Document 850bpadYjKZ6E0G82KdjxREB

. A226 (908 Bladder Cancer inPerfluorooctanesulfonyl Fluoride Manufacturing Workers Final Report Bruce H. Alexander, PhD Divisionof Environmental Health Sciences SchoolofPublic Health University of Minnesota November21,2004 1 36 ---- y OBI | Abstract A mortality study of employees of a perfluorooctanesulfonyl fluoride (POSF). manufacturing facility reporteda significant excess riskofdeath from bladder cancer in workers who`held jobs `with exposureto high levelsofto perfluorooctanesulfonate (PFOS). To further investigate this finding a study to ascertain all cases ofbladder cancer was conducted. Membersofthe original cohort were contactedbymail to inform themofthe study and invite them to complete a brief questionnaire that ascertained any history of bladder cancer. Nonrespondentswere contacted by telephone to ensure receiptofthe study material and offered the chance to complete the questionnaire by telephone. Validationofreported cancers was attempted through medical records for those cases consenting to release ofmedical records. Death certificates were obtained for all cohort members identifiaesd deceasedandcoded for underlying and contributing causesofdeath. The ratesof bladder cancer were compared to the expected population based rates from the Surveillance Epidemiology and End Results (SEER) data published by the National Cancer Institute. Analyses were conducted by estimated cumulative exposure to PFOS. The riskof bladdercancerwas `compared between workers with varying degrees ofexposure. `The questionnaires were retuned by 1,400ofthe 1895 cohort members presumed alive during the study period, ofwhich 1,137weremen and 263 were women. The questionnairreespondents contributed 36,982 person-yearsoffollow-up to the analysis. One hundred cighty-eight cohort `members were identified as deceased. The overall response rate was 74%, with 77% for women and 73% for men. There were 838 respondents and 120 deceased cohort members who held jobs einxphiogsheo.r low PFOS exposed jobs, with 624 and 82ofthese holding at least one job with high A totalofeleven cases of primary bladder cancer cases were identified from the surveys (N=6) and death certificates (N=5). The age,genderand calendar period adjusted Standardized Incidence Ratio (SIR) for the entirecohortwas 1.28 (95% CI=0.64-2.29). The SIR for ever working (N=6), and working for more than ayear (n=3) in a high exposed job were 1.74 (95% CI=0.64-3.79) and 1.12 (95% CI = 0.23-3.27) respectively. Compared to employees in the lowest 2 3Y cumulative exposure category the relative riskof bladder cancer was 0.83, (95% C1=0.15-4.65), 1.92 (95% CI=0.30-12.06), and 1.52 (95% CI=0.21-10.99). Overall, the results ofthis study do not confirm the high excess riskofbladder cancer reported in the mortality studyofthis populationoffluorochemical manufacturing workers. However, the possibility remains for a smaller risk (approximately 1.5to 2 fold)inthe higher exposed workers, but the limited sizeofthe population prohibits a conclusive exposure response analysis. 3 xs Background `The 3M facility in Decatur Alabama was oneoftwo major production sites of `perfluorooctanesulfonyl fluoride (POSF, CsF7SO,F) based specialty chemicals. These specialty chemicals have a wide rangeofapplications, including surface treatments, paper and packaging Protectants, and performance chemicals." POSF based chemicals can degrade or be metabolized 0 perfluorooctanesulfonate (PFOS, CyF1705). PFOS is also usedas a primary component of limited number of specialty chemical applications `The presence ofPFOS in non-occupationally exposed populations and wildlife, particularly `marine mammals and piscivorous birds, has raised concerns about the environmental and health effects of PFOS. PFOS is now recognized asapervasive compoundthat will persist in the environmeandt can accumulate in wildlife." Moreover, PFOS has been detected at levels of 30 10 40 parts per billion in the general population**. These facts prompted the phase-outofthe `production of POSF-based chemicals by the major producer (3M Company). Toxicological studiesofrats and cynomolgus primates have shown that high dosesof PFOS induced enlargement of liver and apparent alterations in metabolic processes, including reduced serum cholesterol levels." PFOS was not found to be a developmental toxicant in ratsor rabbits.* Higher maternal doses of PFOS increased neonatal mortality, absorptions, resorptions, and reduced weight gain in rat pups. There were no effects on post-natal `neurological development or on fertility and estrous cycling in offspring in multigeneration studies. Multiple genotoxicity assays indicate PFOS does not present a hazard from interaction with genetic material." Although the mechanism of toxicity in laboratory animals is not fully understood, it may be due to an effect on fatty acid transport and metabolism, membrane function, peroxisome proliferation, `and/or mitochondrial bioenergetics.