Document gOQg7QbKwb6mMoYQExkG4mjQ

CorporatOeccupationMaeldicine 3M CenterB,uildin2g20-3W-05 St.PaulM,N 55144-1000 65''1,34'2130Telephone 6517339066Fax 5. Fluorochemicals and Human Health: Studies in an Occupational Cohort This isa doctoralthesisby Dr. Frank Gilliland(graduatestudentatthe time in the Division of Environmental and Occupational Health at the University of Minnesota). The dissertationconsistsof two studies:1) an analysisof the 1990 fluorochemical medical surveillancedata of employees who voluntarilyparticipatedin the program at the Chemolite (Cottage Grove, Minnesota) manufacturing plant;and 2) the second update of the retrospectivecohort mortalitystudy atthe Chemolite (Cottage Grove, Minnesota) site. In 1990, the Cottage Grove fluorochemical medical surveillanceprogram analyzed for serum totalorganic fluorine.This was considered a proxy for serum perfluorooctanoate (PFOA) levels.Perfluorooctanoateisthe anion of perfluorooctanoicacid.Gilliland reported no significantclinicalhepatictoxicityassociatedwith PFOA levels.However, the effectof obesity on livertransaminase levelsdecreased as PFOA increased. Gilliland also reported thatthe effectof alcohol on HDL was reduced as serum PFOA levels increased. Serum PFOA (i.e.s,erum totalorganic fluorine)was positivelyassociated with estradioland negatively associatedwith freetestosterone.The negative association between free testosteroneand PFOA was stronger in older men. Thyroid Stimulating Hormone (TSH) was positivelyassociatedwith PFOA. Gillilandconcluded thatthese resultssuggested thatPFOA may affectmale reproductive hormones and thatthe liveris not a significanstiteof toxicitiynhumans atthe PFOA levelsobserved inthiscrosssectionaalnalysis.Note: These observationsreportedby Gillilanidn hisdissertation have been examined inthreesubsequent(biennialm)edicalsurveillanceexaminationsof thisworkforce. Olsen eta]examined 1993 and 1995 fluorochemicalmedical surveillance datawhich specificallayssayed forperfluorooctanoatienthe serum using mass spectrometrymethods. The dissertatiofnindingscould notbe replicated.PFOA was notsi9nificantlpyositivelayssociatedwith estradioalnd TSH, norwas itne4g:a1tively associatedwith free testosterone(see study # 6). Olsen etal examined the 1993, 1995 and 1997 fluorochemicalmedical surveillancdeata,measuring specificallfyor perfluorooctanoatea,nd confirmedthelackof clinicahlepatictoxicitiynthisworkforce forthe serum PFOA levelsmeasured (seestudy#8). They were unabletoobserve,as initiallryeportedby Gillilanda,n associatioonf PFOA tomodulate theeffectof obesity on livertransaminaseclinicaclhemistrytestsor blunttheeffectthatPFOA may have on theeffectof alcoholuse with HDL. Regarding the second updateto theretrospectivceohortmortalitystudy,a paper was publishedin 1993 (seestudy# 4 ) which describedtheresultsfrom theGillilandoctoral thesis. FLUOROCARBONS AND HUMAN HEALTH: STUDIES IN AN OCCUPATIONAL COHORT SUBMITTED A THESIS TO THE FACULTY OF THE GRADUATE OF THE UNIVERSITY OF MINNESOTA BY FRANK DAVIS GILLILAND SCHOOL IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY/ENVIRONMENTAL HEALTH OCTOBER, 1992 UNIVERS rTY OF MINNESOTA ThisistocertiftyhatIhave examined thisbound copy ofa doctoralthesisby FRANK DAVIS GILLILAND and have found thatitiscompleteand satisfactoriynallrespects. and thatany and allrevisionsrequiredby the final examining committee have been made. JACK S. MANDEL PhD. Name ofFacultyAdviser Signatureof FacultyAdvisor Date GRADUATE SCHOOL ACKNOWLEDGEMENTS Iam indebtedto Dr.Jack Mandel whose competent research and career advicewere invaluable.Not onlydidDr Mandel guide me to thisr.esearch' projecth;e directedme tothe NIOSH occupationalmedicine fellowshipthat has enrichedmy clinicamledicine knowledge and supported my research effortosver the laslthreeyears. His generositywfth histime and patienceare deeply appreciated.Dr.Timothy Church. Dr.WilliamToscano, Dr. Ian Greaves, and Dr.Thomas Sellersserved on my commfttee. They deserve a specialthanks fortheirefforts. Iwish tothank the Divisionof Environmentaland OccupationalHealth atthe Universityof Minnesota and the OccupationalMedicine Sectionat St Paul Ramsey Medical Center forthe superb trainingopportunitiesIhave had over the pastthreeyears. Dr.WilliamLohman, Dr.Samuel Hall.and Paula Geiger at St.Paul Ramsey Medical Center providedmuch appreciatedsupportduring the arduous taskof residencyand doctorajtraining.Several members of the Divisionof Environmentaland Occupationalhealthstaffwere instrumentalin the successfulcompletionof thisresearchefforl Sarah Wolgamot and Maraiyn Zappla providedexcellentadministrativseupporl Gavin Watt,Mindy Goisser,Richard Hoffbacica,nd othermembers of Colon Cancer Contmi Study providedoutstandingcomputer and statisticsaulpporl Dr.LarryZobal and Dr.JeffreyMandel of the 3M Corporation'sMedical Department providedadviceand supporl Theirhelp was an essartfial element inthe su=ess of thisprojecl Stan Sorenson, Dr. Roger Perkins,and other3M Medical Department members shared theirinvaluableexperience and knowiedge. Iwould alsoliketo acknovaedge the support ofthe Dow Chemical Corporationover the lasttwo years of my training. Last,but not least,thiswork could nothave been accomplished wfthoutthe lovingsupportof Susan, my Wife. Her understanding and excellenteditorial comments are greatlyappreciated. ABSTRACT Perflucrooctanoiacid(PFOA) has been reportedtobe a nongenctoxic hepatocarcinogenand reproductivheormonal toxininrats.Although PFOA is the M2jor component oftotalfluorinienhumans, littlienformatioinsavailable concerninghuman toxicitieTsh.e healtheffectsof PFOA were assessed intwo studiesconductedinoccupationalleyxposed workers. The associations between PFOA and repmducbve hormones, hepaticenzymes, lipoprotains, hematologyparameters.and loukocytocountswere studiedin115 ma)e employees. Serum PFOA was positivelayssociatedwithastradioalnd negativelayssociatedwithfreetestostemne(TF)butwas notsignificantly associatedwithfuteinizihnogrmone. The negativeassociationbetween TF and PFOA was strongerinoldermen. Thyroidstimulatinhgormone and PFOA were positivelayssociateclPFOA and prolactiwnere positivelayssociatedin moderate drinkers.The effectofadipositoyn serum glutamyiox2Joaceticand glutamylpyruvictransaminasedecreased as PFOA increased.The inductionof gamma glutamyitransferassby alcoholwas decreased as PFOA increased. The offectofalcoholon HDL was reduced as PFOA increaseclA positive associatiobnetween fiemoglobinm,ean cellulavrolume, and leukocytecounts withPFOA was observed. These resultsuggestthatPFOA affectsmaje reproductivheormones and thatthe liverisnota significansthe oftoxicitiyn humans atthe PFOA levelsobserved inthisstudy. However, PFOA appears to modifyhepaticand immune responsestoxenobiotics.A retrospectivceohort mortalitsytudyof2788 majo and 749 femalesworkers employed between 1947-1984 ata PFOA productionplantwas conductecl Overall,therewere no significantilnycreasedcause specifiScMRS. Among men, ton years of employment inPFOA productionwas associatedwitha significantthreefold increaseinprostatecancer mortalitcyompared tono employment inproducfion. Given the smallnumber ofpmsiate cancer deaths and the naturalhistoryofthe disease,theassociatiobnetween produdon work and prostatecancer must be viewed as hypothesisgeneratingand shouldnotbe over interpretedf.fthe pmstate cancer mortafiteyxcess isrelatedtoPFOA, the resultsofthe two studiessuggestthatPFOA may increasepmstate cancer mortalittyhrough endocrinealterations. TASLE@OF coNTEN-TS 1.INTRODUCTION .................................................1.......... 2. REVIEW OF THE LRMRATURE ...................................4........... 2.1 Introductio.n.................................................4........... 2 OrganicFluo=hemicals ...................................4.......... 2.3 PhysicaJProperties ...........................................6........... 2.4 Synthesis....................................................7........... 2.5 Sources Of Organic Fluo6do Exposure ..........................a........... 2.6 Toxicokinaticsof PFOA ...........................................1.1........... 2.7 Toxicodynamics of PFOA .....................................1.6........... 2.7.1Male Reproductive Toxicitie.s.........................1.6........... 2.7.2FemaJo Reproduc*e Toxicities.......................2.0........... 2.7.3Thyroid Toxicftie.s...................................2.0........... 2.7.4Hepatic Toxicities....................................2.1........... .2.7.5Nongenotoxic Carcinogenesis .........................2.3........... 2.7.6lmmunctoxicity..........................................2.3........... 2.7.7Mechanisms of Action ................................2.4........... 2.8 Occupational FluorineExposures At Chemolits ..................2.6.......... 2.9 EpidemiologicalStudies ......................................2.7........... 2.10 Summary ...................................................28........... 3. METHODS .........................................................2.9........... 3.1 Introduction..............................................................2.9........... 3.2 Retmspective Cohort MortW4 Study ...........................3.0........... 3.2.1DefinitionOf The Cohort ..............................3.0........... 3.2.2Study Databases And Files...........................3.1............ 3.2.3Data EcHng ...........................................................3.1....... 3.2-4ValidationOf The HistoricaClohort Information.........3.2..... 3.2.4.1Assessment Of Completeness Of Aseertainment .................................................3..2..... 3.2.4.2ValidationOf Cohort Information...............3.3........ 3.ZS VitalStatus Aseertainment ..............................3.3... 3.2.6Validationof VitalStatus Ascertainment ................3.4.......... 3Z7 Analysis...............................................................3..4..... 3.3 Cross SectionalStudy Of PFOA Exposed Workers ..............3.6........ 3.3.1PopulationDefinitionAnd clacruftment..................3.6.......... 3.3.2Data Collection......................................3.7........... 3.3.2.1Study Logs And Files.........................3.7........... 3.3.2.2Questionnaire ................................3.7........... 3.3.2.3Laboratory Procedures .......................3.7........... 3.3.2.3.1Height and Weight ...........................3.7... 3.3.2.3.2Blood ................................3.8........... 3.3.2.3.2.1Drawing And Handling ........3.8... 3.3.2.3.2.2Assays .....................3.8........... 3.3.2.3.2.3Quality Assurance ............4.0...... 3.3.3AnaJysis ........................................................4..0.... 4. RESULTS ............................................................4.3.......... 4.1Cmss SectionaPlerfluoromrbonPhysiologicEffectsStudy ......4.3... 4.1.1ParticipanCtharacteristic.s..........................4.3........... 4.1.2TotaiSerum Fluohne 44 4.1.3Hormone Assays .............................*.'*.'.*.".*.4".5*.......... 4.1.4Hormone Ratios ....................................4.9.......... 4.1.5CholesterolL,ow Density Upoprotein,High Density Upopmtein, And TMglycerides.............................5.1.......... 4.1.6HepaticParameters .................................5.2.......... 4.1.7Hematology Parameters .............................5.4.......... 4.1.8Summary Of Results................................5.6.......... 4.2The 1990 Chemolhe Retmspective Cohort MortalityStudy .......5.8... 4.2.1StandardizedMortalitRyatios(SMRS) .................5.9.......... 4.2.1.1SMRs For Women ..........................5.9........... 4.2-1.2SMRs For Men .............................5.9.......... 4.2.2StandardizedRate Ratios(SRRS) .....................6.0.......... 4.2.3Mantel-RelativeRisks (RRMH) .......................61........... 4.2.4ProportionalHazard Regression Model Relative Risk Estimates..........................................6.1........... 4.2.4.1ProportionalH2zard Models For Male Workers ..........................................6.1.......... 4.2.4.2ProportionaJHazard Models For Female Workers ..........................................6.3.......... 4.3 PhysiologicEffectsTables ...................................6.4.......... 4.4 MortalitTyables ........:......................................1.5.7.......... 4.5 Figures..................................................1.9.0........... 5. DISCUSSION .............................................................1..9..8............. 5.1 PhysiologicEffectsStudy ...................................1.9.8.......... 5.1.1Introductio.n......................................1.9.8........... 5.1.2Hormones ........................................1.9.8........... 5.1.3CholesterolT,riglyceridesa,nd Lipoproteins...........2.0.2...... 5.1.4Hepatc Parameters. 203 5.1.5Hematology Counts 5.1.6TotalFluorine.....................................2.0.9........... 5.1.7MethodologicalConsiderations............................2.1.0........ 5.1.7.1SelectionSias .............................2.1.0........... 5.1.7.2Information Sias ..............................2.1.1............ 5.1.7.3Confounding Bias ............................2.1.4............ 5.1.7.4Analoc Model SpecificationBias ..............2.1.6...... 5.2 1990 Chemolhe MortalityStudy ................................2.1.7............ S.Zl Introduction...........................................2.1.7............ 5.2.2 ParticipantCharacteristics.............................2.1.7............ 5.Z3 MortalityResults .......................................2.1.8........... 5.2.4 Methodological Considerations ........................2.2.0............ 5.2.4.1 Information Bias ..............................2.2.0............ 5.2.4.2 Confouncring and Selection Bias ...............22.1....... 5.2.4.4 Analytic Model SpecificationSias ..............2.2.3..... 6. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS .............2.2.5...... 6.1 Cmss-Sectional Study ofthe Physiologic Effectsof PFOA .......2.2.5.. 6.2 RetmspectiveCohortMortWityStudy Of The Chemolite Workforce,1947-1990 .........................................2.-.2.6........... REFERENCES ...................................................2.3.0........... APPENDIX 1 ....................................................2.5.5............ APPENDIX 2 ......................................................2.5.9............ APPENDIX 3 .....................................................2.81............ LIST OF TABLES Table 4.1.1Age DislHbutionInFiveYear Age GMUPS ....................6.4.......... Table4.1.2Disiribuii0o1n AlcoholAnd Toba= Use ....................6.5........... Table4.1.3The JointDistributiOofn Toba= And AlcoholUse ............6.6....... Table4.1.4DistributioOnf Age By Smoking And DrinkingStatus...........6.7....... Table4.1.5Pearson CorrelatioCnoefficientBsetween TotalSerum FluorineA,ge, Body Mass Index(Bmi).......................6.8.......... Table4.1.6Body Mass Index Distributi.o.n...........................6.9........... Table4.1.7Body Mass Index By Smoking And DrinkingStatus............7.0........ Table4.1.8The DistributiOofn Age,AlcoholAnd Toba= Use By Body Mass Index........................................7.1.......... Table 4.1.9TotalSerum FluorideDistributi.o.n........................7.2........... Table4.1.10TotalSerum FluorideBy Body Mass Index.Age. Smoking And DrinkingStatus.............................7.3....... Table 4.1.11Age DistributioBny TotalSerum FluorineCategory...........7.4...... Table4.1.12DistributiOofn Toba= Use By TotalSerum Fluoride Category..............................................7.5.......... Table4.1.13DistribLrtiOofnAlcoholUse By TotalSerum Fluoride Category..............................................7.6.......... Table4.1.14Body Mass Index DistributiBoyn TotalSerum F-luorine Category..............................................7.7........... Table4.1.15CoefficienOtf VariationFor Seven Hormono Assays..........7.8..... Table4.1.16The Observed Versus Expected Number Of Workers Wtth Hormone Assays OutsideThe Assay Reference Range ................................................7.9........... Table4.1.17Pearson CorrelatioCnoefficientBsetween Serum Hormones .............................................80............ Table4.1.18Pearson CorrelationCoefficientBsetween TotalSerum FluorideA,ge, Body Mass Index(Bmi),DailyAlcohol Use, DailyToba= Consumpton, And Serum Hormones .....................................................81............ Table4.1.19Bound Testosterone(Tb)By Body Mass Index,Age, Smoking, DrinkingStatusAnd TotalSerum Fluoride..........8.2..... Table4.1.20LinearMufttvariatReegressionModel Of Factors PredictinTghe Bound Testostemne (Ng/Di)Among 112 Male Workers..........................................8.3.......... Table 4.1.21Free Testosterons(T@ By Body Mass Index,Age, Smoking And DrinkingStatusAnd TotalSerum Fluorid.e ..............................................8.4.......... Table 4.1.22LinearMuttivariatReegressionModel Of Factors PredictinTghe Free Testostemne Value (Ng/Dl).............a.s........ Table4.1.23ParticipanEtstracrioBly Body Mass Index,Age, Smoking DrinkingStatusAnd TotalSerum Fluorid.e.................8.6........... Table 4.1.24LinearMultivahateRegressionModel Of Factom PredictinTghe EstradioVlalue(Pg/Di)Among 113 Male Workers .................................................................8.7........... _iv Table 4.1.25L.LiteniziHnogrmone (Lh)By Body Mass Index,Age, Smoking And DrinkingStatus,And Total Serum Fluorine................................................e.s......... Table 4.1.26Unear MuftivariatReegression Model $tlOf Factors PredictingThe LutenizingHormone" Value (Mu/Ml) Among 113 Male Workers .................................8.9......... Table 4.1.27FolliclSetimulatinHgormone (Fsh) By Body Mass Index, Age, Smoking And DrinkingStatus.And TotalSerum Fluorine................................................9.0......... Table 4.1.28Unear MuttivariatReegression Model Of Factors PredictingThe FolliclSetimulatingHormone Value (Mu/Mi)Among 113 Male Workers ..........................91.......... Table 4.1.29ThyroidStimulatingHormone (Tsh) By Body Mass Index,Age, Smoking And DrinkingStatus,And Total Serum Fluorine..................................................9.2.......... Table 4.1.30Unear MuftivariatReegression Model Of Factors PredictingThe ThyroidStimulatingHormone* Value (Mu/Ml)Among 113 Male Workers .........................9.3.......... Table 4.1.31ProlacbnBy Body Mass Index.Age, Smoking. Drinking Status,And TotalSerum Fluorine..........................9.4......... Table 4.1.32Unear MuftivariatReegression Model Of Factors PredictinTghe PmlactinValue (Ng/M) Among 113 Male Workers ................................................9.5......... Table 4.1.33Pearso.nCorrelationCoefficientBsetween Hormone Ratios.AndTotalFluorideA,ge, Body Mass Index, AlcoholAnd Tobacco Consumption ........................9.6.......... Table 4.1.34Pearson CorrelationCoefficientBsetween Pmlacdn Hormone RatiosAnd TotalFluoride,Age, Body Mass Index,AlcoholAnd Tobacco Consurnption..................9.7.......... Table 4.1.35Pearson CorrelationCoefficientBsetween Thyroid StimulatingHormone RatiosAnd Total Fluoride,Age, Body Mass Index,AlcoholAnd Toba= Consumption ........9.8.... Table 4.1.36Pearson CorrelationCoefficientBsetween Follicle StimulatingHormone RatiosAnd Total Fluoride,Age. Body Mass Index,AlcoholAnd,Tobacco Consumption ........9.8.... Table 4.1.37Pearson CorrelationCoefficientBsetween Phuilary Glycopmtion Hormone RatiosAnd Total F:IuorideA,ge, Body Mass Index,AlcoholAnd Toba= Consumption ........9.9.... Table 4.1.38Unear MultivadateRegression Modell Of Factors PredictingThe Bound-Free Testosterone Ratio Among 112 Male Workem ..........................................1.0.0......... Table 4.1.39Unear MuftivariatReegression Model2 Of Factors PredictingThe Bound-Free Testosterone Ratio Among 112 Male Workers .......................................i.o.i......... Table 4.1.40Unear MufttvariatReegression Model 01 Factors PredictingThe Estradiol-BoundTestostemne Ratio Among 112 Male Workers ................................1.0.2......... v Table 4.1.41 Unear MuftivahateRegression Model Of Factors PredictingThe Estradiol-FreTeestostemne Ratio Among 112 Male Workers .................................1.0.3............ Table 4.1.42 Unear MuftivariatReegression Model Of Factors PredictingThe Estradiol-Lh+Ratio Among 112 Male Workers .................................................1.0.4........... Table 4.1.43 Une2t MuftivahatRegression Model Of Factors PredictingThe Bound Testosterone-Lh+ Ratio Among 112 Male Workers ........................................1.0.5............ Table 4.1.44 Une2r MultivaHatReegression Model Of Factors PredictingThe Free Testostemne-Lh+ Ratio Among 112 Male Workers ............................................1.0.6........... Table 4.1.45 Unear MuftivadateRegression Model Of Factors PredictingThe Bound Test=omne-Prolactin Ratio Among 111 Male Workers ..................................1.07.......... Table 4.1.46 Unear MultivariatReegression Model Of Factors PredictingThe Free Testosterone-Pmlactin Ratio Among 111 Male Workers ..................................l.o.8........... Table 4.1.47 Unear MultivariatReegression Model Of Factors PredictingThe Estradiol-PmlactinRatio Among 111 MaJe Workers . ...........................................l.o.g........... Table 4.1.48 Unear Multiva@ateRegression Model Of Factors PredictingThe Pmlactin-Fsh@ Ratio Among 111 MWe Workers .................................................i.i..o........... Table 4.1.49 Unear MultivariatReegression Model Of Factors PredictingThe Pmtactin-Lh"* Ratio Among 111 Maie Workers .................................................................. Table 4.1.50 Unear MuftivariatReegression Model Of Factors PredictingThe Pmlactin-Tsh+ Ratio Among 111 Ma)e Workers .........................................................1.1.2...... Table 4.1.51.Unear MuftivariateRegression Model Of Factors PrecrictinTghe Bound Testostemne-Tsh+ Rao Among 112 Male Wo*ers ...................................................1.1.3............ Table 4.1.52 Unear MuttivariatReegression Model Of Factors PredictingThe Free Testostemne-Tsh+ Ratio Among 112 MaJo Workers ........................................1.1.4............ Table 4.1.53 Unear Mul@varialeFlegressionModel Of Factors PredictingThe Estradiol-Tsh+Ratio Among 112 Male Workers .................................................1.1.5........... Table 4.1.54 Unear MuftivariatReegression Model Of Factors PredictingThe Bound Testosterone-Fsh+ Ratio Among 112 Male Workers .....................................................1.1.6......... Table 4.1.55 Unear MuftivariatReegression Model Of Factors PredictingThe Free Testostemne-Fsh+ Ratio Among 112 Male Workers ........................................1.1.7........... Table 4.1.56 Unear MultivatiatReegression Model Of Factors PredictingThe Estradiol-Fsh+Ratio Among 112 Male Workers ............................................................ vi Table 4.1.57LinearMuftivariaRtegressionModel Of Fa=rs PredictinTghe Bound Tsh-Fsh+ RatioAmong 112 MaJo Workers.............................................1.1.9........... Table 4.1.58LinearMuftivariatRegressionModel Of Factors PredictinTghe Tsh-Lh+ RatioAmong 112 Male Workers..............................................12.0........... Table 4.1.59LinearMultivahateRegressionModel Of Factors PredictinTghe Bound Lh-Fsh+ RatioAmong 112 Male Workers..............................................12.1........... Table 4.1.60Pearson CorrelatioCnoefficientBsetween TotalSerum FiuorideA,ge, Body Mass Index(Bmi),DailyAJcohol Use, DailyTobacco Consumption,And Upopmteins .........1.2.2.... Table 4.1.61LinearMultivadatReegressionModel Of Factors PredictinTghe CholesierofAmong 111 Male Workers........1.2.3.. Table 4.1.62LinearMUltiV2ri2tRGegression Model Of Factors PredictinTghe Low DensityUpoprotionAmong 111 MaJe Workers .........................................1.2.4.......... Table 4.1.63LinearMuftivariatRegressionModel Of Factors PredictinTghe HighDensityUpopmten (Hdl)Among 111 Male Workers .....................................1.2.5........... Table 4.1.64LinearMuftivari2tRegressionModel Of Factors PredictinTghe T@glyceddes Among 111 Male Workers.......1.2.6. Table 4.1.65Peamon CorrelatioCnoefficientBsetween TotalSerum F-luoridAeg,e, Body Mass Index(Bmi),DailyAJcohol Use, DailyToba= Consumption,And Hepatic Parameters...........................................1.2.7.......... Table 4.1.66Pearson CorrelatioCnoefficientBsetween Hepatic Enzymes, Serum Hormones, And Upoproteins.............1.2.8........ Table4.1.67Pearson CorrelatioCnoefficientBsetween Hepatic Parameters .........................................................1.2.9........... Table 4.1.68Serum GlutamicOxaloaceticTransaminase (Sgot GlutamicPyruvicTranS2minass (Sgpt),Gamma GlutamylTransferass(Ggt),And AlkafinePhosphatase (Akph)By TotalSerum Fluorine..........................1.3.0.......... Table4.1.69Serum GlutamicOxaloaceticTransaminass (Sgot)By Body Mass Index,Age, Smoking And DrinkingStatus........1.3.1... Table4.1.70Serum GiLdarnicPyruvicTransaminass (Sgpt)By Body Mass Index,Age, Smoking And DrinkingStatus.............13.2........ Table4.1.71Gamma GlLnamylTransferase(Ggt)By Body Mass Index,Age. Smoking And DrinkingStatus..................1.3.3.......... Table4.1.72A)kafinePhosphatase (Akph)By Body Mass Index,Age, Smoking And DrinkingStatus............................1.3.4........... Table4.1.73aLinearMultiva@ateRegression Model I Of Factors PredictinTghe Serum GlutamicOxaloacetic Transaminase (Sgot)Among 111 Male Workers............1.3.5..... Table 4.1.73bLinearMultiva6ateRegression Model 2 Of Factors PredictinTghe Serum GlutamicOxaJoacetic Transaminass (Sgot)Among 111 MaJo Workers............1.3.6..... vii Table 4.1.73cUnear Muftiva6ateRegressionModel 3 Of Factors PredictinTghe Serum Gl=mic OxaJoacetic Transaminase (Sgot)Among 11i M2Je Workers...........1.3.7...... Table 4.1.74aUnear MultivaHateRegressionModel 1 Of Factors PredictinTghe Serum GlutamicPyrvvicTransaminass (Sgpt)Among 111 Male Workers.................:.............1..3..8....... Table 4.1.74bUnear Muftiva@ateRegressionModel 2 Of Factors PredictinTghe Serum GlutamicPyrLrviTcransaminass (Sgpt)Among 111 Male Workers........................1.3.9......... Table 4.1.74cLinearMuftivadateRegression Model 3 Of Factors PredictinTghe Serum GlutamicPyruvicTransaminass (Sgpt)Among III Male Workers.........................1.4.0......... Table 4.1.75aLinearMuftivariatReegression Model 1 Of Factors PredictinTghe Gamma GiLitamyTlmnsfierass(Ggt) Among 111 Male Workers..............................1.4.1......... Table 4.1.75bLinearMuftivadateRegression Model 2 Of Factors PredictinTghe Gamma GlutamytTransferass(Ggt) Among 111 Male Workers..............................1.4.2......... Table 4.1.7ScUnear MuftivariatReegression Model 3 Of Factors PredictinTghe Gamma GlutamytTransferass(Ggt) Among 111 Male Workers..............................1.4.3......... Table 4.1.76LinearMultivariatRegressionModel 1 Of Factors PredictinTghe AlkalinePhosphatass (Akph)Among III Male Workers.........................................1.4.4......... Table 4.1.77Pearson CorrelatioCnoefficientBsetween TotalSerum Fluc@d6.Age, Body Mass Index (Smi),DailyAlcohol Use, DailyTobacco Consumption,And Hematology Parameters ...........................................1.4.5......... Table 4.1.78LinearMuftivariatReegression Model Of Factors PredictinTghe Hemaglobin Among 111 Male Workers.......1.4.6... Table 4.1.79Unear MuftivadateRegressionModel Of Factors PredictinTghe Mean CorpuscularHemobloVin (Mch) Among 111 Male Workers.................; .........................1.4.7..... Table 4.1.80LinearMultivariatReegressionModel Of Factom PredictinTghe Mean CorpuscularVolume (Mcv)Among III Male Workers......................................1.4.8......... Table 4.1.81LinearMultivariatReegression Model Of Factors PredictinTghe White Blood CollCount (Wbc)* Among 111 Male Workers ..............................................1.4.9......... Table 4.1.82LinearMuftivarialReegressionModel Of Factors PredictinTghe PolymorphonuciearLeukocute Count (Poly)Among 111 Male Workers. ........................I.S.O......... Table 4.1.83Unear MuftivariatReegressionModel Of Factors PredictinTghe Band Count (Band) Among 111 Male Workers..............................................1.5.1......... Table 4.1.84LinearMuftivariatReegressionModel Of Factors PredictinTghe Lymphocyte Count (Lymph) Among 111 Male Workers ..........................................................1.5.2......... viu Table 4.1.85LinearMuftivariatReegression Model Of Factors PredictinTghe Monocyte Count (Mono) Among 111 MaJe Workers . ................................... 153 Table 4.1.86LinearMultivadateRegression PredictinTghe EosinaphilCount (Ecs).....................1...5.4......... Table 4.1.87LinearMultivariatReegression Model Of Factors PredictinTghe PlateleCtount (Plate)Among 111 MaJa Workers ...............................................1.5.5.......... Table 4.1.88Unear MultivaHateRegression Model Of Factors PredictinTghe BasophilCount (Baso) Among 111 Male Workers .........................................1.5.6.......... Table 4.2.1CharacieristicOsf 749 FemaJe Employees, 1947-1989 .........1.5.7... Table 4.2.2CharacteristicOsf 2768 MaJo Employees, 1947-1990 ..........1.5.8.... Table 4.2.3VitalStatusAnd Cause Of Death Ascertainment Among 749 Female Employees, 1947-1990 .......................1.5.9........... Table 4.2.4VitalStatusAnd Cause Of Death Ascertainment Among 2788 Male Employees, 1947-1989 .........................1.5.9.......... Table 4.2.5Numbers Of Deaths And StandarcfizedMortalityRatios (Smrs) Among 749 Female Employees, 1947-1989 ..........i.s.o... Table 4.2.6Numbers Of Deaths And StandardizedMortalityRatios (Smrs) By DurationOf Employment Among FornaJe Employees .............................................1.6.1........... Table 4.2.7Numbers Of Deaths And StandanfizedMortalityRatios (Smrs) By Latency Among Female Employees, 1947- 1989 ...................................................16.2........... Table 4.2.8Numbers Of Deaths And Standardized MortalityRatios (Smrs) By Any Employment tnThe Chemical Division Among Female Employees, 1947-1989 .....................1.6.3.......... Table 4.2.9Numbers Of Deaths And StandardizedMortalityRatios (Smrs), Based On U.S.White Male Rates. .................1.6.4.......... Table 4.2.10Numbers Of Deaths And Standarched MortalityRatios (Smrs), Based On Minnesota White Male Rates, Among 2788 Male Employees, 1947-1989 ..................1.6.5.......... Table 4.2.11Numbers Of Deaths And Standardized MortalityFtabos (Smrs) By Latency, Based On Minnesota White Male Rates,Among Male Employees, 1947-1989 ................1.6.6......... Table 4.2-12Numbers Of Deaths And Standardized MortalityRatios (Smrs) By Latency, Based On Minnesota White Male Rates,Among Male Employees, 1947-1989 ................1.6.7......... Table 4.2.13Numbers Of Deaths And StandardizedMortalityRatos (Smrs) By Latency, Based On Minnesota White Male Rates,Among Male Employees, 1947-1989 ................1.6.8......... Table 4.2.14Numbers 01 Deaths And Standardized MortalityRabas (Smrs)By DurationOf Employment, Based On Minnesota White Male Rates, Among Male Employees, 1947-1989 . ............................................1.6.9........... Table 4.2.15Numbers Of Deaths And Standardized MortalityRatios (Smrs) By DurationOf Employment, Based On MinnesotaWhite Male Rates,Among MaJe'Employees, 1947-1989...........................................1.7.0......... Table4.2.16Numbers Of Deaths And StandardizedMortalitRyatios (Smrs)By DurationOf Employment Based On Minnesota White MaJe Rates,Among MaJo Employees, 1947-1989............................................17.1.......... Table 4.2.17Numbers Of Deaths And StandardizedMortalitRyabos (Smrs),Based On MinnesotaWhite Male Rates, Among 1339 Male Employees Ever Employed InThe Chemical Division1,947-1989....................................1.7.2......... Table4.Z 18 Numbers Of Deaths And StandardizedMortalitRyatos (Smrs),Based On MinnesotaWhite Male Rates, Among 1449 Male Employees Never Employed InThe Chemical Division1,947-1989 ...........................1.7.3......... Table4.2.19Numbers Of Deaths And StandardizedMortalitRyabos (Smrs)By Latency,Based On Minnesota White Male Rates,Among Male Employees Never Employed In The Chemical Division1.947-1989.......................1.7.4......... Table 4.2.20Numbers Of Deaths And StandardizedMortalitRyatios (Smrs)By Latency,Based On Minnesota Whfte Male Rates,Among Male Employees Ever Employed InThe Chemical Division1.947-1989...........................1.7.5......... Table 4.2.21Numbers Of Deaths And StandardizedMortaittRyatios (Smrs)By DurationOf Employment, Based On MinnesotaWhite Male Rates,Among MaJo Employees Ever Employed InThe Chemical Division1,947-1989.........1.7.6 Table4.2.22Numbers Of Deaths And StandardizedMorW4 Ratios (Smrs)By DurationOf EmploymeM Based On Minnesota White Male Rates,Among Male Employees Ever Employed InThe Chemical Division1,947-1989........1.7.7... Table4.2.23Numbers Of Deaths And StandardizedMortalitRyabos (Smrs)By DurationOf Employment. Based On WinnesotaWhite M21e Rates,Among Male Employees Never Employed InThe Chemical Division1,947-1989.......1.7.8.. Table4.2.24Numbers Of Deaths And StandardizedMortalitRyabos (Smrs)By DurationOf Employment, Based On Minnesota White MaJe Rates,Among Male Employees Never Employed InThe Chemical Division1,947-1989.......1.7.9.. Table4.2.25Age AdjustedStandardizedRate Ratios(Sffs)For All Cause, Cancer, And CarcgovascularMortalitByy DurationOf Employment, Among Male Employees, 1947-1989 ............................................18.0......... Table4.2.26Age AdjustedStandardizedRate Ratios(Sffs)For All Cause. Cancer, Lung Cancer,Gi Cancer,And CardiovascularMortalitByy Ever/Never Employed In The Chemical DivisionA,mong Male Employees. 19471989........................................................... x Table4.2.27Age StratifieYde,ars Of Follow-UpAdjusled Rate Ratios (Rrmh)ForAllCause, Cancer,And Cardiovascular MortalitByy Ever/NeverEmployed In7be Chemical DivisionA,mong Male Employees, 1947-1989..............1.8.2........ Table 4.2.28 Age StratifieYde,ars Of Follow-UpAdjustedRate Ratios(Rrmh) For AllCause, Cancer,And CardiovascularMonalityBy DurationOf Employment in The Chemical DivisionA,mong Male Employees, 1947- 1989 ................................................1.8.3.......... Table4.2.29ProportionaHlazard RegressionModel Of Fa=rs PredictinTghe AllCause MortalitAymong 2788 Male Workers..............................................18.4.......... Table 4.2.30ProportionaHlazard RegressionModel Of Factors PredictinTghe CardiovascularMortalitAymong 27aa Male Workers .........................................1.8.4......... Table4.2.31ProportionaHlazard RegressionModel Of Factors PredictinTghe Cancer MortalitAymong 2788 Male Workers..............................................18.5.......... Table4.2.32ProportionaHlazard RegressionModel Of Factom PredictinTghe Lung Cancer MortalitAymong 2788 Male Workers ..............................................18.5.......... Table 4.2.33PropordonalHazard RegressionModel Of Factors PredictinTghe Gi Cancer MortalitAymong 2788 Male Workers..............................................18.6.......... Table4.2.34PmporfionalHazard RegressionModel Of Factors PredictinTghe ProstateCancer MortalitAymong 2788 Male Workers .........................................1.8.6.......... Table4.2.35Pmpor6onal Hazard RegressionModel Of Factors PredictinTghe PancreWic Cancer MortalitAymong 2788 Male Workers.............................................1.8.7.......... Table4.2.36ProportionaHlazard RegressionModel Of Factors PredictinTghe DiabetesMellftuMsortalitAymong 2788 Male Workers................................'.........1.8.7........ Table4.2.37ProportionaHlazard RegressionModel Of Factor*s ..*.*'**' PredictinTghe AllCause MortalitAymong 749 Female Woftrs ..............................................l.a.a.......... Table4.2-38ProportionaHlazard RegressionModel Of Factors PredictinTghe CardiovascularMortalitAymong 749 Female Workers..............................................l.e.a.......... Table4.2-39Pmpordonal Hazard RegressionModel Of Factors PredictinTghe Cancer MortalitAymong 749 Female Workers ..............................................1.8.9.......... LIST OF TABLES Figurei.Free Tesiostemne VersusTotalSerum Fluorine.................1.9.0......... Figure2. Bound TestosteroneVersus TotalSerum Fluorine...............1.9.1......... Figure3. EstradioVlersus TotalSerum Fluorine........................i.9..2........ Figure4. LutenizingHormone Versus TotalSerum Fluorine...............1.9.3......... FigureS. FolliclSetimulatinHgormone Versus TotalSerum Fluorine.......1.9.4... Figure6. PmlactinVersus TotalSerum Fluorine.........................1.95......... Figure7.Thymid StimulatinHgormone Versus TotalSerum Fluorine.......1.9.6... Figure8.Bound Testostemne To Free TestosteroneRatioVersus TotalSerum Fluorine.......................................1.9.7......... 1-lNrTROOUCTION Fluorinweasfirisstolataesdan elemenitn1880byMoisse1r.Fiveyearslater he synthesizedthe firsftluorocarbontshroughuncontrolledreactionsofcarbort withelementalfluorineh.was notuntitlhe late1930s thatthe controlled synthesisoffluorocarbonbsecame possible.Inthe 1940s, Fhgidaireand DuPont developed chlorofluorocarbontsh,e firsciommerciallyavailablfeluomcarbons,for use inrefrigerati1o.nDuringthe same periodperiluorocarbonsa,subclassof perfluodnatedorganicfluorocarbonwsithuniquepropertiesw,ere first synthesizedtomeet thespecialneeds ofthe Manhattan pmject 2. The electrochemicaflluorinatimoenthod forperfluorocarbopnroductionmade commercialproductionofpeffluorocarbonpsossibleand opened the door to widespread use of ;>erfluorocarbon3s.-1. Fluorocarbonsare wide rangingintheirstructureasnd uses. Many commercial applicationhsave been developedforchlomfluorocarboncompounds including refrigeratiodne,greasing,aerosoldispensing,polymerizationp,olymerfoam blowing,drugs,and reactivientermediateosrcatalysts.Perfluorocarbons(PFCS) have extensiveapplicationbsecause oftheirunique physicaland chemical propertiesT.hese applicationisncludeuse.as artificiballood substitutes, computer coolants,polymerssuch as teflons,urfactantsl,ubricantsf,oaming agents,skiwaxes, and inan extensivespecialtcyhemical industrywhich produces grtaglz-.tonidlrepellenctoatingsforpaper and c!oth,polymers, insecticideasn,d a varietyofconsumer products.Perfluorocarbonsare currerttly beingtestedas replacementsforchlorofluorocarbonisnindustdajprocessesand products. For many years fluomcarbonswere generallythoughttobe nontoxic. Perfluorocarbonswere consideredtobe particuladnyontoxicbecause theywere chemicallyand physicalliynertand showed low acute toxicitiynanimals4. Recent epidemiologicaalnd experimentalstudieshave associatedexposure to chlorofluorocarbonas,subclassoffluorocarbonspreviouslyclassifieads nontoxicw,ithdirectand indirecatdverse human healtheffects.Subsequently, researchersand regulatortsurnedtheirattentiotno the study ofother fluomcarbons.The discoverythatone perfluorocarbonp,erfluoroocianoiacid (PFOA)w,aspresenitnmeasurablqeuantitienrseside=ofseveraUl.S.cities 5-7.therecognititohnatsome perfluorocarboinscludinPgFOA have longhalf livesinthe humans a and the observationsthatPFOA prodbced toxicoffects in animals,includinghepatotoxiciteyn,docrine toxicityi,mmunotoxicity,and carcinogenesis9,has ledtoa r"valuation ofthe toxicpotentialof. perfluomcarbons,particularPlFyOA, in humans. Despitewidespread exposure to perfluorocarbons,littliesknown about their effecmon human health.h was apparent thatadditionasltudiesdesigned to exploretheirphysiologiceffectsand potentialadverse healthoutcomes and conducted inan occupatonal cohortwith high exposure to PFCA, were necessary. The 3M Chemolhe PlantlocatedinCottage Grove, Winnesota isone ofa few PFC productionfacilitiiensthe worid. Biologicalmonitoringdata from studiesofthe Chemaliteworkforceshowed thatemployees have had high levels and long durationsof exposure to PFOA 8-10. This occupatonal cohortprovided theopportunitytostudytheeffectsof PFOA on humans. The specificgoals and objectivesof thisstudy were: GOAL 1)To quantifythe human offects ofperfluomoctanoicacid on the followingphysiologicparameters: a) Hormones: freeand bound testosterone,estradioll,utenwng hormone, thyroidstimulatnghormone, prolactina,nd follicslteimulating hormone. b) Serum lipidasnd lipoproteinsc:holesteroll,ow densitylipopmtein, high densitylipopmtein,and tdglycerides. c) Hematologicparameters:hemoglobin, mean corpuscularvolume, whiteblood cellcount,polymorphonuciear leukocytecoum band count, lymphocyte count,monocyte count,platelectount,eosinophflcount,and basophilcount. 2 d)Hepatiecnzymess:erumglutamiocxaloacettircansaminassee,rum glutamipcynivitcransaminasgea,mma glutamyttransferaasned,alkaline phosphatase. OBJECTIVE1:toconducatcross-sectisolnuadlyofproductiwoonrkertso estimateherelationshbieptsweentotaslerumfluoriades,urrogataessay forprefluorooctancoiidac,ndphysiologpiacrameters. GOAL 2)Toquantithyemortaliitnaynoccupationcaolhorwttthlongterm exposurteoperfluorooctaancoiidpcroduction. OBJECTIVF2:toconducat retrospectciovheorotccupationsatludyto assessthemortalietxypedenceofworkerussingexpectemdortalibtaysed on Minnesotmaortalirtaytes. 3 5)-g;zvlfw c)FTHF LRTERAT[JRE 2-1 Intrneitiction The presence ofsmallamounts offluorideinhuman blood was recognized in 1856 11.More than loo years laterT,aves 5,6 presented evidence thatfluohne existsintwo major forms inhumans and animals;ina freeionicstateand ina covalentlbyound organicstate.Priorto thisreport,itwas assumed thatfluorine existedprimarilays inorganicionicfluoridienbiologicaslystems. Taves, observationshave sincebeen confirmedby severalother investigator1s2-16.The discoverythatorganofluorinecompounds consthLAsthe majodty of fluorinefound inhumans focused researchon characterizintghese undefined compounds. Guy. iderytifiaepderfluo6natedcompound, ;>erfluorooctanoaiccid (PFOA), as a major cons-@tuenotfthe serum organicfluorinteracbon7.17.Perfluomoctanoic acid (PFOA) isthe onlyorganicfluo6necompound to be identifieidn human serum is. The recognitioonfhuman and animaltoxicitieasssociatedwith portluomchemicws 9.19,has renewed interesitnunderstanding the human health effectsof perfluamearbons(PFC), particularPlFyOA. 2 Omanic Fluoroebemicats Organic fluorochemicalso,therwisereferredto as fluorocarbons,are compounds composed offluorinec,arbon and otherelements such as oxygen, nitrogenand sulfur.Perfluomcarbonshave structuresanalogous to hydrocarbons, except the hydrogens areexhaustivelyreplacedby fluorine20.A limitednumber oforganic fluomchemicalsoccur innature 21-23h,owever no PFCs occur naturall2y4.2S The firsrieportofthesynthesisof a fluomcarbon was publishedin1890 when Moissan claimed to have pu4fied carbon letrafluoridfet.islikelyhe isolated fluomgraphfteh,owever 1. Pure carbon tetrafluoriwdaes not obtaineduntil1930 26.Work by Ruffand the Belgian chemist,Swans, inthe late1Sth and early2Mh centudes laidthefoundationof organicfluodde chemistry.Midegly and Henne exiended Swarts'work and reportedthe synthesisof dichlomdlfluormethane, 4 C12F2i,n193027-Thischlorofluorocwairtbhotnhetradename Freon12isan inertn,on-toxicrefhgerariwthich was vastlysuperiorto otherrefrigerants availablienthe 1930s. AftercommercialproductionofFreon 12 began in1936, rapidlbyecame a majorindustricahlemical2-21. A number of choloffluommethanesand chlomfluomethaneshave been produced on a commerciajscaleinmany regionsoftheworld.These chlorolluorocart)ohnasve been used inlargeamounts as aerosolpropellantasnd degreasers,inadditionto theiruse as refrigerantCsu.rrentlyt,heirproductonisbeing reduced as a result oftheirozone depletingproperties in1937,Simons and Blockdevelopeda method toproduce laboratorqyuantities ofperlluorocarbons,uch as C3F8, C4FlO, cycioC5Flo and cycloc@6Fl22-3. The analysisofthesecompounds ledtotheunderstandingthatmany ofthestructures ofsaturatedhydmcart)onscouldbe replicateidnthe form ofperfluomcarbons. Fiesearchinthe area ofperfluorocarbonwsas stimulatedby two developments. FirstP,lunkettdiscoveredthepolymer,polytetrafluomethylenoer,Teflon1. Second, thedevelopment ofperfluorocarbocnhemistrywas stimulatedby the U.S.effortto develop atomicweapons duringWorld War 11under the Manhattan Pmject.The 235U isotopeofuranium was requiredforthedevelopment ofatomic bombs. One method ofuranium isotopeseparationwas gaseous diffusionT.he onlyvolatilueraniumcompound availableforuse inthisdiffusiopnrocesswas uranium hexafluodde,UFS, an extremelyreactivegas. Materialswere needed tor use as coolants,lubricantse,alersand buffergases inequipment exposed to thishighlyreactivegas 1.Z 26. Perfluorocarbonpsrepared by Simons were found tobe inerto UF6. This discoveryledtoa researcheffordtirectedtoward understandingthepropertieosfa varietyofperfluorocarbonasnd developing commercialmethods forpreparationofperfluomcarbons.The development by Simons oftheelectrochemicaflluorinati(oEnCF) was a major milestoneinthe fluomchemicalindustryS.ince World War 11therehas been much interesatnd work inthisnew branch oforganicchemistrybased on perfluorocarbons. The use ofSimons'ECF method has allowedthe productionofa wide varietyof perfluomcarbonsincludingperfluorinateadlkanes,alkenes,ethers,esters, amides,sulfonamidesand compounds withcyclicand ringstructure2s. The 'inert' perfluomearbonsare compounds made up ofonlycarbon and fluorineT.his class ofcompoundsrangesfromcarbontetrafluortiodceomplexmultiplreing structuresuch as perfluorodecaliPne.rfluo6natedsurfactantsincludecarboxylic acids,sulfonicacids,and theirderivativesT.hese compounds form the basis of an extensivefluorochemicalindustryA. varietyof;>erfluorinatepdolymers and . elastomersexistT.he most widelyused are polytetrafluoroethyieannsd Kel-F,a slastomerofvinyidienefluohdeand hexafluorapropylens. 