Document Gn58mpxM0NJjq2BopzjQ7D7N

CorporatOeccupationMaeldicine 3M CenterB,uildin2g20-3W-05 St.PaulM,N 55144-1000 IV571374230Telephone 651 7339066Fax 4. MortalitAymong Employeesofa PerfluorooctanoAiccidProductionPlant Thispublishepdaper(GillilaannddMandel;J OccupMed 1993;950-95i4s)thesecond update of theoriginalretrospectivceohortmortalitystudy(seestudies# 2 and #3).This mortalityanalysisisthe second-halfof thedoctoraldissertatiobny Gilliland(seestudy# 5).The cohortconsistedof 1339 male and 245 female workers employed between 1947 and 1993 inthe chemicaldivisionofthe3M CottageGrove (Minnesota)manufacturing site.Also analyzedfortheirmortalityexperiencewere 1449 male and 504 female employees who worked inthe non-chemicaldivisionatthissite.Vitalstatuswas ascertainedthrough 1989. There were no significantliyncreasedcause-specific standardizedmortalityratiosfor eithermen orwomen. Ten yearsof employment (or more) inthechemicaldivisionwas associatewdith a 3.3foldincrease(95% Cl 1.010.6)inprostatecancermortalitycompared tothenon-chemicaldivision.Gillilanadnd Mandel suggestedthe resultsbe interpretecdautiouslydue tothe factthattherewere only fourprostatecancerdeathsinthe chemical divisioncohort.Based on a subsequentreview ofwork historyrecords,only one of the fouremployees inthe Chemical Divisionwho diedfrom prostatecancerappeared to have worked directlyinthe PFOA production building(OlsenetalJOEM 1998;40:614-622).Thispointsoutanotherlimitatioonf this studyinthatthe entirework historyrecordswere not abstracteduringthe originasltudy (seestudy4 2).This abstractioonf work historyrecordswas recentlycompleted.An ongoing retrospectivceohortmortalitystudywillutiliztehisinformationtodevelop a more refinedexposure matrixforPFOA. Vitalstatuswillbe updated through 1997. A finalreportisscheduled for September 2000. 950 Gillila&ndMandel-Mortalitiyna PFOA ProductioPnlant MortaliAtmyong Employeesofa PerfluorooctanAociicdProduction Plant FrankD.GiiiiiaMnDd, PhD erfluorooctaancoiidc(PFOA)andits JackS.Mandel,PhD,MPH salta,mmonium perfluorooctanoate, P are perfluorinatesdurfactantsB.ecauseof theirunique surfaceactive propertiesthey are used in a large Perj7uoroocianoiaccid (PFOA) hasbeenfoundatlowlevel(s10to100 numberofindustraipapllicatiaonnds partsperbillioinn)seraofthegeneraplopulatioand athighelrevelisn consumerproductisncludinpglasti- occupationaelxlpyosedworkersA.lthougPhFOA hasbeenreportetdobea cizerlsu,bricanwtest,tinaggentsa,nd promoterofrodenhtepatocarcinogenaensditsoalterreproducthiovremones emulsifiersD.e's-p'itetheirwideinhumans and rodentst,hereislittilneformatioonnhuman healtehffectsspreaudse,littilseknown aboutpo- associatweidthPFOA exposureT.hepresensttudyexaminedtherelationshiptentiaadlversheealtehffects. betweenPFOA and mortalituysinga retrospecctoihvoermtortalidteysign. PFOA inducemdarkedhepatomeg- Thecohorctonsistoefd2788maleand749femalweorkeresmployedbetween alyand peroxisomperoliferatiinon 1947and1983ataplantthatproducedPFOA. Theall-causetsandardizedrodentlivers.T"he chemicalldyi- mortalirtaytiwoas.75(95% confidenicneterv[aClII.,56to.99f)orwomen 'ersg'roupo.fxenobiottihcastinduce and .77 (95% Cl, .69 to .86)for men. Among men the cardiovascular Pe'OX'SomesI ofconcernbecauseof