>!! `The potential health effects of PFOS exposure have been studiedinoccupationally exposed populations. PFOS has not been shown to affect clinical blood and urine chemistry analyses." A studyof health insurance claims filed by workers atthe Decatur site 1993-1998 evaluated the relative frequencyofepisodesofcare for specific conditions between employeesofthe chemical 4 34s plant (fluorochemical exposed) and the film plant (fluorochemical `nonexposed)." Workers employed inthearoeftahe plant where PFOS exposure may occur had more frequent claims for biliary tract disorders and cystitis recurrence, which were included in the list ofapriori conditions from the 424 categoriesofconditions identified. Claims for benign colon polyps, malignant colorectal tumors and malignant melanoma were also more frequent in the exposed population. A cobort mortality study of current and former employeesofthe 3M Decatur facility `was conducted to evaluate the health of workers exposedto POSF based fluorochemicals.TM The cohort included 2083 workers who were employed for a minimum of 1 year at the Decatur ffaicnidliintyg,farnodmtthheemsotrutdayliwtaysexapnereixecnecsesowfasdeaastcherftoaimnebdltahdrdeorugchanDceecr.eTmbheers,ta3n1d,a1r9d9i8z.edThmeormtaaliinty ratio (SMR) forbladder cancer for the entire workforce was 4.81 (95% CI=1.0-14.1), and the SMR for employees who held highexposure jobs was 12.8 (95% CI=2. 6-37.4). However, these refsorulstesvewrearleyebaarssedinohnigohnleyxptohrseuerecajsoebss.,`hTohweeevmeprljooybesewsewrheoprdiimeadriflryommabilnatdednearnccaenacenrd awlal`stweo:rked water treatment, and not specifically related to fluorochemical production. Twoofthe cases were `maintenance workers and the third worked a majorityofhis career in the wastewater treatment plant and incinerator. Subsequent to the completionofthe study an additional death from bladder cancer was identified outside the study period, and nother employee informed the 3M Medical Department that he had been diagnosed with bladder cancer. Neitherofthese cases worked in PFOSexposed jobs, but they were long time employees at the Decaturfacility. aInndrepsapsotnusseetooftkhne oewlenvaotresdursipsekoctfebdlbaldaddedrercacnacrecriniongtehnesmaotrttahilsitfyacsiltiutydyw,asa croenvdiuecwtoefdt.heFocuurrrent `materials, 4,4 methylene-dianiline, orthotoluidine, benzidinesalts,and `butyl benzyl phthalate, `were formerlyusedat the plant and the useofthese materials endedin the 1960s and 1970'. `The chemicals were useinformationwas not widely used inthe plant; very limited. Melamine has however the available exposure been usedforthelast decade in monitoring an epoxy and capsule (non-fluorochemical) product line. exposures to melamine were low based on Qualitative exposure assessments indicated short exposure task durations. Chloroprene was also usedinseveral manufacturing processes in the chemical plant in the 1960's and 1970's. Chloroprene is considered a possible human carcinogen, however the evidence foar role in 5 x7 bladder cancer is equivocal.' This review indicated that the excessofmortality from bladder cancer in the high PFOS exposed sub-cohort could not be clearly attributteod exposure to a known bladder carcinogen. Bladder cancer is a relatively rare malignancy that has been linked to smoking ' and several occupational exposures. Aromatic amines encountered in the textileandother industries are classified as known human carcinogens. Other occupational exposures associated with bladder cancer and classified as possible human carcinogens by the Intemational Agency for Research on Cancer include polycyclic aromatic hydrocarbons (PAHS), plasticizers such as acetamide and di2-ethy! hexyl)-pthalate, flame retardants such as thiourea, tris (2,3-dibromo propyl phosphate) and antimony trioxide.'" Several other occupational exposures are inconsistently reported to be associated with bladder cancer, including, metal cutting fluids, diesel exhaust, polybrominated biphenyls, and solvents.'* Bladder cancer oceurs more frequently in men than women. The Surveillance Epidemiology and End Results programofthe National Cancer Insitute estimates annual age adjusted incidence rates for 1992-2001of 36.1/100,000 and 9.2/100,000 for men and women respectively. Age adjusted mortality rates for the same period are substantially lower (men =7.7/100,000 and women =2.3/100,000), however survival rates decrease with an increasing age at diagnosis." Due to the large difference between incidence and mortality rates, mortality study will not fully ascertain the burdenofbladder cancer in a population. The goal ofthe present study was to determine whether the association observed in the mortality is representativeofthe entire bladder cancer experienceofthe employees at the 3M Decatur facility. This was accomplished through acase finding effort using direct contact with the population. 6 3Y Methods Study Population `The study site was the 3M facility in Decatur, Alabama. The plant is divided in two major sections, which are approximately 300 yards apart, The chemical plant produces specialty chemicals, including the POSF lineof chemicals. The other section is the film plant, where a variety offilmsare produced, but litle or no occupational fluorochemical exposures occur. The populationofinterest included all current, reired, and former Decatur employees who were eligible for the original cohort mortality study. All current employees who were not partofthe original cohort, those hired after January 1, 1998, were also included in the case finding exercise, but not a primary focusofthe analysis as their exposure had occurred so recently that its contribution to bladder cancer is unlikely. The study required direct contact with the participants. A roster of all known addresses and telephone numbers were obtainedthrough 3M personnel or retiree records. The address information and vital status was updated through a varietyoftracing resources sources available to the