2-3 Physic;tPfrol2erties Perfluorooclanoiaccidisa straighcthaineightcarbon carboxylicacid witha mole=iar weightof414.1S. The mehing pointof POFA is59-60*C. Itsboiling pointis189*C atstandardconditions30. Perfluorooctanoicacidisproduced as a complex mixtureof branched chainisomers. Inpractice,alleightcarbon carboxy(icacidisomersare reforedto as PFOA. The ammonium sattof PFOA (APFOA) isthe common industrialulsyed form of PFOA. h isa whitecrystalline powder thateasilybecomes airbome and sublimes at 130'C. Perfluomearbonshave unique chemical and physicalproperties20.26.31.32T.he importanceof perfluodnatioinnproducingthese propertiescannot be overemphasized. Perfluorocarbonsare notjustanother hydmcarbon-like molecule.Chemically,perfluorocarbonsare remarkably inerlThey are stableto boilinignstrongacidsand bases. Very few oxidizingor reducingagents react appreciablywithperfluorocarbonsP.erfluomcarbons thalcontainotherorganic moleculessuch as nitrogen,Mgen and sulfurwfllparticipatienreactionatthe siteofthese molecules.For instance.perfluoroctanoyslulfonicacid wdl reactand form thesuffonamidedertvativeT.he amide portionofthismolecule can then be conjugatedwithmany otherorganiccompounds. The perfluorinatepdortionof these largermolecules remains non-reactive. Perfluomcarbons are heat stable.They can be heated to greaterthan 250*C withoutbreakdown. At high temperatures,greaterthan 400*C, some compounds willbreakdown. For example, PTFE, breaks down to perfluomisobutyiene(PFIB), an extremelytoxicgas 1.Because most perfluomchemicals are heat stablethey are used inhightemperature applications. 6 The inertperfluor=rbons areexcellenitnsulatorsP.olymem, such as PTFE, and inertsPFCS. such as perfluorohexanea,reused inelectricajpplications because oftheirsuperiordielectrpircopertiesT.heirheat stabilitaynd insulation propertiemsake perflucrocarbomnaterialtshe insulatorosfchoice20. Perfluorinatesdurfacesare the most non-wettableand non-adhesivesurlaces known 20.26. Flucrochemicalsurfactantasre some ofthe mosi poter?sturface activeagentsyetdiscovered31.Very low concentrationsofftuorochemical surfactantesffeclivelryeduce the surfacetensionat interphaseboundaries. Most penluorocarbonsare poorlysolubleinboth aqueous and organicsolutions. They form a group oflluomphiliccompounds, however some perfluorocarbons withfundonal groups such as thesaltsofPFOA, are highlywater soluble31.32. Perfluoror-arblainquiddsissolveoxygen avidly.This unique propertyisthe b Is forthe use ofperfluomcarbonsas bloodsubstimes 33. Perfluodnatedcarboxyliacnd sulfoniaccidsare some ofthe strongestorganic acids.known31. The pKa ofPFOA is2.5 34. Thus, when inphysiologiscolutions, theyexistinprimarilaynionicforms.Tne anionicforms have a strongpropensity toform complex ionpairs*. Inthepast,some investigatorhsave assumed thatthe chemicaland physical propertieosfmany fluomcarbonsissynonymous withlackofactivitiynbiologic systems -Is3-6. However, abundant evidenceexiststhattheirchemicaland physicalinertnessdoes notimplybiologiicnertnessI'-30.37.38.. 2-4 Synth@sis Synthesisoffluorocarbonshas been accomplishedusing fourmajor methods; electrochemicaflluorinati(oEnCF),directfluorinatiotno,leomerizationa,nd catalytimcethods usinghighvalenceheavy metals.The ECF was developed by Sirnonsin1941 3. The Simons processistheoldestcommercial techniqueand remainsa commercial method toobtainmany perfluorocarbons.A solutionof PemonalcomrmnicationiromJames Johnson.3M Corporation 7 organicsubsirateisslOCtrOlyzeidnanhydrous HF ata low voh2ge, high current, nickelanode.The productsoftheseelectrolysciesllreactionsare largely perfluodnatedT.he spectrum of materialproduced by the ECF process isdefined by thestartinmgaterialC.ommercial productsfrom thisprocess include perfluomalkanesp.erfluoroalkeytlhers,perfluoroalkenesp,erfluorowkylesters, perfluorotrialakmyilnes,pemucroc-arboxyliaccidsand perfluomsulloniaccids2. ProductsofECF oftenincludea significanptroportionofcomplex isomers and fragmentatiopnroducts.For example,ECF productionofPFOA from straight chainoctanoicacidproduces30% complex branchchain isomers 39. The mixture ofproductsfrom each ECF run isunpredictablvyariableT.hese isomericmixes are difficutfotseparateand pu6ty-'*.Workers producingPFCs usingECF may be exposed to a complex mixturethatchanges compositionovertime. Directfluorinatiiosnanothermethod used toproduce perfluorocarbons.Itisnot subjectedtothe impuritpyroblems associatedwiththe ECF process.Direct fluorinatiroenactsfluoringeas withhydrocarbonsubstrate.Because fluoringeas isextremelyreadve, directfluorinatiiosna technicalldyifricupfrtocess and has onlyrecentlybeen pilotestedforcommercialproductionoffluomcarbons. World productionoffluorocarbonsislimitetdo a handfulof commercial plants. The 3M CorporationoperatesPFC productionplantsinMinnesota,Illinois, Alabama and Antwerp,Belgium.A plantinItaloywned by a Japanese and Italian .-consorbLtm-pfpducelsimiteadmount offluorocarbonsP.erfluorocarbonsare also produced inGermany- and have been produced,inthe past,inthe formerSoviet Union. 2.5Sources Of Croanic Fluo6de Exnosure Guy 17 presentedpossiblecandidatesforthe organicfluodneconstituentosf human bloodbased on observationmade duringthe isolatioonfPFOA fmm serum. The organicfluodnewas notlikeltyobe a macmmolecule such as a proteinornucleicacid,because ofitssolubiliitnyorganicsolventssuch as ether orchlorofomi/methanolI.twas not covalentlbyound toalbumin sinceItwas removed on charcoalatpH 3 atroom temperature.The solubilicthyaracteristics suggestedthatmultiplceompounds existedwithdifferenptolaritiesT.he major compound was a polarlipildikeMoleculethatwas identifieads PFOA. other lesspolarcompounds appeared to be present.Thisdata suggeststhat flucrocompounciostherthan PFOA were bound toalbumin.These compounds were not estersOfCl 3 -18 fattyacidsand were lesspolarthan PFOA. Perfiuorooctansyullfonamide(PFOS) and itsderivativceompounds ftthis descriptioannd may be constituentosfthe organicfluorinteractionA.lthough exposure isprobablylow,thepropertieosfPFOS suggest thatitmay accumulate tomeasurable levels. incontrastoionicfluoridel,itthlaes been reportedconcerningthe organic fluorinceontentofwaterand beverages.The fluo6necontentofgmund wateris essentiallayllinionicform. Some fluorochemicalss,uch as the perfluorinated carboxyliaccidsurfactartatnsd theirsalts,are solubleinwater. Such water solublecompounds may locallcyontaminatesurfaceand ground water near industripalantsthatuse these compounds. Other perfluorin2tecdompounds such as the Wkanes, alkenes,and ethersare fluomphiliacnd are insolublein aqueous solutionsA.Jthoughdataon the oralorganicfluorinientakeislimitedi,t isunlikeltyhatwaterand beverages are significasnoturces oforganicfluorinien humans. The criaats a sourceofthe organicfluorinfeound inhuman serum has been the .subjecotfspeculations.s.18.40. Non-pernuorinatedfluomcompounds have found inbiologicaslystenz,-Maralsshowed-thatfluoroacetatweas the compound responsiblfeortoxicitfyrom thepoisonous plantDichapetalumCYMCSUM 41. Otherinvestigatohrasve found plantspeciesthatsynthesizefluomacetate, fluamcitrataen,d monofluorinatefdattyacids. Petersreportedthata few toxic plantsproduce fluoroacetat4e2. Fluoroacetateand fluorocitrahtaeve been found inbeans gmwn inhighfluoridseoil23. Peters21 and Lovelace st al.22 have reportedthe occurrenceoffluorocitraitnea few plantsand foods. Inanimals,the metabolicactivatioonffluoroacetatiento(-)-erythro-fluorockbmleocksthe transporotfcitratientothe mitochond6a and citratbereakdown by aconhass 4z '43.Other omega-fattyacidswitheven numbers ofcaeoon atoms are highlytoxic as a resultofoxidatiotnhatproducesfluomacetate.Fluorocitrataelsoundergoes rapiddefluarinatiionnratliveirnthepresence ofglutathion(eGSH) ". Given the low envimnmental levelst,he infrequenotccurrence,the toxicitayn,d the rapid metabolisomfthesecompoundsinm2mmalianspeciesi,tisunliketlhy2tthese monofluodn2ted cornpounds contributesubstantialltyothe organicfluorine contentinhumans. Taves measured the organicand inorganictuohns in93 food hems 45. No significanotrganicfluorinweas found inthe testedfoods. Ophaug and Singer testeda market basket offood. They concluded thatthere was no significant organicfluorinecontentinfood. Althoughfood and beverages generallydo not containPFCS, h ispossiblethatthey may be contaminated by fluorochemicaj packaging materials.Water and grease repellenctoatingsinpackaging material could leachintofood items insmallquantitiesT.his could occur when materials thatare notdesigned formicmwave use are used inmicrowave ovens. Studies have not been reportedthatquantityhuman exposures from food packaging sources. Perfluorocarbonsare containedinmany consumer products.Fluorocarbon surfactantsuch as PFOA, PFOS, and itsderivativesare present inwindow cleaningproducts,floorwaxes and polishes,fabricand leathercoabngs and carpetand upholsterytreatments20. Additionalltyhese compounds are used to coat food wraps and are incorporatedintoplasticfood storage bags. Fluorocarbonsare the basisfora new generationof=ss countryskiwaxes. :Teflonand Teflonrelatedpmducts are widely used as lubricantse,lectrical insulatorsh,eat and chemical stablegaskets and liningsand in non-sock cookware. Fluomalkanes such as perfluomhexane are being evaluatedas CFO replacements.Ifperfluomhexane or otherfluorocarbonsare used as replacementsforCFC'S, consumer exposure from aerosolsand otherpmducts willincreasedramatically.PFC's have severalexperimentalmedical uses includinguse as blood substitutesx,-rayand magnetc resonance imaging contrasiagents ". vitreousreplacementand inliquidverttilatitohnerapeutic methods 47.Recently,a potentfluomcarbon insecgcidehas been marketed to controlflreants 48. Perfluorocarbonshave a varietyofindusidaluses. Teflonand otherpolymers are used where heat stableand chemicallyinertlinersg,askets and lubricantasre necessary.Inaddhion.they are used as electricailnsulatorsboth insolidand 10 liquidform and used as inertnon-conductiveliquicdoolantsinelectricadlevices such as Cray supercomputers.Perfluorinatesdurlactantsare importantfire suppressionmaterials.Perfluorocarbonshave been used to controlthe metal vapors inelectroplatipnrgocesses and to preventthe releaseoftoxicgases\ from landfills2POe.rfluerocarbonsare being considered to replaceCFC's in many processes such as refrigerationp,olymer loam blowing and buildinginsulation. New applicationsare being continuallydeveloped forthese unique compounds, making increased exposure to workers probable. 2-6 Toxicokinetics of PFOA Since Taves and Guys observations, perfluomcarboxylic acids, PeAuomsuffonic acids and theirderivativeshave been the subject of numerous toxicokinsticand toxicodynamic studies inanimals. These studies have focused primarilyon two compounds, PFO& and perfluorodecanoic acid (PFDA). Perfluorooctanoicacid or itssaltsare well absort>ed by ingestion.inhalationor dermal exposure. Absorpton has been studied primarilyin rats.although a number of other spades have been studied. Five mais and fivefemale ratswere exposed to airbome APFOA for one hour. In thisexperiment the nominal airconcentration of ammonium perfluomoctanoate was 18.6 mgA.-No animals died during the inhalationexposure or the 14 day post exposure observation period. Pooled serum samples contained 42 ppm of organic fluorineforma)es and 2 ppm forfemales. Inorganic fluoridecontent was 0.02 ppm for males and 0.01 ppm forfemales 9. Kennedy and Hall 38 studied the inh2lationtoxicityin male rats of ammonium perfluoroctonate using both single dose and repeated dose schedules. They found a LCSO of 980 mg/m3 fora 4 hour exposure placing PFOA in the moderately toxic by inhalationcategory. Following ton repeated doses at levels of 1.0,7.6, and 84 mg/m3 blood ammonium PFOA levelswere obtained. At the 1.0 mg/m3 level PFOA levels were 13 ppm, atthe 7.6 mg/m3 levelPFOA levelswere 47 ppm and at 84 mg/m3 levelPFOA levelswere 108 ppm. Therefore itappears that PFOA iswell absorbed by inhalation.Itshould be noted that the exposures were to APFOA dust the likelyform foroccupationaj exposu're. Ammonium p-erfluorooctonoaitnefood and PFOA administersd by gavage in propylensglycolorcorn oilvehiclesarewellabsorbed inrats'.In an acute oral LOSO study 9.ratsdisplayeda dose dependent spectrum oftoxicitieisncricating thatPFOA was absorbed afteringeslionP.FOA levelswere not measured inthis study.Ina subacute oraltoxicitsytudy,ratswere fed PFOA for90 days 9. Serum concentrationoforganicfluorineshowed a dose response relationshiipn both sexes. A marked gender differenceinorganicfluorinelevelswas observed. Males had organicfluohns50 times higherthan females at each dose level. Studieshave sincedemonstrated excellentoralabsorptionof PFOA ina variety ofspeciesincludingrats,mice, guinea pigs,dogs, hamsters and monkeys 9-19 Of most immediate.relevanceto humans have been stucfieisna small number *of rhesus monkeys 9.Ina 90 day oraltoxicitsytudy,monkeys were given3, 10,and 30 mg/kg/day doses ofAPFOA. Inmonkeys althe 3 mg/kg/day dose, mean serum PFOA was 50 ppm in males and 56 ppm infemales.At the same dose, ma)es had 3 ppm and females 7 ppm inliversamples. At 10 mg/kglday doses, maje monkeys had a mean serum PFOA of63 ppm and females 75 ppm. Uver levelswere 9 and 10 ppm formales and females,respectivelyB.ecause allbut 1 monkey died atthe 30 and 100 mg/kg/day dose levelso,nly 1 serum sample from a male monkey inthe 30 mg/kg/day dose group was available.Inthismonkey the serum levelofPFOA was 145 ppm. Inthe30 and 100 mg/kg/day dose gmup .-mean liver.levewlesre greaterthan 100 ppm. Thus, the oralmute ofabsorption May be a significanctontributotro the body burden of PFOA Inexposed workers. Dermal absorptionofPFOA has been studiedinratsand rabbits.Ammonium perfluorooctanostiesa finewhite powder thatmay come intocontactwithswn and be absorbed.In mts dermally exposed toammonium perfluomctonateai4 dose levels,PFOA was absorbed ina dose dependent fashion37 . Insingledose dermal exposure experiments using rabbits,PFOA appeared to be absorbed. Levelsoffluorinewere not measured, butdose dependent toxicchanges were noted 9. Ina mufti-doseexperiment,ten male and ten female rabbitswere injectedermallywitha 100 mg/kg dose of PFOA on a fiveday a week schedule fortwo weeks. Totalserum fluorinelevelswere increasedina dose-dependent fashion.Dose-dependent changes inweight were noted 49.Fmm these studies, 12 itappearsthatdermalexposuretothesaftsofFFOA are absorbed inanimals. In the past,Chemoliteworkershave been exposed tolargedermal doses of ammonium perflucrocionatIet.appearsthatdermal exposure may have playeda significarotlteintheabsorptionofFFOA inthese workers.Upon recognitionthat PFOA couldbe absorbed dermally,work practiceswere changed and engineeringcontrolswere adoptedthatreduced dermal exposures.The rolethat dermal exposurescurrentlpylayinPFOA absorptionat Chemolite has not been wellstudied. Once absorbed,FFOA entersthe plasma pmbably by diffusinags a neuiralion pair.Inplasma,PFOA isstronglybound toproteinsinthe serum withmore than 97.5percentinbound form -50.itislikeltyhatalbumin isthe major she forhigh affinibtiyndingr"7.-',O-@T4h]e.re does notappear to be a sex differenceinprotein binding50-u. HanhijarveitaJ.have suggestedthatproteinbindingissaturans in raisw. Usinghuman serum, Ophaug and Singer39 found thatPFOA was 99% proteinbound atPFOA levelsup to16 ppm totalfluorineh,owever. Guy suggestedthatperfluomcarboxyliacidsbindtoalbumin ina similarfashionto fattaycids24. Thishypothesisisconsistenwtiththe resultsofseveralstudies. Taves observedthatthe organicfractioonf serum co-migmted withalbumin duringelectrophoresi6.s DiaJysisand ultraftitrasttiuocnrieosbserved the retentioonforganicfluodneduringctalysiasnd ultrafiltrat7i.o1n75.6. Belisleand Hagen reportedthatPFOA appeared to be stronglyproteinbound inhuman sonims'. Extractio-onf.,FPOA- from acidifie2dwaterisq.uantitativelcyomplete usinghexane.When PFOA isextractedfrom plasm2. recoveryisonly35 percent.Plasma appeared tocomplex PFOA and PFDA. The partitionionfgthe bound intoorganicphase duringextractiownas more difficuafntd necessitated theuse ofmore polarsolvents.Kievens53 suggested thatCF2 and CF3 groups complexwithpolargroups thatare presentinthe amino acidsinproteinssuch as albumin.Inproteinprecipitatisotnucrieussingbovine serum albumin,PFOA bound toalbuminatan estimated28 bindingsitesper molecule52. Nordby and Luckstudiedthe precipitatioofnhuman albumin by PFOA- Under acidicpH conditionsP,FOA produced reversiblperecipitationf albumin 57 by bindingto highaffinistiytes.These studiesdo notruleout signific2nbtindingtoother plasma proteinsorerythrocytceomponents. instucrieussing serum protein electmphoresist,he proteinbound organicfluorinweas distdbLAedina diffuse 13 patter6,n17suggestinthgalPFOA proteinbindingmay be nonspecificT.he large amount bound to albumin may reflectthe abundanc*e of albumin in plasma and serum. in rats,PFOA isdistriblrtetdo alltissuesstudied except adipose tissue. The highestconcentrationsof PFOA are inthe serum, livera,nd kidneys. Ylinen st a]. u studiedthe dispositionof PFOA in male and femaje ratsaftersingleand 28 day oraldosing. Aftera singledose of 50 mg/kg, PFOA was concentrated inthe serum. Twelve hours afterdosing 40% ofthe PFOA dose was found inthe serum of ma)es and 10% in females. Males retained3.5% of the dose in serum atterl4days. PFOA was retainedinthe liverformuch longer than inserum. In females,the half-lifoef PFOA in liverwas 60 hours compared to 24 hours in serum. In males the half-lifweas 210 hours in liverand 105 hours in serum. itis noteworthy that PFOA was not found in acfiposebssue in detectable quantities. After28 days of PFOA treatment,PFOA was dist@buted to the followingsitesin decending amounts: serum, liverl,ung,spleen,brain,and testis.Again, no PFOA was found inadipose tissue.The distributioonf PFO from serum tothe tissuesoccurred ina dose dependent manner forfemales. In maje rats,the concentrationsof PFOA in testisand spleen followed a dose dependent trend. The levelsin male rat*terumand liverwas the same forthe 10 mg/kg and 30 mg/kg dose gmup,. Johnson and Gibson I'-19 studied the distributioonf 14C Icibeleadmmonium pernuomoctonoate aftera singleivdose In rats.Their findingswere similarto those of Ylinen et al. The primary shes of distribution were the liverk,idneys,and plasma. Other sites,includingadipose tissue,had lessthan 1% of the administered dose. The levelof PFOA In the testisof male ratswas not reported. As cfjscussedpreviously,the 90 day oraltoxicitsytudy In rhesus monkeys showed thatthe relativeamounts of PFOA in serum and liver was differenItn monkeys compared to rats.In the low dose group of monkeys (3 and 10 mg@kg/day) serum had 5 to 10 times the PFOA levelsfound inliver. However, at higher dose levels.the PFOA levelswere equallydistributed. Additionallyn,o sex cfille,encewse,,,noted inthe monkeys livorand serum PFOA levels. ' There isno evidence thatpemuorinated compounds includingPFOA are biotransformedby livingorganisms. Several studies have examined whether 14 PFC)A isconjugatedor IncorporateIdntotissueconstftuentsuch as tn'glycerides orlipidsY.linenetal.foundno evidenceinWistarr= formetabolism or incorporatioonfPFOA intolipid-s".Afthoughthe lipidcontentinPFOA treated ratswas difefrenithan thatinuntreatedrats,Pastoor.etal.did notfindevidence. forPFOA incorporatioinntolipidosrofmetabolism60 Vander Heuval'etal. showed thatPFOA was notincorporateidntotdacylglycerolpsh,ospholipidso,r cholesteroelstersinthe liverk.idney,heart.fatpat.or testisof maje or temajo rats61.No evidencehas been foundthatPFOA isconjugated inphase 11 metabolism61. KuslikiestaJ.studiedtheformationof activatedcoenzyme A (CoA)derivativeosfPFOA usingratlivemricrosomes. They foundno evidence fortheformationofa CoA derivative. Sex relateddifferenceisnthe toxicokinetiocfsPFOA have been reportedforrais. The mechanism ofPFCA excretionappearstobe species-dependentsincethese genderdifferenceasre notseen inmice,monkeys, rabbitso.r dogs 9.62.The half-lioffePFOA infemaler2tshas been estimatedto be lessthan one day 39, whereas the half-lioffePFOA inmales isfiveto seven days 34-M. Itisofnote thatPFOA does notexhibithisgender differenc9e3. ftishypothesizedthatthe sex differenceisnsensitivittoythe toxicitioefsPFOA are as a resultofthe slower 01 excretioonfPFOA inmale ratscompared tofemale rats.Investigatorhsave reportedthatratshave an estrogen-dependentactiverenalexcretionmechanism torPFOA which can be inhibitebdy pmbenecid Sc)--@A4s.noted previously, femaleshave a much shorterhalf-litfhean male rats.The half4ifienmales can be reduced by castratioonrestrogenadministratioInt.can be reduced tothe femalehalf-libfyea combinationofcastratioannd estrogentreatment.Estmgen administratoanloneisalmostas effectivaes thecombinationofc=raton and estradiotlreatmentinreducingthe PFOA half-lifTeh.istreatmentincreasedthe renalexcretioonfPFOA inmale ratstothose observed infemale mts. Other investigatohrasve reportedthatthe gender differenceinhalf-lidfeepends on a testosteronmsediatedincreaseinPFOA tissuebincring-6T4h.is hypothesisis consistenwtiththe gender differenceintissuehalf-licffeiscussedpreviously34. Johnson has suggestedthatthe primarymethod ofexcretioninintactmales is viathehepatobiliarmyute sg-sg.He reportedthatcholestyramineenhanced the fecaleliminatioonfcarbon 14 labeledPFOA inmale rats.These data suggest therewas biliareyxcretionwithentemhepatc circulatioonf PFOA, particuladiyn male rats.HowevGr, ina maig worker withhigh serum PFOA levelswho was treatedwithcholestyraminel,ittliefany change inexcretionof PFOA was noted. Inthisstudy PFOA was excretedslowlyinthe urine. Inhumans, the half-lioffePFOA appears tobe extremely long and isnotsex dependeryt.Ubel and Griffit8hreportedkineticdata forone highlyexposed worker.At the time he was removed from exposure hisserum organicfluorine was 66 ppm, 80 percentofwhich was PFOA. Over the next 18 months his organicfluorinleeveldecreased to39 ppm. UrinaryexcretionofPFOA fellfrom 387 micrograms/24 hours to 80 micrograms/24 houm. The declineinorganic fluo@nelevelswas consistentwithtwo compartment kineticsw,ith a calculated half-lioffe2 to 5 years.Additionalunpublished biologicamlonitoringdata from threeChemoliteworkers isconsistentwiththe 2 to 5 year half-lifIen.the Chemoliteworkforce,male and femaje workers employed injobs withsimilar PFOA exposure have increasedPFOA levels.Since men and women with similarexposures have similarlevels,a largegender differenceinPFOA toxicokineticisunlikelyT.herefore,the relevanceof the ratdata inassessing the effectsof PFOA in h.umans isquestionable. 7.7Toxicedynamics ofPFO ?.7-1Male Renmducgv@ Toxicities Both PFOA and PFDA have been found to produce significanttoxicitieisnthe reproductivesystems ofmaie rodents 19,63-65.The testishas been reportedas thetargetorgan oftoxicitfyorboth PFOA and PFDA Additionalevidence existssuggestingthatthese compounds affectthe functionofthe hypothalamic- pftuftary-gonaadxis(HPG) 19-65. Pernuorodecanoic add, but not PFOA. has been shown to pmduce degenerative changes inratseminiferoustubulesthatcould pmgress totubularnecrosis.Van Ratelghem etal.reportedthata singleipdose of 50 mgag of PFDA, produced degenerativechanges inratseminiferoustubules 8 days afterinjectio6n6.Similar butlesserchanges were noted inthe seminiferoustubules ofhamsters and guinea pigstreatedinthe same manner. They reportedno such change in 16 treatemdiceB.ookSt2ffand Moore61didnotobservesimilacrhangesinrats treatewdith20-80mg/kgofPFOK They useda differesnttraionfr2isintheir expedmerttwshich islesssusceptibleto testiculatroxicantsthan those used by Van Ratelghem et al. Thus, the effectsof perfluomcarboxylic acids on seminiferoustubulesmay be limitedto a specificompound, PFDA. ina specific strainof rats.The effectsobserved by Van Rafelghem st a].in other spe<@Gs were not consistent and did not demonstrate a dose-response relationship.In monkeys treated orallywith PFOA, no compound related histopathologic changes inthe seminiferous tubules were noted 8. In a two year ratfeeding study, PFOA treatedanimals were observed to have increased numbers of Loydig celltumors*. Male and female r2ts were fed PFOA containing dietsresultingin a mean intake of 1.5 and 15 mg/kg/day. A siatisticajlsyignificanitncrease in Loydig celladenomas of 0%, 70/o,and 14% in the control,low dose, and high dose groups, respectively,was observed at the end of the two year study. The resultwas statisticalsliygnificantas a resuh of the unexpectedly low number of adenomas incontrolanimals. HistoricallyC,D rats experience a lifetimemean Leydig callincidence of 6.3 percent whh a range of 2 to 12 percent. The high dose group incidence is outside the expected range and may represent a compound relatedeffect.Although the evidence was not definftlvei,tsuggested that PFOA may alterthe histologyas well as the function of Loydig cellsinr2ts.Penluorcoctanoic acid was not mutagenic in the standard testsincludingthk Am!ps assay using f@ive.@p@!cioefs Salmonella t),Phimudum and insaccharomycescgrevisiae Mammalian celltransformation 2SS2yS using C3H 1OT 1/2 cellswere also negative b'7. These data suggest that PFOA is not a genotoxic xenobiotic. The increase in Leydig celltumors may be the resuftof an epigensticmechanism. The observation that rats fed PFOA for 2 years had an increased incidence of Leycrigcelladenomas prompted researchers to examine the hormonal effectsof PFOA in majo rats 19.Adult male CD rats were treated orallywith PFOA in doses of 1 to 50 mg/kg. Serum estradiollevelswere elevated in the ratstreated wfth more than 10 mg/kg of PFOA. In the highest dose group estradiolwas 2.7 times Report3:M RikerLaboratoriesT.wo Year OralToxie-fty/CarcinogenSitcufdtyofFC143 InRats 9281CROO12,1983 17 greaterthan the estradiollevelsinpairfed controlgroup rats.Serum testosterons levelswere significantdleycreased ina dose dependent manner when compared withad libitumfeed controlanimals.No significandtifferenceswere observed between the highdose ratsand theirpairfed controlsh,owever. No significant differencewsere notedinserum luteinizinhgormone (LH)levels.Additionallyt,he accessorysex organ relativweeights ofthe highestgroup were significantleyss than those oftheirpair-fedcontrols. Inordertoclahfythe siteof PFOA action,Cook 19 conducted a set of challenge experimentsinPFOA treatedmts. The resultsofthese experiments demonsiraie thatthe Wteredtestostemne levelswere PFOA related.Human chorionic gonadotmpin (hCG) challengecan be used toidentifaybnormalitiesinthe stedodogenicpathway. Human chohonic gonadotmpin binds to the LH receptors on Leydigcellsand stimulatessex steroidhormone synthesis". Abnormaitties inLeydig cellfunctioncan be detected by challengingLeycrigcellswith hCG and measuring steroidhortnoneproduction.Similarlya,bnormalitiesinphuitary secretionofgonadotmpins can be identifieudsing a gonadotropin releasing hormone (GNFIH) challengethatstimulatesLH release 69. Hypothalamic dysfunctioncan be identifieudsing a naloxone challenge to stimulateGNRH release70. inratstreatedwithPFOA for14 days at the same dose levelas the initieaxlperiment,theLeydig cellproductionoftestosteronswas significandy bluntedafterhCG challengeinthe highestdose group compared to ad libitumfed controlsA. small,non-significanbtluntng ofthe testoslemne productionin responseto GnRH.and naloxone was-observed. FollowingGNRH and naloxons stimulationL,H levelswere not significantdliyfferenitnthe treatmentand contml animals.The hCG challengeshowed thatthe decrease intestostemne inPFOA treatedratsresultedfrom alteredsteroidogenesisinthe Leydig call.The results from theGNRH and naloxone stimulationwere not definitiveT.he resuttswere compatiblewithan efefctat the pituitarlyevelas wellas at the Leydig celllevel. Cook etal.examined thesiteat which testosteronsstemidogenesis was affected by PFOA. Progestemns, 17 alpha-hydmxyp mg estem ne and androstenedione were measured afterhCG challenge. Progesterone and 17 alphahydmxypmgestemne were unaffected.Androstenedions levelswere significantly decreased inPFOA treatedr2tscompared to controls.Given thatthe conversion of17 alpha-hydmxyprogestemne to androstenedione by C17/20 lyassis necessaryforiestosieronseynthesist,heso resultsuggestthatdecreased testosteroniestheresultofa blockinthisconversion'stepI.nhCG stimulatedrai Leyerigcellst,he 17 alphahydroxyiase/C-17/2l0yaseisinhibitebdy estradiol. Taken togethert,hesedata are consistentwiththe hypothesisthatthe elevated estradiollevelsassociatedwithPFOA treatmentinhibitthe C-17/20 lyassenzyme and therebydepresstestostemnelevelsC.ook etal.suggested thatthe blunted responseofLH tolowtestosteronemay be mediated,inpart.by elevated astradiollevelsA. subtlehypothalamicor pituitareyffecmtay alsobe present. however. The mechanism forthe estradioellevatiownas not stucried. Perfluorodemnoicacidaltersreproductivheormones inmale ratsina fashion similatroPFOA- Inmale ratstreatedwithdoses ofPFOA rangingfrom 20 to80 mg/kg,givenas a singleipdose,PFDA decreased plasma androgen levelsina dose dependent fashion65.Both plasma testosteronasnd -@-alpha dihydrotestosteronwere significantrleyduced.Compared to ad libitumfed controlratvalues,mean plasma testastemne was decreased by 88 percentin PFDA treatedanimalsand DHT was decreased by 82 percent These changes were reflecteidnaccessorysex organ weightand histologyT.he changes in accessorysex organsafter-PFDAadministratiownere found to be reversedby testosteronreeplacement.The PFDA decrease inandrogens was the resuftof decreasedresponsivenessofLeydigcellsto LH. There was no evidencefor akeredmetabolismoftestosteroneA.dditionallpyl,asma LH concentrationsdid notincreaseappropriateliynthe face oflow plasma testostemneconcentrations. ThissuggeststhatPFDA may afterthe normal feedbackmechanisms ofthe HPG axis. h isofinteresttonotethat2,3,7,8tetrachlorodibenzo-p-dio(xTiCnOD), which, likePFOA, isa nongenotoxicratcarcinogen,a peroxisome proliferatorasn,d an inducerofP-450 system,has been shown to produce hormonaj effectsinmale ratssimilatrothose observed forPFOA and PFDA. Moore eta).71stucrietdhe effecotfTCDD on slemidogenesisinratLeydigcellsE.xposure ofcelltoTCDD resulteidndepressiontestosteroneand 5-alpha-DHTconcentrationswithout alterinLgH concentratioonrtestosteronemetabolism.Moore concluded that TCOD treatmentinhibittshe earlyphase ofthe synthetipcathway and the mobilizationfcholesterotlo cytochrome P450scc. However, Moore et c-d. observed decreased estradiolT.COD has been shown to increase the estrogen mediated feedback inhibitioonf LH secretion72 Additionallyi,nstucriesusing MCF-7 breasttumor cellst,he antiestrogeniecffectofTCDD was mediated by afterationisnthe cytochrome P450 metabolism ofestradiol73. The decreased testosterone in ratscould be mediated by the effectof TCDO on Leydig cells directlyb,y alterationisntestosteronemetabolism,or through increased negative feedback atthepituitaroyr hypothalamiclevel.Recenty, reportsfrom occupationalstudiesofTCOD exposed workers have associated TCOD exposure with.hormonal alterationsin human males. Egeland et al.74 reported that men with high TCOD levelshad significantldyepressed serum testoslerons levels. The changes intestosterone were not associated with altered LH values. Esiradiolvalues were not reported. They concluded that dioxin has a similar effectsin men and male rodents. The obvservations that PFOA, PFDA. and TCOD have overlapping spectrums of mdent toxicitiesuggests that pemxisome proliteratorsi,nducers of the P-450 system and non-genotoxic carcinogens may also ajterthe hypothalamic -pituitary-gonadfunction in male animals. 2.7.2 Female Rer)roductiveToxicities Inthe two year mt feecringstudy, female ratstreated with PFOA were observed to have an increased number of mammary fibroadenomas compared to control animals. Allmammary carcinomas occurred in controlanimals. Hyperplasia of the ovarian simma was observed, but specifichistopathologicalstudies were not reported*. No Informationisavailableconcerning the effectof PFOA and PFOA on HPG axis in women or female animals. 2-7-3Thyroid Toxichigl Afteredthyroidhormone dynamics have been observed in rats exposed to PFDA 73'78.A singleip dose of PFDA in rats resuftsin a rapid and persistent decrease inthyroxin(T4) and T3 78. Gutshall reported that the decrease in thyroid hormones occurred as early as eight hours aftertreatment and persistentforat least90 days 72.These changes were associated with a hypothyroid-like state in Flepor3LM Rker LaboratoriesT.wo Year OralToxicfty/CarrinogenkfStyudy ofFC143 inRas r2alCROO12,1953. 20 thetreatedrats.The afterationisnserum thyroidlevelsoccurtedatdose levels thaldid notproduce a hypothyroisdyndrome 78. Animals withdepressed T4 levelswere foundto be metabolicalleyuthyroid77. Replacement ofT4 resultedin norrnalfood intake,butdidnotreversethe hypothymid-likseyndrome of hypothermiaand bradycardia76. ThissuggeststhatPFOA has a marked effect on cellulamretabolismthatisindependentofitseflecton thyroidhomeostasis. The low T4 was thoughttobe a resultoftwo mechanisms. FirstP,FDA readily displacesT4 from albuminwhich resultsinincreasedmetabolicturnover ofthe hormone. Second, theresponseofthehypothalamic-pftuitary-thy(rHoPiTd) axis appeared tobe depressed as assessed by thymtmpin releasinghormone simulatiotnestingn. Inthesestudiest,he animalshad increasedlevelsof thyroidresponsivehepaticenzyme activitiseusggestingthatthe PFDA treated ratswere notfunctionallhypothyroid.The histologicaplpearance ofthe thyroid glandswere unremarkable,althoughtreatedratshad significantlloywerthyroid weights.TSH levelswere notstudied.No similasrtudiesare availablfeor PFOA. PFOA has been notedtoproduce a transienwteight lossintreatedrats 30. The hypothyroid-liskyendrome observed inPFOA treatedratshas notbeen stucfieidnPFOA treated@ats,however. Sincethe thyroidhormone effectsof PFDA do notcause the hypothyroid-liskteateinrats,PFOA may alterthe HPT axiswithoutproducingthissyndrome. 2.7.4 Hegatic loxieffies The primarysiteofPFOA toxicitiynrodentsisthe liver.Peroxisome proliferation (PP),inductionof enzymes involved in 8-oxidationof fattyacids,and inductionof cytochrome P450 occur aftera singlePFOA dose. Marked hepatomegaly has been noted coincident to the PP and enzyme induction. Increased liversize was the resuh of a combination of both hypertrophy and hyperplasia. Cell hypertrophy predominated afteran initiablurst of cellproliferationT.he inital hyperplasia is evidenced by large hepatocytes and markers of DNA synthesis Areas of increased necrosis in the pe6portal regions have been observed The relationshipbetween hepatic enlargement peroxisome pmiifer-ationa,nd increased i3@-oxidatioinsunclear. Xenobiotic induced changes in one specific pemxisomal enzyme are not necessarily linked to changes in other peroxisomal 21 enzymes or hepaticenlargement 82. Stucriehsave suggested thatxenobiotc induced hepatomega)y and PP may be relatedtothe endocn*ne statusof experimentalanimalsorto oxidativestress90.93-81.Adrenal and thymid hormones may playa roleinperoxisomal proliferatio8n0.-25. Studiesof ciofibratea,PP, have shown thatendocrine manipulationcan mocrityftshepatic effects. Inadrenalectomizedand thryoidectomizedrats,ciofibrate-induced hepatomegaly was reduced compared tothe effectincontrolrats83,". Conversely,inthyroidectomizedor hypophysectomized rats,ciofibratienduced peroxisomal 3-oxidatioennzymes were increasedcompared to normal rats93. Thoitasseryetal.compared the PFOA-induced hepatomegaly in normal rats, adrenalectomizedratsand adrenalectomizedratswithcortisarleplacement go. They found thathepatomegaly was cor*jsodlependent and was primarilya resun ofhepatocytehypertmphy. Hyperplasticresponses were also cordsofdependent and were noted inperipor%arjegionsof the liver.Peroxisomal pmliferaliodnid not depend on cordsofand was observed incerrtdlobularregions.They concluded that PFOA-induced hepatomegaly and pemxisome proliferatiwoenre separate pmcesses. inoralteedingstudies,PFOA and otherPP were reportedto cause increased hepatomegaly inmales compared to females. This differencecould be reduced by exogenous estradioaldministratioonr castrationand eliminatedby castration and estradioaldministratio"n. These observationsmay be explainedby an estrogendependent renalexcretionmechanism or a testostemne mediated increasein tissuebinding$5,$7. 88 lssemann and Green have cloned a mouse PP activatedreceptor.MPPAR, a member ofthe nuclearhormone receptorsuperfamilyof ligand-acbvated transcdptiofnactorsthatisactivatedby pemxisome proliferator8s9.This receptordirecgymecftatestheeffectsofperoxisome pmliferators(PPs). Tugwood has shown thatPPs activatedPPAR recognizes a specificresponse uniton the Acyl-CoA oxidassgene promoter in a manner similarto thestemid hormone receptor90.The actionof PFOA and otherPPs may be mediated by a familyof cytosolicreceptorsthatregulategene transcdpton ina manner similar toothernuclearhormone receptom. 22 2.7-5Nonnenotoxic CaMinonene Ininitatiosne,lectiona.nd promotionexperimentsinrats,PFOA produced an increasednumber of hepatocellulcaarrcinomas 11.12 Severalmechanisms for PFCA associatednongenotoxiccarcinogenesishave been suggested. Perfluorooctanoiaccidisan archatypajmember ofa unique sub classofPPs that are not metabolized.Reddy has argued thatthe structura)dliyverse;>emxisome proliierato(rPsP)area distinctlassofnongenotoxiccarcinogensas.Reddy proposed thatPPs induceoxidativsetresswhich resultsinincreasedtumor formation. Accordingtothistheory,theobserved increaseinhydrogen peroxide formationassociatedwithincreasedB-oxidatioinsnot associatedwithan increaseofsimilamragnitudeindetoxifyincgatalaseactivit"y. Oxidabye attack by hydrogen peroxideand otherreactiveoxygen specieson cellconstituentasnd membranes leadstoDNA damage and increasedcellproliferatioInn.creased proliferatiionnconcertwithDNA damage produces increasedcalltransformagon and malignancies. StudiestestingthetheorythatPFOA induces HCC by increasingoxidativsetress have leadtoconflidngresults.Takagistal.observed an increaseinahydmxydeoxyguanosine inliverDNA fmm ratsexposed toPFOA. They concluded thatrathepatocyteswere under increasedoxidativestress93 Handleretal.found no increaseinhydmgen peroxideproductioninirdacWters exposed to PFOA 94. Lake et al.%iledto findan associationbetween hepatic tumor formationand pemxisome proliferati9o6n.Thottasseryetal.observedthat the PFOA induc@on of3-oxidatiownas independent of adrenalhormone status. A PFOA associatedincreaseincatalaseWivity depended on cortiso8l0. Therefore,the hormonalstatusinanimajs used inexperimentscould confound studiesofoxidativsetressand account forthe conflictinrgesufts. 2-7-6lmmunotoxicity Inthe 90 day monkey feedingstudy,bone marrow and lymphaidtissuewere a siteofhistopatholog9y.Treated monkeys inthe highesttwo dose groups were observed tohave moderate hypocelluladtoyfthe bone marrow. Specific 23 hiSIpOatohlogirfailndinwgesrenotreporteAdt.rophoyflymphoifdolliciines lymph nodes and thespleenwere notedinthe sam'e'treatmengtroups. No follow-upstudiesoftheseobservationshave been reported.StudiesinPFOA treatedratshave notshown histologiccahlanges inthe immune system 7.7 Mechanisms ofAction The mechanism oftoxicitoyfperfluorinatseudrfactantmsay be mediated by their effecton cellmembranes. Olson and Andersen = suggestedthatPFOA may aftermembrane functionthroughchanges infattyacidcompositionand oxidabon siatusL.evittand Uss hypothesizedthatthe effectofperfluodnatedsurfactantiss mediatedby theirafteratioonfmembrane organizatioonr fluidi9t6y,97. Shindo32 reportedthatmiscibiliotfyfluorocarbonand hydrocarbonsurfactants depends stronglyon carbon chainlength.A carbon chainlengthgreaterthan eightcarbons isnecessaryforimmiscibilitPye.rfluomcarbonsurfactantwsith eightor fewercarbonatoms are misciblewithhydrocarbonsurfactantwsith carbonchainlengthsup tonine.These observationscouldhave impormm implicatiofnosrbiologicaslystems thalcontainfluorocarbonsurfactantsC.ellular membranes are a phase boundary,usuallybetween a lipipdhase and an aqueous phase. Surlactantwsillsegregatetothisphase boundary. Two iinmiscibsluerfaciantmsay form two coexistentmonolayers on the insideand oumide ofthe membrane whereas rnisciblseurfactantwsillform onlyone such monolayer.The presenceoftwo monolayerswillmaximallyreducethe surface tensionatthe boundary.whereas a singlemonolayer willafted surfacetensionto a lesserdegree. Changes insurfacetensionmay aftermembrane fluidiatnyd affecittsfunctioninsuch processesas signalrecognitioannd transduction.itis intereslitnognotethatthe change inmiscibiliitnyShindo'sexperimentalsystem occurredforfluomcarbonsurfactantwsithcarbon chainlengthsgreaterthan sighlThischange inmiscibilidteypended on hydrocarbonsurfactancthain lengthas well. The effectsofPFOA and PFOA on experimentalmembrane systems and cellular membranes have been investigated.Inoue studiedthe differentieaflfectsof oclanoicacidand perfluomocianoicacidon experimentalcellmembrane 24 properties98.The phase transitiotnemperatureofdipajmhoylphosphatidyicholine vesiclesdecreased linearlays PFOA increasedinconcentrationup to one mM and then reacha plateau.This suggested thatPFOA may form aggregates in the membrane above a criticcaolncentration.Such a phase separationis observed tooccurinmicelles32. The partitiocnoefficienbtetween water and the membranes forPFOA, K - 8910, was largerthan the coefficientfor ionized ocianoic acid,K - 135, possibly because ofthe differencein hydrophobicity between hydrocarbon and fluorocarbon chains in aqueous solution. The differencesbetween the toxicokineticsand toxicodynamics of PFOA and PFOA may be the resultof theirdifferingmiscibilitiewsith callmembrane surfactants. Levittand Uss investigatedthe effact of PFOA and PFDA on the plasma membranes of callsfrom F4 human B-lymphoblastoid celllineusing the dye memcyanine 540 (MC540) 97. The dye binds to phospholipids that are loosely packed on the outer cellmembrane, but does not bind to highlyorganized lipids and does not penetrate the membrane of healthy cells99. A large decrease in MC540 cellsurface binding was observed aftertreatment with sub-lethal concentrationsof PFOA and PFDA but not other non-perfluodnated fattyacids. Albumin or serum reduced the change in MC540 binding. This effectmay be a resultof the strong protein binding of PFOA and PFDA by albumin 50.These observationssuggest that PFOA and PFDA eitherinteractdirectlywith MCSQ lipidbinding sitesor afterthe structureof the lipidsinthe membranes. Inexperiments examining functionalchanges in the lymphoblastoid celllines, Levittand Uss observed that PFOA and PFDA could cause crirecdtamage to cellsresuftinginthe release of membrane bound cellproteins and immunoglobulins insoluble form 96. PFDA was significantlmyore potent than PFOA insolublizingproteinsand killincgells.This may be the resultof different miscibilitiiensthe cellmembrane of these compounds. However, neither PFOA nor PFDA reduces the abilitoyf surface immunoglobulins to migrate and undergo capping afterantigen recognftion 17;. Inthe PFOA concentration ranges that decreased MC540 binding,PFOA did not affect immunoglobulin migmton and capping. Capping involves the cytoskeletalmediated polar migration of immunoglobulins withinthe plane of the membrane 100. Apparently, the PFOA 25 and PFOA associatedmembrane changes do notaffectmembrane characteristitchsatare importantforreceptormigration. The membrane effectosfPFDA have been studiedingreaterdetail.Pilcheretal. reportedthata singleinjectioonfPFOA inr-atssignificantrleyduced the apparent number of B adrenergicreceptorsincardiaccells101.This change innumber of receptorswas reflecteidnthe diminishedresponseofadenytylcyclase(AC) to epinephrineinPFDA treatedratcardiaccellsT.he intrinsipcropertiesofAC were notaltered.The actionofPFOA was on the epinephrinereceptor.The fattyacid compositionofthetreatedratcardiaccellmembranes was significantalfytered 101.Paimitic(16:0)acidwas elevated13 percent,eicosotrienoi(c20:3WS) was elevated71 percent,and docosahexaenoic acid(22*.w63) was elevated18 percent.Arachidonicacid(20:4)was reducedby 18 percent.Severalother investigatorhsave reportedchanges inmembrane functionfollowingPFDA exposure.Wigierand Shaw 112demonstratedthatPFOA inactivatead membrane transportchannelfor2-aminopurineinL 5178 Y mouse lymphoma cells.Invitreoxperimentsreportedby Olson etal.103showed thaterythrocytes exposed toPFDA exhibitedecreased Osmotictragiw and increasedfluidity. Taken togethert,hesestudiesindicattehatperfluorinatesdurfactantesxerttheir effectosn cellmembranes. The effectsappear tobe limitetdothe outerportion ofthemembranes as the resultofdifferentipaalrtitioniwnigthinthe membrane or bindingtospecifimcembrane constftuentsA.lthoughPFOA and PFDA can be cytotoxiacs a resulotftheirdetergentactionon membranes, theirmembrane effectastlowerdoses are not reialedtotheirdetergentaction.From available data.itappearsthatfunctionamlembrane changes may be limitetdospecific receptormediated functions. 