standardizedmortalityrate was .68 (95% Cl, .58 to .80)and the all- itsassociatiownithnongenotoxiche- gastrointestindailseaseswas .57 (95% Cl,.29to.99).Therewas no signifi- patocarcinogenesisP.F'O-A" did not cantlyincreasedcause-specifsitcandardizedmortalityratiofor eithermen partoodcueclelaunlcaiarnrccrienaosmeandsuimnbae2r-yoefahrerpa-t or women. Ten yearsofemployment inexposedjobswas associatewditha feeding study.'However, biphasic 3.3-folidncrease(95% CI, 1.02to 10.6)inprostatecancermortalitycom- (initiatiaond promotion)and tri- pared tono employment inPFOA productionT.here wereonlysixprostate phasic(initiatisoenl,ectioann,d pro- cancerdeathsoveralalndfour among theexposedworkers;thus,the results motion)hepaticcarcinogenessitsud- must be interpretecdautiouslyI.f prostatecancermortalityis relatedto iesinrodentshaveshown significantly PFOA, PFOA may increasperostatecancermortalitbyyalterinrgeproductive increasendumbers of carcinomasin hor?nonesinmale workers. thePFOA-treatedmts.","Ithasbeen suggestedthatthemarked rodentbep- atomegaly produced by PFOA isa marker for carcinogenicpotential." The observationosfincreaseLdeydig celltumorsina 2-yearratPFOA feed- ingstudyand of disruptioonf the hypothalamic-pituitary-gaoxniasdin PFOA-treatedrat!aereconsistenwtith thehypothesisthatPFOA-associated tumorsaremediatedby a hormonal nongenotoxicmechanism. From theDivisionof Environmentaland OccupationalHealth,Schoolof FlubliHcealth, PFOA has a long half-liifne hu- UniversitoyfMinnesota,MinneapolisM.innesota(DrGillilanDdr, Mandel);and Department mans. A studyof occupationalleyx- of InternalMedicine,Occupationaland EnvironinentWMedicinSeectioSnt,PaulRamsey MedicaClenterS,tPaulM,innesot(aDrGifliland). posedworkersshowed thatthe half- Addrescsorrespondetnoc:eFrankD.GillilaMnDd, UniversiotfyNew MexicoSchooolf lifienmen isgreatetrhan 1.5years." MedicineN.ew MexicoTumorRegistr9y0,0CaminodeSaludNE,AlbuquerquNeM, 97131. Hence, accumulationof PFOA may 0096-1'736/93/3509-0950$03.00/0 occur from small,frequentPFOA Copyrigh0tbyAmericaCnollegoefOccupationaanldEavironrnenMteWdicine doses.PFOA intheserum ofthegen- JOM Volume35,Number 9,September1993 951 er;ilpopulationsof industrialized death certificatefsor coding by the men based on US and Minnesota ctitiiitriesi"s-l"ikelyto be the result same nosologistI.n the 25 deathcer- white male mortalityratesfor three til@'inaccumulation of small PFOA tificatefsrom 1970 to 1989 resubmit- latencyintervals1(0,15,and 20 years) t.itises. ted to the nosologistfor ICD coding, and three categorieosf duration of Nk) health problems relatedto therewere no changes in the major employment (5, 10, and 20 years). PFOA exposure were observed in a categoriesof cause of death. The SMRs were calculatedusing the cross-sectionasltudy among workers Workers were categorizedas ex- program developedby Monson." employed at the PFOA production posed or unexposed to PFOA based The relativerisk(RR) and 95% plant." Cross-sectionasltudies of on theirjobhistoriesE.xposed work- confidenceinterval(CI) for deaths PFOA-exposed workers at this plant ers were defined as allworkersem- from allcauses,cancer,cardiovascular have shown thatPFOA was associated ployed for I month or more in the diseases,and