c o n tro ls in th e d is tri b u tio n o f th e se v a r io u s e x p o s u r e s to ta lc w e re e x a m in e d b o th q u a lita tiv e ly a n d q u a n tita tiv e ly . T h e in flu e n ce o f c o n fo u n d e r s a n d effect m o d ifie rs w a s a ss e s se d first t h r o u g h stra tific a tio n a n d th e n u s in g u n c o n d i tio n a l lo g istic re g re ssio n .9 T h e p rim a ry m a tc h in g v a r i able, a g e , w a s re ta in e d in e a c h lo g is tic m o d e l. T h e y test fo r lin e a r tre n d w a s c a lc u la te d b a se d o n th e c h a n g e in d e v ia n c e in m o d e ls w ith a n d w ith o u t c o n tin u o u s e x p o su re v a ria b le s.9 Results T a b le 1 s h o w s th e p ro p o rtio n o f case s a n d c o n tro ls w ith a n y re p o rte d p e rin e a l ta lc e x p o su re , a n d th e a sso c ia te d c ru d e e x p o su re o d d s ra tio (O R ), b y c e rta in d e m o g ra p h ic a n d re p ro d u c tiv e su b g ro u p s. O v e ra ll, a 20 H arlow et al Talc a n d O v a r ia n C ancer O bstetrics b G ynecology Table 2. History of Talc Exposure by Types of Application, Brand of Powders, Years and Frequency of Use, and Era of Use Cases Controls Adjusted OR* 95% c i No genital talc application Any genital talc application Type of application Only via sanitary napkins and, or underwear Via partner or applications to diaphragm' Via dusting powder to perineum' Applications of talc per month <5 5-29 230 Years of talc use <10 10-29 tIoi>V Age (y) at first talc use <20 20-25 >25 Years since last talc use Within last 6 mo Between 6 mo-10 v 10 v or more Era of use' Exclusive use after 1960 Any use before I960 Brand of application4 Brand or genenc baby powder Deodorizing or other scented powders * 121 (51.5%) 114 (48.5%) 9 (3.8%) 20 (8.5%) 85 (36.2%) 32 (13.6%) 24 (10.2%) 58 (24.7%) 14 (6.0%) 49 (20.9%) 51 (21.7%) 66 (28.1%) 27 (11.5%) 21 (8.9%) 48 (20.4%) 36 (15.3%) 30(12.8%) 29(12.3%) 75 (31.9%) 91 (38.7%) 16 (6.8%) 145 (60.7%) . 94 (39.3%) 12 (5.0%) 21 (8.8%) 61 (25.5%) 28 (11.7%) 25 (10.5%) 41 (16.7%) 15 (6.3%) 39 (16.3%) 40(16.3%) 50 (20.9%) 26 (10.9%) 18 (7.5%) 27(11.3%) 39 (16.3%) 28(11.7%) 30 (12.6%) 57 (23.9%) 72 (30.1%) 17(7.2%) 1.0 1.5 1.1 1.2 1.7 1.5 1.2 1.8 1.2 1.6 1.6 1.7 1.2 1.6 2.3 1.1 1.4 1.1 1.7 1.6 1.2 1.0-2.1 0.4-2.8 0.6-2.4 1.1-2.7 0.8-2.7 0.6-2.2 1.1-3.0 0.5-2.6 1.0-2.7 1.0-2.7 1.1-2.7 0.6-2.2 0.8-3.2 1.3-J.O 0.7-1.9 0.8-2.6 0.6-2.1 1.1-2.7 1.1-2.5 0.6-2.5 OR = odds ratio; O * confidence interval. Adjusted for pants- (0, 1-2, >2), education {<12 years, >12 years), manta! status (never married, ever married), religion (Jewish, non-Jewish), use of sanitary napkins (no. yes), douching (no, yes), age (continuous), and weight (<140 lb, 140 lb). ' Includes combination* with sanitary napkins or underwear. ! Restricted to women older man 10 years in 1960; 100 cases and 118 controls were unexposed and used as the referent group. 4Excludes seven cases and five controls with unknown powders. Seven cases and four controls reported combinations of more than one brand and were classified according to the brand used most frequently and lor the longest penod. Specific brands mentioned by cases were: Johnson and Johnson, 71; genenc baby powder, 20; Shower to Shower, four; other scented, 12. Specific brands mentioned by controls were: Johnson and Johnson, 54; genenc baby powder, 18; Shower to Shower, three; other scented powder, 14. gre a te r p e rc e n ta g e o f c a se s (4 8 .5 % ) th a n c o n tro ls (3 9 .3 % ) re p o rte d a n y p e r in e a l e x p o s u r e to ta lc c o n ta in in g p o w d e rs. S u b g r o u p s o f c o n tro ls in w h ic h e x p o s u r e to talc a p p e a r e d to b e m o r e c o m m o n w e re w o m e n o ld e r th a n 50 y e a r s (P = .002), e v e r m a rrie d (P = .12), J e w ish (P = .08), a n d p a r o u s (P = .05). C o n t r o ls w h o re p o rte d n o o ra l c o n tra c e p tiv e (C X I) u se a lso re p o rte d gre ate r ta lc u se . H o w e v e r , th is in te ra c tio n m a y b e e x p la in e d b y th e o ld e r a g e d istrib u tio n a m o n g c o n tr o ls w h o re p o rte d p e r in e a l a p p lic a t io n o f talc. W e o b se rv e d s tro n g e r a s s o c ia tio n s b e tw e e n ta lc a n d o v a r ia n c an c e r risk in th e s u b g r o u p s o f c a se s a n d c o n tro ls y o u n g e r th a n a ge 50, le ss w e ll e d u c a te d , h e a v ie r th a n 140 lb 5 y e a rs b e fo re d ia g n o s is , a n d re p o rtin g a h isto ry o f o n e o r tw o liv e b irth s. T h e n u m b e r o f liv e b irth s w a s the o n ly facto r th at p ro d u c e d sta tistic a lly sig n ific a n t h e te ro g e n e ity in th e O R s fo r th e ta lc a n d o v a ria n c a n c e r a sso c ia tio n . A g e , e d u c a tio n , m a rita l sta tu s, re ligio n , w e ig h t, a n d p a rity w e re c o n sid e re d c o n /o u n d e rs a n d w e re in c lu d e d a s c o v a n a te s in su b se q u e n t m u ltiv a ria te m o d e ls. U s e o f O C s d id n o t c o n fo u n d th e ta lc -o v a ria n c a n c e r a sso c ia tio n . T a b le 2 e x a m in e s th e a s s o c ia tio n b e tw e e n ta lc u s e a n d o v a ria n c a n c e r b y v a ria b le s re la te d to th e sp e c ific ty p e s a n d la n d s o f a p p lic a tio n s, a n d m e a su re s o f d u ra tio n in c lu d in g le n g th , fre q u e n c y , a n d p e rio d o f u se . C o m p a r e d w ith w o m e n w ith n o g e n ita l ta lc e x p o su re , w o m e n e x p o s e d to ta lc o n ly th r o u g h u s e a s a d u stin g p o w d e r o n sa n ita ry n a p k in s o r u n d e rw e a r h a d n o a p p re c ia b le in c re a se d risk fo r o v a ria n can cer. T h e re w a s a lso n o su b sta n tia l in c re a se in risk a m o n g w o m e n e x p o se d to ta lc o n ly t h r o u g h a h u s b a n d 's u s e o r u s e in the sto ra ge o f d ia p h ra g m s, o r in c o m b in a tio n w ith a p p lic a tio n s to sa n ita ry n a p k in s or u n d e rw e a r. T h e m o s t fre q u e n t m e t h o d o f p e r in e a l talc e x p o s u r e w a s u se a s a d u s tin g p o w d e r d ire c tly to th e p e rin e u m , VOL 80. N O . 1. JULY 1992 Harlow et al Talc a n d O v a ria n C a n cer 21 Table 3. Estimated Total Lifetime Perineal Applications of Talc-Containing Powders in Cases and Controls Applications Cases Controls Adjusted OR* 959c a Total applications None <1000 1000-10.000 >10,000 JT2 ldf test for linear trend * 2.85, P *= .094' Applications excluding use after hysterectomy or tuba] ligation None <1000 1000-10.000 >10,000 X2 ldf test for linear trend = 3 19, P = .077* Applications excluding use after hysterectomy or tubal ligation, and use dunng noncrvulatory months' None <1000 1000-10.000 >10,000 X1 ldf test for linear trend = 6.15. P = .015' 121 18 54 42 121 19 57 38 124 24 55 32 145 19 44 31 145 19 46 29 149 23 51 16 1.0 1-3 0.7-2.7 1.5 0.9-2.4 1.8 1.0-3.0 1.0 1.4 0.7-2.9 1.5 0.9-2.4 1.7 1.0-3.0 1.0 1.5 0.8-2.9 1.3 0.8-2.0 2.8 1.4-5.4 Abbreviations as in Table 2. Adjusted for parity (C. 1-2, >2), education (< 12 years, >12 years), marital status (never married, ever married), religion (Jewish. non-Jewish), use ol sanitary napkins (no, ves), douching (no, yes), age (continuous), and weight (<140 lb, <140 lb). ' Trend test based on actual applications as a continuous variable. 1 Excludes exposures while taking oral contraceptives, while pregnant or breast-feeding, or occurring after menopause. There were three cases and four controls who moved from 'exposed'' to ' nonexposed" in this category, as all of the exposure occurred during oral contraceptive use, pregnancies, or alter menopause a lo n e o r in c o m b in a tio n w ith e ith e r a p a r tn e r 's u s e o r u s e in th e storage of d ia p h r a g m s . T h is e x p o s u r e o c c u rre d in 85 c a se s (3 6 .2 % ) a n d 61 c o n tro ls (25.59c) (P = .01). O f th e a p p lic a tio n m o d e s s tu d ie d , d ire c t p e r in e a l a p p lic a tio n p ro d u c e d th e gre a te st risk ( O R 1.7, 959c c o n fid e n c e in te rv a l [CJ] 1.1-2.7). W e a lso e x a m in e d the ta lc -o v a n a n can ce r a sso c ia tio n b y fre q u e n c y a n d y e a rs o f u se (T a b le 2). W h e n m o n th ly fre q u e n c y w a s c o n sid e re d a s a c o n tin u o u s v a ria b le in th e lo g is tic m o d e l, th e if lin e a r te st o f tre n d w a s 4 .0 6 (P = .046), in d ic a t in g th a t th e risk fo r o v a r ia n c a n c e r in c re a se d sig n ific a n tly w ith in c re a sin g fre q u e n c y o f a p p lic a tio n s p e r m o n th T h e c a te go ric a l a n a ly s is s h o w e d th a t re la tiv e to n o n -u se rs, th e risk w a s g re a te st in w o m e n w h o a p p lie d ta lc at le a st o n c e p e r d a y . W h e n y e ars of u se w a s in c lu d e d a s a c o n tin u o u s v a r i a b l e , t h e t e s t f o r l i n e a r t r e n d w a s 3 . 3 2 . ( P s= .07). T h e c a te g o r ic a l a n a ly s is s h o w e d th a t re la tiv e to n o n -u s e r s , w o m e n w h o a p p lie d ta lc fo r m o re th a n 10 y e a rs w e re a t 6091 gre a te r risk fo r o v a ria n can cer. L ik e w is e , p e r in e a l a p p lic a tio n s o f ta lc e a rly in life (b e fo re a ge 20) o r a p p lic a tio n s w ith in 6 m o n th s of d ia g n o s is (re fe re n ce a g e fo r c o n tro ls) p ro d u c e d th e stron ge r O R s. T o a sse ss w h e th e r th e risk o f o v a ria n c a n c e r w ith p e rin e a l e x p o s u r e to ta lc w a s a ffe cte d b y th e tim e w h e n ta lc -c o n ta in in g p ro d u c ts w e re m a n u fa c tu re d , w e e x a m in e d th e a sso c ia tio n se p a ra te ly in w o m e n w h o o n ly u se d ta lc u m p o w d e r afte r 1960 a n d in w o m e n w h o re p o rte d a n y u se o f ta lc u m p o w d e r b e fo re 1960 (T a b le 2). A ft e r re stric tin g th e p o p u la t io n to w o m e n o ld e r th a n a ge 10 in 1960 a n d a d ju s tin g fo r a ge , p a rity , a n d a n u m b e r o f o th e r d e m o g r a p h ic c h a ra c te ristic s, w e fo u n d th a t th e a sso c ia tio n o f ta lc a n d o v a ria n c a n c e r w a s g re a te r in w o m e n u s in g ta lc p r o d u c ts b e fo re 1960 (P = .025). T h e la st e n try in T a b le 2 s h o w s th e risk s b y b ra n d o f p o w d e r u se d . N o su b je c ts c o u ld re c a ll e x c lu siv e u se o f sta rch -b a se d p o w d e rs. B ra n d o r g e n e ric "b a b y p o w d e r " w a s u se d m o st fre q u e n tly a n d w a s th e c ate go ry a sso c ia te d w ith a sta tistica lly s ig n ific a n t risk fo r o v a r ia n can cer. W ith re sp e ct to o th e r p o w d e rs, fo u r case s a n d th re e c o n tro ls re p o rte d p rim a r y u se o f d e o d o riz in g p o w d e rs, a n d 12 case s a n d 14 c o n tro ls re p o rte d p ri m a r y u se o f o th e r s c e n te d p o w d e r s . It s h o u ld b e a p p re c ia te d that, b e c a u se th e p e rio d o f e x p o su re o fte n o c cu rre d o v e r d e cad e s, v e rific a tio n o f b ra n d s w a s n o t p o ssib le . T a b le 3 e x a m in e s th e o v a ria n c a n c e r risk a sso c ia te d w ith th e total n u m b e r o f a p p lic a tio n s to th e p e rin e u m , e stim a te d b y c u m u la tin g fre q u e n c y a n d y e a rs o f u se fo r th e v a rio u s k in d s o f e x p o su re s. A n 8 0 % e x ce ss risk w a s a sso c ia te d w ith a n e stim a te d e x p o su re o f m o re th a n 10,000 a p p lic a tio n s (e q u iv a le n t to d a ily u se fo r 30 22 H arlow et al Talc a n d O varian Cancer O bstetrics & Gynecology Table 4. Adjusted Odds Ratios and 95% Confidence Intervals for Ovarian Cancer by Any Perineal Exposure to Talc* and Indicators of Ovulation and Tubal Occlusion Cases C ontrols Total Talc exposure Total Talc exposure A d ju sted OR' 95% C l A ll subjects 235 114 (48.5% ) 239 94 (39.3% ) 1.5 1 .0 -2 .1 M id -c y d e pain No Yes 181 88 (4 8 .6 % ) 184 7 7 (4 1 .9 % ) 54 26 (4 8 .2 % ) 55 1 7 (3 0 .9 % ) 1.4 2.0 0 .9-2.2 0 .8 -5 .2 Regular period No Yes 26 12 (4 6 .2 % ) 34 16 (4 7 .1 % ) 209 102 (48.8% ) 205 78 (38.1% ) 1.1 1.7 0 .4 -3 .4 1 .1 -2 .5 H D or ectopic pregnancy No 226 113 (50.0% ) 230 92 (40.0% ) 1.6 1 .1 -2 .4 Yes 9 1 (1 1.1% ) 9 2 (22.2% ) 0.1 0 .0 1 -7 .0 P ID = p elvic in fla m m a to ry disease; o th e r abbreviations as in Tab le 2. ' Sources o f p erin eal talc exp o su re include; d u s tin g o f u n d e rw e a r, d ia p h ra g m s , o r san itary napkins; use by p a rtn e r on his p erin eal area; use as a bod y p o w d e r. - ' A d ju s te d fo r p a rity (0. 1 -2 . > 2 ) , edu cation (< 1 2 years, > 1 2 years), m a rita l status (n ever m a m e d , eve r m a rrie d ), relig io n (Jew ish, n o n -J e w is h ), use of s a n ita ry n a p k in s (n o , yes), d o u c h in g (n o , yes), age (c o n tin u o u s ), a n d w e ig h t ( < 1 4 0 lb, 2 1 4 0 lb). y e a rs) a s c o m p a re d w ith n o n -u se rs. W h e n c o n sid e re d a s a c o n tin u o u s v a ria b le in th e lo g istic m o d e l, th e lin e a r te st o f fr e n d w a s 2 .8 5 (P = .094). T h e r e m a in in g e n trie s in T a b le 3 s h o w h o w c o n d itio n s th at e ith e r c lo se th e u p p e r g e n ita l tract o r are a sso c ia te d w ith a n o v u la tio n affect th e d o se re sp o n se o f n u m b e r o f a p p lic a tio n s o n o v a ria n can ce r n sk . T h e se co n d e n try s h o w s th e effect o f c e n so rin g a p p lic a tio n s th at o c cu rre d afte r tu b al lig a tio n o r h y ste re cto m y. N o a p p re c ia b le c h a n g e in th e O R s o r th e d o se re sp o n se w a s n o te d . T h e th ird e n try s h o w s th e effect o f c e n s o r in g a p p lic a tio n s after h y ste re c to m y a n d tu b al liga tio n a n d u se d u r in g p re su m e d n o n o v u la to ry p e n o d s. E x c lu d e d w ere e x p o su re s o c c u rrin g w h ile ta k in g O C s , w h ile p r e g n a n t o r b re a st-fe e d in g , a n d afteT m e n o p a u s e . T h e n s k a ss o c ia te d w ith fe w e r th a n 10,000 a p p lic a tio n s w a s n o t s u b s t a n tia lly a lte re d . H o w e v e r , th e n s k a sso c ia te d w ith m o re th a n 10,000 a p p lic a tio n s w a s n e a rly th re e f o ld a n d s t a t is t i c a ll y s ig n i f i c a n t . T h e x* te s t f o r a li n e a r tre n d o f risk , b y n u m b e r o f a p p lic a tio n s a s a c o n tin u o u s v a r ia b le , in c r e a se d to 6 .1 5 (P = .015). T a b le 4 s h o w s th e a s s o c ia tio n o f ta lc e x p o s u r e a n d o v a ria n c a n c e r b a se d o n o th e r c lin ic a l facto rs th a t m a y p re d ic t e ith e r o v u la tio n o r tu b a l o c clu sio n . T h e O R s w e re g re a te r in w o m e n w ith a h isto ry o f m id -c y c le p a in o r re g u la r p e rio d s -- p o te n tia l c lin ic a l p re d ic to rs o f o v u la to ry c y c le s. A n a ss o c ia tio n b e tw e e n ta lc a n d o v a r ia n c a n c e r w a s a b se n t in w o m e n w ith a h isto ry o f e ith e r p e lv ic in fla m m a t o r y d ise a se o r e c to p ic p r e g n a n c y -- p o te n tia l m a rk e rs o f a c lo se d g e n ita l tract. H o w e v e r , o n ly n in e c a se s a n d n in e c o n tro ls re p o rte d a h isto ry o f p e lv ic in fla m m a t o r y d ise a se o r e c to p ic p re g n a n c y , a n d t h e Cl o n t h e e x p o s u r e O R w a s c o r r e s p o n d i n g l y w id e . T a b le 5 s h o w s th e re la tiv e risk fo r a n y p e rin e a l u se o f ta lc w h e n re stric te d to sp e c ific h is to lo g ic ty p e a n d g ra d e o f o v a ria n tu m o rs. T h e g re a te st a sso c ia tio n w a s fo u n d in w o m e n d ia g n o se d w ith a n e n d o m e trio id tu m o r o r b o rd e rlin e o v a n a n tu m o r. Discussion A n im a l a n d e p id e m io lo g ic stu d ie s h a v e a d d re sse d th e p la u s ib ility o f a n a ss o c ia tio n b e tw e e n ta lc a n d o v a r ia n can ce r. In tra p e rito n e a l in je c tio n o f ta lc in ro d e n ts p ro d u c e d p a p illa ry c h a n g e s in th e su rfa ce e p ith e liu m n o t in c o n s is te n t w ith th e first sta g e in th e d e v e lo p m e n t o f s u r f a c e p a p i l l a r y e p i t h e l i a l n e o p l a s m s . 10 H o w e v e r , b e c a u se th e o v a rie s o f sm a ll ro d e n ts are s u rr o u n d e d b y a Table 5. History of Talc Use by Histologic Type and Grade H istologic type A n y use o f talc No use of talc A d ju s ted OR* 95% C l C ontrols 94 145 1.0 H istologic type Serous 60 64 1.4 0 .9 -2 .2 M uanous 17 25 1 .2 0 . 6 - 2 .5 E ndom etnoid 18 11 2 .8 1 .2 - 6 .4 O ther 19 21 1.6 0 . 8 - 3 .3 Histologic gtade B orderline 32 30 2.4 1 .2 -4 .5 G rade 1 6 11 1.0 0 . 3 - 2 .8 G rade 2 21 22 1.5 0 . 7 - 3 .0 G rade 3 27 24 1.5 0 .8 -2 .8 U n d ifferen tiated 28 34 1.2 0 .7 -2 .2 A b b re v ia tio n s as in T a b le 2. * A d ju s te d fo r p a rity (0. 1 -2 , > 2 ) , e d u c a tio n ( < 12 yea rs , > 1 2 years), m arital status (n ever m a m e d . eve r m a rrie d ), relig io n (Jew ish, n o n -Ie w is h ), age (c o n tin u o u s ), a n d w e ig h t ( < 1 4 0 lb, 2 1 4 0 lb). VOL. 80, N O . 1, JULY 1992 H arlow et al Talc a n d O v a ria n C a n c er 23 Tabic 6 . O d d s R a t i o s W i t h 9 5 % C o n f i d e n c e I n t e r v a l s o f O v a r i a n C a n c e r i n R e l a t i o n t o A n y P e r i n e a l E x p o s u r e t o T a l c a s R e p o rte d in P re v io u s E p id e m io lo g ic Stu d ie s Cases Controls Authorfs) (year) Total Tile exposure Total Tile exposure Crude OR 95% Cl Cramer et at* (1982) 215 92(42.8%) 215 61 (28.4%) 1.9 1.3-2.9 Hartge et al (1983)' 135 67(49.6%) 171 100 (58.5%) 0.7 0.4-1.1 Whittemore et als (1988) 188 98 (52.1%) 539 248(46.0%) 1.4 0.9-2.0 Harlow and Weiss6 (1989)' 116 49 (42.2%) 158 64 (40.5%) 1.1 0.7-2.1 Booth et al7 (1989) 217 141 (65.0%) 434 256 (59.0%) 1.3 0.9-1.9 Harlow et al (1992) (current study) 235 114 (48.5%) 239 94 (39.3%) 1.5 0.9-1.8 All studies' 1106 561 (50.7%) 1756 823 (46.9%) 1.3 1.1-1.6 Abbreviations as in Table 2. * Hartge P, Hoover R, Lesher LP, et al. Talc and ovarian cancer (letter). JAMA 1983^50:1844. Odds ratio may indude nonperineal talc exposure as well. ' Restncted to borderline ovarian tumors. 