2 R QccunationalFluorineFxl2osur@sAt Chemolits Inworkersemployed influorochemicaplroductionplants,blood organicfluorine has faroutweighedionicfluo@de8.1Z 14,51.56M.ore than 98 percentofthe total fluo6neinthese groups has been reportedtobe organicfluorineT.herefore,the use oftotalfluoridleevels,which consistpredominantlyoforganicfluorine compounds, isa validsurrogatefororganicfluorinienoccupationalleyxposed groups.Inworkersatthe Chemolhe plant.PFOA has been identifieidnthe serum 26 ofthese workers and was es-bmatedto account for90 percent of organicfluorine found inthe serum samples 8.Inthiscohortofworkers,totalfluorineisa good surrogatemeasure forPFOA. Industriahlygiene measurement offluorochemicalshave been conducted at the Chemolits plarrstincethe 1970s 8.These measurements inciudearea samples, personalbreathingsamples and surfacewipe samples. In1977, a comprehensive efforattevaluatingfluorochemicalexposures was conducted at theChemolite plant.Duringcertainoperationsbreathingzone PFOA concentrabonswere as highas 165 ppm. Afterextensiveengineeringcontrol alterationtsh,e plantwas seriallrye-surveyed.In general,airborneexposures were below the recommended limitof 0.1mg/m3. However, there was evidence ofsurfacecontaminationinproductionbuilcring8s. In 1986, airbornePFOA, as wellas breathingzone samples were lessthan 0.1 mg/m3 based on 8 hour time weightedaverages.Levelsas high as 1.5 mg/m3 were measured inbreathing zone samples duringcertainclean-upand maintenance zone samples. Perfluorobulydaccidwas also found,butin much lower concentrations.Spray dryeroperatorshad cons.i.stentlhyigherexposures, even followingextensive equipment improvements. n appears thatairborneexposure to PFOA was low formost workers.Spray dry 'operatorasnd workers involvedincioan up and maintenance activitiehsave higherintermittenetxposures. Although personalprotectiondevices are required inhighexposure jobs,worker compliance has not been evajuated.The rolethat surfacecontaminationplaysinworker exposure has not been defined'. The mute ofPFOA exposure inworker has not been cleariyidentified. 2.9E-r)idemiologicEatludies A retrospectivceohortmortalitsytudy ofemployees atthe Chemolite Plantinthe periodof1948-1978 was conducted by Mandel and Schuman 2.Of the 3,688 male employees who were employed foratleast6 months, 159 deaths were identifieTdh.ere'was no excess mortalityinthe employees as compared to all Personaclommnication fromStan Sorenson,3M CorporateMecricaDlepanffwrd personacloffmnicawn imm Stan Sorenson,3M CorporateMecricaDlepartnvnt 27 cause orcause specifimcortalitiynthe U.S.whitemale population.The subcohortof allchemicaldivisiownorkersdidnotshow any allcause or causespecifiecxcess inmortality. Startinign 1976 medicalsurveillanceexaminationswere offeredto Chemolite employees inthe Chemical divisio0n.Approximately90 percentofthe workers participateidnthe program.No healthproblems relatedto the exposure to fluorocarbonswere encounteredinparticiparrStesr.iallcyonducted surveillance examinationshave faed torevealany relationshibpetween blood levelsof organicfluorinaend clinicaplathology 2 10 SLimrnary Animalstudieshave suggestedthatthereare fiveareas oftoxicitayssociated withPFOA exposure.These includehepatotoxicitiym,mune system aiterabons, reproductivheormone alterationLse,ydigcelladenomas, and non-genotoric hepalocarcinogenic@yT.oxicitystudieshave primariluysed rodents.There is considerablveariabilibteytween strainsofratsforsome ofthe toxicendpoints such as Leydigcelladenomas. Adcritonallsyo.me ofthe effectsseen inmts have notbeen seen inotherrodentspeciessuch as mice,hamsters or guinea pigs. The limiteddata availabloen PFOA exposed rhesusmonkeys and occupationally exoosedworkerssuggeststhatany extrapolationfthe resultsfrom mdent experimentstohumans requiresmdie inforiiiataobnoia the mechanism ofPFOA toxicityF.rom thisdataitdoes not appear thatthe liverisa major siteforPFOA toxicitiynhumans. Of greaterhuman healthconcern are the potentiaelffectasn theimmune system and the reproductivehormones. Inthepasi,workers have been found tohave significanbtlood levelsofPFOA. Many workershave levelsabove one ppm. These blood levelsare 50-1000 timesbackgmund levelsinthe generalpopulation.These levelsmay be high enough toproduce toxiedieisnoccupationalleyxposed humans. A confident estimateofriskcannot be made untiflurtherinformationon the adverse health effectosf PFOA exposure in humans isobtained. Personaclomnwnicatiolnmm LarrZyobel3;M Corporatimoen(ficaDlepannwr4 -28 bd;zTHODS q j introduction The effectsof perfluoro=anoicacid(PFOA) exposure on human healthwere studiedinemployees ofthe 3M Chemolitsplant(hereaftereferredto as Chemolite)locatedinCottage Grove, Minnesota. Two studieswere conducted to investigatoefthe human healtheffectsassociatedwith PFOA exposure. First, mortalitayssociatedwithoccupationalPFOA exposure was studiedusing a retrospectivceohortdesign. Second, a crosssectionalstudy design was used to estimatethe relationshipbsetween PFOA exposure and selectedphysiologic parameters. A retrospectivceohortstudywas designed to examine mortalityamong workers. Ailworkers ever employed atthe Chemoliteplantforgreaterthan sixmorrthswere includedinthe cohort.AJIcauses and cause-specificmortalitywere compared to expected mortalityE.xpected mortalitwyas calculatedby applyingsex and race specifiqcuinquennialage, calendarperiod,and cause-specificmortalitryatesfor the UnitedStatesand Minnesota populationsto the distributioonfobserved person-time104- 'Os.Age adjustedstandardizedrateratioswere calculated106.A relativreisk(RR) forPFOA exposed workers compared to unexposed worksm was calculatedusingproportionahlazard regressionmodels 107. "The RR were stratifibeyd gender and adjustedforage atfirsetmployment. du@atio-o-fn employment and calendarperiodoffirsetmploymenl Any significandtifferences between observed and expected cause-s;>eciflmcortalitywere tobe explored usingnestedcase controlstucries.Case studieswere completed forcauses of deathwith5 or more deaths and standardizedmortalityratesgreaterthan 1.5. Each deceased individual'rsecordwas examined forcommonaftles injob history inforinatiionncludingage atfirsetmployment, calendar periodof employmem yearsinthe Chemical Division,and durationof employmenl Selectedphysiologiceffectsof PFOA exposure were studiedusing a cmss sectionasltudy design. The relationshipbsetween totalserum fluorinaend biochemicalparameters includingreproductivehormones, hepaticbiochemical parameters,lipidand lipoproteipnarameters,and hematologic parameters,were 29 explored.A sample ofthework forceemployed on November 1, 1990 was invited toparticipateA.llemployees inhighexposure jobs we'*raesked to participateA. sample ofworkers employed inlow exposure jobswas frequency matched to the age and sex distributionfthehigh exposure group. Each participanctompleted a questionnairewhich includedmedical historyand informationconcerni*ng alcohol,tobacco,and medicationuse. The questionnaireisprovided inAppendix 3-1. Blood was drawn fordeterminationofhematologicand biochemical parameters.Totalserum fluorinef,ree(Fr)and bound testoslemne (Bl@. esiradio(lE).thyroidstimulatnghormone (TSH),folliclsetimulatinhgormone (FSH),prolacti(nP)and luteinizihnogrmone (LH)were assayed. The PFOAhormone dose-response relationshifporeach hormone was estimated using linearregressiontechniquesto adjustforthe effectsofage, sex, body mass, alcoholconsumption.tobacco use, and otherpotentiaclonfounders. The PFOAhortnonedose-response relationshiwpas furtherexploredby fittinlginear multvariatemodels to hormone ratios.Allunique ratiosbetween the seven hormones were defined.Twenty-one hormone ratioswere calculatedforeach participanTth.e prevalence ofhormone valuesoutsidethe laboratoryreference range formen was compared to the expected prevalence assuming a normal distributifoonrassay values. 3.2Retrospecttv@Cohort Mortaft Study S.2-1DefinitioOnf Th@ Cohort The Chemolfte faciliotpyened in 1947. Individualwsho were employed atthe Chemoliteplantbetween January 1. 1947 and December 31, 1983 were identified fromcompany records.Workers with fewer than sixmonths employment were excluded.InOctober 1951 largescale commercial PFOA productionfacilities became operational(Abe 1982). Because largescale PFOA productioncid not beginuntil1961, a second cohortwithpotentiallsyignificanPtFOA exposure was definedas those workers employed between October 1, 1961 and December 31, 1983. Subjettswithgreaterthan sixmonths employment were includedinthis second PFOA cohort. 30 The cohortwas initialalsysembled in 1979. Subsequently,the cohortwas updated to includenew employees through 1983. Pamonnel recordsfor employees working priorto 1979 were coded to@demographic hems and work historyby trainedabstractorsC.ompute6zed corporatepersonnel databases were L4ilizetdoprovideinformatiofnorworkers employed inthe 1979 to 1983 period. Abstractedwork historyincludedyear offirsetmployment, year of last OMPIOYMGnt, years employed atChemolfte,and months worked inthe chemical division.Individuajlob historiewsere not absiractedbecause job titlewsere definedby wage grades and did not correspond to s;>ecifjiocbs or locationswithin the plant. 3.2.2Study Datah2sps And Files A Chemolitecohortdatabase was created on a VAX computer using ingres software.Data storedon magnetictape were transferredtothe VAX Duplicate recordswere identifieadnd removed. Missing data were identifiedT.he Ingress update functionwas used fordata editing.Finalanalyticfilesforthe Monson program, SAS programs,and custom programs were constructedusingthe Ingressreportwriter. 3-2.3Data E-diting The lngresrelationadlatabase allowed extensiveinternalconsistencychecks to be made. Alldateswere checked forplausibilitTyh.ose recordswithimplausible, inconsistento,rimproperlyformatteddates were editedand correctedif informatiownas available.Ftecordsof workers withfewer than sixmonths employment were flaggedand excluded from the analyticdata set. A random check of 50 ofthe364 workers withfewer than sixmonth employment found no errorsinclassificatoonfemployment length. Extensiveattempts were made to obtainallmissingdata items. Sources of informationIncludedplantpersonnel records,corporatepersonneldatabases, t>enefirtecords,archivedcorporaig records,plantmedicalrecords,and death certificatesN.o individuaelmployees or nert-of-kiwnere contacted.Four employees were excluded from thecohon as a resultof missingdemographic data hems. 31 3 ?,,Ydalid2:@Oofn-i-n-e-HistCno@hcnarlt-l-nfnrmation 1-? A-1 A5r.-ossrTieOnftC rnnletenessOf AcceMainrnent The cohortwas initialdleyfinedfrom personnelrecordsstoredatthe Chemolite plant.Complete recordswere maintainedon allworkers ever employed at the plant.Houdy and saladed workers were incjudedinthese filesa,s were all transferredt,erminatedand retiredformeremployees. Records forworkers first employed inthe 1947-1978 periodwere abstractedfrom documents, coded and computerized. A corporatecomputehzed database was used to update the cohort through December 1,1983. Since insufficieinntductiontime had lapsed between 1983 and 1989, no new employees orwork historyinformationwas added to the cohortdatabase forthe post 1983 periodforthisstudy. Verifyingtheascertainmentofalleligiblceohortmembers was problematic.The assumption thatthe personnelrecordsrepresenteda complete mster was difficult to check because ofa lackofindependent information.Several sources were used toexcludemajor errorsinthe enumeration ofthecohorl The historicapllant hiringpatternbased on senioritdyates was compared withthe distributionf dates offirsetmployment. Qualitativeldya.tes ofmajor plantexpansion correspondedto peaks inthe distributioonf dates offirsetmployment and to senioritdyates. Large increasesinhiringdue to new plantopenings were reflecteidnpeaks inthedistribwanof startindgates inthe cohorl A sample of 25 annuitybeneficiariersetirefdrom the Chemolits plantwere obtainedfrom the corporatepersonneloffice.Ail25 were found to be includedinthe enumerated cohorl Severalplantpersonnelrecordsystems were randomly sampled. Separate files were maintainedforactiveworkers, retireest,ransferredand terminatedworkers, and workerswhose employment at Chemolits ended priorto 1960. A sample of recordsforcurrentemployees withstartdates priorto December 31, 1983 was compared tothe cohort.AJI12 records from the 194-r--1960periodforstartdates were found inthe cohortdatabase. Of 30 recordssampled from the 1961-1969, 28 (93%) were includedinthe cohort. Fiftytwo recordshad startindgates inthe 1970-1978 pedod. Of these52 records.fortyseven (90%) were found inthe -32 database.Inthe1979.1gso period18 of44 (41%)recordswereinthedatabase. Lasty,inthe 1981 through 1983 period,36 of37 recordswere intho database (97%). The low ascertainmentforworkers firsetmployed inthe 1979-1980 period was furtherexamined. Of the 34 workers notinthe cohortdatabase, 16 (47%) were firsetmployed inthe 7/79-1/80period.These omissions occurred inthe transitiopneriodbetween document abstractinagnd electronicupdatingof the cohort.Using senioritlyists4,4 workers currentleymployed were hiredbetween 1979 and 1980. They representapproximately1% ofthe totalnumber of individualisnthe workforceand lessthan 0.5% of thetotalperson time at riskfor thecohort.Records forretiredworkers were sampled from filescontainingall workersretirefdromChemolfte. Fortyseven ofthe 48 (98%) sampled records were presentinthedatabase. A sample ofthefilescontainingthe personnel recordsofemployees completing employment before1960 was randomly drawn. Of the67 selectedrecords,65 (97%) were inthe database. Finallyf,iles containingrecordsofalltransfertedt,erminated,or disabledemployees were randomly sampled. Of the 120 sampled records,116 (970/ow)ere presentinthe cohortdatabase. 3.2.4-2ValidationOf Coho rtInformation Informatioinnthe editeddatabase was compared to informationinthe personnel records.A random sample of25 recordswas drawn from the personnelfiles. Database n"ames,'sbcia@l@e6jfrtnyUtnbeir:@'(SSNd)a,tes of birth(DOB), and dates ofemployment were Yerifiedagainstrecordinformation.The sole erroroccurred incodingthe lastdigitofone SSN. Allotherinformationwas correctlyentered intothe database. The reliabiloiftyICD8 =cring ofdeath certificatfeosrunderlyingcause of death was evaluatedby resubmittinga sample ofdeath certificatefsorcocringby the sarnenosologislThe sample consistedof 25 death certificatfersom 1970 -1989. No change inthe major categoriesofcause ofdeath was noted. Allcancer deathswere coded concordantly.Withincardiovascularcauses ofdeath,two certificatweesre discordant. 3-2-5VitmiSta us Ascprtpinment 33 The vitasltatuswas ascertainedfrom the SocialSecurityAdministration(SSA) and the NationalDeath Index (NDI).Allindividualwsith unknown vitalstatuswere tracedsuccessfullyand vitalsiaiusdetermined. Vitalstatusdeterminationinthe 1979-1989 periodwas obtainedthroughthe NDI. Death cer6ficateswere requestedfrom the appropriatestatehealthdepariments forthose individuals identifieads,or presumed to be, deceased. A professionalnosologistcoded the death certificatfeosrunderlyingcause of death accordingto Intemational ClassificatioofnDiseases,8th revision(IC08). Informa@on concerning the date and cause oftwo deaths which occurredoutsidethe United Stateswas obtained from familymembers or otheravailablesources. Date of death and the ICD8 code forthe underlyingcause ofdeath were entered intothe database. 12 6 Vpfid2t!cngfVitqlStatus Agcgrtainme The vitasltatusdeterminationprocedures forthe cohortwas evaluated. Corporatebenefitrecordswere utilizeads an independent source forvitalstatus among theretrees.Vitalstatusfrom the database was compared tovitalstatus incorporaterecords.A listofallretireeisnthe 1947-1984 cohortwas sent to3M benefitdsepartment These individualwsere matched to retireewsho had received3M death benefits.3M recordswere not complete forpedods priorto -1975.Inthe pri@-1983period,4 deaths in refirseswere identifiebdy 3M records. -ViW status-wascorrectlyascertainedby the SSA matching procedureforonlyone oftheseretirees.Inthe 1983-1989 period,34 deaths inrebroes were identifieidn 3M records.The NDI matching procedureascertainedall34 ofthese deaths. The NDI was notavailablefor1990. 3M re=rds indicatethat8 retireecsried during1990. The incomplete SSA aseartainmeritinthe period1975 to 1983 resulteidnextendingthe NDI search to include1979 to 1983. All3M ideftfied deathswere also identifieidnthe subsequent NDI search coveringthe 1979 to 1983 period. 3.2.7Anafvsis Analyticmethods employed inthisstudy were appropriateforcohortstudies.The relativreiskwas estimated by calculatingan adjusted standarcrizemdortwityratio 34 (SMR) 105.Thisstudyused bothnationaalnd Minnesota mortalitryatesfor comparisons. Mortalitfyorman inthe Chemolhe cohortwas compared to expected nationaland Minnesotamortalitya,djustedforage, calendarperic@ds,ex and race.The use ofmonalityratesintheruralcountiessurroundingthe plant were notconsideredto be stableformany causes of death and were not.used. Sincelessthan one percentof plantemployees are non-white,white maie and femalerateswere used forcomparison.For women, only U.S. rateswere used because cause-and calendarpehod-specifiMcinnesota rateswere notavailable. SMRs were calculatefdorallcause,allcancer,and cause-specifimcortalityT.he effectosfdiseaselatency,durationofemployment, durationoffollow-upa,nd work inthe Chemical Divisionwere examined usingstratifieSdMR analyses. Three additionamlethods ofanalysiswere used toassess the validitoyfthe SMR contrastsT.he threemethods were:standardizedrateratios(SRR) 106,Mantel Haenszoladjustedrelativreates(RRMH) 106,and proportionahlazard regression adjustedRR 107. Umited exposuredata were availablferom plantrecords.Exposed workerswere definedas allworkerswho worked for1 month or more inthe chemicaldivision. Exposed and unexposed workers'allcause,allcancer,and cause-specific mortalitwyas compared usingstratfieSdMRS, SRRs 106,and strafifiMeadntel Haenszelanalysis118-109.Additionalltyh.esame summary measures were calculatecdontrastintghe ratesforworkers withatleastten years durationof employmerttand thosewithlessthanten years employmiinl The relativreisk(RR) and 95% Cl fortheRR fordeaths from allcauses,cancer, cardiovasculadriseases,and selectedspecificcauses were estimatedusing a proportionhaalzard model (PH) 107.109.The timeto everttorcene-ohngwas definedas timefrom firsetmployment toeventor December 31, 1989.InPH modelsforspecificcauses ofdeath,deaths from othercauses were censored at thetimeofdeath. Exposurewas quantifiedby months ofchemicaldivision employment. Covahates includedinthe models were age atfirsetmployment, yearoffimiemploymenl and durationofemployment. The analyseswere stratifibeydgender.The appropriatenessofthe proportionahlazard assumptions weretestedusingstratifimeoddels vathgraphicalanalysisoflog(-lo(gsurvival)) 35 versusfollow-uptimerelationshiapnsd models thattestedthe significancoef a productterm between exposureand log(follow-utpime)102- 1-3Cro rgionELSItudy Of PFOA Fxncsed Wc*ers 1-3.1Pnnul2iionr)efinftiAonnd Recruitmpnt Medicalscreeningofworkersemployed attheChemolite plartotccurs everytwo years. The generalmedicalscreeningprogram includeda medicalquestionnaire (Appendix3-1).measurement ofheight,weightand vftwsigns.pulmonary functionevaluationu,dnalysiss,erum assays,and hematology incricesT.his screeningprogram offeredan opportunittyo assess the physiologicoffectsof PFOA exposure inworkersengaged incommercial productionof a limited spectrumofPFCS. Of particulairntereswtere the effectsofPFOA. the primary fluomchemicalfound inthe serum ofChemolhe workers.(Griffiatnhd Ubel,1980). Participatiionthe PhysiologiEcffectsStudy requiredthe subjectsw'illingnestso undergo hormonal and biochemic@Wtesti.nagnd to have an additiona1l5 mi of blooddrawn fortotalfluodneassay. Inthe cross-sectionasltudy,exposure classificatiwoans based on the potentiaflorPFOA exposure ina workersjob and plantlocation.Allworkersengaged inany facetofPFOA productioninthe .previousfiveyearswere consideredto have potentiallhyigh PFOA exposure. Th6 jadtdoftideredtohave highexposurepotentiailncludedalljobsinthe productionbuilding(sbidg6 and 15),allmaintenance workerswho were assigned tothe PFOA productionareas,and allmanagement jobs requirinpghysical presenceinthe productionbuilding.Plantrecordsand job historyinformatiownas used toassignexposure statustoindividuawlorkers. A random sample of workersinjobswithlow exposure.potentiawlas frequency matched to the age and sex distributiofnthe highexposure workers. Workers withlow exposure potentiawlere definedas those assignedtojobs not involvedinthe pmduefion of PFCs foratleastfiveyears.A rosterofworkers meeting the low exposure potentiawlas definedfrom plantrecordsand knowledge ofplantpersonnelabota thelocatioonfhighexposurejobs. A gender stratifiseadmple from the group of workersinlow exposurejobswithan age (5yearstrata)distributisoinmilartothe exposed group was identifieadnd invitetdo participateI.fa workerina jobwith 36 low exposuredeclinedtoparticipataen,otherworkerinthe same age and sex s@ratumwas randomlyselectedand invitedtopartcipate.Inallcases inforrned consentwas obtained.Participatiionnthisstudywas voluntary. Data Collection 3 3-2-1Study Loos And Files A rosterofparticipanwtass maintainedby the plantoccupationalhealthnurses. A logforbiologicaslample informatiownas completed by the laboratortyechnician. The dateand timeofample coolectionwas recorded.Qualityassurancesamples were recordedon a separatelog.Resultsreportedon paper recordswere maintainedas medicalrecords.Resultsforothertestswere transmitted electronicaltloya complnedzed database and coded as SAS datasets.All recordswere storedwithemployee medicajrecordsor inthecorporatemedical officefsorconfidentialpiutryposes. Printedlaboratoryresultsand questionnaire datawere enteredintoa SAS dataset. 3.3.2.2Ouestionnalre Each participacnotmpleteda medicalqueslionnairepriorto reportintgothe plant medicaloffice(.Appencri3x-1) Items includeddemographic information. symptoms, illneshsistoraynd diagnoses,and medicationusage. Detailed questionsconceming tobaccouse and alcoholuse were included.Workers were notre-contactetdo obtainmissinginformationor tocorrectinconsistencies. Ftesponseswere notvalidated.Two plantoccupationalhealthnursescollected thequestionnaireasnd returnedthem tothe corporatemedicaldepartment.Inthe corporatemedicalofficed,ata were coded and enteredintoa SAS data bass. [:;omory ProcedUres 3@3-2-3-1Height and Weicht 37 U Pon reportingtothe plantmedical officep.articipanthsad theirheightand weightdetermined by an occupationalhealthnurse. Heightand weight were measured once on the same calibratedscale. -2,1? Blood @1-22-3 2 1 Drawino And Handlinei Fourva=ainers ofblood were drawn from a singlevenipunctureby a laboratory technologistT.wo 15 mi redtop vacutainersof blood were drawn and allowed to clot.One 10 mi purpletop vaculainerwas drawn forhematology studies.A speciahyprepared fluorinferee15 ml vaclnainerwas used to collectblood fortotal serum fluorindeetermination.Venipunctureswere scheduled to occur at the same time ofday and on the same shiftforeach worker. Allblood was drawn between 6:30 and 8:00 am. Workers inthe Chemical Divisionof the Chemolao plantrotateshiftson a weekly basis.Blood was drawn aftera worker was assignedto the day shiftforatleast3 days. Allspecimens were refrigerateadt the plantpriorto transportto the appropriate laboratoryC.lottedredtop vawainer specimens were centrifugedfor12 minutes toseparateserum from callsbeforetransporttothe contractlaboratory.In order torenderthe totalserum fluorinsepecimens non4nfecflous,serum fortotal fluorinaessays was etherextractedinthe corporatemocricaldepartment priorto sencrintghe samples tothe 3M Chemical Divisionanalyticlaboratories. 3.3.2.3.2-2Assayfi Serurrsoamples were analyzed fortotalserum fluorineh,epaticbiochemical parameters,cholesteroll,ipoproteinsa,nd seven hormones. Assayed biochemical parameters includedserum gildamicoxaloacetictmnsaminase (SGOT), serum glutamaticpyruvictransaminass (SGPT), gamma glutamyltmnsfemse (GGT), and alkalinephospatase (AKPH). The followinghormones were assayed: bound testosteronef,reetestosterone,estradiolp,rolactinl,uteinizinhgormone (LH), follidsetimulatinhgormone(FSH), and thyroidstimulatinhgormone (TSH). EDTA Preservedwhole blood samples underwent routinehematologic analysisincluding 38 complete blood courrtwitherylhrocytoindicesand leukocytedifferenticaelllcount (CBC). Analyses were done withoutknowledge oftftsubjectstatusor purpose ofthe study. Totalserum fluorinweas determined in3M's Chemical Divisionanalyticlaboratory usingthesodium biphenylextractiomnethod (Venkateswariu,1982). The accuratedeterminationoftotalfluorinienthepartsper million(ppm) range requiredspecializedequipment, procedures,and personnel. Assays were completed ina dedicatedlaboratoryfollowingtestedprotocols. Upon receiptofextractedserum samples dividedaliqum were tmzen at-70 degrees centigrade.Afterallsamples had been received,batches of 15 samples were assayed on successiveworking days. Each batch inciudedhigh and low qualitcyontrolsamples. Each sample was assayed twice. H the crifferencien assayed valueswas greaterthan 1 ppm, the sample was re-assayed. The total serum fluorinvealue was reportedas a mean value and a munded integervalue. Serum glutamicoxaloacetictransaminase (SGOT), serum gilgamaticpyruvic transaminase(SGPT), gamma glutamyltransterase(GGT), and alkaline phospatase(AKPH) were assayed by the United Health ServicesLaboratoryin Apple Valley,Liinnesotausing clinicaclolorimetriacssays. CBCs were determinedusingautomated Cauftercounters.Lightmicmscopy was utilizefdor Estradiolp,rolactint,hyroidstimulatinhgormone (TSH). luteinwng.hormone (LH), and follicslteimulatinhgormone (FSH) were assayed by the United Health Serviceslaboratoryusing racriolmmunoassay(RIA)and enzyme linked immunasorbent assay (ELISA). FSH, LH, and prolacfinwere assayed using AbbottlaboratorieIsMX microparticleenzyme linkedimmunoassays. TSH was assayed usingLondon Diagnosticschemiluminescense immunometdc assay. Estr2cgowlas determined using DiagnosticProducts Corporation'sCoat-a-count assay. 39 Testosteronewas assayed by the Mayo CliniclinicallaboratoriesT.otal lesiosieronewas determinedby RIA usingproprietariymmunoglobulins.Free and bound leslosleronweas determinedusingequilibriudmialysis.111. :3.2-2@3@2-3Ouality Assurance Two methods were used to assess the accuracy and reliabiliotfythe laboratory assays. The laboralohesroutinelyfollowedqualityassurance programs. Three standards were run with each batch. Itthe controlvalues were outside two standard deviationsof the intraassay mean value foreach standard.the assay was repeated. If10 controlswere outside1 standard deviationof the mean, the assay was flaggedforreview. The reliabiliotfyeach of these assays was assessed. For each assay, fivespecimens were randomly selected and splitinto two ajiquots.The aliquotswere labeledwith differentidentifieresnsuring thatthe assays were carriedout in a blindedfashion. Both aliquotswere submitted on the same day to the laboratory.The coefficienotf variationwas calculatedforeach hormone. 3.3.3 Anal There were two analyticstrategies.Firsta,ssay resultswere treatedas contnuous parameters and modeled using regression methods. Models were fit to-assess-therelationshibpetween assay.resultsand totalfluorineb,ody man index,alcoholconsumption, and smoking. Second, hormonal assay resultswere dichotomized intothose withinthe referencerange and those outsidethe referencerange. The hormone assay categorieswere based on published sex specificnormal referencevalues foreach assay. The purpose of this dichotomizationwas to evaluate the possibilittyhat highlysusceptibleincrraduals may be affectedat lower levelsof exposure and not followthe adjusted doseresponse curve. The relationshipbsetween totalserum fluorineand the assayed parameters were estimated by fittinlginearmultivariatreegression models to the data. The clinical parameters and ratiosof selected parameters were firstmodeled as functionsof nominallycategodzed exposure and covadates. Dependent variablesthatwere 40 not normallydistributewdere approp4iateltyransformed.Totalserum fluorinewas categorizedintomutuallydistincctategories.Cutoffvalues forthe catecories were chosen to assure adequate numt>er'sineach category while maintainingthe fullesriange ofexposure valuespossible.Accordingly,totalserum fluorinelevel categorieswere definedas the following:lessthan I ppm, greaterthan I ppm tolessthan 4 ppm, 4 ppm to 10 ppm, greaterthan 10 ppm to 15 ppm, and greater than 15 ppm. Ifinsufficiennutmbers ofevents occurred withinindividual categories,the number ofcategorieswas reduced by combining adjacent categories.Additionallym,odels were ftnedwithtotalserum fluorinentered as a continuousvariableusinglinears,quare,square roottransformations. Age, body mass index (BMI),alcoholuse and tobacco use were includedinthe model as potentiaclonfounders. Age was includedinthe models as both a categoricalvariableand a continuousvadable. Age was grouped intofourten year age categories.Age was treatedas a continuous variableusinglinear, square, square mot, and logtransformations.BMI was entered inthe models as a categoricavlariableand as a continuousvariable.BMI categorieswere less than 25 kgIM2, 25-30 kg/m2, and greaterthan 30 kg/m2. AdditionallyB,MI was dichotomized intoobese,*greaterthan 28 kg/m2, and non-obese, lessthan or equal to 28 kg/m2. The cotitinuousvariablewas entered as linears.quare,log, and square transformationsA.lcoholuse was categorizedInto3 categories:less than 1 drinkper day, greaterthan one to 3 drinksper day. and non response to the questionnaireitem. Smoking was categorizedas currentnonsmokers and currentsmokers. A nonresponse categorywas notincludedsince onlytwo individualwsere inthiscategory.These two individualswere excluded from anajysesthatrequiredsmoking history.Smoking was quantifiedas cigarettes smoked per day. Unear, square and square roottransformationsofcigarettesper day were used in regressionmodels. The choiceofthe finalmodel was somewhat subjective.For each dependent variable,othercovariateswere includedinthe finalmodel ifthey were potential confounders. Other potentiaclonfounding hormones and biochemicalparameters were includedinthe models ifthey produced significancthanges ineffect estimates. 41 Totalserum fluorineand confoundingcovahates were entered Intomodels as continuousvariablesS.ignificanntonlineardose-resppnse relationshipwsere evaluatedby comparing model fitand parameter estimatesusing categorical variablesand continuousvariablesS.quare, square mot. exponential,and logarithmitcransformationwsere used ifthetransformed variablespmduced models ofsuperiorpredictivpeower as assessed by model fitAlltwo way interactionbsetween totalserum fluorinaend the includedcovariateswere evaluated. Interactiotnerms were includedinthe finalmodel ifthe parameter . estimateforthe iriteracbotnerm was significanattthe alpha -.1 0 level. The potentiaflorsusceptibilittoyconfound the relationshibpetween PFOA exposure and the assayed parameters was examined by comparing the observed prevalenceofassay resultsoutsideofthe referencerange withthe expected prevalence.The prevalence ofabnormal assays was based on publishedreferencevaluesforthe adultmale US population.Reference ranges fortestparameters were definedas being within2 standard deviationsabove or below the mean valueforthe parameter. The laboratorymaintainslaboratory and assay specificreferencerange foreach assay. Given thatthe distribta!onf valuesisappmximately normal,about 2.5% ofindividuavlaluesare expected to falalbove the upper limiatnd 2.5% below the lower lim!Lh followsthatthe prevalencefora hightestis.025. The prevalence fora low value is.02S. Using ihese provalences,an expected number oftestsoutsideofthe referencerange can be defined. A priorhiypotheses based upon animal and invitrostudies definedthe expected directioonf the effecl The calculatioonfan observed to expected ratioallowedtheestimaton ofthe relativeprevalence fora testoutside ofthe normal range inthe study subjectsas compared to the generalpopulation. The 95% Cl forthe ratiowas calculatedassuming thatthe expected number isa constantand the observed number isa random variablewitha Poisson distribution. 42 4-.-RESULTS A I Cross S-ectionalperfluorocarbo-n-physiologicEffects Study In October 1990, atthetime ofthecrosssectionalstudy,the workforce'at Chemolite consistedof880 salariedand hourlyemployees. There were 50 men and 2 women in highexposure potentiajlobs. Since therewere only 2 women in thisgrolr,,the studywas restrictetdo majes. Forty-eigh(t96%) ofthe SO male workers inhigh exposure potentiajlobs agreed to participateT.he exact number oflow exposure workers invitetdo participatienthe study was not recorded. However, few individualisnthisgroup refusedto participateT.hus, itisestimated thatover 80% oflow exposure workers participated. 4.1.1Particir)aCrhtafr,2ctedstl= Since frequency matching forage was used toselectstudy participantst,he overallage disiribubonreflectetdhe age distributioonfworkers inhigh exposure potentiajlobs (Table4.1.1)A.ges ranged from 24 to 59 years,witha median age of37 years and a mean age.of39.2 years. Table 4.1.2presentsthe alcoholand toba= use profiloefthe study participants. The lightdrinkerscategoryincluded22 participantwsho reportedno alcoholuse. Consumption of one to threeounces ofethanolper day was reportedby 20 (18.7%) participantsN.o participantrseporteddrinkinggreaterthan threeounces ofethanolper day. Eightworkers (7.0%)did notcomplete thishem ofthe questionnaire.There were 28 (24.8%) smokers who smoked an average of21.7 cigarettesper day. Smoking statuswas notavailablefortwo workers (1.8%). The associabon between smoking and alcoholconsumption ispresented inTable 4.1.3.Thirteen(15.3%)of 85 nonsmokers and seven (25.0%) of 28 smokers reportedmoderate drinking(p..24).Table 4.1.4cgsplaysthe age cristdbutifoonr alcoholand tobacco use categories.There were no significandtifferenceisn mean ages among smoking or drinkingcategories. 43 Totalfluo6nweas notsignificacnotrlryelatweidthage.BMI.alcohoolr,tobacco use (Table4.1.5).BMI and age were correlate(dr.@.-P.-2.60.05).AJcoholuse and tobaccouse were notsignifir-anctolryrelate(dr..08;p>.7). SMI ranged from 18.8to 40.5 kg/m2 witha median value of26.3 kg/m2 and a rne2n of26.9kgjm2 (Table4.1.6).Haltofallworkershad BMis between 25 and 30 kg/m2. The mean BMI insmokers was notsignificantdyifferenftrom thatof nonsmokers (Table4.17).The mean BMI formoderate drinkerswas not significantdliyfferenftromthe BMI oflighdtrinkersS.moking statusand BMI were notsignificantalsysociated(Table4.1.8).There was a significalnitnear relationshbieptweenSMI and age (3..10SE(B)..035).Thisrelationshwiaps not substantiajlaylteredafteradjustingforsmoking status,alcoholuse,and total serum fluorinleevel. 4-1.2TotalgerurnFlUoHno The totalserum fluo6nevaluesranged from zero to26 witha median valueof two ppm, a mean of3.27ppm and a standarddeviationof4.68ppm (Table 4.1.9).The inter-assacyoefficienotfvariatiownas 66% calculatefdrom repeated assays on crifferednatys. TworTty-thre(s20.0%)ofIIS workers had totalserum fluorinvealueslessthan one porff.T-hisgroup included-eighwtorkers valuesreportedas zero ppm (below the limitosfdetection)E.ighty-eighwtorkers(76.5%)had levelslessthanor equaltothreeppm. Six(5.2%)of I15 workers had valuesbetween 10 and 15 ppm and five(4.4%)had valuesgreaterthan 15 ppm. Allworkerswithlevels greaterthan ten had worked inBuilding15,the primaryPFO productionarea at the Chemolhe Plant. There were no significadnitfferencesintotalserum fluoridmeean valuesamong the SMI, age,alcoholuse and tobacco use categories(Table4.1.10)N.o statisticaslilgynificadnitfferenceisnmean age between totalfluorinceategories were observed(Table4.1.11). 44 Participantwsithlessthan one ppm totalfluorinsemoked the least(16.3)number ofcigarettesperday (Table4.1.12).Those withone ppm to threeppm total fluohnesmoked thegreatestnumber of cigarettepser day (24-5).This difference was statisticaslilgynifica(npt<.005).As estimatedina regressionmodel, the linearrelationshibpetween totalfluodneand smoking status,adjustedforage and BMI, was smallinmagnitude (3-0.10,SE(B)-0.062,p-.09).Smokers average totalserum fluorinewas estimatedto be 0.1 ppm higherthan nonsmokers. The number of cigarettesmoked per day was weakly correlated withtotalserum fluodne(Table4.1.5). Drinkingstatuswas notassociatedwithtotalfluorine(Table4.1.13).Overall, eight(7.0%) participantdsid notrespond to thisquestion. Four had lessthan one ppm totalserum fluorine. Table 4.1.14presentsthe distributioonf BMI inthetotalfluorinecategories definedpreviously.SMI mean values were notsignificandtifferencesamong the totalserum fluorinceategories.The [!nearrelationshibpetween SMI and total fluorinea.djustedforage, smoking, and alcoholuse, was weak and not significan(t8--.016SE(B@-.069,p>.5). 4-1.3Hormone Assa3M The intra-.*sscaoyefficienotfvariation(CV) forthe bound and freetestostemns. estradiolT,SH, LH, pmlactin,and FSH assays are providedinTable 4.1.15.The estradioalssay had thehighestCV, 18.3%. The pmlacflnassay had the lowest CV of 3.1%. Table 4.1.16 presentsthe observed and expected number of hormone assays outofthe assay referencerange,the observed to expected (O/E)mtio, and the 95% confidencerimitsT.he O/E ratiowas significantglryeaterthan one for estracrioflr,eetestosteroneb,ound testostemne and prolacfinT.he O/E ratiosfor LH, FSH, and TSH were notsignificantldyifferenftrom one. The Pearson correlatiocnoefficientasmong the seven hormones qss:tyedin study participantasre presented inTable 4.1.17. As expected,estradiowlas 46 correlawtietdhtreetest=eron(er..40p,..0001a)nd boundtestostemne(r-.32. p..0006).Bound tesiosieronewas correlatewdithfreetestostemne (r..74 p..0001),LH (r..28p,a.003)and FSH (r..16p,..04).LH and FSH were significantcloyrrelated(r-.63,pm.0001).FSH and TSH were significantly correlate(dr..23p,..01). As shown inTable 4.1.18,totalfluorinweas significantcloyrrelatedwfth pmlactin (r..19p...045)and TSH (r..26,p..005).Age was negativelycorrelatewdith estradio(lr.-.25p,..01),freetestosteron(er.-.45;,@-..0001b)o.und testosterans (r.-.24p,..01),and prolacti(nr.-.19p,..01).Age was positivelCyorrelatedwith FSH (r..33p,-.0003).As expected,SMI was negativelycorrelatedwithfreeand bound testostemne(r..-.26p,..005 and r.-.36,p..0001 respectively)B.MI was correlatepdositivelwyithLH (r...20p,..03).Alcoholconsumption was significantcloyrrelatedwithFSH (r..-.2p4..01). Bound testostemne ranged from 141 to 1192 ng/dlwitha mean of672 ng/diand a median of561 ng/di(Table4.1.19).The standarddeviationswere large.The mean bound testosteronsv-alues were not significantdliyfferenatmong the total serum fluoringeroups. As expected,the mean bound test=emne decreased significantalsyBMI increased.The mean bound testosteronevalueswere significantdliyfferenatmong the age categories(p-.Ol6). Therewas a significannto-nliFiiir-r-elatiob-eitiwietegnp'totW-serum-fluodneand bound testostemne(ST)inthe finalregressionmodel (Table 4.1.20).Bound lestosteronsw,hich was positivelayssociatedwith both LH and estradiol, decreasedas both age and BMI increased.Alcohol and cigaretteuse were wealdyassociatedwithST. There was a significanitnteractiobnetween age and totaslerum fluorineT.here was a negativeassociationbetween bound testostemneand totalserum fluorinienyoung workers than inolderworkers. In workersgreaterthan 45 years of age, totalserum fluodne was associatedvvitah slighitncreaseinBT. The relationshibpetween bound testostemne and tow serum fluoriniespresented forfourdifferenstets ofcova4ate value (Figure2 Dose-responsecurves forbound testosteronewere plottedforyoung, lean individualasged 30 withSMis of25, young obese individualsaged 30 with SMis of35,middleaged lean incrividuaalgsed 50 with BMis of 25, and middle aged 46 obese individualasged 50 withBMIs of 35. Each of the relationshipissfor nonsmoking, lighdthnking men withthe sample mein LH value (5.4mUA) and mean estradiovlalue(33.4pg/mi).In30 yearoldworkers,bound testosterone decreased as totalserum fluohneincreasedinboth BMI groups. The doseresponse relationshifpor40 yearoldworkerswas appmximately flat(notshown). inworkers greaterthan 50 year ofage, ST increasedas totalserum fluorine increased. Totalserum fluoridweas not significanytaissociatedwithfreetestosierone(FT) (Table4.1.21).WithinBMI categories'f,reetestostemne was highestinthe less than 25 kg/m2 group and lowestinthe greaterthan 30 kg/m2 category.The differencienmean FT among BMI categorieswas statisticalsliygnifican(tP-.03). There was a significantonlineardose-response relationshibpetween total serum fluorinaend FT Inthe finalregressionmodel (Table4.1.22).As total serum fluoriniencreased.freetestosteronedecreased. There was a significant interactiobnetween age and iota)serum fluorine.Figure4.2 Illustrattehse modifyingeffectofage on the totalserum fluorinfereelestosteronerelationship. The covadate vectors(nonsmoker, lightdrinker,mean LH and estradjola,ge-30 and BMI..25 or 25,age=50 and BMI-25 or 35) were the same as used Figure-1. Lean orobese 50 year oldmen had low freetestosterons(lessthan 9 no/dl)for allvaluesoftotalserum fluorine.In30 year olds,freetestostemne decreased asymptoticalltyoward the 50 year oldvalues. Inthismodet.-a 50 year old,obese, moderate drinkerwithany totajserum fluorinelevel(thelower limitofassay sensitiviwtays appmximately 1 ppm totalserum tuodne) had freetestostemne below nine ng/cH. As shown inTable 4.1.23,the estradiolmeans in the three BMI groups were not significary#dJiyfferent(p-.88). As the age of participantsincreased, mean estradiolevelsdecreased. Inthe greater than 30 to 40 year age gmup, mean estracriowlas 36.8 pg/mi compared to 25.9 pg/ml in the greater than 50 to 60 yearage group. The age group means were significantclfyiafrferit(p..Ol8). There was a nonsignificanptositiveassociationbetween mean estradioalnd total serum fluorine. 