other selectedcauses %%-itdhecreased free testosteroneand Chemical Division.Unexposed work- were estimated using proportional increasedestradiol.20 ers were employees who eithernever hazard models."',"The time to event To determine whether mortality worked in the Chemical Divisionor or censoringwas definedastime from from any cause was associatedwith worked in the Chemical Divisionfor firstemployment to eventor to De- occupationalexposure toPFOA, a ret- lessthan I month. Cumulative expo- cember 31, 1989.In models forspe- rospectivecohortmortalitystudy was sureto PFOA was estimatedusingthe cificcauses of death,deaths from conducted at a plant that has pro- surrogatemeasure of months of other causes were censored at the time duced PFOA since 1947. Chemical Divisionemployment. The observed numbers of cause- of death. Age at firstemployment, year of firstemployment, and dura- Methods specifidceathswere compared to the expected numbers of deaths obtained tionof employment were includedas covariatesin the model. The analyses The plant consistsof severaldivi- by applying sex- and race-specific were stratifiebdy gender.The appro- sions,with PFOA production re- quinquennial age, calendar period, priatenessof the proportionalhazard strictetdo the Chemical Division.A and cause-specifimcortalityratesfor assumptions was testedusing strati- number of other specialtychemicals the United Statesand Minnesota pop- fiedmodels with graphicalanalysisof have been produced in thisdivision. ulationsto the distributioonf ob- log (-Iog(survivalv]e)rsusfollow-up The studycohortconsistedof workers who were employed atthe plantforat least6 months between Jan 1, 1947, served person-time.21.21 Bemuse less than I% of plantemployees were nonwhite,white male and white female time relationshipasnd models that testedthe significancoef a product term between exposure and log fol- and Dec 31, 1983. Data were ab- rateswere used for comparison. For IOW_UP tiMe.21.21 Proportionalhamd stractedfrom plantpersonnelrecords, which were maintained on allworkers women, only United Statesrateswere used because cause- and calendar calculationswere conducted using SAS.25 ever employed at the plant.Vital period-specifiMcinnesota ratesfor status was ascertainedfrom the women werenotavaijablTeh.e effectsResults SocialSecuritAydministrationforthe period 1947 to 1982 and from the of latency,durationof employment, A totalof 3537 workers employed and work in the Chemical Division at the plantbetween Jan 1,1947 and NationalDeath Index for the period were examined usingstratifiesdtand- Dec 31, 1983 were identifiefdrom 1979 to 1989. AU workers with un- ardizedmortalityratio(SMR) anal- company records.Six workers who known vitalstatuswere tracedusing yses.Cause-spe6fic mortalityrates had incomplete employment records a varietyof tracingstrategiessuch as directoryassistance,Metronet and were compared between exposed and were excluded from the study.The unexposed workers using stratified cohortconsistedof 2788 (79%) men TRW searches,reverse directories, SMRs.13 SMRs were calculatedfor and 749 (21%) women (Table1).Men motor vehicleregistratiolnistsc,on- tactingneighborsand relativesa,nd the post offices.Death certificates TABLEI were obtainedfrom