Document gbXbBMkmzzM8wzn63B3xrdOBL

CorporatOeccupationMaeldicine 3M CenterB,uildin2g20-3W-05 St.PaulN,lN55144-1000 651'3/'j4230Telephone 6517339066Fax 6. An EpidemiologiIcnvestigatioofnReproductivHeormones inMen with OccupationaElxposuretoPerfluorooctanoAiccid ThispublishepdaperbyOlsenetal(JOccupEnv Med 1998;40:614-62f2u)rther examined the observationinitialrleyportedby Gillilandinhisdoctoralthesisthatserum PFOA levels(asmeasured by totalorganicfluorine)may be associatedwith reproductive hormone changes. In 199') and 1995,the fluorochemicalmedical surveillancperogram had III and 90 male productionworkers(CottageGrove, " voluntarilpyarticipate. Unlike the 1990 program which analyzedforserum totalorganicfluorine,serum PFOA was measured in 1993 and 1995 by mass spectrometrymethods. Serum PFOA levels were then compared to severalreproductivehormones. Serum PFOA was not significantlayssociatedwith estradioolr testosteroneA. 10% increaseinmean estradiol levelswas observed among. employees who had the highestlevelsof serum PFOA although thisassociationwas confounded by body mass index. Neitherwas PFOA consistentlayssociatedwith any of the othermeasured hormones. 614 Epidomfologic Study of Perfluorooctenoic Acid o Olson at al An Epidemiologic Investigationof Reproductive Horinones in Men with OccupationalExposure to Perfluorooctanoic Acid Geary W. Olsen, DVM, PhD lthough fluoride(inorganicionicflu- Frank D. GillilandM,D, PhD MicheleM. Buriew,MS Jean M. BurrisM,PH Jack S. Mandel, PhD JeffreyH. Mandel,MD A oride) was identifiedin human blood 140 years ago,'-2 the presence of fluorinein a covalently bound or- ganic state was firstreported in .3-4 1968 Guy subsequently identi fied perfluorooctanoic acid (PFOA, C7F,5CO,H) as a major component Perj7uorooctanoic acid (PFOA),apotenstynthetiscurfactanutsedin indust7iaalpplicationiss,a peroxisomeprol@ferattohrathas resulteidn dose-relateidncreasesin hepatic,pancreaticacinar,and Leydig cell adenomas in laboratoryanimalsI.n addition,PFOA increasedserum estradiolevelsthrough theinductionof hepaticaromatase activityI.n 1993 and 1995, we conductedtwo cross-sectiosntauldiesofI II and 80 production workers, respectivelya,nd specifzcallmyeasured theirserum PFO,A in relationtoseveralreproductivehormones to determine whether such. an effect occurs in humans. PFOA was not si@mificantly associated un th estradz*oolr testosteronein eitheryear'sstudy. A 10% increasein mean estradiolLevelswas observed among employees who had the highest levelsof se?um PFOA, although this association was confounded 6y bodv mass index. Neither was PFOA consistentlyassociated with the other measured hormones. Our resultsProvide reasonable assurance that, in this production setting,therewere no significanthormonal changes associatedwith PFOA at the serum levelsmeasured. Limilatiotisof this investigatioinnclude itscross-sectiondaelsign,thefezv sitt@erei.s-cposaetdthehighestlevelsa,nd thelowerlevelosfserum PFO,-L @ne(isiirecdo,mpared with those levelsreportedto cause effectsin labot-alora-nYimal studies. f the serum organic fluorinefrac- 0tion5.Ammonium perfluorooctano ate,a potent synthetic surfactantused inindustriaaplplicatiornasp,idldyissociatesin aqueous solutionto PFOA. In laboratoryanimals,PFOA acid, or itssalts,isabsorbed by ingestion, inhalation,or dermal exposure6- 3 and is not metabolized.9 -12 PFOA is distributedprimarilyin the plasma and liverof male ratsand the liver, plasma, and kidney in female rats." The major routeof eliminationisvia untieand feces.In the female rat, thereisa tenfold-greaterrenalex,etionrate.1'.11.14 Castratedmale rats treatedwith estradiolhave PFOA urinaryexcretiornatesimilar tothoseof femalerats.0'-11 Peroxisome proliferatorsl,ike PFOA. area diverseclassof chemi- cals thatcause hepaticperoxisome proliferatioand enzyme induction. liverhyperplasiaa.nd, in some in- From iheMedicalDepartment.3NI Company. St.Paul.Minn. (Dr Olsen,,%IBsurlew.ivisBurrisD,r J.S.\Iandel)t.he Epidemioloiaavnd Cancer ControlProgram.Schoolof Medicine.Universityof New .%IexicoA.lbuquer4uc.N,%t(DrGilliland)a.nd theDivisionof Environmentaland OccupationalHealth. Schoolof PublicHealth.Universitoyf Minnesota.Minneapolis.Minn. (Dr J.H.%Iandel). A,Idrcscsorre,,,pL)ndteon:cGeeary Olsen.DV,'@tP.hD. MedicalDepartment.3NI.220-3W-05. St. PULII.