UniversityofMinnesota, including TRW/Experian, Lortan Data,and National Change ofAddress.If a cohort member was noted to have died since the mortality study was completed acopyofthe death certificate was obtained fromthestate of record. Questionnaires The primary purposeofthe questionnaire was to identify cases of bladder cancer and the year of diagnosis. The questionnaire also recorded historyofsmoking; aknown risk factor for bladder cancer. Although bladder cancer was the primary focusofthis study, additional questions pertaining to diseases and other conditions were included. These diseases and conditions were selected based on the toxicology studies of PFOS and astudyofepisodesof care conducted in the Decatur plant workforce. The results for the non-bladder cancer endpoints will be presented elsewhere. Inaddition to bladder cancer, the questionnaire ascertained diagnosesofmelanoma, liver, prostate (men only), breast (women only), colon or rectal cancer, non-cancer conditions including liver disease, gall stones, gall bladder infections, stomach ulcers, bladder infections, colon polyps and other diseasesofthe prostate (men only). Abrief pregnancy outcome history was askedofthe women. Several other questions were asked related to routine screening procedures that may be related tothediagnosis ofprostate disease or colon polyps or cancer. 7 Recruitment `The UniversityofMinnesota Institutional Review Board approved the study protocol. Prior to recruitment a series ofmeetings were held with current employees and retirees to inform them of the upcoming study and allowthemto ask questions about the study. Recruitmentofall presumed living cohort members was initiated witah letter and brochure that outlined the reasons for the study, the scopeofthe study, what s requiredofparticipants, and assurances of confidentiality. The study questionnaire, with cover letter and postage paid return envelope, followed the recruitment letter by approximately one week. A reminder post card was sent to all nonrespondents two weeks afte the questionnaire and a second was mailed an additional two weeks after that. Ifmailings were returned with undeliverable addresses, the address information was re-entered into the search engines to identify possible alternate addresses and the mailing `was re-sent, Cohort members who did not respond to the second questionnaire mailing after one `month were contacted by telephone to verify receiptofthe questionnaire and inquire about intent 0 participate. At that time the respondent was offered the opportunity to complete the questionnaire by telephone ifthey preferred. All questionnaires were reviewed upon receipt and double entered into an electronic database. Validationofthe diagnosisofthe self-reported casesof bladder cancer was attempted through `medical records. Participants reporting these conditions were contacted by letter, with telephone follow-up, to request permission to contact their physiciaton verify the diagnosis. Participants who agreed provided signed consent and medical release forms, and the name and addressof the `physician or clinic of reference. Copiesof theconsent and medical release forms were sent to the `physician or clinic along with a request for pathology reports, surgical notes, or any other information pertaining to the diagnosisofthe reported bladder cancer. If no response was received from the clinic or physician they were contacted by telephone to assure receipt ofthe material and encourage appropriate response. Exposure Assessment The exposure assessment followed the previously described method used in the mortality study." This method created job specific exposure categories based on job titles, departments, 3 370 and datesofemployment identified in the participant's individual work histories, and potential for PFOS exposure. The relative differencesinserum PFOS by job were determined by a comprehensive assessment, which is detailed elsewhere." Briefly, serum PFOS concentrations were measured in 186 employees (n = 126 chemical plant;60 =film pian) from arandomly selected sample of232 employees. The geometric mean serum PFO level (95% confidence interval) for chemical plant employees was 0.94 ppm (95% CI=0.79-1.1an3d) for film plant employees it was 0.14 ppm (95% CI=0.11-0.16). The exposure to film plant employees is thought to be influenced by environmental exposure from proximity to the chemical plant. Chemical plant jobs were classified into eight categories: cell operators, chemical operators, maintenance workers (primarily mechanics and electricians), mill operators, waste treatment `plant operators, engineers/laboratory workers, supervisors/managers and administrative assistants. The highest geometric mean levelofserum PFOS was observed in cell operators (2.0 ppm) followed by the waste operators (1.5), chemical operators (1.5 ppm), and maintenance workers (1.3 ppm). Supervisors/managers (0.9 ppm), mill operators (0.6 ppm), engineer/lab `workers (0.4 ppm) and administrative assistants (0.4 ppm)had lower geometric mean serum PFOS levels. Because production processes have remained constant over time, a simple exposure matrix was developed basedonthe work history recordsofthe study cohort. The work histories used inthe exposure analysis covered the period from when the plant opened, 1961,till 1997 when the work histories were collected for the mortality study. With the knowledgeofthe majorjob-specific serum PFOS levels, a company industrial hygienist and epidemiologist assigned each unique job 9 So 37 and department combination in the work history records to oneofthe following three major exposure categories: 1. No direct workplace exposure to POSF-based fluorochemicals (encompasses film plantjobs). 