Original article. Occupational exposure to endocrine-disrupting compounds and biliary tract cancer among men

Transcription

1 Originl rtile Snd J Work Environ Helth 2007;33(5): Ouptionl exposure to endorine-disrupting ompounds nd biliry trt ner mong men by Wolfgng Ahrens, PhD, 1 Chinr Mmbetov, MD, 2 Niole Bourdon-Rverdy, 3 Agustin Llopis- González, MD, 4, 8 Psl Guénel, MD, 5 Lennrt Hrdell, MD, 6 Frno Merletti, MD, 7 Mri Morles-Suárez-Vrel, MD, 4, 8, 9 Jorn Olsen, PhD, 10 Håkn Olsson, MD, 11 Mogens Vyberg, MD, 12 Pol Zmbon, MD 13 Ahrens W, Mmbetov C, Bourdon-Rverdy N, Llopis-González A, Guénel P, Hrdell L, Merletti F, Morles-Suárez- Vrel M, Olsen J, Olsson H, Vyberg M, Zmbon P. Ouptionl exposure to endorine-disrupting ompounds nd biliry trt ner mong men. Snd J Work Environ Helth 2007;33(5): Objetives This study investigted the ssoition between ner of the extrhepti biliry trt nd exposure to endorine-disrupting ompounds. Methods Altogether 183 men with histologilly onfirmed rinom of the extrhepti biliry trt nd 1938 mthed ontrols were interviewed between 1995 nd 1997 in the frme of n interntionl multienter se ontrol study in six Europen ountries (Denmrk, Frne, Germny, Itly, Spin, nd Sweden). Selfreported job desriptions were onverted to semiquntittive vribles (intensity, probbility, nd durtion of exposure) for 14 endorine-disrupting ompounds. The ses were ompred with 1421 popultion ontrols nd 517 olon denorinom ptients. Odds rtios (OR) nd 95% onfidene intervls (95% CI) were obtined with unonditionl logisti regression nd djusted for ge, ountry, nd gllstones. Results Ouptionl exposure to endorine-disrupting ompounds resulted in n OR of 1.4 (95% CI ) with no dose effet reltionship for umultive exposure (low: OR 1.3, 95% CI ; medium: OR 1.5, 95% CI ; high: OR 1.4, 95% CI ) (only index prtiipnts). The elevted risk ws restrited to extrhepti bile duts nd mpull Vteri (OR 1.7, 95% CI ). The djusted OR for ner of the extrhepti biliry trt fter exposure to polyhlorinted biphenyls ws 2.8 (95% CI , only index prtiipnts). Conlusions The dt show some ssoitions between exposure to endorine-disrupting ompounds in the workple nd the risk for ner of the extrhepti biliry trt mong men, prtiulrly for the extrhepti bile dut nd mpull of Vter. Polyhlorinted biphenyls ould possibly be strong risk ftor. Key terms epidemiology; rinom of the extrhepti bile dut; se ontrol study; gllbldder rinom; xenoestrogen. Cner of the extrhepti biliry trt onsists of gllbldder rinom, rinom of the extrhepti bile dut, nd ner of the mpull of Vter. These ners our the most often in the elderly nd re often ftl (1 3). They gretly differ s to their gender distribution (1, 2, 4). Gllbldder rinom is more ommon in women 1 Bremen Institute for Prevention Reserh nd Soil Mediine, University of Bremen; Institute for Medil Informtis, Biometry nd Epidemiology, University of Essen, Germny. 2 Tshkent Reserh Institute of Hemtology nd Trnsfusiology, Tshkent, Uzbekistn. 3 Registre du ner de l Somme; CHU Nord. Bt. de Snté Publique, Amiens, Frne. 4 University of Vleni, Are of Preventive Mediine nd Publi Helth, Vleni, Spin. 5 INSERM (Ntionl Institute of Helth nd Medil Reserh) U754 - IFR69, Villejuif, Frne. 6 Deprtment of Onology, University Hospitl, Orebro, Sweden. 7 Unit of Cner Epidemiology, University of Turin, Turin, Itly. 8 Centros de Investigión Biomédi en Red (CIBER) CB06/02/045, CIBER tionsepidemiology nd Publi Helth, Institute Crlos III, Spin. 9 Clinil Epidemiologi Reserh Unit, Dr Peset University Hospitl, Vleni, Spin. 10 University of Arhus, Denmrk. 11 Deprtment of Onology, University Hospitl, Lund, Sweden. 12 Institute of Pthology Alborg Hospitl, University of Arhus, Denmrk. 13 Veneto Cner Registry, University of Pdu, Instituto Onologio Veneto (IOV), Itly. Correspondene to: Professor W Ahrens, Bremen Institute for Prevention Reserh nd Soil Mediine (BIPS), University of Bremen, Germny, Linzer Strsse 10, D Bremen, Germny. [E-mil: hrens@bips.uni-bremen.de] Snd J Work Environ Helth 2007, vol 33, no 5 387

2 Endorine-distributing ompounds nd biliry trt ner thn in men, with n overll gender rtio between 1.2 nd 1.8 for most Europen ountries (5), wheres higher frequeny mong mles is often found for ners of the extrhepti bile dut nd the mpull of Vter, the femle-to-mle rtio rnging from 0.6 to 1.1 (2). A history of gllstones, the most importnt risk ftor for extrhepti biliry trt ner (1, 2, 6), ws observed in bout 60% to 80% of ptients with gllbldder rinom (7) nd in 13 40% of ptients with ner of the extrhepti bile dut or the mpull of Vter (8, 9). The preponderne of gllbldder rinom mong women my be promoted by femle sex hormones diretly or indiretly through holelithisis (10). Gllbldder ontrtility is redued during the seond hlf of the menstrul yle; residul volume nd biliry stsis during the lst trimester of pregnny re inresed (11). Endogenous (12) nd exogenous (13) estrogens nd pregnny (14) inrese the sturtion of bile with holesterol nd impir biliry motility. Prity (15, 16), first birth before the ge of 25 yers (17), brest feeding of more thn one infnt (18), nd hormone replement therpy (19) my inrese the risk of gllbldder rinom, s my the use of orl ontreptives (9, 20). However, the ltter findings remin ontroversil (20). Some ouptionl exposures with possible hormonl effet, suh s exposure to pestiides nd polyhlorinted biphenyls (PCB), hve been reported to inrese the risk of heptobiliry ner mong men (21 23). None of these studies, however, hs been ble to exmine the risks for ner of the extrhepti nd intrhepti biliry trts seprtely. Severl xenobiotis disrupt norml funtioning of the endorine system through interferene with hormones nd the wy hormones ontrol growth, metbolism, nd body funtions. These gents re termed endorine-disrupting ompounds. As prt of the Europen multienter se ontrol study on rre ners (24, 25), this report explores the influene of ouptionl exposure to endorine-disrupting ompounds on the risk of extrhepti biliry trt ner mong men. Tble 1. Number nd perentge of index prtiipnts who ompleted job-speifi questionnires. Number of job-speifi questionnires per person Cses Controls N % N % Totl Men Study popultion nd methods Histologilly onfirmed ses of rinom of the extrhepti biliry trt in men 35 to 75 yers of ge, inluding gllbldder rinom, ner of the extrhepti bile dut, nd ner of the mpull of Vter (Interntionl Clssifition of Diseses, 9th revision, odes 156.0, 156.1, 156.2, 156.8, 156.9), were dignosed, nd the men were interviewed fe-to-fe between Jnury 1995 nd June 1997 in the frme of n interntionl multienter se ontrol study in six Europen ountries (25). The ontrols were rndomly drwn from the generl popultion in Frne, Denmrk, Germny, Itly, nd Sweden nd were mthed by ge nd study enter. Hospitl ontrols (ptients with olon denorinom) were seleted s the only ontrol group in Spin nd s n dditionl ontrol group in Denmrk. About 33% (60 of 183) of the se interviews nd less thn 4% (67 of 1938) of the ontrol interviews were onduted with next-of-kin or other surrogte person, sine the index prtiipnt hd died or ws too ill for n interview. Sine ouptionl informtion on pst ontt with gents nd the use of personl protetive equipment would presumbly be poorly reported by proxies, the risk estimtion ws restrited to the index interviews. The ore questionnire inluded strutured setions for soiodemogrphi ftors (edution), ouptionl history, physil hrteristis (onstitution), exposures (smoking, lohol, pestiides, hormones, nd oils with PCB, light, het, X-rys, et), nd medil history (inluding the history of gllbldder diseses). It ended with heklist of ouptions, industries, nd job tsks to trigger the pplition of job-speifi questionnires, of whih 20 ddressed exposure to endorine-disrupting ompounds (tble 1). Answers were entered into ommon dtbse by eh study enter, where editing nd error heking ws performed. Dt from the three job-speifi questionnires onerning dentistry, tnneries, nd rilwy working ontined no positive nswers to questions relted to endorine-disrupting ompounds. The numbers nd perentges of the prtiipnts who nswered the jobspeifi questionnires re listed in tble 1. Qulittive exposure ssessment The gents under investigtion were seleted s representing the most importnt potentil ouptionl endorine-disrupting ompounds (26 28). They were estrogens, lkylphenols, phthltes, oils with PCB, bisphenol pa, hlorophenols or penthlorophenol, phenylphenol, pestiides, nd other endorine-disrupting ompounds (furnni nd phenoli resins, moth- or rotproofing gents, impregntion gents, nd hrdeners). 388 Snd J Work Environ Helth 2007, vol 33, no 5

3 Ahrens et l A speifi setion in the ore questionnire ddressed pst ontt with pestiides, its durtion, nd pplied tehniques. The setion on hemils with n estrogeni effet ddressed the hndling of hormonl gents nd the hndling of trnsformer oils with PCB. The use of PCB ws restrited to losed environments like trnsformers nd ooling systems in Europe sine the beginning of the 1980s. Only the prtiipnts who reported pst exposure to estrogeni hormones or oils with PCB nd who relled the type of gent were tully lssified s exposed to endorine-disrupting ompounds. On the bsis of nswers to speifi questions in the job-speifi questionnires bout prtiulr hemils or relted job tivities, the prtiipnts were lssified s exposed if they hd positively reported the pst use or pplition of n gent. The perentge of the prtiipnts who were lssified s unexposed beuse they did not nswer the prtiulr questions in the job-speifi questionnires (missing vlues) vried from 0% to 32%. The job-speifi questionnires were not mutully exlusive for given exposure when the gent ould be used in different irumstnes. Quntittive exposure ssessment Exposure reords for eh gent were onverted into semiquntittive vribles (probbility, intensity, durtion). Eh of the 14 hemils reeived numeril ode (estrogens, lkylphenols, phthltes, oils with PCB, bisphenol A, hlorophenols, phenylphenol, herbiides, insetiides, fungiides, furnni or phenoli resins, moth- or rot-proofing gents, impregntion gents, nd hrdeners). The exposure