%IN 55144. 1076-2752/9814007-0614$3.(X)/O stances,hepatocarcinogenesiisn rats and mice."-2' Two-year feeding studiesinCrl:CD BR (CD) ratsata maximum amount of 300 partsper million(ppm) PFOA showed liver adenomas and an increasedinci- C,)p@ri@-,(hDiby American Collegeof Occupuiionaland Environmental'Vledicine dence of pancreasacinarcelladeno- JOEM -Volume 40,Number 7,July1998 615 mas 2:!.2 3 and Levdig cell adenbmas. 17.2:! PFOA is not mutagenic6 and thus the inductionof thesetumors most likelyoccurs through nongenotoxic mechanisms such as oxidativestress.19-24 To determinewhether the Leydig celladenomas were the resultof an endocrine-relatemdechanism, Cook etal"'gavagedCD ratsfor14 days with up to 50 mg/kg of ammonium perfluorooctanoatAe.significanitncrease in serum estradioland decrease in testosteronelevelswere observed.The estradioilncreasemay be due to an inductionof hepatic aromataseactivity.T"he decreasein serum testosteronleevelsmight be the resultof reducedconversionof 17-alphahydroxyprogesterone(17HP) to androstenedion(eviathe inhibitionof the C-17,20 lyaseenzyme). However, Biegeletal'9were unable to replicattehe negativetestosteroneassociation. CD ratsfed 100 ppm PFOA fora maximum of 13 weeks showed increasedestradioblut nottestosterone levels2.5.26 Elevatedestradiotlevels were found among CD ratsfed 300 ppm duringa 2-yearbioassay,with no dose-relatedifferencesfor testosteronel,uteinizinhgormone (LH), and folliclset-miul ating hormone (FSH).22 The 3M Company has conducted medical surveillanceof employees involvedin PFOA production.Levelsof serum organicfluorine(1.0071.00ppm) inproductionemployees were 10-to 50-foldgreaterthanvalues (0.01-0.13ppm) reportedfrom human sera.27 Since PFOA could be an endocrinemodulator,'Ba crosssectionalstudyamong workers with potentialexposure to PFOA was conducted in 1990.28Among 115 men engaged in PFOA production, totalserum fluorinelevels(sodium biphenyl extraction method:19) ranged from 0 to 26 ppm (mean = 3.27 ppm*,standarddeviation[SDI = 4.68 ppm). Ithad been estimatedthat 80%-90% of human totalserum fluorine levelsconsistedof PFOA .27 Adjustingfor potentialcovariates and hormones of a prioriinterest Sixty-eightemployees participated resulteidna positivneonlinear(qua- inboth years.The surveillancceon- dratic)associationwith estradiol. sistedof a medical questionnaire: positivelinearassociationwsith pro- measurement of height.weight.and lactinand TSH. and negativenonlin- pulmonary function.standardbio- ear (squareroot)associationwsith freeor bound testosteron2eB. chemical and urinaivsitsestsP.FOA deterrnniation-a.nd ;everalhormone A mortalitystudyofemployees in assays. the chemical division,which in- cluded the PFOA productionbuild- PFOA Analysis ings,found no significantliyn- A thermospray mass spectropho- creased cause-specificstandardized tometry assaywas used to determine mortalityratiofor eithermale or serum PFOA levelsin 1993 and femaleemployees.""There were four 1995 .32 The range of serum PFOA deaths from prostatecancer,com- was 0 to80 ppm in 1993 and 0 to115 paredwith 1.97expected(95% con- ppm in 1995.The upper limitof fidenceinterval[CI],0.55-4.59). detectionin 1993 was 80 ppm, Only one employee had worked di- whereas therewas no upper limitof rectlyinthePFOA productionbuild- detectionin 1995. Levels were ings.An associationbetween PFOA highlycorrelateadmong the68 em- exposure and prostatecancer was ployees studiedin 1993 and 1995 consideredbiologicallyplausible (r = .91,P = 0.0001).There was based on theanimal and human data, also high correlatiobnetween total which showed associationbsetween PFOA and reproductivehormones. serum fluorinelevelmeasured with the 1990 studY28 and thePFOA mea- The purpose of thisreportis to suredin 1993 (r = .72,P = 0.0001, describetheresultsfrom two cross- n = 94 subjects)and in 1995 (r = sectionasltudiesfrom thesame plant .84,P = 0.0001,n = 63 subjects). thatwere done in 1993 and 1995. These findingswere not unexpected, Methods PFOA Production because of the estimated18- to 24- month half-lifeof PFOA in humans.27 PFOA productionatthisplantbe- Hormone Assays gan in 1947.PFOA, a white powder, Serum samples were analyzed by is produced by an electrochemical theUniversityof Minnesota'sEndoprocess.3 'The productsof