47 As shown inTable4.1-24a,stracrainoditotaslerumfluorinwere positively associatedinthe finalregrosr@onmodel. Totalserum fluorinefolloweda nonlinearrelationshiwpithesiradiolN.o interactiotnerms were statistically significantA.s expected,freetestosteroneand estradiowlere positively associated(B-.85p..0007).The relationshibpetween totalserum fluorinaend estradioilsillustratiendFigure3. The plottedcurvesdepictthe relationshifpor lean(25 kgIM2) and obese (3Skg/m2) male workers who were 30 years oldwith sample mean freetestosteroneand who were nonsmokers and lightdrinkers.As totalserum fluoriniencreasedover the observed range, estracrioilncreased quadraticallyI.nobese men (BMI-35 kg/m2 ) aged 30, estmcriolexceeded 44 pg/miwhen totalserum fluorinewas between 15 and 20 ppm. The highest estradiollevelswere inyoung, obese smokers who consumed 1 to 3 ounces of ethanolper day. LH was notsignificantalsysociatedwithserum fluorineb,utwas negatively associatedwithSMI (p-.003)and positivelayssociatedwithsmoldng (p..025), age, and ST. There was no associationbetween totalserum fluorinaend FT. (Table4.1.25,Table 4.1.26,and Figure4). FSH was notsignificantrleylatedto totalserum fluodnelevelsbut was positively associatedwithage (p-.Ol4) (Table4.1.27.Table 4.1.28).The finalregression model forFSH isIllustratiendFigureS. The relationshiwpas essentiallfyledover thetotalfluadnemngs. TSH was positivelayssociatedwithtotalserum fluorineinboth univariateand multvariataenalyses(Table4.1.29.Table 4.1.30and Figure 7). TSH was not significantrleylatedto age, BMI, alcoholuse, smoking, and other hormones. Proladn was positivelayssociatedwith totalserum fluorineand smoking (Table 4.1.31.Table 4.1.32).Moderate drinkershad a differerpvmtlacdn-totalserum fluorinreelationshicpompared to lightdrinkersand nonrespondents. Figure6 illustrattehse relationshiopf prolactinwithtotalserum fluorineand the modifying effecotfalcoholuse. Totalserum fluchnswas weakjy associatedwithproi=in inlighatnd moderate drinkers.However. inmoderate drinkers(1-3oztday). therewas a positiveassociationbetween prolactinand total'serumfluodne. 48 The univariatdeistributionosfthe 21 ratiosare provided inAppendices 4.1 and 4.2. A tableispresentedforeach ofthe 21 ratiosshowing the number of participantmse,an ratiovalue withthe standard deviation,median ratiovalue, and the range of ratiovalues ineach ofthe previouslydefinedcategoriesof BMI, age,alcoholuse, tobacco use, and totalserum fluorine Correlationbsetween totalserum fluorin(eppm),age (years),BMI (kg/m2), alcoholuse (oz/day)a,nd cigaretteconsumption (cigarettes/daayn)d allpossible ratiosbetween E, freetestostemne TF, TB, and LH are displayedinTable 4.1.33. The esiradiotlo bound testostemne ratio(EITB)and esiradioltofreetestosterone ratio(E/TF)were significantcloyrrelatedwithBMI (rm.32.p-.001 and rn.27. p..004 respectively)T.he estradiotloILrteinizihnogrmone ratio(E/1-H)was negativelcyorrelatedwithage (r.,-.2p6-,.005),and positivelcyorrelatedwfthBMI (r..8l,p..06).The bound testostemneto luteinizinhgormone ratio(TB/LH) followeda differenptatternas compared to EA-H. The correlaliocnoefficient between TBILM and age -was -.32(pm.001)whilethe coefficienbtetween TB/1-H and BMI was -.14.(p..l3). The freetestosteroneto luteinizinhgormone ratio .(TF/LH)had thesimngest correlatiownithage (r.-.40p,...0001)butwas not .significanctolryrelatedwithBMI. The bound testosteroneto freetestostemne ratio(TBn7) followeda unique patternT.B/TF was posifivelcyorrelatedwithage (r..24p,..01).and negativelycorrelatedwith BMI Prolactirnatioswithbound testosterone(TB/P),freetestostemne (TFIP), estradio(lEIP),follicslteimulatinhgormone (FSH/P),luteinizinhgormone (P/LH), and thyroidstimulatinhgormone (P/TSH) are presented inTable 4.1.34.None of theprolactin-hormoneratioswere significantcloyrrelatedwithtotalserum fluorine orSMI. Allexcept P/TSH were significantcloyrrelatedwhh cigarette consumptiorl Table4.l.3Spresents,thePearson correlationcoefficientfsorthe bound testosterontso thyroidstmulatinghormone (TB/TSH) ratiot,he freetestosterone tothyroisdtimulatinhgormone (TF/TSH), and the estracriotlo thyroidstimulating 49 hormone (E/TSH).ToW serum fluorinaend TF/TSH were negativelycorrelated (r,.1B,P..05).Allthreeratioswere significanatnldy negativelycorrelatewdith age. TB/TSH and TF/7SH were negativelcyorrelatewdithSMI, (r.-.24p..Ol and r.@.23p.-01 respectively). The Pearson correlatiocnoefficienftosrthe bound tesiosteronetofollicle stimul2tinhgormone (TS/FSH) ratiot,hefreetestostemnetofollicslteimulating hormone (TF/FSH).and theastradiotlofollicslteimulatinhgormone (E/FSH) are providedinTable 4.1.36.Age was theonlycovariatethatwas significantly correlatewdiththethreeratios. The correlatiocnoefficienftosrselectedratiosbetween pituitargylycopmtein hormones,TSH, LH, and LH, are presentedinTable 4.1.37.The thyroid stimulatinhgormone tofollicslteimulatinhgormone (TSWFSH), thethyroid stimulatinhgormone tolutainizihnogrmone (TSH/LH),and the follicslteimulating hormone tolutsinizihnogrmone (FSKIH) arepmvided. Age was significantly correlatewdiththe FSH/LH ratioand the TSHIFSH ratio.AJcoholconsumption was correlatewdithbothTSHIFSH and TSK/LH. As shown inthe flnalregressionmodels,theTBMF ratioincreasedas tota)serum fluoriniencreased(Tables4.38and 4.39).Alcoholconsumption,cigarette consumption,estradiolp,rolactina,nd TSH were notsignificantryelatedtothe TB/TF ratioineithermodel. These covariatesdo not substantiallaylterthe estimatedrelationshbieptween totalserum fluo6neand TB/TF ratiowhen includedinthe regressionmodel. The quadraticincreaseofthe TB/TF ratioover theobservedrange oftotalserum fluoriniesillustratiendFigure4.8. The covarialeu3sed were: nonsmoker, lessthan one ounce ofalcoholconsumed per day,30 yearsofage, and a SMI of 30 kg/m2. Table4.1.40presentsthe fullregressionmodel forthe estmcriotlo bound lest=emne ratio(EITB).Totalserum fluorinweas notsignificantalsysociated withtheE/TB ratio.BMI was a determinantofthe E/TB ratio.Free testostemne was negativelryelatedtothe E(rB ratio. 50 The fulrlegressionmodel forestradiotlofreetestosieroneratio(E/TF)is displayedinTable 4.1-41.There was a significarprotsftive*dose-response relationshibpetween the E/TF ratioand totalserum fluorine.Although the doseresponse relationshifporfreetestosteronewas modified by age, the doseresponse relationshifporthe ratiowas not modifiedby age. As shown inTables 4.1.42,4.1.43and 4.1.44,totalserum fluorinewas not significantalsysociatedwith EILH and TBA-H, but was positivelayssociationwith theTFILH ratio(3-.05, p--.09).Bound testosteronsand FSH were associated withtheTFA-H ratio(Sm.003.p=.0001)and (3-.33,p-.0001). Cigaretteconsumption and freetestosteronewere stronglyand significantly relatetdo theTBIP ratio(3-1.49,p..02 and 3-3.93.p..008 respectively()Table 4.1.45).Cigaretteconsumption and bound testostemne were significaryrtellyated totheTF/P ratio(B..04.p...03and 8-.002,p..03 respectively()Table4.1.46). Only cigaretteconsumption was significantrleylatedto the E/P ratio(13=.0l. pa.005)(Table4.1.47). Tables4.48 through4.50 presentfulrlegressionmodels forthe ratiosofpmlactin toFSH (P/FSH),pmfacfinto LH (PILH),and pmlacfinto TSH (P/TSH). Ineach of thethreeregressionmodels totalserum fluorinewas positivelaynd significanoy associatedwiththe prolactin-hormoneratio.Moderate drinkershad a significantdliyfferenrtatiototalserum fluorinedose-response relationship compared tothe relabonshipsinlightdrinkerand nonrespondents. The fulrlegressionmodels forthe gly=proteinhormone ratiosare presentedin Table 4.1.51to 4.1.59. As shown intable4.1.-r,t2o.talserum fluorinweas significantrleylatedto TF/TSH (B.-.28,p..03) and bound testosteroneand FSH were significantrleylatedto the TFITSH ratio(B..Ol,p..006 and B..68,P..04 respectively)T.ota)serum fluorinweas not significantayssociatedwiththe other glycopmteinhormone ratios. 4-1-5Cholester@olL-ow Density LionDrotain-Hich QenSity Lipcomtein-And T6gly.aerides 51 Table A.1.60providesthecorrelatiocnoefficientfsorserum lipid!ssp4ecifically cholesterolo,w dens@y lipoprotei(nLDL).and high-aensitlyipopmtein(HDL),with totasjerum fluo@ne,age,BMI, alcoholconsumption,and cigarettceonsumption. Totalserum fluorinweas notsignificantcloyrrelatedwithcholesterolL,DL HDL@ orthglyce6des.Cholesteroalnd thgiyceridewsere correlatewdithage (r-.25, P,.008and r..19.p..04,respectivelya)n,d SMI (r.1.9.p-.04and r..27,p..004. respectivelyC)i.garettsemoking was positivelaynd significantcloyrrelatewdith cholester(orl-.35p,..0001)L,DL (r..28p,..002),and triglyce6de(sr..9l,p..04). HDL was notsignificantcloyrrelatewdithany variable,althoughthe correlation withalcoholconsumptionwas suggestive(r-.18,p-.06). Totalfluohnswas notsignificantalsysociatedwithcholesterolL,OL or thglyceride(sTables4.1.61,Table4.1.62,Table4.1.64).Smoking, age,and GGT were poshivelyand signifir-anatslsyociatedwithcholesterol.Smoking and prolactiwnere positivealnyd significantalsysociatedwithLDL Smoking and free testasteronwsere positivelayssociatedand bound testosteronswas negatively associatedwitht6glycerides. The finarlegressionmodel forHDL, displayedinTable 4.1.63,presentsa differepnitcture.HDL decreased as totalfluoriniencreasedinmoderate drinkers. Inlighdtrinkerst,herewas a negligiblcehange inHDL as totalfluoriniencreased. Self-reportemdoderatealcoholconsumption was positivelayssociatedwithHDL Additionallbyo,und testostemnewas positivelayssociatedwithHDL, whiletree testostemnewas negativelayssociated. A 1 .6 HepaticPararnetersS-erum GltjtamicOxaloaceticTmnsaminase (SGONSerim ClutamicPvmic Trar-saminase(SGPT)- AlkglfinPehgsghatase (AKPHI. Gamma GILnarnylTransferase(GGT). Table4.1.65presentsthe correlatiocnoefficientbsetween the hepatic parameters,SGOT, SGPT, GGT, AKPH, and totalserum fluorinea,ge, BMI, alcoholconsumption,and cigarettceonsumption. The hepaticparameterswere notsignificantcloyrrelatewdithtotalserum fluorine.SGOT was notsignificantly correlatewdithany ofthe parflcipanctharactehstos. SGPT and GGT were correlatesdignificantolnylywithBMI (ru.20,p..02 and r..27,p..004 52 respec@vely).AKPH was significardcyorrelatedwithage, SMI, alcohol consumption,and cigarettceonsumption. The correlatiocnoefficientbsetween the hepaticparameters and cholesterol, LDL. HDL. thgiycehdes,estradiolT,F, TB, and prolactinare displayed-inTable 4.1.66.SGOT and AKPH were significantcloyrrelatedwithprolactinS.GPT was correlatewdithcholesieroafnd thglycerides.GGT was correlatedwith cholesterolt,hglyceridesa,nd freetestosterone.As expected,SGOT, SGPT, and GGT were highlycorrelate(dTable4.1.67).AKPH was only correlatedwith GGT. The SGOT, SGPT, GGT, and AKPH mean values were not significantdliyfterertt among thefivetotalserum nuodne categories(Table4.1-68).SGOT and SGPT mean valueswere not significantdliyfferenftorSMI, age, alcoholuse. and smoking (Tables4.1.59to4.1.72). Mean GGT was significanthliygherinthe greaterthanthirtByMI group (p..03).As shown inTable 4.1.72,mean and median AKPH valueswere significanthliygher insmokers compared to nonsmokers (pn.012). Tables4.1.73A, S,and C presentthreelinearmultipleregressionmodels for SGOT. Innon-obese workers (SMI-25).SGOT decreased as totalfluorine increased.Inobese workers (BMI- 35),the associationbetween totalserum fluorinaend SGOT was inthe oppositecrirecfionM.odel 2 includedGGT as a covahate(Table4.1.73B). The associationbetween totalfluorinaend SGOT, as wellas theeffectmodificatiobny BMI, were presentafteradjustingforGGT.. When SGPT was includedinthe regressionmodel (Table4.1.73C),the associatiobnetween totalfluorineand SGOT was weak and nonsignificant.The effecmtodificatiobny SMI was no longerpresent. AKPH had littloeffecton the regressionestimateswhen includedinthe model. Three linearmultipleregressionmodels forSGPT are providedinTables 4.1.74 A, S. and C. Innon-obese workers (SMI-25), SGPT decreased as totalfluorine increased.However, inobese workers (BMI- 35).the associationbetween total serum fluorinaend SGPT was inthe oppositedirection.Uttlechange occurredin theestimatesafteradjustingforGGT. As seen inTable 4.1.74C, the association was significanatl,thoughweaker instrength,afteradjustingforSGOT. The effect S3 modificatiobny SMI was Present.When AKPH was includedinthe model,effect estimatesdidnotchange significantly. The finarlegressionmodels forGGT, providedinTables 4.75A. B, and C, presenta differenpticture.GGT decreased as totalfluoriniencreasedin moderatedrinkers.InlighdtrinkersG,GT decreased lesssteeplyas totalfluorine increased.ControllinfgorSGOT and SGPT (model2 and 3)didnotsignificantly altertherelationshbieptween totalfluorinaend GGT. Moderate alcohol consumptionwas positivelayssociatedwithGGT. Table4.1.76presentsthe finalregressionmodel forAKPH. Innonsmokers, total serum fluorinweas negativelayssociatedwithAKPH. As the number of cigarettessmoked per day increasedtomore than fiveper day,AKPH increased as totaslerum fluoriniencreased. A.1.7HernatolocyParameters*Hernoglobin.White Blgod Count PolyrnomhomuclearLeuko=e Count, Band Count, E-gsinog2bCioluntLymphocyteCount.Mono2de Count.PlateleCtount,And BasophilCount, Table4.1.77presentsthecorrelatiocnoefficientbsetween the nine hematology parametersand totalserum fluorinea,ge, SMI, alcoholuse, and cigarette consumption.The onlyparameterthatwas significantcloyrrelatewdithtotal serum fluorinweas lymphocytecount (r-.91, ps..04)M.onocyte count was correlatewdithSMI (rm-.22p,a.04)and alcoholconsumption,(r-.21.pa.03).A,ll theparameters,exceptthe basophiland band counts,were stronglyassociated withcigarettceonsumption.Alcoholconsumption was correlatewdith hemoglobin(,r.-.20p,m.04),and band count(ru.26p,..005). The finarlegressionmodels forhemoglobin and the erythrocytiendices,mean corpusculahremoglobin(MCH) and mean corpuscularvolume (MCV), are presentedinTables 4.1.78,4.1.79,and 4.1.80respectivelyT.otalserum fluodne was significantalsysociatedwithhemoglobin..The associationhemoglobin and MCV were modifiedby smoldng. Insmokers who smoked seven or more cigarenesperday,hemaglobin and MCV increasedsignificantalsytotalfluorine increased.Innonsmokers, hemaglobin and MCVdecreased as totalfluorine 54 increased.The associatioonftotaflluorinweithMCH was modifiedby smoking and by alcoholuse. The increaseinMCH as totalfluorineincreasedwas enhanced withincreasedsmoking. Inlighdtrinkerst,otalserum fluorinehad a weak associatiownithMCH. Inmoderate drinkersM,CH increasedas total fluoriniencreased.There was a positivaessociatioonfboth MCH and MCV with alcoholconsumption.None ofthe estimatedassociationsare ofclinically significamratgtnitude overthe range oftotalfluorinvealues. The whilebloodcellcount(WBC) increasedsignificantilnynonrespondents as totaflluoriniencreasedabove 2ppm. increasedlessinmoderate drinkers.and increasedthe leastinlightdrinkers(Table4.1.81)A.s expected,cigarette smokingintensitwyas positivelayssociatedwithWBC. PMN increased significarirntallycoholuse nonrespondentsas totalfluorineincreasedand increasedlesssteeplyinmoderate drinkers(Table4.1.82).Inlightdrinkerst.otal serum fluoridaebove 10 ppm was associatedwitha decreased inPMN. Cigarettsemoking was positivelayssociatedwithPMN. As shown inTable 4.1.83,thefinalregressionmodels forband countprovideslittelveidence that totaflluorinweas associatedwithband count Moderate alcoholuse was estimatedtoreducethe band count. Smoking was positivelayssociatedwith band counl -The negativeassociatiobnetween totalfluorinaend lymphocyte countwas modifiebdy adipositya.lcoholconsumption,and cigarettsemoking (Table 4.1.84)T.he decreaseinlymphocytecountwas sm2lleras BMI increased.The decreaseinlymphocytecountassociatedwithtotalfluorinaebove 3 ppm was greateirnmoderate drinkerscompared tononrespondents. As cigarette consumptionincreased,the decrease inlymphocyte count increased. The positivaessociatiobnetween totalfluorinaend monocyte count(MONO) was modifiedby adipositCyrable4.1.86).As SMI increased,the associatiownith MONO was weaker Cigarettesmoking and LH were positivelayssociatedwith MONO. Alcoholconsumptionwas negativelayssociatedwithMONO. The associatiobnetween totalfluorinaend sosinophilcount (EOS) was negativefor nonsmokers,butwas positiveas more than ten cigarettepser day were smoked 55 (Table4.1.86).As smoking increased.the PFOA associateddecrease in BASO was smaller(Table4.1.88). The associationbetween totajfluorinaend platelectount (PLAT) was modifiedby adipositaynd cigarettsemoking intensit(yTable4.1.87.).In loan participants (BMI-25),PLAT increasedas totalfluoriniencreased. Inobese participants (SMI-40),the PLAT decreased as totalfluoriniencreased. As smoking increased,the rateof increaseinPLAT associatedvath totalfluorineabove 10 PPM decreased. A 1-8 Summary Of Results The serum fluorinleevelsinChemolite workers were 20-100 times higherthan expectedinworkers notdirectliynvolvedinPFOA production.Allworkers with levelsabove 10 ppm fluorinework inPFOA productionareas. Smoking was associatedwitha smallincreaseinserum fluorine.Age was not associatedwith serum fluorinleevels.The two women employed inthe PFOA productionareas had totalserum fluorinleevelssimilarto men. Alcoholuse, smoking, age, BMI, and hormones had the expected associations withperipheraleukocytocounts,hematology parameters, cholesterolH,DL, LDL, and hepa:ficenzymes. The main hormone resultsare: 1.The number ofmale workers withhormone values outsideofthe labormoryreferencerange was greaterthan expected forestradiolf,ree testostemne,bound lestosteronea,nd prolactin. 2.Totalserum fluorinweas negativelyassociated with freetestostemne and positivelayssociatedwith estradiol.No associationwas noted between totalserum fluorinaend LH. 3.ETTF and TB/TF, but not EITB. were positivelayssociated withtotal serum fluorine. 4. EILH and TB/LH were not associatedwithtotalserum fluorine.However, the relationshibpetween totalserum fluorineand TF/LH was suggestive. 56 5.TSH was Positivelayssociatedwithtotalserum fluorine.TF/TSH was negativelyassociatedwithtotalserum fluorineT;S/TSH and E/TSH were not. 6. Prolactiannd totaslerum fluorinweere positivelayssociatedin moderate drinkersb,ut notinlightdrinkers. 7.P/FSH, P/LH, P/TSH were positivelayssociatedwithtotalserum fluorine. TEIP, TFIP,and EfP were notassociatedwithtotalserum fluorine The main hepaticparameter resultsare: 1. The increaseinSGOT and SGPT levelsassociatedwithacripasitwyas enhanced by totalserum fluorine. 2.The inductioonfGGT by alcoholwas decreased as totalserum fluorine increased. 3.The inductionofAKPH by smoking was increasedby increasinglevelsof totalserum fluadne. The main cholesteroalnd lipopmtsinresultsare: 1.Cholesteroland t6glyceridleevelswere notassociatedwithtotalserum tuodne. 2. LDL was*notassociatedwithtotalserum fluorine. 3.The positivaessociatiobnetween moderate alcoholuse and HDL levels was reduced as totalserum fluoriniencreased. The main hematologyparameterand peripheraloukocytocount resultasre: 1.The effectofsmoking on hemoglobin and MCV was enhanced by total serum fluorine. 2.Totalserum fluorinweas negativelyassociatedwithallperipheral leukocytocountsexceptPMNs and MONOS, which were poshively associated. 3.The associationbsetween cellcounts and totalserum fluorinweere modifiedby smoking,drinkinga,nd adiposity. 57 -dP Thp IPgr)QhernotitoRetrosneeliveCohort mortaft Stu@y A totalof3,537 individualwsho were employed atthe Chemol!tL,plantbetween January 1,1947 and December 31, 1983 were iderctififerdom company records. The cohortconsistedof2,788 (79%) male and 749 (21%) females employees (Tables4.2.1and 4.2.2)T.he majorityofwomen (67.3%) never worked inthe Chemical Division.Of the 19,309 person years (PY) observed forwomen, 68.8% occurredinthose who were never employed inthe Chemical Division.The mean follow-upforwomen was 25.8 years inthe overallcohort.24.6 years inthe ChemicalDivision(CD) cohort and 26.4 yearsinthe non-CD cohorl The distributioofnfollow-upperiodswas similarinthe women's CO and non-CD cohorts. The women's mean age atfirsetmployment was 27.6 years. Sixty-eighptercent were lessthan 30 years oldat employment, 9.7% were olderthan 40 at first employment atChemolfte.The CO cohortwas slightloylderthan the non-CO cohort.The CO and non-CD distributionosflatencytimes were notstatistically differen(tp-.66)T.he mean durationofemployment forwomen was 8.7 years and ranged from sixmonths to 41.4 years.The cristributoonfyears of pmployment was significantdliyfferenftorCD and non-CO women (p@e-0001).Of non-CD women, 11.9% were employed formore than twenty years.Of 245 women intheCD cohort,51 (21.1%) were employed formore than twenty years. As shown inTable 4.2.2.the 2,788 men who were ever employed formore than sixmonths atChemolitecontributeda totalof71,117.7 PY which was about equallydividedbetween the CO and non-CO cohorts. The mean follow-upforthe overalmlajo cohortwas 25.5 years. The distributioonffollow-upperiodsand distributioofnyear offirsetmployment was similarinthe m;We CD and non.CD cohons. The average age atdeath was higherinthe male non-C D group, 58.1 years,compared totheCD group, 54.2 years. The durationof em'ployment for men (mean 13.6years,median 9.8 yeam) was longerthan forwomen. The distributioofnyears ofemployment was significantldyifferenftorCD and non-CO men (p<.0001).-Ofnon-CO men, 25.5% were employed forlongerthan twenty 58 years. Of men inthe CD coholl,38.0% were employed forlongerthan twenty years. vitalstatuswas obtainedfor100% ofthe women's cohort(Table4.2.3).Among the 749 women therewere 50 deaths;11 inthe CO cohort and 39 inthe non-CD cohorl Vitalstatuswas obtainedfor100% ofthe men's cohort.Among the2788 men therewere 346 deaths;148 deaths inthe CD group and 200 inthe non-CO group. Sixindividualwsho had employmertt recordsthatwere missing informatiownere excluded from the cohortand theirvitalslatuswas not ascertained.Death certificatweesre obtainedfor99.5% ofdeaths. Two deaths occurredoutsidethe U.S.and causes ofdeath were ascertainedby othermeans. A 2 1 gtnnd2rdi2edMortaiilRy2tins(SMRR) A@2-1@lSMEs For oman The numbers ofdeaths,the SMRs and 95% confidence interval(sCI)among women inthe 1947-1989 follow-upperiodare shown inTable 4.2-5.The SMRs forallcauses ofdeath (SMR-.75, 95% Cl .56-.99)a,nd cancer (SMR..71, 95% Cl .42-1.14)were significantlloywerthan expected incomparison tonational rates.No associationwas found withdurationof employment or latencyfor deathsfrom allcauses, cancer,and cardiovasculardiseases (Tables4.2.6and 4.Z7). SMRs forCO women and non-CO women are displayedinTable 4.2.8. The estimatedSMR fortheCD cohortofwomen were lessthan expected. In CD women, the allcauses SMR was .46 (95% Cl .23,.86)and the cancer SMR was .31(95% Cl .07,1.05).The SMRs forthe non-CD women were closerto unity. 4.2-1.2SMRs For Men The number of.ma)e deaths,the expected number of male deaths based on U.S. nationalwhitemale rates,and age and calendar periodadjustedSMRs with associated95% Cis are presented inTable 4.2.9.The SMR forallcauses (.73. 95% Cl forcardiovascularcfiseases(SMR..71, 95% Cl 60..48)f,orall gastrointestin(aGlI)diseases(.50,95% Cl .26..87)and forallrespiratory diseases(.50,95%Cl .27,.86)were significantleyssthan one. None ofthe S9 cause-specifSiMcRs were largenorwere theestimatessignificantdliyfferent fromone. AS shown inTable 4.2.10,the resultswere similarwhen the expected numbers ofrnajodeathswas based an Minnesota whitemaje rates. Table4.2.11,Table 4.2.12,and Table 42.13 presentadjustedSMRs and 9S% Cl formales based on Minnesotamortalitryatesforthreelatencyinterval1s0, 15, and 20 yearsrespectivelyT.he threelatencyintervaltshe allcauses SMR ranged trom.75to.77. For allcancers,SMRs ranged from 1.06to1.12and were nonsignifir-anAtm.ong men therewas no associationbetween any cause ofdeathand durationofemployment (Table4.2.14,Table 4.2-15,and Table 4.2.16). Table4Zl 7 and 4.2.18displaythe SMRs and 95% Cl forCD and non-CD male workers.The allcauses SMRs were .69 (.59,.79f)orthe non CO group and .86 (.72,1.01f)ortheCO group.The SMRs forprostatecancer,based on a comparisonwithMinnesotapopulationrates,were 2.03 (95% Cl .55.4.59i)nthe Co groupand .58(95% Cl.07,2.09)inthenon-CD cohort. There were 4 observeddeathsfrom prostatecancer compared to 2 expected inthe CD group. The latencyanalysisfornon-CD and CD men are presented inTables 4.2.19and 4.2.20.There was no associationbsetween any cause ofdeath and latencyin eithergroup. As shown inTable 4.2.21and 4.2.22,male CO cohortmembers withmore than 10 or more than20 yearsof employment had SMRs thatwere lessthan one for allcauses ofdeath,allmalignancy,cardiovasculacrriseasesand allrespiratory diseases.Among male non-CO cohortmembers withmore than ten years of employment or more 20 yearsof employment. the SMRs forallcauses, cardiovasculadriseaseand allrespiratordyiseaseswere significantlleyssthan expected(Table4.2.23and 4.2.24).There was no associationofany cause of deathwithdurationofemployment atChemolitsineitherCD ornon-CD groups. 4-2-2Standardlz@dR2t@ Ratios(SRRsi Age adjustedstandardizedrateratios(SRRS) were calculatedforallcauses,all cancer,and cardiovasculadriseasesmortalitcyompahng men employed atthe so plarnfortenyearsormore tomen employed forlessthan ten years. The SRRs are presentedinTable 4.2.25.The 95% Cls forallcauses,allcancer,and all cardiovasculadriseaseswere wideand includeone. Confoundingvariablesuch as yearoffirsetmployment and lengthoffollow-upwere notcontrolleidnthis anajysisdue tosmallnumbers and unstablerateswithinthe largenumber of strata. Table4.2.26presentsthe age adjustedSRRs forallcauses.allcancers,lung .cancerG.i cancer,and allcardiovasculadriseases mortalitcyomparing men ever employed inthe CO withmen neveremployed inthe CD. AllSRRs were slightly greaterthan one,however, none was statisticaslilgynificarTL 4.2-3MarftelR-elativeRisks (RRMH) Age stratifiReRdMH, contras;fitnhge ratesinmen ever employed intheCD compared to the ratesinmen neveremployed inthe CD. were calculatedforall causes,allcancer,and allcardiovasculadriseases mortalitaynd are displayedin Table4.2.27.The estimatedRR forCO employment versus non-CD employment didnotfollowa monotonic patternand the 95% Cls includeone foreach ofthe threeendpoints. Table4.2.28presentsthe RRMH formen employed forlessthan ten yeam to thoseemployed formore thantenyears.The allcauses RRMH (2.16,95% Cl 1.52,2.70)inthe30 to39 yearage atfirsetmployment stratawas reflacedin boththe RRMH forallcancers(1.75,95% Cl.95,3.21)and cardiovascular diseases(3.53,95% Cl 1.68,6.21)T.he RRMH were notadjustedforimportant timecova@ates such as the year offirsetmployment. 4.2.4Pron"ional Ha--ardRegressionModel RelativeRiskEstimates 4.2.4.1Prn2grtionaHla--ardModels For Male Workers Table4.2.29to4.2.36show thefinalproportionahlazard (PH) model fordeath fromallcauses,cardiovasculadriseases,allcancers,lung cancer,Gi cancer, pmstatecancer, pancreaticcancer,and diabetesamong the 2788 male workers 61 ever employed atChemoliteforgreaterthan sixmonths. There was no evidence forviolatioonfthe PH assumptions or forsignificarnfotnlinearassociations between theindependentvariablesand mortajityA.s expected.age at first employment was positivelayssociatedwithallcauses ofdeath. The RR fora* one year increaseinage at firsetmployment was 1.082 (95% Cl 1.069,1.094). Year offirsetmployment and durationofemployment were negativelyassociated withallcauses mortalityT.he hsk ofdeath associatedwith months inthe Chemical Divisionwas smalland nonsignificant Forcardiovasculadriseases monality,the RR fora one year increaseinage at firsetmployment was 1.126 (95% Cl 1.069,1.094).Year offirsetmployment was negativelayssociatedwithcardiovasculardiseases mortality.Time inthe CD was not associatedwithdeath from cardiovasculardiseases. Age atfirsetmployment was positivelayssociatedwithcancer mortality.The RR fora one year increaseinage of employmertt was 1.08 (95% Cl 1.06,1.10). Durationofemployment was negativelyassociatedwith cancer. The RR was .972(95% Cl.96,.99f)ora one year increaseinemployment There was no associatioonfcancer mortalitywithemployment time inthe CD. The finalprostatecancer mortalitypmporficnalhazard model formale cohort -members isshown inTable 4.2.34.Time inthe Chemical Divisionwas posftively and-significariatsfsyociatedwith prostatecancer mortaltty.-Ther,elativreiskfora one yearincreaseinCO employment time was 1.13 (95% Cl 1.01.1.43).Age at firsetmployment was positivelayssociatedwith prostatecancer mortalitryisk.A one year increaseInage at firsetmployment was associatedwitha RR of 1.09 (95% Cl .99.1.19).The RR forlungcancer mortalitwyas 1.07(95% Cl .03.1.12) fora one year increasein2gG of employment. Months inthe chemicaldivision was notsignificanvayssociatedwithlung cancer mortalityT.able 4.2.33shows thefinalproportionahlazard (PH) model forallGi cancer mortalityT.he estimatedRR fora one year increaseinage at firsetmployment was 1.14 (95% Cl 1.09,1.19)Y.ear offirsetmployment. durationofemployment and time employed inthe CD were not associatedwithGi cancer risk.Age atfirst employment was positivelayssociatedwith pancreaticcancer mortal4. The othercovariateswere weakly associated with pancreaticcancer riskand were 62 notsignificantdliyfferenftrom one. A one yearincreaseinage atfirst employment was positivelayssociatedwithdiabetesmortality(RR m 1.10.95% ci 1.01,1.19). -42 A-2 PronnriinnalH2i2rd ModLaISFor F@mate Workers Table4.2.37,4.2.38and 4.2.39 show the finwPH model fordeath from all causes.cardiovasculadriseasesand allcancersamong the 749 female cohort members. Age atfirsetmployment was positivelayssociatedwithallcauses mortalityT.he RR forallcauses ofdeath among women employed fortwo toten years(3.72)and among women employed for greaterthantenyears(2.33)were significantglryeaterthanthe allcauses mortalitiynwomen employed forless thantwo years.Time inthe CO was notrelatedtomortalityT.he RR fordeath fromcardiovasculadriseasesassociatedwitha one year increaseinage atfirst employment was 1.13(1.07,.118).The year atfirsetmployment, durationof employment,and timeinthe CD were notsignificantalsysociatedwithfemale cardiovasculadriseasesmortalityT.he RR fordeath from cancer was associated withage atfirsetmployment A one year increaseinage atfirsetmployment increasetheRR fordeathfromcancer (1.09(1.04,1.14)T.he yearatfimt employment,durationofemployment, andrime inthe chemicaldivisiownere weaklyand non-significantalsysociatedwithfemale cancer mortality. 63 A n PhY@q!ClCgiECfter"-Tplbles TABLE 4.1.1AGE DISTRIBUTION IN FIVE YEAR AGE GROUPS 3M CHEMOLITE PLANT, C017AGE GROVE, MINNESOTA AGE NUMBER PERCENT 21-25 2S-30 31-35 3S-40 41-45 46-50 51-55 56-60 TOTAL MEAN SD MEDIAN --RANGE 3 is 26 22 is 9 13 6 115 392 8.91 37 24-Sg 2.6 15.7 22.6 19.1 15.7 7.8 11.3 52 100.0 64 TABLE 4.1.2DISTRIBUTION OF ALCOHOL AND TOBACCO USE 3M CHEMOLITE PLANT, C077AGE GROVE, MINNESOTA USE STATUS NUMBER PERCENT TOBACCO USE CURRENT SMOKER 28 NONSMOKER 85 MISSING VALUES 2 TOTAL 115 ALCOHOL USE -clcz ETHANOL/DAY' 87 1-3oz ETHANOLIDAY 20 MISSING VALUES a TOTAL 115 'Includes22 nondrinkers 24.3 73.9 1.8 100.0 75.6 17.4 7.0 100.0 65 TABLE 4.1.3THE JOINT DISTRIBUTION OF TOBACCO AND ALCOHOL USE 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA ALCOHOL USE 1oz/day 1-3ozlday missing TOTAL SMOKER TOBACCO NONSMOKER USE MISSING TOTAL 19(67.9%) 7(25.0%) 2 (7.1%) 28(100-/.) 67(78.8%) 13(15.3%) 5 (5.9%) 1 (50.0%) 87(75.6%) 0 (0'/-) 20 (17.4-/.) 1 (SO.Oo/.) 8(7.00/.) 85 (10()*/0) 2 (1000/.) 11s (ioo-/.) se TABLE 4.1.4DISTRIBUTION OF AGE BY SMOKING AND DRINKING STATUS. 3M CHEMOLITE PLANT, COTTAGE GROVE. MINNESOTA AGE(years) N MEAN SD MEDIAN-- RANGE- TEST# Alcohol clovd 97 39.9 9.31 37 24-59 i-3cvd missing 20 37.5 6.95 37 27-SI ';)-.29 8 36.6 8.70 35 27-S4 *pw.17 Tobacco smoker 28 40.4 7.Sg 39 28-S4 monsrnoker as 39.0 0.3S 37 24-S9 'p-.47 rnesing 2 32.S 3M 32 30-35 TOTAL 115 392 8.91 37 24-Sg "Studentt test,Prob>T,referencegroups <1oz/day,srnoker 67 TABLE 4.1.5PEARSON CORRELATION COEFFICIENTS BETWEEN SERUM FLUORINE. AGE BODY MASS INDEX (BMI). DAILY ALCOHOL USF, AND DAILY TOBACCO CONSUMPTION. 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA TOTAL TOTAL FLUORINE AGE smi ALCOHOL TOBACCO TOTAL FLUORINE (Pam) AGE (yosrs) ISLO(kg/mz) DD4 .0002 1 .26 D..Oos ALCOHOL (ovday) -.007 -.14 .08 1 TUBAC-CO (Cig3wday) .006 .15 ..04 .08 1 68 TABLE 4.1.6BODY MASS INDEX DISTRIBUTION 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA BMI (k2/m2) *15-20 .v2D-25 w2S-30 W30-M w3S-" TOTAL MEAN ON so MEDIAN BAM RANGE NUMBER 1 40 57 is 2 115 26.9 3.4 26.3 PERCENT 0.9 34.8 49S 13.0 1.8 100.0 69 TABLE 4.1.7BODY MASS INDEX BY SMOKING AND DRINKING STATUS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA BMI(kg/m2) N MEAN so MEDIAN RANGE TEST# Alcohol ,clcvd 1.3cztd rnissing 87 26.9 3-r%4 26.1 20 27.2 3.10 27.0 22.8-33.7 *p..71 a 2S.9 3." 26.1 21.3-30.4 Tobacco smoker nonsmoker mining 28 26.6 3.63 26.3 18.8-28.2 as 27.0 2.99 26.6 21.4-M.7 0;)..V 2 26.2 2.87 26.2 24.1-28.2 Total 115 26.9 3.45 26.3 laz4os *Studenttteslttest;@-valuer,eterencegroups <1 oztday,smoker -70 TABLE 4.1.8THE DISTRIBUTION OF AGF- ALCOHOL AND TOBACCO BODY MASS INDEX *- 3M CHEMOLITE PLANT. COTTAGE GROVE, MINNESOTA USE BY TOBACCO USE SMOKER NONSMOKER MISSING TOTAL ALCOHOL USE cl cz/day 1-3oz/day MISSING TOTAL AGE c4D years 3.40years TOTAL -c25 11(26.8%) 29(70.7%) 1 (2.5%) 41(100-/.) BMI Mg/kg2 25-30 15 (26--"/.) 41(71.9%) 1 (1.8%) 57(100%) 31(75.6%) 6(14.6%) 4 (9.8%) 41(100%) 31(75.6%) 10(24.4%) 41(100%) 43(75.4%) 11(19.3%) 3 (5.3%) 57(100%) 26(49.1%) 29(50.9%) 57 (100%) 2030 2(11.8-,4) is (BBZA) 0 (0%) 17(100-/,) 13(76.4%) 3(17.7%) 1 (5.9%) 17(100%) 6(35.3%)11(64.7%) 17 (100%) -1testp..005 71 TABLE 4.1.10TOTAL SERUM FLUORIDE BY eoay L4ASS INDEX, AGF, SMOKING AND DRINKING STATUS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA FLUORINE (ppm) N MEAN SD MEDIAN --RANGE TEST# BMI ,c2S 41(35.7) 2.8 3.74 2 25-30 57(49.6) 4.0 S.47 2 :o3O 17(14.8) 2.1 3.51 1 AGE -c3l 21(18.3) 3.7 4.95 2 31-40 48(41.7) 3.2 4.08 2 AL-SO 27(23.S) 32 4.26 2 51-60 19(lS.S) 3.0 6.42 1 Alcohol 41OVd 87(7S.8) 3.4 s.is 2 1-3cVd 20(17.4) 3.2 2.97 2 missing 8(7.0) 2.1 2.S3 1 Tobacco smoker 28(24.9) 3.6 4.36 2 nonsmoker 85(75.2) 3.2 4.13 2 missing 2(l.7) 3.0 424 3 TOTAL 3.3 4.97 2 #univariatAenova .*Studentt test,Probr@xT 0-19 0-26 0-14 0-20 0-14 0-19 0-26 (@-26 0-12 0-6 0-20 0-26 04 C@-26 Ful.478 Pa.24 Fa.108 Pam pm.66* 73 TABLE 4.1.11AGE DISTRIBUTION BY TOTAL SERUM FLUORINE CATEGORY. 3M CHEMOLrrE PLANT, COTTAGE GROVE, MINNESOTA AGE 20-25 26-30 3143 36-40 41-4S 49-50 51-55 56-60 TOTAL TOTAL SERUM -el 1-3 NUMBER FLUORINE 31.@10 (PERCEWM (ppm) v,15-26 1 (4.4) (1-5) 3(13.0) 10(iS.4) 6(2S.1) 13(20.0) 4(17.4) 12(18S) 2 (8.7) 13(20.0) 0 (0) 7(10.7) 6(26.1) a (03) 1 (4.3) 3 (4.6) 23(100) 65(loo) 0 (0) 4(2S.0) 4(25.0) S(31.2) 2(12.S) 0 (0) 0 (0) 1 (U) 16(loo) 1(16.7) 0 (0) 0 (0) 2(33.2) 1(20.0) 1(20.0) 0 (0) 1(20.0) i(le.7) 0 (0) i(le.7) 1(20.0) l(IS.7) 0 (0) 0 (()) 1(20.0) 6(100) 5 (100) MEAN AGE SD MEDIAN AGE AGE RANGE 39.9 10.2 37 2S-59 39.6 as 38 24-56 36.0 7.5 35.5 27-57 39.3 11.1 37.S 25-54 41.6 10.S 40 30-57 74 TABLE 4.1.12DISTRIBUTION OF TOBACCO USE BY TOTAL SERUM FLUORIDE CATEGORY. 3m CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Tobacco use Smoker Nonsmoker Missing Total TOTAL SERUM FLUORINE (ppm) cl 1-3 z-@10 2010-15 3,15-26 TOTAL NUMBER(%) 3(13.0) 16(24.6) 19(82.7) 49(75.4) 1(4.3) 0(0) 23(100) 65(100) 6(37.S) 9(56.2) i(E.3) 16(loo) 2(33.3) 4(66.7) o(O) 6(100) 1 (20.0) 28 (24.3) 4(80.0) aS (73.9) 2(1.7) 5(100) 115(100) Cigarettes/day (among smokers) MEAN 162 24.5* lao 20 so 14.0 8.8 9.9 0 MEDIAN 17 20 20 20 RANGE 2-30 7-40 3-30 20 *significwtdliyfferenftrom -clppm mean (pc.005) 20 21S 10.1 2() 20 20 2-40 75 TABLE 4.1-13DISTRIBUTION OF ALCOHOL USE BY TOTAL SERUM FLUORIDE CATEGORY. 3M CHEMOLITE PLANT, C077AGE GROVE. MINNESOTA ALCOHOL USE -clcz/day 1-3ovday MISSING TOTAL SERUM FLUORINE (ppm) -cl 1-3 2.@10 3010-15 2,15-26 NUMBER (PERCENT) 17(73.9) 2 (8.7) 4(17.4) 51(78.5) 13(M.0) 1 (1.5) 9(56.3) 4(25.0) 3(18.7) 5(83.3) l(IG.7) 0(0) 0 (0) TOTAL 23(100) GS(IOO) 16(loo) 6(100) 5(100) 76 TABL6 4.1.14BODY 3M CHEMOLRM MASS INDEX DISTRIBUTION FLUORINE CATEGORY. PLANT, COTTAGE GROVE, BY TOTAL SERUM MINNESOTA TOTAL SERUM FLUORINE (ppm) -cl 1-3 31@io SMI(kgIM2) P,lS-20 3-20-25 ,v2S-30 x3C)-35 2,3S-40 p 40 4 TOTAL MEAN BAM so MEDIAN BMIRARGE 1 (4.4) 9(39.1) 5(21-7) 7(30.4) 0 (0) 1 (4.4) 23(100) 27.6 sa 27 18.840.S NUMBER (PERCE@M 0(0) 0(0) 21 (32.3) B(SO.0) 39(60.0) 5(31.2) 5(7.7) 3(18.8) 0(c) 0(c) o(O) 0(0) 65(loo) 16(loo) 26.6 2.5 26.8 22-r@-M.7 26.3 3.3 25.7 21.4-32.5 O(D) 1(16.7) 4(66.6) a(c) 1(16.7) 0(0) 6(100) 29.4 3.7 29.8 24-c-3S.S 0(0) 1 (20.0) 4(80.0) a(D) o(O) 0(0) 5(100) 26.0 1.4 2S.s 24.1-27.6 77 TABLE 4.1.15COEFFICIENT OF VARIATION FOR SEVEN HORMONE ASSAYS. 3M CHEMOLFTE PLANT, COTTAGE GROVE. MINNESOTA HORMONE BOUND TESTOSTERONE FREETESTOSTERONE ESTRADIOL TSH LH PROLACTIN FSH cv 10.6% 12.1% 1"% 10.0% 8.6% 3.1% 5.6% 78 TABLE 4.1.16THE OBSERVED VERSUS EXPECTED NUMBER OF WORKERS WITH HORMONE ASSAYS OUTSIDE THE ASSAY REFERENCE RANGF3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA OBSERVED Estracriol 17 >=44 pg/ml Testosterone 13 bound <=300 ng/di Testosterone 11 free <zg ng/cg Prolacgn 10 >cl5 ng/mi LH 3 2-12mU/ml FSH 1 1-12 mulml TSH 1 :o.4.6mU/mi EXPECTED 2.8 2.8 2.8 2.8 2.8 2.8 2.8 OIE* 95% cl- 6.0 (3.6,9.8) 4.5 3.9 (2.0,7.1) 3.5 (1.8,6.7) 1.1 (0.3,3.3) .4 (0.1,2-0) .4 (0.1,2-0) ME -OBSERVED M EXPECTED RATIO Cl-95% CONFIDENCE INTERVAL 79 TABLE 4.1-'1P7EARSON CORRELATION COEFFICIENTS BETWEEN SERUM 3M CHEMOLITE PLANT. COTTAGE GIIOVE. MINNESOTA. ESTRADIOLO EITRA FR TE AO P-.0001 TEST. .32 pm.0000 CTIU4-LnLH++ .16 .06 P..Oa FREE TESTOSTERONEI BOUND TESTOSTEROHE* PROLACTINO 774- --713P-.0001 1 21 PO-E.303 .10 .2 P-.0003 .155@ HORMONES. FS"+ -.14 P-.15 -.05 .16 P-.04 .004 LnUi+* FSH* @fPrgiwg" onw"ld #*LO(3 LUTENIZING HORMONE (mLiVmQ # FOLLICLE STRA"TM 140AMONE (mU/rno #LOG THYROID STOMULATM HORMONE (mW" .63 p-.oooi TABLE 4.1.18PEARSON CORRELATION COEFFICIENTS BETWEEN TOTAL SERUM FLUOFIIDF, AGE, BODY MASS INDEX (BMI),DAILY ALCOHOL USE, DAILY TOBACCO CONSUMPTION, AND SERUM HORMONES. 3M CHEMOLRTE PLANT, COTTAGE GROVF-, MINNESOTA TOTAL FLUORINE AGE (yeam) a mi ALCOHOL TOBACCO (kWrr,2) (ovday) (cig&tday) ESTRADIOLO .13 .25 -.01 p..16 P0.01 FREE TESTOMRONE" BOUND TESTOSTERONEO PROLACTINS .03 .08 .19 c@..OAS -.45 o..0001 .24 00.01 -.19 0..Ol ..26 o..005 -.36 c@n.0001 -.06 LnLm#* .04 .11 20 on.03 FSW -.03 m -.08 ow.= LNTSHO .26 .09 I opgtft envd ONOW **LOG LURENIZWG HORMONE (rnLftO # FOLLICLE STIMULATING HORMONE (mU/" #LOG THYROID STIOMLLATING HORMONE (mLi/mQ .05 12 1 P-m-2 -.08 os -.is on.1 I .03 -.14 ..24 22.1s .11 ---- -.16 pa.09 .18 o..06 .17 $)No.(* ..w TABLE 4.1.19BOUND TESTOSTERONE C77B)BY BODY MASS IND EK, AGE, SMOKING. DRINKING STATUS AND TOTAL SERUM FLUORIDE 3M CHEMOLrrE PLANT, COTTAGE GROVE, MINNESOTA- TS(Vdi) N MEAN SD MEDIAN RANGE TEST# SMI (KWMZ) .c2S AO(3S.4) 641 2c--30 3-30 56(49.6) sss 17(15.0) 436 Age ,c3l 31-40 41-50 5140 20(17.7) 598 48(42.5) 634 26(23.0) 512 19(is.8) 470 Alcohol <I*Vd 140z/d miuing 86(76.1) Sal 19(16.8) 494 8(7.1) 690 Tobacco smoker non"r 27(23.9) 84(74.3) 659 2(l.8) 432 Total Fluorine 'elppm 1-3 31@io 3.10.16 vlS-26 Total Z3(20.4) 584 64(56.6) 567 IS(13.3) S30 6(5.3) Goo 5(4A) am 113(100) sn #univariateAnova 242.9 196.8 172.7 232.8 214.1 las.8 226.1 212.S 215.1 272.8 177.7 233.0 97.6 592 275-1192 FuS.64 Sao 141-954 P.M@006 438 2104303 673 278-1192 Fm3.60 605 275-1189 P..016 498 141-947 409 210-Vj4 574 210-1192 F.I.23 417 141-1039 pa.27 602 409-1101 617 379-1039 Fal.69 SSG 141-1102 pa.20 432 363-601 295.4 2OU 189.3 234.6 149.6 220.7 438 27S-1192 FwO.39 m 141-1039 pa.82 674 '210-810 S93 2"-947 eso S17-NO Sel 141-1192 82 TABLE 4.120 LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE IBOUND TESTOSTERONE (ng/cfAlM)ONG 112 MALE WORKERS. 3M CHEMOLITF- PLANT, COTTAGE GROVF, MINNESOTA Variable SE(B) tvalue Intercept 1027 190.7 .0001 TotalFluorine(ppm)* -148 Age (years) .9 Age X TotalFluoride' 3 672 3.3 1.6 .05 .009 .04 SMI (kg/M2) -16 5.4 .003 Smoker** 74 45.0 .28 AJeohol(.colztday)# 89 Estradio(lpg/ml) 2 LH (mU/ml) 116 47.5 .11 1.0 .02 6.1 .004 Prolacdn(ng/mi) 8 4.1 .04 R2s .39 *Square roottransfortneonoftotw serum fluoridmeeasured inppm. **Referencecategoryisnonsmokem #Reference categoryismoderate drinkerswho consume 1-3 cz ethanoliday. -Z3 TABLE 4.121 FRSS TSSTOS'TERONE (TF)BY BODY L4ASS INDE)@ AGF-. SMOKING AND DRINKING STATUS AND TOTAL SERUM FLUORIDE. 3M CHEMOLRTE PLANT, COTTAGE GROVF, MINNESOTA N(OA) MEAN TF(ng/cu) so MEDIAN RANGE TEST# SMI kgVM2 as 25-30 3.30 40(35.4) 56(49.6) 17(15.0) Age yeam <W 31-40 41-90 si-a 20(17.7) 48(42S) 26(23.0) 19(le.8) Alcohol -clazld 1-3 oVd missing 86(79.1) is(iG.B) 8(7.1) Tobacco smoker nonsmoker missing 27(23S) 84(74.3) 2(l.8) Totw Fluctine 41ppm 1-3 3k@io 2,10-15 3,15-26 Total 23(20.4) 64(56.6) 15(132) 6(52) 5(4A) 113(100) 17.4 15.1 13.7 18.7 17.0 14.1 iis is.8 142 18.1 16.6 15.4 15.9 16.4 15.6 152 151 152 15.7 sm 4.13 6.08 7.64 3.75 4.73 3.78 5.36 4.79 6.40 3.71 SA4 -4.45 8.4 4A 3.8 52 2.2 5.4 16.7 15.8 13.5 7.4-45.3 32-23.8 S.6-M.5 Fu3Za P..03 16.7 9-14452 F=9.14 17.1 7.4-2237 pu.oool 14.3 32-232 11.5 5.6-19.0 15.2 5.645.3 F@.i.45 15.3 32-23.9 P-.23 172 11.0-29.7 17.1 1S.3 15.9 8.4-24.3 Fa.95 32-45.3 P-.33 12.7-19.0 13.9 15.8 15.3 17.0 14.1 is 6.4-4SJ 32-30.5 7.1-io.7 5.6-19.9 13-1-182 3.2-464 F=0.13 pa.97 #univariatAenava 84 TABLE 4.122 LINEAR MULTIVARLATE REGRESSION MODEL OF FACTORS PREDICTING THE FIREE TESTOSTERONE VALUE (ng/di) AMONG I11 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable 1 -9 E(-B) p-value iniercept 29.72 4.57 .0001 TotalFluorine(ppm)* -3.56 1.62 .03 Age (years) -.34 .08 .0001 Age X TotalFluoride* .07 .04 .05 BMI (kg/m2) -21 13 .11 Smoker** 1.46 1.03 .16 AJcohol(<Icz/day)# 1.65 1.14 .15 Fztradio(lpg/ml) .10 .03 .003 LH (mU/ml) .18 .15 .20 F12a .39 *Square roottransformatioonftotalserum fluoridmeeasured inppm. **Referencecategoryisnonsmokem #Referencecategoryismoderate drinkerswho consume 1-3 oz ethanovday. TABLE 4.1.23PARTICIPANT ESTRADIOL BY BODY MASS INDEX, AGE, SMOKING DRINKING STATUS AND TOTAL SERUM FLUORIDE. 3M CHF-MOLRTE PLANT, COTTAGE GROVE, MINNESOTA N(%) MEAN ESTRADIOL LOWMI) SD MEDIAN RANGE TEST# SMI (kWM2) .c2S 25-30 .!o,30 .AGE .c3O 31-40 41-SO 51-M Alcohol -cIoz/d 1.3 oz/d missing Tobacco sm&or nonsmokor missing Total Fluorine 41 ppm ->-1-3 >15-26 Total 40(3S.4) 56(49.6) 17(IS.0) 20(17.7) 48(42.5) 26(23.0) 19(is.8) 66(76.1) 19(is.9) 6(7.1) 27(23.9) 64(74.3) 2(l.8) 23(20A) 64(56.6) 15(13.3) B(S.3) 5(4A) 113(100) 34.1 33.2 322 34.4 36.8 31.6 25.9 33.0 31.8 41.1 36.3 .32.S '30.S 36.2 31.4 32.8 38.2 41.2 33.4 #univariateAnova 12.91 40 13.89 33 12.36 Z7 10.1s 34 ll.S4 36 18.48 28 7.93 24 11.78 33 16.61 30 18.20 40 17.40 34 11.63 32 13." 30 13.1 34 13.6 30 10.6 34 IS.2 3ss 11.4 42 132 33 "9 8-83 1&67 io-ss 1249 $-a IS-47 "S 8-69 23-83 14-83 "S 21-40 F..13 pm.se Fm3-SO pomi.018 Fa.14 p..?l Fa.13 P-Age 1440 8-63 10-Sa 22-66 2rD-SS 8-M Fol27 PS22 86 TABLE 4.124 LINF-AR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE ESTRADIOL VALUE (pg/dlA)MONG 113 MALE WORKERS. 3M CHEMOLITE PLANT. COTTAGE GROVE, MINNESOTA Variable IntercePt a smu@ 12.89 SE(s) 12-13 p-v ue .29 TotalFluorine(ppm)o .03 .01 .03 Age (years) ..22 .15 .14 SM I(kg/M2) Si .34 .14 Cigarettes/day .16 .11 .15 Aj=hol(<Ioz/day)# .09 .11 .98 FreeTestostemne (ng/dl) .85 24 .0007 R2u .24 *Squaretransformatioonftota)serum fluoridmeeasured inppm. #Reference categoryismoderate drinkerswho consume 1-3 cz ethanol/day. 87 TABLE 4.2SLUTENIZING HORMONE (LH) BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS. AND TOTAL SERUM FLUORINE 3M CHEMOLITE PLANT, COTTAGE GROVF@ MINNESOTA LH (MU/Mi) N(%l MEAN SD -MEDIAN RANGE TEST# SMI mWkg2 C25 25-30 @,,30 40(35.4) 56(49.6) 17(15.0) Age years -c3O 31-40 &I-so 51-60 20(17.7) 48(42.S) 26(23.0) 19(16.8) Alcohol <Ioztd 1-3oVd missing 86(76.1) 19(16.a) 69.1) Tobacco moker nott@rnoker m"ing Total Fluorine -clppm "1-3 31@io :1-10-15 >15-26 Total 27(23.9) 84(74.3) 2(l.8) 23(20.4) 64(56.6) 15(13.3) B(S.3) 5(4.4) 113(100) 5.49 5.84 3.72 421 5.49 5.33 5.90 5.60 4.69 4.86 6.30 S.Os 7.45 S.0 6.6 5.1 5.4 5.3 SA 3.06 325 1.21 2.26 3.14 1.64 4.73 3.34 1.30 1.00 3.7a 2.71 2.47 2-1 3.6 2.7 0.3 1.3 3.0 4.60 5.15 3.60 4.45 4.7S 5.15 4.10 4.70 421 4.05 5.30 4-S2 7.4S 2.6-21.7 1.7-23.0 2.0-7.2 F=G. 19 Pa-003 1.7-10.1 2.4-21.7 2-r.9.6 2.0-23.0 Fa.69 po -94 1.7-23.0 Fml.24 2-1--10.1 pm-27 3.4-62 2.6-21.7 1.7-23.0 S.7-9.2 F=5.16 pa-025 4A 2-S-9.3 F.O.is 4.8 1.7-ZSo pa.se 4.9 2.0-13.9 4.9 3.7-7.5 5.5 3.7-72 4.7 1.7-23.0 #univariatAenova TABLE 4.1.26LINEAR MULTIVAFTIATE REGRESSION MODEL FACTORS PREDICTING THE LUTENIZING HORMONE- VALUE AMONG 113 MALE WORKERS. 3M CHEMOLFTE PLANT. COTTAGE GROVE. MINNESOTA #1 OF (mU/mi) --Variable 8 SE(z, p-value Intercept 1.26 .40 .002 TotalFluodne (ppm)* .001 .008 .93 Age (years) .01 .005 .03 SMI (kg/m2) -.02 .01 .15 Smokers** .24 .23 .29 Aicohol(<Icz/day)# .06 .10 .60 Bound Testosterone .001 .0002 .008 (ng/dl) R2=.28 *logarithmitcransformatioonflt4enizinhgormone (LH). *0Reference categoryisnonsmokers. #Reference categoryismoderate drinkerswho consume 1-3 oz ethanouday. ag TABLE 4.1.27FOLLICLE STIMULATING HORMONE (FSH) BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS, AND TOTAL SERUM FLUORINE 3M CHEMOLrrE PLANT, COTTAGE GROVE. MINNESOTA N(%) MEAN FSH(rnuiml) SD MEDIAN RANGE TET#- BMIMgAg2 42S 25-30 jp3O 40(35.4) 5.02 S6(49.6) S.39 17(15.0) 421 Age yam wm 31-40 41-50 51-W Alcohol <lcvd 1-3oz/d missing 20(17.7) 328 48(42.S) 4.W 26(23.0) 5.95 19(16.8) 6.22 86(76.1) 5.37 19(lG.8) 4.18 8(7.1) 4M Tobacco smoker nonsmoker M"ng Total Fluorfne 41 ppm 1-3 3.