the appropriate Characteristicosf Female and Male Employees, 1947-1989 statehealth departments for those identifieads,or presumed to be,de- Chemical Division Non-Chemical Division Total ceased.Information concerning the data and cause of two deaths which occurred outside the United States was obtained from family members. A nosologist coded the death certificatesfor underlyingcause of death accordingto the InternationaCllamificatioonf Diseases,8th revisionT.he reliabiliotfythecoding was evaluated by resubmittinga random sample of Female male Numberofworkers Person-yearsof ob- servation Mean follow-up(y) Mean age atemploy- ment(y) Mean yearofdeath Mew yearofdeath Mean " atdeath(y) 245 6029.0 24.6 28.8 1965.0 1981.3 58.7 1339 33385.3 24.8 25.6 1963.8 1978.3 54.2 Femle su 13280.4 26.4 26.9 1962.8 1979.2 54.4 male 1"9 37732.4 26.0 28.9 1962.3 1978.1 58.1 Fomia 749 19309.4 25.8 27.6 1963.5 1979.6 55.4 male 2788 71117.7 25.5 27.3 1963.0 1978.2 56.4 952 Gillilan&d Mandel - Mortalityina PFOA Production Plant contributed71,117.7person-yearsof SMR was .36(95% Cl. .07to 1.05). tatecancerwere 2,03(95c-Ccl..55to observationw.hich were equallydi- The all-causeSsMR for the non- 4.59)intheChemical Divisiongroup videdbetween theChemical Division Chemical Divisionwomen was .91 and .58 (95@-cCf. .07 to 2.09)in the and non-Chemical Division.Women (95% Cl,.64to 1.24)and thecancer non-ChemicalDivisioncohon. In the contributed 19,309.4 person-years, SMR was .91 (95% Cl, .49to 1.52) Chemical Divisiongroup,therewere two-thirdosf which were in thenon- Chemical Division. Vitalstatuswas obtainedfor 100% (datanotshown). Using Minnesota ratesforcompar- ison,the SMR formen forallcauses, 4 observed and 2 expecteddeaths from prostatecancer.There was no significanatssociationbetween any ofthecohort(Table2).There were 50 deathsamong thewomen (II in the Chemical Divisioncohortand 39 in the non-Chemical Divisioncohort) and 348 deathsamong the men (148 deaths in the Chemical Division group and 200 in the non-Chemical forcardiovasculadriseasesa,nd forall gastrointestindailseaseswas significantlylessthan I (Table4).None of thecause-specifiScMRs was largenor was any significantdliyfferenftrom 1. The resultwsere similarwhen the expectednumbers of male deathswere causeof death and latencyin either exposuregroup.For theChemical Divisioncohort,the prostatecancer SMR was 1.61(95% Cl,.32to4.70) in the greaterthan 15-yearlatency group. Table 5 presentsthe finalpropor- Divisiongroup).Death certificatesbased on US mortalityratesF.or the tionalhazardmodel forall-.auseasl,l- were obtainedfor99.5% of deaths. For women, theSMR forallcaises ofdeath(SMR - .75;95% Cl,.56to .99)was significantlloywer than ex-pected(Table3).There was no associatiownith durationof employment or latencyfor deathsfrom allcauses, cancer,and cardiovasculardiseases (datanot shown).Mortalityamong Chemical Divisionwomen was less than expected.In Chemical Division women, the all-causeSsMR was .46 (95% Cl, .23to .86)and the cancer threelatencyintervalst,he SMRs for deathsfrom aD causesranged from .75to .77.For allcancers,the SMRs rangedfrom 1.06to 1.12 and were nonsignificant. Among men, therewas no associationbetween any causeof deathand durationof