1 Meta-analysis.1' p e rito n e a l b u rsa w h ic h o fte n b e c o m e s o c c lu d e d a n d d iste n d e d w ith fo llic u la r flu id afte r in tra p e rito n e a l in je ctio n o f fo r e ig n b o d ie s ,1 it is d iffic u lt to d is tin g u is h w h e th e r th e p a p illa ry c h a n g e s a re th e effects o f fo r e ig n -b o d y e x p o su re o r b u rsa l d iste n tio n . It w o u ld be w o r th w h ile to re p e a t th e ta lc e x p e rim e n ts in g u in e a p ig s o r ra b b its, th e a n im a ls u se d in th e o n g in a l re search of G ra h a m an d G ra h a m .1 Is th e re e v id e n c e to s u g g e s t th a t ta lc c a n tra n slo c a te fro m th e v a g in a to th e p e rito n e a l c a v ity ? It is k n o w n th a t re d c e lls a n d e n d o m e tria l tissu e are c a p a b le o f r e t r o g r a d e f l o w f r o m t h e f a l l o p i a n t u b e s . 11 E x p e r i m e n t s in rats c o n f ir m e d t h e p r e s e n c e o f t a lc in th e o v a r ie s a fte r t h e in t r o d u c t io n o f a ta lc s u s p e n s io n into th e v a g in a a n d ce rvical o s . ': Ln h u m a n s , tw o stu d ie s o b s e r v e d t h e m i g r a t i o n o f i n e r t c a r b o n p a r t i c l e s 13 a n d r a d i o a c t i v e l y l a b e l e d h u m a n a l b u m i n m i c r o s p h e r e s 14 fro m th e v a g in a to th e fa llo p ia n tu b e s. In a d d itio n , se ve ral in v e stig a to rs h a v e o b se rv e d b ir frin g e n t c ry s t a ls e m b e d d e d i n o v a r i a n t i s s u e . 15-17 H o w e v e r , c r it ic s h a v e a rg u e d th a t th e se fin d in g s re su lte d fro m p o o rly d e sig n e d stu d ie s o r la c k o f p re c isio n in m e a su rin g p a rtic u la te s, d u e to le a c h in g o f ra d io n u d e o t id e m a r k e rs fro m th e te st m a te ria ls o r th e in tro d u c tio n o f c o n t a m i n a n t s d u n n g t i s s u e p r o c e s s i n g . 18 l n s i x m u l t i p a ro u s c y n o m o lg u s m o n k e y s se p a ra te ly c age d , n o tr a n slo c a tio n o f ta lc w a s o b s e r v e d a fte r 3 0 c o n se c u tiv e d a y s o f d o u c h in g w ith a su sp e n sio n o f n e u tro n a c tiv a te d ta lc c o u p le d w ith w e e k ly in je c tio n s o f o x y to c i n . 18 H o w e v e r , t h i s s t u d y w a s n o t a b l e t o a d d r e s s t h e effects o f lo n g -te rm u se o r c o itu s, w h ic h m ig h t fac ili tate talc tr a n s lo c a tio n . S e v e r a l e p id e m io lo g ic s t u d ie s 4-7 h a v e a d d r e s s e d th e g e n ita l t a lc -o v a r ia n c a n c e r a s s o c ia t io n (T a b le 6). E a c h s tu d y h a s c o n s is te n tly re p o rte d little o r n o a ss o c ia tio n w ith th e u se o f ta lc -d u ste d d ia p h ra g m s , a n d stro n ge r a s s o d a t io n s w ith d ire ct p e rin e a l a p p lic a tio n . D o s e re s p o n s e o f p e rin e a l ta lc e x p o s u r e fro m a ll s o u r c e s b y fre q u e n c y o r y e a rs o f u se w a s n o t a v a ila b le in th e stu d ie s b y H a rtg e et al (H a rtg e P, H o o v e r R , L e sh e r L P , e t a l. T a lc a n d o v a r ia n c a n c e r [le tte r]. J A M A 1983;250:1844), H a r lo w a n d W e is s ,6 o r C r a m e r et a l.4 W h itte m o re et a l5 re p o rte d n o sig n ific a n t d o se re sp o n se b y ye ars or freq u en cy of use, w h e re a s B o o th et a l7 re p o rte d a m a rg in a lly sig n ific a n t tre n d w ith fre q u e n c y o f u se . U s in g th e te c h n iq u e s o f m e ta -a n a ly sis, in w h ic h O R s fro m m u ltip le stu d ie s are w e ig h te d b y t h e i r v a r i a n c e s , 19 w e c a l c u l a t e d a s t a t i s t i c a l l y s i g n i f i c a n t O R o f 1.3 fo r a n y p e rin e a l ta lc e x p o s u r e a n d o v a ria n c an c e r risk (9 5 % C l 1.1 -1 .6 ) fro m th e v a rio u s stu d ie s. W e th e re fo re c o n d u d e th at th e re is a n a sso ria tio n , a lb e it m o d e st, b e tw e e n o v a ria n c a n c e r a n d p e r in e a l ta lc u se . H o w e v e r , in s u ffid e n t d e ta il h a s b e e n a v a ila b le to ru le o u t a stro n g e r a sso c ia tio n in c e rta in su b gro u p s of users. O u r s t u d y in c lu d e d g re a te r d e ta il o n g e n ita l ta lc u se , in c lu d in g m e th o d s, fre q u e n c y , a n d y e a rs o f u se (T a b le 2). T a lc a p p lie d a s a d u s t in g p o w d e r d ire c tly to th e p e rin e u m c a rrie d a g re a te r risk th a n le ss d ire ct e x p o s u r e v ia a p a r t n e r 's u s e o r th e d u s t in g o f u n d e r g a r m e n ts, sa n ita ry n a p k in s, o r d ia p h ra g m s. C u rre n t u se o f ta lc w a s a s s o d a t e d w ith a g re a te r o v a ria n c a n c e r risk th a n p a st u se . D a ily v e rs u s le ss th a n d a ily ta lc u se , a n d ta lc u se fo r m o re th a n 10 y e a rs v e rs u s le ss th a n 10 y e a rs, w e re a sso d a te d w ith gre ate r risk fo r o v a ria n can ce r. M o s t su b je cts re p o rte d u se o f "b a b y p o w d e r." W e w e re u n a b le to c o n firm a p re v io u s fin d in g th a t p o w d e rs w ith "d e o d o r iz in g " a g e n ts w e re a ss o d a te d w ith p a rtic u la r risk .6 T h u s, th is s tu d y fa ile d to a n s w e r a k e y issu e in th e ta lc -o v a ria n c an ce r a sso ria tio n : w h e th e r th e r is k p e rta in s to a ll c o sm e tic ta lcs o r o n ly to c e rta in 24 H arlow et al Talc a n d O v a ria n C ancer O bstetrics & G ynecology \ [ 1i f IN .1 i I p re p a ra tio n s lik e ly to b e c o n ta m in a te d b y a sb e sto s. C r a l l e y e t a ]20 a n d R o h l e t a l21 f o u n d c o n s id e r a b le v a ria tio n in th e p u rity o f c o sm e tic ta lc s, w ith fib e rfo rm c o n te n ts v a r y in g fro m le ss th a n 1 to 3 0 % . M o s t o f th e se p r o d u c t s w e r e m a n u fa c tu r e d b e fo re 1970 a n d it is lik e ly th a t th e a sb e stifo rm c o n te n t h a s d e c re a se d sin c e 1976, w h e n m a n u fa c tu re rs in stitu te d v o lu n ta ry g u id e lin e s o h a sb e sto s c o n ta m in a tio n . O u r fin d in g o f a lo w e r o v a ria n can ce r risk in w o m e n w ith e x c lu siv e u se o f ta lc a fte r 1960 m a y s u p p o r t th is th e o ry . B e c a u se o f th e d iffic u lty in o b ta in in g a c o m p le te a n d d e ta ile d h isto ry o f p o w d e rs u se d , th e issu e o f w h e th e r risk p e rta in s o n ly to a sb e sto s-c o n ta m in a te d p o w d e rs m a y n e e d to be se ttle d b y a n im a l e x p e rim e n ts. In o u r a n a ly sis, w e first c a lc u la te d a ll g e n ita l a p p li c a tio n s o f ta lc b a se d u p o n fre q u e n c y a n d y e a r s o f u se . A s a c o n tin u o u s v a ria b le in a m u ltiv a ria te m o d e l, n o sig n ific a n t d o se re sp o n se w a s o b se rv e d b e tw e e n total g e n ita l a p p lic a tio n s o f talc a n d o v a r ia n c a n c e r risk . B e ca u se the "tra n slo c a tio n " th e o ry a ss u m e s a n o p e n g e n ita l tract, w e th e n e x c lu d e d a p p lic a tio n after tu b a l lig a tio n o r h y ste re c to m y , b u t o b se rv e d n o a p p re c ia b le c h a n g e in th e d o s e re s p o n s e . F u r th e r re s tr ic tin g talc e x p o su re to m o n th s w h e n th e w o m e n w e re lik e ly to be o v u la to ry ', w e o b se rv e d a s ig n ific a n t d o s e re s p o n s e , su c h th at w o m e n w ith a n in tact g e n ita l tract a n d m o re th a n 10,000 a p p lic a tio n s d u r in g o v u la to r y c y c le s h a d n e a rly a th re e fo ld in cre a se in risk fo r o v a ria n can cer. S o m e a d d itio n a l e v id e n ce th a t m ig h t su p p o r t a n in te r a c tio n w ith o v u la tio n in c lu d e s s tr o n g e r a ss o c ia tio n s in w o m e n w ith re g u la r p e n o d s a n d m id -c y c le p a in . C r a m e r et a l4 s u g g e s te d th a t talc c o n ta m in a tio n a r o u n d th e tim e o f o v u la t io n m ig h t le a d to th e in c o r p o r a t io n o f ta lc p a rtic u la te s in to in c lu sio n c y sts th a t m a y fo rm w ith o v u la tio n . E x p e n m e n tso n fo re ig n -b o d y tu m o rig e n e sis h av e s h o w n th at im p la n ta tio n o f fo re ig n b o d ie s in to th e lu m e n s o f e p ith e lia l-lin e d o r g a n s p r o v id e s a fa v o r a b le e n v ir o n m e n t f o r c a r c i n o g e n e s is . 22 A lt e r n a t iv e l y , M o s t a f a e t a l 16 s p e c u l a t e d t h a t f o r e i g n - b o d y e x p o s u r e m ig h t p ro d u c e co rtical "g r a n u lo m a s " w h o s e lin k to stro m a l h y p e ra c tiv ity (a n d h o rm o n a lly re la te d ca n c e rs) is a r g u e d i n o l d e r li t e r a t u r e . 23 W e o b s e r v e d th a t th e talc a s s o c ia t io n w a s s tr o n g e s t in w o m e n w ith e n d o m e trio id o r b o rd e rlin e o v a ria n tu m o rs. H o w e v e r , a n e a rlie r s t u d y b y C r a m e r et a l4 re p o rte d n o su c h v a n a tio n in risk b y h isto lo g ic su b - ty p e . It w a s n o te d th at a g re a te r p r o p o r t io n o f w o m e n w ith e n d o m e trio id tu m o rs th a n w ith o th e r h isto lo g ic ty p e s o f o v a ria n c an c e r re p o rte d m o re th a n 10,000 life tim e a p p lic a tio n s o f ta lc d u r in g o v u la to r y c y c le s w h ile h a v in g a n in ta ct g e n ita l tract (34 v e r s u s 1 6 % ). A lt h o u g h th is m a y e x p la in in p a r t th e s t r o n g talco v a ria n ca n c e r a sso c ia tio n n o te d in w o m e n w ith e n d o m e t r io id t u m o r s , it d o e s n o t e x p la in th e s t r o n g a sso c ia tio n n o te d in w o m e n w ith b o rd e rlin e o v a n a n tu m o r s , o f w h o m o n ly 1 3 % r e p o r t e d lo n g -t e r m talc e x p o su re . T h is v a ria tio n in risk a m o n g h isto lo g ic su b ty p e s m a y re fle ct a c h a n c e f in d in g o r a n e e d to e x a m in e e n d o m e trio id a n d b o rd e rlin e tu m o rs m o re c a re fu llv fo r e v id e n c e o f a fo r e ig n -b o d y effect. A n u n u s u a l o b se rv a tio n w a s th e s tro n g a sso c ia tio n b e tw e e n ta lc u s e a n d o v a r ia n c a n c e r in th e s u b g r o u p o f w o m e n w ith o n e o r t w o p r e g n a n c ie s b u t a lo w e r r isk in w o m e n w ith e ith e r n o c h ild re n o r th re e o r m o re c h ild re n . A lt h o u g h c h a n c e m a y b e th e m o s t lik e lv e x p la n a tio n fo r th is fin d in g , w e w o n d e r w h e th e r th is p e c u lia r in te ra c tio n m ig h t re fle ct, in p a rt, a n effect o f p re g n a n c y o n th e d e g re e o f o p e n n e s s o f the c e rvica l os a n d its a b ility to a llo w tra n slo c a tio n o f v a g in a l p a rtic u late s. T h e p a r o u s c e rv ix h a s a la rg e r o s th a n the n u llip a r o u s ce rv ix , a n d th is c o u ld e x p la in th e gre a te r risk in w o m e n w ith o n e c h ild c o m p a re d w ith w o m e n w ith n o liv e b irth s. A t th e o th e r e x tre m e , m u ltip le p re g n a n c ie s m a y o ffe r o th e r p ro te c tiv e m e c h a n ism s (su c h a s re d u c e d o v u la tio n a s d isc u sse d a b o v e ) that c o u ld o u t w e ig h t h is e ffect. C le a r ly , th is is a v e ry sp e c u la tiv e h y p o t h e s is b u t o n e th a t m ig h t b e te ste d in e x p e r im e n t a l s t u d ie s (ie , r e p e a t in g th e v a g in a l talc e x p e rim e n ts in n u llip a r o u s a n d p a r o u s m ice o r p ri m ates). N o n c a u s a l e x p la n a tio n s a re p o ss ib le in a n y e p id e m i o lo g ic re se arch . W e c a n n o t ru le o u t th e p o ssib ility of d iffe re n tia l o v e r - o r u n d e r -r e p o r t in g o f talc e x p o s u r e in o u r case s a n d c o n tro ls, e sp e c ia lly in th o se w ith re p ro d u c tiv e e v e n ts th a t e n h a n c e O R s . In a d d itio n , th o u g h w e w e re su c c e ssfu l in in te rv ie w in g 6 9 % o f e ligib le o v a ria n can ce r c a se s a n d 8 1 % o f e ligib le c o n tro ls c o n tacted, w e c a n n o t a sse ss w h e th e r th e case s a n d c o n tro ls n o t in te rv ie w e d c o u ld h a v e se le c tiv e ly d iffe re d in th e ir re p ro d u c tiv e c h a ra c te ristic s o r in th e ir u se of ta lc -c o n ta in in g p o w d e rs. B e c a u se o u r a sso c ia tio n s are b a se d u p o n re sp o n se s fro m p a rtic ip a tin g case s a n d c o n tro ls, th e v a lid ity o f o u r re su lts d e p e n d s u p o n the a ssu m p tio n th at re sp o n d e n ts a n d n o n -re sp o n d e n ts w e re s im ila r w ith re sp e c t to ta lc a n d o th e r re le v a n t e x p o su re s, o r th at th e m a g n itu d e of a n y re sp o n d e n tn o n -re sp o n d e n t d iffe re n ce w a s sim ila r for case s a n d c o n tro ls. B e c a u se th e in te rv ie w p ro v id e d th e o n ly so u rc e o f "e x p o s u r e " in fo r m a tio n , w e w e re u n a b le to a sse ss the lik e lih o o d o f th is a ss u m p tio n . T h e e xten t of th is b ia s, h o w e v e r , is lik e ly to b e s m a ll b e c a u se the re p ro d u c tiv e c h a ra c te ristic s a n d h is to r y o f talc e x p o su re in p a r tic ip a tin g c a se s a n d c o n tr o ls are, fo r the m o st p art, re a s o n a b ly c o n s is te n t w ith e a rlie r e p id e m i o lo g ic stu d ie s o f o v a ria n can ce r. F in a lly , in o u r a tte m p t to p re se n t th e m o st a cc u ra te O R s , w e m a d e a va rie ty of a d ju stm e n ts to a c c o u n t fo r th e c o n fo u n d in g in flu e n ce o f facto rs a sso c ia te d w ith b o th o v a ria n can ce r risk a n d VOL, 80. N O 1. |LLV 1992 H arlow et al Talc and O varian Cancer 25 1 talc exposure. Nevertheless, we cannot rule out the presence of other unknown factors that might have influenced, in part, our observed associations. Because the overall association between genital use of talc and ovarian cancer remains weak, it is unlikely that this exposure-disease pathway is the principal one involved in ovarian cancer etiology. We have previ ously discussed the role of dietary and metabolic factors in a model for ovarian cancer involving gonad otropin stimulation of failing ovaries.8 Even if an etiologic association were to pertain in the subgroups of daily users or users with more than 10,000 applica tions during ovulatory months, we calculate that by applying these ORs to the exposure rate among cas es,24the proportion of ovarian cancer incidence attrib utable to this level of talc exposure is about 10%. Nevertheless, given the poor prognosis for ovarian cancer, any potentially harmful exposures should be avoided, particularly those with limited benefits. For this reason, we discourage the use of talc in genital hygiene, particularly as a daily habit. References 1. Graham J. Graham R. Ovarian cancer and asbestos. Environ Res 1967;1:115-26. 2. Parmley TH, Woodruh JD. The ovarian mesothelioma. Am J Obstet Gynecol 1974,120.234--il. 3. Longo DL, Young RC. Cosmetic O k and ovarian cancer Lancet 1979x:349-51. 4. Cramer DVV. Welch VVK, S c u lly RE, Wojaechowski CA Ovarian cancer and talc Cancer 1982;50:372-6 5 VVhittemore AS, Wu ML. Paffenbarger RS, et a]. Personal and environmental charactensncs related to epithelial ovarian cancer. II Exposures to talcum powder, tobacco, alcohol, and coffee. Am J Epidemiol 1988.128:1226-40. 6. Harlow BL. Weiss N'S. A case-control study of borderline ovarian tumors The influence of penneal exposure to talc. Am J Epide miol 1969,130.390-4. 7. Booth M. Beral V, Smith P. Risk factors for ovarian cancer: A case-control study Br J Cancer 1989;60:592-8. 8. Cramer DW, Harlow BL, Willett WC, et al. Galactose consump tion and metabolism in relation to the risk of ovanan cancer. Lancet 1989;ii:66-71. 9. Breslow- N"E, Dtv N"E Statistical methods in cancer research. Vol 1. The analysis of case control studies. 1ARC scientific publication no 32. Lyon: international Agency fot Research on Cancer. 1980. 10. Hamilton TC. Fox H. Buckley CH. Henderson WJ, Griffiths'K. Effects of talc on the rat ovary. Br J Exp Pathol 1984;65:101-6. 11. Sampson JA. The development of the implantation theory for the origin of endometriosis. Am J Obstet Gynecol 1940;40:549-57. 12. Henderson WJ, Hamilton TC, Baylis MS, et al. The demonstra tion of the migration of talc from the vagina and posterior uterus to the ovary in the rat. Environ Res 1986;40:247-50. 13. Egli GE, Newton MD The transport of carbon particles in the human female reproductive tract. Fertil Steril 1961;12:151-5 14. Venter PF, Iturralde M. Migration of particulate radioactive tracer from the vagina to the peritoneal cavity and ovaries S Afr Med J 197955:917-9. 15. Henderson WJ, Joslin CAF, Turnbull AC. Griffiths K. Talc and cardnoma of the ovary and cervix. J Obstet Gynaecol Br Commonw 1971:78266-72. 16. Mostail SAM, Bargeron CB, Flower RW, Rosenshein NB, Parm ley TH, Woodruff JD. Foreign body granulomas in normal ova ries. Obstet Gynecol 1985;66:701-2. 17. Griffiths K, Henderson WJ, Chandler JA, Joslin CAP. Ovanan cancer Some new analytical approaches. Postgrad Med j 1973:49: 69-72. 18. Wehner AP, Hall AS. Weller RE, Lepel EA. Schirmer RE. Do partides translocate from the vagina to the oviducts and beyond? Food Chem Toxicol 198523:367-72. 19 Greenland S. Quantitative methods in the review of epidemio logic literature. Epidemiol Rev 1987;9:1-30. 20. Cralley LJ, Key MM, Groth DH, Lainhart WS, Ligo RM Fibrous and mineral content of cosmetic talcum products. Am Ind Hyg Assoc J 196829:350-4. 21. Rohl AN, Langer AM, SeLkoff IJ, et al. Consumer talcums and powders: Mineral and chemical characterization J Toxicol Envi ron Health 19762255-84. 