@io @,10-15 3-IS-26 Total 27(23.9). 5.77 84(74.3) 4AS 2(l.8) 6.10 23(20.4) 4.4 64(56.6) 5.4 15(13.3) 4.8 6(5.3) 5.4 5(4.4) 4.9 113(100) 5.1 #univariatAenova 2.39 2.71 1.75 1.86 2.24 2.55 3.01 2.62 1.92 1.49 2.46 4.49 0.42 1.95 2.75 2m 2.14 2.36 2.49 4.6 is-10.3 Fol27 4.S 1.4-14.8 p=29 3.9 1.6-8.3 3.6 1.4-92 F4.72 4.6 1.6-10.3 pw.014 4.6 2.1-14.8 5.0 2.7-14.8 4.6 1.4-14.8 F=3.47 3.9 2.0-9.8 pm-Des 4.8 2-64.4 4.9 2.6-11.9 F-2.so 42 1.4-14.8 pm.09 6.1 SA-6.4 4.4 I.&IDJ F.O.75 4.8 1.4-14.8 p=M 4.9 LI-9.7 4.4 3-r@-8.9 3.7 2.6-7.7 4Z 1.4-14.8 90 TABLE 4.1.28LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE FOLLICLE STIMULATING HORMONE VALUE (mU/mi) AMONG 113 MALE WORKERS. 13M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable 8 SE(S) P-value Intercept 120 1.62 .46 TotalFluorine(ppm)* .004 .04 .91 Age (years) SMI (kg/m2) .08 .02 .0006 -.04 .05 .41 Cigarettes/day .02 .02 29 AJcohol(<Ioz/day)# .45 .48 .34 TSH (mU/ml)* ..43 .22 OS LH (mU/ml)" .44 .06 .0001 R2* 48 "logarfthmitcransformatioonffoiricsitsimulatinhgormone (FSH). #Reference categoryismoderate drinkerswho consume 1-3 cz ethanovday. Ca)ThyroiSdtimulatingHormone ##LLitenizinHgormone TABLE 4.1.29THYROID STIMULA-NNG HORMONE (TSH) BY BODY MASS INOM AGE, SMOKING AND DRINKING STATUS, AND TOTAL SERUM FLUORINF3M CHEMOLRTE PLANT, COTTAGE GROVF, MINNESOTA TSM(mwmi) N MEAN SD MEDIAN RANGE TEST# BMI mg&02 <2s 40(3S.4) im 0.66 1.04 0.37-3.14 FaM 2S-30 2o3O 56(42.6) 1.64 1.01 1038 0.4S-6.80 p..70 17(IS.0) 1.72 0.71 iss 0.62-3M Age yean W20 31-40 41-50 si-so Alcohol <Iovd 1-3oVd missing Tobacco smoker nonwnoker M"no Total fluorine -clppm 30=1-3 N3-10 2-10-15 .%16-26 Total 20(17M 48(42S) 26(23.0) 19(16.8) 86(76.1) 19(16.8) 89-1) 27(23.9) 84(74.3) 2(l.9) 23(20.4) 64(58.6) 1S(132) 6(Si) 5(4A) 113(100) 1.43 1.68 1.64 1.70 I.S7 1.93 IAO im 1.66 126 1.5 1.$ 1.6 2.4 22 1.6 Q.Ss 1.04 0.75 0.74 0.70 1.39 0.63 0.61 0.92 0.42 0.64 0.94 0.67 CZ7 1.66 0.85 1.42 1.46 1.34 im 1.40 I_ss 1.61 1.37 1.49 128 0-14&2m 0.37-6.80 0.75-3.W 0.62-3.09 Fa.47 po.70 0-1&3.56 0.60-6.80 0.37-2.22 Ful.22 pm27 0.61-3-03 OJ7-6.80 O.W 1.57 Fe.09 po.76 is 0.3-32 Fa230 la 0.4-6.8 pa.08 1.4 O."A 2.S 0.83-3.S 2.1 1.7-3.5 1A 0-"A *univariateAnova 92 TABLE 4.1.30LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE THYROID STIMULATING HORMONE' VALUE (mU/mi) AMONG 113 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVF- MINNESOTA Variable no@ IntercePt 5 -.190 SE(B) .465 _p-value .68 TotalFluorine(ppm)* .027 .009 .004 Age (years) .006 .005 .29 SMI (kg/m2) -.002 .013 .89 Cigarettes/day -.001 .004 .74 Alcohol(<3oz/day)# -.140 .194 26 Free Testostemne** .020 .009 .04 FSH## .060 .019 .003 R2=.30 *lo;arithmtircansformatioonfthyroidstimulafinhgormone (TSH). #Referencecategoryismoderate drinkerswho consume 3 oz ethanovday. " ng/co ##Folliclsetimulafinhgormone mU/fni 93 TABLE 4.1.31PROLACTIN BY BODY MASS INDF-X,.AGE, SMOKING, DRINKING STATUS, AND TOTAL SERUM FLUORINE 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA N(%) MEAN PROLACTIN SD (nwmi) MEDIAN RANGE TR=ST# SMI (kWM2) as 25-30 :-30 40(35.4) 56(49.6) 17(lS.0) Age ,c3O 31-AO 41-SO 51-60 20(17.7) 48(42.5) 26(23.0) 19(16.8) Alcohol -cIoVd 1-3oVd missing 86(76.1) 19(16.8) 6(7.1) Tobacco 3-Moker monsmokor mining 27(23.9) 64(74.3) 2(l.9) Total Fluortne @clppm 3001-3 3@@io z-10-15 3-15-26 Total 23(20.4) 64(66.6) 15(13.3) 6(5.3) 5(4.4) 113(100) 9.10 8.71 7.4S 9.63 9.01 aze 7.16 8.61 9.46 725 6.97 9.13 ll.GS 7.9 8.5 4.9 15.1 8.3 8.7 5.13 S.18 3.08 4.20 sm S-S2 3.37 4.S7 a.67 2.33 3.14 S.is .9.40 3.19 4.34 1.15 11.01 4.16 4.90 8.4 2.7-24.3 Fa.69 7.8 1.2-33.7 p..Sl 72 2-r-13.8 S.S 3.9-18.3 F..gs 8.7 12-33.7 pn.Sl 6.2 2.9-22.5 6.8 2.5-15.1 7.8 1.2-24J Fm.44 a.7 2.9-33.7 pm.50 62 43.-10.5 6.8 8.5 11.7 12-12.8 2-c-M.7 S.C@-193 F=4.18 po.043 7S 2-c-18J Fu3.D2 8.1 1.2-24.3 p-.02 SA 1.4-18.1 9.4 6".7 7.7 3.8-15.1 7.7 12-33.7 #univafiateAnova 94 TABLE 4.1.32LINEAR MULTIVARIATE REGRE@SION MODEL OF FACTORS PREDICTING THE PROLACTIN VALUE (ng/mi)AMONG 113 MALE WORKERS. 3M CHEMOLITE PLANT, CO-T7AGE GROVE, MINNESOTA Variable 8 SE(s) p-value Intercept 7.41 4.14 .07 TotalF-iuorin(eppm) 1.43 .36 .0002 Age (years) -.D4 .05 .41 SMI (kWM2) -.08 13 .53 Cigarettes/day* -.08 .04 .08 Estradio(lpglml) .06 .03 .07 AlcoholUsoo Ught (<loz/day) Nonresponse (NR) Ught X totalfluoride NR X totalfluoride 321 2-14 -1.67 -1.34 1.65 2.69 .77 .37 .05 .43 .03 .0006 R2. 22 *#Reference categoryismoderate drinkerswho c-onsume 1-3 oz ethanovday. Nonrespondants(NR) failetdocomplete the alcoholuse questonnaire hems. Ught X totalfluoridaend NR X totalfluorideare interactiotnerms foralcohol categoriesand totalserum fluoride. 95 TABLE 4.1.33PEARSON CORRELATION COEFFICIENTS BETWEEN HORMONE RATIOS AND TOTAL FLUORIDE, AGE, BODY MASS INDEX, ALCOHOL AND TOBACCO CONSUMPTION 3M CHF-MOLRM PLANT, COTTAGE GROVE, MINNESOTA EITSO EITP TB/LW TFILM4.0 Tsfrr TOTAL FLUORINE (OC)MI -.01 .11 .001 .002 -.09 .16 ow.09 AGE (Y"ra) IBMJ(kWM2) .0064 .32 pm.001 .15 c*.15 -.26 0..OOS ..32 00.001 ..40 on.0001 .27 pa.004 is t@-.06 -.14 Pa.13 ..02 .24 16 0-.Ol I P-M..Oe ALCOHOL (Ovfty) .05 .01 .06 -.01 .03 -.12 TORACCO (eigs/day) .05 .01 .04 -.01 .03 .09 @F-STRADIOL TO BOUND TF-STOSTERONE RATIO *ESTRADIOL TO FREE TESTOSTERONE ESTRADIOL TO LMNIZING HORMONE RATIO *BOUND TF-STOSTF-RONE TO LUTF-NIZING HORL40NE +*FREE TF-STOSTERONE TO LUTENMNO HORMONE BOUND TESTOSTERONE TO FREE TF-STOSTEFIONE RATIO FLATIO FTATIO 96 TABLE 4.1.34PEARSON COFIRELATIO *NC*OEFFICIE* NTS E-m7WEEN PROLACTIN HORMONE RATIOS AND TOTAL FLUORIDE, AGF- BODY MASS INDEX, ALCOHOL AND TOBACCO CONSUMPTION 3m CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA 70TAL FLUOPJNE AGE (years)BL41(kg/rn2) ALCOHOL TOBACCO (ovday) (cigs/day) Ta/Pe -.05 -.04 -.13 -.03 24 pa.cl ..Os .06 22 EIP# FSKFP" -Pit-M" P/Tsm-- .03 -.05 .03 -.09 .37 .004 on.0001 .11 .24 .09 c@0.003 .07 17 .07 1 D..07 .007 -.13 t@o.I a .15 ou.1I .17 D..07 25 ou.008 21 c@-.02 .22 cw.t* 09 @Free tostmtoronetoprowcfinratio *Fre*test=orono toprowctinratio *Esraioi toproladn mlio "Folidestmulathv hwmons toproLwM rahe "Pmlacdn toluteniz@ hormorw ratio ++Protactinto #7mid stimulatinhgormone ratio 97 TABLE 4.1.35PEARSON CORRELATION COEFFICIENTS BETWEEN THYROID STIMULATING HORMONE RATIOS AND TOTAL FLUORIDE, AGE, BODY MASS INDEX, ALCOHOL AND TOBACCO CONSUMPTION 3M CHEMOLFTE PLANT, COTTAGE GROVE, MINNESOTA TSITSH# TF(RSW EirsHo* TOTAL FLUOFTINE (pom) -.13 -.16 ow.05 -.13 AGE (YearS) IJM (kgtm2) .23 P..Dl . '@4 Da-0002 .24 Z)N..Dl .24 0..Dl .23 on.01 -.C)5 ALCOHOL (ovday) TOBACCO (cig3vday) -.16 .03 I:hs.09 -.13 01 ..05 - #Sound lostosteranteotvaid stimulatnghom,4n* rabo 'Fr" to&Umurame tothyroidstimulatinhgormone ratio *Estmdicttothyroidstimulatinhgormone ratio TABLE 4.1.36PEARSON CORRELATION COEFFICIENTS BETWEEN FOLLICLE STIMULATING HORMONE RATIOS AND TOTAL FLUORIDE.AGE, BODY MASS INDEX, ALCOHOL AND TOBACCO CONSUMPTION 3M CHEMOLrrE PLANT, C017AGE GROVE, MINNESOTA TBIFSH# TF/FSW FJFSH@ TOTAL FLUOFUNE (ppm) .07 -.01 .04 AGE (y"ra) Bkg (kg/m2) ALCOHOL (OVday) -.43 -.16 .06 po.0001 pa-08 -.47 .04 .08 c@-.0001 -.36 .07 .04 ow.0001 I I #80und testeslorane to fouicio cumulating hormone ratio *Free tost=orona to toflicleStimulating hormone ratio *Estradiol to tollide stmutating hamons rato TOBACCO (elga/day) -.06 -.12 ..02 I I 98 TABLE 4.1.37PEARSON CORRELATION COEFFICIENT8 BETWEEN PITUITARY GLYCOPROTIEN HORMONE RATIOS AND TOTAL FLUORIDE, AGE, BODY MASS INDEX, ALCOHOL AND TOBACCO CONSUMPTION 3M CHEMOLFTE PLANT, COTTAGE GROVF, MINNESOTA TSWFSHO TSHILHO FILM* TOTAL FLUOPJNE (COM) .12 .09 -.05 AGE (years)SMI (kg/M2) ALCOHOL TOBACCO (ozfday) (clgwday) -.16 D4 24 -.14 pa.ce on.01 -.02 .15 .21 -.14 oa.03 .26 .13 -.14 .05 cm.003 I I I i @Thyroid stimulabmghormone to follidestimulatinghomwm ratio *Thyroidstmulating harmonotolutenizhionrgmoneratio *FoUicisstmulating hormone tolutenizinghormone ratio -gg TABLE 4.1.38LINEAR MULTIVARIATE REGRESSION MODELI OF FACTORS PREDICTING THE BOUND-FREE TESTOSTERONE RATIO AMONG 112 MAL.E WORKERS. 3M CHEMOLrrF- PLANT, C07TAGE GROVF-, MINNESOTA Variable a SE(S) pvalue Intercept 36.60 6.87 .0001 TotaJFluorine(ppm)* .02 .008 .02 Age (years) SMI (kg/M2) .19 .101 .07 -.48 .244 OS LH+ .12 .337 .73 FSHO .92 .440 .04 R2= 21 *squaretransformatioonftotwserum fluoride +Uienizinghormone mU/ml @ foll!csjtemulatng hormone mu/ml 100 TABLE 4.1.39LINEAR MULTIVARIATE REGRESSION MODEL2 OF FACTORS PREDICTING THE BOUND-FREE TESTOSTERONE RATIO AMONG 112 MALE WORKERS. 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA Variable is SE(B) p-value Intercept 37.3 6.97 .0001 TctalFluorine(ppm)* .02 .009 .03 Age (years) SMI (kg/m2) .2S .097 .009 -.52 mo .03 LH# .55 271 .05 p2..17 "squaretransformatioonftotwserum fluoride +lltteinizihnagnnone mU/mi @ follicsitsimuiabnghormone mU/ml -101 TABLE 4.1.40LINEARMULTIVARIATEREG;KESSION MODEL OF FACTORS PFIEDICTING TI-IEESTRADIOL-BOUND TESTOSTERONE RATIO AMONG 112 MALE WORKERS. 3M CHEMOLTTE PLANT, COTTAGE GROVE, MINNESOTA Variable SE(S) ;@-value Intercept TotaJ Fluorine(ppm) Age (years) SMI (kg/m2) Cigarettestday AJcohol (<Ioztday)# Free Testosterone* LH+ .05 .00001 -.0004 .002 -.00001 .003 -.001 .0001 .027 .00001 .0004 .0007 .00002 .007 .0006 .0006 .05 .74 .29 .006 .96 .63 .008 .94 FSHO TSH++ -.002 .001 .12 -.003 .003 .30 Prol=in"* .0001 .0005 .78 R2= 21 #Reference category ismoderate * ng/dl +Iuteinizinghormone mU/ml drinkerswho @ follicjsetimulatinghormone mU/mi ++ Thyroid stimulatinghormone (mU/mi) prolactinngimi consume 1-3 oz ethanovday. -102 TABLE 4.1.41LINEAR MULTIVARIATE REGRESSION PREDICTING THE ESTRADIOL-FREE TESTOSTERONE MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MODEL OF FACTORS RA710 AMONG 112 MINNESOTA Variable 6 SE(s) tvalue Intercept 1.31 .880 .15 TotalFluorine(ppm) .002 .001 .03 Age (years) .012 .011 .34 SMI (kg/m2) .048 .026 .07 Cigarettes/day .005 .008 Si Alcohol(<laztday)# .090 .730 .70 Bound Testosterone* -.001 .0004 01 LH* .012 .035 .73 FSHO -.059 .046 .21 TSH-O+ -.204 .110 .05 Prolacbn*" o27 .018 .15 R2,.22 #Reference categoryismoderate drinkerswho 0 ng/dl +lLrfieni7ihnogrmone mU/mi @ follicslteimulatinhgormone muftl ++ Thyroidstimulatinhgormone (mU/mQ prolactnngftni consume 1-3 oz ethanovday. 103 TABLE 4.1.42LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE ESTRADIOL-LH* RATIO AMONG 112 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable 8 p- alue Intercept 3.07 3.58 .39 TotalFluorine(ppm) .02 .07 .80 Age (years) BMI (kg/m2) -.03 .05 .39 27 .10 .008 Cigarettes/day .009 .03 .77 AJcohol(.cclz/day)# .37 .90 .68 FreeTestostemne' .13 Bound Testosterone* .001 .10 21 .002 .71 FSHO TSH4.* -.75 .15 .0001 -.39 .42 .35. Prolacdn" -.03 .07 .72 R2= .34 +estradjotloJuteni7inhgormone (mU/mQ rago#Reference categoryismoderate drinkerswho consume tngtdi @folliclsetimulatinhgormone mU/mi ++ Thyroidstimulatinhgormone (mU/mQ prolacdnngftl 1-3 oz ethanouday. 104 TA]BLE 4.1.43LINEAR MUL11VARIATE REGRESSION MODEL OF FACTORS PREDICTING THE BOUND TESTOSTERONE-LH+ RATIO AMONG 112 MALE WORKERS. 3M CHEMOLTTE PLANT, COTTAGE GROVE. MINNESOTA Variable SE(S) P-value lnterceprt 74.46 48.53 .13 TotalFluorine(ppm) 27 .98 .79 Age (years) SMI (kgVM2) 29 .62 .64 -.43 1.35 .75 Cigarettes/day AJcohol(.cclz/day)* -.15 7M .44 .73 12-1 .54 Free Testosterone* 5.96 1.01 .0001 estradiol@ -.28 .38 .45 FSHOO -B.SS 2.03 .0001 TSH- -.23 S.69 .97 Prolac:Un*" .11 .9s .90 R2u.43 +bound testosteronteofutenizinhgormone (mUhnQ mtio #Reference categoryismoderate drinkerswho consume 1-3 oz ethanovday. * ng/cs @ pg/mi @@folliclestmulafinghormone mU/mi ++ Thyroidstimuwng hormone (mU/ml) prolactinng/mi 105 TABLE 4.1.44LINEAR MULTIVARIATrz REGRESSION MODEL OF FACTORS PREDICTING THE FREE TESTOSTERONE-LH+ RATIO AMONG 112 MALE WORKERS. 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA' Variable 5 SE(B) p-value Intercept TotalRuorine (ppm) 3.00 -.05 1.38 .03 .03 .09 Age (years) BMI (kg/m2) Cigarettes/day Alcohol(<loztday)# .001 .07 -.007 .30 .01 .91 .04 .08 .01 .58 .36 .41 Bound Testosterone* .003 .0007 .0001 ;:-ztradiol@ .001 .01 gi FSHCQP TSH++ Pmlacdn** -.33 .18 -.05 .06 .0001 17 .30 .03 .08 R2w.46 +freetestosterontsolutenizinhgormone (mU/mQ ratio #Reference categoryismoderate drinkerswho consume ng/di @ pg/ml @@folliclestmulatng hormone mU/ml ++ Thyroidslimulatnghormone (MU/Mi) prolactnng/ml 1-3 cz ethanovday. io6 TABLE 4.1.45LINEAR MUL-NVARIATE REGRESSION M OOEL OF FACTORS PREDICTING THE BOUND TESTOSTERONF-.PROLACTIN RATIO AMONG 111 MALE WORKEFTS. 13M CHEMOLITE PLANT, COTRAGE GROVE, MINNESOTA Variable SE(B) P-value Intercept 60.05 68.04 .38 TotalFluorin(eppm) -.15 1.38 .91 Age (years) .84 .88 .34 SMI (kgIM2) -l.S4 1.92 .42 Cigarettes/day 1.49 S2 .02 Aicohol(.colz/day)# 13.9 17.2 .42 Estradiol** ..22 .53 .68 Free Testoslemne Ljq*o 3.93 -2.23, 1.45 2.63 .008 .40 FSHO -2.55 3.50 .47 TSH+ -.95 .80 .24 R2=.17 +-+pg/tnl #Referencecategoryismoderate drinkerswho ng/di 'luteni7inhgormone mU/ml 0 follicsltemulatng hormone mU/mi + Thyroidstimulatnghormone mU/ml consume 1-3 cz ethanovday. 107 TABLE 4.1.46LINEAR MULTIVARIATE REGRESSION PREDICTING THE FREE TESTOSTERONE-PROLACTIN L4ALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MODEL OF FACTORS RATIO AMONG 111 MINNESOTA Variable B SE(B) P-value Intercept 2.41 1.76 .17 TotalFluorine(ppm) -.03 .04 .35 Age (years) -.004 .02 .95 BMI (kg/M2) -.004 .05 .93 Cigarettes/day .04 .02 .03 AJcohol(.colzlday)# -.03 .76 .97 Estradiol- -.0001 .01 .99 Bound Testosterons .002 .0001 .03 LHO* -.08 .07 .24 FSH'LB TSH+ -.12 -.18 .09 .21 .21 40 R2= .15 ++pg/mi #Reference categoryismoderate drinkerswho ng/di futeni7inhgormone mU/ml follicslteimulatnghormone mU/rnl + Thyroidstimulatinhgormone mulml consume 1-3 cz ethanovday. 108 TABLE 4.1.47LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE ESTRADJOL-PROLACTIN RATIO AMONG 111 MALE WORKERS. 3M CHEMOLRTE PLANT, COTTAGE GROVF-. MINNESOTA variable SE(OF .2-Value Intercept 2.65 4.01 Si TotalFluorine(ppm) .005 .081 .95 Age (years) .01 .05 .80 SMI (kgVM2) .07 .116 .53 Cigarettes/day .10 .036 .005 AJcohol(<loztday)* .86 1.01 .40 Bound Testosterone* -.001 .003 .95 Free Testostemne .12 .12 .31 LH** -.13 .15 .39 FSHO TSH+ -.29 .21 .17 -.67 .47 .16 R20.16 #Reference categoryismoderate drinkerswho ng/dl **lutenizinhgormone mU/mi @ follicslteimulatnghormone mU/ml + Thyroidstimulatinhgormone mU/mi consume 1-3 cz ethanovday. 109 TABLE 4.1.48LINEAR MULTIVARIATF- REGRESSION MODEL OF FACTORS PREDICTING THE PROLACTIN-FSHO RATIO AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVF-, MINNESOTA Variable SE(s) p-value Intercept 2.SS TotalFluorine(ppm) .31 I.S2 .11 .09 .008 AJcohol low (.col7Jday) .81 .52 .13 nonresponse(NR) .19 as .82 low X Ruodde -.31 .12 .01 NR X Fluoride -.08 .29 .78 Age (yeam) -.05 .02 .01 SMI (kg/m2) .01 .04 .86 Cigarettes/day -.03 .01 .03 Estradiol++ .02 .01 .06 Bound Testosterone* -.001 .001 .92 Free Tesiostemne .01 .04 .75 LHOO -.07 .05 .15 TSH+ .31 .17 .07 R2--.31 +* pgftl #Referencceategoriysmoderatedrinkerwsho consume 1.3cz othanovday. knonizinhgormonemu'ffd 0 ldfides*nufatinhgormonemLlIM + ThymidstimulatihnogrmonemU/rW -110 TABLE 4.1.49LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PFIEDICTING THE PROLACTIN-LH" RATIO AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable B SE(s) p-value intercept 1.07 TotalFluorine(ppm) .34 1.27 .09 .38 .0003 AJcohol # low (.1coztday) .68 .43 .11 nonresponse (NR) .41 .69 low X Fluoride -.35 .09 NR X Fluoride -.39 .20 s5 .0004 .05 Age (years) SMI (kg/m2) -.02 .01 .12 .05 .04 .17 Cigarettes/day -.02 .01 .09 Estradiol++ .003 .009 .76 Bound Testosterons' .001 .0008 17 Free Testosterone -.04 .03 .30 FSHO -.11 .05 .03 TSH+ .15 .14 .29 R2..31 lLftnWnghormom .**pgvng #Referencceattgoriysmodemit drink*mwho consuffm1.3cz othanoyday. *ng/dl 4)falwottknulanghormonemU/rrd Thyroisdtimulatihnaim;sonsmLlIrW -411 TABLE 4.1.soLINEAR MULTIVARIATE REGRESION MODEL OF FACTORS PREDICTING THE PROLACTIN-TSH+ RATIO AMONG 111 MALE WORKERS. 3M CHEMOLFTE PLANT, COTTAGE GROVF-. MINNESOTA -variable onam@ Intercept TotalFluorine(ppm) SE(S) p-vaiue 5.06 4.s3 .30 1.61 .37 .0001 Alcohol# low (<Iozlday) 4.16 1.67 .01 nonresponse (NR) low X Fluoride 3.85 -1.76 2.72 .38 .16 .0001 NR X Fluoride -2.11 .77 .008 Age (years) BMI (kg/M2) -.19 .06 .003 .11 .14 .43 Cigarettes/day Estradiol4-+ Bound Testosterone* -.06 .03 .008 .04 .14 .04 .46 .003 .02 Free Testosterone -.35 .14 .01 FSHO Si .20 .01 A2..34 +4 pgvfrj ivreforencceategoryismoderatedrinkerwsho consume 1-3cz othancilday. ng/dl @ failds&6mulatinghamorw mLYmi Thyroidstimulalinhgormone mU/tW TABLE 4.1.51LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE BOUND TESTOSTERONE-TSH- RATIO AMONG 112 MALE WORKERS. 3M CHEMOLRTF- PLANT, COTTAGE GROVE, MINNESOTA Variable SE(B) e:value Intercept 559.5 3SO.7 .12 TotalFluorin(eppm) 37.7 109.8 .73 Age (years) SMI (kg/m2) -1.1 62 .85 -9.8 8.9 .27 Cigarettes/day -.66 2.9 .82 AJcohol(<lozlday)* 74.9 78.3 .34 Free Testosterons' Estradiol@ FSHOO LH-* Proladn** 12.5 -1.6 47.1 5.7 -1.1 6.6 2.5 15.9 122 62 .06 .51 .004 .64 .85 R2=.29 +bound testasteronteothyroidstimulafinhgormone (mU/mo ratio #Reference categoryismoderade drinkerswho consume 1-3 cz ethanol/day. ngvdi @ pg/mi @@folliclestimulatinhgormone mU/mi ++ Thyroidstmulatinghormone (mU/mi) prolacfinnglmi ii3 TABLE 4.1-52LINEAR MULTIVARIATE REGRESSION MOOF-L OF FACTORS PREDICTING THE FREE TESTOSTERONE-TSHWORKE,Q.me RA'NO AMONG 112 FAALE 3M CHEMOLrrE PUKNT, C07TAGE GROVE, MINNESOTA Variable 0 SE(S) p-v-alue Intercept TotalFluorine(ppm) 15.65 -28 6.34 .02 .13 .03 Age (years) SMI (kgIM2) -29 .08 .003 -.01 .19 .94 Cigarettes/day Al=hol (<Ioztday)# Bound Testosterone* -.03 1.50 .01 .06 .65 1.64 .36 .003 .006 EstradiolO -.01 .05 .80 FSHOO .68 .33 .04 LH++ -.001 .25 .99 Prolacbn** -.18 .13 .17 R2= .37 +freetest=emne tothyroidstmulafinghormone (mU/rnQ r;ato #Referencecategoryismoderate drinkerswho consume 1-3 oz othanovday. ngtdl @ pg/mi @@foflicissUmula6ng hormone mU/m++Thymid stimulatinhgormone (mU/mi) prolacgnng/Mi 114 TABLE 4.1.53LINEAR MULTIVARIAT E REGRESSION MODF-L'OF FACTORS PREDICTING THE ESTRADIOL-TSH* RATIO AMONG 112 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable a SE(S) P-value Intercept 30-80 16.10 .06 TotalFluorine(ppm) -.425 .31 .18 Age (years) SMI (kg/M2) -.53 20 .01 .32 .46 zo Cigarettes/day .06 .14 .70 Alcohol(<loz/day)# 2-36 4.00 .55 Free Testosterone* -.28 .46 .55 Bound Tesiosterone' .009 .01 .42 FSHO .81 .81 .31 LH++ 20 .62 .75 Prolacfin*o -.07 .32 .83 R2= .15 +estracrtiotlhyroidstimulatinhgormone (mU/rnQratio #Referencecategoryismoderate drinkerswho consume 0 mg/di @folliclsetimulatinhgormone mU/ml ++ thyroisdtimulatinhgormone (mU/rni) prolaclinng/mi 1-3oz ethanovday. TABLE 4.1-54LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE BOUND TESTOSTERONE-FSH* RATIO AMONG 112 MALE WORKERS. 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA Variable B SE(s) p-value Intercept TotalRuodne (ppm) Age (years) BMI (kg/M2) Cigarettestday AJ=hol (<lcz/day)# Free Testosterone' LHOO Estradiol@ TSHProlaedn** 101.89 .66 -.13 -1.08 -.37 .29 6.87 -7.61 .77 8.90 -.03 6125 124 .75 1.70 ss 15.30 1.28 1.93 .47 7.03 1.20 .10 .60 .14 .53 .50 .98 .0001 .0002 .11 21 .97 R2=.So +bound testosteronteo follicslteimulatnghormone (mU/ml) rdo #Referencecategoryismoderate drinkerswho consume 1-3 cz othanoilday. *ng/di @Iuteinizinhgormone mU/mi @@estacfiolpgftl ++ Thyroidstmulgng hormone (mU/mQ prolacdnng/mi 116 TABLE 4.1.55LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE FREE TESTOSTERONE-FSH+ RATIO AMONG 112 MALE WORKERS. 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA Variable a SE(S) p-value Intercept 4.31 2.01 .03 TotalFluorine(ppm) -.04 .04 .27 Age (years) -.10 .02 .0001 SMI (kgVM2) .06 .06 .28 Cigarettes/day Alcohol(<lcz/day)# Bound Testosterone* LHOO Estradiol@ TSH++ Prolacfin** -.02 .18 .003 -.25 .03 .49 -.05 .02 .52 .001 .07 .04 .24 .04 .31 .74 .02 .0003 .27 .04 .23 R2= .43 +freetestosterontsofollicslteimulatinhgormone (mU/mQ #Reference categoryismoderate drinkerswho consume 0 ng/dl @Iuteinizinhgormone mU/mi @C&estracfiolpg/mi *+ Thyroidstmulafinghormone (mU/mi) proladn nglml rato 1-3 oz ethanovday. 117 TABLE 4.1.56LINEAR MULTIVARIATE REGRESSION PREDICTING THE ESTRADIOL.FSH- RATIO AMONG 3M CHEMOLRTE PLANT. COTTAGE GROVE. MODEL OF FACTORS 112 MALE WORKERS. MINNESOTA Variable intercept TotalFluorine(ppm) Age (years) SMI (kg/m2) -B 6.91 .006 -.19 .27 SE(S) 5.69 .006 .07 .16 R:value .23 .34 .008 .10 Cigareftes/day Aicohol(<1cz/day)# Free Testosteroneo Bound Testosterans' .03 .05 .57 .52 1.42 .71 .26 .16 .11 -.002 .004 .62 LHO TSH++ Proladn** -.49 .18 .009 .08 .65 .90 .04 .11 .70 R2a .26 +estracritooljoilicle.stmulatihnogrmone (mU/mo rauo #Reference categoryismoderate drinkersWho CONSUMO 0ng/di @Iuteirdzinhgormone mU/mi ++ Thyroidstimulatinhgormone (mU/mi) prolactinnglml 1-3 oz ethanovday. lie TABLE 4.1.57LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE BOUND TSH-FSH* RATIO AMONG 112 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable SE(S) p-value intercept .72 .33 .03 TotalRuorine (ppm) .01 .006 .14 Age (years) .002 .004 .59 SMI (kg/m2) .00007 .01 .94 Cigarettes/day Alcohol(<lozlday)# Estradiol-1-4 Bound Testosterone* Free Testosteronee Prolactn"O LMO -.003 -.16 -.001 -.0002 -.01 .002 ..03 .003 .08 .003 .0002 .009 .007 .01 .37 .05 S6 .28 15 .73 .005 R2= .26 ++pg/mi #Reference categoryismoderate drinkerswho consume 1-3 cz othanovday. ng/dl prolacdnngftnl lldenizinhgormone mU/ml + thyroidstimulatinhgormone (mU/mo tofollicslteimuiatnghormone (mU/mQ ratio TABLE 4.1-58LINEAR MULTIVARIATE REGRES'S'ION MODEL OF FACTORS PREDICTING THE TSH-LH-o RATIO AMONG 112 MALE WORKERS. 3M CHEMOLrrE PLANT. COTTAGE GROVE, MINNESOTA Variable SE(S) p-value intercept .32 .25 .21 TotalFluorine(ppm) .006 .005 .21 Age (years) .004 .003 .26 SMI (kg/M2) .008 .007 .27 Cigarettestday AJcohol(<Icz/day)* EstradiolBound Testosterone* FreeTestostemns" Proladn** FSHO -.001 -.07 -.004 -.0001 .007 .001 -.05 .002 .06 .002 .001 .007 .005 01 .53 26 .07 .84 .32 .91 .0001 R2u .26 ++pgftnl $Reference categoryismoderate drinkerswho consume 1-3 cz ethanovday. ng/cg **prolacdnngvml @ follicsltemulatng hormone mu/mi + Thyroidstimulafinhgormone (mU/ml)toluteni7inhgormone (mU/mQ ratio -1-20 TABLE 4.1.59LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE BOUND LH-FSH+ RATIO AMONG 112 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable 8 SE(i3) p-value intercept .60 .43 .17 TotalRuorine(ppm) -.0001 .009 .98 Age (years) BMI (kg/M2) .009 .005 .09 .01 .01 .40 Cigarettestday .0001 .004 .82 AJcohol(.cloz/day)# .04 .11 .71 Estradiol++ .004 .003 .18 Bound Testosterono* .0001 .0002 .18 Free Testostemne* -.004 .01 .78 Prolacfin** -.005 .009 .57 TSHO -.05 .05 .29 R2=.12 ++Pg/ML-#Reference categoryismoderate drinkerswho ng/di prolactn gftl thyroidstimulatinhgormone mU/ml consume 1-3 oz ethanovday. + ILnenizinhgormone (mU/mi) tofollicslteimulatinhgormone (mU/m I)ratio TABLE 4.1.60PEARSON CORRELA'NON COEFFICIENTS BETWEEN TOTAL SERUM FLUORIDE, AGE, BODY MASS INDEX (SMI).DAILY ALCOHOL USE, DAILY TOBACCO CONSUMPTION, AND LIPOPROTEINS am CHEMOLFTE PLANT, COTTAGE GROVE., MINNESOTA TOTAL FLUOFUDE fpom) AGE (years) BUI (kgtm2) ALCOHOL T08ACCO (ovday) (cigwday) CHOLESTEROL* LDL" .07 .2S c)-m.008 .02 .13 HOLS -.01 .03 TFUGLYCEFUDES" .09 .19 1 ow.04 .mgvdi 'low dons4 lipoprotain #high densitylipaprotein is 0-.OS .09 as om.0001 .06 -.008 28 --- D-.002 -.13 .18 Pm.06 -.09 .27 .07 .19 ow.004 I I -122 TABLE 4.1.61LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE CHOLESTEROL AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA V2riable s SE(s) p-vaiue Intercept 107.30 33.00 .002 TotalFluoride(pprn) .52 .67 .44 Cigarettes/day SMI (kg/M2) 1.12 .31 .0005 1.44 1.01 .16 Age (years) Ajcohol .77 .38 .05 low (<Ioz/day) -5.50 8.71 .53 nonresponse (NR) -13.53 14.75 .35 GGT (IU/dl)- .41 .12 .001 Bound Testosterene*" .03 .02 .07 $/Fteforrve category is moderate *garnma gluuvnyt transforass -ng/di drinkers who consuma 1-3 cz edwal/day. 123 TABLE 4.1-62LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE LOW DENSITY LIPOPROTIEN AMONG 111 MALE WORKEFIS. 3M CHEMOLITE PLANT. COTTAGE GROVF- MINNESOTA Variable 8 SE( intercept 73.93 32.00 .03 TotalFluodde(ppm) Z2 .65 .73 Cigarettestday BMI (kg/M2) .69 .30 .02 1.26 .95 ig Age (years) .37 .37 .32 AJcohol low (<Ioztday) -3.02 8.33 .71 nonresponse(NR) -10.85 13.93 .43 Prolacti(nng/ml) -1.sg .66 .02 Bound Testosterone(ng/dl) .04 .02 .0071 R2..ig *Roforar4coategoriysmodamtodrinkerwsho consunm1-3czstuncliday. -124 TABLF_ 4.1.63LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE HIGH DENSITY LIPOPROTIEN (HDL) AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGR= GROVE, MINNESOTA Variable a SE(S) value Intercept TotalFluoride(ppm) Alcohol# 65.00 -1.61 10.07 .77 .0001 .04 low (<lo7Jday) -9.92 3.51 .006 nonresponse(NR) -6.77 5.73 .24 low X Fluoride 1.62 .80 .04 NR X Fluoride' 2-05 1.63 .21 Age (years) -.004 12 .97 SMI (kg/M2) -.31 .29 .28 Cigarettes/day -.12 .09 .18 Bound Testosterons** .018 oo7 .009 FreeTestostemne" -.77 .28 .008 F12..17 *Referencceategoriysmodamto drinkerwsho =nsunn 1-3cz etwoltday. *interacttioomns betweentotaf)luoridaend aJ=hoJ=tegcfy ngtdl -125 TASLS 4.1-64LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE TRIGLYCERIDES AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVF- MINNESOTA Variable SE(B) tvalue Intercept -114.50 117.20 .33 TotalFluoride(ppm) 2.38 2.31 .15 Cigarettes/day 2.2B 1.05 .03 BMI (kg/M2) 6.07 3.39 .08 Age (years) A)cohot # 2.32 1.44 .11 low (<I oztday) -11.48 P-9.4 .70 nonresponse(NA) -19.94 49.03 .69 Free Testosterons" 7.34 3.37 .03 Bound Testosterone, -.21 .08 .009 freforoncceategoriysmoderatedankorswho =nsume 1-3cz et=ol/day. ng/dl 126 TABLE 4.1.65PEARSON CORRELATION CCREFFICIENTS BETWEEN TOTAL SERUM FLUORIDE. AGF- BODY MASS INDEX (SMI),DAILY ALCOHOL US-=, DAILY TOBACCO CONSUMPTION, AND HEPATIC PARAMT-TERS 3M CHEMOLRTE PLANT, C077AGE GROVE, MINNESOTA TOTAL FLUORINE AGE (Y"M) SMI fkgtm2) ALCOHOL TOBACCO (clvday) (eigs/day) SGOT' 12 -.11 SGFT" oi .01 20 .03 -.11 ow.02 GGTO ..04 .12 27 .15 .03 Da.004 AKPH" -.03 .27 .19 -.10 .26 c*.004 ow.04 *SERUM GLURAMIC OXALOACENC TRANSAMINASE IU/di SERUM GLLRTAMIC PYRUVIC TRANSAMINASE IU/di *GAMMA OLLTTALIYL TRANSFERASE PJ/d "ALKALINE PHOSPHATASF- ILI/di 127 TABLE 4.1.66PEARSON CORRELATION COEFFICIENTS BETWEEN ENZYMES, SERUM HORMONES, AND LIPOPROTEINS 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA HEPATIC graot SGPT GGT AKPH CHOLESTEROL' LDL" .07 25 .19 .09 [)..DOB o.-OS .02 13 DS ..Oos HDL# TRIGLYCEFDOES* ESTRADIOL+ FREE TESTOMRONE" BOUND TESTOSTEF;TONE" PROLACTIW .mg/di -low densitylimpraftin *highdensity lipoprotain *pg/mi ngldl -.01 .09 -.16 Dw.09 -.12 is .20 P-.03 .03 .19 ..04 -.14 -.10 -.15 -.13 .27 M ..23 DO.OI -.12 -.16 pw.Dg .07 -.DM -.03 -.12 .20 P-.03 TABLE 4.1.67PEARSON CORRELATION COEFFICIE14TS BETWEEN HEPATIC PARAMETERS 3M CHEMOLITE PLANT. COTTAGE GROVE, MINNESOTA sgor TG- PT'* GGT,V AKPH" S -G-0T SGPT .68 P..oooi GGT .43 Pa.0001 PM.*O6w0l A KFH- .04 I Pa.02 *SERUM GLIJTAMIC OXALOACENC TRANSAMINASE RJ/di SERUM GLUTAMIC PYRUVIC TRANSAmiNASE IU/di $GAMMA GLLJTAWYL TRANSFERASE IU/di "ALKAUNE PHOSPHATASrm ILI/di 129 TABLE 4.1.68SERUM GLUTAMIC OXALOACENC" TRANSAMINASE (SGOT), GLUTAMIC PYRUVIC TRANSAMINASE (SGPI),GAMMA GLUTAMYL TRANSFERASE (GGT), AND ALKALINE PHOSPHATASE (AKPH) BY TOTAL- SERUM FLUORINE 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA N TOTAL FLUORINE <1 PPM 23 >=1.3 es >3-1 0 16 >10-15 6 >1 5-26 5 TOTAL 115 <1 23 "1-3 as >3-10 is ,010-15 6 >1 5-26 5 TOTAL 115 <1 ppm 23 ;,=1-3 es ;..3-10 16 >1 0-15 6 >1 5-26 5 TOTAL 115 <1 ppm 23 >=1-3 es >3-10 is >1 ()-IS. 6 >1 5-26 s TOTAL 115 #univanate Anova MEAN so MEDIAN SGOT OU/dQ 22.5 4.1 22 24.1 8.6 23 25.8 14S 22.5 25.7 11.3 22-S 22.2 S.i 22 24.0 6.9 23 SGPT OU/dl) 47.7 10.7 46 51.3 302 45 53.0 14.0 SOS 732 532 52.5 44.6 8.6 42 51.7 26.8 47 AlkalinePhosphatase OWCO) 66.1 25.6 as OS.9 19.9 so 77.9 20.3 71-S 872 34.0 75.5 89.0 42.1 84 83.3 22-9 so GGT (IU/dt) 372 29A 27 32.4 26.7 2S 35.4 35.4 26 38.3 16.7 36.5 22.2 11.5 20 33.7 27.6 26 RANGE 13-29 10-74 17-77 17-47 14-27 10-77 3"9 4-263 29-40 38-177 34-54 4-263 43-IS3 36-137 64-123 61-193 41-153 39-IS3 6-117 &174 10-158 19-60 11-37 -r-174 TEST# Fwo.41 Fol.io pa.32 F.O.43 pa.78 FmO.39 po.81 130 TABLE 4.1.69SERUM GLUTAMIC OXALOACE71C TRANSAMINASE (SGOI) BY BODY MASS INDEX, AGE. SMOKING AND DRINKING STATUS 3M CHEMOLITE PLANT, COTTAGE GROVE. MINNESOTA N(%) -SGOT (luidl) MEAN SD-MEDIAN RANGE TESTN smi ,c2S 25-30 )..30 AGE 430 31-40 41-50 SI-60 Alcohol -clovd 1-3*Vd M"ing Tobacco smoker nonsmoker mosing TOTAL 41(35.7) 24 57(49.6) 23 17(14.8) 27 21(le.3) 25 48(41.7) 24 27(23.5) 22 19(16.S) 26 87(al.3) 26 20(18.7) 24 a 23 28(24.8) 24 85(75.2) 24 2 20 115 12.4 22 13-77 Fa.92 S.8 23 10-42 po.40 8.1 26 17-47 12.7 23 17-77 Fu.78 9.1 23 10-74 PE.51 5.4 23 13-40 7.8 23 1447 13.5 22 16-77 Fa.61 8.0 23 lt).74 P.." 4.3 21 19-31 S.-4 23 13.@n F-.02 11.0 22 10-42 Pam 3.5 20 17-47 *uravanate Anova 131 TABLE 4.1.70SERUM GLUTAMIC PYRUVIC-TRANSAMINASE (SGPT) ey BODY MASS INDEX, AGF- SMOKING AND DRINKING STATUS 3M CHEMOLITE PLANT, COTTAGE GROVF-. MINNESOTA N(%) MEAN SGPT(ILVdi) so MEDIAN RANGE TEST smi 42S 25-30 .o3O 4 1(35.7) 49 57(49.6) so 17(14.8) 64 AGE 40 31-40 41-50 51-60 Alcohol 'CloVd 1-3oVd Missing 21(16.3) 42 4s(41.7) 53 27(23.S) 47 19(le.5) 57 87(81.3) 53 20(la.7) 47 a 51 Tobacco smoker 28(24.8) 48 nonsmoker 85(752) 53 missing 2 49 TOTAL 115 35.4 41 14.2 49 32-8 55 lis 45 33.6 47 15.2 46 32.0 50 2935 47 16.9 46 10.9 52 ls2 47 29.6 48 2.S.5 49 29-2S3 4.9S 38-177 Fa2.1 p..12 31430 29-263 4-99 34-177 Fu.gi Fmm--Sl 29-293 4-99 3S-67 Fu.sa P...41 4-90 3C@-293 31-67 Fa.76 p...39 #univariatAenova 132 TABLE 4.1.71GAMMA GLUTAMYL TRANSFERASE (GG-n BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS 3M CHEMOLFTE PLANT, COTRAGE GROVE, MINNESOTA N(%) GGT (JU/di) MEAN SD MEDIAN RANGE TEST# smi -c2S 25-30 :b.30 41(35.7) 28 57(49.6) 34 17(14.8) 48 AGE ,c3C 31-40 41-50 si-so 21(18.3) 32 48(41.7) 31 27(23S) 33 .19(16.S) Alcohol .4io7Jd i,'4oz/d missing 87(81.3) 40 20(18.7) 32 a 41 Tobacco smokor 28(24.8) 36 nonsmoker BS(752) 32 missing 2 as TOTAL 115 31.1 23.1 28.6 23.4 32.7 17.2 29.3 2S.S 2S.3 50.4 21.3 26.3 103.2 17 6-174 F-3-c;4 Is B-158 p-.03 44 19-117 25 11-111 - FmlM 22 5-174 pm.36 29 8-72 35 11-117 3S 8-as F=1.64 26 6-174 pj6 23 12-lSS 33 S-89 FN-CZ 25 6-174 pa.46 as 12-158 #univariatAenova 133 TABLE 4.1.73ALINEAR MULTIVARIATE REGRESSION MODEL 1 OF FACTORS PREDICTING THE SERUM GLLITAMIC OXALOACETIC TRANSAMINASE (SGO-0 AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable 8 SE(i3) _p-value Intercept 26.71 7.1 .0003 TotalFluorine(ppm) -3.23 1.31 .02 IBMI(kgVM2) -.0004 23 .99 SMI X T. Fluorine* .12 .05 .015 Age (years) -.003 .08 .97 Alcohol low (<Ioz/day) .70 1.85 .71 nonresponse (NR) -1.10 3.10 .72 Cigarettes/day -.09 .07 .16 Prolactn(nglml) -.37 is .01 R2..17 #Reference category ismoderate drinkerswho cormuffw 1.3 cz adwal/day. interacdonterm bet"en totw "rum fluoride SMI. 135 TABLE 4.1.738 LINEAR MULTIVARIATE REGRESSION MODEL 2 OF FACTORS PREDICTING THE SERUM GLUTAMIC OXALOACETIC TRANSAMINASE (SGOT) AMONG 111 MALE WORKER$. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable a SE(S) p-value Intercept 27.71 6.22 .0001 TotalFluorine(ppm) -2.70 1.23 .02 SMI (kg/m2) -.09 .06 .11 SMI X T. Fluorine* .10 .04 .02 Age (years) Cigarettes/day -.02 .07 .74 -.11 .06 .11 AJeohol# low (<Iex/day) 1.84 1.61 .28 nonresponse(NR) -1.3 2.7 .64 Prolactn(nglml) GGT (IU/ol)" -.27 .13 .04 .13 .02 .0001 F;I;JS #Roforwrmcategoriysmoderatedrinkerwsho consuffw1-3 lnieracltieornmbetweentotwserum fluariadnod BMI Gamma gluulmynlmfomso et=ovday. .136 TABLE 4.1.73C LINEAR MULTIVARIATE REGRESSION MODEL 3 OF FACTORS PREDICTING THE SERUM GLUTAMIC OXALOACENC TRANSAMINASE (SGOT) AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA V2riable 13 SE(g) p-V21ue Intercept 120.60 4.0 .002 TotalFluorine(ppm) .63 .78 .42 SMI (kg/M2) -.07 .13 .58 SMI X T. Fluorine' -.03 .03 .34 Age (years) Cigarettes/day .06 .05 .23 -.02 .04 .45 Alcohol# low (Icoiztday) -.65, 1.03 .53 nonresponse (NR) -1.40 1.72 .42 ProlacUn(ng/mi) SGPT (IU/cl)" -.09 .06 29 24 .01 .0001 R2..74 #Rofof*rm catnory is modamto drinksis who consurro 1-3 intoracten term bohv"n total"rum fluoride and SIM Serum gkitamipcynivivcwulamimse otmwi/day. 137 TABLE 4.1.74ALINEAR MULTIVARIATE REGRESSION MODEL 1 OF FACTORS PREDICTING THE SERUM GLUTAMIC PYRUVIC TRANSAMINASE (SGP'T) AMONG 111 MALE WORKERS. 3M CHEMOLTTE PLANT, COTTAGE GROVE, MINNESOTA Variable 0 SE(B) Intercept TotalFluorine(ppm) BMI (kgVm2) SMI X T. Fluorine* Age (years). AJeohol# 58.13 -15.80 .30 .62 -24 24.26 4.58 .82 .17 .28 .02 .0008 .72 .0004 .39 low (.cclz/day) 5.54 6.36 .39 nonresponse (NR) 1.31 10.63 .90 Cigarettestday -27 23 24 Prolacfl(nngtrnl) -1.18 Si .02 ,vRefamr4mauegoryismoderatedrinkerwsho comuffw 1.3cz od=olday. intaracootnermbetweentotwserumfluorkiasnd BMI. 138 TABLE 4.1.74BLINEAR MULTIVARIATE REGRESSION MODEL 2 OF FACTORS PREDICTING THE SERUM GLUTAMIC PYRUVIC TRANSAMINASE (SGPT) AMONG I11 MALE WORKERS. 3M CHEMOLITF- PLANT, COTTAGE GROVE, MINNESOTA Variable a SE(B) o-vatue Intercept TotalFluorid(eppm) BMI (kg/m2) 62.09 -13.70 -.70 19.63 3.64 .66 .002 .0003 .30 SMI X T. Fluorine* S4 .14 Age (years) -.33 22 Cigareftestday -.027 .18 AJeohol# low (<loz/day) 10.02 5.09 nonresponse@(NR) .48 8.44 Prolacti(nngtml) -.74 .41 GGT (IU/dl)" .56 .07 A2.Sl ireforencceatnoryismoderatedrfnksmwho consumv 1-3cztriano". Interacttieornmbotw"n totw"rum fluoriadnedSMI Gamma glutamyfltmforno .0001 .14 .14 .05 .95 .07 .0001 139 TABLE 4.1.74C LINEAR MULTIVARIATE Fiffi@RESSION MODEL 3 OF FACTORS PREDICTING THE SERUM GLUTAMIC PYRUVIC TRANSAMINASE (SGPT) AMONG 111 MALE WORKERS. 3M CHEMOLrrE PLANT, COTTAGE GROVE, MINNESOTA Variable 8 SE(e) p-vatue Intercept TotalFluorine(ppm) BMI (kg/m2) SMI X T. Fluorine* Age (years) Cigarettes/day Alcohol# -1825 -6.65 .30 27 -.23 -.001 14.36 2.61 .45 .10 .16 .13 21 01 -51 .007 .14 .99 low (<Ioztday) 3.55 3.53 .32 nonresponse (NR) 4.39 5.91 .46 Prolacbn(ng/ml) -.11 29 .72 SGOT (IU/cn)- 2.85 .19 .0001 R2@-.7ra i#Roforonccoategoryismoderatedrinkerwsho =nsuffw 1-3cz etwol/day. lnt&racton term between totaslerum fluorweand IBM sorum glutamicexakmmtic Vanumk%aso -t4O TABLE 4.1.75ALINEAR MULTIVARIATE REGRESSION MODEL FACTORS PREDICTING THE GAMMA GLUTAMYL TRANSFERASE AMONG 111 MALE WORKERS. 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA 1 OF (GGT) Variable a SE(B) _p-vaue lnterr.ept -12.59 22.S2 Sa TotalFluorin(eppm) -1.93 2.11 .3S Alcohol low (<loz/day) -12-37 9.50 .20 nonresponse(NR) -28.13 IS.46 .07 low X Fluorine* I.sg 2.18 .47 NR X Fluodne' 13.90 4.48 .003 Age (years) .29 .30 .33 BMI (kgft2) 1.71 .76 .03 Cigarenes/day .09 .24 .72 R4.is areforenccaetegoIrsymoderatderlr*owrhso con*Uff1w-3czetwovday. linteracftaomnubetweetnotafl uoriadnedajcohcaltegory -141 TABLE 4.1.75B LINEAR MULTIVARIATE REGRESSION M**ODEL 2 OF FACTORS PREDICTING THE GAMMA GLUTAMYL TRANSFERASE (GGT) AMONG 111 MALE WORKERS. 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA Variable ommonomm@ - intercept n TotalFluoride(ppm) -58.78 -1.79 SE(S) 21.55 1.83 p-value -M .008 .33 Aicohol# low (<Ioz/day) -9.04 8.25 .28 nonresponse(NR) -20.08 13.49 .14 low X Fluorine* 1.39 1.90 .47 NR X Fluorine' Age (years) 12.18 .15 3.91 .002 26 .57 SMI (kg/M2) 1.30 .66 .05 Cigarettes/day .01 23 -96 Cholestero(lmg/dl) .15 .06 .02 SGOT (IU/di) 1.18 .24 .0001 R2. jo xroforon=categoryismodOM4 drinwm wft consurm 1.3 ormouday. 'interacttieorn= bohveontotaflluame and alcoholcategory 14r TABLE 4.1.75C LINEAR MULTIVARIATE REGRESSION MODEL FACTORS PREDICTING THE GAMMA GLUTAMYL TRANSFERASE AMONG 111 MALE WORKERS. 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA 3 OF (GGT). Variable s s (8) P-value Interoept TotalFluorine(ppm) Alcohol -32.39 -1.93 18.47 .08 I.SO .31 low (<Icz/day) -13.58 7.17 .06 nonresponse(NR) low X Fluorine" -26.68 .92 11.75 1.66 .025 .58 NR X Fluorine" Age (years) 12.04 2s 3.41 .0006 23 27 SMI (kgIM2) Si .59 .38 Cigarettestday .09 .20 .65 Cholestero(lmg/di) .12 .06 .04 SGPT (IU/cl)-- .59 .07 .0001 R2. M #Roforonccsategoriysmoderatedrinmrswho consunw1-3= oemal/day. inforac@toenrmsbet--,o,ntotaflluarkained alcohoclategory S@RJMglUtaMiPCYMNICnnsaminaso 143 TABLE 4.1.76LINEAR MULTIVARIATE REGRESSION MODEL 1 OF FACTORS PREDICTING THE ALKALINE PHOSPHATASE (AKPH) AMONG 111 MALE WORKERS. 3M CHEMOLFTE PLANT, COTTAGE GROVE. MINNESOTA variable a SEIB) kvalue intercept 24.50 15.69 .09 Tcmi Fluodne(ppm) -1.03 .43 .02 Cigarettestday -.06 .22 .79 Cigarettes/daXy Fluadne' m SMI (kg/m2) 1.10 OS .0001 .55 .05 Age (years) Alcohol .54 .22 .02 low(cloz/day) 5.78 4.90 .24 nonresponse(NR) 8.12 8.13 .32 R2.jl #Roforomccaotegoriysmoderatderinkewrhso consume1.3czothamovday. interacteornmbetweentotaslerumfluoriadnedcigareft*Wday. 144 TABLE 4.1.77PEARSON CORRELATION COEFFICIENTS BETWEEN TOTAL SERUM FLUORIDF- AGE. BODY MASS INDEX (BMI),DAILY ALCOHOL U-SE. DAILY TOBACCO CONSUMPTION, AND HEMATOLOGY PARAMETERS 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA HEMAGLOSIN' wacPMN COUNT* EOSINOPHILS LYMPHOCYTES MONOCYTES PLATLETS BASOPHILS BANDS TOTAL FLUORINE (Dom) -.07 .10 .05 -.10 .19 c@m.04 .05 .10 .04 AGE (yomrs)SMI (kWM2) ALCOHOL TOBACCO (Cztday)(clgwday) -.03 .04 -20 .20 C..04 c@0.008 .07 c7 -.07 .70 D..Oool .08 .09 -.10 .64 c@-.0001 .13 .05 .02 -- .23 -.05 .04 .15 .28 1 22-002 .0.4 .22 .21 .32 ow.02 - pm.03 C@m.0004 -.13 -.11 .05 .29 P..002 -.08 -.02 -.14 -.05 .26 -.14 ca.005 @whitsbloodcallcount 4 potymorphonuctealroukoc)nceount 145 TABLE 4.1.78LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE HEMAGLOBIN AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable 8 s E(B) p-value Intercept TotalFluorine(ppm)* 14.51 -.002 .67 .0009 .0001 .02 AJcohol# low (<Iovday) .22 @20 .27 nonresponse(NR) .56 .33 .09 Age (years) .001 .009 .88 SMI (kg/M2) 01 .02 .65 Cigarettes/day .01 .007 .20 Cigs/dayX Ruodne2*o .0003 .0001 .0005 Estradia(lpg/ml) 01 .006 .07 R4.23 'squworansfonmtcnoftotaRlucrido #Referencceategoriysmodoralsdrinkerwsho consumw1.3CZ othanaliday. interacttieornmbetweencigarettpeesrdayandsquam ransformatioonftotwtuarift 146 TABLE 4.1.79LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE MEAN CORPUSCULAR HEMOBLOBIN (MCH) AMONG 111 MALE WORKERS. 3M CHEMOLRTE PLANT. COTTAGE GROVE, MINNESOTA Variable 8 SE(S) o-value Intercept TotalFluorine(ppm) Al=hol# 31.65 .15 .95 .0001 .09 .10 IOW (<IOZ/d2Y) -29 .65 .65 nonresponse (NFI) .03 .01 .02 low X Fluorine -.16 .09 .08 NR X Fluorine -.04 .19 .80 Age (years) .03 .01 .02 BMI (kgVM2) -.07 .03 01 Cigarenes/day .02 .01 .13 Cigs/dayX FluoMe* .006 .003 .03 R2..24 #ReferencceategoryIsmod*raled6rkerswho consurm 1-3cz etwovday. "interacftieornms;alcohoclategoryby totaflluoridcei.garettepserday by totaflluoride 147 TABLE 4.1.1.80LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS pplEDICTING THE MEAN CORPUSCULAR VOLUME (MCV) AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVF-, MINNESOTA able 8 SE(B) e-value Intercept TotalFluorine(ppm) 8.74 -.04 2.SO .07 .0001 .52 AJeohol# low (,colz/day) -.61 nonresponse (NR) -.9s Age (years) .11 EIMI(kVM2) -.06 .78 .43 1.27 .46 .03 .002 .08 .05 Cigarettes/day .04 Cigs/dayX Fluorins' .02 TSH (mU/mi) .38 .03 21 .007 .004 .35 .29 RL 28 #Referencceategoriysmoderatediinkerwsho consunm1-3czod%anovday. interacdtoanffnc;igarellp"erday bytotwfluoride 148 TABLE 4.1.81LINEAR MUL71VARIATE REGRESSION MODEL OF FACTORS PREDICTING THE WHITE BLOOD CELL COUNT (WBC)- AMONG 111 MALE WORKERS. 3M CHEMOLrrE PLANT, COTTAGE GROVE, MINNESOTA Variable 6 SE(S) c@-value Intercept ToW Fluorine(ppm) Alcohol# 2.87 1.32 .03 .07 .10 .49 low (.cclz/day) .46 nonresponse(NR) -1.08 .74 low X Fluorine -.04 .10 NR X Fluorine .59 .21 Age (years) -.007 .02 SMI (kg/m2) .07 .04 Cigarettes/day .13 .01 FreeTestostemne(ng/cff) .04 .03 Lme .10 .04 .33 .15 .68 .006 .64 .0.5 .13 .02 RZ. .67 vacnooo #Rofofonccsategoriysmoderge drinkerwsho =nsume lutenizihnogrmonemU/ml 1-3cz *&mnol/day. 149 TABLS 4.1.82LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE POLYMORPHONUCLEAR LEUKOCUTE COUNT (POLY) AMONG 111 MALE WORKERS. 3M CHEMOLFTE PLANT, COTTAGE GROVE, MINNESOTA Variable a SEJB) P-.value Intercept 368 1151 75 TotalFluorin(eppm) 16S es .06 A)cohol low (.clczlday) 746 399 .06 nonresponse(NR) -49 651 .94 low X FluoHne -161 90 .08 NR X FluoHne 370 iss .05 Age (years) 6 14 .66 BMI (kg/m2) 45 33 .17 Cigarettestday 95 10 .0001 LH (mUlmQ++ 79 36 .03 Bound Testosteronse Free Testosterons -1.62 84 .8 .04 32 .01 FIZ-M.** UAoniZingharmcra'#Roftronm ca*gory.itmoderatodrinksmwho consuffw14 cz ee=ovday. rigvd) ISO TABLE 4.1.83LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE BAND COUNT (BAND) AMONG 111 MALE WORKERS. 3M CHEMOL[TE PLANT, COTTAGE GROVE, MINNESOTA variable S E(B) P-Val6e interoept -11.4 129.6 .93 TotalRuorine(ppm) -3.4 3.2 .30 AJcohol low (<lozlday) 78.2 40.3 .05 nonresponse(NR) 14.9 67.8 .83 Age (years) 1.0 1.8 .56 13MI(kg/m2) 2.2 4.6 .63 CigarettL*s/day 4.2 1.5 .005 R2..12 #Referencreategariysmoderatedrinkerwsho consuma 1-3cz ot=oitday. TABLE 4.1.84LINEAR MULTIVAFTIATE REGRESSION MODEL OF FACTORS PREDICTING THE LYMPHOCYTE COUNT (LYMPH) AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA Variable SE(B) P-vaiue intercept TotalFluorin(eppm) 2205.6 -342.7 611.1 125.3 .0005 .007 Alcohol# low (<loztday) -526.6 222.7 .02 nonresponse(NR) -977.1 355.7 .007 low X Ruorins 189.0 S2.3 .0005 NR X Fluodne 247.9 103.9 .02 Cigarettes/day 34.0 6.9 .0001 Cigs/dayX Fluorine* -3.3 1.45 .02 SMI (kg/M2) 1.58 19.6 .94 SMI X Fluorine" 7.15 4.1 .08 Age (years) -16.1 8.6 .06 Prolactn(ng/mi) 38.5 142 .008 TSH (mU/ml)* 170.4 772 .03 RZ. M #Referencceategoriysmoderatedrink*mwttcconsume 1-3oz othamovday. 'intorac@toenrmsalcohoclategorbyy totwfluoridcei;gamttssldbay totwguoride, SMI by totwfluoride. thyroisdfimulatihnogrmonemLYrAl IS2 TABLE 4.1.85LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE MONOCYTE COUNT (MONO) AMONG 111 MALFWORKERS. 3M CHEMOLTTE PLANT, COTTAGE GROVE, MINNESOTA Variable 5 SE(s) -0-value Intercept 397.4 198.9 .05 TotalFluorine(ppm) 110.4 38.6 .005 Ai=hol* low (<lcz/day) 132.1 53.8 .02 nonresponse(NR) 40.1 89.1 .66 Age (years) -.37 2.4 .88 BMI (kg/M2) -2.66 7.0 .70 SMI X Fluorine' -4.0 1.42 .006 Cigareftes/day 7.0 1.9 OOD4 LHO 13.9 6.8 .04 R2. M #Referencceategoriysmoderatedrinkerwsho consunm1-3czot=cvday. *interactieornmS.MI bytotaflluorkie. lutenizihnogmionsmU/rW 153 TABLE 4.1.86LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE EOSINOPHIL COUNT (EOS) AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GROVF, MINNESOTA Variable SE(B) P-Value Intercept SO.45 122.30 .68 TotalFluorine(ppm) -7.31 3.35 .03 AJcohol low (<Ioz/day) -12.10 37.91 .75 nonresponse(NR) 21.79 6225 .73 Age (years) 1.56 1.67 .3S SMI (kg/M2) 2.10 4.13 Si Cigarettes/day 3.04 1.69 .08 CigstdayX Fluorine' .62 .35 .08 TSHO 30.1 17.1 .08 R2..Is #Referencceatopfyismodamto drinkerwsho =nsurne 1-3cz atmiavday. Iinteracbioonrm.cigarettpeesrday by totwfluorido CDThyroisdtirnuwnghormonemU/rW 154 TABLE 4.1.87LINEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE PLATELET COUNT (PLATE) AMONG I II MALE WORKERS. 3M CHEMOLITE PLANT, COTTAGE GFIOVE, MINNESOTA Variable B SE(B) g-value Intercept TotalFluorin(eppm) AJeohol# low (<1cz/day) nonresponse(NR) Age (years) BMI (kg/m2) BMI X Fluohne' Cigarettes/day Cigs/dayX FlucHne* Prolact(ing/ml) Bound Testosterone** 264.7 29.8 8.2 .9 -1.3 1.1 -1.0 2.7 -.3 2.6 -.04 54.8 .0001 9.5 .002 13.3 .54 22.5 .97 .6 .04 1.7 .53 .4 .004 .6 .0001 .1 .04 .03 .09 .03 .10 R2..28 #Referencecategoryismoderatedrinkerwsho consuma 1-3cz othanovday. 'Interactitoenrms.SMI by totaflluoridcei.garettepserday by totalfluoride. mg/di 155 TABLE 4.1.88LINEAR MULTIVARIATE REGRESSION PREDICTING THE BASOPHIL COUNT (BASO) AMONG 3M CHEMOLFTE PLANT, COTTAGE GROVE, MODEL OF FACTORS 111 MALE WORKERS. MINNESOTA Variable s SE(s) o-vaiue Intercept 44.35 !>4.19 .42 T=l Fluorin(eppm)* -.03 .06 .61 Aicohol low (<loz/day) -1.73 13.61 .90 nonresponse.