thiselec- crinologyLaboratory(Minneapolis, trolysicsellreactionare highlyflu- MN) or the EndocrineScienceRef- orinatedcompounds, with the end- erence Laboratory(Tarzana,CA). product defined by the starting Eleven hormones were assayed:cor- materialP.roductioninvolvesa four- tisold,ehydroepiandrosteronseulfate stageprocess:isolatinagnd convert- (DHEAS), estradiolF,SH, ]7a- ing the chemicalto a saltslurty, hydroxyprogesterone(17-HP), free convertingthe slurryto a saltcake, testosteronet,otaltestosteroneL,H. dryingthe cake,and packaging.The prolactin,thyroid-stimulatihnogr- greatestlikelihoodfor exposureto mone (TSH) and sex hormone- PFOA occurredin the dryingarea. bindingglobulin(SHBG). All but SubjectSelection SHBG were analyzedattheUniversity of Minnesota'sEndocrinology General medical surveillancies Laboratory, performed bienniallfyor employees Cortisolwas assayed usinga fluo- at thisplant.There were III male rescence polarizationimmunoassay employees in 1993 and 80 male em- (AbbottTDx. North Chicago.IL). ployeesin 1995 who participateidn Radioimmunoassays (RIA) were medical surveillanceh.ormone test- used for DHEAS (Pantex,Santa ing,and serum PFOA determination. Monica, CA). estradio()modified 616 EpidemiologsitcudyofPerfluorooeftnAoclied-01"n stof P@iiite1.7x-)H.P(modifieCdIS)and (ol,itlestosteron(eCoat-A Count. Di@ct@ ,,nt)sPtriocductCorp..Los Angeles, CA). Free lestosteronweas deteriiiinedusing equilibriumdialysis. LH. FSH and prolactinwere assayed usina,,a microparticleenzyme immu- comparing parameterestimatesu,sing indicatorand continuousvariables. Stepwise selectionprocedureswere alsoused.Study resultswere analyzed by theSAS System.-13 We did not examine hormone changes between the two examina- Results Serum PFOA levels were not highlycorrelatewdith eitherthecovanates or the hormones. These correlatiocnoefficient(sinparentheses) for 1993 and 1995 data,respectively, noassay(Abbottlmx).TSH was de- tionsbecause of theestimatedhalf- forPFOA and the variableof interest termined usinga chemiluminescence lifeof PFOA (approximateltywo were thefollowinga:ge (-.22,.14); immunometric assay (Nichols,San years)and intraindividuvaalriability alcohol(.10,.18);BMI (.11,.10); Juan Capistrano,CA). SHBG was assessedviaa radioimmunoassay afterchromatographicsample purification (EndocrineScience Reference Laboratory)B.ound testosteronweas calculatedas totaltestosteronleess in hormones. Since theresultsforthe 68 employees who participateidn both yearswere similarto thoseobtainedforthe entirestudy,only the resultsfor allemployees are presented below. cigarettes(.05, .11); cortisol(.07. .05);DHEAS (.13,.12);estradiol (.12,.15F)S;H (-.12,-.13);17-HP (.11,.30)L;H (-.06,.13)p;rolactin (.04,-.04);SHBG (-.07,.03); bound testosteron(e.01,.02);free freetestosteroneT.he same assays were used for both 1993 and 1995 analyses. IGO 1993 n=lll DataAnalysis Simple and stratifieadnalyses,anal- ysisof variance(ANOVA), Pearson correlatiocnoefficientasn,d ordinary E multivariable regression were used to 7'0 evaluateassociationsbetween PFOA and each hormone, with adjustmentfor potentialconfoundingvariablesF.or ui so stratifiaendalyses,employees were di- videdintofourPFOA catecories0:-1 L- 40 ppm, 1-<10 ppm, 10-<30 ppm, and @30 ppm in ordertodetentiiniefan effectexistedat the highestsenim 0 la 30 20 30 40 50 so 70 so so 100 Ito 120 levels.For multivariableevaluation, PFOA(ppm) PFOA, age,body mass index(BNU), alcoholuse,and cigarettuese were examined as both categoricaalnd con- too 1995 n=80 tinuousvariablesA.lcohol usewas an- alyzedas lessthanI drinkperday,1-3 drinksperday, and non-responsetothe So questionnairietem.Cigaretteuse was recordedas eithercurrentsmoker or .9 70 nonsmoker. Re(yressiomnodels we,. fittedwith PFOA enteredas a contin- uous variableu,sing linears,quare,and LU % square roottransformationisnorderto so assure that associationswere not missed.The possiblenonlinearassoci- 40 ationof estradiolf,ree testosterone, and bound testosteronweas evaluated. Nonlineardose-responserelationships were examined by model fitand by 0 10 30( 20 30 40 so 40 70 so 90 too 110 120 PFOA(ppm) Fig.1.Scatterplooftsenimestradi(oplg/mlb)y perfluorooctaancoidc(PFOA.inpartsper millio[nppm])foremploy"s in1993and 1995. JOEM Volume 40,Number 7,July1998 617 1993 nzlll 30 25 20 10 0 0 10 20 30 