ssessment ws blinded with respet to the se ontrol sttus, nd it followed n priori strtegy strting with the job-speifi questionnires. The probbility of exposure ws grded ording to the use of protetive equipment with 0 = no exposure, 1 = use of effetive protetive equipment (filter msk nd overll, bin, et), 2 = use of less protetive equipment (only gloves or only msk without filter, et), nd 3 = no use of protetive equipment. In helth re, the exposure probbility for those who injeted estrogeni drugs ws presumed to be very low nd ws grded s 1, while orl dministrtion or the pplition of plsters ws set to 0. The exposure intensity ws lssified ording to the type of proess s 0 = no exposure, 1 = no personl pplition of the hemils, 2 = personl pplition, but utomtilly (distnt opertion or distribution by mhines with good protetion), 3 = mnul pplition with or without mhines. Job-speifi questionnires for ll of the workers exept those in frming, niml husbndry, nd forestry did not ontin questions on the method of pplition; therefore, the exposure intensity ws grded ording to the frequeny of the orresponding tivity: 1 (low) when work proess ws seldom rried out, 2 (medium) when work proess ws rried out severl dys week or severl dys month, 3 (high) when work proess ws rried out dily. For the exposures to hormones, PCB, nd pestiides, whih were identified through the ore questionnire, but not through job-speifi questionnires, it ws not possible to ssess the intensity nd probbility of the exposure diretly. This ws the se for one prtiipnt (16.7%) exposed to hormones, for 17 respondents (22.4%) exposed to PCB, nd for 55 respondents (18.4%) exposed to pestiides. In these ses, when the prtiipnts reported the use of hormones, PCB, or pestiides in the ore questionnire but did not fill out ny of the orresponding job-speifi questionnires, the intensity nd probbility were estimted individully by evluting the tsk desription in the ore questionnire. If the tsk desription ws missing in the ore questionnire, the verged vlues of the orresponding vribles were obtined from the dt in relevnt job-speifi questionnires of ll of the other exposed prtiipnts [seven job-speifi questionnires ddressing exposure to pestiides nd two job-speifi questionnires ddressing exposure to PCB (see the ppendix vilble on the homepge of the Sndinvin Journl of Work, Environment & Helth)] to derive vlues for probbility nd intensity. In seond step, the semiquntittive exposure vribles for eh exposure were multiplied by the durtion in yers nd summed up for eh prtiipnt. An exposure index ws lulted for eh person s the produt of the probbility, intensity, nd durtion for eh endorine-disrupting ompound s follows: = =1 K Indexexp k P k I k Dk, where Index exp is the umultive exposure index with P k (probbility), I k (intensity), nd D k (durtion, in yers) in job period k of K job periods. The exposed prtiipnts were tegorized in tertiles of the joint distribution of the ses nd ontrols ording to the exposure index s low, medium, nd high. For the presenttion of the exposure intensity, the mximum intensity rehed during worker s entire worklife ws used. Sttistil methods The study popultion ws strtified ording to ge (9 groups with n ge rnge from 32 to 72 yers) nd ountry. Snd J Work Environ Helth 2007, vol 33, no 5 389

4 Endorine-distributing ompounds nd biliry trt ner The first step of the exposure ssessment ws mde mnully using Mirosoft Exel, nd the seond step ws performed implementing n SAS progrm (SAS In, Cry, NC, USA). The referene group for the ombined nlysis of ll of the endorine-disrupting ompounds ws formed by ll of the prtiipnts who were unexposed to ny of them, while the referene group for gent-speifi risk estimtion omprised those who were unexposed to the endorine-disrupting ompound of interest. The SAS sttistil softwre pkge (8th relese) ws used for the sttistil nlyses. The odds rtios (OR), bsed on unonditionl multiple logisti regression models, nd their 95% onfidene intervls (95% CI) were lulted for eh exposure gent s the mesure of the ssoition between the speifi exposure nd extrhepti biliry trt ner. The following onfounders were onsidered in the djusted nlysis: OR 1 ws djusted for ge (ontinuous) nd ountry. OR 2 ws djusted for ge (ontinuous), ountry (one dummy vrible per ountry), nd gllstones (ever onfirmed by physiin, dummy vrible). The finl nlyses were onduted with the exlusion of the ontrol interviews with next-of-kin or other surrogte person, s we ssumed tht the detiled informtion on whih the exposure ssessment ws bsed ould only be provided by the index prtiipnts. Results The response proportion ws 72% for the ses, providing totl of 183 interviews. The response proportion for the ontrols vried by ountry, being high in Frne (78%) nd Itly (73%) but lower in Denmrk, Germny, nd Sweden (54 57%). Overll, 1421 of 2343 eligible popultion ontrols nd 517 of 612 hospitl ontrols prtiipted. Next-of-kin interviews were obtined for 60 ses nd 70 ontrols. Detils of the responses hve been given erlier (25). Altogether 59 ses