(NR) -.56 22-54 si Age (years) BMI (kg/M2) -.07 -.61 .68 .92 1.58 .70 Cigarettestday -.80 S2 .12 CigstdayX Fluo6ne2** .02 .007 .007 Bound Testosterone## -.07 .04 .06 FreeTestasterone## 2.5 LHO 5.5 1.6 .11 1.8 .002 $squar&eansfarmagmoftotw"nornfluar;do $Referencceategoriysmoderatedrinksmwho consunw 1-3ox odianovday. *interacttieornm,cigarettpee&rday by totaf)luoride. songidl 0 lutenizihnogrmonemLYml 156 4 4 Morta&Tphfe--q TABLE 4.2.1CHARACTERISTICS OF 749 FEMALE EMPLOYEES, 1947-1989. Chemica) Division Non chemicw Division Total number ofworkers 245 504 749 personyears of observation mean follow-up (years) mean age at employment (years) mean year of employment (years) mean year of death(years) mean age at death (years) 6029.0 24.9 28.8 1965.0 1981.3 58.7 13280.4 26.4 2S.9 1962.8 1979.2 54.4 19309.4 2S.8 27.6 1963.5 1979.6 55.4 157 TABLE 4.2.2CHARACTERISTICS OF 2788 MALE EMPLOYEES, 1947-1990. ChemicW Division Non cmemicaj Division number ofworkers 1339 1449 2788 personyears of observaton mean follow-up (years) mean age at employment (years) mean year of employment (years) mean yearof death(years) mean age at death (years) 33385.3 24.8 2S.6 1963.8 .1978.3 54.2 37732.4 26.0 28.9 1962.3 1978.1 Se.1 71117.7 25.5 27.3 1963.0 19782 56.4 TABLE 4.2.3VITAL STATUS AND CAUSE OF DEATH ASCERTAINMENT AMONG 749 FEMALE EMPLOYEES, 1947-1990. Vitalstatus chemicaj Division ra. % Non memical Division No. % Total No. % Alive 234 95.3 465 91.6 699 93.3 Dead, 11 4.7 39 8.4 so 6.7 Total 245 100.0 504 100.0 749 100 'Twodeaths o=urred outsidetMe U.S.witncause ofdeath ascefn@e d mt@m sources otherthan death certificates. TABLE 4.2.4VITAL STATUS AND CAUSE OF DEATH ASCERTAINMENT AMONG 2788 MALE EMPLOYEES, 1947-1989. Vitalstatus Chemical '@Division No. % Non chemical Division No. % ToW No. % Alive 1191 88.9 1249 86.2 2440 87.5 Dead* Total 148 11.1 200 13.8 348 12.5 1339 100.0 1449 100.0 2788 100.0 'mo deatnsoccurredoutsidethe U.S.withcause of death ascertainedfrom sourcesotherthan death certificates. TABLE 42.5 NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRS) AMONG 749 FEMALE EMPLOYEES, 1947-1989. Cause of Death obs Exo SMR 9 1 AJIcauses so Cancer 17 Gastrointestinal 2 Respiratory 4 Breast 3 GenhW 2 Lymphopoiefic 3 Heart disease 10 Cerebrovascular 3 Gastrointestinal 3 Injuries 4 Suicide 1 66.74 23.04 4.54 4.72 5.87 3.37 2-04 12-39 3.51 3.41 6.23 1.78 .7S .71 .9s Si .59 1.47 Si .86 .88 .64 .56 sr-,-.gg .42-1.14 .05-1.sg .26-2.43 .10-1.49 .07-2.14 .30-4.29 .49-1.29 .01-4.80 .18-2.57 .17-1.64 .01-3.13 160 TABLE 4.2.6NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRS) BY DURA-NON OF EMPLOYMENT AMONG FEMALE EMPLOYEES, 1947-1989. -Cause ot Deatm Obs ExD SMR 95% ci Duration510 years Alicauses 50 Cancer 17 Cardiovascular 18 66.74 23.04 22.00 .75 .56-.99 .71 .42-1.14 .82 .48-129 Dumtlon.-olDyears Allcauses 20 Cancer 6 Cardiovascular 8 26.62 9.42 10.27 .75 .64 .23-1.39 .78 .34-1.54 Abt)reviationused are:Obs, observed;Exp, expected;Cl,confiden@irtterv@al. 161 TABLE 42.7 NUMBERS OF DEATHS AND STANDARDIZED MORTALFTY RATIOS (SMRS) BY LATENCY AMONG FEMALE EMPLOYEES, 1947-1989. Cause of Death obs Exo- SMR 95% cl Latency;-.1y0ears Allcauses Cancer 41 56.94 .72 .52-.98 is 20.93 .76 .44-1.24 Cardiovas'cular 13 19.86 .65 .35-1.12 Latency@olSyears Allcauses 37 Cancer 14 Cardiovascular 13 49.37 18.25 17.79 .75 .53-1.03 .77 .42-1.29 .73 .39@-125 Latencp2O years Allcauses 29 Cancer 11 Cardiovascular 10 39.20 .74 .49@-1.06 14.47 .76 .38-1.36 14.67 .68 .33-1.25 Abbreviationussed are-O.bs, observed;Exp, expected;Cl,confidenceinterval. -462 0 TABLE 42.8 NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRS) BY ANY EMPLOYMENT IN THE CHEMICAL DIVISION AMONG FEMALE EMPLOYEES, 1947-1989. (:ausfoDteath ODS -Exo SMR 95% ci Not employed in co Allcauses 39 Cancer 14 Cardiovascular 13 Heartdisease 8 A)]Gi 2 Allrespiratory 2 Injuries 3 43.OS 15.46 13.82 7.69 2.23 2.23 1.48 .91 .64-1.24 .91 .49--l.52 .94 .50-1.61 1.04 .45-2.05 .90 .10-3.23 .90 .10-3-23 2.02 .41-5.90 employed inCO Allcauses 11 Cancer 3 Cardiovascular 5 Heartdisease 2 AllGi 1 Allrespiratory --1 Injuries 1 23.69 8.38 8.19 4.69 1.18 1.28 1.98 .46 .23-.83 .36 .07-1.05 .61 .20-1-43 .43 OS-1.54 .85 .01-4.73 .78 .01-4.81 Si .51-2.81 Abbreviabonsused are:Obs, observed;F-xp,expected;Cl,confidenceinto al; CD, ChemicalDivision. if)r3 TABLE 4.2.9NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRs), BASED ON U.S.WHRTE MALE RATES, AMONG 2788 MALE EMPLOYEES, 1947-1989. Cause of Death obs SMR 95% ci AJIcauses 347 Cancer. 103 Gastrointesbnal 24 Colon 9 Pancreas 8 Respiratory 31 Lung 29 Prostate 6 Testis 1 Bladder 3 Lymphopoietic 13 Cardiovascular 145 CHO 110 Cerebrovascular 10 AJIGastrointestinal 12 Ailrespiratory 13 Diabetes 8 Injuries 38 Suicide 473.56 107.80 25.94 9.11 5.33 40.53 38.72 5.10 .82 2.20 11.42 203.31 147.04 19.92 23.99 25.89 6.53 46.56 17.10 .73 .95 .93 .99 1.50 .76 .75 1.18 1.22 1.36 1.14 .71 .75 .50 .50 .50 1.23 .82 .70 .66-.81 .77-1.15 Sg-1.38 .45-1.88 .65-2.96 .52-1.09 .50-1.08 .43-2.S6 .02-6.80 .27-3.98 .54-1.64 .60-.84 .61-.90 .24-.92 .26-.87 .27-.86 .53-2.42 .58-1.12 .32-1.23 Abbreviationussed are:Obs, observed;Exp,expected;Cl,conticenfein-terval; CHD, coronaryand atherosclerotcheartdisease. 464 TABLE 42.10 NUMB ERS OF DEATHS AND STANDARDIZED MORTALRTY RATIOS (SMRs), BASED ON MINNESOTA WHRTE MALE RATES, AMONG 2788 MALE EMPLOYEES, 1947-1989. -Cau-seof Death Ot)S Exo SMR 95% Cl Allcauses 347 Cancer 103 Gastrointestinal 24 Colon 9 Pancreas 8 Respiratory 31 Lung 29 Prostate 6 Testis 1 Bladder 3 Lymphopoiefic 13 Cardiovascular 14S CHO 110 Cerebrovascular 10 AllGastrointestinal 12 Allrespiratory 13 Diabetes a Injuries 38 Suicide 12 450.79 97.29 26.78 9.42 5.58 30.42 28.94 6.07 .92 2.18 12.07 212.19 159.09 24.66 21.13 21.75 S.S2 47.74 15.09 77 1.05 .90 .96 1.43 1.02 1.00 .99 1.09 1.37 1.09 .68 .69 .60 .57 .60 1.23 .80 .79 .69-.86 .86-1.27 .57-1.33 .44-i.si .62-2.83 Ml.45 .67-1.44 .36-2.15 .01-6.OS .28-4.01 .57-i.s4 SS-.80 .57-.83 .32-1.02 .29-.99 .32-1.06 S3-2.42 .56-1.08 .41-1.39 Abbreviationsused are:Obs, observed; Exp, expected,Cl,confidin-c-e-'inte-rval; CHD, coronaryand atherosderoteheartcrisease. TABLE 42.11 NUMBERS OF DEATHS AND STANDARDIZED MORTALrry FLATIOS (SMRS) BY LATENCY, BASED ON MINNESOTA WHRTE MALFRATES, AMONG MALE EMPLOYEES, 1947-1989. --Ca-useot Deatm LATENCY @!10 YEARS Ot)S Exo SMR 95% ci Allcauses 299 Cancer 98 Gastrointestinal 24 Pancreas 8 Respiratory 29 Lung 27 Skin 3 ProsMe 6 Bladder 3 Lymphopoietc 11 Cardiovascular 130 AllGastrointestinal 8 AJIrespiratory 11 Diabetes 8 Injuries 21 Suicide 11 398.27 88.71 24.78 5.20 28.81 27.44 l.S3 5.94 1.75 10.03 195.91 18.58 20.16 5.37 27.61 10.19 .77 1.10 .97 1.54 1.01 .98 1.96 1.01 1.72 1.10 .66 .43 .55 1.49 .76 1.08 .68-.86 .90-1.35 .62-1.44 .66-3.03 .67-1.45 .65-1.43 .39-5.73 .37-2.20 .34-S.ol S5-1.96 .55-.79 .19-.86 27-.98 .64-2.94 .47-1.16 .54-1.93 At)t)reviatiounssed are:Obs, ooserved; Exp, expe=ec; Cl,contioe-n-c-e'i"-nte-rval; CHO, coronaryand atheroscieroteheartcrisease. --166 TABLE 42.12 NUMBERS OF DEATHS AND STANDARDIZED MORTALrrY RATIOS (SMRS) BY LATENCY. BASED ON MINNESOTA WHITE MALE RATES, AMONG MALE EMPLOYEES, 1947-1989. Cause of Death LATENCY?. 15 YEARS o@b-S Exo SMR 95c-/I. AJIcauses 266 Cancer 90 GastrointestinaJ 24 Pancreas 8 Respiratory 27 Lung 25 Skin 3 Prostate 5 Bladder 3 Lymphopoietc 9 Cardiova= tar 119 AJIGastrointesiinal a Allrespiratory 9 Diabetes 7 lniu6es 23 Suicide 9 344 80.64 22-63 4.72 26.71 25.45 via 5.73 i.i.57e.w 8.68 178.25 16.17 18.60 4.54 29.21 7.47 .77 1.12 1.06 1.69 1.01 .98 2.33 .87 1.53 1.04 .67 .49 .48 1.54 .79 1.21 .68-87 .90-1.37 .68-1.51 .73-3.32 S7-1.47 .64-1.45 .47-6.80 .28-2.0.4 .37-4.47 .47-1.97 -55-.80 .21-.97 .22-.92 .62-3.18 -SO-1.16 .55-2.29 @@b-r@e-@i'abuosnesd are:Obs, observed; Exp, expected; Cl,confide-n-ca-"int-erva); CHb, coronatyand atheroscierotihceartdisease. 167 TABLE 42.13 NUMBERS OF DEATHS AND STANE)ARDIZEE) MORTALrrY RATIOS (SMRS) By LATF_NCY, BASED ON MINNESOTA WHITE MALE RATES, AMONG L4ALE EMPLOYEES, 1947-1989. -Cause of Death LATENCY 2:20 YEARS obs Ext) SMR 95% Cl Allcauses 216 Cancer 73 Gastrointestinal 15 Pancreas 4 Respiratory 25 Lung 23 Skin 2 Prostate 5 Bladder 3 Lymphopoietic 7 Cardiovascular 99 AllGastrointestinal a AJIrespiratory 9 Diabetes 7 Injuries 13 Suicide 7 286.9 68.74 19.33 4.06 23.06 21.06 .98 5.29 1.75 7.01 151.80 12.90 16.3 3.65 19.47 5.01 .75 1.06 .77 .99 1.08 1.05 2.02 .95 1.72 .99 1.06 .62 .55 1.92 S7 1.40 .66-.86 .83-1.34 .43-1.28 .27-2.52 .70-1.60 .6rz,--5l7 .23-7.34 .30-2.21 .34-S.01 .39@-2-03 .83-1.34 .27-1.21 25-1.05 .77-3.95 .36-1.14 .56-2.80 Aborevi'alionused are:Obs, observed; Exp, expectecl;Cl,confidiin-cnetervai-, CHO, coronaryand atheroseleroteheartcrisease. -168 TABLE 4.2.14NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRS) BY DURATION OF EMPLOYMENT, BASED ON MINNESOTA WHITE MALE RATES, AMONG MALE EMPLOYEES, 1947-1989. Cause of Dealn DURATION 2 5 YEARS Obs F--Xo SMR 95% ci Allcauses 256 Cancer 80 Gasirointestinal 22 Colon a Pancreas 7 Respiratory 25 Lung 23 Prostate 4 Bladder 2 Brain 3 Lymphopoietic 6 Cardiovascular 114 CHD 90 Cerebrovascular 6 AllGastrointestinal 7 Allrespiratory 9 Diabetes a Injuries 29 Suicide 9 321.20 72.21 20.21 7.10 4.22 23.72 22.10 4.47 1.68 2.51 8.41 159.50 120.20 18.44 15.20 IS.30 4.53 36.60 8.81 .80 1.11 1.09 1.13 1.66 1.08 1.04 .84 1.19 1.20 .71 .71 .75 .33 .46 ss 1.77 .79 1.02 .7(@-.90 .88-1.38 .68-1.6S .49-2= .66-3.42 .70-1.59 .66-1.56 .23-2.15 .13-429 .24-1.50 .2r-P-1.55 .59@-.86 SO-.92 .12-.71 is-.95 .25-1.05 .76-3.48 .53-1.14 .47-1.94 Abbrev-iatiounssed are:Obs, observed;Exp,expecied;Cl,contidenceinterval; cmo, coronaryand atherosclerothiecartcrisease. TABLE 4.2.15NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRS) BY DURATION OF EMPLOYMENT, BASED ON MINNESOTA WHITE MALE RATES, AMONG MALE EMPLOYEES. 1947-1989. -Cause of Death DURATION z 10 YEARS Obs Exo SMR 95% Ui- Ailcauses 203 Cancer 67 Gastrointestinal 20 Colon 7 Pancreas 6 Respiratory 22 Lung 20 Prostate 4 Bladder 1 Brain 3 Lymphopoietc 5 Cardiovascular 92 CHD 75 Cerebmvascular s AllGastrointestinW 4 Allrespiratory 7 Diabetes 8 Injuries 19 Suicide 257.30 59.36 16.7S 5.92 3.50 19.38 18.47 4.20 1.44 1.89 6.58 132.13 99.75 15.49 11.96 13.80 3.49 23.46 5.88 .79 1.13 1.19 1.18 1.71 1.13 1.08 .9s .69 1.59 .76 .70 .73 .32 .33 Si 2.29 .68 1.36 .66-.91 .87-1.43 .73-i.S4 .47-2.44 .63-3.71 .71-1.72 .66-1.67 .26-2.44 .01-3.IB5 .32-4.64 .24-1.77 .56-.85 .57-.92 .10-.75 .09-.86 .2(@-1.05 .99-4.51 .34-1.22 .59-2.68 Abbrevirionsused are:Obs, observed;Exp,expected;Cf.confidenceinterval; CHD, coronaryand atherosclerotihceartcfisease. 1'70 TABLE 4.2.16NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRS) BY DURATION OF EMPLOYMENT. BASED ON MINNESOTA WHITE MALE RATES, AMONG MALE EMPLOYEES, 1947-1989. -dau-seof Deatri DURATION a 20 YEARS Obs P-xo SMR 95% ci Allcauses 10.4 Cancer 35 Gastrointestinal 10 Colon 5 Pancreas 1 Respiratory 11 Lung 10 Prostate 2 Bladder 1 Brain 1 Lymphopoietic 4 Cardiovascular 48 CHO 39 Cerebrovascular 1 AllGastrointestinal 2 Allrespiratory 5 Diabetes 5 Injuries 2 Suicide 3 152.36 37.31 10.52 3.77 2.21 12.69 12.10 2.83 .94 1.03 3.82 80.6 61.25 9.13 6.87 8.61 1 *94 6.61 Z54 .68 .94 .95 1.33 .45 .87 .83 .71 1.06 .97 1.05 s8 .64 .11 .29 Ss 2-58 .30 i.ia .56-.83 .43-3.09 .01-2.52 .43-1.55 .40-1.52 .08-2.55 .01-5.91 .01-5.40 .28-6.02 .19-1.36 .45-.87 -OD-.61 .03-1.OS .19-1.36 .83-6.02 .03-1.09 24-3.4S Abbreviationsused are:Obs, observed; Exp, expected; Cl,confidenEi-i@nt-erval; CHO, coronaryand atherosclerotihceartcgsease. 171 TABLE 42.17 NUMBERS RATIOS (SMRS), BASED 1339 MALE EMPLOYEES OF DEATHS AND STANDARDIZED MORTALRTY ON MINNESOTA WHITE MALE RATES, AMONG EVER EMPLOYED IN THE CHEMICAL DIVISION, 1947-1989. Cause of Death obs Exo SMR 95% C) Allcauses 148 Cancer 40 Gastmintestinal 9 Colon 4 Pancreas 4 Respiratory 12 Lung 11 Prostate 4 Testis 1 Bladder 1 Lymphopcietic 5 Cardiovascular 54 CHD 43 Cerebrovasailar 4 AllGastrointestinal a AJIrespiratory 7 06t*tes 3 Injuries 31 Suicide 10 172.96 36.31 9.77 3.46 2.04 11.26 10.70 1.97 .44 .75 4.76 76.65 57.74 8.S3 8.27 7.770 2.55 31.72 6.99 .86 1.10 .92 1.15 1.96 1.07 1.03 2.03 2.2S 1.33 1.05 .70 .74 .47 .97 .91 1.18 .98 1.43 .72-1.01 .79@-1.50 .42-1.75 .31-4.01 .53-5.01 .55-1.86 .51-1.84 .55-4.Sg .03-12.66 .02-7.40 .34-2.45 .53-.92 .54-1.00 .13-1.20 .42-1.91 .36-1.87 .24-3.44 .66-1.39 .68-2.63 Abbreviabonsused are:Obs, observed; Exp, expected; Cl,contidegc-e-i"n-te-rval; CHO. coronaryand atheroscierotcheartcrisease. 172 TABLE 42-18 NUMBERS OF DEATHS AND STANDARDIZED MORTALrrY RATIOS (SMRS). BASED ON MINNESOTA WHr7E MALE RATES, AMONG 1449 MALE EMPLOYEES NEVER EMPLOYED INTHE CHEMICAL DIVISION, 1947-1989. Cause of Death ObS Ext) SMR 95% Cl Ailcauses 200 Cancer 63 Gastrointestinal is Colon 5 Pancreas 4 Respiratory 19 Lung 18 Prostate 2 Testis 0 Bladder 2 Lymphopoietic a Cardiovascular 91 CHD 67 Cerebrovascular 6 AllGastrointestinaJ 4 AJ)respiratory 6 Diabetes 5 Injuries 23 Suicide 2 29125 67.56 16.46 5.79 3.37 25.58 24.44 3.45 .43 1.45 6.89 129.77 93.84 12.93 14.56 16.77 4.05 38.28 9.26 .69 .93 .91 .89 1.19 .74 .74 .58 .00 1.38 1.16 .70 .71 .46 27 .36 1.24 .60 .60 .59-.79 .72-1.19 .51-1.50 28-P-.Ol .32-3.04 .4S-1.16 .44-1.16 .07-2-09 .00-8.45 .164.99 .50-229 .56-.86 .55-.91 .17-1.01 .07-.70 .13-.78 .40-2.88 .38-.98 .02-.78 Abbreviationsused are:Obs, observed; Exp, expected;Cl,confidenR"iln-te-rval; CHD, coronaryand athemscieratehW cfisease. !t73 TABLE 42.19 NUMBERS OF DEATHS AND STANDARDIZED MORTALRT-Y RATIOS (SMRS) BY LATENCY, BASED ON MINNESOTA WHITE MALE RATES, AMONG MALE EMPLOYEES NEVER EMPLOYED IN THE CHEMICAL DIVISION, 1947-1989. Cause of Death LATENCY k 15 YEARS obs Exo SMR 95% Cl AJIcauses 161 Cancer 56 Gastrointestinal is Colon 5 Pancreas 4 Respiratory 17 Lung is Prostate 2 Lymphopoietic 5 Cardiov:,=lar 75 AllGastrointestinal 4 Allrespiratory 4 Diabetes 5 injuries 7 216.10 50.70 14.37 5.13 2.99 16.73 15.94 3.86 5.40 113.60 9.80 12.14 2.82 11.17 .75 1.10 1.05 .98 1.34 1.02 1.00 S2 .93 .66 .41 .33 1.77 .63 .63-.87 .83-1.43 .59-1.73 .31-2.28 .36-3.43 .59@-1.67 .57-1-63 .06-1.87 .30-2.16 .52-.83 .11-1.05 .09-.84 .57-4.14 .25-1.29 At)breviatiounsed are:Obs, observed;Exp, expected;01, confidenceinterval; CHO, coronaryand atherosclerotihceartdisease:CO. ChemicW Division. 174 TABLE 4.2.21NUMBERS OF DEATHS ANO STANDAROIZED MORTALITY RATIOS (SMRS) By DURATION OF EMPLOYMENT, BASED ON MINNESOTA WHITE MALE RATES, AMONG MALE EMPLOYEES EVER EMPLOYED IN THE CHEMICAL DMSION, 1947-1989. -Cause of Death DURATION z!10YEARS C)bs Exo SMR 95% ci AJIcauses 90 Cancer 27 Gastrointestinal 6 Colon 3 Pancreas 2 Respiratory a Lung 7 Prostate 3 LymphopoieVc 4 Cardiovascular 38 AJIrespiratory 3 Diabetes 3 Injuries 7 108.7 24.4 6.92 2.47 1.46 8.16 7.78 1.55 2.84 54-60 5.42 1.51 8.11 .83 1.08 .87 im 1.37 .98 .90 1.94 1.41 .70 .55 1.99 .86 .67-1-02 .71-1.58 .22-1.89 .24-3.55 .7S-4.86 .42-1.93 .36-1.86 .39-5.66 .38-3-61 .50-.97 .11-1.62 .40-5.80 .35-1.78 Aboreviationussed are:Obs, observed; Exp, expected;Cl,conficenceinterv;ui,. CHO, coronaryand atheroscierobeheartdisease;CO, Chemical Division. IY6 TABLE 4.2.22NUMBERS OF.DF-ATHS AND STANDARDIZ=D MORTALITY RATIOS (SMRS) BY DURATION OF EMPLOYMENT. BASED ON MINNESOTA WHITE MALE RATES. AMONG MALE EMPLOYEES EVER EMPLOYED IN THE CHEMICAL DIVISION, 1947-1989. e ot Deatri mom@ Allcauses Cancer Gasirointestinal Colon Pancreas Respiratory Lung Prostate Lymphopoietc Cardiovascular AU respiratory Diabetes Injuries DURATION 2 20YF-ARS obs Exo SMR 45 6629 .68 16 1621 .99 3 4.53 .66 3 1.61 l.B4 0 .96 .00 5 5.57 .90 4 5.31 .75 2 1.10 1.82 3 1.67 1.79 18 34.48 .52 2 3.52 .57 2 S4 2.37 2 3.27 .61 95% rl .50-.91 .56-120 .13-1.94 .37-5.38 D-3.84 29-2.09 .2D-1.93 .2D-6.58 .36-S24 .31-.83 .06-2.52 .27-e.56 .07-2.21 Aboreviationussed are:Obs, observec:Exp, expected:Cl,contidenceinterval; CHD, coronaryand atheroscierotcheartdisease;CD, Chemical Division. TABLE 4.2.23NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRS) BY DURATION OF EMPLOYMENT, BASED ON MINNESOTA WHITE MALE RATES. AMONG MALE EMPLOYEES NEVER EMPLOYED IN THE CHEMICAL DIVISION, 1947-1989. Cause of Deatm DURATION z!LOYF-ARS Obs F-XD SMR 9 5 r/.-Ci Allcauses 113 Cancer 40 Gastrointestinal 14 Colon 4 Pancreas .4 Respiratory 14 Lung 13 Prostate 1 Lymphopoietc 1 Cardiovascular 54 Allrespiratory 4 Diabetes 5 Injuries 4 148.60 34.43 9.82 3.45 2.04 11.22 10.69 2.65 3.47 78.31 8.35 1.98 7.96 .76 1.16 1.43 1.16 1.96 125 1.22 .38 27 .69 .48 2-S2 .50 .63-.91 .83-1.58 .78-2.39 .31-2.97 s3-s.ol .68-2.09 .65-2.08 .01-2.10 .01-1.49 S2-.90 .13-127 .81-3.a7 0.14-129 Abbreviationsused are:Obs, observed; Exp,expected;Cl,confidenceinterval; CHO, coronaryand atheroscierotcheartdisease;CD, Chemical Division. 178 TABLE 42.24 NUMBERS OF DEATHS AND STANDARDIZED MORTALITY RATIOS (SMRS) BY DURATION OF EMPLOYMENT. BASED ON MINNESOTA WHITE MALE RATES, AMONG MALE EMPLOYEES NEVER EMPLOYED IN THE CHEMICAL DIVISION, 1947-1989. -Cause of Death DURATION a 20YF-ARS Obs Exo SMR 95% Cl Allcauses 59 Cancer 19 Gastrointestinal 7 Colon 2 Pancreas 1 F;espiratory 6 Lung 6 Prostate 0 Lymphopoietic 1 Cardiovascular 30 Allrespiratory 3 Diabetes 3 Injuries 0 86.1 21.09 5.99 2.IS 1.25 7.12 6.79 1.73 2.15 46.14 5.43 1.10 3.34 .69 .90 1.17 .93 .80 .84 .88 .00 .46 .65 .59 2.74 .00 .52-.88 .54-1.41 .47-2.41 .10-3.32 .01-4.45 .31-1.83 .32-1.92 .0-2.12 .01-2.58 .44-.93 .12-1.72 .55-8.00 .00-1.14 Abbreviationussed are:Obs, observed;Exp,expected;C), confidenfien-terval; CHO, coronaryand atherosclerotihceartcriseaseC;O, Chemical Division. 179 TABLE 4.2.25AGE ADJUSTED STANDARDIZED RATE RATIOS (SRRS) FOR ALL CAUSE. CANCER, AND CARDIOVASCULAR MORTALITY BY DURATION OF EMPLOYMENT. AMONG MALE EMPLOYEES, 1947-1989. Cause ofdeath allcauses allr-ancers allcardiovetci,ar SRRO .81 1.04 .91 95%Cl .63-1.03 .67-1.61 .62-1.34 Abbreviationsused are:SRR, standardizedrateratio;Cl, confidenceinterval. lessthan 10 years ofemploymerttas referentcategory TABLE 4.2.26AGE ADJUSTED STANDARDIZED R-ATE RATIOS (SRRS) FOR ALL CAUSF- CANCER, LUNG CANCER. GI CANCER, AND CARDIOVASCULAR MORTALITY BY EVERINEVER EMPLOYED IN THE CHEMICAL DIVISION, AMONG MALE EMPLOYEES. 1947-1989. Cause of death allcauses allcancers lung cancer GI cancer allcarcriovascular SRRI 1.18 1.10 1.09 1.16 1.05 95%Cl (.95,1.47) (.74.1.65) (.67.2.31) (.50,2.69) (.76,1.48) Abbreviationussed are:SRR, standardizedrateratio;Cl,confidenceintervalG;i, gastrointestnal. Never employed intheChemical Divisionas referentcategory TABLE 4.2.27AGE STRATIFIED, YEARS OF FOLLOW-UP ADJUSTED RATE RATIOS (RRMH) FOR ALL CAUSE, CANCER, AND CARDIOVASCULAR MORTALRTY BY EVER)NEVER EMPLOYED IN THE CHEMICAL DIVISION, AMONG MALE EMPLOYEES, 1947-1989. Ace at emoloyment RRMH* 95%Cl Allcauses 15-19 years 20-29 years 30-39 years 40-65 years AJIcancers 15-19 years 20-29 years 30-39 years 4045 years AJIcarcriovascular 15-19 years 20-29 years 30-39 years 40-65 years_ 1.22 .95 .95 1.02 .95 .72 1.10 .66 1.40 .86 .78 1.11 (.62.2.40) (.68,1.32) (.611.50) (.72-1.44) (.21,4.34) (.38,1.35) (.62.1.90) (.27,1.60) (.39.5.03) (.44,1.67) (.44,1.29) (.73.1.82) Abbreviatonussedare:RRMH, Mantel-Haenszealge acrjusterdateratioCl, confidenceinterval. Adjustedforyears offollow-upand stratfiebdy fourage categories. Never employed inthe ChemieW Divisionas referentcategory -182 TABLE 4.2.28 AGE STRATIFIED, YEARS OF FOLLOW-UP ADJUSTED RATE RATIOS (RRMH) FOR ALL CAUSE. CANCER, AND CARDIOVASCULAR MORTALITY BY DURATION OF EMPLOYMENT IN THE CHEMICAL DIVISION, AMONG MALE EMPLOYEES, 1947-1989. Ace at emoloyment RRMH* 95%Cl AJIcauses 15-19 years 20-29 years 30-39 years 40-65 years Allcancers 15-19 years. 20-29 years 30-39 years 40-65 years A)lcardiovascular 15-19 years 20-29 years 30-39 years. 40-65 years 1.30 1.16 2.16 1.69 2.17 .84 1.75 2.67 .88 1.38 3.53 1.50 (.58.3.28) (1.52,2.70) (1.07.2.60) (.40,11.61) (.44,1.Sl) (.95.3.21) (.995.7.14) (.25,3.33) (.73.2.60) (1-68,6.21) .1.81,2.79) Abbreviationussed are:RRMH, Mantel-Haenszelage acrjustedrateratio;Cl, confidenceinterval. Acrjustefdoryearsoffollow-upand stabfiedby fourage categories. lessthan 10 yearsemployment as referentcategory 183 TABLE 42-29 PROPORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE ALL CAUSE MORTALITY AMONG 2788 MALE WORKERS. Variable 0 SE(S) p-value RR# Year offirsetmployment Age atfirsetmployment* Durationof employment* -.55 .079 -.34 .009 .006 .001 .0001 .0001 .0001 .946 1.082 .967 Months inchemical division .001 .001 24 1.001 Abbreviationsused are: B,regressionparameter;SE(B),stanc3arderrorofthe slopeparameter;RR. relativreisk. relativreiskforone unitchange inindependentvariable years TABLE 4.2.30PROPORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE CARDIOVASCULAR MORTALRTY AMONG 2788 MALE WORKERS. Variable Year offirsetmployment Age atfirsetmployment* Duraton ofemploymento Months inchemical division- 6 -.075 .119 .230 .0002 SE(B) .016 .009 .294 .001 p@-value .001 .0001 .45 .85 RRO m@ .928 1.126 .852 1.00 Abbreviabonsused are: B,regressionparameter-S,E(B),stanclarceirroroftne slopeparaineterR,R, relativreisk. relativreiskforone unitchange inindependent variable years 184 TABLE 4.2.31PFIOPORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE CANCER MORTALITY AMONG 2788 MALE WORKERS. Variable 8 SE@B) p-vaiue RRS Year offirsetmployment Age atfirsetmployment' ourabon of employment' Months inchemical division -.031 .078 -.028 .002 .019 .011 .009 .001 .11 .0001 .002 .20 .969 1.081 .972 1.002 At)t)reviatiounssed are: S. regressionparameter;SE(B), manciarcierrorof me slopeparameter; RR, relativreis@- relativreiskforone unitchange inindependentveable years TABLE 42-32 PROPOR71ONAL HAZARD REGRESSION FACTORS PREDICTING THE LUNG CANCER MORTALITY MALE WORKERS. MODEL OF AMONG 2788 Variable a SE(S) p-vaiue RRO Year offirsetmployment -.019 .042 .65 .981 Age atfirsetmployment* .070 .021 .001 1.072 Duraton ofemployment* -.062 .133 r>4 .940 Months inchemical division -.026 .016 .11 .975 Abbreviationsused are:8,regressionparameter;SE(B), standard errorofthe slope parameter; RR, relativreisk. relativreiskforone unitchange inindependent variable year3 iss -- TABLE 42.33 PROPORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE GI CANCER MORTALITY AMONG 2788 MALE WORKERS. Variable 5 SE(S) P-value RR* Year of firsetmployment .015 .035 Age atfirsetmployment* .130 .021 Durationofemploymento .005 .020 Months inchemical division .001 .002 .71 1.015 .001 1.139 .82 1.005 .56 1.001 Abbreviationsused are:(31Z-a,a-strointesbna8J,;regressionparameter;SE(B), standarderroroftheslopeparameter,RR, relativreisk. relativreiskforone unitchange inindependentvariable years TABLE 4Z34 PR OPORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE PROSTATE CANCER MORTALITY AMONG 2788 MALE WORKERS. Variable 5 SE(S) p-value RRO Year offirsetmployment .010 .081 .90 1.011 Age at firsetmployment' .082 .045 .06 1.085 Duraton of employment* -.070 .052 .18 .932 Months inchemical division .010 .005 .03 1.010 Abbreviationussed are: 8,regressionparameter;SE(a),standarderrorofihe slopeparameter,RR, retailvreisk. relativreiskforone unitchange inindependentvariable years TABLE 42.35 PROPORT10NAL HA7-ARD REGRESSION MODEL OF FACTORS PREDICTING THE PANCFIEATIC CANCER MORTALITY AMONG 2788 MALE WORKERS. Variable s SE(B) p-value RRID Year offirsetmployment Age atfirsetmployment* Durationof employment* Months inchemical division D46 .136 -.012 -.002 .066 .034 .035 .006 .48 .0001 .73 .73 1.047 1.146 .988 .998 Abbreviationussed are: 8,regressionparameter;SE(B),stancarcierrorofthe slopeparameter;RR, relativdesk. relativreiskforone unitchange inindependentvariable *years TABLE 42.36 PRO PORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE-DIABETES MELLRTUS MORTALRTY AMONG 2788 MALE WORKERS. Variable 8 SE(13) p-value RRO Year offirsetmployment -.40S .221 Age atfirsetmployment* .092 D44 Durationof employment* .009 .030 Months in chemical division -.001 .004 .06 .667 .04 1.096 .75 1.009 .76 .999 Abbreviationussed are: 8,regressionparameter;SE(B),standard errorofthe slopeparameter;RR, relativreisk. # relativreiskforone unitchange inindependentvariable *years 187 TABLE 4.2.37PROPORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE ALL CAUSE MORTALITY AMONG 749 FEMALE WORKERS. Vahable a SE(B) p-value RRO Year of firsetmployment -.02 .03 .41 .977 Age atfirsetmployment* .08 .02 .0001 1.08 Duraton ofemployment' 2-10 years >1 0 years Months inchemical division 1.31 -54 01 3.72 .8s .57 .14 2.33 -.003 .004 .48 .997 Abbreviationussed are:8,regressionparameter;SE(B),stanclarcejrrorofthe slopeparameter:RR, relativreisk. relativreiskforone unitchange in independentvariable years TABLE 42.38 PROPORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE CAFIDIOVASCULAR MORTALITY AMONG 749 FEMALE WORKERS. Varia-ble 0 SE(G) p-vaiue RRO Year offirsetmployment -.034 .048 Age atfirsetmployment' .119 .024 Durationofemploymento -.011 .025 Months inchemical division -.015 .017 .4a .0001 .67 .37 .966 1.126 .986 .985 Abbreviationussed are: B.regressionparameter;SE(B),stancard errorofthe slopeparameter;RR, relativreisk. relativreiskforone unitchange inindependent variable years lea TABLE 4.2.39PROPORTIONAL HAZARD REGRESSION MODEL OF FACTORS PREDICTING THE CANCER MORTALFTY AMONG 749 FEMALE WORKERS. Variable mo@ Year of firsetmployment Age atfirsetmployment' s -.043 .085 SE(B) .053 .025 p-value .42 .001 RR# .958 1.089 Durationofemployment* -.021 .025 .65 .980 Months inchemical division .001 .005 .87 1.001 Atoreviationussec are: 8,regressionparameter:SE(B),stancard errorottne slopeparameter;RR, relativreisk. relativreiskforone unitchange inindependentvariable years LLI z 0 ui t; 0= 0 Co LU LU ui U. FIGURE 1.Free testosteroneand totalserum fluorine 1990 3M Chemolite study AGE-30 SM-25 20-1 AGE-30 5Mo3S AGE-SOOMWZS ls- AGE-SO SMlm3S 10. ------------------ S. 0I 0 io io 30 TOTAL FLUORINE (ppm) 190 Figure2.Bound testosteroneand totalserum fluorine 1990 3M Cheniolltestudy 600. Soo. LU z 0 LLI 400. 0 LUL2 300. 0- C& 0 c1.0 do z 0 200. AGE.W SMI.25 AGE=30BM633 AGE.50 SMI-25 AGESSO SM145 0 10 2@ 30 TOTAL FLUORINE (ppm) 191 Figure3.Estradioland total3erum fluorine 1990 3M Chemolite study 60- so- 0. a 40- 30' 20 AGE=3 0 BMW25 AGE=30 BM-3S 1 ;o TOTAL FLUORINE (ppm) 192 Figure4.Lutent7Jnghorrmne and totalserum fluorine 1990 3M Chemolite study 10. 9- 7- CL 41 32 0 20 30 TOTAL FLUORINE (ppm) 193 Figure5. Folilefsetimulatinghormone and totalserum fluorine 1990 3M Chemolite study 4- 3- co 2U. 0i 0 0 io 30 TOTAL FLUORINE (ppm) 194 Figure6. Prolactinand totalserum fluorine 1490 3M Chemilts study so- 40- z 30- 8 0 zo. 10 0-. 0 ...... L4oderatedrinkers Lightdrinkers Nonmspondents 10 io 30 TOTAL FLUORINE (ppm) Figure7.Thyroidstimulatinghormone and totalserum fluorine 1990 3M Chemollte study 31 E 0 000000 0 io io --"30 TOTAL FLUORINE (ppm) 196 FigureS.Bound to treetestosteroneratioand totalserum fluorine 1990 3M Chemlite study 45. 1 40. 0 cc LU x u. ui 0 ui I- AGGEI=30 BMlxL30 NS LD 25 0 10 20 30 TOTAL FLUORINE (ppm) 197 prototyPpeeroxisomperoliferaMatyorr.egulasttoeroidogenesbiysbindintgoa member ofa new familyOfCytosolicreceptors(PPAR) belongingto thenuclear hormone receptorsuperfamilyand tranS2CIivatint;he transcriptioonfgenes involvedin steroidsynthesis119-121. PFOA was positivelayssociatedwiththe TSITF and E/TF ratios.PFOA binding tosex hormone bindinfglobulin(SHBG) may have produced changes inthe bound to freetestosteroneratio.However, thiswould resultin a change inthe TB17F thatisinthe oppositedirectiotno the observed associationbetween PFOA and TB/TF. The associationsofPFOA withthese ratiosare consistent witha mechanism thatinvotyesdecreased productionoftestostemne and increasedproductionof estradiol. The HPG axisofoldermen appeared to be more susceptibleto PFOA compared tothatofyounger men. No animaldata has been reportedconceming age relatedsensitivitoythe effectsof PFOA. However, theonset ofLeydig cell tumors has been reportedto occur lateintwo year ratfeecfingstudies122. This findinmgay representincreasedsusceptib!14forhormonal alterationisnaged rats,Furtheranimal researchisneeded todefineany age relatedsusceptibility factors. Pmlactinlevelswere positivelayssociatedwithtotalserum fluorideinparticipants who reportedmoderate drinking(1-3 drinkstday).Since the functionof pmlacon inmen isuncertain,the clinicaslignificancoefsuch an associationIsunclear. Alcoholingestionisa stimulusforpmiactinsecretion.The mechanism ofthis effecatppears to be mediated by alterationisncalcium mediated signal iransductionpathways 123.This suggests thatthe elevationofprolacdn associatedwithPFOA and alcoholmay be mediated by alterationsincalcium mediated eventssuch as transmembrane signaltransductionpathways. Thyroidstimulatinghormone was positivelayssociatedwithtotalserum fluoride. Animalstudieshave shown thatperfluomdecanoic acid depressed peripheral thyroidhormone levelswithoutproducing a hypothymid response 75,79, ioi. In the present study, peripheral thyroid hormone levels were not assayed. Therefore, itis not possible to assess whether the observed associafion between 200 PFOA and TSH couldbe a direchtypothalamiceffact.a phu" regulatory affecto,ran effectmediatedby changes inpe@pheralthyroidhormone levels. insummary, thisisthefirsrteportofhertnonalchanges associatedwithPFOA in humans. The presentfindingisnhumans are consistenwtiththose previously reportedinanimalstudies'g..Tne consistenftindingisncludelow free testosteronei,ncreasedestradioaln,d unchanged LH. Rodent and human reproductiveendocrinesyslems diffegrreatlyy,etthesuggested effectsof PFOA aresimilarI.nlighotfthe observedsimilaritiienseffecti.tistemptingto speculatethatPFOA may effecthe humans and rodentsreproductivendocrine systemthroughthesame mechanism. A hypothesisthatPFOA altersa calcium mediatedcellulasrignaltransductiopnathway,such as the CAMP or inasitol thphosphatemediated second messenger response,may providea unified mechanism forthe multiplleociofputativeffects. No adversehealtheffectshave been observed inexposed 11.The presentstudy didnotexamine adversehealtheffectsa,lthoughseveraladverse outcomes associatedwithhormonal alterationasre possible.The etiologyofa number of cancersinciudinagdenocarcinomas ofthe prostatee,ndometrium, colon,rectum, pancreasand breast,have been linkedtochanges inendogenous hormones 124. Cancers inthisetiologiccategoryinciuds. Perfluamocianoiaccidisnota genotoxiccarcinogeninstandard assays 9 However,PFOA isa nongenotoxicmdent carcinogen.Inmts exposured to PFOA overa two yearperiod,therewas associatedincreasein Leydigcell tumors 125. Leycrigcelltumors have been observed inassociationwithother pemxisome proliferatoirnsrats122. tthas been hypothesizedthatchronically elevatedL.Hproduced testiculanreoplasms 10-122. However, in PFOA treated rats.LH was notelevated.Thismay be due to estmgens feedback inhibitioans discussed-previousloyr,due toinsufficieenxtperimentalinductiontime le. Alternativelayn,othermechanism may have been operativeInproducingLoydig celltumors.Exogenous esiradioplroduces Leydigcelltumors inmice 126. High estradjollevelsare associatedwithLeydigcelllumars inboth ratsand humans 127.121. The tissuesurroundingthe Leydig celladenomas alsoproduces increasedestrogens127.High estradiolmay be a stimulusforLeycrigcell 201 proliferaboannd tumor formation.Thishypothesisissupportedby the observatiotnhat eslracfisotlimulateTsGF-A secretioinnLeydigcells TGF-A bindsto EGF receptorsexpressedon Leycricgells129 and stimulatescell pm[iferaton.The hormonalchanges associatedwithPFOA may be a mechanism fornongenctoxiccarcinogenesisT.he roleofPFOA inhuman nongenotoxiccarcinogenesisneeds to be clarified. Adequate androgen levelsare necessaryformaintenanceof potency, spermatogenesis,libidaond male reproductivoergans. Low testosteroneand highestrogensmay decreaselibidoa,nd fert!Winmajes Im. Decreased male fertilmiatyy be one potentaladverse outcome ofPFOA. The reproductive toxicitoyfPFOA has notbeen exiensivelsytudied.No studieshave been conductedinhumans. PFOA was nottoratageniicnrats 9.131,132 . No,gtdverse effectosn fertilwietrye notedforfemaleratsina teralogenesistudy9. male mts were notstudied.No otherreproductivsetudiesinanimalshave been reported. Studiesofhuman reproductivfeunctionare needed sincehuman reproductive processesarethoughtto be more sensitivteo xenobioteinsuftscompared to otheranimalspecies 33 9 1.3Chnle!@terotT.dglycL*@d@,cnnd Upt)prot@ins Cholesterolt,6glyceddes,and LDL were notsignificantalsysociatedwithPFOA. The lackofassociatioonfPFOA withcholesteroolrt6glycerideissconsistent withobservationsinexperimentalanimalmodels. No animalstucrieosf PFORS effecotn LDL are availablfeorcomparison. The are no studiesinhumans concerningthe relationshiopf PFOA withLDL, cholesterolo,rtriglyoorides. InlighdtrinkersP,FOA had littelfefecton HDL levels.Inmoderate drinkers, increasinPgFOA reducedHDL The putativeffectsofPFOA and alcoholmay be mediated by alteratioonfa common HDL regulatorpyrocess. The findings arelimitedby the smallnumber ofexposed workers,the*rjmiterdange oftotal fluoridvealues,and the limitatioonfsthe study design.The conclusionand suggestedmechanism must be consideredpreliminary. 202 The mechanism by which PFOA modifiesthe alcohol@HDL relationshicpould be mediatedby alterationisnfattyacid metabolism or fat*tyacid binding.Alcohol intakeinducesspecificP450 metabolicenzymes incjuding2EI and afterslipid metabolism l-"4.PFOA induces a specificP450 Al familyofmetabolicenzymes and alterslipidmetabolism inrodents.The jointeffectofalcoholand PFOA on P450 mediated lipidmetabolism could alterHOL dynamics. The primary structuroef PFOA suggests thatPFOA could affectthe ligandbindingoffatty acidinhepatocytesand HDLS. The competitionforNEFA bindingshes could reduce the effectofalcoholon HOL levels.Studiesofthe jointeffectofPFOA and alcoholon HDL may clariftyhe regulatorymechanisms forHDL The decrease inHDL associatedwithincreasingPFOA levelsmay be ciinically significantI.na meta-analysisof 12 prospectivestudiesof the relationship between HDL levelsand coronaryheartdisease (CHO), Gordon estimatedthat the change inCHD riskassociatedwitha one mg/di change inHDL levelis approximatelythe same as the change inriskassociatedwitha 2-4 mg/di change inLDL level135.The predicteddrop inHDL fora moderate drinkingparbciparit witha totalfluoridoef20 ppm is30 mgldl. A change ofthisorderof magnitude may have a measurable impact on theoccurrence of cardiovasculardisease. in theretrospectivmeortalitsytudy,therewas no increase inmortalityfmm cardiovasculadrisease. However, thereare a limitednumber ofworkers with totalserum fluohnslevelsof20 ppm ofmore. Any increaseinriskfor cardiovasculadriseasesamong a smallgmup of highlyexposed workers may not be readilaypparent.ina study of allChemolite or CD employees. Further researchisneeded toconfirmand ciariftyhe associationbetween PFOA and HOL level.Futurestudiescould tesithe hypothesisthatPFOA and alcoholjointly alterNEFA metabolism resultingina decrease in HDL and an increasein cardiovasculamrorbidityand mortalityrisksforexposed workers who drink alcohol. 5-1.4Henatic Parameters Changes inSGOT (AST) and SGPT (ALT) appear to be associatedwithtotal serum fluoridteh.rough an interactiownith adiposity.In obese participantsb,oth SGOT and SGPT increasedwithincreasingPFOA. However, theredid not 203 appearlo be an independerdeffectofPFOA on SGOT afteradjustingforSGPT. The findingsare limitebdy the smallnumber ofexposed workers,the limited range oftotalfluohdevajues,and the previousldyiscussed limitaiionosfthe studydesign.The conclusionand suggestedmechanisms must be considered preliminary. Compared toSGOT, SGPT isa relativeslpyecifimcarker forhepaiocyte disruptio1n31 . The lackofassociatioonf SGOT withPFOA afteradjustinfgor SGPT suggeststhatthe liveirsthe primarysource forthe smallPFOA associatedchanges intransaminases.SinceSGPT isa enzyme associatedwith theER membrane, the increaseinSGPT may have been the resultofPFOA associatedER proliferatiohnm.ay indicatae disruptioinnthe integritoyf hepaiocylemembranes which allowsincreasedreleaseofcytosolihcepatic enzymes. The tissuespecifiecffectsuggestedforhepatocytomembranes could be due to a higherhepaticconcentrationof PFOA. Uver injuryisgenerallyconsideredto be a muhifactodalprocess. There ifs evidencethatinteractionbsetween andogenous and exogenous factorsplaya roleinhopatotoxicftoybserved inworkers.137.The modfica:fioonfthe adposity. SGPT associatiobny PFOA suggests thatthe mechanisms oftransaminase elevatiomnay be linked.Obesityhas been associatedwithelevationof transaminasesas wellas clinicalilmyportanthepatiti1s3S.139.The observation thatsome obese individualesvidence littaldeiposityeffectwhileotherobese individuadlesvelophepaticfibrosihsas notbeen explained.fthas been hypothesizedthatmetabolicpolymorphisms orotherhopatotoxinexposure may playa role140. Animal studiesand limitedhuman data suggest thatxenobiotics, such as certainsolventsand alcohol,may potentiattehe offectsof other hepatotoxins141.142. Following-thimsodel, PFOA may directloyr incfirecdy potentiattehe hepatotoxiceffectof obesity. A mhochondrialsiteofPFOA actionmay occur.The mitochondriaplaysan essentiarloleintatmetabolism. Disruptionofmitochondrialfunctioncan produce impairmentofmitochondriaoixidationoflongchain and medium chainfattyacids. StudiesoffattyacidmetabolisminPFOA exposed humans have not been carried out VaJproicacid,an eightcarbon branched chainfattyacid(2pmpyl-pentanoic -204 acid)thatimpairsmftochondriaflunctionand fattyacid metabolism, Isan example ofa hepatotoxicxenobioticofsimilarcart)onrructure to PFOA 113. Commercial graciePFOA cornainsisomers withc@=on backbones idenbcajto valproicacids_ structur3e9. The valproate-likiesomers ofPFOA could produce toxicitsyimilar to that ofvalproate.The modificationofthe associationbetween PFOA and the transaminasesby adipositycould be mediated by disturbancesof mitochonddai fattyacidmetabolism inhumans. GGT increasedas alcoholuse increased.The increaseinGGT was smalleras PFOA increased.This associationwas independent ofchanges inSGOT, SGPT, and AKPH. Perfluorooctanoiaccid may inhibitthe hepatotoxiceffectsof alcohol.The GGT-alcohol dose response relationshiipsthought to be secondary totheinductionand increasedreleaseof GGT. Increasedserum GGT levels indicateproliferatioofnthe endoplasmic reticulumand inductionof cytochrome P450 system, leakagefrom hepatocytes,or injuryto otherfissues144-147. Perfluorooctanoiaccidmay decrease serum GGT by attiadngcallmembrane permeabilityb,y reducingthe alcoholmediated inductionofGGT, or by changing alcoholoxidationpathways and reducingthe productionoftoxicintermediates such as aestaidehyde. Perfluorooctanoiacidwas negatvely associatedwithAKPH innon-smoksm. In workerswho smoke greaterthan fivecigarettepser day, PFOA was positively associatedwithAKPH. The associationofAKPH with PFOA was independent of GGT, transaminases,and hormones. Smoking has been reportedto elevate AKPH 148.The mechanism of this effect is thought to be the resuft of AKPH inductionby compounds incigarettesmoke. The jointeffectof smoking and PFOA could increase the inductionof AKPH. Insummary, the associationsbetween PFOA and hepaticenzymes are weak and are notclinicallsyignificantI.nthe retrospectivmeortalitystudy.therewas no increase'dinmortalityassocaftewdithliverdisease. Future studiesof the effects of PFOA may elucidatepossiblemechanisms ofactionof nongenotoxic hepatic carcinogens.The hepaticenzyme resultsare illustratiovfethe problem of extrapolatinfgindingsobserved inmdent animal models to otherspecies, includinghumans Inhumans, PFOA does not cause the dramatichepatic 20S effectsobserved inmdents. Instead,the observed associationsmay resultfrom PFOA modificatioonfthe hepaticeffectsofobesity,alcoholconsumption, and smoking. Each ofthese factorsare independentlyassociatedwith hepatotoxicity. Furtherstudiesofthe jointeffectsof PFOA and SMI, alcohol,and smoking on hepaticenzymes are needed. 5.1.5HernatolocyCounts and Pararneters PFOA was weakly,butsignificantalsysociatedwith hemoglobin levels,MCV, and MCH. The associationsbetween PFOA and erythrocyteindicesappeared to be mediated throughinteractionwsithsmoking, and ;>ertws alcoholconsumption. The findingsinanimal studies9-"-10are consistentwitha decrease inred cell volume and a largerdecrease inred cellnumber. Together,these changes produce an increaseincellulahremoglobin concentration.The estimated changes inerythrocytesindicesare notof clinicaslignificanceover the range of iota)serum fluoride.However. these findingssuggest thatfurtherstudiesof the effectofPFOA on red cellregulationand functionare needed. The findingsare limitebdy thesmall number of exposed workers,the limitedrange oftotalfluoride values,and the previouslydiscussed limitationosfthe study design. Pharmacologicaldoses ofandmgens increaseerythrocytenumber and mass but produce littclheange inMCV or MCH Is'-152. The mechanisms by which androgens increasehemoglobin appear to.mediatedby modulatingthe erythmpoietirnesponsivenessof mufti-potentisatlem cellsand by stlmulating erythropoletpirnoduction151.153-155. Inphysiologicdoses, the effectof testostemne on erythrocyteincriceisscontroversialP.alacios ot a].and Cunningham et W. reportedthattestasteroneisassoci2tedwith a small Increase inhemoglobin,bLnno change inMCV or MCH 156-157. Mauss et al.reportedno change inredcellindicesforphysiologiclevelsoftestosterons158. inthe presentstudy,the testosteronelevelwas not stmngly or significantrleylatedto theredcellincricesE.stradiolwas wealdy associationwithHGB but not MCV or MCH. The effectofphysiologicestradialevelson the mwe hematological system ispoorlyunderstood.Tellet al.reportedthatthe effectofsmoking on red cellindiceswas differenitn male than infemale adolescents 159.This suggests thatestrogenlevelsmay play a roleinthe effeelof xenobictieson red cell 206 indices.Taken togethert,he evidencesuggeststhatthe ar-sociatiobnetween PFOA and erylhrocytiendiceswas notmediated by the PFOA associated changes intestosteroneb,ut may have been mediated in partby changes in estradiol. Thyroidhormone was associatedwithchanges inHGB and MCV. A decreased availabiloifttyhyroxin(T4)to myxedma levelsproduces a mildmacrocytic anemia inhumans. The increasedcellvolume isdue toajterationisnlipid depositioinnerythrocytmeembranes thato=rs duringineffectiveerythropoiesis 160.TSH confounded the assodationbetween PFOA and MCV. Decrease inT4 couldexplainsome ofthe increaseinMCV and TSH. However, PFOA appeared tohave an independentand oppositeeffecton MCV. Therefore.the association between PFOA and changes inred cellindiceswas probablynot relatedto changes inthyroidfunction. The immune system effectsassociatedwithPFOA presenta complex picture. As expected,smoking had a strongeffecton loukocytocounts. Smoking modifiedtheassociatiobnetween callcouritand PFOA fortotallymphocytes, eosinophilsp,lateletasnd basophils.However, smoking didnot modifythe estimatedPFOA effecotn WBC, PMN, band count,or monocyte count. AJeohol modifiedtheassociatiobnetween PFOA and cellcountforWBC, PMN, and lymphocytecount Adipositymodifiedthe associatiobnetween PFOA and lymphocytecount,monocyto count,and platelectounl Taken togethert,his preliminardyata suggeststhatPFOA isassociatedwithchanges inperipherajy feukocytecounts.The negativeassociatiownithlymphocytecount isconsistent withthelymphocyteseffectsobserved inprimatestudies.PFOA couldmodulate cellcountsby alterintghe effectsofsmoking,alcoholconsumption,and adiposity on peripheraleukocytecounts. The magnitude ofthe WBO and PMN associationwsere not ciinicals,ilgynificant froman infectioudsiseaseperspective.IncreasedWBC ispositvelyassociated withmortalitfyrom allcauses, cardiovasculadriseases,cancer and myocardial infarcti1o6n1-198f.tisunclearifthe alteratioinnWBC isa consequence of,orthe cause of,ongoing pathologicaplrocesses. Judgment as tothe clinicarlelevance ofthePFOA associatedchanges inWBC must awah furtherstudy. 207 low& Adipositymodifiedthe associationbetween cell-countand PFOA formonocytes. Alcoholand cigarettceonsumption were independent determinantsinthe present study.Monocyte countshave been reportedtobe low inmassivelyobese individual1s69.The biologicablasisforthese effectsare notclear.The univahateand jointeffectsofadiposityand PFOA on monocyle count may a fruitfaurlea forfutureresearch. Inthepresentstudy,thecomplex relationshipbsetween lymphocyte =unt, PFOA, alcoholuse, cigaretteuse, and body mass may have been the resultof thedifferentieaflfecton T cellsubsets. Inorderto clarittyhese associations, specifiscubsets need to be measured. The associabon oflymphocyte subsets withdisease endpointshave yettobe cia6fied.The interpretationf the observed associationrequiresfurtherresearch. Smoking was negativelyassociatedwithbasophilcourd. As PFOA level increased,the smoking effectwas diminished.Tayloret al.reportedan increase inblood basophilsinsmokers compared to nonsmokers 170. Walteret al.studied smokers and nonsmokers and found thataWe smoking causes degranulation and lossofbasophils.However, chronicsmoking Isassociatedwithan elevamed basophilcounl 171-174N.o attemptwas made to prohibitsubjectsfrom smoking priortothe timeofblood sampling. The negativeassociationobserved Inthli' studymay reflecrtecentsmoking by participantpsriorto blood drawing. The apparentreductioninthe degranulatingeffectofsmoking suggests thatPFOA may interacwtiththe basophildegranulafflonprocess. Exposure to PFOA may be associatedwithchanges inimmune functionbeyond simplechanges incellnumber. The avid oxygen binding by PFCs may afterthe effectivenesosfperoxidatickillinbgy PMNS. Cytokine signalingisimportariitn immune functionand could be afteredby PFOA exposure 173. The response to antigenbindingdepends upon rearrangement of membrane proteins.Changes inthemembrane physicalcharacteristicpsroduced by the poteristurfactant actionof PFOA could alterimmune responses. More research isneeded inthe areaof PFOA immunotoxicity.The findingsof the present study need to be confirmed.Lymphocyte could be immunophenotyped using wellestablishedflow 208 cytometrymethods 171. 