2. Low potential workplace exposure to POSF-based fluorochemicals (includes suchjobs as engineers, quality control technicians, environmental, health and safety workers, administrative assistants and managers). 3. High potential workplace exposure to POSF-based fluorochemnicals (includes cell operators, chemical operators, maintenance workers, mill operators, waste operators and crew supervisors) Hereafterthese three categories will be referred to as the non-exposed, low exposed and high exposure. An additional classification for cumulative exposure assigned the non-exposed, low exposure, and high exposurejobs relative POSF-basedjob exposure value of 1,3 and 10 respectively based on biomonitoring data. The years spent in eachjob were multiplied by the relative weights to develop a cumulative quantitative exposure metric. This exposure metric assumes a continually increasing accumulation of PFOS exposure. However, the half-life for PFOS is prolonged, thus exposures to high concentrations can resinuhligth body burdens for a long time after exposure ceases. To account for this possibility the study participants were also classified as being in an exposed job, high or low andhighonly, foratleast 1 year. Analysis `The estimated incidenceof bladder cancer was compared to the expected incidence based on rates derived from the Surveillance Epidemiology and End Results (SEER) at the National Cancer Institute." The cohort members contributed person-time to the analysis until the diagnosis ofa bladder cancer, death, or the endofthe study(December 31, 2002). All self reported casesofprimary bladder cancer from the questionnaire and bladder cancers identified by death certificate were included in the analysis. If participants checked `Unsure' instead of "Yes" for bladder cancer the case was not included. The age, gender, calendar-year, and exposure-specific person-timeofthe cohort was tabulated using the Life Table Analysis System for the personal computer (PCLTAS) developed by the National Institute for Occupational Safety and Health*. SEER referent data were only available for this program from 1970 through 10 372, 1999. Accordingly, the follow-up period for the incidence analysis began in 1970 and the referent rates for 1999 were appliedfor the years 2000-2002. Standardized incidence ratios (SIR) were estimated for participants by exposure group and by weighted exposure. Cut-points for the `weighted exposure were selected to correspond to 1, 5, and 10years of employment in high exposed jobs. `The expected numberof bladder cancer cases was also determined for the non-respondenttos the questionnaire by the above weighted exposure categories. For this analysis, the time at risk was `estimated from the beginningofemployment till the endofthe study, which estimated the `maximum numberofexpected cases, (based on prevailing rates). These estimates were used to evaluate potential effectsofselection bias from non-participation. Estimatesofbladder cancer risk by relative cumulative exposure using the study population as an internal referent were made using Poisson and logistic regression. Standardized rate ratios, adjusting for age, gender, and calendar period were estimated with Poisson regression. This `analysis accounted for the time-dependent nature of the PFOS exposure, as participants accrued PFOS exposure at different levels over the courseofemployment. A summary analysis was conducted using logistic regression to estimate the riskof cumulative exposure at follow-up, gender and smoking habit. The precision ofthe estimates for all analyses are described with 95% confidence intervals. 1 373 Results Of the 2,083 original membersofthe cohort, 188weredetermined to be deceased at the time the questionnaire was sent and 1,400 responded to the questionnaire. The remaining 495 did not respond either because they declined to participate or were not reachable. Overall 73.9 percent ofthose eligible responded. For eligible cohort members with no occupational PFOS exposure, only low exposure or high exposure for lessthanone year, and high exposure for one year or more the response rates were 75.8% (563of 741), 81.4% (358of440) and 67.2% (480 of714) respectively. The response rate for women was slightly higher than men and the respondents were older and less likely to have a history of working in PFOS exposed areas of the plant (Table 1-2). Eleven bladder cancer cases were included in the analysis. Fiveofthe cases were identified on the death certificates and 6 were reported on the questionnaire. (Table 3) The diagnosisofthe selfreported bladder cancers were confirmed for two cases, but four self-reported cases declined to sign the consent forms permitting validation. One self-reported case indicated on follow-up that the bladder cancer wasnotaprimary tumor, but declined to provide access to medical records; this casewas not included inthe analysis. An additional nine individuals marked the question pertaining to a historyofbladder cancer diagnosis as `unknown', These respondents marked several typescanceras unknown, thus it was assumed that bladdercancerwas not diagnosed. Twoofthe bladder cancer cases were women and all were over 50 yearsof age at the time of diagnosis or death. (Table 4) The median ageofthe bladder cancer cases was 63. (Table 5) Compared totherestofthe questionnaire respondents, the bladder cancer cases were more likely to smoke, with 83 percent havingever smoked regularly, compared to 56 percentofthe non. cases. Two (18%)ofthe bladder cancer cases never worked in PFOS exposed areas while 9 `worked at some time ina low or high exposedjob, and sixofthese worked for a year or more in thesejobs. Only three bladder cancer cases worked in a high exposurejob for a least one year. `These three are the same cases identified in the mortality study. 