40 so so 70 so 00 100 110 120 PFOA 1996 n=80 30 2S 220 0 10 20 30 40 so Fig.2.ScatterpolfosterumfreeteStOStCTone and 1995. a 70 w w loo 110 120 PFOA (ng/dib)y PFOA (ppm)foremployeeisn1993 testosteron(e.09,.01);and TSH (.03, were 206 ng/dL of 17-HP and 80 .15). ppm (upper limitof detectionin FiguresI and 2 are scatterplotfsor 1993)forPFOA. estradioland free testosteronfeor Table I providesthe mean, me- each year.Simple linearregression dian,standarddeviation,and range of the naturallog of each hormone of the covariatesand severalhor- with PFOA, treatedas a continuous mones, by four levelsof PFOA cat- variabler,esultedin no statisticallyegorizatio(n0-<I,1-<10, 10-< 30, significanctoefficientisn 1993 for and 2:30 ppm). Seventy-fivepercent any hormone and only one in 1995: of theemployees with serum PFOA 17-HP (betacoefficien=t 0.006, levelsat 10 ppm or greaterpartici- P = 0.03,R2 = .06)T.hisresultwas pated in both years.From Table 1, dependent upon one person.In 1995, severalobservationsare noteworthy. thispersonhad a levelof 198 ng/dL Firstt,he mean of the PFOA ppm of 17-HP and 114 ppm of serum categoriesdifferedby two ordersof PFOA. In 1993,thisperson'svalues magnitude between the lowest and highestcategoriefsorbothyears. Second.theL,30-ppmPFOA category had the youngest mean employee age inboth years.Third.BMI was thegreatestamong employees in the !,30-ppm PFOA categoryin 1995. Fourth,mean estradiolevels were not significantdliyfferenbtetween PFOA levelsin eitheryear. althoughthe a30-ppm PFOA categorieshad mean estradiolevelsthat were 10% higher than the other PFOA levels.Fifth.therewere no discernibletrends between PFOA and eitherbound or freetestosterone. Sixth,17-HP levelswere hiLc-hestin the a30-ppm PFOA group in both years. No significanatssociations were observedforcortisolD,HEAS, FSH, LH, and SHBG (datanot shown). As expected,34 estradiolwas highlycorrelatewdith BMI (1993: r = .41,P < 0.001;1995:r = .30, P < 0.01a)ndfreteestostewriotnhe age(1993r: = -.48,P < 0.001; 1995:r = -.40,P < 0.001)t;hus Table 2 providesmean estradioalnd free testosteronvealuesstratifiebdy BMI and age,respectivelyI.tshould be noted thatallfiveemployees in 1995 with serum PFOA levels2:30 ppm had BMIs @t28. Linearand nonlinearrelationships, taking into account potentiaclonfounders(especialalgye and BMI) as well as othercovariatesthatmay be on the biologicpathway of effect, resultedin no significanatssociationswith PFOA exceptfor 17-HP in the 1995 analyses(datanot shown). Again. thisassociationwas dependent on the one employee discussed earlier. Because a primary hypothesisof thepresentstudywas whetherPFOA increasedestradioland decreased testosteronseerum levelsin a nonlinear fashion,we replicatetdheseprior models2" with our 1993 and 1995 data. PFOA was not significantly associatedwith serum estradiolf.ree testosteroneo,r bound testosterone (datanot shown).There was no significantassociatio(ndatanot shown) between PFOA and prolactinamong GIs - EpidomiOlOgSItcudyofPerfluarooetanAociicdoOlsonatat TABLE1 Mean,MedianS,tandarDdeviati(oSnD)ofMeanandRangeofPerfluorooctaAncoiid(cPFOA)D,emographiacndHormonal ValuesbySerumPFOA Levelsa,ndYearofDataCollection* 1993Data 1995Date PFOA(ppm) PFOA(ppm) o-<i 1 1-<10 10-<30 @30 Mean Median so Range Mean Median 0.48" 3.34* 16.26* 60.13* F = 253.25,P 0.47 2.49 15.40 64.45 0,0001 0.27 2.17 3.39 24.01 0.00-0.99 1.03-8.92 11.90-21.00 31.60-80.00 0.31 3.03 17.11* 55.96$ F = 77.57,P 0.2 2.4 14.3 42.4 0.0001 SD Range 0.32 1.84 6.90 33.29 0.00-0.90 1.10-8.20 10.30-28.20 34.20-114.10 Age (yr) 0-<l 1-<10 10-<30 2t30 43.6 39.2 39.9 33.3 F 3.67,P 45.0 38.0 39.5 32.5 0.01 9.2 27.0-61.0 42.0 41.0 7.7 27.0-60.0 41.3 40.0 4.2 34.0-45.0 45.1 46.0 7.4 25.0-43.0 38.2 35.0 F 0.88,P 0.46 8.3 29.0-60.0 8.6 24.0-58.0 7.4 30.0-55.0 9.2 27.0-50.0 Alcohol(drinks/day) 0-<l 0.4 0.3 0.5 0.0-1.9 0.5 0.3 0.7 0.0-2.9 1-<10 0.7 0.5 0.7 O.D-3.4 0.5 0.4 0.5 O.D-1.9 10-<30 0.9 0.7 0.6 0.4-2.1 0.8 0.7 0.7 0.0-2.1 2t30 0.9 0.7 0.8 0.0-2.0 0.5 0.4 0.6 0.0-1.4 F 3.05,P 0.03 F 0.94,P 0.43 BMI (kgtm2) 0-<l 1-<10 10-<30 at30 28.0 26.8 29.1 28.5 F 1.60,P 27.5 26.3 28.8 28.4 0.19 4.2 20.9-42.0 27.6 26.8 2.5 21.6-32.5 28.6 27.9 1.8 27.1-32.0 27.8 27.7 1.6 26.9-30.2 29.8 28.9 F = 0.77,P 0.52 4.2 21.9-45.2 3.4 22.1-38.3 4.0 21.2-34.8 1.8 28.2-32.6 Cigarefte(scigareftes/day) 