plantemployment. The all-musesSMRs were .86 (95% Cl, .72to 1.01)fortheChemical Division groupand .69(95% Cl,.59to .79)for the non-Chemical Divisiongroup (datanot shown).The SMRs forpros- cancer,and prostate-cancemrortality among the 2788 male workersemployedformore than 6 months. The estimatedrelativeriskforall-cause mortalityfora 1-yearincreasein age at firsetmployment was 1.08(95% Cl, 1.07to 1.09).Year of firsetmployment and durationof employment were negativelyassociatedwith deathsfrom allcauses.The riskassociatedwith months employed in the Chemical Divisionwas small and nonsignificant. In the finalprostatecancermortal- TABLE 2 VitalStatusand Cause of Death Ascertainmentamong Female and Male Employees,1947-1989 Chemical Division Non-Chemical Division Total vital Status Fwulo Maio Female male Female Male No. % No. % No. % No. % No. % No. % AN" D"d Total 234 95.3 1191 88.9 465 91.6 1249 86.2 699 93.3 2440 87.5 11 4.7 148 11.1 39 8.4 200 13.8 50 6.7 348 12.5 245 100.0 1339 100.0 504 IDO.0 1449 100.0 749 100.0 2788 100.0 itymodel, lengthof employment in the Chemical Divisionwas positively and significantalsysociatewdithpros- tatecancerrisk.The relativreiskfor a 1-yearincreasein Chemical Divisionemployment time was 1.13(95% Cl, 1.01 to 1.27)For 10 years'employment in the Chemical Division, the relauveriskwas estimatedto be 3.3(95% Cl, 1.02to 10.6)compared with workers neveremployed in the Chemical Division.Age at firsetm- TABLE 3 Observed (Obs)and Expected (Exp)Deaths,StandardizeMdortalitRyatios(SMR) WW 95% Confidenceinterval(sCI)for749 Female Employees Cause of Death Obe Exp SMR 95% Cf AJ Causes carxw so 66.74 0.75 0.56-.99 17 23-04 0.71 0.42-1.14 ployment was positivelayssociated withprostatceancermortalityL.ength of time employed in the Chemical Divisionwas not significantrleylated to mortalityfrom lung cancer,gastrointestincaalncer,pancreaticcancer,or diabetesmellitus. Gas"ntestir.w 2 4.54 0.44 0.05-1.59 Respiratory Breast Gerutal 4 4.72 0.95 0.26-2.43 3 5.87 0.51 0.10-1@49 2 3.37 0.59 0.07-2.14 Discussion Thiswas thefirsrtetrospecticvoe- Lymphopoetic 3 2.04 1.47 0.30-4.29 hort mortality study of workers em- Cardiovascular Corebrovascular Gasvointestinal initr@a sucas 10 12,39 0.81 0,49-1.29 3 3.51 0.86 0.01-4.80 3 3.41 0.88 0.18-2.57 4 6.23 0.64 0.17-1.64 1 1.78 0.56 0.01-3.13 ployed in a PFOA production plant. Mortality from all causes in both men and women was signiricantly less than expected. Because of the healthy worker effem, internal comparisons 4 JOM Volume35,Number 9,September1993 953 TABLE4 Deaths and StandardizedMortalityRatios(SMR) Based on Minnesota White Male Rates,Among 2788 Male Employees, 1947-1989, and 1339 Men Ever Employed intheChemical Division1,947-1989 Causes ofDeath AllMale Employ*es Obs Exp SMR 95% Cl Obe Men Employ*d in Chemical Division Exp SMR OS% ei Allcauses 347 Cancer 103 Gastrointestinal 24 Colon 9 Pancreas 8 Respiratory 31 450.79 97.29 26.78 9.42 5.58 30.42 0.77 0.69-0.86 148 1.05 0.86-1.27 40 0.90 0.57-1.33 9 0.96 0.44-1.81 4 1.43 0.62-2.83 4 1.02 0.69-1.45 12 172.96 36.31 9.77 3.46 2.04 11.26 0.86 0.72-1.01 1.10 0.79-1.50 0.92 0.42-1.75 1.15 0.31-4-01 1.96 0.53--S.Oi 1.07 0.55-1.86 suggestedassociationbetween PFOA exposure and prostatecancer must