22. Brand KG, Johnson KH, Buoen LC. Foreign body tumorigenesis. Crit Rev Toxicol 1976;4:353-94. 23 Woll E, Hertig AT, Smith GVS, et al The ovary in endometnal cardnoma with notes on the morphological history of the aging ovary. Am J Obstet Gynecol 1948:56:617-33. 24. Rothman KJ Modern epidemiology Boston: Little, Brown, 1986 35-40. Address reprint requests to: Bernard l . Harlow, PhD Obstetrics and Gynecology Epidemiology Center Brigham and Women's Hospital Harvard Medical School 221 Longwood Avenue Boston, M A 02115 Recerved January 6, 1992. Received in revised form March 23, 1992. Accepted March 23, 1992. Copyright C 1992 by The American College of Obstetricians and Gynecologists 26 H arlow et al Talc a n d O va n a n C ancer O bstetrics b Gynecology  HARTMAN LIBRARY TEL:908-707-9860 v*box-ow e onooloov 45, 2ft-25 (3992) Jul- 13 '92 13:24 No.004 P.02 ~ T A tC L _ Mineral Fiber Exposure and the Development of Ovarian Cancer K a r i n A. R o k k n b l a t t , Ph .D .,*-' Morses S z k l o , M .D ., D r.P.H.,* Neil. B. R o x k n s h e i n , M .D .' -T 'D epartm ent o f E pidem iology. The John* H opkins School o f lly g h n e and Public H ealth, Baltim ore, M aryland 21218,; and t D epartm ent o f G ynecology end O bstetrics. The John* H opkins H ospital, Baltim ore, M aryland 21218 R e c e iv e d J u ly IQ, 3991 A hospital-baaed ra te control study of the association between fiber exposure and th e development of epithelial ovarian cancer wee perform ed a t th e Johns Hopkins H ospital in B altim ore, M ary land. G enital and respiratory fiber exposures were ascertained from incident cases ( N = 77) and age-race m atched controls (V 46) using a structured questionnaire. Cases were ascertained between 1981 and 1985- An increased risk was observed for ex posure tn talc on sanitary napkins (OR 4.79, 95%Cl, - 1.29 17.79), genital fiber exposure from different sources for a long (cum ulative exposure *37.4 years) length o f tim e (OR 2.35. 95%Cl " 0.93-5.80), and occupational fiber exposure in relatives (OR * 2.81, 95%Cl 0.90-8.75). A negative association was observed for antecedent tubal ligation (OR = 0.15, 95% Cl * 0.027-0.88). Findings from this study should be confirm ed in larger iaveatigalkxu. c met a.mi s Im. INTRODUCTION Occupational exposure to asbestos has been identified ak a risk factor for ovarian cancer in several studies [ l 3], Similarly, fiber-containing substances, such as tale, have also been implicated as risk factors 14-9]. A matched case-control study of epithelial ovarian cancer was con ducted, in which the role of both genital and respiratory source of fiber was assessed, to confirm these findings. STUDY POPULATION Cases and controls were ascertained from the Johns Hopkins Hospital between 1981 and 1985. This analysis is restricted to the 77 cases who were matched to 46 hospital controls and treated for conditions other than gynecologic or malignant diseases. Originally 140 newly diagnosed cases of epithelial ovarian cancer who met the 1 Prcient sddrea: Department of Health and Safety Sunlici, tinivenity of Illinois at Urhana--Champaign, 120 Huff Hall, 12U6 S. Fourth St., Champaign, IL 61820 (Correspondence ui Or. Roaenhlatt at thia address), eligibility criteria were ascertained from the Johns Hop kins Hospital. One hundred eight (77.1%) of these cases were successfully interviewed. These castes were diag nosed within 6 months of admission, were pathologically confirmed hy examination of the ovaries, were admitted as in-patients for treatment or diagnosis, and were resi dents ol the United States. Controls were in-patient fe males without gynecologic or malignant conditions who were initially matched to cases by age (within 5 years), race, and date of diagnostic admission (within 1 year). Since it was difficult u> find controls meeting ail of the matching criteria, a control could not be found for all cases. Unmatched cases were therefore matched a pos teriori to controls, to form matched triplets of 2 cases and 1 control. A posteriori matching was performed within the same 5-year interval, by race and by closest date of diagnostic admission. The matching procedure allowed for the inclusion of 77 cases in the study. There were 46 matched sets, of which 31 consisted of 2 cases and 1 control. No matched control could be found for 13 cases, which were excluded from thf analysis. MliJ'HODS Data were ascertained primarily from a questionnaire that was administered to participants both by telephone and in the hospital. Information on previous abdominal and gynecologic operations was also ascertained from medical records. The questionnaire elicited information on the presence and length of genital fiber and respiratory fiber exposure, reproductive factors, estrogen use, family history of can cer in first-degree relatives, and other contraceptive use. Descriptions of the questions used to ascertain fiber ex posure are listed in Appendix 1. Fiber exposure was de fined as exposure to asbestos, talc (which may contain asbestos), and fiberglass. An attempt was made to estimate the overall effect of 0O#U-M i/92 11.50 Copyritht 6 1442by Araikmk Prca, lac. AM r i | t a ol ttptoOucOtm in any (m b m o v e d . 2ft HARTMAN LIBRARY TEL :908-707-9860 Jul 13*92 13:25 No.004 P.3 Fib e r s a n d o v a r ia n c a n c & r 21 " dooc" or length of genital and respiratory fiber exposure. This was accomplished by adding the num ber of years of each type of genital o r respiratory exposure from all sources. Since many of these exposures occurred at the TABLE 1 Distribution of Matched I and Contrais Caaes C ontrol* sam e tim e, these variables should he considered as crude N %N% m easures of dose rather than the true length o f exposure. C onditional logistic regression [10J w as used to d eter mine the strength of the association. O dds ratios were calculated to describe the relationship of genital and res Age (yesrt) Leas tttan 30 3 0 -3 9 4 0 -4 9 3 1 11 3.9 1.3 14.3 2 4.4 1 2.2 8 17.4 pirato ry fiber exposure to ovarian cancer. T h e o d d s ratio 50-39 21 27.3 lit 21.7 is an estim ate o f th e relative risk for relatively rare dis 6 0 - 35 45.4 21 45.6 eases such ax ovarian cancer and, in this paper, is referred to as the relative risk estim ate (R R ). 7 0 -7 9 SOur higher Total 3 1 77 6.S 3 1.3 1 46 6.3 2.2 T he following variables were assessed as potential confoundcrs: Race white 70 90.9 41 89.1 Tobacco use Num ber of cigarettes per day Black Total 7 9.1 5 10.9 77 46 O vulatory time period Num ber of pregnancies C ancer in m other or father Obesity I year prior to diagnosis Obesity 20 yean prior to diagnosis Obesity at highest weight during the 20 years prior to diagnosis Obesity according to average weight during the 20 years prior to diagnosis H u sb an d 's and su b ject's education Previous cancer If a potential confoundcr changed the relative risk es tim ate associa ted with a fiber ex posure by m ore than 15%, it was retained in the m ultivariate m odel. Confounders were sequentially added to m ultivariate modeix, depend ing on the level with w hich they changed the odds ratio of a fiber related exposure. Interaction between fiber exposures and potential confounders was evaluated by com paring the deviance of models with and w ithout the interaction term [12]. M arital status Religion RESULTS Use of oral contraceptives Use of contraceptive foams, cream s, and jellies by them selves Use of an IUD On the basis of the strength of their relative risk es tim ates and the difference in frequency with which they w ere observed in caxex and controls, th e follow ing vari Age and Racial Distribution Table 1 lists the age and racial dixtributioa of cases and controls. Most eaxes and controls w ere in the age groups 40 to 69. It wax difficult to obtain controls that did not have a chronic disease, resulting in a low er than expected num ber of controls. ables w ere identified ax potential confounders: m easures Diagnoses of Controls o f o besity, socioeconom ic status (su b ject's ed ucation; RR - 0.5, 95% C l = 0.2-1.1), and religion (Jewish, RR = 2.9, 95% C l = 0 .6-13.8; C atholic, R R - 0.5, 95% Cl - 0 .2-1.5), reproductive status (total live births, 1-2, R R - 0.7, 95% Q = 0.2-1.9; live births > 2 , R R - Controls were selected so that they did not have their primary diagnostic condition for more than 1 year. Tabic 2 lists the primary diagnosis o f the controls included in the study. 0.4, 95% C l " 0 .2 -1 .3 ), and oral contraceptive use (O R Genital Fiber Exposure - 0.9, 95% C l ~ 0.3-23.0). O besity 1 year p rio r to diagnosis (R R - 2.3, 95% Cl D ifferent sources o f genital fiber exposure w ere ex * 0 .9 -6 .1 ), obesity 20 y e a n prior to diagnosis (R R = amined (Table 3). E xposure from any of the potential 2 .1,9 5 % C l - 0.7--6.9), and obesity according to highest sources of genital fiber was highly prevalent (91.1% in w eight durin g th e 20 y e a n prior to diagnosis (K R - 1.5, controls) and not found to be related to ovarian cancer 95% C l - 0 .7 -3 .2 ) w ere also found to be p o tential con- (R R = 1.0. 95% C l *- 0 .2 -4 .0 ). Since it wax felt that founders. O besity was defined as a h eight/w eight index this index did not accurately characterize cum ulative ex g reater th an th at exhibited by w om en in the H5th p er posure, the median length o f ex p o su re from all genital centile [11] of the population. sources (with the tim e since tu b al ligation subtracted) wax HARTMAN LIBRARY T E L :908-707-9860 Jul 13*92 13:25 No.004 P.04 22 ROSENBLATT. SZK tO . AND ROSHNSHEIN TABLE 1 nuributioai at Matched Control* by Primary Diagnorit asbestos showed an increased risk of lung cancer |16|, asbestos has been observed in cosmetic face powdeni [17 - Diagnosis of restricted control* N % 19|, and papillary growth, has been observed after im plantation of asbestos [20] and talc [21] into the peritoneal Infectious and parasitic disease* Disease* of the digestive system Opthamologic disorders Diseases of the circulatory system Symptoms, signs, and iU-dehncd condition* Disease* of uiuscidoskclctti and connective tissue Diseases of the endocrine system nr metabolic or immunoioaic disorder* Diseases of the respiratory system Benign neoplasms injury and poisoning Total 1 2.2 cavity. 6 1.1.0 We found an increased relative risk (4.8) for ulc use M 11 8 32.6 23.9 17.4 on sanitary napkins with a smaller effect for geniud bath talc exposure (RR - 1.7). This is in accordance with the 1 2.2 original finding of a significant increased risk for perineal talc exposure (RR = 1.9. 95% Cl = 1.3-2.9) by Cramer 1 2-2 ei til. [4]. Preliminary findings from a Chinese study also 1 1 1 2.2 2.2 2.2 suggest that perineal application of talc-containing dusting powder increased the risk of epithelial ovarian cancer (RR - 3.9. 95% Cl - 1.1-13.8) [5]. A nonsignificant effect 46 100.00 for genital talc exposure (on genitals, sanitary napkins, or underwear) was detected in the study of Hartge et ul. |A) (RR -- 2.5, 95% Cl - 0.7-10.0). Whinemore et ul. used to categorize the dose of genital liber exposure. A |7| detected an increased risk (RR = 1.4, P - 0.06) for relative risk estimate of borderline significance was seen perineal exposure. In a study of borderline ovarian tu for exposure above the median length of time (37.4 years. mors, an increased risk was also observed with talc ex RR = 2.4, 95% Cl = 1.0-5.8). A significant negative posure from use on sanitary napkins (RR - 1.9, 95% association was observed for previous tubal ligation (RR Cl = 0.9-6.9, 8). - 0.2, 95% Cl - 0.03-0.9). We also investigated the relationship with cumulative A history of several gynecologic and abdominal oper duration of ail forms of genital fiber exposure (with the ations, given by responses u> the questionnaire, was ex time sinee tubal ligation subtracted). The positive asso amined to determine if exposure to talc from surgeon's ciation (RR - 7.35) in our study for exposure longer gloves increased risk |I3]. No statistically signiticunt re than the median length of time contrasts with the lack of lationships were detected but, with the exception of ovar association with duration observed by Whiucmorc et al. ian biopsies, relative risk estimates were generally below [7], Whinemore et at. [7], however, did observe a positive 1. An attempt was made to combine information from dose-response relationship with frequency of exposure the questionnaire and medical records, with no important (1-20 times per month, RR = 1.3, 95% Cl - 0.8-2.0; changes in the relative risk estimates. >20 tunes per month, RR = 1.4, 95% Cl > 0.9-2.2). Moderately elevated relative risk estimates, whieh were Although Booth et a!., [9] did not observe a significant not statistically significant, were observed with use of con (P - 0.05) trend of increasing risk'with more frequent doms, diaphragms (when powder was used), and genital use, weekly genital talc use (RR = 2.0, 95% Cl - 1.3 bath talc. The level of association observed with exposure 3.4) was associated with an increased risk of ovarian to talc on sanitary napkins (RR = 4.8, 95% Cl - 1.3 cancer. 18.0) was significantly greater than unity. The results of our study and others suggest that genital Respiratory fib er Exposure fiber exposure may be associated with an advene effect [4-8] but further study is needed to determine if this Numerous sources of respiratory fiber exposure were relationship is causal in nature. evaluated (Table 4): the only such source showing an The present study also showed a significant negative important effect was occupational fiber exposure in rel association with tubal ligation (RR - 0.15). Harlow [22] atives (RR - 2.8, 95% Cl - 0.9-8.8). This relative risk (RR = 0.3, 95% Cl = 0.3-1.1), Booth et al. [9] (RR estimate was calculated after three cases who were both - 0.2, 95% Q - 0.1-0.6), Mori et al. |23] (RR = 0.4, exposed to asbestos themselves and received asbestos ex 95% Cl - 0.2-1.0), W hittcm ore a al. [7] (RR = 0.6, posure from their relatives were excluded. P - 0.07), and Irwin et al. [24] (RR - 0.7, 95% Cl = 0.5-1.0) also observed negative associations. Although DISCUSSION the findings of our study d iffe r from the positive associ ation observed by Koch el al. [25] the com panion rates Genital talc exposure has been proposed as an e t i o l o g i c [2A] used in Koch's cohort study may have been agent of ovarian cancer [14] because talc was o b s e r v e d undcrcstimated. more frequently in cancerous ovaries [15] than in non- Tubal ligation may protect against ovarian cancer by canecrous ovaries, occupational studies of talc-containing inhibiting the carcinogenic action of talc through bkickagc .HARTMAN LIBRARY [ ' TEL :908-707-9860 Jul 13*92 13 :26 N o .004 P .05 FIBERS AND OVARIAN CANCER 23 T A R I.E 3 Num b and Frequency Distribution of C u te and Control with Odds Ratio* for fim im pggr FrnrwM* and Other Related Variable* hxpoture interval Ocnitai fiber nsc Length o f use o f gciuial fiber* (yean ) (M edian o f c u e * and coniro b ) Ovarian biopsie* U nilateral oophorectomy Tubal ligation H ytrerecrom y Condom ute Diaphragm use with ' powder O cnitai hath talc Sanitary naplun with talc exposure A ttribute Ye* No M itrili* *3 7 .4 < 37.4 M itring M edian Ye* No Yc* No Yea No Yea No Yet No M itring Yes No M itring Ye* No M atin g Yet No M itring Case* -- N 67 87.(1 10 13.0 0 39 55.7 31 44.3 7 41.9 6 7.8 71 92.2 8 10.4 69 89.6 4 5.2 73 94.8 19 24.7 58 75.3 35 49.3 37 50.7 S 14 18.9 (0 81.1 3 22 28.9 54 71.0 1 21 .10.11 49 70.0 7 Controls N '% 40 88.0 5 11.1 1 16 39.0 15 61.0 5 24.0 3 6.5 43 93.5 5 10.9 41 89.1 6 13.0 40 H7.fi 12 26.1 34 73.9 22 51.2 21 48.8 3 5 11.4 39 8 8 6 2 8 18.6 35 81.4 3 6 13.6 38 86.4 2 O dd* ratio 1.0 2.4 9 5 confide* interval 0 .1 - . * t.o -s .r 1.1 0.3 4.4 0.8 0 .2 -2 .5 * 0.2 0 .3 -0 .9 ** 0.7 o.3 t.r 1.6 0 .6 -3 .9 ~ 3.0 0.8-10.8*-* 1.7 0 .7 -3 .9 4.8 1.3-17.8' * A fte r subtraction of the time since tubal ligation. for those who had ligation, * Adjusted for number of Uve births, ' Adjusted for religion. " Adjusted for yean of education on subject. ' Adjusted for highest weight 20 years prior Ut diagnosis. ' Adjusted fot highest weight 1 year prior to diagnosis. of the fallopian tube or through a "screening" effect [27], The effect of antecedent tubal ligation should be eval uated in future studies of ovarian cancer to determine if it negative axnodaoon is consistently observed and to de termine the reason for this. Several cohort studies of women with respiratory ex posure to asbestos detected an increased relative risk for ovarian cancer [1-3]. Wc elicited information about em ployment by relatives in occupations witii asltcslos or fi berglass exposure [28] and the relative risk was found to be elevated (RR --2.8). In a previous ease--control study of ovarian cancer, the relative risk for occupational as bestos exposure in relatives was not elevated [29], al though the authors did not describe how the asbestos exposure was ascertained. Our initially suggestive findings regarding asbestos or fiberglass exposure in relatives should be evaluated further in additional studies. In summary, our study shows that the development of ovarian cancer may be associated with genital fiber ex posure (especially talc on sanitary napkins) and occupa tional exposure to fibers in relatives. Given its small sam ple size and the potential selection bias stemming from inclusion of patients from only one hospital, further re search needs to he performed in order to confirm our findings. APPENDIX 1 Questions Asked to Ascertain Fiber Exposures Genital Fiber Exposure 1. Have you had any of the following operations prior ro your hospitalization in ____ ? --Biopsy or removal of part of an ovary --Removal of one ovary --Removal of uterus (hysterectomy) , HftRTMPN LIBRARY TEL:908-707-9860 Jul 1 3 '92 13:27 No.004 P.06 24 R O S E N B LA TT, SZJCLO, A N |) R O SEN S H E IN TABLE 4 Number and Frequency Distribution of C asa and Controls with Odds Ratios tor Trlrrtrit Charecterwtic* n-- a to Respiratory Fiber Exp e w e Exposure interval Respiratory fiber exposure Cosmetic face powder use Insolation insulted at residence Living in tha vicinity of a fiScr-cmiUmg industrial esubtubm aM Applied bath talc to body (respiratory exposure) Kibcr exposure in relatives Use o f speckling and taping compounds A ttribute Yes No Yes No Missing Yes No Missing Yes No Missing Yes No Missing Yes No Ye* No Missing Cates N 16 69 89.6 fi 10.4 3H 50.7 37 49.3 2 30 39.3 46 60.5 1 7 9.2 W 90.fi 1 47 61.K 29 3H.2 1 18* 24.3 56 7S.7 20 29.