0-<l 2.6 0.0 7.5 1-<10 6.0 0.0 10.5 10-<30 2.5 0.0 7.1 :@30 5.0 0.0 10.0 F 1.26,P 0.29 0.0-30.0 3.8 0.0 9.4 0.0-40.0 0.0-40.0 2.6 0.0 6.0 0.0-20.0 0.0-20.0 9.1 0.0 15.2 0.0-40.0 0.0-20.0 6.0 0.0 8.9 0.0-20.0 F = 1.26,P 0.30 Estradio(lpg/mL) 0-<l 1-<10 10-<30 2t30 54.7 56.0 54.8 62.8 F = 0.54,P 17-HP (ng/dL) 0-<l 1-<10 10-<30 @30 106.8 120.2 97.9 126.5 F = 1.55,P Prolacti(n@Lg/L) 0-<l 8.2 1-<10 8.8 10-<30 15.06 2t30 7.5 F = 3.67,P Bound testosteron(eng/dL) 0-<l 1-<10 528.7 609.7 lo-<30 @30 485.2 569.6 F = 2.48,P 53.0 55.0 58.0 63.0 0.66 106.0 115.5 105.5 123.0 0.21 8.0 8.0 9.0 7.5 O@01 513.7 609.2 477.5 596.5 0.07 13.5 12.0 11.6 8.4 34.9 41.5 28.4 66.8 3.5 4.6 15.2 0.6 178.0 168.2 113.9 81.6 32.0-83.0 34.0-81.0 39.0-70.0 55.0-70.0 68.1 65.2 67.1 73.2 F = 0.69,P 66.0 62.0 66.5 75.0 0.56 11.7 45.0-94.0 14.9 45.0-96.0 9.1 54.0-79.0 6.7 64.D-81.0 44.0-203.0 91.6 94.0 32.2 39.0-190.0 45.0-249.0 110.6 105.5 35.6 54.0-179.0 54.0-134.0 110.3 85.5 77.5 46.0-297.0 54.0-206.0 123.0 102.0 54.7 72.0-198.0 F = 1.67,P 0.18 2.0-18.0 2.0-22.0 6.0-51.0 7.0-8.0 10.9 11.8 12.9 9.4 F = 0.66, P 10.0 10.0 14.0 9.0 0.58 220.9-1059.5 212.2-1021.6 301.0-651.6 450.9-634.4 534.8 518.5 567.7 564.9 554.4 549.7 567.8 623.0 F = 0.26,P = 0.85 S.i 4.0-23.0 6.0 5,0-28.0 5.3 3.0-21.0 2.7 7.0-14.0 150.4 152.3 185.5 155.0 278.7-1059.5 216.4-898.4 238.1-823.8 341.7-703.0 JOEM -Volume 40,Number 7,July19H 619 TABLE1 Continued 1993Data 1995 Data PFOA(ppm) Mean Median SD Range Mean Median SD Freetestosteron(eng/dL) 0-<l 15.0 14.8 4.0 6.1-28.1 14.2 14.0 2.7 1-<10 16.6 16.5 4.4 6.8-28.4 14.2 14.4 3.1 10-<30 14.2 13.5 2.5 10.0-17.4 13.2 14.1 3.5 2--30 17.4 18.5 4.7 11.1-21.6 14.4 16.0 4.3 F 1.81,P 0.15 F 0.31,P 0.82 TSH (mU/L) 0-<l 1.4 1.3 0.8 0.2-4.3 1.7 1,5 0.8 1-<10 1.4 1.2 0.7 0.5-3.1 1.7 1.5 0.9 10-<30 2.1 2.2 0.8 1.2-2.9 2.9 2.5 1.1 2:30 1.2 1.1 0.4 O@8-1.8 1.7 1.3 0.6 F 2.21,P 0.09 F 5.47,P 0.002 BMI,body mass index;17-HP,17-alphahydroxyprogesteronTeS;H, thyroid-stimulathionrgmone. Samples sizes:0-<l ppm: 1993,n 53;1995,n = 39. 1-<10 ppm: 1993,n = 46; 1995,n = 26. 10-<30 ppm: 1993,n = 8;1995,n = 10. a30 ppm: 1993,n = 4;1995;n = S. 4:Mean significandtilfyferen(tBonferronti-testp,< .05)thanthethreeotherPFOA ppm levels. 0 Mean levelsignificandtilfyferen(tBonteffotn-itestp,< .05)thanthe0-<l ppm and 1-<10 ppm PFOA categories. Range 9.3-20.5 5.6-19.4 7.0-18.0 7.3-18.0 0.6-4.0 0.5-3.7 1.9-5.8 1.1-2.5 TABLE 2 Mean, Median, Standard Error(SE)of Mean and Range of Estradiolby Body Mass Stratifiebdy Serum PFOA Leveland Year of Data Collection 1993 Data BMI (kg/mlby PFOA (ppm) n Mean Median SE Range n Estradio(lpg/mL) BMI<28 0-<l ppm 30 48.4 47.0 1.7 32.0-68.0 23 1-<10 30 55.0 55.0 2.2 34,0-81.0 13 10-<30 3 54.3 56.0 5.5 44.0-63.0 5 2t30 2 62.5 62.5 7.5 55.0-70.0 0 BMI z,28 0-1 ppm 23 63.0 66.0 2.9 32.0-83.0 16 1-<10 16 57.8 55.5 3.1 34.0-79.0 13 10-<30 5 55.0 60.0 6.1 39.0-70.0 5 at3O 2 63.0 63.0 7.0 56.0-70.0 5 Freetestosteron(eng/dL) Age <40 0-<l Ppm 20 17.3 16.8 0.9 10.5-28.1 18 1-<10 28 16.8 17.4 0.7 10.1-24.4 13 10-<30 4 15.2 14.9 1.0 13.4-17.4 2 @30 3 19.5 20.1 1.4 16.9-21.6 3 Age a-40 0-<l ppm 33 13.6 13.4 0.6 6.1-21.2 21 1-<10 18 16.2 15.8 1.3 6.8-28.4 13 10-<30 -z30 4 13.2 13.1 1.4 10.0-16.6 8 1 11.1 11.1 - 2 Index and Free Testosteroneby Age, 1995Data Mean Median SE Range 66.0 66.0 2.2 48.0-87.0 62.0 62.0 3.7 48.D-91.0 64.6 64.0 4.1 54.0-79.0 - - - 71.1 72.0 3.2 45.0-94.0 68.3 65.0 4.5 45.0-96.0 69.6 72.0 4.1 55.0-78.0 73.2 75.0 3.0 64.0-81.0 15.3 15.2 0.6 11.3-20.5 14.7 14,6 0.6 10.8-17.8 15.7 15.7 2.4 13.3-18.0 15.9 16.0 1.2 13.8-18.0 13.2 13.4 0.6 9.3-18.3 13.7 14.3 1.0 5.6-19.4 12.6 14.1 1.2 7.0-16.2 12.2 12.2 4.9 7.3-17.0 moderate drinkers,as was previously reported.2m Discussion We conducted two cross-sectional studiesof PFOA productionworkers to investigattehe relationbetween exposure and estradiolm,ean estraserum PFOA levelsand severalre- diollevelswere 10% greateramong productivehormones: in particular, employees with the highestserum estradioalnd testosteronAel.though PFOA levels(2:30ppm); however. we did not observe a significantly thiswas confounded by BMI, and positiveassociationbetween PFOA any interpretatioins limitedby the 620 EpidemiologiSctudyofPtrfluorooctanoAiccid-Olsonotal few subjectast thisPFOA level. vides u plausiblemechanism for a thosewith the highestserum PFOA Gillilanadlst)observedan approxi- possiblethresholdrelationshibpe- exposure levelsin 1993 alsopartici- iiiii1t0e% incre;tsien mean estradiol tween PFOA and estradiol."'