177. The standard immunctoxicologicassessment deflned by the NationaJToxicologyProgram 178 should be carriedout forPFOA. Smoking has been observed toincreaseplatelentumber, survival,adhesiveness, activationa,nd aggregationwhen exposed toADP 179-184. Adhesiveness may change as a resultofthe effectsofsmoking on noneste6fiedf2ttyacids (NEFA). Smoking increasesNEFA which may compete with PFOA forplatelemtembrane bindingsites.Such competitioncouldajterthesmoking associatedincreasein platelectount.This hypothesizedmechanism could be testedby invitro modelingofplateleftunctioninthe presence ofNEFA and PFOA- The relationshibpetween obesityand platelectount has notbeen wellstudied.BMI has been reportedto be negativelyassociatedMth plateletcount 185.The mechanism forthiseffectisnot clear,but may be relatedtochanges inNEFA associatedwithobesity.Thus, theeffectmodificationofthe PFOA offac by smoking and obesitymay have resultedfmm a common effecton NEFAChanges inplatelectount have been associatedwith riskforcardiovascular disease 186.187. Directand indirectmechanisms have been hypothesized for the observed increaseindisease occurrence. Thus, PFOA associatedchanges in platelectount may be a marker forincreasedcardiovasculardisease 6sk. Furtherstudy ofpotentialeffectsof PFOA on plateletcount and functionare needed. S.1.6 Total Fluorine Smoking and totalserum fluorinweere weakly associatedinparticipantsT.he adjustedestimateforthe differencein mean fluorinebetween smokers and nonsmokers was small(0.1ppm) and probablynot ofbiologicaslignificance. Smoking intensitwyas not significantcloyrrelatedwithtotalserum fluorinelevels. h isunlikelythat smoking affectsthe pharrnacokinetics of serum fluohne or PFOA- R isunlikelythat smoking was a primary mute for absorpton of PFOA. Exposure reduction does not need to awah the resultsof haure studies. In rodents,removal from exposure resultsinthe reversaj of the marked hepatic responses to PFOA 188. Interventionto reduce the PFOA body burdens of employees would prevent any potentialadverse effectsin the future. The reductionof exposure is especially important since PFOA has an unusually long 209 biologichaallf-lifAe.significanrteductioinnbody burden wfllrequireyears of reduced exposure. F.17 Mpttir)dolocicCAolnsiderations Sip-s Giventhe occupationasltudysettingt,hevoluntaryparticipatioann,d the re-quirementfsorbloodsample collectiotnh.e overallparticipatiwoans unexpectgdlyhigh.Past medicalscreeningprograms atChemolits had participatiroantesof60% to70%. The presentstudys participatiroante exceeded 80%. Given the highparticipationno,n-responsebiasislikeltyobe small. Selectionbiasisan importantvaliditiyssueforcross sectionajstudies189.Only activeChemical Divisiownorkerswere includedinthisstudy. Workers not includedmay have had a differenrtesponse p2ftem thanthose who were included.Ifcontinuedemployment depended on response toexposure and the exposurewas associatedwiththe ondpointof interest,hen selectionbiasmay have-occurred.A fincrinogfthe presentstudywas thatPFOA was associated withdecreasedfreetestosteronaend increasedestradiolI.fworkerswho had highsusceptibiltfotythe effectof PFOA changed jobs,than the ovemffslopeof thedose responsecurvecouldbe underestimated.--Converselyi,fworkers with lowtestostemneassociatedwithPFOA changed jobs lessoften,thenthe overall dose responsecurvemay be overestimated.Migrationout ofthe highexposure jobsisunlikeltyobe the resultofsubclinicaclhanges inhormone level& AD currentChemical Divisionemployees who worked inhigh exposure jobsover the lastfiveyearswere includedinthe sample. Many workers who had been amp'loyedinthe highexposure jobs,butwho changed jobs were includedas participantTsh.e vastmajorityofworkers who had fsignificaenxtposure over the previousfiveyearswould be includedinthestudysample as the lum-overmte in Chemolfteemployees was low (threepercentper year)and thestudyincludedall currentemployees -witahppropriatjeob historiesS.electionbias isnot a likely explanatiofnorthe findingsinthisstudy. -210 j;1-7 InformgtioPni;;q No worker was unexposed. The lowestpotentialexposure group had significantly elevatedlevelsoftotalserum fluorine.Inview of this.the observed effects may representan underestimate ofthe trueeffect. Totalserum fluo@newas used as a surrogatevariableforPFOA exposure. The use oftotalserum fluorinehas been validatedinpast biologicamlonitoringinthe ChemolftePlantand otherplantsusing PFOA 8. Directmeasuremerft of PFOA usinggas chromatographictechniques have been highlycorrelatedwithtotal serum fluorinienChemolfte workers. Approximately 90% oftotalserum fluorine inChemolitsworkers was reportedto be inthe form ofPFOA 8.12. The validitoyf usingthissurrogatemeasure was not directlayssessed inthe currentstudy due tocost.Small amounts of PFCs otherthan PFOA may have been present in serum. The half-lioffePFC compounds isdirectlryelatedto molecularweight. Compounds withsixor lesscarbon backbones are likeltyo be rapidlyexcreted by exhalation190.Shortchain PFCs are unlikeltyo contributeappredably to total serum fluorine.Longer chain PFC, such as perfluorodee2noicacid (PFDA), are notproduced atChemolite. The hightoxicitoyfPFDA excludes itfrom commercic-Japplication$s3.77.79.101.102. Longer chain PFC are unlikeltyo be a significacnotmponent of totalserum fluorine.Other organicfluorinceontaining compounds existinbiologicaslystems and the environment. However, the small amounts absorbed from the envimnmertt intheform ofdrugs or plantproducts are rapidlymetabolized and excreted 191.Inorganicfluorinewas not a large constituenotfthe totalfluorinelevels.Serum ionicfluorinelevelsinthe 1-5 ppm range are associatedwith death inunintentionaolccupationalexposures 192. Totalserum fluoriniesa good surrogatemeasure forPFOA inthiscohorl The coefficienotfvariatiofnortotalserum fluohnewas 66%. The repeatabilitoyf theassay was betterattotalserum fluorinelevelsabove fiveppm. At the low end ofthe spectrum (< 1 ppm), where the assay islimitedby sensitivittyh,e totaj senim fluo@nevalues may overestimatethe truevalue. These measurement errorsare likeltyolead to an underestimate ofthe effectof PFOA on the physiologicencjpoints. 211 Commercial PFOA isa complex mixtureof isomers and relaetd compounds 39. h isclearthatsiru=rallyretatedcompounds, such as valproicacid,exhibit toxicitfyorcertainisomericforms,but not others143-123.Itiswidelyrecognized thatdifferendtrug enantomers have differenptharmacokineticand pharmacodynamic properties191.Thus, diflerenitsomers of PFOA may have differenttoxicitiesI.fone isomer ofPFOA isassociatedwithtoxicityt,hen the use oftotalserum fluorineortotalPFOA levelscould have produced an underestimateofthe truestrengthof association.However. inanimal studiesusing straighcthainPFOA, thespectrum oftoxicitesissimilarto those observed in siudiesusingmixed isomerof PFOA 37.M. a8.194. Furtherresearch isneeded to clarittyhe roleofPFOA isomers. The toxicokineticosfPFOA inhumans are differenftrom those observed in rodents.Extrapol2tintghe tissuedistributiofn PFOA from animals to humans may not be valid.No data existon the relationshibpetween serum and tissue PFOA distributiornbody burden inhumans. The use ofserum levelsto extrapolatteothe concentrationsatthe siteof PFOA actionmay have been inappropriateO.btainingpharmacokineticdata inhumans or appropriateanimal models isan importantarea forfutureresearch efforts. The temporalvariabiliotfyphysiologicparameters isrecognized. The uhraridan. circacriaann,d eircannualvariabiliotfythe stud ondpointswas notassessed directlyI.nstead,blood samples were drawn at the same trne ofday, on the same shiftforallparticipantsO.ne sample was drawn to estimatemean parametervalues. Considerablemeasurement errorisinherentinthisprocedure forhormones withshortpulsatilientervalssuch as LH, FSH, and testostemne. However, studieshave shown thatone sample isas good as threesamples in estimatingmean values 195.The use ofa singlesaimpleto estimatemean hormone levelpmduced random measurement errorand would be expected to attenuatet.he observed relationshipsM.ean serum values ofthe assayed hormones may not representthe biologicalliym;>ortariqtuantitiesatthe she of action. Validatinostudies of self-reported smoking status, using biochomiC@W ma*9r3 such as 'exhaled carbon monoxide, serum and urinethiocyanate,and serum -212 thiocyanateh,ave shown thatsmokers underreportjheisrmoking 196.Smoking is ae_socjatewdithchanges inphysiologicparameters such as hematologicaj=unts 197-199,cholesteml200.lipoprotein's-10210.2, and hepaticenzymes 1". The strengthand directionofthe associationbetween self-reponedsmoking inforrn2tiaond these parameters can be used toindirel-dayssess the validitoyf thesmoking information203. Inthisstudy,smoking statusand intensitwyas stronglyand significantly associatedwithfeukocytecount,band count,oosinaphilcount,plateletcount, and monocyle count. As expected,smoking intensitwyas negativelyassociated withbasophilcount. No participanrteporteddrinkingmore than 3 ounces of alcoholper day. This may reflectthe company's success indiscouragingheavy alcoholconsumption in employees,a reportingbias,or the factthatheavy d6nker-smay not be able to continueemployment due tothe demands ofChemical Divisionjobs. Alcohol=nsumption isassociatedwithchanges inphysiologicparameters such as hepaticenzymes 204, erythrocytemean corpuscularvolume, thgiyceddes,and highdensitylipoprotei1n44.As withsmoking, the strengthand directionofthe associatiobnetween self-reporteadlcoholconsumpton informationand these parameterscan be used to incrjrectalsysess the validitoyfthe alcohol informationI.ncreasedserum HDL isassociatedwithmoderate alcoholintake IA4.205.206.The expected relationshibpetween alcoholintakeand HOL was observedinthisstudyinindividualswithlow PFOA levels.As expected,there was a positiveassociationbetween alcoholintakeand t6glycefides.AJcohol has a directoxiceffectupon erythrocytesize,maturation,and number 207.M. The specificiotfyMCV is90% in identifyinaglcoholicsfrom socialdrinkerswitha positivperedictivvealueof 96% 209. Inthe present study alcoholconsumption of 1 to3 drinksper day was associated withan increasein MCV ofthe same order as reportedpreviously208 Alcoholinduces GGT. The senshivityof GGT in deleclingalcoholuse variesfrom 52% to 94%. GGT ishighlynon-specificfor alcoholconsumption or forhepaticabnormalities1". 210. Heavy dhnking oftwo tofivedrinksper day over one week or more are necessary to induce GGT 145208 GGT may be the onlycommonly assayed hepaticenzyme to increasewfth 213 heavydrinkingA. significanptositivaessociatiobnetween GGT and 3elf-reported alcoholuse was observed.The presence ofthese P-ssociatioinnsdicatesthatthe misciassificatoofnalcoholuse was unlikeltyoproduce a biaslargeenough to explaintheobserved associations. Alcoholuse was weakly associatedwithhepatictransaminases.SGOT and SGPT are lesssensitiviendicatorosfalcoholuse than GGT. Inalcoholinduced livedriseass,SGOT may t>eslightleylevatedand SGPT littclheanged. SGPT has be'en shown todecrease insome cases ofalcoholinduced liverdisease211. Consideringthe relationshikpnown to existbetween alcoholuse, SGOT and SGPT, littallecoholassociatedchange intransaminaseswould be expected. The observed weak associationisnot an unexpected findingand therefore probablydoes notreflecmtisclassificatioofnalcoholuse. Nonrespondentstothe alcoholftemwere criefrfentthan respondents.They were treatedas a separategroup inthe analysissincethe differencceould notbe explainedby me,@-suredcovariates.Althoughthepower ofthe study is diminishedby treatinaglcoholinformatioans a nominalcategoricavlariablet,he potentiaflorbiaswas reduced. 5 1.7-3Confoundina Bias Informatioonn thedurationofemployment inexposed jobswas notcollected. Plantrecordsdidnotcontainsufficienitnformatiotno reconstructexposures more thanfiveyearsinthe pasl The durationofexposure may be an important determinanotfPFOA effecl Durationofemployment may be relatedtoPFOA levelsincePFOA has the potentiaflorbioaccumulation.The durationof exposuremay have been a confounder forpeptidehormonal ondpairrtisnthis study.Inmdents,steroidhormonal and hepaticenzyme offe= of PFOA expasureoccuraftertwo weeks ofexposure,whereas peptidehormonal effects may re-quirelongerexposures 19. Uydig celltumors may requirea considerable lengthof exposure or latencyto develop 122. There are many compounds incomplex androgen-estmgon system. Th's'prosent studymeasured onlya few ofthem. Other biologicalilmyportantstemid 214 horrnonesindude cortisola,ndrostenedione,dihydroepiandrosilenedionseuffaie (OHEAS). estrone,esthol.esirogeniccaiechols,and dihydrotestosteron(eDHT). A totalestrogenindexor estrogentotestosteronsratio(EfT)may be more importantthan assays of individuaclompounds 212. Sex hormone binding globulin(SHBG). a major determinantoftheestrogen to testostemne bajance at the tissue level 213, was notassayed. More researchisneeded to cJaHfythe potentialroleof these hormones as confounders of the observed associations. The relationshifporbound testosteronsmay have been confounded by steroid hormone binding globulin(SHBG). Sex hormone binding globulinisan important determinant of test=erons and estradiollevelsin differenttissues as well as theirmetabolism 213. Plasma SHGB levelsare posftivelyassociated with estrogens and negatively associated wfth androgens 214. Thyroid hormone levels affectSHBG 215. The ratiosof estradiolto testostemne and testastemne to DHT may be regulated by SHBG levels213. The association between PFOA and bound testosteronemay have been, inpart,relatedto estradioland thyroid hormone changes in SHBG levels. Adult ratsdo not express SHBG 216. The declineintotaltestostemne observed in ratsisnot the resuftof changes in the amount or binding charactedste of SHBG. The observed depression of free testosteronein men isanalogous to changes intotaltestostemne in rats and is pmbably not significantlryelatedto changes inSHGB binding. Major stresses,such as surgicalprocedures, have been shown to markedly affecthormones in men 217. h isunlikelythat major physical stresses were associated with PFOA. Therefore, siresswas not a significantconfounder in the present study. Shiftworkhas been shown to affecta varietyof physiologic endpoints including biochemical parameters, hematologic incrices,and hormones 21". Study participantsrotatedweekly through three shifts.Allsamples were collected on the day shiftat leastthree days post shiftchange. Given the rotatingshiftsand standard day shiftsampling, itis unlikelythat shiftwork and PFOA were associated. Shiftwork did not appear to be a significantconfounder of the estimated dose response relationships. 215 SeveraJdietaryfactorsare determinantsofthe ondpoints considered inthis study.The effectsofdietaryfatand cholesterooln serum lipoproteinasnd lipicis iswidelyappreciated200. Dietaryc;Woriest,at,and carbohydrates affectstemid hormones 211.220. Dietcan alsoaffectthe metabolism ofsteroidhormones 221=2 Since itisunlikeltyhatdietisassociatedwith PFOA, itispmbably not a con*foundingcovadale inthisstudy. Physicalac@vityaffectsmany physiologicparameters includinghormones ZM, enzymes 224, lipopmteins 2m, and hematologic indices. For hormones, only maximal exercise produced an effecl No effectwas noted forsubmaximal physicalactivity.R isunlikelythat many participaritesngaged In maximal physicalactivity.Therefore, inthisgroup, itisunlikelythat physical activityis a determinant of the hormonal ondpoints under study. Physica) activitymay effect HOL levels,but itis unlikelythat physical activitywas associated wfth PFOA. Therefore, physical activitwyas unlikelyto be a si;nificantconfounder in this study. Medication usage and diseases such as diabetes mellitus are important determinants for some of the physiologic parameters measured in this study 123. 195.Questionnaire hems concoming mecricaflonuse and medicat historywere incomplete and were not validated. PFOA exposure has not been associated withany medicaj concritions8.Ifthe use of medication or the diagnosis of a medical conditionthataffectsone of the physiologic endpoints is asscdaled with PFOA exposure, then confounding may occur. However, no such relationships have been described. Inflammatory processes, which are major determinants of WBC, were not assessed inthisstudy. There is no evidence that inflammato ry processes are relatedto totalserum fluorineor serum PFOA. Therefore, these determinants of le.ukocytocount are unlikelyto confound the estimated relationships. 5-1-7.4Annlvtic Model Specification Sia-q The analyticmultivahate approach used inthisstudy assumed that a linear model with additiveeffectswas an adequate model with which to summarize the --216 data.A normal errorterm was used. Similarmodels of physiologicvariables have been extensivelyused inthepast and theirassumptions tested225. The model form was partialldyefineda pdoa based on a biologicahlypothesis.The choice of a finalmodel was based on biologicalknowledge plus best predictive power. The va@able transformations used were not based on a specific biologicalmechanism, but instead reflecthe basic form of dose response relationshipsobserved in nature. 2 1990 Chernolite MortalityStud 5 2 1 Intmduction This was a retrospectivecohort study of mortalityin workers employed in a PFOA productionplantforgreaterthan sixmonths duringthe periodfrom January 1. 1947 to December 31, 1989. Completeness of the cohort was assessed from independent sources. Demographic and work historydata were collectefdrom plantrecords and verifiefdrom independent sources where possible.Cohort members were notindividual=lnytacted foradditional informatioonn confoundingvariablessuch as smoking. Vitalstatuswas confirmed for 100% of the cohorl Cause of death was obtained from death certificatefsor99.6% of deaths and other sources for 0.4% of deaths. Cause of death was coded by ICD-8 catego6es by a nosologisl Reliabilitoyf death certificatceoding was assessed by random resubmission of.death certificatefsor recoding. The concordance was 100% forthree digftICD-8 codes. 5.2.2 ParticinantChammLaristics The 749 women were observed for 19,309 person-yeam, had a mean age at fiw employment of 27 years and mean follow-upof 26 yeam. The number of expected events given the age and size of the cohort was small. The study had limitedpower to detect moderate increases in cause-specific monaftiy. The 2788 men were observed forover 70,000 person years. The mean age at firsetmployment was 27 yeam,the mean length of follow-up was 25 years and a the mean age of death was 56 years. Non-CD men were older on average than --217 CO men and had more person-yearsinthe olderage groups where mortalitywas thehighest.IniemaJcomparisons were confounded by age as wellas othertime correlatedfactorssuch as lengthoffollow-up. F5-2.3MortalityResutts infemales,6.7% were deceased compared to 12.5% inthe males. Given that themean age alfirsetmployment and mean lengthof follow-upwas similarfor majes and females,thisreflecttshe expected survivaladvantage ofwomen. For both males and females the propordon ofdeaths was smallerinthe CO cohort. Employment intheChemical Divisiondidnot produce a largesurvival disadvantage. The allcauses, allcancer,and allcardiovascularmortalityamong women was lessthanexpected inthe overallcohorl The SMRs were remarkablystable when stratifioend tenyear exposure groups,and ten.fifteena,nd twenty year latencyperiods.The allcauses SMR was .75 inthe totalcohort .75 inthose employed forat leastten years or forthose employed longerthan tenyears,and 75 inallthreelatencyperiods.Cardiovasculardiseases and cancer mortality -followead similarpattern. inmales,the allcauses, cardiovasculardiseases,allgastrointestinaaln.d all respiratordyiseases SMRs were significantlleyssthan one. The allcauses SMR was .77using Minnesota mortalitryatesand .73using nationalmtes. The low SMRs are most likelay resultofthe healthyworker effect(HWE). As expected, the cancer SMR islessaffectedby the HWE. The allcauses SMRs were .75for allthreelatencygroups. Latency did nothave a strong relationshiwpiththe HWE. The allcauses SMR was .80inthe greaterthan fiveyear employment durationgroup and .68inthe greaterthan 20 year employment group. The low allcauses SMR inthegreaterthan 20 year durationgmup suggests thatworking for20 or more years was associatedwithcontinued selectionbased on good health.The allciausesSMR decreased withdurationofemployment inone meta-analysisof retrospectivceohortstudies226 but increased inthe metaanajysisby Fox and Collier227. 218 The SRRs forWicauses,allcancera,nd allcardiovasculdairseasesforlessthat ton years employment to more than ten years employment were not significantly differenftrom one. Because the rateswere based on small numbers of events,the95% Cl were wide. Due tothe smallnumber ofevents inthe females.SRRs were not calculatedT.he SRRs are similatrothe SMRs forthe lessthan ton year employment and greaterthan ton year employment groups. The SRRs forCO versusnon-CD male workers forallcauses, allcancer,and all cardiovasculardiseaseswere not significanatnd were similartothe SMRS. Working inthe CD didnotsubstantiallaylterthe ratesof death. The small number ofevents observed forrarecauses-ofdeath or specificcauses of death make itunlikeltyhatmoderate increasesinratescould be detectedinthiscohort forthe follow-upperiodthrough 1989. More follow-uptme wfllbe needed to allowsufficienptower to detectmoderate increasesinratesforspecificauses of death. The resultsfrom the adjustedRRMH contrastingthe mortalityratesforall causes, allcancer,and allcardiovasculardiseases between CD and non-CD ma)e workers were similatrothose forthe SRRs and SMRS. None of RRMH pointestimateswere statisticadlilfyferenftrom one. The contrastof rates between lessthan ten years of employmerrt and greaterthan ten years of employment presenteda differenpticture.Allcause RRMH were significardly elevatedinthe oldesttwo age groups, whilethe RRMH forcardiovascular diseaseswas significantellyevatedinthe 30 to lessthan 40 year age group.The allcancer RFIMH displayeda trendtoward a statisticalsliygnificanetlevationin theoldesttwo groups. The RRMH were notadjustedforyear offirst employment They may have been substantiallcyonfounded by changes of exposure over time sinceyear offirsetmployment. As seen inseveralPH regressionmodels, year offirsetmployment was significantlayssociatedwiththe mortatiftyA.fterage and lengthoffollow-upc,alendar time isthe strongesttime factorassociatedwith mortalit1y89.Hence, itislikeltyhatthe elevatedRRs for composite categoriesofcause ofdeath inthe oldestgroups were a resultof uncontrollecdonfoundingby calendarperiod.Given the small number ofevents instratai,twas notfeasibleto furtherstratiftyhe data on year offirtt employment. 2i-9 916.- inthePH regressiaonnalysipsr,ostactaencermortaliwtays-poshivealnyd significartrtelyatedtotimeinthe Chemical Division.Ten yearsof employment in theCD was associatedwitha 3 foldincreaseinprostatecancermortality compared tomen neveremployed inthe CO. Thistrendwas evidentinthe SMR anajysissiratifibeydCD and non-CO employment. This associationwas independentofdurationofemployment and year offirseimployment. As expected,age atfirsetmployment was positivelryelatedto pmstate cancer mortalitryate.The interpretatioofnthisestimatedrelativreateistempered by a number offactors.The estmates were based on sixpmstate cancer deaths,four intheCD cohortand two inthe non-CD cohorl A change ofone case could significanatltyertheestimates.Ascertainmentofallprostalecancer deaths may have been incomplete.Diagnosismay have been more complete inthe CD cohort.Given thatdeathcertificactaeuse ofdeath informatioinsknown tobe imperfectm,isciassificatoifoonne ormore deathscouldoccur.The use of mortalitays the eventofinteresftoretiologiscludiesofprostatecancer isnot the beststudyendpointbecause ofthe longnaturalhistoryand low mortalitoyf prostatecancer.The majorftyofincidentprostatecancers do not progressand cause death 221. m. For localizeddisease,an 80% ten year survivalinuntre2ted patentshave been reported2". Studiesofpmstatecancerincidenceinthis workforceare needed toclariftyhesuggested increaseinprostatecancer risk. The findingosfhormonalalterationisnPFOA exposed men suggestsa possible biologimcechanism forthe increaseinprostatecancer morwfty 231.Incidence studiesofotherdiseasesthatare hormonallymediated may be indicatedH the PFOA associatedhormonal changes are confirmed. 5-2.A MethoclolocicaClonsiderations 5.2.4.1InformationSias The use ofdeath certficatetso =egorize cause of death imperfect232*=5. The sizeofthepotentiabliasdepends an the cause ofdeath. Inone study cancer as a cause of deathwas under-reportedby 13% 232. Leukemias and lymphomas were underreportedin19% ofautopsy confirmedcases. Colorectalcancers were underreportedin12% ofcases. Therefore,itappears thatcancerdealhs were notseverelymisclassifiedA.l.lcardiovasculadriseasesas a gmup may --220 havebeeninaccuratIen.dividduiasleasweiththewholemay be severely rnisclassificaatneddmay produce largebiases.For example, specificauses of deathinthecardiovasculagrroup,such as cerebrovasculardisease,are inaccurateldyesignatedon death certificates. Three measures of PFOA exposure based on job historywere used inthisstudy. Firstt,hecohortwas dichotomizedintothosewho ever worked inthe CO and thosewho neverworked inthe CO. Second,the number ofmonths worked inthe CD until1985 was used as a continuousparameterforPFOA exposure. Third, thetotadluraton ofChemoliteemployment was used as a continuousparameter fortheeffectofwork ina plantproducingPFOA among a largenumber of products.Each ofthesesurrogatevariablemsay produce a differerstptectrum of misciassificatoCna.tegorizatioonfworkersintoeverversus never employed in theCD may notreflectthe biologicae)ffectivdeose ofPFOA- Many CD jobs do notentaiPlFOA exposure.A number ofworkerswere employed inthe CD for shortperiodsinthe distanptasl, Theirexposure may not have been significant. Thiscatego@zationmay misclassifuynexposed workers as exposed. Conversely,PFOA exposurewas widespreadamong Chemical Division(CO) employeesworkinginjobswithno exposure toPFOA. No exposure measurements have been done innon-CD employees. Itispossiblethatnon-CD employeeshad significabnotdy burdens ofPFOA. Ifthiswas thecase, exposed workers-muld..havebeen classifieads unexposed. Such misclassificatiwoonuld be expectedtobiasthe effectestimatestowardthe null.The months of employment inthe Co was the best availablestimateofPFOA dose. Not allCO jobshave PFOA exposure.The misclassificatipornoduced by classifying unexposed workersas exposed couldhave biasedthe estimatetoward the null. The use ofdurationofemployment atChemoliteas a continuousexposure parameterislessspecififcorPFOA thantime inthe CO. Ifanotherxenobictic exposureinthe planthas modulated diseaseoccurrencerates,the use of durationmay produce lessmisclassificatitohnan use ofdurationinthe CD. 5.2.4.2Confouncrtmaan@ SelectignBias The healthyworkereffectstronglyatlectsthe validitoyfmany occupational studies182-2m. h i'as complex biasthatresultsi,npart,from the selectionof 221 individualfsoremployment who are healthietrhan those inthe comparison population.The HWE isusuallystrongerforcardiovasculardiseases and respiratordyiseases. Because cardiovasculardiseases mortality unts fora significapnotrtionof allcauses mortalityt,he HWE usuallyreduces the allcauses SMR. The age atfirseimployment. age atrisk.lengthof follow-up,and duration ofemployment arefourtime factorsthatare associatedwithchanges inthe HWE 169.Generally,the HWE diminisheswithage and time. Collectioonfconfounder informationforindividualissdifficuifntretrospective cohortmoralitystudies.The presentstudy includedworkers followedformore than40 years. h was notfeasibleto collectindividuailnformationon such covahatesas smoking, healthstatus,medical historyo,r dietaryhabits.The proportionotworkers at Chemolite who smoke has been lower than inother facilitioewsned by thesame corporation.Inrecenthealthmaintenance studies, theself-reportesdmoking prevalence(25%) islower than the s=evade smoking prevalence.The observalionthatallrespiratordyiseases and lung cancer rates arelowerthan expected may be the resultofhistoricalloyw smoking prevalence. The low smoking prevalence may depress theallcauses SMR, allcancer SMR, and allrespiratordyisease SMR. The use ofinternalcomparison groups may reduce thissmoking relatedbias237 Time factorsuch as age at 6sk,age at firsetmployment, year of firm employment, and duration of employment are associated with the occurrence of many diseases 189. The use of an internalcomparison group may reduca certain selectioneffects,but may not controlconfounding ifthe exposure defined internal comparison groups have differentdistributionsof these time factom. Although the mean age at fimtemployment and mean year of firstemployment are similar inthe CO and non-CO cohorts of men and women, the comparisons of the rates ofcftseaseare confounded by differences in the distribldionof age at 6sk. These time factorsare stmngly correlated, with some being exact linearcombinations of others. The relationshipbetween measures of exposure and disease occurrence may be complex functions of these inter-relatetdime factors. Adjustment fortime factorsmay reduce the effectsof confounding, but may not controlconfounding 2-u Ifthe disease occurrence relationshipisdefined in terms of cumulabve 22-2 exposure,the trueeffectofexposure may be bi&ad toward the nullby uncontrolledconfoundingdue tothe complex time factors189. Some workers were exposed tomany otherpotentialldyisease causing xenobioticss,uch as benzene and asbestos,duringtheiremployment at Chemolits. Adjustment fortheireffectswas notpossibleinthisstudy. Even if informationwas availablee,xposures areoftenhighlycorrelatedmaking the separationof individuaelffectsimpossible. 5-2.4.4AnalZic Model SpecificatioBnias Comparison of SMRs and RRMH between exposure groups May notbe strictly valid.However, ifthe distributiofnthe person time inthe comparison groups is notstronglydiscordant,then such a comparison may be useful.Inthe current study,the person-timedistributioanrse differenitnthe exposed groups.However, thedifferencesappear to be ofa magnitude thatmakes usefulcomparisons of SMFis possible. Althoughthe proportionahlazard (PH) model has been used frequentlyforcohort studiesand clinicatlrialsi,thas notbeen widelyused inoccupationalstudies.In thepast.ithas been suggested thatPoisson regressionwas the analyticstrategy ofchoicebecause computationalcostswere lessand the conceptualizatioonf the model straigtftorward 129.However, PH models are now easilyrun with standardcomputer packages 109.Theirwide applicationinclinicatlrialasnd cohortstudieshas fosteredthe understandingofthe PH models. Poisson models appear lessfrequentlyinthe literaturaend may not be as well understood. Poison regressionand PH models have theoreticalinksand have been shown to givesimilarresultswhen used toanalyze the same data set Cox PH regressionwas chosen as the muftiva@atemodel to employ inthisstudy. The validitoyfthe proportionalhazards assumptions was examined using the two standardtechniques.The assumptions didnot appear to be grosslyviolatecl However, inanalyses involvinga smallnumber of events,the assessment ofthe validitoyf assumptions may be limitedT.he use ofthe factorsas continuous variableswas based on lack ofstatisticeavlidence fora significanntonlinear 223 effe'ctA.lthoughthissirategyhas been widelyused forcontrolofconfounding, has notbeen extensivelvyalidaledInsimulationstucries. 224 6- SUMMARY -RONCLUSIONS AND PFCOMMEN.D ATIONS I Crnsc.SectionpjStt_,doyfthg physiolonicEffe&s of PEOA This was a cross-sectionasltudy of selectedphysiologiceffeclsof PFOA, as quantifiebdy totalserum fluorine.Participantwsere recruitedfrom workers employed duringNovember 1990 inthe Chemical Divisionofthe 3M Chemolite PlantinCottage Gmve, Minnesota. AJIcurrentworkers who were employed in high exposure jobsat any time duringthe previousfiveyears and an age maiched sample ofworkers employed inlow exposure jobs were invitedto participate. Participantcsompleted a corporatemedical historyquestionnaireand had vital parameters measured by an occupationalhealthnurse. Blood was drawn for assays oftotalserum fluorines,even hormones irtvotveidnthe hypothalamicpituftary-gonadajxis,serum lipidsl,ipoproteinsh,epaticfunctionparameters, and hematology indices.Blood was drawn inthe moming afterworkers were assigned tothe-dayshiftforat leastthree days. inpast studies,the majodty oftotalserum fluorinefound inChemolhe workers was intheform ofPFOA. Thus, totalserum fluo6neisa validsurrogatemeasure ofPFOA inChemolhe employees. For 93% ofworkers.totalserum fluorine levelswere 20 timesgreaterthan community and corporatebackground levels. Findingsinthe currentstudyare consistentwith otherdata suggestingthatPFOA has a long biologicahlajf-liifneboth men and women. The long half-lioffePFOA may resultinsignificanbtioaccumulationfrom smallfrequentdoses or large, infrequentdoses. The hormonai findingsfrom thisstudy are consistentwiththe hypothesisthat PFOA depresses thehuman hypothalamic-pituitary-gonadaaxlis.The results show thatlow levelsofserum PFOA (20@LM)depressed freetestostemne and elevatedMradiot withlittloebserved change inLH levels.Inoldermen, free testosteronewas depressed below 9 ng/diatserum fluorinelevelsbelow one ppm (estmated PFOA levelsbelow 1 gM). --225 Mean pmlactinlevelswere positivelayssociatedwith PFOA inmoderate drinkers, bljinot inlighdtrinkers.Since the func@on ofprolactininmen isuncertain,the clinicaslignificancoef thisfindingisunclear. Mean thyroidslimulatinhgormone was positivelayssociatedwith PFOA. Since peripheratlhyroidhormone levelswere not assayed. itwas not possibleto assess whether the observed associationbetween PFOA and TSH was the result ofa directeffecton the hypothalamus,pituitartyh,yroidgland,or peripheral thyroidhormone metabolism. Cholesterolt,@glyceddes,and LOL were not significantlayssociatedwith PFOA. PFOA was negativelyassociatedwithHOL inmoderate drinkers. PFOA was notassociatedwhh the marked hepaticchanges inhumans that have been observed inmdents. PFOA appeared to alterthe hepaticresponse to endogenous factorsand xenobiotics. PFOA was significantalsysociatedwithhemoglobin levels,MCV, and MCH. The estimatedchanges inerythrocytesare notofclinicaslignificanceover the range ofobserved totalserum fluodne. The changes inloukocytocounts associatedwith PFOA exposure presented a complex picture.For example, the negativeassociationbetween PFOA and lymphocyteswas increasedby smoking more than 10 cigarettesper day and decreased by alcoholuse and adiposity.The magnhude ofthese associations are notclinicalsliygnificanftrom an infectioudsisease perspective.However, elevatedWBC has been associatedwithincreasedallcauses, cardiovascular diseases,and cancer mortalitays wellas increased incidenceof myocardial infarction. 6-2 FR@tmsnectiveCohort Mortaft Study Of The Chemolite Workforce- 1947im Thiswas a retrospectivceohortstudy of mortalityinworkers employed In a PFOA productionplant.Allcauses mortalitiyn both male and female Chemothe employees were significantlleyssthan expected based on comparisons to the 226 -A mortaliteyxpe6ence ofthe Minnesota and UnitedStates population.The SMRs forseveralothercauses ofdeath includingallrespiratordyiseases were lessthan expected. Since the healthyworker effectwas apparentlystronginthe Chemolitscohort,intemalcomparisons ofSMRs were made between Chemical Division(CD) and non-Chemical Division(non-CD) employees. These comparisons did notsuggest any significanetxcesses inmortalitiynCD or nonCD employees. Generally,thefindingsfrom thisstudy provideno evidence thatemployment at Chemoliteresultsinelevatedmortalitryatesfrom any cause. However, pr=ate cancer mortalitmyay be associatedwfth lengthofemployment inthe Chemical Division.Ten years ofemployment inthe CD was associatedwitha significant threefoidincreaseinprostatecancer mortalityT.here was no association between prostatecancer mortalityand employment (ever/never)inthe Chemical Division.Given the smallnumber ofdeaths from prostatecancer inthisstudy and the naturalhistoryofthe disease,the associationbetween employment in theCD and prostatecancer must be viewed as hypothesisgeneratingand should notbe over interpretedH.owever, the biologicapllausibiliftoyrany association between CD employmerttand prostatecancer isincreased by animal and human toxicologicadlata suggestingan associationbetween PFOA and steroidsex hormone changes. .3 Conclusion Perfluorooctanoiaccidwas associatedwith reproductivehormonal changes in exposed workers. The clinicaslignificancoefthese findingsareunknown. The associationsofPFCA wfth hormones, HDL. hematology parameters.prostate cancer mortalitiynmen indicatesthe need forfurtherresearch. 6.4Rp mm@ndations Research isneeded infiveareas. 227 1.An assessmert%ofthe hormonal effectofPFOA Inwomen isneeded. A crowsectionalsiudy should be conducted usingspecifictgsaysforPFOA and accountingfortemporal hormonal variations. 2.The clinicaslignificancoefthe associationosf PFOA with the physiologic parameters need clarificatioSni.nce morbidityfrom diseases such as prostate cancer isreflecteidnmortalitya,n update ofthe retrospectivmeoralitystudy is needed infiveyears. Morbiditystudiesshouldbe conducted of endpoints that may be produced by hormonal changes . Since exposed workers are relatively young and are limitedinnumber, thefeasibleendpointsfora shortterm morbidity studyare limited.Poolingof workers from a number ofplantscould increasethe number ofexposed workers and allowendpointswithlower incidenceto be studied.The morbiditystudyshould be a longterm which would allowthe study of endpointsthatoccur at higherfrequencyinolderage groups. In men. endpoinisshould includetheincidenceofbenign prostatichypertrophyand prostatecancer. The feasibiliotfyincjudinginflammatorybowel disease and colorectaclancer as endpointsshould also be evaluated. Inwomen. endpoints shouldincludethe age of menopause, the incidenceofosteopomsis and related fracturesu,terinefibmids,and cholelithiasiIsf.thereare a sufficienntumber of events,ondometdal cancer and inflammatorybowel disease should be evaluated.Ifthecmss-sectionalhormonal studyinwomen findsno association between PFOA and hormones, then the morbiditystudycan be limitedto men. 3.Stucrieosf reproductiveoutcomes inboth men and women are needed. Ubido, potency,and lertlitayre directlayssociatedwithsteroidhormones levels.The feasibiliotfya retrospectivsetudyof reproductiveondpoints or a pmspective studyoftime-to-pregnancyneeds to be explored. 4.The mechanisms ofactionof PFOA need tobe studiedconcurrentlywith morbidity.Mechanisticstudiesare needed todefinethe relevance of animal studiesforhumans and providea firmbiologicablasisforthe findingsof the mortalitym,orbicritya,nd reproductionstudies. Invitroand celllinestudiescould clarittyhe mechanisms of actionof PFOA on thepituitarsyecretionof LH, FSH, TSH, and prolactin.Pituicyteculturesmay be --228 AM& helpfuilnevaluatintghe directeffectofPFOA pituitarfyunction*The effectof PFC)A on otheraLnoc6ne or paracrinefactorsuch as TGF-A, TGF-3, FGF, and TNF couldalsobe evaluated.Human adiPtOcYtGculturescouldbe used tostudy theeffectofPFOA on aromalass activityA.dditionallsyt,udiesare needed to clariftyhe relationshbieptween PFCA and the temporal variabiliotfyrepmduc@ve hormones. S. Studiesare needed to betterdefinethe PFOA exposure profiloefallworkers employed at Chemolite,to ascertainthe source oftheirPFOA exposure and routeofcontinuedabsorptionand to ciarittyhetoxicokinebosand toxicodynamics ofPFOA inhumans. --.haw PFFFRENCES 1. Kirk-Othmer.EncyclopediaofChemicalTechnology.2 ed. Vol.9.Now York: intersdencePublishers1,966. 606-647. 2. Abe T, Nagass S.Electrochemicaflluorinati(oSnimons Pmcess) as a routeto ;>erfluorinatoerdganiccompounds ofindustriailnterestI.n:Bankes R, eds. Preparadon,Propetliesa,nd IndustriaAlpplicadonsof Organofluorine Compounds. Now York:John Wiley& Sons.1982. 3. Simons J, Bryce T. ElectrochemicaflluorinatioInn.:Simons J.eds. Fluorine Chemistry. New York:Academic Press.1954:340-377. 4. Simons J.FluorineChemistry.New York:Academic Press,1963. 5. Taves D. Evidence thattherearetwo forms offluoridienhuman serum. Nature1968;217:1050-51. 6. Taves D. Electmphoreticmobilitoyfserum fluorideN.ature 1968;220:582583. 7. Guy W, Taves 0, Brey W. OrganicfluorocompoundsInhuman plasma: prevalenceand charactehzation.In:FilleRr. eds. Biochemistryinvomng carbon-fluotinbeonds.ACS Symposium Sedes. New York:American Chemical Society.1976:117-134. 8. Ubel F.Sorenson S.Roach D. HealthStatusofPlantworkers exposed to fluomchemicals:a PreliminaryReporL Am IndHyg Assoc 1980*,41:584-589. 2. GritfdFh,Long J.Animal ToxicityStudieswithAmmonium perfluorooctanoate. Am IndHyg Assoc 1980;41:576-583. 10.ZobelL 3M CorporateMedicalDepartment experienceinmedicalscreening. 1990, -230 11. NicklesM. Presence du fluordans lasang. COmpt Rend 1856;43: 885. 12. VenkateswaMu P. Sodium biphenylmethod fordeterminationofcovalently bound fluorinienorganiccompounds and biologicamlaterials.Anal Chem 1982,54:1132-1137. 13. Yamamoto G, YoshitakeK, Sato T, et al.DistribLnioannd forms offluorinein whole bloodofhuman males.Anal Siochem 1989:182:371-376. 14. BelisleJ,Hagen D. Method forthe determinationofthe totalfluorinerontent of whole blood,serum/plasma, and otherbiologicaflluidsA.nal Siochem 1978;87: 545-555. 15. Chinba K. Tsunoda K, Haraguchi H, Fuwa K. Determinationof fluorinein urineand bloodserum by aluminum monofluoridemolecularabsorption spectrometryand witha fluoridelectrode.AnaJ Chem 1980;52:1582-1585. 16. Singer L,Ophaug R. Concentrationsofionic,total,and bound fluoridein plasma. ClinChem 1979:25:523-525. 17. Guy W. Fluorocompounds of Human Plasma: Analysis,Prevalence, purificatioann,d CharacterizationP.h.D. thesis.Rochester, NY:' Universityof Rochester. 1972. 18. Taves D, Guy W. Brey W. Organic Fluorocarbons inHuman Plasma: Prevalence and CharacterizationI.n:FilleRr. eds. BiochemistryInvolving Caton-Fluctine Bonds. Washington, DC: American Chemic.-JSociety. 1976:117-134. 19. Cook J, Murray S, Frame S. Hurtt M. Induction of Loydig celladenomas by ammonium perfluoroocianoateA:possibleendocrine relatedmechanism. Tox Appl Pharm 1991:113:209-213. 20. Bryce H. Industriaaln.d UtilitariAasnpects of Fluodne,Chemistry. In:Simons J,ods. FlucifneChemistry. Now York: Academic Press. 1964:297-492. 231 21. PetersP. Shorthouse M. Fluorocitratienplantsand food.Photochemistry 1972:11: 1337-1338. 22. Lovelace J,MilleGr, Welkis G. The accumulationof fluoroacetataend flucrocitraitneforagecrops collectednear a phosphate plant.Atmos Envimn 1968:2:187-189. 23. Cheng J,YU M. MillerG. Fluoro-organicacidsinsoybean leavesexposed to fluorideE.nv Sci Tech 1968;2:367. 24. Guy W. Ionicand organicfluorineinblood.In:Johansen E, Taves D, Olsen T, eds. ContinuingEvaluationoftho Use of Fluotide. Boulder.CO: Westview Press.1979. 25. Taves D. Comparison of"organic'fluorideinhuman and nonhuman serums. J Dent Res 1971;50:783. 26. Bankes R. Fluorocarbonsand theirDerivativesL.ondon: MacDonald Technical& Scientifi1c9,70. 27. Midg@ey T, Henne.A. organic fluo6des.asrefrigerantsI.nd Eng Chem 1930=: S42-646. 28. Moore C. Industryresponse tothe Moritrealprotocol.Ambio 1990;6-7:320323. 29. Rowiand R. Stratospherioczone depletionby chlorofluorcarbonsA.mbio 1990,*-7:281-292. 30. Olson C, Andersen M. The acutetoxicitoyf perfluoro=anoic and perfluorodecanoicacidsinmale ratsand theireffectson fattytissue.Taxicol Appl Pharmacol 1983;70:362-372. 232 31. Guenthner R, VietorM. Surface activematerialsfrom ;>erlluorocarboxylic and perfluorosulfonaiccids.I&EC Product Research and Development 1962;1:165-169. 32. Shindo K, Nomura T. Miscibiliotfyfluorocarbonand hydrocarbon surfactants inmicellesand liquidmixtures.Basicstudiesofoilrepellentand fire extinguishinaggerns.J Phys Chem 1980;84:365-369. 33. RiessJ, LeBlanc M. Preparationof pernuamchemical emulsions for biomedicaluse: phnciples,materialsand methods. In:Lowe I<,eds. Blood SuststFwtes. London: EllisHorwood. 