be viewed as hypothesisgeneratingand should not be overinterpretedT.he associationmay be real,may have been a chance rinding,or may be the resulto an unrecognizedenviron- mental factor.However, the biologic plausibilitfyor any associationbe- tween PFOA employment and prostatecancer isProvided by animal toxicologicand human epidemiologic data thatshow an associationbetween Lung 29 Prostate 6 Testis 1 BWder 3 Lymw*poietil-- 13 Cardiovascular 145 CMD* 110 CArebmvascular 10 AJ gasrointes*W 12 AN respiratory 13 Diabetes 8 Injuries 38 Suicide 12 28.94 6.07 0.92 2.18 12.07 212.19 159.09 24.66 21.13 21.75 6.52 47.74 15.09 1.00 0.67-1.44 11 0.99 0,36-2.15 , 4 1.09 0.01-6.05 1 1.37 0.28-4.01 1 1.09 0.57-1.84 5 0.68 0.68-0.80 54 0.69 0.57-0.83 43 0.60 0.32-1.02 4 0.57 6.294.99 8 0.60 0.32-1.06 7 1.23 0.53-2.42 3 0.80 0.56-1.06 31 0.79 0.41-1.39 10 10.70 1.97 0.44 0.75 4.76 76.65 57.74 8.53 8.27 7.77 2.55 31.72 6.99 1.03 2.03 2.28 1.33 1.05 0.70 0.74 0.47 0.97 0.91 1.18 0.98 1.43 0.51-1.84 0.5-r@4.59 0.03-12.66 0.02-7.40 0.34-2.45 0.53-0.92 0.54-1.00 0.13-1.20 0.42-1.91 0.36--1.87 0.24-3.44 0.66--l.39 0.68-2.63 PFOA and reproductivehormone changes.20 The all-causesa,ll-cancera,nd allcardiovascular mortality among women was lessthan expectedin the overallcohort.The low SMRs am most likelytobe a resultof thehealthy worker effect.Latency and duration of plant employment did not have a strongrelationshiwpith the healthy worker effect. The interpretatioonf thisstudyre- *CHD,coronarwy4 aewoscierotihceartdisease. quiresconsiderationof methodologi- cal issues.SMRs for the subgroups of workers are not strictlcyomparable. were made between Chemical I)ivi- to assessthe associationbetween We attempted to calculatestandard- sionand non-Chemical E)ivisionem- PFOA and prostatecanceroccuffence ized rate mtios; however, the rates ployees.Them were no significantly isproblematic.Age-adjustedprostate wen based on small numbers and pro- elevatedSMRs in Chemical Division cancer mortalityratesfrom 1983 to duced unstable mtios.Esdmates of or non-Chemical ]Divisionemployees. 1989 (949 per 100,000)were only PFOA exposure were based on job However, prostatecancer mortality 25% of the incidencerates(99.4)."' history,and categorizatioonfworkers was associatedwith lengthof employ- This low proportion of deathsamong into ever versus never employed in ment intheChemical Divisioninpro- casesattributedto prostatecancerre- the Chemical Divisionmay not mflect portionalhazard analysis.Ten years flecttshe high riskof death forcom- the biologiceffectivdeose of PFOA. of employment in the Chemical Di- petingcausesforthisdisam ofelderly PFOA exposurewas apparentlywide- visionwas associatedwith an esti- men. Given the small number of ob- spread among employees not directly mated 3.3-foldincrease(95% Cl, 1.02 serveddeaths from prostatecancer in exposed to PFOA,"' and the exposure to 10.60)in prostatecancer mortality. thestudy,and the observed difference categorizatiomnay misclassifwyork- The useofprostatecancer mortality in incidenceand mortalityrates,the ersas unexposed when they were ex- TABLE5 ProportionaHlazard Regression Model of FactorsPredictingMortalitaymong AilMale Employees* Varlobl* AllCauses of Death 0 $E(8) p RRt