were oded s gllbldder rinom, 55 s ner of the extrhepti bile dut, 58 s ner of the mpull of Vter, nd 11 s ses with overlpping sites. The men ge of the ptients in these four groups ws lmost the sme (60.9 yers, 57.3 yers, 59.4 yers, nd 59.5 yers, respetively). The men ge nd the distribution of the prtiipnts by ountry re given in tble 2. The distribution of ses nd ontrols by number of job-speifi questionnires is shown in tble 1. Among ll of the index prtiipnts, 1420 (71.2%) were unexposed. Altogether 574 (28.8%) index prtiipnts were lssified s exposed, 56 (2.8%) of them only on the bsis of the ore questionnire [4 ses (3.3%), 52 ontrols (2.8%)]. Of ll the index prtiipnts, 16 (0.80%) reported exposure periods in the ore questionnire in ddition to doing so on the job-speifi questionnires. There were 1235 exposed job periods overll, 1156 of whih were only bsed on the job-speifi questionnires. Altogether 444 (22.26%) study prtiipnts were exposed to just one endorine-disrupting ompound, while 83 (4.16%), 38 (1.91%); 6 (0.30%), nd 3 (0.15%) were exposed to two, three, four, or five or more endorine-disrupting ompounds, respetively. A summry of the key questions ording to the job-speifi questionnire on whih the exposure ssessment ws bsed nd the number of index prtiipnts lssified s exposed to the orresponding endorine-disrupting ompounds n Tble 2. Number of interviewed prtiipnts. Cses Popultion ontrols Hospitl ontrols All ontrols N % Men SD Rnge N % Men SD Rnge N % Men SD Rnge N % Men SD Rnge All interviews Denmrk Sweden Frne Germny Itly Spin Age (yers) Index interviews Denmrk Sweden Frne Germny Itly Spin Age (yers) Only interviews with index prtiipnts. 390 Snd J Work Environ Helth 2007, vol 33, no 5

5 Ahrens et l Tble 3. Risk of ner of the extrhepti biliry trt mong the men reporting tivities entiling exposure to endorine-disrupting ompounds in job-speifi questionnires (only index prtiipnts). (OR = odds rtio, 95% CI = 95% onfidene intervl) Job-speifi questionnire Exposed OR 1 95% CI b OR 2 95% CI d Cses Controls (N) (N) Helth re 5444 Cooking nd food preprtion Frming Animl husbndry Forestry 12 Foundries Wood prodution 9 Pper prodution Textile industry Met proessing Shoe or lether prodution 7 Eletril work Rubber industry Plsti prodution Pinting Pint mnufturing industry 6 Chemil industry 10 OR 1 djusted for ountry nd ge. b 95% CI for OR 1. OR 2 djusted for ge, ountry nd gllstones; the referene group for eh OR ws formed by ll of the prtiipnts not belonging to the respetive tegory. d 95% CI for OR 2. be found in n ppendix tht is vilble on the homepge of the Sndinvin Journl of Work Environment & Helth. Ouption-speifi nlysis For the 17 speifi job-speifi questionnires with vilble reports on endorine-disrupting ompounds, djusted risk estimtes were inresed for the index prtiipnts bout twie or more s often for those working in foundries, the textile industry, met proessing, eletril work, nd plsti prodution (tble 3). Anlysis ording to exposure to endorine-disrupting ompounds Exposure to endorine-disrupting ompounds mong the index prtiipnts resulted in n djusted OR (djusted for ge, ountry, nd gllstones) of 1.4 (95% CI ). The highest OR vlues were observed for exposure to oils with PCB (djusted OR 2.8, 95% CI ), bisphenol A (djusted OR 2.1, 95% CI ), phthltes (djusted OR 2.0, 95% CI ), nd lkylphenols (djusted OR 2.0, 95% CI ) (tble 4). The odds rtios were, in generl, lower in the nlysis for ll of the study prtiipnts (dt not shown), while restrition to the speifi dt from the job-speifi questionnires resulted in higher odds rtios (djusted OR for ll endorine-disrupting ompounds ombined 1.7, 95% CI ) (tble 4). Tble 4. Adjusted odds rtios (OR) for exposure to speifi endorine-disrupting ompounds (only index prtiipnts). (95% CI = 95% onfidene intervl) Endorine-disrupting ompounds Unexposed Exposed OR 1 95% CI b OR 2 95% CI d Cses Controls Cses Controls (N) (N) (N) (N) Hormones Phthltes Alkylphenols Pestiides (job-speifi questionnires only) Pestiides e Oils with polyhlorinted biphenyls (job-speifi questionnires only) Oils with polyhlorinted biphenyls e Bisphenol A Chlorophenols Phenylphenols Other endorine-disrupting ompounds All endorine-disrupting ompounds e All endorine-disrupting ompounds (job-speifi questionnires only) OR 1 djusted for ountry nd ge. b 95% CI for OR 1. OR 2 djusted for ge, ountry nd gllstones. d 95% CI for OR 2. e Estimtion of the risk of ner of the extrhepti biliry trt due to exposure to endorine-disrupting ompounds, evluted on the bsis of the ore questionnire nd the job-speifi questionnires; the referene group for eh OR ws formed by ll of the prtiipnts not belonging to the respetive tegory. Snd J Work Environ Helth 2007, vol 33, no 5 391