2 3K For the respondents and decedents, 8.6 casesof bladder cancer were expected in the period of follow-up (SIR=1.28, 95% C1=0.64-2.29).(Table 6) The SIR for women was 6.42 (95% Cl= 0.78-23.18), but this was based on only 2 cases. The women who reported bladder cancer did not work in the exposed jobs. The SIRs ranged from 1.1 to 23 for the various exposure groups with the highest being the participants ever employed in low exposed jobs. The highest SIRS for the exposure specific strata based on the cumulative exposure score was 2.7 (95% CI=0.55-73.95) corresponditnog 5-10 yearsofemployment in the higher exposedjobs. The SIR for the workers who held a high exposure job for at least one year was 1.12 (95% C1 0.23-3.27). There were 495 eligible cohort members who did not complete a questionnaire. Based on the U.S. population rates from SEER andthenon-respondents age, and gender, an additional 1.93 casesofbladder cancer would be expected to occur in the non-respondent group during the study period (Table 7). The breakdownofthe expected numberof cases by cumulative exposure category is also presented in Table 7. The riskofbladder cancer was similar for men and women. (Table 8) Those who had ever smoked regularly had a higher risk of bladder cancer, but fiveofthe eleven cases were identified by death certificate (smoking status coded as missing) sothe overall impactofsmoking is hard to characterize. There was no clear association between employment in PFOS exposedjobs and bladder cancer. Similar results were observedforthecumulative exposure estimataets the end of follow-up (table 8) and the time dependent exposure analysis (Table 9). Compared to the cohort memberswho never held a PFOS exposed job, those with moderate PFOS exposure had nominally higher riskofbladder cancer. It must be noted in allofthese analyses that the estimatesare very unstable as therewereonly eleven bladdercancercases. Fluctuations in the point estimates, particularly when the data are adjusted for other covariates, such as smoking, gender, and age, may be artifactsofthe analysis. 13 378 Discussion The primary objectiveofthis study was to evaluate whether the finding fromthemortality study ofa twelve-fold risk ofrisk of bladder cancer associated with ever working in a high PFOS exposedjob was representativeofthe overall bladder cancer experienceofthe cohort, The resultsofthis study suggest that bladder cancer incidence in this cohort is similar to the incidenceof bladder cancer in the U.S. population with the same age and gender. Moreover, the isk of bladder cancer in this analysis does not appear to be significantly influenced by `employment in jobs where PFOS exposure is more likely. Clearly the greatest uncertainty in these results is the completenessofcase ascertainment. This study used a postal-questionnaire with telephone follow-up to identify all additional cases of bladder cancer in the population. This approach was taken because there is no population-based cancer registry available for this geographic area or time period. While the overall participation vias reasonable for a mailed survey, it is quite possible that some bladder cancer cases were missed. An additional two cases of bladder cancer were expected in the non-respondents based on SEER data. The potentially missed cases would only affect the resultsifthe rate inthenonparticipants was higher than the prevailing rates in the population. Asensitivity analysisofthis effect can describe the potential effectofunderestimating the true number ofcases.Ifthe rate in the non-participants were twice the population expected rate, the net effect would produce an overall SIR of 1.41 (95% CI=0.79-233). The SIR for the weighted cumulative exposure strata (Table 6) would be 1.27, 1.11, 2.57, and 1.53; all with wide confidence intervals that include the. null. To reach nominal statistical significanace the p<0.1, 0.05, and 0.01 levels, the rate of bladdercancerin the non-respondents would need to be 2.07, 2.60, and 3.63 times thatofthe rate ofthe reference population respectively.Ifthe two highest cumulative exposure strata are combined an SIR of 2.00 (95% CI = 0.75-4.29) would resultifthe rateofbladder cancer in the. non-participants was twice the general population. The rates in the non-respondents for this `combinedstratawould need to be 1.63, 3.27, and 4.91 timesthe population rates to reach significanceatthe p<0.1, 0.05, and 0.01 levels. The likelihood that cohort members who experienced bladder cancer were more or less likely to respond is speculative. However, as described in the sensitivity analysis the overall effect would be marginal, unless there were 14 37 substantially higher bladder cancer rates in the nonrespondent group than the reference. population. The validityofthe cases ascertained by the postal questionnaire is also a potential limitation of this study. Only two of six