(1 49 71.0 X C o n tro l* M% 41 89.1 5 10.4 24 54.3 20 45.4 2 17 37.8 2a 62.2 i 4 8.9 41 91.1 I 24 55.fi 19 44.2 3 5 10.9 41 89.1 14 31.1 31 68.9 1 fM ds ratio 1.3 1.1 1.2 J.O 1.6 2.8 1.6 4 3 confide Mental 0.3-3.6* n .4 -2 .7 * 0.3 4.6' 0 .3 -3 .4 0 .6 -2 .7 ' 0 .9 -8 .8 0 .5 -3 .4 '-' * Adjusted (or highest weight 1 year prior to diagnosis. * Adjusted for years of education on subject. ' Adjusted for number o f live births. 4 Three case* were excluded became they had been directly exposed to fibers and had a household member who Had been occupationally exposed to fibers. Fiber exposure included asbestos, talc, or fiberglass exposure. ' Adjusted for religion. --Tubal ligation --Any other abdominal operation (These items were confirmed by examination of medical records) 2. Did you use a diaphragm prior to your hospitalization in ___ .? If yes: Did you use a powder to dust and dry the diaphragm? If yes: Was it talc? 3. Did you and your sexual partner ever use a condom prior to your hospitalization in ____ ? 4. Have you regularly applied talcum powder to your body after bathing or for other reasons prior to your hospitalization in ____ ? Did you commonly apply the talcum powder to your genital area? 5. Have you used talc on sanitary napkins or any other sanitary products used during your menstrual period prior to your hospitalization in ____ ? Respiratory fib er Exposure 1. Prior to your hospitalization in ____had you ever regularly used cosmetic face powder? 2. Have you ever had insulation installed in a place that you lived in prior to your hospitalization in -------? 3. Have you ever lived in the vicinity of a shipyard, an asbestos or talc mine, or an asbestos, tale, or fiberglass processing plant prior to your hospitalization in ____ ? 4. Have you regularly applied talcum powder to your body after bathing or for other reasons prior to your hospitalization in ____ ? Did you commonly apply the talcum powder to your face, upper torso, or legs and feet? 5. Have you or anyone who has ever lived in your household (including your husband) been employed in any of the following industries prior to your hospitali zation in -------? (a) Installation or removal of insulation materials (b) Brake lining manufacture (c) Automobile repair involving brake repair (d) Rooting using asbestos materials (e) Asbestos milling or mining (f) Asbestos textile or paper manufacture (g) Building construction (h) Other industries where asbestos is used (i) Talc mining and milling (j) Other industries where talc is used (k) Fiberglass or mineral wool manufacture (l) Other industries where fiberglass or mineral wool was used HR-RTMflN LIBRARY V TEL:908-707-9860 Jul 13*92 13:27 No.004 P.07 FIB ER S A N O V A R IA N C A N C E R 25 6. Did you ever use speckling end toping compound? If the respondent answered yes to any of these ques* tions, they were asked when the exposure started and topped. If they could not answer this question, they were asked how long the exposure occurred. REFERENCES 1. Achcsoa. E . D ., G ardner, M . J ., Pfpparh. E . C ., and G rim e, 1- F. M ortality o f two group 01 women who manufactured gar mark fmm chrysolite ami auudoU le asbestos: A ft year foilow-up, Ur. J. Ind. Mod. 29, 344-348 (1982). 2. Ncwhousc, M . L ., B erry, G ., W agner, J. C ., and Turok, M . L . A rtudy of the m ortality o f fem ale asbestos workers. Br. J. ind. Med. 29, 134-141 (1972). 3. W ignail, B . K ., and Fox, A . J. M ortality o f female gar mark i f semhicr, Ur. J. Ind. Med. 39, 34-38 (1982). 4. Cram er, D . W ., W elch, W . R -, Scully, K . K ,, and Wojcieehowski, C . A . Ovarian cancer and talc: A case control study. Cancer S t, 372-376 (1982). 3. Chen, Y .. and W u, B. Z . Risk factors for epithelial ovarian cancer in China-- A ease-control study (abstract). A m . J. Epidemiol. 132, 777 (1990). 6, Hartge, P .. H oover, R ., Lcsher, I - F ., and McGowan, L. Talc and ovarian cancer, J. A m . Med. Assoc. 254, 1884 (1983). 7. W hiuem ote, A . 8 ., W u, M . L ,, Paffcnbarger. R . S .p Jr., Sarics, D . L , Kam pcrt, J. R ., Groswa, S .. Juii, D . L ., Ballon, S ., and Hendrickson. M . Personal and environmental characteristics related to epithelial ovarian cancer. II. Kxpoaurcs to talcum powder, toh.icco. alcohol, and coffee, A m . J. Epidemiol. 128, 1228-1240 (1988). M. Harlow, H. I.., and Wei, N . S. A case control study o f borderline Ovarian tumors. The influence o f perineal exposure 10 talc. A m . J. Epidemiol. 3d, 39(3-394 (1989). 9. Dooth. M ., Reral, V ., and Sm ith, P. Risk (acton (orovarian cancer: A case control study. Dr. J. Cancer 66, 392-598 (1989). ill. Smith. P. O ., Pike, M . C .. H ill, P. A .. Breslow, N. H.., and Day. N , E . M ultivariate conditional logistic analysis of stratum-matched case-control studies, A p p i. Slot. 36, 19(3-197 (1981). 11. National Center for Health Statistics. Obese and overweight adults in the United States, in National Health Survey Series No. 230, V o l. 11, H yatlsvillc. M D . (1983) 12- Breslow. N . E .. and D ay. N . E. Statistical methods in cancer re search. in ARC scientific publications No. 32. International Agency (or Research on Cancer, l.yon (1980). 13. Tolbert, T . W ., u d Braw n, J. L_ Surface powders cm MMpeat power. Arch. Surg. 11, 729-732 (WHO). 14. Inagy, D . L .. sad Young, R. C . Cosmetic talc and ovarian cancer, Lamcm t , 344-331 (1979). 13. Henderson, W . J ,, Joslia, C . A . K , TUrnbult, A . C , and O riflW w , K . Talc and carcinoma o i the ovary and o n b , J. OhettL iTj n w rrf Ur. C om m tm w .'n, 266 272 (1971). 16. International Agency for Research on C r o w . S ttea and 10UM Ml* icaiea, in IARC monographs on th* evaluation o f the carcinogenic risk o f chemicals to human, V ol. 42, pp. 1SS-224, Lyons, France. (1987). 17. C rallcy, I.. J ,, Key. M . M ., Crash, D . H ., Iaaah art. W . S ,, and I Jgo. R . M . Fibrous and mineral content o f cosmetic talcum prod ucts. A m . Ind. Ilyg. Assoc. /. 26, 330-354 (1966). 18- Paoietli. L ., Caiazza, S-, D onslli, C ., and Pncchaari, F . Evatuation by electron microscopy techniques of aahealoe cuolamination in industrial, cosmetic, and pharmaceutical talcs. Reg. Toxicol. Phar macol. 4 , 222 233 (1984). 19. Kohl. A . N ,, U n g e r, A . M ., Selikotf. I. J ., T urdini, A .. Klim entid it, R ., Bowes, I) . K ,, and Skinner, D . L . Consumer talcums and powders: M ineral and chemical characterization, J. TnxicnL E n viron. Health 2 , 235-285 (1976). 20. Graham , J ., and Graham , R . Ovarian cancer and asbestos, Environ. Rat. I, 113-128(1967). 21. Ham ilton, T . C , Fox, I I . , Rueklcy. C. H ,, Henderson, W . J ,, and G riffith , S. K . Kffcct of talc on the rat ovary, Ur. J. E xp. Pathol. <5, 101-1116(1984). 22. Harlow , U. I.. An epidemiologic study o f borderline Ovarian tu mors, Doctor of Philosophy Dissertation. University o f Washington (1987). 23. M o ri. M ., Harahuctii, I. , M iyake. H ., Casagrande, J. T ., Hender son. H. K ., and Rosa, R . K . Reproductive, genetic, and dietary risk factor feu ovarian cancer, Am. J. Epidemiol. 126, 771-777 (1988). 24. Irw in , K .. Weiss, N ,, Lee. N ., and Peterson, H . Tubal sterilization, hysterectomy, and the subsequent occurrence o f epitheMal ovarian cancer (abstract), Am. J. Epidemiol. 132, 777 (1990). 15. Koch, M .. Slarreveld, A . A ., H ill, G . B ., and Jenkins, H . The effect o f tubal ligation on the incidence of epithelial cancer o f the ovary. Cancer. Detect. Prev. 7, 241-245 (1984). 26. Koch, M .. Starreveid. A . A ., Brown, L . B .. and G acdkc, H . Sur vival trends in women with epithelial cancer o f the ovary in Northern A lberta, 1960- 1979. Clin. Invest. Med. 5 , 121 124 (1982). 27. Weiss. N . S ., and Harlow . B. L . W hy docs hysterectomy without bilateral oophorectomy influence the subsequent incidence of ovar ian tau ter. Am. J. Epidemiol. 124. 836-838 (19HA). 28. Kusnen, S .. and Hutchison, R. K . A guide to the work-rtlaudnetx o f disease. U S D H F W , N IO S H . Cindnnattl, O H . (1979). 29. Ncwhouse. M . L ., Pearson, R. M .. Fullerton, J. M -, Bocscn, E. A . M .. and Shannon, H . S. A case-control study o f carcinoma of the ovary, ttr. J. Prev. Soc. Med. 31, 148-133 (1977). A m er ic a n J o u r n a l o p E p id e m io l o g y Copyright'?; 1988 by T h e Johns H o pkins U n iversity School of Hygiene and Public H ealth A ll rights reserved 7 TV<^ Vol. 128. No. 6 Printed in U.S.A. PERSONAL AND ENVIRONMENTAL CHARACTERISTICS RELATED TO EPITHELIAL OVARIAN CANCER n . EXPOSURES TO TALCUM POWDER, TOBACCO, ALCOHOL, AND COFFEE ALICE S. WHITTEMORE,' MARION L. WU,' RALPH S. PAFFENBARGER, Jr.,1 DORIEN L. SARLES,1 JAMES B. KAMPERT,1 STELLA GROSSER,' DEXTER L. JUNG,' SAMUEL BALLON,2 a n d MICHAEL HENDRICKSON3 Whittemore, A S. (Stanford U. School of Medicine, D ept of Health Research and Policy, Stanford, CA 94305-5092), M. L Wu, R. S. Paffenbarger, Jr., D. L. Sarfes, J. B. Kampert, S. Grosser, D. L Jung, S. Ballon, and M. Hendrickson. Personal and environmental characteristics related to epithelial ovarian cancer. II. Exposures to talcum powder, tobacco, alcohol, and coffee. Am J Epidemiol 1988;128:1228-40. Vaginal exposures to talc and other particulates may play an etiologic role in epithelial ovarian cancer. Surgical sterilization may protect against ovarian cancer by blocking entry of such particulates into the peritoneal cavity. The authors assessed histories of talcum powder use, tubal sterilization, and hysterectomy with ovarian conservation in 188 women in the San Francisco Bay Area with epithelial ovarian cancers diagnosed in 1983-1985 and in 539 control women. To investigate the roles of blood-borne environmental exposures on ovarian cancer risk, they assessed lifetime consumption of coffee, tobacco, and alcohol in these women. Of the 539 controls, 280 were hospitalized women without overt cancer, and 259 were chosen from the general population by random digit telephone dialing. Ninety-seven (52% ) of the cancer patients habitually used talcum powder on the perineum, compared with 247 (46%) of the controls. Adjusted for parity, the relative risk (RR) = 1.40, p = 0.06. There were no statistically significant trends with increasing frequency or duration of talc use, and patients did not differ from controls in use of talc on sanitary pads and/or contraceptive dia phragms. Fewer ovarian cancer patients (7% ) than controls (13% ) reported prior fallopian tube ligation (RR, adjusted for parity, = 0.56, p = 0.06), and fewer patients (20% ) than controls (28% ) reported prior hysterectomy (RR = 0.66, p = 0.05). The protective effect of hysterectomy was confined to those who underwent this surgery 10 or more years prior to interview and to those who had not undergone prior tubal sterilization. Consumption of cigarettes and alcohol did not differ between cases and controls. By contrast, 11 (6% ) cases never regularly consumed coffee, compared with 31 (11%) hospital controls and 26 (10%) population controls (RR, adjusted for smoking, = 2.2, p = 0.03, for the comparison using all controls). Overall, ovarian cancer risk among women who had drunk coffee for more than 40 years was 3.4 times that of women who had never regularly consumed coffee (p < 0.01). However, the data exhibited no clear trends in risk with increasing consumption. Although risk ratios relating duration of coffee drinking to ovarian cancer were unaffected by adjustment for several characteristics, further study is needed to exclude potential confounding by other unmeasured characteristics. alcohol drinking; coffee; environmental exposure; hysterectomy; ovarian neo plasms; talc; tobacco; sterilization, tubal Received for publication M ay 29, 1987, and in fin a l 1Department o f H ealth Research and Policy. Stan- lorm A p ril 11. 1988. ford University School o f Medicine, Stanford. CA 1228 OVARIAN CANCER AND ENVIRONMENTAL EXPOSURES 1229 S e v e ra l in v e s tig a to r s h a v e h y p o th e siz e d a c a rc in o g e n ic ro le fo r e x p o su re s o f th e o v a ria n e p ith e liu m to e n v iro n m e n ta l a g e n ts th a it e n te r th e p e lv ic c a v it y th r o u g h the v a g in a l c a n a l (1 -7 ). A tte n tio n h a s fo cused o n h y d ro u s m a g n e siu m silic a te s su ch as ta lc a n d a sb e sto s b e c a u se o f th e s im ila r ity b e tw e e n e p ith e lia l o v a r ia n c a n c e rs a n d m e so th e lio m a s, w h ic h a re c a u se d b y e x p o sure to a sb e sto s. T h e h y p o th e sis p re d ic ts th a t ta lc u m p o w d e r u se o n th e p e rin e u m , on s a n it a r y n a p k in s , a n d o n c o n tra c e p tiv e d e v ice s in c r e a se s o v a r ia n c a n c e r risk . It a lso p re d ic ts th a t tu b a l ste riliz a tio n a n d h y ste re c to m y w ith o u t b ila te ra l o o p h o re c to m y p ro te c t a g a in st th e d ise ase b y p re v e n tin g e n v iro n m e n ta l c a rc in o g e n s fro m c o n ta c tin g th e o v a r ia n e p ith e liu m . T h e re are few d a ta re g a rd in g th e se p re d icte d c o n se q u e n c e s o f th e h y p o th e sis. C ra m e r e t a l. (8 ) re p o rte d th a t o v a r ia n c a n cer p a tie n ts w e re s ig n ific a n tly m o re lik e ly th an n o n h o sp ita liz e d c o n tro l w o m e n to use ta lc u m p o w d e r o n th e p e rin e u m o r o n s a n ita ry n a p k in s . H o w e v e r , H a r t g e e t a l. (9 ) fo u n d n o sta tistic a lly sig n ific a n t d iffe r ences in p rio r ta lc u se b e tw e e n c a se s a n d a se rie s o f h o s p it a l c o n tr o ls . A c o h o r t s tu d y (10) o f \y o m e n w h o h a d u n d e rg o n e tu b a l lig a tio n n o te d fo u r o v a ria n c a n c e rs v e rsu s 1.45 e x p e c te d , a ft e r 2 2 ,0 0 0 p e r s o n -y e a r s o f fo llo w -u p ( p = 0 .0 6 ). T h e fo u r p u b lish e d c a se -c o n tro l stu d ie s th a t h a v e e x a m in e d the e ffe c ts o f h y ste re c to m y w ith o u t b ila t eral o o p h o re c to m y fo u n d c ase s to h av e a lo w e r p re v a le n c e o f h y s te r e c to m y th a n c o n tro ls, in d ic a t in g th a t h y ste r e c to m y is a s s o c ia te d w it h d e c re a se d r is k o f o v a r ia n c a n c e r (1 1 -1 4 ). W e is s a n d H a r lo w (1 5 ) su g g e ste d th a t th is a sso c ia tio n m a y be d u e to scre e n in g fo r m a lig n a n t o r p re m a lig n a n t o v a ria n c o n d i t i o n s a t h y s t e r e c t o m y b y p h y s i c i a n s *1 ' 444 H igh Street. Palo A lto. CA. 1Department o f Pathology, Stanford University School o f M edicine, Stanford, CA. Reprint requests to Dr. Alice S. W hittem ore, Stan ford U nive rsity School o f Medicine, Departm ent o f Health Research and Policy, H R P Building, Stanford, C A 94305-5092. T his w ork was supported by NIH G rant CA 35067. w h o d o n o t ro u tin e ly re m o v e th e o v a rie s. A c c o r d in g to th is h y p o th e sis, w o m e n w h o p a s s su c h sc re e n in g w o u ld h a v e re d u ce d su b se q u e n t o v a ria n c an c e r risk w h e n c o m p a re d w ith w o m e n w h o w ere n o t su b je cte d to h yste re cto m y. D a t a re la tin g o v a ria n c a n c e r to c o n su m p tio n o f co ffe e , to b a cc o , a n d a lc o h o l a lso are sp a rse a n d c o n flic tin g . O n e c a se -c o n tro l stu d y h a s sh o w n a sta tistic a lly sig n ific a n t a sso c ia tio n b e tw e e n co ffe e c o n su m p tio n a n d in c re a se d r isk o f e p ith e lia l o v a ria n c a n c e r (1 6 ). T h is s tu d y c o m p a re d c a se s w ith h o sp ita liz e d c o n tro l w o m e n , w h o se c u rre n t coffe e c o n su m p tio n m a y n o t re p re se n t th a t o f w o m e n in th e g e n e ra l p o p u la tio n . F o u r o th e r stu d ie s u s in g h o sp ita liz e d c o n tro ls (1 7 -2 0 ) a n d o n e u s in g n o n h o sp ita liz e d c o n tro ls (2 1 ) fo u n d w e a k , n o n s ig n ific a n t p o si tiv e a sso c ia tio n s b e tw e e n risk o f th is d is e ase a n d a m o u n t o f u su a l coffe e c o n su m p tio n a t in te rv ie w . It is im p o r ta n t to d e te rm in e to w h a t e x te n t th e se c o n flic tin g fin d in g s m a y b e d u e to d iffe re n c e s in re ce n t co ffe e c o n s u m p tio n b e tw e e n h o sp ita liz e d a n d p o p u la tio n -b a se d c o n tro l g ro u p s o r to o th e r so u rc e s o f b ia s. C ig a re tte s m o k in g h a s b e e n a sso c ia te d w ith in c re a se d o v a ria n c a n c e r risk in o n e p ro sp e c tiv e s tu d y (2 2 ) b u t w a s u n a sso c ia te d w ith it in se v e ra l c a se -c o n tro l stu d ie s (16, 17, 2 1 ). In d e e d , th e c a se -c o n tro l stu d ie s h av e fo u n d s m a ll (n o n sig n ific a n t) re d u c tio n s in ris k a m o n g sm o k e rs. G e n e ra lly , s u c h s tu d ie s h a v e fo u n d n o re la tio n b e tw e e n a lc o h o l c o n su m p tio n a n d o v a ria n c a n c e r, a lth o u g h a re c e n t la rg e stu d y fo u n d a re d u c tio n in risk a sso c ia te d w ith h e a v y d r in k in g (23). W e p re se n t th e re su lts o f a c a se -c o n tro l stu d y o f h is io lo g ic a lly v e rifie d e p ith e lia l o v a ria n c a r c in o m a in w h ic h c a se s' p rio r h isto rie s o f ta lc u se , tu b a l lig a tio n , h