Cancer Deaths 0 SE(P) p RRt ProstateCancer Deaths 0 SE(O) p RRt Yew offirst employment -0.55 0.009 0.0001 0.946 -0.031 0.019 0.11 0.969 0.010 0.081 0.9 1.011 Age atfirsetm- 0.079 0.006 0.0001 1.08 0.078 0.011 0.0001 1.081 0.082 0.045 0.06 1.085 ployment(y) Durationofam- -0.34 0.001 0.0001 0.967 -0.028 O.DD9 0.002 0.972 -n n7 0.052 0.18 0.932 plo@@t (y) Months in 0.001 0.001 0.24 1.001 O.DO2 0.001 0.2 1.002 0.01 0.005 0.03 1.01 cheniiewdivis.ion Abbreviationussed am: p.roWes3k)n perammer.SeA, sWOard oTm of ft Sk)Peparameter;RR, relativreisk. t Relativreiskforone unitchange inkxkp*ndsm v@ftble. 954 Gillilan&d Mandel. Mortalitiyna PFOA ProductionPlant posed.Such misclassificatwiounld beexpectedtobiastheeffectestimates toward the nullifincreasedexposure increases death rates. Months employedintheChemical Divisionmay betterreflecthe biologiceffective dose becausecumulativeexposure reflecttshe bioaccumulationof PFOA. Workers were exposed to many other xenobioticss,uch as benzene and asbestos,duringtheiremployment at theplant.However, none of these materialhsas been associatedwith prostatecancer. Although themean age at fintemployment and mean year of firstemployment aresimilarin the Chemical Divisionand non-ChemicalDivision cohortsof men and women, thecomparisons of the rates of disease are confounded by differences in the distributioonfage at risk.The use of an internaclomparison group may reduce,but not eliminatec,onfounding if the internalcomparison groups have differendtistributionosf these time factors.Bemuse the diseaseoccuffencerelationshiips defined in terms of cumulativeexposure,the true effectof PFOA exposure may have been biasedtoward or away from the nuU by uncontrolledconfounding by time factors."-29 Furtherresearchisneeded to evaluateand confirm the associatiobnetween PFOA and prostatecancer.The findingsin thisstudy arebased on a smallnumber of casesand could have resultefdrom chanceor unrecognized confoundingfrom exposureto other factorsS.tudiesof prostatecancerincidence in this and other PFOA-exposed work forcesmay clariftyhe suggestedincreaseinprostatecancer risk. Acknowledgments Thiswork was supportedinpartby NationalInstituftoerOccupationalSafetyand HealthGrantT150H07098-16.and the3M Corporation. References 1.Guenthner R. VietorM. Surfaceactive materialfsrom Pertluorocarboxylaincd rwrtlui)rosull'oancicds.I&EC' Pri)dut-i 14. Ubcl F. Sorenson S. Roach D. Health Rt,.%turt-ainid Dt,t@tloptpieri1t9.62;1: 165-169. statusof plantworkers exposed to fluo- rocht:micalas:preliminaryreport..4m 2. GrifritFh.Long J.Animaltoxicitsytud- lndHi@g.4.vs(x1-9J80:41:584-589. ieswithammonium perfluorooctanoate. 15.Guy W. Fluorocompounds of human .4mIndlf,@'9.4ssL1x9'8J0-@41:576-583. plasma:analysisp.revalencep,urifica- 3. OlsonC, AndersenM. The acutetoxicity tion,and characterizatiPohnD, thesis. ofperfluorooctanoaincd perfluorodeca- RochesterN.Y: UniversitoyfRochester; noicacidsin male mts and theireffects 1972. on fatttyissuTeo.xicoAlppiPharmacol. 16.Guy W, TavesD, BreyW. Organicfluo- 1983;70:362-372. rocompounds in human plasma:preva- 4. PastoorT. Lee K, PerriM. GilliePs., lenceand characterizatiIonn:.FilteRr. Biochemicaland morphologicasltudies ed.Biochemistryinvolvincgarbon-fluo- of ammonium perfluorooctanoate-in- rinebondsA.CSSymposium SeriesN.ew duced hepatomeg4y and peroxisome York: American Chemical Society: proliferatEixopn.Mol Pathol1.987;47: 1976:117-134. 