@"'piiiedin 1995.This reducedsample levelsfrorihiislowest(0-1 ppni)to W hileresponsesto peroxisome pro- size resultedin lower statistical Iii,,he(s1t5-26 ppin)totalserum or- literatorlsi,kePFOA, are readilyob- power. Third, we specificallmyea- ,,@inifcluorinelevelsamong these production employees.2m Unlike the presentstudy,thispreviousreport alsoobserved a significanntonlinear positiveassociationbetween estra- served in ratsand mice. otherspecies-including humans-have shown no such responsesto many typesof peroxisome proliferatoarts equivalentdose levels-.1%-41 sured serum PFOA levels.Use of totalserum organicfluorinemay representotherperfluorocarbonsw,hich could be peroxisome proliferators, although data suggest that PFOA dioland totalserum organicfluo- We did not observeany significant would representthe greatestfraction rine.28Possiblereasonsfor the dif- associationbsetween PFOA and free ferentresultsincludethe following: or bound testosteroneH.owever, we of totalserum organicfluorinelevels in this employee populatio.n27.29 (1)use of differenmteasurementsof did observea significanptositiveas- Fourth,therecould be measurement exposure(totalserum organicfluo- sociationbetween 17-HP and PFOA errorin importantconfoundingvari- rinein 1990 and serum PFOA in in the 1995 but not 1993 analyses. ables.Analysis of the 68 subjects 1993 and 1995);(2)the possibility We examined 17-HP. a precursorof thatthe multivariatmeodel used in testosteroneb,ecause Cook et al'8 1990 may have transgressedthe ho- suggestedthatPFOA may affectthe who participatedin both years showed good correlatiofnortheconfounding factorsof BMI (r = .93, moscedasticityassumptionof regres- conversionof 17-HP to testosterone P = 0.0001) and the self-reported sionanalysi3s5 ; (3)possiblemisclas- viainhibitioonf 17,20-lyaseI.fthis aspectsof alcoholconsumption(r= sificatioonf confounding variables enzyme was inhibitedt,heexpected .67,P = 0.0001)and cigarettsemok- (co, the expected relationshibpetween BMI and estradiowlas not resultwould be an increasein 17-HP levels,which was observed in both ing (r = .84,P = 0.0001).Fifth,the qualityof medical surveillancdeata observed in 1990: correlatiocnoef- years'studiesa,lthoughsignificantly can be evaluatedby whetherknown ficien=t -.01);(4)differenstub- in only the 1995 study.Recent labo- associationsare observed.42 In this jectsanalyzed(94 employees partic- ratorywork suggeststhattheremay regard,we observed various ex- ipatedin both theoriginal1990 and be an accommodation by the CD rat pectedassociation(seg,estradioalnd the 1993 surveys,compared with 61 to the initiatlestosterone-loweringBMI, freetestosteronaend age).Fi- employees who participateidn 1990 and 1995);and (5)differenceisnthe effectof PFOA.19 A previousre- nally,thepulsatilneatureofsome of port2llobserved a significanntega- thehormones studied(eg,FSH, LH, estradiolassays. tive nonlinearassociationbetween testosteroneh)as resultedin prior Dose. thresholdeffecta,nd species totalserum fluorineand free or recommendations that mean hor- sensitivitmyay accountfor the ap- bound testosteroneT.his observation mone measurements should be the parentdifferencesbetween the ani- was dependent upon one influential resultof pooledblood from multiple mal and human studiesW.e did not datapoint,thatof an employee who samples takenat shortintervalS.4I3n observea significanatssociatiobne- had no detectabletotalserum organic our study,multiplesamples were not tween estradioland PFOA but did fluorinelevelbuthad thehi-hestfree feasiblebecause of the low probabil- observea 10% increaseatthehighest testosteronleevelmeasured. ityof employees voluntarilgyiving serum levelosf PFOA. Serum PFOA Several methodological issues threesamples over a 45- to 60- in theseworkers was likelybelow should be consideredin evaluatina minute time period. theobservableeffectlevelsinanimal the resultsfrom thisstudy.First.the In summary, we conducted two Studiest-h.e observableeffectlevelin cross-sectionadlesicyndoes not allow cross-sectionasltudiesin 1993 and the CD rat is somewhere above a fora directanalysiLsof the