y ste r e c to m y w ith o u t b ila te ra l o o p h o re cto m y , a n d c o n s u m p tio n o f coffee, to b a cco , a n d a lc o h o l w e re c o m p a re d w ith th o se o f w o m e n fro m th e g e n e ra l p o p u la tio n , a s w e ll a s w ith th o se o f h o sp ita liz e d c o n tro l w o m e n . A ll su b je c ts re p o rte d life tim e h a b its o f ta lc u se , co ffe e d r in k in g , c ig a re tte s m o k in g , a n d a lc o h o l c o n su m p tio n . S u m - 1230 WHITTEMORE ET AL. m a ry m e a su re s o f life tim e e x p o s u r e s a re e v a lu a te d in re la tio n to o v a ria n c a n c e r risk . Materials and methods S tu d y subjects C a s e s w ere re sid e n ts o f n o r th e rn C a lifo r n ia a ge d 18 to 74 y e ars w h o w ere d ia g n o se d d u r in g th e p e rio d J a n u a ry 1 9 8 3 to D e c e m b e r 1985 a t o n e o f th e se v e n h o sp ita ls in S a n ta C la ra C o u n ty o r a t th e U n iv e rsity o f C a lifo rn ia , S a n F ra n c isc o , M e d ic a l C e n te r. T h e p re se n t re p o rt is re stric te d to 188 w o m e n w ith p rim a ry e p ith e lia l o v a ria n c an ce r. T w o gro u p s o f con trol w o m e n w ere se le c te d . T h e fir s t g r o u p c o n s is te d o f w o m e n w h o w ere h o sp ita liz e d in o n e o f th e h o s p i ta ls to w h ic h c ase s w ere a d m itte d . T h e s e c o n d g ro u p w a s se le c te d fro m th e g e n e ra l p o p u la tio n u s in g ra n d o m d ig it d ia lin g te le p h o n e contacts. B o th gro u p s o f w o m e n w e re m a tc h e d to c a se s o n a ge (w it h in fiv e y e a r in te r v a ls), ra c e (w h ite , b la c k , o r ie n ta l), a n d a d d itio n a l c rite ria d e sc rib e d in th e a c c o m p a n y in g p a p e r (24). A to ta l o f 188 o v a r ia n c a n c e r p a tie n ts, 2 8 0 h o sp ita l c o n tro ls, a n d 2 5 9 p o p u la tio n -b a se d c o n tro ls p a r tic i p a te d . E xposure data and statistical analysis S tu d y su b je c ts p a rtic ip a te d in stru c tu re d in te rv ie w s in th e ir h o m e s c o n d u c te d b y tr a in e d in te rv ie w e rs to a ss e s s m e n s tr u a l a n d re p ro d u c tiv e h isto ry , m e d ic a l a n d fa m ily h isto ry , a n d e n v iro n m e n ta l e x p o su re s. Su b je cts w ere ask e d w h e th e r th e y h a d ever u se d ta lc u m p o w d e r o n th e p e rin e u m , o n sa n ita ry p ad s, o r o n d ia p h ra g m s. S u b je c ts w h o re sp o n d e d a ffirm a tiv e ly to a n y o f th e se q u e stio n s w ere a sk e d a b o u t fre q u e n c y a n d d u ra tio n o f use. S u b je c ts a lso w ere a sk e d w h e th e r th e y h ad ever d ru n k m o re th a n 10 c u p s o f co ffe e in a n y o n e ye ar. S u b je c ts w h o re sp o n d e d a ffirm a tiv e ly re p o rte d th e a ge s w h e n co ffe e d r in k in g sta rte d a n d sto p p e d , th e to ta l n u m b e r o f y e ars o f co ffe e d rin k in g , a n d th e n u m b e r o f c u p s u su a lly c o n su m e d p e r d a y o r p e r w e e k e ith e r c u rr e n tly o r p rio r to sto p p in g . S im ila r q u e stio n s w ere a sk e d a b o u t c ig a re tte s m o k in g , p r o v id e d th a t th e su b je c t h a d sm o k e d a t le a st 1 0 0 c ig a re tte s d u r in g h e r life , a n d a b o u t c o n s u m p t io n o f a lc o h o lic b e ve rage s, p ro v id e d th a t sh e h a d ever co n su m e d m ore th a n 10 su ch b eve r a ge s in a n y o n e year. E lig ib ilit y c rite ria , p a r tic ip a tio n ra te s, a n d d e ta ils o f th e s ta tis tic a l a n a ly s is a re p ro v id e d in th e a c c o m p a n y in g p a p e r (2 4 ). O d d s ra tio s are c a lle d re la tiv e risk s, a n d a ll p v a lu e s a re tw o -ta ile d . Results C h a ra c te ristic s o f th e c a se s a re c o m p a re d w ith th o se o f th e tw o c o n tro l g ro u p s in ta b le 1. P o p u la t io n c o n tr o ls w e re s o m e w h a t y o u n ge r, b e tte r e d u cated , a n d m o re lik e ly to be p re m e n o p a u sa l th a n w ere case s a n d h o sp ita l c o n tro ls. F u rth e rm o re , c a se s w ere le ss lik e ly to h a v e u se d o ra l c o n tra c e p tiv e s a n d h a d a n e a rlie r a ge a t m e n a rc h e a n d T able 1 Characteristics of study participants, San Francisco Bay Area, 1983-1985 C h a ra c te ris tic Cases (n = 188) (%) Controls Hospital in = 280) (% ) Population (n = 259) 1%) Age (years) <40 4 0 -4 9 5 0 -5 9 60+ Race W h ite E d u catio n (years) >12 N o. o f term pregnan cies* 0 1 -3 4+ Age (years) a t m e n arche 12 o r less M en o p au sal status Prem enopausal N a tu ral menopause S u rg ical m enopause O ra l contraceptives E v er used 10 30 26 35 95 59 21 60 19 48 34 45 22 46 * 20 or m ore w eeks gestation. 10 29 28 33 97 59 17 59 24 44 32 33 35 50 15 29 27 29 94 67 10 67 23 46 41 38 21 58 a t the re tte s io n o f e had le v e r- ra te s, s are (24). n d a ll ja re d ta b le w hat ik e ly and w ere tiv e s and :isco ation 259) .) | ) OVARIAN CANCER AND ENVIRONMENTAL EXPOSURES 1231 few er te rm p r e g n a n c ie s t h a n e it h e r c o n tr o l group. T h e la tte r fin d in g s a re re p o rte d in detail in th e c o m p a n io n p a p e r (2 4 ). R e la tiv e r is k s a ss o c ia te d w ith ta lc u se , tubal lig a tio n , a n d h y ste re c to m y w e re s im ilar w h e n c a s e s w e r e c o m p a r e d w it h h o s p i tal c o n tro ls a n d w ith p o p u la tio n c o n tro ls. T herefore, fo r th e se v a ria b le s, w e re p o rt re sults o n ly fo r c a s e s v e r s u s th e c o m b in e d group o f a ll c o n tro ls. Talc use A gre ate r p ro p o rtio n o f c a se s (5 2 p e r cent) th a n c o n tr o ls (4 6 p e r c e n t) re p o rte d p rior u se o f t a lc u m p o w d e r o n th e p e r in e u m (re lative r is k ( R R ) = 1.4 0 , p = 0 .0 6 ). H o w ever, th e re w a s lit t le d iffe r e n c e b e tw e e n cases a n d c o n tro ls in u se o f ta lc o n s a n ita r y pads a n d d ia p h ra g m s , w ith re la tiv e risk s o f 0j93 ( p = 0 . 7 6 ) a n d 0 . 9 5 ( p = 0 . 8 6 ) , r e s p e c tive ly. T a b le 2 s h o w s t h e d is t r ib u t io n s o f cases a n d c o n tr o ls a n d re la tiv e ris k s fo r talc u se d ir e c t ly o n th e p e r in e u m , o n s a n i tary p a d s, a n d o n c o n tr a c e p tiv e d ia phragm s, sin g ly a n d in c o m b in a tio n . T h e tab le p r o v id e s n o e v id e n c e o f e le v a te d r is k a sso c ia te d w ith m o re t h a n o n e fo r m o f ta lc use. N o n e o f th e w o m e n in th e s t u d y re ported p rio r o c c u p a tio n a l e x p o su re s to ta lc or a sb e sto s fib e rs. W e n e x t e x a m in e d w h e th e r risk in creased w ith in c re a se d d u ra tio n o r fre quency o f u se o f ta lc u m p o w d e r o n th e p e rin e u m . S in c e tu b a l lig a tio n a n d h y ste r e ctom y p re ve n t co n ta ct betw een th e o v a r ia n e p ith e liu m a n d e x o ge n o u s a g e n ts in th e v a g in a l c a n a l, w e e x c lu d e d a n y p e rin e a l ta lc u se a fte r th e d a te o f tu b a l lig a tio n o r h y s te re cto m y in c a lc u la tin g d u ra tio n o f use. A s se e n in ta b le 3, 55 p e r c e n t o f c a se s v e rsu s 59 p e r c e n t o f c o n tro ls re p o rte d su c h u se fo r le ss t h a n o n e y e a r. T h e r is k fo r ta lc u se b e tw e e n o n e a n d n in e y e a rs, re la tiv e to th a t a m o n g u s e r s o f s h o r te r d u r a tio n , w a s 1.60 (p = 0 .0 5 ). H o w e v e r, a n in c re a sin g d o se re sp o n se p a tte rn w a s n o t a p p a re n t, w ith risk a m o n g lo n g -te rm u se rs o f 10 o r m ore y e a r s o n ly 1.11 tim e s t h a t o f th e n o n u s e r s o r u se rs o f le ss t h a n o n e y e a r (p = 0.61). A c c o rd in g to th e lo g istic m o d e l fit to th e d a ta , th e o v e ra ll in c re a se in risk fo r a n y 10y e a r in c r e a s e in d u r a t io n o f u s e w a s 1.01 (p = 0.5 6 ). T able 3 Ovarian cancer risk by length of talcum powder use on the perineum ,* San Francisco Bay Area, 1983-1985 Years of talc use* Cases n% Controls n% R e la tiv e r is k t 95% confidence interval None 103 55 3 2 0 59 1.00 1 -9 34 18 72 13 1.60 1 .0 0 -2 .5 7 10+ 5 0 27 147 27 1.11 0 .7 4 -1 .6 5 U nknow n 1 100 T o tal 188 100 539 100 * Prior to tubal ligation or hysterectomy, t Adjusted for parity. T able 2 Ovarian cancer risk by type of talcum powder use. San Francisco B ay Area, 1983-1985 T y p e o f talc use None Perineum only Sanitary pads only Diaphragm only Any two of perineum , pads, and diaphragm All three of perineum , pads, and diaphragm Incomplete data Cases n% 75 40 22 12 53 95 67 36 11 95 C o n tro ls n% 230 * 55 28 19 43 10 5 4 168 31 92 30 6 Relative risk* 1.00 1.45 0.62 1.50 1.36 0.35 95% confidence interval 0 .8 1 -2 .6 0 0 .2 1 -1 .8 0 0 .6 3 -3 .5 8 0.9 1 -2 .0 4 0.04 -2 .9 4 Total 188 * A djusted fo r p a rity a n d o ra l c o n tra c e p tiv e use. 100 539 100 1232 vvhittemore et al. T h e re la tio n b e tw e e n o v a ria n c a n c e r risk a n d u su a l fre q u e n c y o f ta lc u se o n th e p e r in e u m is s h o w n in ta b le 4. W o m e n w h o u se d ta lc a n a v e ra g e o f o n e to 20 tim e s p e r m o n th w e re n o t a t s ig n ific a n tly a lte re d risk fr o m th o se w h o u se d it le ss fre q u e n tly ( R R = 1.27, p = 0 .2 9 ). T h o s e w h o c u s to m a rily u se d ta lc o n th e p e rin e u m 20 o r m o re tim e s p e r m o n th w e re a t 1.45 tim e s th e risk o f w o m e n in th e lo w e st u se c a te go ry ( p = 0 .09). T h e o v e ra ll in c re a se in risk a sso c i ate d w ith 30 a p p lic a tio n s p e r m o n th w as 1.30 ( p = 0 .1 9 ). T ubal ligation a nd hysterectom y T o in v e stig a te fu rth e r th e h y p o th e sis o f a ro le fo r v a g in a l e x p o su re to e n v iro n m e n ta l c a rc in o g e n s in th e e tio lo g y o f o v a ria n can ce r, w e e x a m in e d th e e ffe ct o f tu b a l lig a tio n a n d h y ste re c to m y o n risk fo r th e d ise a se . S e v e n p e r c e n t o f c a se s a n d 13 p e r c e n t o f c o n tro ls re p o rte d a h isto ry o f tu b a l lig a tio n ( R R = 0 .53, p = 0 .0 5 ). R e la tiv e r is k s fo r tu b a l lig a tio n w e re le ss th a n o n e a m o n g n u llip a ro u s, u n ip a ro u s, a n d m u ltip a ro u s w o m e n , in d ic a tin g th a t th e p ro te c tiv e e ffe ct o f su c h su rg e ry o n o v a ria n c a n c e r risk c a n n o t be e x p la in e d b y c o n fo u n d in g d u e to its g re a te r p re v a le n c e a m o n g p a ro u s w o m e n w h o are at re d u ce d risk o f the d is ease. T h e o v e ra ll re d u c tio n in risk a sso c i a te d w ith tu b a l lig a tio n , a d ju ste d fo r p a rity , w a s 0 .5 6 ( p = 0.0 7 ). C a se s w ith tu b a l lig a tio n te n d e d to u n d e rgo th is su rg e ry a t y o u n ge r age s (m e a n a g e ( s ta n d a r d e rro r) = 31.9 y e a rs ( 0 .5 )) th a n d id c o n tr o ls (3 4 .2 y e a rs ( 0 .1 )). T h is d iffe re n c e d o e s n o t su p p o rt th e h y p o th e sis th a t e a rly tu b a l lig a tio n c o n fe rs gre a te r p ro te c tio n to th e o v a rie s b y e a rly te r m in a tio n o f e x p o su re fro m th e v a g in a l c a n a l. R e la tiv e to w o m e n w ith o u t th is su rge ry , th e risk fo r w o m e n w ith tu b a l lig a tio n w ith in 10 y e a r s o f in te rv ie w w a s 0 .35 (9 5 p e r c e n t c o n fid e n c e in te rv a l ( C l) 0 .1 2 -1 .0 2 ), w h ile th e risk fo r w o m e n w h o u n d e rw e n t th e s u rg e ry m o re th a n 10 y e a rs b e fo re in te rv ie w w a s 0 .6 9 (9 5 p e r c e n t C l 0 .3 2 -1 .5 0 ). T a b le 5 s h o w s th a t h y ste re c to m iz e d w o m e n e x p e rie n c e d 0 .6 6 tim e s th e risk o f th o se w ith o u t su c h su rg e ry ( p = 0 .05), b u t it d o e s n o t sh o w a n y tre n d o f d e c re a sin g p ro te c tio n w ith tim e sin c e h y ste re c to m y . S u c h a tre n d w o u ld be e x p e cte d if th e p r o te c tiv e e ffe c t o f h y ste re c to m y re fle c te d m e re ly th e re m o v a l o f h ig h -risk w o m e n fro m th e h y ste re c to m iz e d p o p u la tio n v ia se le c tiv e o o p h o re c to m y a s su g g e ste d b y W e is s a n d H a r lo w (1 5 ). O n th e c o n tra ry , ta b le 5 s h o w s th a t p ro te c tio n is lim ite d to w o m e n h y ste re c to m iz e d 10 o r m o re y e a rs p r io r to in te rv ie w . O v e ra ll, c a se s a n d c o n tro ls d id n o t d iffe r sig n ific a n tly in th e age s a t w h ic h h y ste re c to m y w a s p e rform e d . T h e m e an age at h yste recto m y am o n g cases w ith s u c h p r io r s u r g e r y w a s 40.1 y e a r s ( 0.3 ), w h ile th e c o rre sp o n d in g age fo r c o n tr o ls w a s 39 .7 y e a rs ( 0 .1 ). T h e h y p o th e sis th a t h y ste re c to m y p r o te c ts a g a in s t o v a ria n c a n c e r b y b lo c k in g T able 4 O vanan cancer risk by usual frequency'of talcum powder use on the perineum , San Francisco Bay Area, 1983-1985 A pplications of talc per month None 1 -2 0 20+ U nknow n Cases n ci 97 52 41 22 44 23 63 Controls n 312 114 101 12 58 21 19 2 R e la tiv e risk* 1.00 1.27 1.45 95% confidence interval 0 .8 2 -1 .9 6 0 .9 4 -2 .2 2 T o tal 18 100 539 100 O v e ra ll tre n d fo r 30 uses per m o n th ' A d ju sted for p a rity . 1.30 0.88 -1 .9 2 OVARIAN CANCER AND ENVIRONMENTAL EXPOSURES 1233 exogenous a ge n ts' a cce ss to th e o v a rie s p re d icts th a t su c h s u r g e r y c o n fe rs n o b e n e fit on w o m e n w h o se o v a rie s a re a lre a d y p ro tected b y p rio r tu b a l lig a tio n . T a b le 5 sh o w s re la tiv e r is k s a s s o c ia te d w ith h y s te r e c to m y a m o n g w o m e n w ith a n d w ith o u t p rio r tu b a l liga tio n . A s p re d ic te d , h y ste r e c to m y fa ile d to p ro te c t w o m e n w h o h a d u n d e r g o n e p r io r tu b a l lig a tio n ( R R = 2 .56, p = 0 .4 5 ), b u t it d id p ro te c t th o se w h o h a d n o t ( R R = 0.57, p = 0 .0 2 ). T a b le 6 sh o w s th e e ffe c ts o f p e rin e a l ta lc u se se p a ra te ly a m o n g w o m e n w ith a n d w ith o u t p rio r tu b a l lig a tio n o r h y ste re c to m y . T h e h ig h e st risk w a s e x p e rie n c e d b y ta lc u se rs w ith o u t su c h su rg e ry . T h e risk in T able 5 Ovarian cancer risk after hysterectom y without bilateral oophorectomy, by time between hysterectomy and interview and by absence or presence of prior tubal ligation. San Francisco Bay Area. 1983-1985 Cases n% Controls n% Relative risk 95% confidence interval No hysterectomy Hysterectomy Time (years) between hysterec- tomy and interview 1-9 10-19 20+ 151 80 37 2 0 15 8 11 6 11 6 389 150* 42 66 41 72 28 8 12 8 1.00 0 .6 6 1.01 0.47 0.63 0.43-1.00 0.54-1.89 0.24-0.92 0.31-1.29 No p rior tubal ligation No hysterectomy Hysterectomy Prior tubal ligation No hysterectomy Hysterectomy 141 81+ 33 19 10 71t 4 29 * Date of hysterectomy was unknow n for one woman, t Per cent o f cases or controls w ith no p rio r tubal ligation, t Per cent o f cases or controls w ith p rio r tubal ligation. 333 135 56 15 71 + 1.00 29 0.57 79 1.00 21 2.56 0.36-0.90 0.23-29.12 T able 6 Ovarian cancer risk by perineal talc use and by history of surgical sterilization, t San Francisco Bay .4rea, 1983-1985 Talc use Surgical sterilization* Cases n <7 Controls n +r Relative risk* 95% confidence interval No No Yes No No Yes Yes Yes 70 37 182 34 1.0 0 71 38 151 28 1.33 21 11 110 20 0.50* 26 14 96 18 0.75 0 .88-2.01 0.29-0.88 0.43-1.29 No Total 91 48 292 54 1.00 Yes T o ta l 97 52 247 46 1.37 0.97-1.95 Total Total No Yes 141 75 333 62 1.00 47 25 206 38 0.53* ' 0.36-0.79 * d < 0 .0 1 . + Tubal ligation or hysterectomy. t Adjusted for parity. Adjusted for parity and surgical sterilization. || Adjusted for p a rity and talc use. 1234 WHITTEMORE ET AL. th is g ro u p w a s 1.33 tim e s th a t o f w o m e n w ith n e ith e r h is to r y o f ta lc u se n o r h isto ry o f su rg e ry ( p = 0 .18). B y c o n tra st, risk w as lo w e st a m o n g w o m e n w ith a h is to r y o f tu b a l lig a tio n o r h y ste re c to m y w h o n e v e r re g u la rly a p p lie d ta lc to th e p e rin e u m ( R R = 0.50, p = 0 .0 2 ). T a b le 6 s h o w s th a t, re g a rd le ss o f ta lc use, w o m e n w h o u n d e rw e n t e ith e r tu b a l lig a tio n o r h y ste re c to m y w ith o u t b ila te ra l o o p h o re c to m y e x p e rie n c e d a r is k o f 0.5 3 (p = 0.002) c o m p a re d w ith w o m e n w ith o u t su c h su rg e ry . C a se s a n d c o n tro ls d id n o t d iffe r s ig n ifi c a n tly in th e p re v a le n c e o f b a rrie r c o n tra ce p tiv e u se , d e v ic e s th a t p re v e n t e n try o f se m e n in to th e p e rito n e a l c a v ity . R is k s fo r w o m e n w h o h a d u se d d ia p h ra g m s a n d c o n d o m s w e re 0 .8 1 ( p = 0 .2 2 ) a n d 0.91 ( p = 0.59), re sp e c tiv e ly , re la tiv e to risk a m o n g n o n u se rs. T h e r e w a s n o tr e n d in risk w ith e x p o su re o f th e o v a rie s to se m e n , d e fin e d a s se x u a lly a c tiv e tim e b e fo re tu b a l lig a tio n o r h y ste re c to m y , m in u s d u ra tio n o f use o f c o n d o m s a n d d ia p h ra g m s. Coffee, tobacco, a n d alcohol T a b le 7 sh o w s th e d istrib u tio n o f case s a n d c o n tr o ls a n d re la tiv e ris k s b y co ffe e c o n su m p tio n sta tu s (e ve r vs. ne ver). T h e re la tiv e r is k fo r a n y c o ffe e c o n su m p tio n , a d ju s te d fo r c ig a r e tte s m o k in g , w a s 2.21 fo r b o th c o n tr o l g r o u p s c o m b in e d ( p = 0.03). T h e re la tiv e r is k w a s 2 .13 (p = 0 .0 6 ) fo r c a se s v e r s u s h o s p it a l c o n tr o ls a n d 1.59 (p = 0 .2 4 ) fo r c a se s v e rsu s p o p u la tio n c o n tro ls. T h e c o rre s p o n d in g u n a d ju ste d re la tiv e r is k s w e re 2.03 fo r c o m b in e d c o n tro ls, 1.9 0 fo r h o s p it a l c o n tr o ls , a n d 1.51 fo r p o p u la tio n c o n tro ls (n o t sh o w n ). T a b le 7 a lso sh o w s th a t risk a m o n g coffee d r in k e rs in c re a se s w ith in c re a sin g d u ra tio n o f c o ffe e c o n s u m p tio n . T h is tr e n d is e v i d e n t in b o th th e c o m p a riso n b ase d o n h o s p ita l c o n tr o ls a n d th e o n e b a se d o n p o p u la tio n c o n tro ls. O v e ra ll, sm o k in g -a d ju ste d can cer rate s a m o n g w o m e n w h o h ad co n su m e d co ffe e fo r m o re th a n 40 y e a rs w ere 3.4 tim e s th o s e o f w o m e n w h o h a d n e v e r c o n s u m e d co ffe e ( p < 0.0 1 ). A m o n g coffe e d rin k e rs, e a c h a d d itio n a l 10 y e a rs o f coffe e d r in k in g c o n fe rre d a n 11 p e r c e n t in cre a se in risk , a c c o rd in g to th e lo g istic fu n c tio n fit to th e d a ta ( p = 0 .3 7 ). T h e s e fin d in g s c o u ld be c o n fo u n d e d b y a ge d e sp ite th e m a tc h in g w ith in fiv e -y e a r a ge g ro u p s. H o w e v e r, th e re su lts w e re s im ila r w h e n age w a s a d d e d to th e re g re s sio n s a s a c o n tin u o u s v a ria b le . R e la tio n s b etw e e n o v a ria n can ce r risk a n d a m o u n t o f u su a l coffe e c o n su m p tio n are s h o w n in ta b le 8. T h e ta b le p ro v id e s n o e v id e n c e fo r a p o sitiv e tr e n d in risk w ith T able 7 O vanan cancer risk by years of coffee consumption. S a n Francisco Bay Area, 1983-1985 Years o f coffee consumption None Any 1-14 15-24 2 5 -3 9 40+ U n sp ecified Cases n% Controls Hospital Population n %n% i i 6 31 u 26 10 177 94 249 89 233 90 18 10 27 10 34 13 3 2 17 43 15 3 6 14 62 33 105 38 98 38 65 35 73 26 64 25 00 10 10 Relative ris kt Cases vs. hospital controls Cases vs. population controls Cases v9. all controls 1.00 2.13 1.57 1.81 2 .3 6 3 .4 5 ' 1.00 1.59 0.65 1.69 1.70 2.54 1.00 2.21 1.45 2.18 2.26 3 .4 1 * 95% confidence interval^ 1 .1 0 -4 .4 1 0.5 9 -3 .5 7 1.00-4.7 9 1.06-4.8 5 1.46-7.9 6 O verall tre n d p er 10 years am ong coffee d rin kers 1.16 * p < 0.01. A d justed fo r s m o k in g (life lo n g n o n sm o ker vs. ever sm o ke r). t Cases versus a ll con trols. 