98-109. 17. Guy W. Ionicand organicfluorinien 5. JustW, Gorso V. Hart)F.Heinemann P. SalazaMr, SchimassekH. Biochemi- blood.In:JohansenE,TavesD, Olsen T, ed.Continuingevaluatioonftheuse caleffectasnd zonal heterogeneitoyf offluoridBeo.ulder,CO: Wcuview Press; peroxisomeprouferatioinnducedby per- 1979. fluorocarboxylaiccidsin rat liver. 18. Taves D. Evidence that there are two HepatoloV. 1989;9:570-581. forms of fluoridein human serum. Na- 6. Kennedy G. Dermal toxicitoyfammo- zure.1968;217:1050-1015. nium perfluorooctinoTaotxei.coAlppi 19.TavesD. Comparisonof"organicf"lu- Pharmacol1.985;81:348-355. orideinhuman and nonhuman serums. 7. Kennedy0, HallG, BritteUJi,Chen H. J Dew Res.1971;50:783. Inhalatiotnoxicitoyf ammonium per- 20. GWiland F. Fluorocarbonsand human fluorooctanoFaotoed.Che7nToxical. healtsht:udiiensanoccupatiocnoahlort, 1986;24:lj25-1329. PhD thesisM.inneapo4 MN: Univer- 8. Cook J,Murray S,Frame S,HurttM. sityofMinnesota;1992. Inductioonf Leydigcelladenomas by 21. BerryG. The analysiosfmortalitbyy the ammonium perfluorooctanaAtep:ossi- subject-yemmethod.Biometrics1.983; bleendocrinerelatemdechanism.Toxi- 39:173-194. colAppiPharmacol.1991;113:209-213. 22. Monson R.Analysisofrelativseurvival 9. Reddy J,AzarnoffD, HigniteC. Hypo- and proportionalrnorwity.Compta Hpidemichepaticperoxisomprolifem- Biomed.Res.1974;7:325-332. torsform a novelclassof chemicalcar- 23. Miettinen0. Standardizatioofnrisk cinogenNsa.ture1.980;283:397-398. ratiosA.m J Epidemiol.1972;96: 10. Reddy J,LalwaniN. Carcinogenesbiys 383-388. hepaticperoxisomeproliferatoervsa:lu- 24. Cox D. Regressionmodels and fife ationoftheriskofhypohpidemicdrugs tables.J R Stat Soc (B). 1972;34: and industriapllas;ticizteorshumans. 197-220. CritRey Toxicol1.9831;2:1-85. 25. SAS InstitutSeA.S User'sGuide:Statis- 11. NiLwn P, BeiieB, PreatV, Erixonr,. tics6.thed.Cary,NC: SAS InstitutIenc; Ramel C. On the mechanism of the 1990. hepatocarcinogenicoiftyperoxisome 26. KalbfleiscJh,Pmntice R. The statistical profiferatorCsh.em BiolInteract1.991; analysisoffailuretimedata.New York: 78:235-250. John Wiley& Sons-,1980, 12. AbdellatAi,fPreatV. Vam=q 1,Ndsson 27. MillerR. Cancerstatistriecvsiew1973- R, RoberfroidM. Peroxisomeproliferationand modulationof ratlivercarci- 1989.Bethesda,Maryland.US Department of Healthand Human Services, nogenesibsy 2.3-dichlorophenoxyacetic NationalInstitutoefsHMth. National acid,2,4,5-trichlorophenoxyaacceitdi,c CancerInstitutNeI.H Pub.No.92-2789, perfluorooctanoaiccidand nafenopin. 1992. Carcinogenesi1s9.90;11:1999-1902. 28. Checkoway H. PierceN, Crawford- 13. TakagiA.SaiK, Umemura T. Hasegawa R. Kurokawa Y. Short-termexposureto Brown D. Researchmethods inoccupationalepidemiologyN.ew York:Oxford the peroxisomeproliferatopresr.fluo- UniversityPress;1989. moctanoic acidand perfluorodecanoic 29. Robins1.A new approachto causalin- acid,causessignificanitncreaseosf 8- ferenceinmortalitsytudiewsitha sus- hvdrox,ydeoxyguanosinienliver'DNAof tainedexposureperiod.Vaih Vodeling. ratsC.ancerLett.1991@5":55-60. 1986;7:1393-1512.