self-reported cases consented to release their medical records pertaining to this diagnosis to confirm this diagnosis. The reasons for declining to consentto the. validation protocol, be they concens for personal privacy or the validityofthe self-reported diagnosis, could not be determined. Nevertheless, the other self-reported cases were included in the analysis. Self-reportofbladdercancerhas been shown to be quite valid in other`populations. Arecent study conducted by the University of Minnesota andtheNational Cancer Institute using the same approach employed for the Decatur studyhadone-hundred percent ofthe 59 bladder cancers for which medical records confirme>d! A significant non-occupational causeof bladder cancer is smoking, which is believed to be due ptrooPblAeHm'astiicn. tThheeciognalryetstemoskmionkge.i"nfAosrcmearttiaoinnamveaniltaobflesmwoaksirnegpohratbeidt oinn tthhee sqtuuedstyipoonpnuailraet,iotnhuwsanso smoking information was available for deceased cohort members, which accounted for five of eleven bladder cancer cases. Five ofthe sx cases reported on the questionnaire had a history of smoking. Adjusting for smoking and vital status inthe logistic regression models did not.change the overall resultsofthe analysis; however the analysis is statistically unstable with so few subjects in each category. The prevalenceofsmoking in the cohort may also confound the comparison to the general population. The questionnaire asked whether the participants ever smoked regularly (more they stopped fthey had. than 100 Overal, cigarettes), the lifetime the number prevalence ofyears they smoked, and ofregular smoking was 56 at whatage percent based on the questionnaire responses. The prevalenceof current smoking was 17 percent (13 percentofmen and 16 percent of women). Data from the Centers for Disease Control and Prevention indicate that nationally 25 percent ofmen and 21 percent ofwomen wereregular smokers in 2001, but the prevalence ofcurrent smoking was as high as 52 and 34 percent for `men and women respectively in 1965. While it is possible that smoking was under-reported by the participants, it does not population when compared appear that smoking to the national data. would account for any excess risk in this 1s Toxicological studies on laboratory animals have not shown effects in the urinary bladder due to exposure to PFOS and related compounds Two year feeding bioassays in rats ofN- ethylperfluorooctanesulfonamide alcohol (N-EtFOSE) and PFOS have not shown an increased riskofbladdertumors."#3 The former compound can metabolize to an undetermined degree to PFOS. Most bladder carcinogens are genotoxic and/or precipitate in the urine. PFOS and related chemistries are neither genotoxic nor insoluble in urine at the levels measured in employees '. A recently completed POSF inhalation studyofrats wes conducted in response to the finding of bladder cancer from the mortality study. The resultsofthis study have been reviewed by Dr. S. Cohen ** and indicate thereisno evidence that the POSF exposure, which results in high PROS body burdens, leads o treatment related changes in the urinary bladder histology. Overall the results ofthis study do not confirm the twelve-fold excess risk of bladder cancer that was reported in the mortality studyofthis population of fluorochemical manufacturing workers.'* However, the possibility remains for a smaller risk (approximately 1.5 to 2 fold) in the higher exposed workers. Bladder cancer is arelatively rare disease and the population occupationally exposed to these compounds is limited, thus the power to detect excess risks in this study is inherently low and prohibitead convincing exposure response analysis. Future follow-up of this cohort may be limited to mortality studies, which will need to consider these: results when interpreting any findings on bladder cancer. 16 References 1. 3M Company. SIDS initial assessment report: perfluorooctane sulfonic acid and its salts: USEPA Public Docket AR-226. Saint Paul: 3M Company, 2000. 2 HansenKJ, Clemen LA, Ellefson ME, Johnson HO. Compound-specific, quantitative characterizationoforganic fluorochemicals in biological matrices. Environmental Science and Technology 2001;35:766-770. 3. Giesy JP, Kannan K. Global distributionofperfluoroctane sulfonate and related perfluorinated compounds in wildlife. Environmental Science and Technology 2001;35:1339-1342. 4. Olsen GW, Church TR, Larson EB,vanBelle G, Lundberg JK, Hansen KJ, Burris IM, Mandel JH, Zobel LR. Serum concentrationsofperfluorooctanesulfonate and other fluorochemicals in an elderly population from Seattle, Washington. Chemosphere. 2004;54(11):1599-611. 3. Olsen GW, Church TR, Miller JP, BurrisJM, Hansen KJ, Lundberg JK, Armitage JB, Herron RM, Medhdizadelkashi Z, Nobiletti JB, O'Neill EM, Mandel JH, Zobel LR. Perfluorooctanesulfonate and other fluorochemicals in the serumofAmerican Red Cross adult blood donors. Environmental Health Perspectives 2004;111(16):1892-901. 6. Olsen GW, Church TR, Hansen KJ, Burris JM, Butenhoff JL, Mandel JH, Zobel LR. Quantitative evaluationofperfluorooctanesulfonate (PFOS) and other fluorochemicals in the serum ofchildren. J Children's Health 2004;2(1):53-76. 7. Scacat AM, Thomford PJ, Hansen KI, Clemen LA, Eldridge SR, Elcombe CR, Butenhoff JL. Sub-chronic dietary toxicityofpotassium perfluorooctanesulfonate in rats. `Toxicology. 2003;183(1-3):117-31. 