1.14 1.11 0.89 -1 .3 8 fre th i qu cu on we nu for tre qu wa wi> dis Ove d d hos; 0 l-3 ( 3 1 -f 61-S 90+ Unk Over ye dr + + OVARIAN CANCER AND ENVIRONMENTAL EXPOSURES 1235 fre q u e n cy o f coffe e d r in k in g , re g a rd le ss o f the) c o n t r o l g r o u p u s e d fo r c o m p a r is o n . F r e quency w as m easu red as u su al n u m b er of c u p s c o n su m e d p e r d a y p rio r to d ise a se o n se t fo r c ase s a n d h o sp ita l c o n tro ls w h o w ere c u rre n t co ffe e d rin k e rs, a n d a s u su a l n u m b e r o f cu p s p e r d a y p rio r to sto p p in g fo r fo rm e r co ffe e d rin k e rs. T h e te st fo r tre n d o f in c re a sin g risk w ith in c re a sin g fre q u e n cy a m o n g th o se w h o c o n su m e d coffe e w a s n o t s ig n ific a n t ( p = 0 .91). S im ila rly , ta b le 9 s h o w s n o tre n d in ris k w ith to ta l c u p s o f co ffe e c o n s u m e d p r io r to d ise a se o n se t (c a se s a n d h o sp ita l c o n tro ls) o r p rio r to in te rv ie w (p o p u la tio n c o n tro ls). W e e stim a te d to ta l co ffe e c o n s u m p tio n b y m u lt i p ly i n g e a c h w o m a n 's r e p o r t e d d u r a tio n o f co ffe e c o n su m p tio n in y e a rs b y h er u su a l fre q u e n cy o f c o n su m p tio n in c u p s p e r d a y tim e s 365. A m o n g co ffe e d rin k e rs, th e te st fo r tre n d in r is k w ith in c re a sin g to ta l c o n s u m p tio n y ie ld e d a p v a lu e o f 0.56. R e la tiv e r is k s c o rre s p o n d in g to th e a b o v e m e a su re s o f co ffe e c o n su m p tio n w ere u n c h a n g e d b y a d ju stm e n t fo r o th e r v a ria b le s p o te n tia lly a sso c ia te d w ith o v a ria n c a n c e r in c lu d in g e d u c a t io n a l le v e l, p a r ity , o ra l c o n tra c e p tiv e u se , e stim a te d y e a rs o f o v u - T able 8 Ovarian cancer n s k by u su a l fre q u en c y of coffee consum ption.* S a n F rancisco B a y A rea, 1983-1985 Cups/day 0 1 2 -3 4+ Cases n% h6 50 27 73 39 54 29 Controls R elative r i s k t Hospital n% 31 n 62 22 94 34 93 33 Population n t"c 2 6 10 58 22 98 38 77 30 Cases vs. hospital controls 1.00 2.21 2.13 2.02 Cases vs. population controls 1.00 1.86 1.63 1.56 Cases vs. all controls 1.00 2.42 2.26 2.07 95% confidence interval$ 1.1 5 -5 .0 9 1.0 9 -4 .6 6 0 .9 7 -4 .3 8 Overall trend per cup/ day among coffee drinkers 1.01 1.01 1.01 0.93-1.08 " Prior to stopping for ex-drinkers, and prior to hospitalization for current drinkers, among cases and hospital controls. t Adjusted for smoking (lifelong nonsmoker vs. ever smoker). X Cases versus all controls. T able 9 Ovarian cancer risk by estim ated total cups of coffee consumed. Sa n Francisco B ay Area. 1983-1985 Cup-years* of coffee con su m p tio n Cases Controls Relative ris k t Hospital n% Population n% Cases vs. hospital controls Cases vs. p o p u la tio n controls Cases vs. controls 95% confidence in te rv a l 0 1 -3 0 3 1 -6 0 6 1 -9 0 90+ U n know n 11 6 31 11 2 6 10 41 22 5 0 18 60 23 32 17 46 16 32 12 2 7 14 3 8 14 32 12 77 41 114 41 108 42 00 10 10 l.o p 2.30 2.21 2.03 2.42 1.00 1.54 2 .3 0 1.96 1.60 1.00 2 .3 0 2.64 2.46 2.28 1 .09-4.8 6 1.2 1 -5 .7 5 1.10-5.5 1 1.08 -4 .7 8 Overall trend per 10 cupyears among coffee drinkers 1.01 1.01 1.01 0.99-1.03 One cup-year equals 365 cups of coffee, t Adjusted for smoking (lifelong nonsmoker vs. ev;r smoker) %Cases versus all controls. 1236 WHITTEMORE ET AL. la tio n , a n d d u ra tio n o f c o n tra c e p tiv e -fre e m a rria ge . T h e risk fo r c ig a re tte s m o k in g re la tiv e to life lo n g n o n s m o k in g , a d ju ste d fo r coffe e c o n su m p tio n , w a s 0 .7 3 ( p = 0.08) fo r th e c o m p a riso n b a se d o n c o m b in e d co n tro l g ro u p s. C o rre s p o n d in g re la tiv e risk s fo r th e h o sp ita l a n d p o p u la tio n c o m p a riso n s w ere 0.6 5 ( p = 0 .0 4 ) a n d 0 .8 4 ( p = 0.3 9 ), re sp e c tiv e ly . U n a d ju s te d re la tiv e r is k s a n d c o n fi d e n ce in te rv a ls w ere sim ila r. W h ile th e o d d s ra tio s a sso c ia te d w ith s m o k in g w ere le ss t h a n u n ity , fe w o f th e m a c h ie v e d s t a tistic a l sig n ific a n c e . C a s e s d id n o t d iffe r fro m e ith e r c o n tro l g ro u p in th e p re v a le n c e o f p rio r a lc o h o l c o n su m p tio n . F o r th e c o m p a riso n b ase d o n a ll c o n tr o ls , th e re la tiv e r is k w a s 0.74 ( p = 0 .1 4 ). F u rth e rm o re , th e re w a s n o e v id e n c e o f a tre n d in risk w ith in c re a sin g d u ra tio n o r a m o u n t o f a lc o h o l c o n su m p tio n . W o m e n w h o d r a n k h e a v ily (2 0 o r m o re d rin k s p e r w e e k ) h a d a r is k 0 .6 6 tim e s th a t o f n o n d rin k e rs, b u t th e re d u c tio n w a s n o t s ta tis tic a lly s ig n ific a n t ( p = 0 .3 4 ). N o n e o f th e se o b se rv a tio n s w e re a lte re d b y a d ju stm e n t fo r c ig a re tte s m o k in g o r co ffe e c o n su m p tio n . Discussion G enital exposures to talc a n d o th er agents T h e ra tio n a le fo r su s p e c tin g ta lc a s a n o v a ria n c a r c in o g e n d e riv e s fro m its c h e m i c a l re la tio n to a n d n a tu ra l o c cu rre n ce w ith a sb e sto s. A sb e sto s c a u se s p le u ra l a n d p e ri to n e a l m e so th e lio m a s (2 5 ), w h ic h are h is to lo g ic a lly s im ila r to e p ith e lia l o v a ria n c a r c in o m a s (6 ). G r a h a m a n d G r a h a m (2). sh o w e d th a t, in g u in e a p ig s a n d ra b b its, a sb e sto s c a n in d u c e o v a ria n e p ith e lia l h y p e rp la sia s im ila r to e a rly e p ith e lia l tu m o rs in w o m e n . E v id e n c e s u g g e s tin g a ro le fo r v a g in a l e x p o su re to p a rtic u la te s in h u m a n o v a r ia n c a r c in o g e n e s is is tw o fo ld . F irst, E g li e t a l. (2 6 ) d e m o n s tr a te d th a t n o n m o tile in e rt c a r b o n p a r tic le s d e p o site d in th e v a g in a p rio r to h y ste re c to m y ca n be re co ve re d in th e fa llo p ia n tu be s. S e co n d , H e n d e r s o n a n d c o w o r k e rs h a v e fo u n d ta lc p a rtic le s e m b e d d e d in b o th n o rm a l a n d m a lig n a n t o v a ria n tissu e (27, 2 8 ). W h ile th e se fin d in g s in d ic a te th a t v a g in a l e x p o su re to p a rtic u la te s c a n le a d to d e p o sitio n o n th e o v a rie s, th e y d o n o t im p lic a te su c h e x p o su re in o v a ria n c a rc in o g e n e sis, a n d d a ta re la tin g d ire c tly to th is p o ssib ility are needed. In a c o m p a ris o n o f e p ith e lia l o v a ria n c a n c e r c a se s a n d n o n h o sp ita liz e d c o n tro ls in B o s t o n , C r a m e r e t a l. (8 ) re p o rte d a re la tiv e r is k o f 1.92 (p < 0 .0 0 3 ) fo r e p ith e lia l o v a ria n c a n c e r a sso c ia te d w ith u se o f ta lc u m p o w d e r o n th e p e rin e u m o r o n s a n ita ry p a d s. B y c o n tra st, th e re su lts o f a c a se -c o n tro l stu d y in W a s h in g to n , D C (9) a n d th o se o f th e p re se n t stu d y sh o w n e ith e r a stro n g n o r a c o n siste n t a sso c ia tio n b e tw e e n g e n ita l ta lc u m p o w d e r e x p o su re a n d o v a ria n can ce r. In th e p re se n t d a ta , re g u la r u se o f ta lc o n th e p e rin e u m w a s a sso c ia te d w ith o n ly a m a rg in a lly s ig n ific a n t e le v a tio n in re la tiv e risk . F u rth e rm o re , th e re w ere n o c le a r d iffe re n c e s b e tw e e n c a se s a n d c o n tr o ls w h e n o th e r fo rm s o f g e n ita l ta lc e x p o su re w e re c o n sid e re d , e ith e r s in g ly o r in c o m b in a tio n . A lth o u g h th e d ata sh o w a tre n d o f in c re a sin g risk w ith in c re a sin g fre q u e n c y o f p e rin e a l e x p o su re , th e tre n d is n o t sta tis tic a lly sig n ific a n t, a n d th e re is n o tre n d w ith d u ra tio n o f e x p o su re . T h u s, w h ile th e se d a ta d o n o t e x o n e ra te ta lc a s a n o v a ria n c a rc in o g e n , n e ith e r d o th e y p ro v id e s t r o n g e v id e n c e t o im p lic a t e it. S e v e ra l so u rc e s o f b ia s m u st be c o n sid e re d a s p o ssib le e x p la n a tio n s fo r th e la c k o f s t r o n g fin d in g s re la te d to ta lc , in c lu d in g t h e s t u d y 's f a i l u r e t o i n t e r v i e w a l l e l i g i b l e o v a ria n c a n c e r p a tie n ts a n d a c o m p le te ly ra n d o m s a m p le o f c o n tro ls, a s w e ll a s th e p o te n tia l p itfa lls o f c o m b in in g th e tw o c o n tro l g ro u p s. A n o th e r so u rc e o f b ia s is c o n fo u n d in g b y d iffe re n tia l ta lc u se a m o n g w o m e n w ith c h a ra c te ristic s p re d ic tiv e o f o v a ria n can ce r. H o w e ve r, su c h c o n fo u n d in g se e m s u n lik e ly . F o r e x a m p le , a lth o u g h W y n d e r e t a l. (1 4 ) re p o rte d th a t c e r ta in m e n str u a l c h a ra c te ristic s d iffe r b e tw e e n w o m e n w ith o v a ria n c a n c e r a n d c o n tro ls. we d iffi any am i irre j p a ir A e rro atte ror t o n ly cont o b sc by ( p e ril e p id i th e r o r p: gene In lo g ic exogf duce< tu be cedui th ro i ta lc a stan c ia n c m en, douci im p lii Th, w ith f w ith i stu d y tu bal enced c re a se stu d y cancel W oi tio n di y e a rs ade of trast, 1 at th e a b ly at sures h OVARIAN CANCER AND ENVIRONMENTAL EXPOSURES 1237 d m a- w e a n d o th e rs (8, 1 3 ) fo u n d n o s ig n ific a n t n o t e w o r th y th a t th e p r o te c tiv e e ffe c ts o f th ese i i d iffe re n ce s b e tw e e n c a s e s a n d c o n tr o ls in h y ste re c to m y n o te d h e re w e re lim ite d to ire to a n y o f se v e ra l m e n s tr u a l c h a r a c te r ist ic s e x su rg e ry 10 o r m o re y e a rs p r io r to in te rv ie w . a the a m in e d , in c lu d in g h is t o r y o f a m e n o rr h e a , T h is fin d in g fa ils to s u p p o r t th e c o n je c tu re ?x p o - irre g u la r m e n stru a l c y c le s, a n d m id c y c le o f W e is s a n d H a r lo w (1 5 ) th a t th e p ro te c d ata p ain . tio n o f h y ste re c to m y is a n a rtifa c t d u e to a re A n a d d itio n a l s o u rc e o f b ia s is r a n d o m se le c tiv e re m o v a l o f p re c a n c e r o u s o v a rie s i e rro r in re p o rte d ta lc u se , w h ic h te n d s to a t th e tim e o f su rg e ry . If su c h se le c tio n m an i a tte n u a te re la tiv e r is k e stim a te s. S o m e e r e x p la in e d th e a ss o c ia tio n , o n e w o u ld e x p e c t tro ls ro r se e m s lik e ly . N e v e rt h e le s s , th e re se e m s to fin d , a s th o se a u th o r s d id , a d e c re a se in ed a o n ly a s m a ll p r o b a b ilit y t h a t th e se d a ta th e le v e l o f p r o te c tio n w ith in c r e a s in g y e a rs ith e - c o n ta in re p o rtin g e rro rs la rg e e n o u g h to sin c e su rg e ry . H o w e v e r, th e p re se n t d a ta ;e o f o b scu re th e tw o fo ld in c re a se in risk n o te d sh o w ju st th e o p p o site tre n d . T h e y a lso san - b y C r a m e r e t a l. (8 ) fo r ta lc u s e o n th e in d ic a te th a t th e b e n e fits o f h y s te r e c to m y of a p e rin e u m a n d o n s a n it a r y p a d s. F u r th e r a re c o n fin e d to w o m e n w ith o u t p r io r tu b a l ' (9 ) e p id e m io lo g ic stu d ie s a re n e e d e d to c la r ify ste riliz a tio n , a fin d in g th a t s u p p o r ts a n e ti- th er th e ro le o f ta lc a s c a r c in o g e n , c o c a rc in o g e n , o lo g ic ro le fo r g e n ita l e x p o su re s to th e o v a b e - o r p r o m o te r o f e p ith e lia l o v a r ia n c a r c in o rie s. a n d ge n e sis. O th e r e x p la n a tio n s a re p o ssib le fo r th e jla r In d ire c t e v id e n ce in s u p p o r t o f a n e tio - p ro te ctiv e e ffe cts o f tu b a l lig a tio n a n d h y s ite d lo g ic ro le fo r v a g in a l e x p o s u r e s to s o m e te re c to m y , if th e e ffe c ts a re n o t d u e to io n i e x o ge n o u s s u b sta n c e s d e r iv e s fro m th e re c h a n c e o r b ia s. F o r e x a m p le , th e se p ro c e n o d u ce d risk a sso c ia te d h e re w ith fa llo p ia n d u re s m a y a lte r th e le v e ls a n d /o r th e c y c lic : on- I tube ste riliz a tio n a n d h y ste re c to m y , p r o v a ria tio n s o f e stro g e n , p ro g e ste ro n e , o r th e sx - ce d u re s th a t b lo c k e n try to th e p e lv ic a re a g o n a d o tro p in s. T u b a l ste riliz a tio n h a s b e e n in th ro u gh th e re p ro d u c tiv e tra c t. A p a r t fro m sh o w n to re d u ce su b se q u e n t se ru m a n d u r i < ra ta lc a n d a sb e sto s, v a g in a lly in tr o d u c e d s u b n a ry e stro g e n le v e ls, p o s s ib ly b y in d u c in g ' re sta n c e s th a t m a y in itia te o r p ro m o te o v a r lo c a liz e d h y p e rte n sio n a t th e o v a r y (2 9 , 30 ). is ian c a n c e r in c lu d e o th e r p a r tic u la te s, se S o m e d a ta (3 0 , 3 1 ) s u g g e s t th a t h y ste r e c no us, as i -o- i m en, sp e rm ic id a l fo a m s o r c re a m s, a n d d o u ch e so lu tio n s. W e a re u n a w a re o f d a ta im p lic a tin g a n y o f th e se su b sta n c e s. T h e re d u ce d risk a sso c ia te d in th e se d a ta w ith p rio r tu b a l s te r iliz a t io n is in c o n s is te n t to m y w ith o v a ria n c o n s e rv a tio n a ls o m a y re d u ce e stro g e n p ro d u c tio n . E le v a te d e stro g e n le v e ls h a v e b e e n lin k e d to b re a st c a n c e r a n d c o u ld be in v o lv e d in o v a ria n c a rc in o g e n e sis, a lth o u g h th e re a re fe w d a ta to s u p ! d Ij c k ag w ith th e re su lts o f a h is to r ic a l p ro sp e c tiv e study o f 666 w o m e n w h o h a d u n d e rgo n e tu b a l lig a tio n (10). T h e s e w o m e n e x p e ri p o rt th is p o ssib ility (32, 33). T u b a l ste r ili z a tio n a n d h y ste re c to m y m a y p ro te c t b y re d u c in g o v a ria n e stro g e n e x p o su re . A lte r i en ce d a slig h t, m a r g in a lly s ig n ific a n t in n a tiv e ly , so m e fo rm s o f tu b a l ste riliz a tio n crease in o v a ria n c a n c e r risk . H o w e v e r, th e h a v e been a sso c ia te d w ith su b se q u e n t in - i ! a stu d y in v o lv e d o n ly fo u r c a se s o f o v a ria n cancer. cre ase cW req u e n cy o f a b n o rm a l o r a n o v u la to ry c y c le s (34, 3 5 ). S in c e e s tim a te d n u m ! i- W o m e n te n d to u n d e rg o tu b a l ste riliz a b er o f o v u la tio n s h a s b ee n a sso c ia te d w ith ig tio n d u rin g th e h e ig h t o f th e ir re p ro d u c tiv e in c re a se d o v a ria n c a n c e r r is k (3 6 , 3 7 ), tu b a l >f ye ars (o n ave rage , in th e e a rly fo u rth d e c ste riliz a tio n m a y p ro te c t b y su p p r e ssin g g ad e o f life , fo r th e p r e s e n t d a ta ). B y c o n o v u la tio n . h trast, h y ste re c to m y is g e n e ra lly p e rfo rm e d n at the e n d o f th e re p ro d u c tiv e y e a rs, p ro b Coffee, tobacco, and alcohol n a b ly a fte r a la rge fra c tio n o f g e n ita l e x p o T h e p re se n t d a ta p ro v id e so m e su p p o rt su re s h a v e a lre a d y o c c u rre d . It is th e re fo re fo r th e h y p o th e sis th a t co ffe e d r iiik in g in - 1238 WHITTEMORE ET AL. c re a se s r isk fo r e p ith e lia l o v a ria n can ce r. C o m p a ris o n o f c a se s v e rsu s th e c o m b in e d g ro u p o f b o th h o sp ita liz e d a n d p o p u la tio n b a se d c o n tro ls su g g e sts th a t c u rre n t o r fo r m e r c o ffe e d r in k e r s e x p e rie n c e d o u b le th e ris k o f n o n d rin k e rs. T h e re la tiv e risk b a se d o n c o m p a riso n o f c ase s v e rsu