temporal- 1995 and did not observe a signifi- mean serum level of 55 ppm itvof an associationG.iven the long- cantlypositiveassociationbetween PFOA. "-25All but threePFOA mea- halflifeof PFOA. itisconceivable PFOA exposure and estradioolr a surements from employees in our thattheremay be some biolo2ical significantlnyegative association study were below 55 ppm PFOA. accommodation to the effect;of with testosteroneO.ur study may not The [Ocl'icncreasein mean estradiol PFOA. as su-cestedby Bic-elet have been sensitiveenoucyhto detect le@'elsobserved amonc, employees with the highestlevelsof serum PFOA (@30 ppm) couldsuggesta al.'9Second. the two cross-sectional whether an associationbetween analysescannot be viewed as inde- PFOA and estradioclould existin pendent populationsbecause 68 em- humans because measured serum thresholdresponse.The discoveryof ployeeswere studiedin both years. PFOA levelswere likelybelow the the conv,eri!encoef peroxisomalpro- Fewer employees participateidn se- observableeffectlevelssue ested in liferatorasnd estradioaltthelevelof rum measurements in the 1995 as- the animal studies.Our reLsultpsro- theirnuclearhon-nonereceptorspro- sessment.althoulhthe majorityof vide reasonableassurancethatsim-- JOEM - Volume 40, Number 7,July 1998 621 nificanthormonal changes among these male production employees 14. HanhijarviH. Ylinen @M,Kojo A. Kosma 27. Ubel FA. Sorenson SD. Roach DE. VM. Elimiiiatia(nidnttixicioif%plerfluo- Healthstatusof plantworkersexposedto were not apparentin relationto their measured serum PFOA le%-els. Acknowledgments The authorsgratefullaycknowledge the contributioonfs thefollowingindividualtso thisstudy:FrancesCurtis.Mary Fowler. Mary Hansen.and Drs James Johnson.Roger PerkinsJ.ames Wolter.and Larry Zobel. References I. Bryce H. Industrialnd utilitariaasnpects of fluorinechemistry.In:Simons J, ed. Fluorine CheiytistnN-e.w York: Academic Press-1.964:297-492. 2. NicklesM. Presencedu fluordans ]a sang.ConiptRend.1856;43:885. 3. Taves D. Evidencethatthereare two formsoffluoridienhuman serum.Nature.1968@217:1050-1051. 4. Taves D. Electrophoretmiocbilityof serum fluorideN.ature.1968;220:582-583. 5. Guy W. Fluorocompounds of Human Plasma:AnalilsiPsr,evalenceP,iirification.and Char"acteri,-ati[oDno.ctoraldisseriationR.o]chesterN,Y: Universitoyf Rochester:1972. 6. GriffitFhD, Long JE. Animal toxicity studieswith ammonium perfluorooctanoate.Am Ind l@vgAssoc J. 1980;41:576583. 7. Kennedy G. Dertnaltoxicitoyf ammonium perfluorooctanoatTeo.xicolAppl rooctanoiaccidduringsubchronicadmin- fluorochemicalsa. preliminaryreport. i.-,tratiinotnhe \k'istarrat.Plitiri?ia(-ol Ain IndHi@,Aiszst)J(.- 1980:41:58-4589. To-vi(-1i9)8l7.:61:66-68. -18. Gilliland FD. Fljioro(-arb(piiasnd Hutrian 15. Ikeda T. Aiba K. Fukuda K. Tanaka M. Health: Sriediesin att 0(-c-itpatioriaClo- The induction (ifperoxisome prolifera- hort. [Doctoral dissertation.]Minneapo- tionin ratliverby perfluorinatfeadtty lis.MN: UniversityofMinnesota:1992. acids.metabolicalliynertderivativeosf 29. Pothapragada V. Sodium biphenyl fattyacids.J Biochein.1985.98:475- method fordeterminatioonf covalently 482. bound fluorinienorganiccompounds and 16. PastoorTP. Lee KP, PerriMA, Gillies biologicamlaterialsA.nal Clietn1.982. PJ.Biochemicaland morphologicalstud- 54:1132-1137. ies of ammonium perfluorooctanoate- 30. GillilandFD, Mandel JS. Mortality induced hepatomegalyand peroxisome among employeesof a perfluorooctanoic proliferatioEnr.p Mol Pathol.1987:47: acid productionplant.J OceiipMed. 98-109. 1993;35:950-954. 17. Sibinski U. Th-o-Year Oral (Diet) Tax- 31. Simons J.BryceT. Electrochemicaflu- iciir-ICarcinogiecnirvStud.%.of Fluoro- orination.In: Simons J. ed. Fluorine che'micalFC-143 inRats.St.Paul.MN: Chemistn-N.ew York:Academic Press; RikerLaboratories1;987. 1954:340-377. 