6 Endorine-distributing ompounds nd biliry trt ner When lssified by mximum exposure intensity, endorine-disrupting ompounds showed no onsistent pttern (tble 5). The djusted risk estimtes for pestiides were redued for low exposure intensity (OR 0.6) nd slightly elevted in the medium nd high tegories (OR 1.6 nd 1.3, respetively). For endorine-disrupting ompounds other thn pestiides the djusted estimtes deresed with intensity, with odds rtios of 2.8, 1.4, nd 1.3 for low, medium nd high intensity, respetively. The djusted odds rtios for PCB exposure vried between 5.2 for low intensity nd 1.3 for medium intensity. All of the endorine-disrupting ompounds ombined showed djusted odds rtios of 1.3, 1.5, nd 1.4 for low, medium, nd high intensity, respetively. Anlysis by umultive exposure index The rnge of the umultive exposure index ws divided into tertiles of the distribution for the exposed index prtiipnts (45 ses nd 529 ontrols), the unexposed group serving s the referene group. Due to the smll number of prtiipnts exposed to single endorinedisrupting ompounds, grouped nlysis ws mde for those exposed to pestiides, endorine-disrupting ompounds other thn pestiides, nd PCB, s well s for the whole group of exposed prtiipnts. No obvious inresing trends by umultive exposure were observed for endorine-disrupting ompounds other thn pestiides (phthltes, PCB, phenylphenol, penthlorophenol, bisphenol A, lkylphenols, nd other endorine-disrupting ompounds), PCB, or ll endorinedisrupting ompounds (tble 6). Cumultive exposure to pestiides showed redued risk in the low exposure tegory nd elevted odds rtios in the medium nd high tegory (tble 6). Sine gllbldder rinom is more frequent in women nd ner of the extrhepti bile dut nd the mpull of Vter re more frequent in men, we strtified the nlysis for ll endorine-disrupting ompounds by tumor site. Gllbldder rinom ws not ssoited with ny risk, while ner of the extrhepti bile dut nd the mpull of Vter showed inresed odds rtios for medium nd high umultive exposure. But these estimtes were unstble due to the smll numbers (tble 7). Disussion Our explortory nlysis showed some ssoitions between exposure to endorine-disrupting ompounds in Tble 5. Odds rtios (OR) for low, medium, nd high intensity of the exposure for seleted groups of endorine-disrupting ompounds (mximum intensity in lifetime, only index prtiipnts). (95% CI = 95% onfidene intervl) Endorine-disruptting ompounds Unexposed Exposed OR 1 95% CI b OR 2 95% CI d Cses Controls Cses Controls (N) (N) (N) (N) Pestiides e Low Medium High Totl All endorine-disrupting ompounds other thn pestiides e Low Medium High Totl Polyhlorinted biphenyls e Low Medium High Totl All endorine-disrupting ompounds e Low Medium High Totl OR 1 djusted for ountry nd ge. b 95% CI for OR 1. OR 2 djusted for ge, ountry, nd gllstones. d 95% CI for OR 2. e Estimtion of the risk of extrhepti biliry trt ner due to exposure to endorine-disrupting ompounds, evluted on the bsis of the ore questionnire nd the job-speifi questionnires; the referene group for eh OR ws formed by ll of the prtiipnts not belonging to the respetive tegory. 392 Snd J Work Environ Helth 2007, vol 33, no 5

7 Ahrens et l Tble 6. Odds rtios (OR) for low, medium, nd high exposure index (produt of durtion, probbility nd intensity) (only index prtiipnts). Endorine-disrupting ompounds Unexposed Exposed OR 1 95% CI b OR 2 95% CI d Cses Controls Cses Controls (N) (N) (N) (N) Cner of the extrhepti biliry trt Pestiides e Low Medium High Totl Missing exposure index 1 3 All endorine-disrupting ompounds other thn pestiides Low Medium High Totl Polyhlorinted biphenyls e Low Medium High Totl All endorine-disrupting ompounds e Low Medium High Totl Missing exposure index 3 Only gllbldder rinom f All endorine-disrupting ompounds e Low Medium High Totl Missing exposure index 3 Only rinoms of the extrhepti bile dut nd the mpull of Vter f All endorine-disrupting ompounds e Low Medium High Totl Missing exposure index 3 OR 1 djusted for ountry nd ge. b 95% CI for OR 1. OR 2 djusted for ge, ountry, nd gllstones. e Estimtion of risk of ner of the extrhepti biliry trt due to exposure to endorine-disrupting ompounds, evluted on the bsis of the ore questionnire nd the job-speifi questionnires; the referene group for eh OR ws formed by ll of the prtiipnts not belonging to the respetive tegory. f Six ses in whih the tumor overlpped more thn one site hd to be exluded from the site-speifi nlysis (gllbldder rinom, extrhepti bile dut nd mpull of Vter). the workple nd the risk of ner of the extrhepti biliry trt for men, prtiulrly for ner of the extrhepti bile dut nd the mpull of Vter. Although mny ouptionl exposures hve been reported to be ssoited with gllbldder rinom nd other biliry trt tumors in the pst, it ws often not possible to disentngle the risk of biliry trt nd liver ner in different ouptions due to the limited dt. Previous epidemiologi studies on ner of the extrhepti biliry trt hve mentioned the smll size of popultions studied nd problems with quntittively ssessing exposure to xenobiotis, prtiulrly endorine-disrupting ompounds. The estimtion of ouptionl exposure to endorine-disrupting ompounds ws bsed on reported ontt with the gents during eh prtiipnt s worklife nd ws therefore prone to mislssifition nd reporting bis. For this reson we exluded the ontrol interviews with next-of-kin or other surrogte person from the nlysis ssuming tht these persons were unble to report the job tivities of index prtiipnts with suffiient detil nd relibility. The pplition of Snd J Work Environ Helth 2007, vol 33, no 5 393