s o n ly th e p o p u la tio n -b a se d c o n tro ls w a s sm a lle r ( R R = 1.59) a n d d id n o t a c h ie v e s ta tistic a l sig n ific a n c e . N e v e rth e le ss, its m a g n itu d e su g g e sts th a t co ffe e m a y b e im p lic a te d in in c re a se d r is k fo r th e d ise ase . T h e h y p o th e sis is a lso su p p o rte d b y th e d o se -re sp o n se re la tio n n o te d b e tw e e n risk a n d d u ra tio n o f co ffe e d r in k in g a m o n g th o se w h o h a d e v e r d r u n k coffee. C o m p a re d w ith r is k s fo r n o n d rin k e rs, s m o k in g a d ju s te d r is k s in c r e a se d fr o m 1.45 fo r fe w e r th a n 15 y e a r s o f c o ffe e d r in k in g to 3.41 fo r 40 o r m o re y e a rs o f c o n su m p tio n . T re n d s o f in c re a sin g risk w ith in c re a sin g y e a rs o f co ffe e d r in k in g w ere e v id e n t in b o th h o s p ita l a n d p o p u la tio n -b a se d c o n tro l c o m p a r iso n s a n d in b o th sm o k in g -a d ju ste d a n d u n a d ju ste d a n a ly se s. Y e t, n o tre n d s w ere e v id e n t w ith in c re a s in g a m o u n t o f u s u a l co ffe e c o n s u m p tio n in c u p s p e r d a y , o r w ith to ta l life tim e c o n s u m p tio n , e stim a te d b y m u ltip ly in g re p o rte d y e ars o f c o n su m p tio n b y re p o rte d a m o u n t o f u su a l c o n su m p tio n p rio r to ill n e ss o r c e ssa tio n o f coffe e d rin k in g . In ste a d , risk re m a in e d a p p ro x im a te ly c o n sta n t a t tw o to th re e tim e s th a t o f n o n d rin k e rs, re g a rd le ss o f a m o u n t c o n su m e d . T h is la c k o f d o se re sp o n se is c o n s is te n t w ith th e n e g a tiv e fin d in g s o f o th e r stu d ie s (1 7 -2 0 ) fo r tre n d in risk w ith in c re a sin g c o ffe e c o n s u m p tio n a t tim e o f in te rv ie w . T h e ab se n ce o f tre n d n o te d here m a y be d u e to in a c c u ra c ie s in re p o rte d u su a l c o n su m p tio n a s a m e a su re o f ave ra ge co ffe e d r in k in g fre q u e n cy . S u c h in a cc u ra cie s (if s im ila r in m a g n itu d e b e tw e e n c a se s a n d c o n tr o ls) c o u ld m a sk e v id e n ce o f a d o se re sp o n se re la tio n . T h e o b se rv e d p a tte rn s o f tre n d w ith d u ra tio n o f coffe e d r in k in g a n d la c k o f tre n d w ith fre q u e n c y o f c o n s u m p tio n a n d to ta l c o n su m p tio n w o u ld be c o n s is te n t w ith a c a u sa l re la tio n if w o m e n te n d e d to re p o rt th e ir d u ra tio n o f co ffe e d r in k in g m o re a c c u ra te ly th a n th e y re p o rte d th e a m o u n t th e y u su a lly c o n su m e d . T h e p re se n t o b se rv a tio n o f in c re a se d o v a ria n c a n c e r r isk a m o n g coffee d rin k e rs c o u ld b e d u e to se v e ra l so u rc e s o f b ia s. T h e fir s t p o ss ib ility is d iffe re n ce s b etw e e n c a se s a n d c o n tro ls in so m e m e a su re d o r u n m e a s u re d c h a ra c te ristic s th a t are c o rre la te d w it h c o ffe e d r in k in g . B o d y siz e is a n u n lik e ly so u rc e o f su c h c o n fo u n d in g , b e ca u se c a se s a n d c o n tro ls w ere s im ila r in se v e ra l a ss e s sm e n ts o f w e ig h t, h e ig h t, a n d w e ig h t fo r h e ig h t, a s re la te d to b o d y siz e b o th a t a ge 2 0 a n d in re ce n t y e ars. R e la tiv e risk s fo r co ffe e c o n s u m p tio n w ere n o t a lte re d in m u ltiv a ria te a n a ly se s w ith o th e r v a ria b le s fo u n d to be p re d ic tiv e o f o v a ria n c a n c e r. T h e s e in c lu d e p a rity , o ra l c o n tra c e p tiv e u se , a n d d u ra tio n o f c o n tra c e p tiv e -fre e m a rria ge . P o te n tia l c o n fo u n d in g b y d ie ta ry fa c to rs m u s t a ls o b e c o n s id e re d . C r a m e r e t a l. (2 1 ) fo u n d a sta tistic a lly sig n ific a n t tre n d o f in c re a sin g risk w ith in c re a sin g c o n s u m p tio n o f a n im a l fat, a fte r a d ju stin g fo r b o d y w e ig h t a n d p a rity . A b se n c e o f d a ta o n d i e ta ry fa t c o n su m p tio n in th e p re se n t stu d y p re c lu d e s e x a m in a tio n o f p o te n tia l c o n fo u n d in g b y th is factor. S e c o n d , th e stro n g e r a sso c ia tio n n o te d w h e n c a se s w ere c o m p a re d w ith h o sp ita l c o n tr o ls c o u ld b e e x p la in e d b y re d u c e d c o f fee c o n s u m p tio n a m o n g th e se c o n tr o ls a fte r th e o n se t o f s u b c lin ic a l d ise a se m a n ife s ta tio n s. H o w e v e r, a sta tistic a lly sig n ific a n t tre n d in risk w ith y e a rs o f coffe e d r in k in g w a s a lso n o te d w h e n c ase s w ere c o m p a re d w ith p o p u la tio n c o n tro ls. F u rth e rm o re , th e tw o se ts o f c o n tro ls h a d s im ila r p re v a le n c e s o f p rio r re g u la r co ffe e d rin k in g . T h e re fo re , th e p o s s ib ilit y o f se le c tio n b ia s a m o n g h o s p ita l c o n tro ls se e m s u n lik e ly to e x p la in the o b se rv a tio n s. T h ir d , c a se s m a y h a v e o v e rre p o rte d th e ir co ffe e c o n s u m p tio n re la tiv e to c o n tro ls. H o w e v e r, th is e x p la n a tio n a lso se e m s u n lik e ly , in v ie w o f th e a b se n c e o f e vid e n ce w om en >f c o ffe e h e y re sum ed, c re a se d irin k e rs ias. T h e ;n cases unm earre late d an uno e c au se se v e ra l w e ig h t o oth at e risk s e re d in ria b le s can ce r, ce p tive ve -fre e facto rs ' ( 2 1 ) .id o f sum pr body on d i stu d y con n o te d sp ita l d co f; afte r fe sta fic a n t n k in g pared e, th e ences ?fo re , h osn the th e ir :ro ls. ; unence OVARIAN CANCER AND ENVIRONMENTAL EXPOSURES 1239 for su c h r e p o r t in g b ia s in c a s e s ' r e p o r t s o f tobacco a n d a lc o h o l c o n s u m p tio n . T o date, se v e n c a se -c o n tro l stu d ie s h a v e e x am in e d th e re la tio n b e tw e e n c o ffe e a n d o varian c a n c e r (1 6 -2 1 , a n d th e p re se n t study). O f th e se , tw o (th e p r e se n t s tu d y a n d one in Ita ly (1 6 )) fo u n d s ta tis tic a lly s ig n if ican t e le v a tio n s o f r is k a m o n g c o ffe e d r in k ers, w ith re la tiv e r is k s r a n g in g f r o m 1.5 to 2.0. T h e r e m a i n i n g f iv e s t u d ie s , o n e i n Greece (1 7 ) a n d fo u r in th e U n ite d S ta te s (18-21), fo u n d s lig h t ly a n d n o n s ig n ifi c an tly e le v a te d r is k s a s s o c ia t e d w ith c o ffe e co n su m p tio n . T h e se d is p a r itie s a re n o t e a s ily e x p la in e d b y d if fe r e n c e s in c o ffe e c o n stitu e n ts o r in w o m e n a m o n g th e d iffe r e n t study are as. It is d iffic u lt to p ro p o se b io lo g ic a lly p la u sib le c a u s a l m e c h a n is m s t o e x p la in t h e p o s itive a s s o c ia t io n s , i f th e a s s o c ia t io n s a r e n o t due to b ia s o r c h a n ce . C o ffe e c o n s u m p tio n has been sh o w n to in c re a se u r in a ry e x cre tion o f c a t e c h o la m in e s (3 8 -4 1 ), s u g g e s t in g that coffee in c re a se s th e a c tiv ity o f th e adren al m e d u lla . C o ffe e c o n s u m p tio n a ls o m ay in c re a se a d r e n a l p r o d u c tio n o f a n d ro ste n e d io n e . S in c e p e r ip h e r a l a r o m a t iz a t io n o f a n d ro ste n e d io n e to e stro n e fo rm s th e p rim ary so u rc e o f e stro g e n in p o s t m e n o p au sal w o m e n (4 1 ), c o ffe e c o n s u m p tio n m ay a lte r o v a ria n c a n c e r r is k b y in c re a sin g estrogen p ro d u c tio n a fte r th e m e n o p a u se . Su ch a m e c h a n ism is sp e c u la tiv e , h o w e v e r, in v ie w o f th e s p a r s it y o f d a t a im p lic a t in g e stro ge n s in o v a ria n c a r c in o g e n e s is (32, 33). C ig a re tte s m o k in g w a s a ss o c ia te d n o n sig n ific a n tly w ith re d u c e d o v a r ia n c a n c e r risk. T h is f in d in g a g r e e s w it h t h o s e o f o t h e r case -co n tro l stu d ie s (1 6 , 17, 2 1 ) b u t d is agrees w ith a n in c re a se d o v a ria n c a n c e r risk fo u n d in a p ro sp e c tiv e s tu d y o f w o m e n w ho sm o k e (22). T h e p re se n t d a ta sh o w n o re la tio n b e tw e e n o v a r ia n c a n c e r r is k a n d d u ra tio n o f c ig a re tte s m o k in g . T h e d a ta a lso p ro v id e n o e v id e n c e fo r a n y re la tio n b e tw e e n a lc o h o l c o n s u m p tio n an d o v a ria n can ce r risk , w h ic h is c o n siste n t w ith re su lts o f o th e r s tu d ie s t h a t h a v e e x a m in e d th is issu e (1 7 , 20, 2 1 ). O n e la rg e stu d y (2 3 ) fo u n d th a t w o m e n w h o c o n su m e d 20 o r m ore d rin k s p e r w ee k h a d h a lf th e risk o f w o m e n w h o d id n o t d rin k . W e a lso fo u n d re d u c e d r isk a sso c ia te d w ith su c h h e a v y d rin k in g , b u t th e a sso c ia tio n d id n o t a ch ie v e s ta tis tic a l sig n ific a n c e . T h e la c k o f sig n ific a n c e m a y b e d u e to sm a ll n u m b e rs in th e p re se n t stu d y ; th u s, the re la tio n o f a lc o h o l c o n s u m p tio n to o v a ria n c a n c e r s h o u ld b e e x a m in e d in la r g e r se rie s. References 1. Blejer JP, Arlon R. Talc: a possible occupational and environmental carcinogen. J Occup Med 1973;15:92-7. 2. Graham J, Graham R. Ovarian cancer and asbes tos. Environ Res 1967;1:115-18. 3. Henderson WJ, Joslin CAF, Turnbull AC, et al. Talc and carcinoma of the ovary and cervix. J Obstet Gynaecol Br Cwlth 1971;78:266-72. 4. Longo DL, Young RC. Cosmetic talc and ovarian cancer. Lancet 1979;2:349-51. 5. Newhouse ML. Cosmetic talc and ovarian cancer. (Letter). Lancet 1979:2:528. 6. Parmley TH, Woodruff JD. The ovarian meso thelioma. Am J Obstet Gynecol 1974;120:234-41. 7. Roe FJC. Controversy: cosmetic talc and ovarian cancer. (Letter). Lancet 1979;2:744. 8. Cramer DW, Welch WR, Scully RE, et al. Ovarian cancer and talc: a case-control study. Cancer 1982;50:372-6. 9. Hartge P, Hoover R, Lesher LP, et al. Talc and ovarian cancer. (Letter). JAMA 1983:250:1844. 10. Koch M, Starreveld AA, Hill BG, et al. The effect of tubal ligation on the incidence of epithelial cancer of the ovary. Cancer Detect Prev 1984; 7:241-5. 11. Annegers JF, Strom H, Decker DG, et al. Ovarian cancer incidence and case-control study. Cancer 1979;43:723-9. 12. Cramer DW, Hutchison GB, Welch WR, et al. Determinants of ovarian cancer risk. I. Reproduc tive experiences and family history. JNCI 1983; 71:711-16. 13. McGowan L, Parent L, Lednar W, et al. The woman at risk for developing ovarian cancer. Gy necol Oncol 1979;7:325-44. 14. Winder EL, Dodo H, Barber HRK. Epidemiology of cancer of the ovary. Cancer 1969;23:352-70. 15. Weiss NS, Harlow BL. Why does hysterectomy without bilateral oophorectomy influence the sub sequent incidence of ovarian cancer? Am J Epi demiol 1986;124:856-8. 16. La Vecchia C, Francheschi S, Decarli A, et al. Coffee drinking and the risk of epithelial ovarian cancer. Int J Cancer 1984;33:559-62. 17. Tzonou A, Day NE, Trichopoulos D, et al. The epidemiology of ovarian cancer in Greece: a casecontrol study. Eur J Cancer Clin Oncol 198420.T045-52. 18. M ille r D R . R o sen b e rg L . K a u fm a n D W , e t al. 1240 WHITTEMORE ET AL. Epithelial ovarian cancer and coffee drinking. Int J Epidemiol 1987;16:13-17. 19. Hartge P, Lesher LP, McGowan L, et al. Coffee and ovarian cancer. Int J Cancer 1982;30:531-2. 20. Byers T, Marshall J, Graham S, et al. A casecontrol study of dietary and nondietary factors in ovarian cancer. JNCI 1983;71:681-6. 21. Cramer DW, Welch WR, Hutchison GB, et al. Dietary animal fat in relation to ovarian cancer risk. Obstet Gynecol 1984;63:833-8. 22. Doll R, Gray R, Hafner B, et al. Mortality in relation to smoking; 22 years' observations on female British doctors. Br Med J 1980;1:967-71. 23. Gwinn ML, Webster LA, Lee NC, et al. Alcohol consumption and ovarian cancer risk. Am J Epi demiol 1986;123:759-66. 24. Wu ML, Whittemore AS, Paffenbarger RS Jr, et al. Personal and environmental characteristics re lated to epithelial ovarian cancer. I. Reproductive and menstrual events and oral contraceptive use. Am J Epidemiol 1988;128:1216-27. 25. Berry G, Newhouse ML. Mortality of workers manufacturing friction materials using asbestos. Br J Ind Med 1983;40:1-7. 26. Egli GE, Newton M. The transport of carbon particles in the human female reproductive tract. Fertil Steril 1961;12:151-5. 27. Henderson WJ, Joslin CAF. Turnbull AC, et al. Talc and carcinoma of the ovary and cervix. J Obstet Gynaecol Br Cwlth 1971;78:266-72. 28. Henderson WJ, Hamilton TC, Griffiths K. Talc in normal and malignant ovarian tissue. Lancet 1979:1:499. 29. Cattanach J. Oestrogen deficiency after tubal li gation. Lancet 1985;1:847-9. 30. Corson SL, Levinson CJ, Batzer FR, et al. Hor monal levels following sterilization and hysterec tomy. J Reprod Med 1981;26:363-9. 31. Beavis ELG, Brown JB, Smith MA. Ovarian func tion after hysterectomy with conservation of the ovaries in pre-menopausal women. J Obstet Gy naecol Br Cwlth 1969;76:969-78. 32. Cramer DW, Welch WR. Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis. JNCI 1983;71:717-21. 33. Weiss NS, Lyon JL, Krishnamurthy S, et al. Non contraceptive estrogen use and the occurrence of ovarian cancer. JNCI 1982;68:95-8. 34. DeStefano F, Perlman JA, Peterson HB, et al. Long-term risk of menstrual disturbances after tubal sterilization. Am J Obstet Gynecol 1985; 152:835-41. 35. Radwanska E, Headley SK, Dmowski P. Evalua tion of ovarian function after tubal sterilization. J Reprod Med 1982;27:376-84. 36. Casagrande JT, Louie EW, Pike MC. et al. "In cessant ovulation" and ovarian cancer. Lancet 1979;2:170-3. 37. Risch HA, Weiss NS, Lyon JL, et al. Events of reproductive life and the incidence of epithelial ovarian cancer. Am J Epidemiol 1983:117:128-39. 38. Levi L. The effect of coffee on the function of the sympathoadrenomedullary system in man. Acta Med Scand 1967;181:431-8. 39. Bellet S, Roman L, DeCastro O, et al. Effect of coffee ingestion on catecholamine release. Metab olism 1969;18:288-91. 40. Edman CD, MacDonald PC. The role of extra glandular estrogen in women in health and dis ease. In: James VHT, Serio M, Giusti G, eds. The endocrine function of the human ovary. New York: Academic Press, 1976:135-40. 41. Klimmer F, Neidhart B, Legeler T, et al. Influence of coffee on the excretion of noradrenaline and adrenaline in urine. Int Arch Occup Environ Health 1984;54:325-34. A m e r ic a n Jc C o p y rig h t ', ; A ll rights rese THE El iY C N. ant / lipi 191 ma yet intc nai cht the les the qu; the thn lun Th. cht sm typ ext ant set t Recent 5) and al Received final form A Epidem Cniversity . Honolulu, H G o o d m an .) This wor CA-33619 ai 55424 from ment of Her The auth. irig hospital Medical Ce Medical Cer Cls H ospital G eneral Ho> American Journal of y d h EPIDEMIOLOGY Volume 136 Number 10 C opyright 1992 by The John Hopkins U niversity School of Hygiene and Public H ealth November 15. 1992 Sponsored by the S o ciety fo r E pidem iologic Research ORIGINAL CONTRIBUTIONS Characteristics Relating to Ovarian Cancer Risk: Collaborative Analysis of 12 US Case-Control Studies I. Methods die. MU ID TX v 0 A :gle j[e\ CA !A Doblic inai ' E. ted 55- Alice S. W hittemore.' Robin Harris.' Jacqueline Itnyre.' Jerry Halpern : and the Collaborative Ovarian Cancer Group3 Data from 12 US case-control stu d ies of ovarian cancer, conducted during the period 1 9 5 6 -1 9 8 6 and representing som e 3.000 c a se s and 10.000 controls were pooled and reanalyzed. Sep arate an a ly ses w ere con d u cted for four sub grou p s of the pooled data: invasive epithelial ovarian cancers in white wom en: epithelial ovarian cancers of low malignant potential in white w om en, epithelial ovarian ca n cers in black wom en, and nonepithelial ovarian cancers. This paper g iv es a brief description of the participating stu d ies and d escrib es the m ethods u sed in the collaborative analysis. A m J E pidem iol 1992:136:1175-83. case-control studies; methods: ovarian neoplasm s T h is is th e first in a sc r ie s o f a rtic le s d e sc rib in e a c o lla b o r a m e c o m b in e d a n a ly s is o l'd a ta fro m 12 c a se -c o n tro l stu d ie s o f o v a rla n c a n c e r c o n d u c te d in the U n ite d S ta te s Received ;c: c -cica tic " August 2 ' 1991 and m linai 'orm July 23 1992 Divisen o ' Eoia em oog. Deoanment ot n e am Re search and Pplic, Stam e ': universi!, Scnooi oi Meoicme Slantorc CA Divisio- S ostai stcs Deoanment ot Heann Re search ano " ove. Sta-'C 'C Uni.e'Sit,. School ot Medicine Staniora CA Members of me Conacoran.e Ovarian Cancer Group D- John T Casagranae Department of Preventive Medr cme University o ' Southern California Los Angeles. C A Dr Danie' Crame- Depanmem of Obstetrics and Gyne coiogy Bngnam ana Women s Hospital Boston MA. Dr Patooa Hange Environmental Epidemiology Brancn Na nonai Cancer institute Betnesaa M D D ' Jenniler L Keisey Division ol Epidemiology Deoanment of Health Research ana P o 'o Sianiora University School ol Medi cine. Slantorc C A Dr Marion Lee Depanment ol Epide midiogy University ol Cantomia San Francisco. San Fran cisco C A Dr Nancy C Lee Women s Health and Fenility Branch Division ol Reproauclive Health. Ceniers tor Dis ease Contro. Aya-ta GA Dr. Joseph L, Lvon Deoanment of Fami'y anc Community Medicine The University of Utah Medicai Cente- Sa : Lane City. UT Dr James R Marshall Department o ` Soda. ano Preventive Medicine State Uni versuv of Nev. Yorv, ai Buttalo Schooi ol Medicine. Buttalo New Yorn D r Lar-. McGowan Division ot Gvnecoioaic Oncoiogy Deoanment ot Obstetrics ano Gynecoiogy George Washington University Medicai Center Wasnmg ton D C Dr Pniiip C Nasca. New York Siate Department ot Health Bureau ol Cancer Epidemiology School oI Public Heaim Deoanment c ` Eoidemioiogv Albany NY Dr Raipn S Paflenparge- J ' Division ot Epidemiology Depanmem ot Healtn Researc- and Policy Stanford University School ol Meoicine Stanford. CA. Dr Lynn Rosenperg Slone Epidemiology Uni: School ol PuDlic Health Bosion uni versity School ot Meaicme. Brookline MA and Dr Noel S Weiss. Deoanmem ol Epidemiology. School ol Public Health and Community Medicine University ol Washing ton. Seattle WA Pro/ect Consultant Dr Genrose D Co pley. Extramural Programs Division ol Cancer Ecology National Cancer institute Bethesda MD 1175 ul s w as u irc c s j w as o f i he tio n a l ancer ra io rs ib ility stu d ses or le e.