8 Case MT, York FG, Christian MS.Ratand rabbit oral developmental toxicology studies with two perfluorinated compounds. Intemational JournalofToxicology 2001:20:101109. 9. Luebker DJ, Hansen KI, Bass N, ButenhoffJL, Seacat AM. Interactions of fluorochemicals with rat liver fatty acid-binding protein. Toxicology 2002;176:175-18s, 17 379 10. Starkov A, Butenhoff JL, Seacat AM, Wallace KB. Structural determinants of mitochondrial dysfunction caused by in vitro exposure to selected perfluorootan] compounds. Toxicologist 2001;60:1658. 11. Wallace KB, Luebker DJ, Butenhoff JL, ethylperfluorooctanesulfonaminde)-ethy! Seacat AM. alcohol are Perfluorooctanesulfonate and 2+(n- peroxisome proliferators in rats, but not guinea pigs. Toxicologist 2001;60:348. 12. Olsen GW, Burris IM, Burlew MM, Mandel JH. Epidemiologic assessmentofworker serum perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) concentrations `and medical surveillance examinations. Journalof Occupational & Environmental Medicine. 2003;45(3):260-70. : 13. Olsen care in GW, Burlew MM, Marshall JC, Burris JM, Mandel a perfluorooctanesulfony! fluoride production facil JH. ity. Analysisofepisodes Journalof Occupation of al & Environmental Medicine 2004;46(8):837-846. 14. A & lexander BH, Olsen perfluorooctanesulp GW, honyl Burris JM, Mandel JH, fluoride manufacturing Mandel facility. JS. Mortalityofe Occupational & mpl oyee s of Environmental Medicine. 2003;60(10):722-9. 15. International Agency for Research on Cancer. IARC `monographs on the evaluation of carcinogenic risks to humans:Re-evaluation ofsome organic chemicals, hydrazine and hydrogen peroxide (Part One). Vol. 48. Lyon: IARC, 1999. 16. Silverman DT, Morrison AS, Devesa SS. Bladder Cancer. In: Schottenfeld D, Fraumen JF, eds. Cancer Epidemiology and Prevention. Second ed. New York: Oxford University Press, 1996;1156-1179. 17. Intemational Agency for Research on Cancer. ARC `monographs on the evaluation of carcinogenic risks to humans: Some flame retardants and textile chemicals and exposures in the textile manufacturing industry. Vol. 48. Lyon: IARC, 1990. 18. Vineis P, Pirastu R. 1997;8(3):346-55. Aromatic ami ann d caencesr. Cancer Causes & Control 18 19. Boffetta P, Jourenkova N, GustavssoPn. Cancer risk from occupational and environmental exposure to polycyclic aromatic. hydrocarbons. Cancer Causes & Control 1997;8(3):444-72. 20. Bonassi S, Merlo F, Pearce N, Puntoni R. Bladdercancer and occupational exposure to `polycyclic aromatic hydrocarbons. International Journal of Cancer 1989;44(4):648-51. 21. Hours M, Dananche B, Fevotte J, Bergeret A, Ayzac L, Cards E, Etard JF, Pallen C, Roy P, FabrJy. Bladder cancer and occupational `exposures. Scandinavian Journalof Work, Environment & Health 1994;20(5):322-30. 22. Steineck G, Plato N, Norell SE, Hogstedt C. Urothelialcancer and someindustry-related chemicals: an evaluation of the `epidemiologic literature. American Journalof Industrial Medicine 1990;17(3):371-91. 23. Weiss NS. Canceinr relation to occupational exposure to perchloroethylene. Cancer Causes & Control 1995;6(3):257-66. 24. Ward EM, Bumett CA, Ruder A, Davis-King K. Industries and cancer. Cancer Causes& Control 1997;8(3):356-70. 25. Anton-Culver H, Lee-Feldstein A, `Taylor TH. Occupation and bladder cancer risk. American Journalof Epidemiology 1992;136(1):89-94. 26. Porru 8, Aulenti V, Donato F, Boffetta P, Fazioli R, Cosciani Cunico S, Alessio L. Bladder cancer and occupation: a case-control study in northern Italy. Occupational & Environmental Medicine1996;53(1):6-10. 27. National Cancer Institute. Surveillance Epidemiology and End Results: Incidence: Urinary Bladder. Vol. 2004 National Cancer Institute, 2004. 28. Olsen GW, Logan PW, Hansen KJ, Simpson CA, Burris JM, Burlew MM, Vorarath PP, Venkateswarlu P, Schumpert JC, Mandel JH. An occupational exposure assessment ofa `perfluorooctanesulfonyl fluoride `production site: biomonitoring. AIHA Journal: 2003;64(5):651-9. 29. National Cancer Institute. Surveillance `Epidemiology and End Results. Vol. 2004 National Cancer Institute, 2004, 19 3a/ 30. National Institutes for Occupational Safety and Health. PC LTAS: Life table analysis systemforuse on the PC. Cincinnati: U.S. DepartmentofHealth and Human Services, 1998, 31. Sigurdson AJ, Doody MM, Rao RR, Freedman DM, Alexander BH, Hauptmann M, Mohan AK, Yoshinaga , Hill DA, Tarone R, Mabuchi K, Ron E, Linet MS, Cancer Incidence in the U. S. Radiologic Technologists Health Study, 1983-1998. Cancer 2003;97:3080-3089. 32. Seacat AM, Thomford PJ, ButenhoJfL.f Terminal observations in Sprague-Dawley rats after lifetime dietary exposure to potassium perfluorooctane sulfonate. Toxicologist 2002;66:185. 33. Thomford PJ, Seacat AM, ButenhoffJL. Terminal observations in Sprague-Dawley rats after liftime dietary exposure to N-ethyl perfluorooctanesulfonamide ethanol. Toxicologist 2002;66:185. 34. Seacat AM, Thomford PJ, J HK, W OG, TCM, L BJ. Subchronic toxicity studies on perfluorooctanesulfate potassium salt in cynomolgus monkeys. Toxicological Science 2002;68:249-264. 35. Cohen'S.Letterto Dr. John Butenhoff. St. Paul, MN, 2004. 