8 Endorine-distributing ompounds nd biliry trt ner job-speifi questionnires should hve further improved the qulity of the job tsk desriptions nd the speifiity of the exposure ssessment. Bsed on nswers to the job-speifi questionnires, the intensity nd probbility of the exposure to prtiulr ompound were ssessed on reltive sle. Therefore, the bsolute vlues of intensity for given gent my not hve been perfetly omprble between different job tsks. This possibility would hve resulted in some mislssifition of the degree of exposure when exposures were summed up over vrious job tsks to estimte umultive exposure. We ssume tht the mislssifition of exposure ws nondifferentil beuse the exposure ssessment ws blinded with respet to the se ontrol sttus. Aording to the explortory design, multiple omprisons were mde tht entiled the problem of mss signifine. The reder should tke this into ount when the results re interpreted. Confidene intervls hve to be onsidered minly s inditors of the stbility of the presented risk estimtes nd not s forml tests of sttistil signifine. Our observtions re onsistent with those of other investigtions. Assoitions between femle hormones nd biliry trt ner were derived from observtions tht prity (15) nd high number of pregnnies (16), s well s ge before 25 yers t first birth (17), brest feeding of more thn one bby (18), nd femle hormone replement therpy (19) mong women inrese the risk of gllbldder rinom. In our study, the risk of extrhepti biliry trt ner ws inresed mong the prtiipnts ever exposed to endorine-disrupting ompounds with known estrogeni effet (lkylphenols, PCB, bisphenol A), nd exposure to PCB ws sttistilly signifint (tble 4). Not ll of the subtypes of ner of the extrhepti biliry trt my be eqully ssoited with endorinedisrupting ompounds. Gllbldder rinom ours t higher rtes in women thn in men, the opposite is true for ner of the extrhepti bile dut nd the mpull of Vter. This sitution would seem to suggest tht gllbldder rinom my be ssoited with hormones but tht ner of the extrhepti bile dut nd the mpull of Vter might not be. It should be noted, however, tht the exess risk for ll endorine-disrupting ompounds ombined ws restrited to the sites extrhepti bile dut nd mpull of Vter, not the gllbldder. This observtion my ontrdit our hypothesis lthough it is bsed on reltively smll numbers. Our finding my lso point towrds different etiology for ner of the extrhepti biliry trt mong men nd women, s the preponderne of gllbldder rinom mong femles ould be medited through their higher risk of developing gllstones. The inonsistent ssoition between pestiide exposure nd ner of the biliry trt should be interpreted with ution in reltion to endorine tivity. We hve grouped together ll pestiide exposures even though not ll of them hve endorine tivity. As it is not possible to obtin vlid nswers to questions on brnd nmes or hemil properties of pestiides in retrospetive study suh s this one, we epted substntil degree of unertinty in this explortory nlysis. The ner inidene nd mortlity of workers in trnsformer mnufturing plnts with extensive use of trnsformer fluid ontining minerl oil reveled n exess of biliry trt ner, prtiulrly of gllbldder rinom (23, 29). From these two lrge ohort studies, Yssi et l reported stndrdized inidene rtio (SIR) for gllbldder rinom of 5.1 (95% CI ) for mle employees of trnsformer mnufturing plnt, nd Gustvsson et l found stndrdized inidene rtio of 2.6 (95% CI ) for liver nd biliry trt ner mong Swedish mle pitor mnufturing workers. Our study reveled high risk for extrhepti biliry trt ner mong prtiipnts exposed to PCB. The estimtes beme stronger when restrited to the index prtiipnts nd to the more speifi dt obtined by job-speifi questionnires. Although we did not observe dose effet reltionship with inresing exposure, our findings my orroborte the ssessment of PCB s hzrdous ouptionl exposure mong eletriins. It should be noted tht, in our study, the prtiipnts exposed to endorinedisrupting ompounds in eletril work formed lrge subgroup, not only in the whole study popultion, but lso mong those employed in ouptions with possible PCB exposure (eletril work nd the rubber industry). Approximtely 70% of the eletril workers were lssified s exposed to ny of the endorine-disrupting ompounds, inluding PCB. Given the lrge number of reported simultneous gent exposures of eletril workers, it n be ssumed tht eletril workers were exposed to endorine-disrupting ompounds to greter degree thn the other workers were. The oinidene of n elevted risk for PCB exposure nd n exess risk mong eletriins my be due to frequent ontt with PCB in this job group. The exposure ssessment in this study ws minly bsed on the detiled nswers in the job-speifi questionnires. As stted previously, the dvntge of suh reords is the possibility to ssess the exposure determinnts semiquntittively. However, the mount of detil bout speifi gents tht n be obtined by job-speifi questionnires is limited. Nondifferentil mislssifition of exposure tht would result in bis of odds rtios towrds unity hs probbly ourred. In ddition, ouptionl informtion bsed on job-speifi questionnires is not free of rell bis reltive to the outome of disese. Another bis ould hve been used by the use of ptients with olon denorinom s ontrols. 394 Snd J Work Environ Helth 2007, vol 33, no 5