\n o sis. .h e lia l n a lig [ a lso scle c n iro ls con 'd fo r n a lig Jed .a r eopoI ) ex tes o r t a n 'S a n d M ost. a ith a 1o t h e r is the tc h e d e and so n ic used it d ie d ir in aly tie ita by . d is- We w ere d not es or a h ie s 'tro l Methods of Analysis 1177 T A B L E 1. C ase-control stu dies of epithelial ovarian cancer a m o n g U S white women Auinors reference no ) Cases C o n tro ls invasive Low malignant potential Year ot diagnosis Place ot diagnosis No.* Source B ye rs e i al (1) Hiiaretn et al (2) M c G o w a n et al (3) W u ei a! |4) R o se noe rg et al (5) Hartge et al. (6) C asaarande et al (7) Cramer et ai (8) N asca ei al (9) W e iss et al (10) 196 59 133 111 115 220 133 177 314 269 C A S H } group |1 1) Whittemore e t ai 11 2 1 303 167 1 9 5 6 -1 9 6 3 Buffalo. NY 795 Case hospitals 3 19 76 -19 79 Connecticut 1.068 Case hospitals 33 1 9 7 4 -1 9 7 7 Washington. D C 165 C a se hospitals 19 75 -19 77 San Francisco Bay 482 Case hospitals Area. C A 8 1 9 7 6 -1 9 8 0 Eastern U S cities} 486 C ase hospitals 41 1978-1981 Washington. DC 288 Case hospitals 1973-1976 Los Angeles. C A 134 C ase neighbor- hoods 41 1978 -19 81 Boston. M A 229 Town directories 27 1 9 7 7 -1 9 8 0 New York State 694 Motor vehicle files 23 1 9 7 5 -1 9 7 9 Utah Washington 700 Household and R D D J phone surveys 107 1 9 8 0 -1 9 8 2 8 S E E R } areas 3.542 R D D phone sur- veys 4 4 1 9 8 3 -1 9 8 6 San Francisco Bay 310 Group 1: R D D Area C A phone surveys Group 2: case hospitals Total 2.197 327 8 893 Numae' mciuaea m me present analysis T ino u cm a B oston M assachusetts New York Nev. York PnuaoeiDnia. Pennsylvania and Baltimore. Maryland, t POD 'a no o m dig i dialing CASH Cancer anc Steroid Horm one Study. SEER Surveillance Epidemiology, and End Results OfOCrarr Atlanta Georgia Detroit Michigan Seattle. W ashington San Francisco Bay Area. California: Iowa: Connecticut: New Mexico, ana u ta r w o m e n w h o h u d o r m ig h t liu v e h u d u b ila terul o o p h o r e c t o m y , T u b le I sh o w s th e n u m bers o f w lu te w o m e n w ith e p ith e lia l o v a ria n c a n c e r (c la ssifie d by tu m o r b e h a v io r) a n d w h ite c o n tr o l w o m e n in c lu d e d in th e a n a ly sis ANALYSIS A n a l y s i s i n v o l v e d t h e f o l l o w 1n u t a s k s : 1 ) c o n s tr u c t in g th e b a s ic v a ria b le s le .g.. ra c e a n d su b ty p e o f d ise a se ) n e e d e d to o rg a n iz e the d a ta in to su b se ts fo r se p a ra te a n a lv sis: 2 ) c h o o s in g the m a jo r h y p o th e se s to be tested: 3 ) d e fin in g th e v a r ia b le s n e e d e d to test th e se h vp o th e se s: 4 ) re v ie w in g the in d iv id u a l q u e stio n n a ire s a n d c o d e b o o k s to d e te rm in e the in fo r m a tio n av a ila b le o n the v a ria b le s o f in te re st. 5 1 u s in g th is in fo r m a t io n to id e n tify a list o f "w o r k in g v a r ia b le 's " to b e u se d in a ll a n a ly se s: 6 i id e n tify in g th e p o sitio n s o n e a ch o f the d a ta ta p e s o f th e in fo rm a tio n n e e d e d to a sse m b le e a c h v a ria b le a n d w ritin g a se p a ra te p r o g r a m to e x tract th is in fo r m a tio n fro m e a ch tape: 7) p re p a rin g d e sc rip tiv e sta tistic s fo r e a c h v a r ia b le a n d e a c h stu d y in se arch o f o u tly in g o b se rv a tio n s a n d o u tly in g stu d ie s: 8 ) m e r g in g th e e x tra c te d d a ta in to a c o m p o s ite file c o n t a in in g d a ta fo r a ll v a r i a b le s a n d a ll stu d ie s: 9 ) e x tra c tin g su b se ts o f d a ta fo r se p a ra te a n a ly se s: 10) a ss e m b lin g a list o f re g re s sio n m o d e ls fo r e a c h v a ria b le : I I ) c o n d u c tin g stu d y -sp e c ific a n d c o m b in e d re g re ssio n a n a ly se s: 12) d o c u m e n tin g a n d c a ta lo gin g o u tp u t: a n d 13) d isc u ssin g a n d in te rp re tin g re su lts a n d p la n n in g fu rth e r fo l lo w -u p re gre ssio n s. T h e se ste p s are d isc u sse d b e lo w . Data organization W c first c o p ie d to a n I B M 3 0 9 0 m a in fra m e c o m p u te r ( IB M . P o u gh k e e p sie . N Y ) o r ig in a l d a ta file s fr o m th e 12 stu d ie s. T h is re p re se n te d d a ta fo r so m e 1 3 .6 0 0 su b je cts a n d c o m p r is e d 3 .9 0 0 v a ria b le s, ra n g in g fro m Methods of Analysis 1179 a m p le s o f w o rk in g v a ria b le s in c lu d e a w o m a ll re le v a n t re g re ssio n s. F u rth e r, w o m e n a n 's to ta l n u m b e r o f te rm p re g n a n c ie s, d e w ith u n k n o w n v a lu e s o f a v a ria b le w ere fin e d a s p r e g n a n c ie s o f at le a st 2 0 w e e k s d e le te d fro m a ll re g re ssio n s c o n ta in in g th at g e s t a t io n a n d c o d e d a s 0 . 1...... 5. > 6 . a n d v a ria b le . T h u s , to ta l c a se a n d c o n tro l n u m h e r to ta l n u m b e r o f fa ile d p re g n a n c ie s, d e b e rs v a ry a c ro ss re gre ssio n s, a n d n u m b e r s fin e d a s m isc a rria g e s, in d u c e d a b o r tio n s, e c p re se n te d in th e fo llo w in g p a rts v a n a c ro ss to p ic p re g n a n c ie s, a n d stillb irth s, a n d c o d e d ta b le s. sim ila rly . O n c e a w o r k i n g v a r i a b l e w a s i d e n t i f i e d , Statistical analysis the in d iv id u a l q u e stio n n a ire s a n d c o d e b o o k s w ere re v ie w e d to d e te rm in e a c o m S e v e ra l p itfa lls m a y re su lt fro m p o o lin g m o n d e fin itio n a n d c o d in g sch e m e . T h is d a ta fro m se p a ra te stu d ie s w ith d iffe rin g p r o so m e tim e s in v o lv e d d iffic u lt tra d e -o ffs b e to c o ls a n d d iffe rin g e x p o su re p re v a le n c e . tw e e n c o d in g d e ta il a n d stu d y in c lu s io n in T h e v a ria b le "s t u d y " c o u ld re p re se n t b o th a o rd e r to o b ta in d a ta fro m a s m a n s stu d ie s stro n g c o n fo u n d in g fa c to r a n d a n effect in a s m u c h d e ta il a s p o ssib le . F o r e x a m p le , m o d ifie r , it is a p o te n tia l c o n fo u n d e r b e w e d e fin e d fa ile d p re g n a n c ie s to in c lu d e s till c a u se b o th th e c a se .c o n tro l ra tio a n d the b irth s (w h ic h a lso are te rm p re g n a n c ie s) in p re v a le n c e fo r a p a rtic u la r e x p o su re m a y o rd e r to use d a ta fro m a stu d y th at h ad v a ry fr o m s t u d y to stu d y . It is a p o te n tia l a gg re g a te d stillb irth s w ith th e o th e r ty p e s o f effect m o d ifie r b e c a u se o d d s ra tio s m a y vary p re g n a n c y fa ilu re . W h e n th e c o n v e n tio n o f fro m o n e stu d y p o p u la tio n to an o th e r. u sin g the c o a rse st d e fin itio n re su lte d in u n T h e re fo re , a n a n a ly sis th at p o o ls d a ta a c ro ss a c c e p ta b le lo ss o f sp e c ific ity o r d e ta il fo r to o stu d ie s c o u ld y ie ld se rio u sly m isle a d in g re m a n y stu d ie s, w e c o n d u c te d a d d itio n a l, su lts. m o re d e ta ile d a n a ly se s u sin g o n ly th o se W e u se d se v e ra l stra te g ie s to a d d re ss th e se stu d ie s w ith a p p ro p ria te d a ta . p itfa lls. F irst, w e stra tifie d a ll a n a ly s e s jo in t ly T o p ro d u c e th e w o r k in g v a ria b le s, se p a b y s t u d y a n d re fe re n c e a g e (< 2 5 . 2 5 - 2 9 ...... rate p r o g r a m s w e re w ritte n to e x tra c t th e 7 5 - 7 9 . a n d > 8 0 y e a r s ) . S t u d ie s 1 0 a n d 11 d a ta fr o m th e in d iv id u a l s tu d s file s. F o r w e re fu rth e r stra tifie d b y th e ir c o n stitu e n t c o m p le x v a ria b le s, su c h a s m e n o p a u s a l sta stu d y ce n te rs. W e u se d c o n d itio n a l lo g istic tu s o r e s tim a te d a g e a t la st o v u la t io n , flo w re g re ssio n (2 2 ). im p le m e n te d o n E G R E T c h a rts (e.g.. fig u re l) h e lp e d to e n su re th a t so ftw a re (2 3 ). to e stim a te o d d s ra tio s a n d je th e (d iffe re n t s tu d ie s c o n tr ib u t e d c o m p a r a b le c a lc u la te (tw o -ta ile d ) s ig n ific a n c e le v e ls a n d in fo rm a tio n . O n c e created, the w o rk in g v a ri c o n fid e n c e in te rv a ls. T h is a p p ro a c h h a s tw o a b le s w e re e d ite d fo r su b tle in te rstu d y in a d v a n ta g e s o v e r an u n c o n d itio n a l a n a ly sis c o m p a tib ilitie s. D e sc rip tiv e sta tistic s w e re c o n ta in in g d u m m y v a ria b le s fo r th e stu d ie s, u se d to id e n tify o u tlie rs a n d o th e r p ro b le m s. th e a g e g r o u p s , a n d th e ir in te r a c tio n s : It iv ie s ch re S o m e o f th e se w e re re so lv e d lo c a lly : o th e rs re q u ire d d isc u ssio n w ith th e c o lla b o r a tin g in v e stiga to rs. S o m e in v e stig a to rs o c c a s io n a lly h a d to re v ie w th e ir o rig in a l d a ta to c o n firm o r re co d e q u e stio n a b le v a lu e s o r to a v o id s u n w ie ld y o u tp u t o f re gre ssio n c o e ffi c ie n ts fo r th e re su ltin g 2 0 0 -p lu s a ge -stu d > v a ria b le s w h ic h a re n o t o f p rim a r y in te re st, a n d it a llo w s in s p e c tio n o f n u m b e r s o f s u b je c ts w h o fa ile d to c o n trib u te to a g iv e n ice ter p ro v id e fu rth e r d e ta il. M o r e th a n 9 0 w o r k in g v a ria b le s w ere c o n stru cte d . a n a ly s is b e c a u se th e ir a ge -stu d y stra tu m la c k e d case s o r c o n tr o ls o r d isc o rd a n t e x p o su re s. H o w e v e r , e ith e r c o n d itio n a l o r stra ti of' seJte Data processing and management O n c e cre a te d a n d e d ite d , th e w o r k in g v a r i fie d u n c o n d it io n a l re g re ssio n c o u ld b e u se d (2 2 ): w e fo u n d th a t th e tw o p ro d u c e d n e a rlv id e n tic a l e stim a te s fo r o d d s ra tio s a n d th e ir jg- a b le s w ere u se d to d e riv e c a te g o ric a l a n d sta n d a rd e rro rs w h e n p e rfo rm e d o n the sam e rol o th e r v a ria b le s fo r re gre ssio n s. W h e n a n a d a ta . T h e sta n d a rd e rro rs a n d c o n fid e n c e if ly z in g a to p ic , w e in c lu d e d o n ly th o se stu d ie s in te rv a ls p ro d u c e d b y th ese re g re ssio n s d o -V th at h a d d a ta fo r a ll v a ria b le s o c c u r r in g in n o t re fle ct v a ria n c e b e c a u se o f in te rs tu d y e-controi a lu a te d se ssio n m o d e l, b la m e d in te re si stu d ie s tin g the a te s Tor o sp ita l.del a n d le i c o n 'coded Methods of Analysis 1181 o n e fo r p a ro u s w o m e n a n d ze ro fo r n u llip a ro u s w o m e n ), w h ile th e e x p a n d e d m o d e l c o n ta in e d P A R a n d fiv e a d d it io n a l d u m m y v a r i a b l e s o f t h e f o r m P A R x S T U D Y . / . ./ = 1......5. w h e re S T U D Y j is c o d e d o n e if a w o m a n p a rtic ip a te d in s tu d y j a n d 0 o th e r w ise . T h e c o e ffic ie n t fo r P A R x S T U D Y j g iv e s the p ro p o rtio n a l a m o u n t th a t th e o d d s ra tio fo r p a rity in stu d y i d ifT e rs fr o m th at o f stu d y 6. a n a rb itra rily c h o s e n re fe re n t stu d y . A n o v e ra ll test o f th e n u ll h y p o th e s is o f n o s tu d y h e te ro g e n e ity is p r o v id e d by th e lik e lih o o d ra tio sta tistic b a se d o n th e m a x i m iz e d lo g -lik e lih o o d s o f sm a ll a n d e x p a n d e d m o d e ls. U n d e r th e n u ll h y p o th e sis, th is sta tistic h a s a p p r o x im a te ly a c h i-s q u a r e d d is tr i b u tio n w ith d e g re e s o f fre e d o m e q u a l to o n e le ss th a n th e n u m b e r o f s tu d ie s b e in g e x a m in e d (22). In a d d itio n to e x a m in in g h e te ro g e n e ity a c ro ss stu d ie s o f o d d s ra tio s fo r a ll m a jo r v a ria b le s a sso c ia te d w ith risk o f in v a s iv e e p ith e lia l o v a r ia n c a n c e r, w e a ls o e v a lu a te d su ch o d d s ra tio h e te ro ge n e ity w ith re sp e ct to p a rity , h isto ry o f o ra l c o n tr a c e p tiv e use. a n d 1 0 -year strata o f re fe re n ce age. O th e r effect m o d ific a tio n w a s e v a lu a te d w h e n m o tivate d by a sp e c ific b io lo g ic a l m e c h a n is m su gge ste d by o n e o r m o re o f the c o lla b o ra t in g in v e stiga to rs. T h ird , w e c o n d u c te d tw o se ts o f c o m b in e d a n a ly se s: o n e fo r th e six stu d ie s th a t in v o lv e d h o s p i t a l c o n t r o l s ( s t u d i e s 1 - 6 i n t a b le 1. h e re afte r c a lle d h o sp ita l stu d ie s (1 -6 )) a n d o n e fo r th e six stu d ie s th at in v o lv e d r a n d o m d ig it d ia l o r n e ig h b o r h o o d c o n tr o ls (stu d ie s 7 -1 2 , h e re a fte r c a lle d p o p u la tio n stu d ie s (7 12)). S tu d y 12 u se d b o th h o s p it a l a n d p o p u la tio n c o n tro ls: w e o m itte d th e h o sp ita l c o n tro l d a ta . S o ftw a re lim it a t io n s o n the m a x im u m n u m b e rs o f study su b je cts an d v a ria b le s in a re g re ssio n m a n d a te d th e sp lit in to h o sp ita l a n d p o p u la tio n stu d ie s, w h ic h fo rtu ito u sly a llo w e d u s to a sse ss th e stre n g th a n d c o n siste n c y o f a n a ss o c ia tio n by c o m p a rin g tw o in d e p e n d e n t se ts o f o d d s ra tio e stim a te s. T h e sp lit a lso p ro v id e d a n o p p o r tu n ity to c o m p a r e o d d s ra tio e stim a te s a n d p re v a le n c e o f c h a ra c te ristic s b e tw e e n h o s p i tal a n d p o p u la tio n c o n tro ls. O v e r a ll, w e fo u n d g o o d a gre e m e n t b e tw e e n o d d s ra tio s o b ta in e d b y th e tw o ty p e s o f stu d y : su c h a g re e m e n t le n d s s u p p o r t to b o th stu d y d e sig n s. F in a lly , w e in c lu d e d y e a r o f b irth a s a c o n tin u o u s v a ria b le in a ll re g re ssio n s to c o n tro l m o r e p re c ise ly fo r re fe re n c e a ge a n d to c o n tro l fo r a n y c a se -c o n tro l d iffe re n c e s in y e ar o f in te rv ie w th a t m ig h t b ia s c o m p a r is o n o f te m p o ra l v a ria b le s su c h a s o ra l c o n tra c e p tiv e use. Limitations and strengths C o m b in e d a n a ly se s su c h a s th is sh a re s o m e o f th e p itfa lls o f m e t a -a n a lv s is (i.e.. th e re v ie w a n d sy n th e s is o f p u b lis h e d fin d in g s). D e sp ite th e c o m m o n d e fin itio n s u se d to re c o d e v a ria b le s a n d a ll e ffo rts to e n su re in te r stu d y c o m p a ra b ility , the d a ta a v a ila b le to a c o m b in e d a n a ly s is n e v e rth e le ss d e riv e fro m q u e stio n s w h o se w o rd in g v a rie d a c ro ss stu d ies a n d th u s c o u ld h a v e e lic ite d d iffe re n t re sp o n se s. In te rp re ta tio n o f th e fin a l re su lts a lso is h a m p e re d b y a n y d e fe c ts in th e o rig in a l stu d ie s, in c lu d in g se le c tio n b ia s in th e e n ro llm e n t o f case s a n d c o n tro ls a n d c o n fo u n d in g by u n m e a su re d o r im p re c ise ly m e a su re d v a ria b le s. P o o lin g d a ta fro m se v e ra l stu d ie s th a t h a v e th e s a m e ty p e s o f b ia s c a n p ro d u c e re la tiv e risk e stim a te s th at a re sta tistica lly h ig h ly s ig n ific a n t b u t n e v e rth e le ss u n c o n v in c in g o f a n u n d e r ly in g c a u s a l re la tio n . (S e e re fe re n ce 2 6 fo r fu rth e r d is c u s sio n o f th e so u rc e s o f b ia s in p o o le d a n a ly se s o f e p id e m io lo g ic d a ta .) O n the o th e r h a n d , a c o m b in e d a n a ly sis o ffe rs se v e ra l b e n e fits. It p r o v id e s la rg e s a m p le siz e s fo r e x a m in in g e ffe c ts o f ra re e x p o su re s. in te ra c tio n s a m o n g e sta b lish e d o r su s p e cte d risk facto rs, c o n siste n c y o f a sso c ia tio n s p re v io u sly su g g e ste d b y so m e stu d ie s b u t n o t c o n firm e d b y o th e rs, a n d effects o f risk fa c to rs b y su b ty p e o f d ise ase . F o r e x a m p le . p a n I V ( 15 ) in th is se rie s d e sc rib e s th e fin d in g th a t o d d s ra tio s re la tin g e p ith e lia l o v a ria n c a n c e r risk to p re g n a n c y a n d o ral c o n tra c e p tiv e u se d iffe r b e tw e e n y o u n g e r a n d o ld e r w o m e n , a fin d in g th a t h a s n o t e m e rge d fro m an y in d iv id u a l study a n d A case r. Am J 1. " Inces:et 1979; R. et al. eprodue.1 Cancer An epin cancer ol 1984: dence o f .1 contra- > o f the nal Instiopmeni ssociated ;d 1987; Jr. et al. s related o talcum J Epide- Characillabora- M. In- Charactllaboralies. III. ntial in >92; 136: CharacillaboraIV. The Am J I. Charcol la bo dies. V. il l Can- et al sk: colol stud y 1992. Xanan Cancer al. The ith the impli- I. et al a n . Br research, Vol I . The analysis o f case-comrol studies. (IARC scientific publication no. 32). Lyon: Inter nationa! Agency for Research on Cancer. 1980. 23. Mauntsen R EGRET software program Seattle. WA: Statistics and Epidemiology Research Corpo ration. I486. 24 Der Simoman R. Laird N. Meta-analvsis in clinical trials. J Control Clin Trial 1986:7:177-88. 75. M iller R The jackknife-- a review Biometrika 1974:61:1-17, ' 26. Clayton D The EURODE.M collaborative re analysis o f case-control studies of Alzheimer's dis ease: some methodological considerations. Int J Epidemiol 1991:20(Suppi. 2LS2-4.