2 380 D`Teacbalteu1r. mGoernbdiedri,tyagset,uadnyd exposure characteristicsofparticipants and non-participants in the Toul Gender M F Ageatendof <30 study 30-39 4049 50-59 60-69 70+ Yearsworked <5 59 044 Is19 20+ PFOS Exposure Group* Questionnaire Respondent Yes No N % NN % 1400 95 Deceased Total N% 188 2083 137 812 416 840 23 188 9 160 177 11 941 1730 59 353 5 65 04 46 3 41 06 83 8 2 4316 U2 127 204 210 144 291 604 431 207 418 40 55 213 478 293 866 352 8 251 57 84 16 170 32 43 27 29 419 U2 17 a7 312 148 106 1179 120 86 584 47 Im 347s 67 135 33 38 77 36 38 77 18 180 364 59 27 66 176 248 138 17s 96 17 314 83 Non Exposed" S62 401 179 362 68 362 809 Low Ever 413 295 121 244 67 356 601 dyer 320 29 78 158 52 277 450 High Ever 624 446 276 558 82 436 9% yer 480 343 234 473 69 367 783 LoworHigh Ever 838 599 316 638 120 638 1274 yew 689 492 273 552 108 574 1070 *Categories overlap ** Non exposed includes workers from the film plant who never worked in the chemical plant. 2 383 `Table 2. Mean, median, andnon-participants in and rangeofagesanddays in the Decatur morbidity study PFOS exposure groups for participants Total Ageatendof study Mem Median Min Max QYueestsionnaire Respondent No 1400 95 556 52.6 289 86.3 527 55.4 288 83.0 Deceased 188 53.7 53.7 18.9 85.7 Total -- 54.7 54.7 18.9 863 Years worked Mean Median Min Max Days in PFOS Exposure groups Non Exposed Mean Median Min Max 15.5 152 1.0 36.6 3461 1248 0 13051 13.9 103 1.0 36.7 2818 802 0 12610 13.3 11.0 1.0 36.2 2336 1120 0 12637 149 132 1.0 36.7 3234 1053 J 13051 Low MMeedainan Min Max 8303 0 13030 4504 11867 9038 10428 704 13030 High . Mean Median Min Max 1383 172 0 12276 1702 0 0 12302 1279 0 0 10804 1450 0 0 12302 High or low MMeedainan 2221m 6 2254526 2525138 242093 Min 0 0 0 0 Never exposed incMluadxes workers fro1m3t3h5e2fim plant wh1o23n0e2ver work1ed32in11the chemi1c3a3l5p2lant. 2 `Table 3. Summaryofbladder cancer cases by sourceofreporting and confirmation status Sourceofreporting Death certificate: Questionnaire confirmed by medical records Questionnaire: Declined consent to medical records N Included in analysis 5 Yes 2 Yes 4 Yes Questionnaire: Reported as secondary tumor 1 No Questionnaire: Reported as unsure if ever had bladder cancer 9. No 2 `Table 4. Demographic and exposure cBhlaraadctdeerirsticsofbladderNcaoncer cases and non-cases. All Total N %N% N i 1577 1588 Vital status Alive Deceased 6 5 S45 1394 88a 4518 116 1400 188 Gender M F 9 2 818 130s 828 82 2m 172 1314 2m Ageatendof study <30 30-39 40-49 50-59 60-69 0+ 0 0 0 13 08 0 86 ss 13 86 0 4 0 34 212 364 655 als 334 659 s 2 455 30 27 182 9 63 395 101 `Tobacco use" Cigarettes Missing Ss 0 83 0 785 13 563 09 79 13 Smokeless Missing 0 1 0 206 148 168 21 15 206 2 PFOS Exposure Never 2 182 628 398 630 Low Ever >1 year 7 5 636 473 300 455 361 233 480 an High Ever 21 year 6 3 S45 700 444 273 s 346 706 49 Lowor High Ever 21 year "Excludes decedents. 9 6 818 949 602 S45 791 502 958 797 2 Tcaasbelsen5o.nM-ceaasne,s.median, and rangeofage and days in each exposure category for bladder cancer -- BBladddeerr NNoo AAlIl Age Mean 6238 553 55.4 Median Min 633 553 505 189 554 189 Max 72 863 863 Days in PFOS Exposure groups Non-exposed Mean Median 4387 4571 3357 1215 3364 1228 Min Max 0 0 0 11286 13051 13051 Low Mean Median IE 7 841 0 845 0 Min Max 0 0 0 7139 13030 13030 High Mean Median 923 21 1374 0 1371 0 Min Max 0 0 0 4963 12276 12276 LoworHigh --_-- Mean Median Min MMaaxx 293 216 216 606 370 mn 0 0 0 77113399 113335522 113353522 25 387 Traatbelsein6.US.tSa.npdoaprudliazteidoInn(cSiEdEenRc)e.Ratios by Exposure Category based on expected bladder cancer OBS Exp SRY 95% CT All 1 86006 128 0.64229 Men 9 82891 109 0,502.06 Women 2 03s 642 078-23.18 Never exposed 2 32995 061 007-219 Ever high 6 341 1m 0.643.79 Ever low 7 30918 226 0914.67 Ever loworhigh 9 si 170 077-322 High! year 3 2688 L112 023327 Highor low 21 year 6 ass 131 0.48285 `Weighted Exposure" 3653 3654-18263 18264-36525 2336562566 2 1847 107 4 420 09s 3 09 27m 22 1414003300 114433 0.12385 025-243 055-73.95 040.6165:50.155 *Standardized by age, calendar year, and gender. an**dY>e1ar0syienaresxopfoseumrpeljooybmsemnutltiinplhiiegdhbPyFeOxSpoesxuproesweedijgohbts..Cutpoints to represent 1, 1-5, 5-10, 2 Table 7. Expected bladder cancer cases among 495 eligible non-respondents to questionnaire based on bladder cancer rates in U.S. population (SEER). CL All non respondents Fepecied Cas 1.9266 `Weighted exposure category* 3653 3654-18263 18264-36525 236526 0.5198 0.7959 0.2995 03114 >*1Y0eayresarison efxepmopsluroeyjmoebnstmuinlthiipglhiePdFbOySexepxopsousreed weight. jobs. Cutpoints to represent 1, 1-5, 5-10, and 27 Table 8. Risk estimates for bladder cancer by gender, smoking history, and cumulative exposure atthe timeoffollow-up. Bladder No -- Total N N OR i 1577 95% CI -- Gender M F 9 2 1305 10 72 107 023490 Ever smoked" NYoes Missing 1 5 596 78 1.0 38 1.0 044326 5 196 152 1761309 Cumulative PFOSTM exposure Never Low o r high <lyear Highfor21 year 2 6 628 403 1.0 39 08202 3 56 18 0384 `Weighted exposure *** Quartiles 2914 Q22915-8055 Q3 8036-20525 Q4220526 2 9510 3 9413 0279 2 4 395 17 39315 09-53 03-84 Weighted exposure categories" 3653 2 20 10 3654-18263 18264-36525 4 3 BT 09 12 26 02:53 04-102 . 236526 2 268 14 02102 Deceasee d pere sons were included in the missing category for smokingeee ee +*+EAxdpjoussutreedsfoarraegwee,igahntdedgednadyesrof employment. + Cutpoints represent 1, 1-5, 5-10,and>10 yearsofemployment in high PFOS exposedjobs. 23 3% Table 9. Estimated population. bladder cancer rates and rate ratios using thecohortas an internal referent Cases Personyeas Rate RRY 95%CL Total i] 3739 0.2515 3We6i5g3hted Exposure" 2 15658 01277 10 3654-18263 4 18955 02110 083 0.15465 18264-36525 3 4077 07358 192 030-1206 236526 2 5048 03962 152 021-1099 *Adjusted for age, and gender **Cutpoints represent 1, 1-5, 5-10,and>10 yearsofemploymenint high PFOS exposedjobs. 2