9 Ahrens et l If the risk for olon ner is influened by exposure, risk estimtes would be ttenuted. Beuse of the smll number of exposed ses, we were unble to evlute whether ountry modified the effet of endorine-disrupting ompounds. Although ommon questionnire ws used, it is possible tht the exposure index mens something different in different ountries, in prtiulr if the exposure ssessment were only bsed on job title nd industry. The speifi exposure questions tht we used nd the use of job-speifi questionnires should, however, vert this problem to lrge degree. The results of this study prtilly onfirm those from previous studies, nd they suggest tht ouptionl exposure to endorine-disrupting ompounds, nd to PCB in prtiulr, my be ssoited with n inresed risk of ners of the extrhepti biliry trt. The pprisl of this onlusion, however, hs to tke into ount the diffiulties in retrospetively ssessing the gents of interest in popultion-bsed study suh s this one. Furthermore, the endorine-disrupting mehnism of the gents onsidered in this study is not yet fully understood for ll of them. The estrogeni effet of lkylphenols (otyl or nonylphenole ethoxyltes), PCB, pestiides [dihlorodiphenyltrihloroethne (DDT), hexhloroylohexne (HCH)], nd bisphenol A hs been shown in mny studies, wheres penthlorophenol seems to exhibit both estrogeni nd nti-estrogeni effets nd phthltes exhibit nti-ndrogeni effets nd wek ffinity to the estrogen reeptor. For this reson the risk estimtes for eh gent need to be onsidered seprtely in ddition to the ombined exposure index. Future studies should try to onfirm or refute our observtions. Aknowledgments The study ws undertken in ordne with the requirements of the ethis ommittees in eh of the prtiipting ountries nd regions. We grtefully knowledge the help of Nils Shmeißer, who refully heked ll of the nlyses nd who orreted the identified errors. We lso knowledge the ollbortion of the ptients, ontrol persons, prtiipting hospitls, nd dt providers, suh s the Frenh Cner Registry Assoited (Résion FRANCIM). The study Ouptionl Risk Ftors for Rre Cners of Unknown Etiology ws finnilly supported by the Europen Commission, DGXII, grnts BMH1 CT nd ERB CIPD CT , nd ntionl funding genies: Denmrk: The Sttegi Environment Progrmme: Frne: Ligue Ntionle ontre le Cner, Fédértion Ntionle des Centres de Lutte ontr le Cner, Fondtion de Frne, ontrt , Institut Ntionl de l Snté et de l Reherhe Médile (IN- SERM) ontrt Réseu en Snté Pulique (Network for Publi Helth) 4R006A, Frenh Ministry of Environment, ontrt ; Germny: Federl Ministry for Edution, Siene, Reserh nd Tehnology (BMBF), grnt 01-HP-684/8; Itly: MURST, Ministry of Lbour, Itlin Assoition for Cner Reserh, Compgni SnPolo/FIRMS; Portugl: Junt Nionl de Investidäo Cientifi e Tenológi, Prxis XXI, no2/2.1/sau/1178/95; Spin: Fondo de Investigión de l Snitrie, Ministerio de Snidd y Consumo,UNIDAD de Investigión Clinio-Epidemiológi, Hospitl Dr Peset, Generlitet Vlenin, Deprtmento de Snidd y Consumo, Gobierno Vso, Fondo de Investigión de l Snitri (FIS), Ministerio de Snidd y Consumo, Ayud l Investigión del Deprtmento de Slud del Gobierno de Nvrr; Sweden: Swedish Counil for Work Life Reserh, Reserh Foundtion of the Deprtment of Onology in Umeå, Swedish Soiety of Mediine, Lund University Hospitl Reserh Foundtion, Gunnr, Arvid nd Elisbeth Nilsson Cner Foundtion, Örebro County Counil Reserh Committee, Örebro Medil Center Reserh Foundtion, John nd August Persson Foundtion for Sientifi Medil Reserh, Bert Kmprd Foundtion for Cner Reserh. The Europen Study Group on Ouptionl Cuses of Rre Cners inluded members from Denmrk (Hermn Autrup, Henrik Kolstd, Preben Johnsen, Lind Kærlev, Elsebeth Lynge, Jørn Olsen, Stein Pulsen, Lisbeth Nørum Pedersen, Svend Sbroe, Peter Stubbe Teglbjerg, Mogens Vyberg), Frne (Psl Guénel, Joëlle Fervotte nd the members of the FRANCIM ssoition: Ptrik Arveux, Antoine Buemi, Pule-Mrie Crli, Gilles Chplin, Jen-Pierre Durès, Céile Dufour, Jen Fivre, Psle Groslude, Anne-Vlérie Guizrd, Mihel Henry-Amr, Guy Lunoy, Frnois Ménégoz, Niole Rverdy, Pul Shffer), Germny (Wolfgng Ahrens, Corneli Bumgrdt-Elms, Sibylle Gotthrdt, Ingeborg Jhn, Krl-Heinz Jökel, Hiltrud Merzenih, Andres Stng, Christ Stegmier, Antje Timmer, Hrtwig Ziegler), Itly (Terri Bllrd, Frno Bertoni, Giuseppe Gorini, Sndr Gostinihi, Giovnn Msl, Enzo Merler, Frno Merletti, Lorenzo Rihirdi, Lorenzo Simonto, Pol Zmbon), Ltvi (Iren Rogovsk, Glin Shrkov, Aivrs Stengrevis), Portugl (Noemi Afonso, Altmiro Cost-Pereir, Soni Dori, Crlos Lopes, José Mnuel Lopes, An Mirnd, Cristin Sntos), Spin (M Adel Snz Agudo, Jun J Aurrekoetxe, Conepiòn Brun, Alii Córdob, Miguel Angel Mrtínez González, Frniso Guillén Grim, Ros Gurh, Agustin Llopis González, Bln Mrín, Ampro Mrquin, Mri M Morlez-Suárez- Vrel, Inés Aguing Ontoso, JM Mrtínez Peñuel, An Purs, Frniso Veg, Mri Auror Villnuev Snd J Work Environ Helth 2007, vol 33, no 5 395