18. Cook JC, Murray SM. Frame SR, Hurtt 32. Johnson ID, Wolter JT, ColaizyGE. ME. Inductioonf Leydigcelladenomas RcthwillPA. NelsonRM. Quantification by ammonium perfluorooctanoataep:os- of perflurooctanoataend perfluo- sibleendocrine-relatmeedchanism. Toxi- colAppi Pharpnacol.192-.113:209-217. rooctanesulfanaitne human serum using 19. BiegelLB. Liu RCM, HurttME, Cook ion-paierxtractioannd highperformance JC. Effectsof ammonium perfluo- liquid chromatography-thermospray rooctanoatoen Leydig cellfunctioni:n mass spectrometrywith automatedsam- i,itrion,vivo,andex vivostudiesT.oxicol plepreparation[.3M EnvironmentalLab- Appi PharmacoL 1995,134:18-25. oratoryReport.]St.Paul,MN: 3M Com- 20. Liu RCM. HurttME, Cook JC, Biegel pany. 1996. LB. Effectoftheperoxisomeproliferator, 33. SAS InstituteI,nc. SAS Users Guide: ammonium perfluorooctanoa(tCe8).on StatisticsV.ersion 6. Cary, NC: SAS hepaticaromataseactivitiynadultmale InstitutIen,c.;1990. CRI:CD BR (CD) rats.Fundam Appl 34. GriffinJE,Wilson JD. Disordersof the Toxicol.1996;30:220-228. testesI.n:lsselbackeKrI, BraunwaldE. Phannacol.1985:81:348-355. 21. Lemberger T, DeSVCTgne B, Wahli W. Wilson JD, MartiniJB.FauciAS, Karper 8. Kennedy G, HallG, BrittelJl.iChen H.. Peroxisome proliferator-activraetceedp- DL, eds.Harrison'sPrincipleosf Inier- Inhalatiotnoxicitoyfammonium perfluo- tors:a nuclearreceptorsignalinpgathway nal Medicine.New York: McGraw-Hill; rooctanoateF.ood Chem Toxicol.1986; in lipidphysiology.Ann Rev Cell Dev 1994:2006-2017. 24:1325-1329. Biol.1996,12:335-363. 35. Montgomery DC, Peck EA. Introduction 9. Ophaug R, SingerL.Metabolichandling 22. Cook JC, HurttME. Frame SR, Bieg'el to Linear RegressionAnal.vsisN.ew of perfluorooctanoaicid in rats.Proc LB. Mechanisms of extrahepatitcumor York: John Wiley and Sons.Inc.;1992: Soc F-rpBiolMed. 1980;163:19-23. inductiobny peroxisomeproliferatoirns 67-117. 10. YlinenM. Koho A, HanhijarviH. Peura Crl:CDBR (CD) ratsT.oxicologis1t9.94; 36. KellerH. DreyerC. Medin J.Mahfoudi P. Dispositioonf perfluorooctanoaicid 14:301(abstract). A. Ozato K. Wahli W. Fattyacidsand in the rataftersingleand subchronic 23. Ogboum JD. Frame SR, EllioGtS, Cook retinoidscontrol lipidmetabolism administratioBnu.ll Ent,iro;Ciontam JC. Pancreaticoncogeniceffectsof throughactivatioonf peroxisomeprolif- To.ricol1.990;44:46-53. 11. Vanden Heuvel1.KuslikisB.Van Refel- Wyeth 14.(>43T.oric-ologi1s9t9.7.36:232 (abstract). erator-activatreedccptor-retinXoidreceptorheterodimersP.rocNatiAcad Sci. ghem M. PetersonR. Tissuedistribution, 24. Rao MS. Reddy JK. Hepat(>carcinogen- 1993-.90:2160-2164. metabolismand eliminatioonf perfluo- esisof peroxisomeproliferatoArnsn.NY 37. HunterJ.Kassam A. Winrow CJ,Rachu- rooctanoicacid. J Biochem Toricol. Acad Sci.1996;804:573-587. binskiRA. Capone IP.Crosstalbketween 1991@6:83-92. 25. Palazzolo MI. 13-Week Dietan- Toxiciti- 12. KuslikisBl.Vanden HeuvelJP.Peterson Stud.ii-t-itTh-5180,Amnionium Perfluo- RE. Lack of evidenceforperfluorode- rooc-iant)a(tCeAS No. 3825-26-1),in cant)vl-or pertluoroocianoyl-coenzvme Male Rats.Madison.WI: Hazelit)Lnab- A f(;rmatioinn male and female'rats. oratories1:991. J Bii)h4i,ipTt;.vi(-1o9l9.2:7:25-29. 26. PerkinsRG. ln%estigilito)nf'ammc)nium 13. Hanhijar%i H. Phaug R. Singer L. The periul(irt)c)ctante)la'tle'eocnl hormone ,@ex-relatLeidit*ferenicneperfluoro4)ctano- levelsand rwrtixi,,otpirita)lliferatiinon ateevcretiti)ntherat.Pro(-.StEpTtl-)Biol the rai. 1992;12:38(ab- Mril.1982:171:.51-55. stract). the thyroidhormone and peroxisome proliferator-activarieecdeptorisnregulating pcroxisome proliferator-responsive genes.Mt)l Cel Endt)t-rino1l9.96.116: 213-221. 38. StringerDA. HeptititP-erorisi)tnPerol@fei-aii(jiimro.nc)graphNo. 17.1Brussels: European Centrefor Ecott)xicologaynd Toxicologyof Chemicals(ECETOC); 1992. 622 EpidemiologSitcudyofPerfluorooctanAociicdoOlsonstal @19D.e lalgiesiFaA. McGuire ED, Haskins JR. LalwaniND. Structuradliversitoyf peroxisomeproliferatoarnsd theireffects on mammalian livecrellsinvivo.AntiNY At-adSt.-1i9.96.804:310-327. 40. Tugwood JD. AldridgeTC. Lambe KG, MacDonald N. Woodyatt NJ.Peroxisome proliferator-activarteecdeptors:structuresand functionA.nn iVY Acad Sci. 1997:804:252-265. 41, Tugwood JD. Lambe KG, Woodyett NJ. lnteri-ndividuadlifferenceisnhuman peroxisomeproliferator-activraetceedptoralphastructuraend functionT.oxicologist.1997:36:236(abstract). 42. 01,,enGW, Kusch GD, StaffordBA. Gudmundsen SL. CurrierMF. The positiveknown associatiodnesign:a quality assurancemethod foroccupationalhealth surveillandcaetaJ. Occup Med. 1991-.33: 998-1000. 43. GoldzieherJW, DozierTS, Smith KD. SteinbergeEr. Improvingthediagnostic reliabiliotfyrapidlyfluctuatinpglasma hormone levelsby optimizedmultiplesamplingtechniquesJ. Clin Endocrinol Herob. 1976.43:824-830.