10 Endorine-distributing ompounds nd biliry trt ner Gurdi), Sweden (Mikel Eriksson, Lennrt Hrdell, Irene Lrsson, Hkn Olson, Moni Sndström, Gun Wingren), Switzerlnd (Jen-Mihel Lutz), nd the United Kingdom (Jnine Bell, In Cree, Tony Flether, Alex JE Foss). Referenes 1. Frumeni JF Jr. Cners of the pnres nd biliry trt: epidemiologil onsidertions. Cner Res. 1975;35: Strom BL, Hibberd PL, Soper KA, Stolley PD, Nelson WL. Interntionl vritions in epidemiology of ners of the extrhepti biliry trt. Cner Res. 1985;45: Crrig MT, Henson DE. Liver, gllbldder, extrhepti bile duts, nd pnres. Cner Suppl. 1995;75(1): Netoux J, Colemn MP. Trends in biliry trt ner. Rev Epidémiol Snté Publique 1993;41: Ztonski W, L Vehi C, Levi F, Negri E, Luhini F. Desriptive epidemiology of gll-bldder ner in Europe. J Cner Res Clin Onol. 1993;119: Ahrens W, Timmer A, Vyberg M, Flether T, Guénel P, Merler E, et l. Risk ftors for extrhepti biliry trt rinom in men: medil onditions nd lifestyle. Results from Europen multi-entre se ontrol study. Eur J Gstroenterol Heptol. 2007;19(8): Frumeni JF, Deves SS, MLughlin JK, Stnford JL. Biliry trt ner. In: Shottenfeld D, Frumeni JF Jr, editors. Cner epidemiology nd prevention. 2nd ed. New York (NY): Oxford University Press; p Chow W-H, MLughlin JK, Menk HR, Mk TM. Risk ftors for extrhepti bile dut ners: Los Angeles County, Cliforni (USA). Cner Cuses Control. 1994;5: Yen S, Hsieh C, MMhon B. Extrhepti bile dut ner nd smoking, beverge onsumption, pst medil history, nd orl ontreptive use. Cner. 1987;59: Diehl AK. Epidemiology of gllbldder ner: synthesis of reent dt. J Ntl Cner Inst. 1980;65: Brvermn DZ, Jonson ML, Kern F. Effets of pregnny nd ontreptive steroids on gllbldder funtion. N Engl J Med. 1980;302: Srgg RK, MMihel AJ, Bghurst PA. Diet, lohol nd reltive weight in gllstone disese: se ontrol study. BMJ. 1984;288: Lyde PM, Vessey MP, Yetes D. Risk ftors for gllbldder disese: ohort study of women ttending fmily plnning linis. J Epidemiol Commun Helth. 1982;36: L Vehi C, Negri E, Frneshi S, Przzini F. Long-term impt of reprodutive ftors on ner risk. Int J Cner. 1993;53: Plesko I, Preston-Mrtin S, Dy NE, Tzonou A, Dimitrov E, Somogyi J. Prity nd ner risk in Slovki. Int J Cner. 1985;36: Ztonski WA, Lowenfels AB, Boyle P, Misonneuve P, Bueno de Mesquit HB, Ghdirin P, et l. Epidemiologil spets of gllbldder ner: se ontrol study of the SEARCH progrm of the interntionl geny for reserh on ner. J Ntl Cner Inst. 1997;89(15): Lmbe M, Trihopoulos D, Hsieh C, Ekbom A, Admi H-O, Pvi M. Prity nd ners of the gllbldder nd the extrhepti bile duts. Int J Cner. 1993;54: Strom BL, Solowy RD, Rios-Dlenz JL, Rodriguez-Mrtinez HA, West SL, Kinmn JL et l. Risk ftors for gllbldder ner: n interntionl ollbortive se ontrol study. Cner. 1995;76(10): Gllus S, Negri E, Chtenoud L, Bosetti C, Frneshi S, L Vehi C. Post-menopusl hormonl therpy nd gllbldder ner risk. Int J Cner. 2002;99: Milne R, Vessey M. The ssoitions of orl ontreption with kidney ner, olon ner, gllbldder ner (inluding extrhepti bile dut ner) nd pituitry tumours. Contreption. 1991;43(6): Coo P, Blir A, Congi P, Sb G, Flore C, E MR, et l. Proportionl mortlity of dihloro-diphenyl-trihloroethne (DDT) workers: preliminry report. Arh Environ Helth. 1997;52(4): Brown DP. Mortlity of workers employed t orgnohlorine pestiide mnufturing plnts n updte. Snd J Work Environ Helth. 1992;18: Gustvsson P, Hogstedt C. A ohort study of Swedish pitor mnufturing workers exposed to Polyhlorinted Biphenyls (PCBs). Am J Ind Med. 1997;32: Kerlev L, Teglbjerg PS, Sbroe S, Kolstd HA, Ahrens W, Eriksson M, et l. Is there n ssoition between lohol intke or smoking nd smll bowel denorinom?: results from Europen multi-enter se ontrol study. Cner Cuses Control. 2000;11: Lynge E, Afonso N, Kerlev L, Olsen J, Sbroe S, Ahrens W, et l. Europen Multi-Centre Cse-Control Study On Risk Ftors For Rre Cners of unknown etiology. Eur J Cner. 2005;41: World Wildlife Fund. Toxis tht tmper with hormones. Wildlife Toxiology Progrm Publition: Egle s Eye 1995;(summer): Pris F, Blguer P, Térounne B, Servnt N, Loste C, Crvedi JP, et l. Phenylphenols, biphenols, bisphenol-a nd 4-tert-otylphenol exibit nd b estrogen tivities nd ntindrogen tivity in reporter ell lines. Mol Cell Endorinol. 2002;193(1 2): vom Sl FS, Hughes C. An extensive new literture onerning low-dose effets of bisphenol A shows the need for new risk ssessment. Environ Helth Persp. 2005;113(8): Yssi A, Tte RB, Routledge M. Cner inidene nd mortlity in workers employed t trnsformer mnufturing plnt: updte to ohort study. Am J Ind Med. 2003;44: Reeived for publition: 14 Deember Snd J Work Environ Helth 2007, vol 33, no 5