1990. 34. YlinenM, kojo A, HanhijarviH, Peum P. Dispositionof perfluorooctanoic acid inthe rataftersingleand sut>chmnicadministrationB.ullEnviron Contam Toxicol1990;44: 46-53. 35. Clark L,BecattinFi, Kaplan S. Obrock U. Cohen 0. Becker C. Peffluomcarbonshaving a shortdwelltime inthe liverS.cience 1973;181: 680-682. 36. SargentJ,SefflR. PropertiesofperlluorinateldiquidsF.ed Pmc 1970;29: 169.9-170.3. 37. Kennedy G. Dermal toxicitoyf ammonium perfluomoctanoate.ToxicolAppi Pharma=l 1985*.Sl3:48-355. 38. Kennedy G, HallG, BdtteilJl.Chen H. lnh-"!atitoonxicitoyfammonium perfluorooctanoateF.d Chem Toxic 1986;24:1325-1329. 39. Ophaug R, SingerL Metabolichandlingof perfluomoctanoicacidin mts. Pmc Soc Exp SiolMed 1980;163:19-23. 40. BelisleJ.Organic fluoridien human serum: Naturalversus industrial sources.Science 1981;212: 1509-1510.. 233 41. MaraisJ.Monofluomaceticacid.thetoxicphncipleof*Glfblaar DlchapetaJum CYmosum. OndersiepoortJ Vet Sic Anim Ind 1944;20: 67-71. 42. PetersR. Lethalsynthesis.Proc R Soc London B 1951:139:143-170. 43. Fanshier D, Gottwald I-Kun E. Studieson specificenzyme inhibitorVsi. Characterizatioannd mechanism ofactinofthe enzyme-inhibitoryisomer of monofluorocftrateJ.SiolChem 1964:239:425-434. Tecle B, Casida J.Enzymatic defluodnationand metabolism of fluoroacetate, fluoroacetamidef,luomethanoland (-)-erythro-fluomcitriantreatsand mice examined by Fl 9 and Cl 3 NMR. Chem Res Toxicol1989;2: 429-435. 45. Taves 0. Dietaryintakeoffluoridaeshad (totaflluoridev). unashed (inorganifcluoridea)najysisof individuafloods.Br J Nutri1983;49:295-301. 46. Long D, HigginsC. Istherea time and placeforradiopaque fluorocarbons. In:Bankes R ads. Preparation,Properti6sa,nd IndustriaAlpplic2dons of OrganofluotineCompounds. Now York:John Wiley & Sons. 1982. 47. FaithfuNl. Potentialapplicationsof perfluarochemicalemulsion inmecricine and researrh.In:Lowe K, eds. Blood Substitutes. Chichester.Ellis Horwood. 1990:130-148. 48. Manning R, BrucknerJ, Mispagel M, Bow J.Metabolism and disposftioonf sulfluramida,unique popwuodnated insecticidei,nthe rat.Drug Metab Disposition1991;19:205-211. 49. McCormick W. Repead application28 day dertnalabsorptionstudy.1983, RikerLaboratories3,M: 50. HanhijarviH, Ophaug R, Singer L The sex-relateddifferencein perfluomoctanoateexcretioninthe rat.Proc Soc Exp BiolMed 1982;171: 50-55. 234 61. Vanden HouvalJ.KuslikiBs,Van Refelghem M. Peterson R. Tissue distribvtimoent,abolism,and e(iminatioonfperfluorooCf2noiacdd. j Siochem Taxicol1991;6:83-92. 62. HanhijarvHi, YfinenM. A proposed speciesdifefrenco inthe renalexcrabon ofperfluorcoctanoaidcd inthe beagledog and raL In:Saynen A. Solieveld H, ods. Now Developments inBiosci4nces:theirlmplicagansforLaboratory Animal Sciences. Dordrecht:MarUnus Nijhoff1.988:409-412- 63. Vanden HeuvelJ,Kusliki8s,PetersonR. Dispositioonfperfluarodgcanoic acidinmale and femalerats.ToxicolAppi Pharmaool lqgl;l07:450-459. 64. Vanden HeuvalJ,DavisJ,Sommers R, PetersonR. Ronal excretionof perfluorooctanoaiccidinmale mts:inhibitoreyffectoftestostemne.J Biochem Toxicol1992;7:31-6. 65. BookstaffR, Moore R, lngalG, PetersonR. Androgenic deficienciynma)s ratstreatde withperfluorodecanoiaccid.ToxicolappiToxicol1990;104:322333. 66. Van Rafelghem M, Matte D, BrunerR. Anderson M. Pathologicaalnd hepaticuftrastructuerfaflectsofa singledose ofperfluoro-n-decanoiaccidin therat,hamster,mouse and guineapig.Fundam Appi Toxicol1987,,95:22540. 67. The 3M CorporationICPD. ProductToxicitySummary Sheet 12251, FC-1 43. 1988,3m corporation: 68. O'DellW. SwercflofRf, Bain J,Wollensen F, Grover P. The effectofsexual maturationon lesticularresponse toLH stimulatioonftestosteronsecretion inthe intacrtatEndocrinof1974,,95:1380-1384. 69. GlassA, MellitRt,VigerskyR, SwerdloffR. Hypoandrogenism and abnormal regulatioonfgonadotropinsecretioninratsfeeda low proteindet. Endocrinol1979;104:438-442. - 236 70. Cicem T, SchainkerB, Meyer E- Endcgenous opiods participatienthe regulatioonfthe ofthe Hypothalamic-pituitary-iutenihzoirnmgone axis and tesicsterone'nsegativefeedback controlofILnenizinghormone. Endocrinology 1979;104:12B6-1291. 71. Moore R. JetcoateC, Peterson R. 2,3,7,8-ietrachlomdibenzo-p-dioxin inhibitsstemidogenesis inthe rattestisby inhibitintghe mobilizationof cholesteroltocytc>chromep450ssc. Toxicolappi PharTnacol 1991;log: 8_c, 97. 72. BookstaftP, Moore R, Petersen R. 2.3,7,8-tetrachlomcfibenzo-p-dioxin increasesthe potency ofandrogens and estmgens as feedback inhibitorosf luieinizinhgormone secretioninmale i-atsT.oxicolAppl Pharmacol 1990;104: 212-224. 73. Spink D, LincolnD. Dickerman H, Gienhy J.2,3,7,8-tetrachlorocribenzo-pdioxincauses an extensivealteratioon 17 beta-estradiomletabolism in MCF-7 breasttumor cells.Pmc NallAcad Sci 1990;87: 6917-21. 74. Egeland G. Serum Dioxin2.3,7,8.TCOD and totalserum test and gonadotmpins inoccupationallyexposed men. inScciefyforEpidemiologic Research,Annual meeting 1992-1992. hfinneapolisM.N: 75. Gutshall0. PilcherG, Langley A. Mechanism ofthe serum thyroideffectof perflucrodecanoiaccid(PFDA) inrats.J ToxicolEnvironHeafth 1989;28: 53-65. 76. GutshallD, PiltcherG. Langley A- Effectsofthymxin supplementationon the response to perfluoro-n-docanoiaccid (PFDA) in rats.J ToxioofEnvimn Health 1988;24:491-498. 77. KellingC. Van Rafeighem M, Menahan L.Petersen R. Effectsof perfluarodecanoicacidon hepaticindicesofthyroidstatusof rat.Biochem Pharmacol 1987*-,36:1337-13". --237 78. Van Rafeighem M. Inhom S, Peterson R. Effectsofperfluorodecanoicacid on thyroidstatusinrats.Toxicoland Appl Pharrn 1987:87: 430-439. 79. GutshailD. The effectsofperfluom-n-decanoicacid on the r2tpituharythyroidaxis.1965,Wright StateUniversity: 80. ThottarseryJ,Winberg L,Youssof J,Cunningham M, Badr M. Regulationof p-erfluorcoctanoaidcd-induced pemxisomal enzyme activitieasnd hepatocellU12grrowth by adrenal hormones. Hep2tology 1992:15: 316-322. 81. Just W, Gorgas K, HartlF, Heinemann P, Salazar M, Schimassak H. Biochemicaleffectsand zonal heterogeneityofpemxisome proliferation induced by perfluorocarboxylaiccidsinratliverH.epatology 1989;9:570-81. 82. Lazamw P, Shio H, Lemy-Houyet M. Specificitiynthe actionof hypolipidemicdrugs:increasesof peroxisomal B-axidationlargely dissociatefdrom hepatomegaiy and peroxisome pmliferatoninthe rat.i Upid Res 1982;23:317-326. 83. EliassenK Osmundson H. Factorswhich may be significanrtegarding regulatioonftheclofibmte-dependentinductionof hepaticpemxisomal Boxidationand hepatomegaly. Siochem Pharmacol 1984;3:i@Ib2@--1'031. 84. Best M. Duncan C. Hypolipidemiaand hepatomegaly from chlorphenoxyisobutyrat(eCPIB) inthe rat.J Lab ClinMed 1964-,64:634642. 85. Kawashima Y, Katch H, Kozuka H. Differentiaelffectsof alteredhormonal stateon theindudon ofacyl-CoA hydmiases and pemxisomal B-oxidation by colfibriaccid.Siochim Biophys Acia 1983;750: 365-372. 86. Reddy J,Azamoff D. HigniteC. Hypolipidemichepaticpemxisome proliferatofrosrm a novelclassofchemical cardnogens. Nature 1980;283: 397-398. -238 87. Osumi T, Hashimoto T. Enhancement offattyacyl-CoA oxidizingactivitiyn ratliverperoxisomes by Di(2-tnhyihexyplh)thalateJ. Biochem 1978;83: 1361-1365. 88. Vanden Heuvel J,NesbitD, StercheleP. InductionofAcyl-CoA binding proteinand fattyacidbindingproteininralhepatocytesby the pemxisome proliferatopresmuorodecanoic acidand ctofibratTeo.xicologist1992:12:S. a9. issemann 1,Green S. Activationofa member ofthe steroidhormone receptorsuperlamilyby peroxisome proliteratorNsa.ture 1990:347: 64_C, 649. 90. Tugwood J.Receptor mediated peroxisome proliferatoarction.Toxicologist 1992;12: 6. 91. AbdellatiAf, PreatV, Vamecq J,NilssonR, RoberfroidM. Peroxisome proliferatiaond modulation of ratlivercan@nogenesis by 2.3cfichlorophenoxyacetiac id,2,4.5trichlomrphenoxyaceticacid, perilucrooctanoiaccidand nafenopin.Careinogenesis 1990:11: 189!;-l902. 92. NilssonR, BeijaB, Preat V, E6xon K. Ramal C. On the mechanism ofthe hepatdcardnogenecityof pemxisome pmliferatom.Chem SiolInteract 1991;78:235-50. 93. Takagi A, Sai K Umemura T, Hasegawa R. Kumkawa Y. Short-term exposure to theperoxisome proliferatomp,onluomoctanoic acidand perfluorodecanoicacid,causes signifir-aintcreasesof 8hydroxydemguanosine inliverDNA of rats.Cancer Lett199 1;57:55-60. 94. Handler J,Seed C, BradfordB, Thurman R. lnducflonof peroxisomes by treatmentvathperfluorooctanoatedoes not increase ratesofH202 productioninintactliverT.oxicolLett1992;60:61-68. 239 Ism 95. LakeS,Gr2yT,SmithA-Hepatipcaroxisomperoliferaatnidoonxidative stress.BiOchOm Soc Trans 1990:18:94-97. 96. Levitt0, Uss A. Toxicityof perfluarinatefdattyacids forhuman and mu@he B celllines.Toxicaland Appi Pharmaeof 1986:86: 1-11. 97. LevittD, Uss A. Perfluorinatefd2ftyazids altermarocyanine 540 dye binding toplasma membranes. J Toxicoland Environ Health 1987;20: 303-316. 98. Inoue T, lwanaga T, Fukushima K, Shimozawa R. Effectsofsodium octanoateand sodium perfluorooctanoatoen the gel-to-liquid-crystalling phase transitioonfdipaJmhoyiphOSp2tidyicholinveesiclemembrane. Chem Physics Upids 1988;46:25-30. 99. Leikes P, Willer1.Porturbatons of membrane Locationand structuraslensitivitoyfMC540 membranes. J Membr Siol1980;52: 1-IS. structureby opticalprobes: 1. bound to phospholipid 100. NicolsonG. The CellSurface:Transmembrane regulationofreceptor dynamics. Prog lmmunol 1977;3:5-7. 101. PilcherG, GutshallD, Langley k The effectsofperflucro-n-decanoiaccid (PFDA) on ratheart8-receptorsa,denylatecydase and fattyacid composition.ToxicolAppl Pharmacol 1987,,90:198-205. 102. WigierP, Shah Y. Perfluorodecanoicacid inactivatioonfa channel for2aminopuhne inthe 15178 Y cellmembrane and recoveryofthe channel. TaxicolAppi Pharmacol 1986;85:456-463. 103. Olson C, George M, Andersen M. Effectofperfluorodecanoicacid on cell composftionand membranes. Toxicologist1983,3:99. 104. Berry G. The analysisof mortalityby the subject-yearsmethod. Biometdcs 1983;39:173-184. 240 105. Monson R.Analysiosfrelatisvuervivajnd proportionmaolrtalityC.ompLn Biomed Res 1974;7: 32-r--332. 106. Miettinen0. Standardizationofriskratos.1972;96:383-388. 107. Cox D. Regressionmodels and lif=eIes. J R StatSoc (B)1972;34: 187220. 108. Mantel N, Haenszel W. StatisticaJspects ofthe analysisofdata from retrospectivsetudiesof disease.JNCI 195922:718-748. 109. SAS. SAS Users Guide: Statsticse.d. InstitutSe. Cary, NC: SAS Insiftuie Inc.,1990. 110. KalbfleiscJh.PrenticeR. The StatisticAanlalysisofFailureTime Data. New York:John Wiley & Sons. 1980. ill.Bammann 8, Coulam C, Jiang N. Totaland freeteststeroneduring pregnancy.Am J Ob Gyn 1980:137: 293-298. 112. Williams-Ashman H. Perspectivesinthe male sexual physiologyof eutherianmammals. In:KnobilE, NeillJ.eds. The Physiologyof Rep,roduction. New York: Raven Press. 1988:727-752- 113. JuniewiczP. OesieriingJ,Walters J. Aromatass inhibitioinnthe dog. 1. Effectson serum LH, serum testostemne concentrationst,estostemne secretiona,nd spermatogenesis. J Uml Igas:l39: 827-831. 114. NishiharaM, Winters C, Buzko E, Waterman M, Dafau M. Hormonal regulatioonf ratLoydigsellcytochrome P450-17a MRNA levels.Biochom Biophys Res Commun 1988.*154:151-158. 115. Tsai-MorrisC. Knox G, Luna S, Dufau M. Acquisitan ofestradiol-mecriated regulatorymechanisms of sisroidogenesisinculturedfetalmt Loycrigcolls. J SiolChem 1986;261:3471-3474. - 241 116. Coffey0.Androgenactioannd thesexaccessarytissueI.n:KnobilE,Neill L.ads.7hePhysioloogfyReproductionN.ow York:RavenPress. 1988:1081-1120. 117. Longsoope C. Kato T, HortonR. Conversionof blood androgens to estrogensinnormal adultman and women. J ClinInvest1969;48:2191. 2-201. 118. Michnovi= J,Fishman J.Increaseoxidatysmetabolism ofcestrogensin male and femalesmokers.In:Waid N, Baron J,ads. Smoking and Hormone-RalatedDisoiders. Oxford:OxfordUniversitPyress.1990:197207. 119. lsserman 1,Green S. Activatioonf a member of the steroidhormone superfamilbyy peroxisome proliferatoNrast.ure 1990;347:645-650. 120. Tugwood J, Isserman 1.Anderson R, BundellK. McPheat W, Green S. The mouse peroxisomeproliferataocrtivatinrgeceptorrecognizesa response elementintheS'flankinsgequence ofthe mt acylCoA oxidassgone. EMSO 1992;11:433-439. 121. Dfe-yer-C,Kroy G, KiifFe Oivif-F-,He-ileFEiF@-G, Wahli W. Controlof pemxisomal B-oxidatiopnathway by a novelfamilyofnuclearhormone receptorC.oll1992,,688:79-ea7. 122. RobertsS,NettT,Hartman H, Adarns T,Sto!!R. SDZ 200-110 induces Leydigcelltumors by increasinggonadotropinsinrats.J Am CollToxicol 1990,,a4:87-505. 123. SchukitM, Gold E. Risch C. Serum pmlactinlevelsinsons ofalcoholics and controlsubjectsA.m J Psychiatry1987:144:854-859. 124. Wald N, Baron J.Smoking and Hormone Relateddisorders.Oxford:Oxford UniversitPyress,1990. 279. 242 125. FiikerLaboratodes.Riker/3M EXP. 0281CROO12 Two year oral(Diet) toxicity/carcinogenisctiutdyy ofFC-143 inrats.1983. The 3M Company: 126. Andervont H, Shimkin M, Canter H. The growth of estrogen-induced interstitciealllesticulatrumors inBALB/c mice. J NallCancer Inst196024: 1219-1237. 127. Due W, Dieckmann K, Loy V, SteinH. lmmu nohistologicadleterminationof oesirogen receptor,progesteronsreceptor,and intermediatefilamentsin Leydi;celltumors,Leydi; cellhyperplasiaa,nd normal Leydig cellsofthe human testisJ. Pathol 1989:157: 225-234. 125. CastleW, Richardson J.Leydig celltumors and metachronous Leycrigcell hyperplasiaa: case associatedwithgynecomastia and elevatedurinary estrogens.J Uml 1986:136:1307-1308. 129. Teerds K,,R,ommerts F. Donington J. lmmunohistochemical detectionof transforminggrowth factoralpha inLeydig cellsduringdevelopment of the rattestisM.ol CellEndocrinol1990,69: Rl -R6. 130. H2ndelsman D, SwerdloffR. Male gonadal dysfunction.ClinicsEndocrinol Metab 1985;14:89-124. 131. StaplesR, Burgess B, Kerns W. The embryo-fetaltoxicitaynd toratogenic potentiaolfammonium pernuorooctanoate(APFO) inthe rat.Fundam Appl Toxicol1984,4:429-440. 132- StaplesR. Improperinterpretationfdata concoming toratogenicitay:case report.Prog ClinSiolRes 1985;163C: 161 -163. 133. Dixon R. Toxic response ofthe reproductivesystem. In:Klassen C, Amdur M, DoullJ,ods. Casareffand Doull'sToxicclogy. New York: Macmillan PublishingCo. 1986:432-477. 243 134. KoopD.Oxidatiavnedreductimvsetabolisbmy.cytochromPe4502EI. FASEB 1992:67:24-730. 135. Gordon D. HDL and coronary heartdisease.In:MillerN, ods. High Density lipoprotains and Atherosclerosis. Publishers B.V.,.1989:3-10. Amsterdam: Elsevier Science 136. StolzA, Kaplowitz N. Biochemical Tests forUver Disease. In:Zakim D, Boyer T, eds. Hepatology.-A 7extbook of LiverDisease. Philadelphia.PA: W.B. Saunders. 1990:637-667. 137. GitlinN. ClinicalAspects of Uver Diseases Caused by Industriaalnd Enviromental Toxins.In:Zakim D,Boyer T, eds. Hepatology.,A Textbook of LivarDisease. Philadelphia,PA: W. B. Saunders.. 1990:791-821. 138. Hodgson M, Van TheilD, Lauschus K, Karpf M. Uver injurytestsin harzardous waste workers: the mis of obesity.JOM 1989;31: 238-242. 139. Ludwig J, Viggiano T, McGill D, OttS. Nonalcoholic Steatohepatitis.Mayo ClinPmc 1980*,55:434-438. 140. DiehlA, Goodman Z, lshak K. Alcohol-likleiverdisease in non-alcoholics. Gastementerology 1988;95:1056-62- 141. Cornish H. Adefin J. Ethanol potentiatioonf halogenated aliphatcsolvent toxicityA.m Ind Hyg Assoc J 1966;27:57-61. 142. Charbonneau M, Tuchweber 8, Plaa G. Acetone potentiationof chronic liverinjuryinduced by repetivieadminstrationof carbon tetrachloride. Hepatology 1986;: 694-700. 143. PallT, SchleiferL Drugs Effectiveinthe Tnerapies of the Epilepsies.In: Gilman A. atal.,eds. The Pharmacological Basis of 7herapeudcs. New York: Pergamon. 1990:450-451. 244 144. Schuckh M, IrwinM. DiagnosisofAlcoholism.Med Clinof N A 1988;72: 1133-53. 145. SchuckitM, GriffithJs.Gamma glutamyitranesrfase values innonalcoholic drinkingmen. Am J Psychiatry1982,139:227-225. 146. Orrego H, SlakeJ,IsraeYl. Ftelationshbieptween gamma glu=yl transpeptidaseand mean urinaryalcohollevelsinalcoholicwshiledrinking and inwithdrawal.AlcoholClinExp Res 19B5;9:10-13. 147. Bates H. GGTP and alcoholisma: sober look.Lab Management 1981*,9l: 1-3. 148. Chan-Yeung M, FerreiraP, FmlichJ.etal.The effectsofage, smoking, and alcoholon routinelaboratortyests.Am J ClinPathol1981;75:321-26. 149. Purchase 1.Inter-speciecsomparisons of carcinogenicitByr. J Cancer 1980:41:454-468. 150. The 3M company R.Two Year OralToxicity/CardnogenicitSytudy ofFC143.1986, RikerLaboratories: 151. Kennedy B.GilbertsoAn. Increasederythmpoiesisinduced by androgenic hormone therapy.NEJM 1957;256:719. 152. StainglassP,Gordon A. CharipperH. Effectofcastrationand sex hormones on thebloodofmts.Pmc Soc Exp BiolMod 1941:48:169. 153. Rishpon-MeyerMoin N, )CilbridTg,eSimons J,FriedW. The effectof testostemneon orythopoistilnevelsinanemic patientsB.lood 1968;31: 453-460. 154. AlexanianR. Erythropoietiand erythopoiesiisnanemic men following androgens.Blood 1969;33:564. 245 155. ShahidiN. Andmgens and arylhmpaiesis.NF-JM 1973;289:72. 156. PalacicsA. CampfieldL,McClure R, SteinerS. SwordioffR. Effectof lestosteroneonarithateon hematopoeisisinnormaj men. Fertilitaynd Sterilit1y983;40:100-104. 157. Cunningham G, SilvermanV, Thomby J,KohlerP. The potentialforan androgen male contraceptiveJ. ClinEndocrinotMetab 1979;49: 520. 158. Mauss J, BorschG, Bormacher K, RichterE, Loyendock G. Nocke W. Effectoflongterm testosteronesnanthate on male reproductivefurwction. Acta Endocrinof(Copenh') 1975;78: 373-384. 159. TellG, Grimm Jr.R, Vellar0, Theodorsen L The relatonshipofwhite cell coum platelectount,and hematocrftto cigarettesmoking in adolescents.-thOeslo Youth Study.Circulatio1n985,72: 971-974. 160. Bunn H. Approach tothe patientwith anemia. In:Thom G. etat.,eds. Hafrfson'sPfincipleosf lntemat Medicine. NewYork:McGrawHill,1.977:1645-1651. 161. Hansen L,Grimm Jr.R. Neaton J.The relationshiopfwhiteblood cellcount and othe cardlovascuairiskfactors.IntJ ofEpidem 1990;19:881-888. 162. de Labry t,Campion E, Glynn F:LVokonas P. Whfte blood cellcount as a predictorofmortalityR:esufts over 18 years from the normativeaging study.J ClinEpidemiol1990;43:153-157. 163. Grimm Jr.R. Neaton J, LugvAg W. Pmgnostic importance ofthe white blood count forcomnary, cancer, and all-causemortalityJ.AMA 1985;254: 1932-1937. 164. ZaJokarJ. RichardJ,Claude J. Leukocyte count,smoking and Myocarcrial infarctionN.EJM 1981 *.304:465-468. -246 1SS. FriedmanG. KlatskAy,SiegelauAb.The leukocylceountas a predictoorf myocardialinfarctioNnE.JM 1974:290:1275-1278. 166. Friedman G. Fireman S. The leukocylecount and cancer mortalityA.m J Epidemiol 1991;133:376-380. 167. Kannel W, Anderson K. Wilson P. White blood cellcount and cardiovasculardisease.Insightsfrom the Framingham Study.JAMA 1992;267:1253-1256. 168. ManttariM, Manninen V. Koshinan P, etal.Loukocytes as a coronary risk factorn a dyslipidemimcale population.Am JeartJ 1992:123: 873-7. 169. K@shman E. Tmst L,Aamns S. Study ofthe functionand maturationof monocytes inmorbidlyobese individualsJ.Surg Res 1982:33: 89-97. 170. Taylor,R, Gross E, Joyce H, Holland F, P6do N. Smoking, allergyand the differentiwahliteblood cellcount.Thorax 1985;40:17-22. 171. WalterS, WalterA. Smoking and blood basophits.Thorax 1986;41:335. 172. WalterS, WalterA. Basophildegranulationinduced by cigarettesmoking in man. Thorax 1982,,37:756-759. 173. WalterS. Blood Basophilcounts insmokers and nonsmokers. IndianJ Mod Res 1982:76:317-9. 174. WalterS, Nancy N. Basopenia followingcigarettesmoking. IndianJ Mod Res 1980;72:422-5. 17S. Youssef J,lqwe 0, Cunningham M. Regulationof hepaticinositol thsphosphatereceptorsby pemxisome proliferatorTso.xicologist1992:12: 37. Z47 176. Robinson J,PfeiferR. Now t&chnologiesforuse intoxicologyswdies: MonitoringtheeffOctSof xenobi0tiCoSn immune functionJ. Am Col Toxicol Io 1990;9: 303-317. 177. TollerudD, ClarkJ, Brown I,ata/.The effectofcigarettesmoking on T cell subsets.Am J Respir Dis 1989;139: 1446-146 1. 178. Dean J,Comacaff J, RosenthalG, LusterM. Immune system: EvaJuationof injuryI.n:Hayes A, eds. Ptincipleasnd Methocfsof Toxi=iogy. NewYork: Raven Press.1989:741-760. 179. Davis J, Davis R. AcL49 effectsoftobacco cigarettesmoking onthe platelet aggregationratioA.m J Med Sci 1978;278:139-143. lao. FusterV, Chesebro J,Frye R, E)veback L Plateletsurvivaland the development ofcoronary arterydisessinthe young aduft:Effectsof cigaretteamoking, strongfamilyhistorya,nd mecricalthreapy.Circulation 1981;63: 546-551. 181. Belch J, McArdle B, Bums P. The effectsofactae smoking on platelet behavior,fibdnolysisa,nd hematology inhabitualsmokers. Thromb Haemostas 1954,51:6-8. 182. Murchison I,Fyfe T, Lowe G, Forbes C. Effectsof cigarettesmoking on serum-lipidsb,lood glucose,and plateleatdhesiveness. Lancet 1966;1:182184. 183. FitzGeral(GJ, Oates J, Nowak J.Cigarettesmoking and hemostatic functionA.m Heart J 1988:115:267-271. 184. Renaud S. Blanche D, Dumont D. Thevenon C. Wissendanger T. Platelet functionaftercigarettesmoking inrelatiotno nicotineand carbon monoxide. ClinPharmacol Ther 1964;36: 389-395. 248 185.GreenM,Peled1,NajensoTn.Genderdifferencienpslatelceotuntandits as.sociatiownithcigarenesmc@kingina largecohortin IsraelJ. Clin Epidemiol1992:45:77-84. 186. Packman M, MustardJ.The roleofplateletisnthe development and complicationsofatherosclerosiSsa.min Hematol 1986:23:8-26. 187. Mehta J. Mehta P. Role of blood plateletsincoronary arterydisease. Am i Cardiol1981:48: 366-373. 188. PerkinsR. Investigatiofnammonium perfluomoctanoateaffecton hormone levelsand peroxysome proliferatiionnthe rat.Toxicologist 1992:12:38. 189. Checkoway H. PierceN. Crawford-BrownD. Research Methods in OccupationalEpiemiology.New York:Oxford UniverstiyPress,1989. 190. RiessJ.Reassessment ofcrite@ifaorthe selectionof perfluorocarbonfsor second generationblood substitutesA.rtifiOcrg 19SA;S:44-SS. 191. Gilman A, RailT, Nies A,TaylorP.The PharmacologicalBasisof Therapeutics.New York:Pergamon, 1990. 192. Burke W. Hoegg U. Systemic fluc@depoisoningresultinfgrom a fluoride skinbum. JOM 1973;15:319-41. 193. Bass N, Ockner R. Drug-InducedUver Disease.In:Zakin M D,Boyer T, ads.Hapatology.-ATextbookofLiverDisease. PhiladelphiaP,A:W.B. Saunders. 1990:754-790. 194. KuslikiSs,Vanden Heuvel J,PetersenR. Lack ofevidence for perfluorodecancylo-rperfluoroactanoyl-CofAormationinmaje and female rats.J Siochem Toxicology1992:7:25-36. 249 195. DaiW, KulliLoLraPortRe,GuW J.Faivo-GeraCr,dCaggiulAa.The epidemiologyofplasma testosteronleevelsinmiddle-agedmen. Am J Epi 1981;114: 804-816. 196. K'IesgesL,KlesgesR, CigrangJ.Discrepenciesbetween self-reported smoking and carboxyhemoglobin:An analysisofthe Second National Healthand NutritioSnurvey.Am J Pub Health1992;82:1026-1029. 197. Sagone A, BaJcerzakS. Smoking as a cause of ErythrocytosisA.nn IntMed 1975;82:512-515. 198. Schwartz J,Weiss T. Host and environmentalfactorsinfluencintghe peripherablloodleukocytocount.Am J Epidem '1991;134:1402-1409. 199. Yamell J,Sweetnam P, Rogers S, otat.Some long term effectsof smoking on the haemostaicsystem.ClinPathol1987;40:909-913. 200. FraserG. PreventiveCardiology.New York:Oxford UniversitPyress,1986. 201. Hames C. Heyden S,TyrolerH, Heiss G, Cooper G. Manegold C.The combined effecotfsmoking and coffeedrinkingon LDL-HDL cholesterol. Am J CarcUal197a;41:40" 10. 2OZ GarrisionR. Kannel W, FeirdabM. stal.Cigarettesmoking and HOL cholesterolA.themscAsmsis 1978;30:17-21. 203. Olsen G, Kusch G, StaffordBA. Gudmyndsen SL, CurrierM,F. The positive known associatiodnesign:A qualityassurance method foroccupational healthsurveillancdeata.JOM 1991;33:998-1000. 204. Lumeng L New diagnosticmarkers ofalcoholabuse.Hepatology1986:6: 742-745. 205. Castel[W!. Doyle J.Gordon T. Alcoholand blood lipidsT:he Cooperative Upopmtein PhenotypingStudy.Lancet1977;2:153-160. 250 AL. 206. SSifrageP, Berg 8, HageMrand 1.AJteration'loifpidmetaboilisminhealthy volunteersduringlong term alcoholintake.Eur i ClinInvest1977;7: 127. 131. 207. Undenbaum J, LeiberC. Hematologic effectsof alcoholinman in absence ofnuldtionadleficienciesN. Eng J Med 1969:281: 333-338. 208. Whitehead T. C12rkeC, WhitfieldA. Biochemical and hematological markers ofalcoholintake.Lancet 1978i: 978-981. 209. Skinner H, HoltS, SchullerR, Roy J,IsraelY. Identificatioofnalcohol abuse usinglaboratorytestsand a historyoftrauma. Ann of lntem Med 1984:101:847-851. 210. Moussavian S, Becker R, Piepmeyer J.Serum gamma-glutamyi transpetidassand chronica)=holism. Dig Dis sci1985;30:211. 211. Matoff D, SelingerM, Kaplan M. Hepatictransanimass activitiynalcoholic liverdisease.Gastroenterology1980;78: 1389-1394. 212. Canik J.The effectofsmoking on hormone levelsinvivoand stemid hormone biosynthesisinvftm.In:Wald N,Bamn J,ods. Smoking and Ho,rmone-ReaftedDisorders. Oxford:Oxford UniversityPress,.1990 208213. 213. PardMdge W. Transportof proteinbound hormones intotssues inVivo. Endo Rev 1981;2:103-123. 214. Anderson 0. Sex-hormone bincringlobulin.ClinEndoc@nol 1974;3:69-96. 21S. Rosner W, Aden 0, Khan M. Hormonal influenceson the secrebon of steroid-bincripnmgtein3 by a human hepatoma derivedcelllineJ. Clin EndocrinolMetab 1984;59: 806-808. 251 21S. Joseph D, HallS, Yarbrough W, CortiM, French F. Structureof the rat androgen bindingprotein.Ann NY Ac-ad Sci Igae;S3e: 31-36. 217. Aono T, Kurachi K, Mizutani S, eta/ Influenceof major surgicajstress on plasma levelsoftestostemne,lldenizinhgormone, and folliclsetimulating hormone. J ClinEndocrinofMetab 1972;35: S3-r-542. 218. Gidlow D, Church J,Clayton S. Hematologicaland biochemicajparameters inan industriaw)orkforce.Ann ClinSlochem 1983;20:341-348. 219. Anderson K Posner W, Khan M, ata/.Dist-hormone interactions: P mtsinicarbohydrateratioajtersrecipmcallythe plasma levelsof testosteroneand cortisoalnd theirrespectivebincftnglobulinsin man. Ufo Sci 1987;40:1761-68. 220. Reed M, Cheng R. Simmonds M. Richmond W, James V. Dietarylipidsa:n additionarlegulatorofplasma levelsof sex hormone bindingglobulinJ. Clin EndocrinofMetab 1987;64:723-729. 221. Bishop 0, Melkle A, SlatteryK Stringham J, Ford M. West D. The effectof nutrffionfaalctorson sex hormone levelsinmale twins.Gen Epid 1988;5: 43-59. 222. WichnoviczJ. Bradlow H. Inductionofestradiolmetabolism by dietary indole-3-carinoilnhumans. JNCI 1990,82:947-949. 223. SuttonJ,Coleman M. Casey J.Androgen responses duringphysical exercise.Br Med J 1973:1:520-2. 224. Nilsson 0. Fords 0, Brenn T. The Tmmso Study. the distributioannd populationdeterminantsof gamma-git4amyl transfemse.A J E 1990:132: 318-326. - --252 225. KleinbaumDG, Kupper LL AppliedRegressionAnalysisand (Dther MuftivaraiTtechniques.Belmont,CA: LifetimeLeaming Publications. 1978. 226. GilbertE. Some confoundingfactorsinthe studymortalitaynd occupational exposures.Am J Epidemiol1982:116:177-188. 227. Fox A, CollieFr.Low mortalitryatesinindustriaclohortstudiesdue to selectiofnorwork and survivalinindustryB.r J Prev Sac Med 1976:30: 225-230. 228. BreslowN, Chan C, Dhom G, of a/.Latentcarcinoma ofprostateat autopsy inseven areas.lntJ Cancer 1977;20:680-688. 229. Nomura A, KolonalL ProsiateCancer A curtentperspectiveA.m J Epidemiol1991;13:200-227. 230. Johnasson J,Adami H. Anderson S, Bergstrom R, Krusemo U, Kra27 W. Naturalhistoroyflocalizedpmstaticcancer.Lancet 1989i: 799@-803. 231. Mei;deA, SmithJ.Epidemiologyofpmstatecancer.Uml ClinN AM --1990;17:-70S@-718@t 232. Percy C, Stanek E,GloeckJerL Accuracy ofcancer death certificataensd itseffecotn cancermortalitsytatisticAsm. J PublicHealth1981,71:242-50. 233. Rosenberg H.lmpmving cause-of-deathstatisticAsJ.PH 1989;79:563-4. 234. )GrrherT,NelsonJ,Burdo H. The autopsy as a measure ofaccuracyofthe death certificaNte.Eng J Med 198S;313:1267-73. 235. CarterJ.The pmblematic death certificatNe.Eng J Med 1985;313:12851286. 253 236. Rothman K Modem Epidemiology.Boston:Urie,Brown and Company, 237. SiemiatyckiJ,Wacholder S. Dewar R, stat.Smoking and degree of occupationaelxposure:are internaalnalysesincohortstudieslikeltyobe confounded by smoking status?Am J IndMed 1988;13:59-69. 238. Robins J.A new approach tocausalinferenceinmortalitsytudieswitha sustainedexposure period.Math Modeling1986;7:1393-1512. 264 0 - 0 APPENDIX 1 .PHYSIOLOGIC EFFECTS STUDY QUESTIONNAIRE 0 i - - -- 255 0 Medical HistoryQuestionnaire 0 256 0 .16.06 Medical HistoryQuestionnaire ED ED MedicalSymptoms 81 Ow row hom km p pw im ftm rwL PW%MWIL*Whwuu%o U"Ical Diagnoses .0@ @ @ - Yes we we W "V wwomm C) cl El El 6--" A--q -may a-m w waspwqpmw@ Ov-v doftm? va - iw cl C3 4L Pwom" meow%" Svm C) 0 7.amo&omwwwm@ 0 C3 LCM"ip&nrfatw@ C) C3 in owe ar ww om *son cl 0 seakmyff"am@" C) C3 0 31 Aod a XL I? C---rv M" de@ ZL Mom atom 2L ' p 31LA-Mwm&msgg 311H."Ndn ILDNFAUWQM boftnft C3 a 'romeonwavyW.wmt.mw owwowCm3 ED Pwo@v womwmm in mma w fte 0 0 IL Plan" 0 77. I A. **GaAs" in wm w we 2Lemme. u" x or. 6. x amem mw cl C3 --- cl cl 0 0 ECD, C3 C3 0 C3 C3 cl 0 C3 C3 C3 d-ftOMAU Onvw &meowocummt INMWO popIV C3 C) -Y. 4L ArAw-aftweandi 4& Lmmopt"mom dotown ".am$OMM 0.~ ALcrawrmw C3 0 C3 C3 C3 C3 C) Cl 0 2p . f fio zo di 3. loo (I Li Ll 0 li 0 L) ci 0 0 0 [1 cl 0 ip 00 f] A 1 1! Is IL qi i ,it APPENDIX 2 TABLES OF HORMONE RATIOS BY BODY MASS INDEX, AGE, SMOKING STATUS, AND ALCOHOL CONSUMPTION 259 Am& TABLE A4.1.1 BOUND TESTOSTERONE TO FREE TR=STOSTERONE (TB/TF) BY BODY MASS INDE)C. AGE. SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOUTE PLANT, COTTAGE GROV;:. MINNESOTA RATIO TBFTF N MEAN so MEDIAN RANGE-, TEST# smi .c2s 25-30 21= AO 372 9.06 37.1 22.3-62-S F.i.A,7 56 37.6 2-31 37.1 19.3-62.4 pm.23 17 33.3 9.18 312 19.7-S2.4 AGE -c3l 31-40 41-50 al-W 20 32.A 6.92 31.2 20.0-A3.6 F-.Z39 AS 37.3 9.37 37.1 19.3-62-6 p-.07 26 37.1 9.30 38.8 19.7-SS.$ 19 39.9 9.90 32.9 22.3-62.4 Alcohol -elcizfd as 37.4 9.90 372 19.3-0-6 Fw42-O6 1-30zfd 19 33.9 6.70 222 22.7-4A.i pm.IS missing 6 38.0 S.70 38.6 26.4-43-S Tobacco smoker 27 nonsmoker 64 missing 2 37.9 7.96 37.2 2S.0-58.6 F-.32 36.7 9.64 36.3 19.3-Me pm.57 27-S I.S7 27.S 26.A-28.6 TOTAL 113 #univariateAnova 260 TABLE A4.1.2ESTRADIOL TO FREE TESTOSTE;;ONE RATIO (EITF) BY BODY MASS INDEX, AGE. SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY. 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA EITF N MEAN SD MEDIAN RANGE TEST# BMI -c25 25-30 3-M AGE .c3l 31-40 41-50 51-W 40 2.07 0." 56 2m o.92 2-17 0.76-5.39.,,30yu,30 17 2.56 0.98 2-42 I.AA-5.31 T-2M p-.02 20 1.94 Q.SS 1.81 I.AA-3.27 Fol.lo AS 2.25 0.81 2.17 0.77-4.18 p-.32 26 2.31 120 2.04 0.77-5.39 19 2-48 1.05 2.42 1.07-5M Alcohol -clazid 1.3or/d missing as 2M 0.92 2.10 0.74@S.31 F..ol 19 2.21 o.76 2.21 o.73-4.is P-.92 2.46 1.96 2-09 1.41-5.39 Tobacco smoker 27 nonsmoker 84 missing 2 2.19 2.27 1.u 0.98 o.92 2.06 0.74-5.39 Fe.15 L19 0.73-S.31 ;@-.70 I.ES-2-11 WOTAL 113 #univariatAenova *Studentt test,Pmb>T T# .70 )3 07 go .08 .98 1.01 .32 261 TABLE A4.1.3ESTRADIOL TO BOUND TESTOSTERONE FLATIO (E/TB) BY BODY MASS INDE)@ AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY. 3M CHEMOLITE PL.ANT, COTTAGE GROVE. MINNESOTA ErrB xioo N MEAN so MEDIAN RANGE TEST* smi -c25 AO 2-9-30 56 3b,= 17 AGE .c3l 20 31-0 48 41-50 26 19 Alcohol .Clo7jd 96 1-30YJd 19 missing 3 Tobacco smoker 27 nonsmoker SA missing 2 TOTAL 113 5.8 2.65 5.7 11.8-13.5 F.3.70 u 2.73 5-S Z2-13.4 P..03 8.0 2-91 7.6 3.7-i4,A 6.1 1.79 5.2 F-.07 6.4 L76 5.6 11.8-13Z pm.98 GZ 3m 5.0 1.7-14.4 6.4 2.76 S.o 2.3-11.6 6.3 2.90 5.7 12-14.4 Fa.08 6$ 1.98 6.8 3.0-10.5 P..Ss 5.5 3.45 5.4 3.9@-13.3 5.9 L73 5.1 11.8-13.5 Fol.01 6.5 2-84 6.0 2.2-13.4 p..32 6.9 I_ss 6.9 3.7-14.4 ounivariateAnova 262 TABLE A4.1.4ESTRADIOL TO LUTENIZING HORMONE RATIO (E/LH) By BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA EILM N MEAN -SD MEDIAN RANGE TEST* BMI -c25 25-30 3-M AC 7.0 3.li 7.3 2.0-16.4 F%2.Sg 5167 7u.0 43..2992 69..58 31..30--1280..46 p.-Ca AGE ,c3l 31-40 41-50 51-M 20 L7 ASI 7.6 2-1-20.6 F-2.si 4S 7.0 3.35 7.6 26 6.A 4.58 5.1 19 5.9 2.89 6.2 Alcohol -clez/d as 1-30ztd 19 missing 8 72 3.95 7.o 1.0-20.6 Fa.04 7.4 4.16 6.7 I.&IG.4 P-.86 8.5 3.19 8.7 4-1-15.4 Tobacco xmokar 27 naftsmoker 94 missing 2 7.0 4.42 62 1-e-IS.4 Fe.34 7.5 3.77 7.1 1.0-20.6 4.7 3.35 4.7 TOTAL 113 #univariatAenova 263 TABLE A4.1.5FREE TESTOSTERONE TO LUTENLZING HORMONE (rF/LH)BY BODY MASS INDEX, AGE. SMOKING AND DRINKING 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA RATIO STATUS TFILM N MEAN SD MEDIAN RANGE TEST9 BMI -c2S 25-30 3.@w 40 3.8 1.74 3.3 1.1-11.3 Fal.47 so 32 1.74 Z9 0.7-2.1 P-.24 17 3.9 1.73 3.4 1.4-7.1 AGE Wal 31-0 41-50 51-W 20 4A 2-9 19 3.9 F.7.21 As 3.6 1.43 3.3 3@3 p-.0002 26 Ls 1.30 2.7 L7 19 2-5 1.14 7-5 2.5 Alcohol -cloZ/d 1-30zfd 19 missing 8 3.A 1.w 32 0.7-11.3 F..06 3J I.A4 u o.7-a.4 p..Bl 3.2 1.43 3.6 2.1-62 Tobacco SMoker 27 nonsmoker 64 missing 2 3.2 1.34 3.2 1.1-6.9 Fml 20 3.6 IJ7 3.2 0.7-11.3 P-28 2-A _u Z3 1.4%U TOTAL 113 #univariatAenova 264 TABLE A4.1.6BOUND TESTOSTERONE TO LUTENIZING HORMONE RATIO (TBA-H) BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA TS/LM N MEAN SD MEDIAN RANGE TEST# BMI .c25 40 25-30 56 2-M 17 AGE .c3l 20 3140 As 41-50 26 51-W 19 Alcohol -clez/d so 1-3azId 19 missing a Tobacco wnoker 27 nonsmoker 64 missing 2 TOTAL 113 131 57.9 116 60.6 122 43.4 147 69.7 133 58.0 101 3&7 n 47.6 122 57.1 IIA 5&1 147 612 116 SZA 125 S&S 64 .34.3 12S 39-298 Fa.79 107 24-288 p-.46 121 55-199 136 39-296 Fu4.72 131 36-288 p-.004 m 29--163 96 24-206 121 24-295 Fa.32 92 29-2M p.S7 142 77-2U 114 41-288 F.-U 121 24-298 P..46 64 39-88 #univariatAenova 265 TAeLr: A4,i.7 THYROID STIMULATING HORMONE TO LUTENIZING HORMONE RAT40 (rSKA.H) BY BODY MASS INDEX AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY. 3M CHEMOLrrE PLANT, COTTAGE GROVE, MINNESOTA TSTVLH xIO N MEAN SD MEDIAN RANGE TR=ST# smi -CIS 25-30 40 32 1.79 32 0.&4.3 F-3.Ao ss 3-S "0 3.o 0.4-17.0 P-.02 17 3 -U 4.4 1.7-13.S AGE -c3l 20 31-40 AS 41-50 26 51-W Alcohol .Clcwd Be 1-3*zfd 19 missing 3.7 2.AS 3.8 3.16 3.4 1.89 3.8 2-43 3S 2-n A.7 4.06 3.3 1.74 3.1 1.0-9.9 Fu.14 3.2 0.4-17.0 p-.93 2-9 C.".3 3.S O.A-1 1.0 3.1 0.4-11.0 F*2-92 3.2 1.0-17.0 P-.09 3.6 O.S-S.$ Tobacco *Mokor 27 nonsmoker 84 missing 2 TOTAL 113 3.0 1.68 4.0 2.89 1.7 0.01 7-3 O.A-7.6 FoP-89 3.3 0.4-17.0 p..09 1.7 1.7-1.7 #univariateAnova 266 TABLE A4.1.eFOLLICLE STIMULATING HORMONE TO LLJTENIZING HORMONE RATIO (FSR") By BODY MASS INDEX. AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA FSWLH N MEAN so MEDIAN RANGE TEST# smi 025 40 2@5-30 56 30= 17 AGE w.31 20 31-40 AS 41-50 26 51-W 12 Alcohol -ciovd 86 1-30ztd 12 missing 8 Tobacco smoker 27 nonsmoker 84 missing 2 TOTAL 113 #univaiiatAenova 1.0 O.A2 0.9 0.4-2.3 Fm2.$A 1.0 OJ7 0.9 0.4-1.9 P-.Oa 1.2 0.46 1.1 Q.A-ZD o.9 0.4 0.8 O.S-1.9 F.3.0-6 o.9 O.Ag o.9 O.A-1.9 p..03 1.1 0.45 1.0 O.A.1.8 12 0.49 1.1 0.-r-Z,3 1.0 0.42 o.9 0.4-2-3 0.9 0.41 0.9 O.A-1.7 F-.48 0.9 0.24 0.9 O-rv,-1 a P-.49 1.0 0.40 0.9 OA-1.8 1.0 0.42 0.9 O.A-2-3 F=0.0 0.9 0.43 0.9 0.7-1.0 pm.98 267 TABLE A4.1.9PROLACTIN TO LUTENIZING HORMONE RATIO (P/LH) BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOFTOCHEMICAL EFFECTS STUDY, 3M CHEMOUTE PLANT, COTTAGE GROVE, MINNESOTA N MEAN P/LM SD MEDIAN RANG-E TEST* smi -c25 2-c-30 i-M AC 1.84 1.14 im C.MS.39 F..73 56 1.78 1.40 1.46 0.39-9.11 p..49 17 Z.21 120 1.93 1.184.91 AGE 431 20 31-AD As 41-50 26 51-W 19 Alcohol .elo7Jd 86 1-3azid 19 missing a Tobacco smoker 27 nonsmoker 94 missing 2 TOTAL 113 2-a 1.70 1.23 1.124.91 Fal-W 1.96 I-Sa 1.97 0.39@-2.1 P=21 1.61 0.93 1.18 0.5&3.70 1.54 0.37 1.39 0.35-3.37 1.78 1.04 1.61 o-u4j9 Fm3.12 2.37 L14 i.7c 0-56-0.11 P..Os 1.57 0.71 1.47 0.8i@-3.00 I.Z7 0.73 1.12 0.39-L74 FeS.25 7-07 im 1.72 0.35-9.11 p..005 1.43 o.78 1.43 O.IW-LOO #univariateAnova 268 TABLEA4.1.10BOUND TESTOSTERONE TO THYROID STIMULATING HORMONE RATIO (fB/TSH) BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOUTE PLANT, C07TAGE GROVE. MINNESOTA N BMI ez 40 2-r--30 SG 31@m 17 AGE .c3l 20 3140 48 41-SO 26 51-60 19 Alcohol -cclzjd as 1-30ztd 19 missing 8 Tobacco wnokor 27 nonsmoker 94 missing 2 TOTAL 113 MEAN Soo 461 296 522 S21 3se 328 488 329 593 ASS "0 371 TSITSM so MEDIAN RANGE TEST# 331 413 l7D-lU2 Fm2.42 364 329 51-2102 p-.09 152 297 87-SB9 367 410 122-1682 FmZS4 398 421 51-2102 203 3AS 131-IMS 231 286 87-1122 352 353 87-21()2 Fo2.74 210 278 si-wo pm.10 326 4s6 1"1154 403 IBA-1195 Fa.07 364 321 51-2102 is$ 371 231-511 #univadate Anova .269 TABLE A4.1.11FREE TESTOSTERONE TO THYROID STIMULATING HORMONE RATIO (rF/TSH) BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS 1990 PERFLUCFIOCHEMICAL EFFECTS STUDY, 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA N smi -c25 40 25-30 56 :,.30 17 AGE ,c3l 20 31-40 AS 41-50 26 51-60 19 Alcohol ,cIaztd 86 1-3ozJd 19 missing a Tobacco smoker 27 nonsmoker 54 missing 2 TOTAL 113 MEAN 13.8 11.7 9-3 TFITSH SD 8.62 7.01 5.05 MEDIAN 10.8 9.7 8.0 RANGE 3.8-43.7 i.7-M.8 2.0-19.7 TEST# snow F.2.43 p..09 15.8 13.,d 9.7 8.1 JZ4 9.5 15.3 9.87 7.rx2 4.Ss 4,46 1.65 S.10 Qzs 10.8 11.9 6.7 7.8 10.0 8.7 12.3 6.1-43.7 1.7-U.0 3.7-ZLC LO-21.3 FuS.36 p-.002 LO-43.7 1.7-20.7 5.0-33.5 Fm2.48 pe.12 12.5 5.53 11.3 S.C@-27.2 F..12 11.9 8.06 9.8 1.7-43.7 pa-73 13.7 7.99 13.7 8.1-19.4 '--#univariaAtneova 270 TABLE A4.1.12ESTRADIOL TO TH YROID STIMULATING HORMONE RATIO (E/TSH) BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EF@FECTS STUDY, 3M CHEMOUTE PLANT, COTTAGE GROVE, MINNESOTA E/TSH N MEAN SD MEDIAN RANGE TEST# BMI -c25 25-30 40 26.7 17.73 Z3.i 3.3-108.1 F-.27 S6 25.4 14.04 21.3 1.6-59.0 p..76 17 23.4 IS.79 18.4 7.7-54.8 AGE .931 20 31-0 48 41-50 26 51-W 19 Alcohol -cioztd 86 1-3oVd 19 missing 8 Tobacco smoker 27 nonsmoker 84 missing 2 TOTAL 113 27.3 12-00 24.6 2.7-50.0 F-3.21 29S 18.40 223 1.8-108.1 po-03 21.8 13.06 12.0 2@3-52.2 18.2 10.80 16.8 7.8-U.8 25.1 13.08 21S 3.3-59.0 F-.S7 22.4 15.30 18.0 1.8-U.2 pe.45 37.8 31.81 28.S 10.4-108.1 26.7 IS.13 23.4 10.3-SD.O Fo.20 25.1 is.as 20.7 I.S.109.1 P-.65 27.1 19.40 27.1 13.4-0.3 #univariatAenava 271 TABLE A4.1.13THYROID STIMULATING HORMONE TO FOLLICLE STIMULATING HORMONE RATIO (TSK/FSH) is BY BODY MASS INC)EX,AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY. 3M CHEMOUTE PLANT, COTTAGE GROVE, MINNESOTA TSH/FSH N MEAN SD MEDIAN RANGE TEST# BMI ,c25 40 25-30 56 .b3O 17 AGE .c3l 20 31-40 AS 41-50 26 51-60 is Alcohol -cioztd 86 1-3oztd 19 missing a Tobacco smoker . 27 nonsmoker 84 missing 2 TOTAL 113 0.39 025 0.41 0.39 0.34 0.08-1.09 F..39 "1 O.Or.-2-u pa.se 0.44 1.29 0.43 0.15-1.26 0.46 031 0." 0.12-1.26 Fa.93 O.AS 0.40 0.35 0.06-2.34 p-.A3 0.37 0.28 0.30 0.09-121 am 0.18 ozi 0.06-0.75 0.38 026 0.32 .06-1.26 F=S.36 O.Sa O-U 0.39 O.lS-2-U P-.02 0.40 0.24 0.39 0.0".76 0.34 1.064 0.26 0.06-0.92 Fm2.39 0.45 228 0.40 0.06-2.34 P-.12 0.21 0.05 0.21 0.17-02S univariatAenova 272 TABLE A4.1.14FREE TESTOSTERONE To FOLLICLE STIMULATING HORMONE RATIO (TFIFSH) - BY IBODY MASS INDEX AGE. SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLITE PLANT, COTTAGE GROVE, MINNESOTA BMI 425 25-30 30-M TF/FSH N MEAN so MEDIAN RANGE TEST# W"NWMN@ 40 AZ 2-se 3.6 1.3-15.6 F-1.03 S6 3.6 2-18 3.0 0.7-11.1 P-.36 17 4.0 2.96 3.1 AGE @c3l 31-Q 41-50 51-60 20 S.8 3-U 5.0 1.'7-IS.6 F.io.3s AS A.2 2.16 3.7 1.7-11.3 p-.0001 26 3.0 1,62 2.S 0.7-6.6 19 2.2 1.37 2.0 0.7-6.3 Alcohol .cloztd 86 1-3CZtd 19 missing a 2.8 2.60 3.1 0.7-IS.6 F-.oi 3.9 2.18 3.7 1.3-10.1 po.91 4.A 1.67 A.7 2.2-7.3 Tobacco smoker 27 nonsmoker 94 missing 2 3.5 1.76 3.0 0.7-7-S Fal.14 4.1 2.67 3.S 0.7-IS.6 p-.28 2.6 0.91 2.6 2.2-7.3 TOTAL 113 #univariateAnova 273 TABLE A4.1.15BOUND TESTOSTEFIONE TO FOLLICLE STIMULATING HORMONS RATIO rrBIFSH) 13YBC)OY MASS INDEX. AGE, SMOKING AND DFTINKING STATus 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLITE PLANT, C;OTTAGE.GROVE, MINNESOTA smi -c25 25-30 3-M AGE Ir.31 314.0 41-50 51-W Alcohol -clez/d 1-3oVd mlssing Tobacco smoker nonsmoker mis3lng TOTAL N 40 56 17 20 48 26 19 as 19 a 2"7 '- 2 113 MEAN 155 126 122 182 154 101 87 133 133 173 130 139 72 TF/FSH SD MEDIAN RANGE 87.3 67.9 7S.7 90.3 74.5 52.9 S3.2 76-S 78.6 80.5 74.4 78.5 70.9 138 39-411 113 23-297 115 3A-306 184 394'11 141 45-348 91 39@-227 78 23-264 120 23-411 112 39-325 190 82-303 106 39-32S 131 23-411 72 57-86 TEST F-2.oo PM.IA F-8.75 P-.0001 F-c.o p-.98 Fu28 p-.60 --.-#utiiyariAantaeva 274 ---OWL TABLE A4.1-16ESTRADIOL TO FOLLICLE STIMULATING HORMONE RATIO (E/FSH) BY BODY MASS INDEX, AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLITE PLANT, C07rAGE GROVE, MINNESOTA N smi @c2S AO 25-30 56 3-M 17 AGE 421 20 3140 48 41-50 26 51-W 19 Alcohol ,clozjd as 1-3ozId 19 missing 8 Tobacco smoker 27 nonsmoker 94 missing 2 TOTAL 113 MEAN 9.7 7.8 9.3 10.8 91 6.9 4.9 8.1 u 10.9 8.1 5.1 E/FSH SD S.Si 5.76 7.o4 sio 6.01 6.13 227 5.83 4.99 7.U 6.82 5.59 ZS6 MED-IAN RANGE TEST# 6.8 1.3-22.3 F-.so 6.6 1.4-29.6 P-.Bl 7.6 2.65-33.1 9.6 3.1-25.o Fm6.oo 7.3 1.4-33.1 pm.003 4.0 1.3-29.6 4.6 1.5-9.3 6.6 1.3-33.1 Fe.01 7.1 3.1-19.1 P-.91 23 4.6-29.6 4.7 1.4-29.6 F-.13 7.o 1.3-33.1 p-.72 S.i 3.2-6.9 #univariatAenova 275 TABLE A4.1.17THYROID STIMULATING HORMONE TO PROLACTIN RATIO (TSHIP) BY BODY MASS INDEX. AGE. SMOKING AN&) DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOUTE PLANT, COTTAGE GROVE, MINNESOTA TSWP N MEAN SD MEDIAN RANGE TEST# BMI 425 25-30 31@w 40 0.22 0.17 0.18 0.06-0-83 F..71 se 0.2S 0.22 0.19 0.02-1.21 P-.49 17 0.28 0.19 026 0.07-0.81 AGE -c3l 3140 41-SO 51-60 20 0.17 0.02 0.15 0.05-0.39 F-1.38 48 0.25 023 0.17 0.02-1.21 P-25 26 0.26 0.18 022 0.06-0.93 19 0.29 0.20 0.20 0.07-0.81 Alcohol -cloz/d 1-307Jd missing as 0.2A 0.19 0.19 0.04-121 F-2.is 19 0.29 024 0.17 0.02-1.00 p..is a 0.21 0.08 020 0.0".30 Tobacco smoker 27 nonsmoker missing 2 TOTAL 113 0.29 0.2S 0.21 o.og6.1.21 Fol.00 om 0.18 0.18 0.02-1.00 p-.32 0.14 0.08 0.14 0.09-0.30 #univahate Anova 276 TABLE A4.1.18FREE TESTOSTERONE To PROLACTIN RATIO (TF/P) By BODY MASS INDEY, AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOUTE PLANT, COTTAGE GROVE, MINNESOTA - TF/P N MEAN so MEDIAN RANGE TEST# BMI -c25 40 25-30, 56 ,*W 17 AGE w.31 20 31-40 As 41-50 26 51-W 19 Alcohol -clc*Jd 86 1-3oVd 19 missing 8 Tobacco smoker 27 nonsmoker 84 missing 2 TOTAL 113 #univariatAenova 2-5 1.47 2.2 0.6-7.8 F..lg 2-3 2.13 1.9 0.5-15.0 p-.&3 2-1 1.12 2-0 0.&4.6 2.A 1.67 zo 0.7-7.9 F-.72 2.6 2.27 zi O.S-lS.O p-.54 zi 1.05 LO 0.7-42 2.0 1.19 1.6 0.8-5.6 2.4 1." 2.0 0.6-iS.O F..ig 2.0 120 1-5 0.5-5.1 p-.37 2.6 o.75 2.7 1.5-3.8 32 2.31 Z4 F=9.58 2-1 1.18 1.9 0.5-7.8 p-.OM 2-2 2-20 2.2 0.7-3.8 277 TABLE A4.1.19BOUND TESTOST-EFIONTEo PROLACTIN RATIOCTB/p) BY BODY MASS INDEX. AGE, SMOKING AND DRINKING STATUS 1990 PERFLUOFROCHEMICAL EFFECTS STUDY, 3M CHEMOUTE PLANT, COTTAGE GROVE. MINNESOTA TB/P N MEAN so MEDIAN FIANGE TEST# BMI -c2S 40 25-30 S6 3-30 17 AGE .c3l 20 3140 48 41-50 26 51-M 19 Alcohol -cloz/d 86 1-30vd 19 missing a Tobacco smoker 27 nonsmoker 84 missing 2 TOTAL 113 87 46.4 92 20-221 F-.60 87 95.0 22-624 P-M 68 3LI 67 26-158 75 47.1 sa 20-206 Fe-93 96 91.0 79 22-624 P-.48 78 41.6 72 23-163 73 34.4 63 34-158 87 ric 72 20-624 F.i.23 68 49.9 49 22-221 P..27 95 2U 100 55-117 121 122.0 27-624 F-i1.31 73 38.7 so 22-206 P..001 so 5&8 w 20-100 #univariatAenova 278 TABLT A4.1.20ESTRADIOL T 0 PROLACTIN RATIO (E/P) BY BODY MASS INDEX AGE, SMOKING ANQ DRINKING STATUS 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLRTE PLANT, COTTAGE GROVE, MINNESOTA EIP N MEAN SD MEDIAN RANGE TEST# ami c25 40 25-30 56 3-M 17 AGE -al 20 31-0 48 41-50 26 51-W 10 Alcohol -tioz/d 1-3oVd missing 8 Tobacco Smoker 27 nonsmoker 94 missing 2 TOTAL 113 4.7 2.90 4.4 1 Fo.19 5.1 4.88 3.9 1.1-32-5 P-.92 S.i 2-U 4J I.M.7 42 220 4.1 1.1-0.0 Foi.oo 5.7 5.18 4.6 1.1-3" p.m 4.7 3= 4.0 1.1-15.1 4J 2.00 3.7 =-Qg 5.0 A.1 1 42 L94 Gi 4.15 4.1 1.1-22-5 pose 3.1 1.1-13.0 pe.45 4.6 3.0-15.1 7.2 5.67 5.2 1.1-= F.1221 A.3 zis 4.0 1.1-10.2 P.L.001 4.6 4." 4.6 1.1-9.0 #univahateAnova 279 TABLE A4.2.3 HORMONE RATIOS BY TOTAL SERUM FLUORIDE: ESTRADIOL/PROLACTIN (EIP) THYROID STIMULATING MORMONEIPROLACTIN (TSI-VP) FOLLICLLE STIMULATING HORMONE/PROLACTIN (FSHJP) PROLACTINILUTENIZING HORMONE (PA-H) 1990 PERFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOUTE PLANT, COTTAGE GROVE, MINNESOTA N TOTAL FLUORIDE PPM -Cl 23 ioal-3 64 *3-10 is 3*lc@-IS 6 o-15-26 s TOTAL 113 -Cl 23 ipal-3 64 3%@ic is 2-10-1s 6 a-is-26 5 TOTAL 1'13 Ici 23 3,=1-3 64 3w@l 0 Is 2-IC@-15 6 P,15-26 5 TOTAL 113 -ti 23 s@al-3 63 3p3-10 is a-It@-15 6 2-IS-25 s TOTAL. 112 #univanate Anova MEAN 5.34 4.75 S.97 3.13 S.SS 4.97 0.22 0.24 0.29 0-24 029 0.24 0.65 0.80 0.97 os2 0.66 0.76 1.71 i.$S 1.79 3.14 is 1.97 so EIP 2.77 4.U 4.65 1.06 1.60 3.94 TSHIP 0.11 021 0.76 0.14 0.09 0.20 FSHV'P OA7 OA2 a." a.= 0.36 OA2 paii 0.65 1.?S 1.20 3.00 a." 129 MEDIAN A.&I 3.77 4.76 3.45 9.17 4.04 021 0.17 0.20 0.25 0.27 0.10 0.60 0." OJT CA7 0.47 0.60 1.64 1.37 1.45 1.86 1.33 1.62 RANGE TEST# IAS-10-22 1.09-32.50 AA7.17" 1.13-4.07 3.11-7.09 1.09-32-W Fm." pm.63 0.07-O.AO 0.04-1.20 0.07-0.BS 0.02-0.41 0.20-4." 0.02-1.20 Fw.40 P-.21 0.18-254 F"9 O.IS-V-ll p@A7 0.15-2A$ 0.13-1.04 C-'A3-1.13 0.13-2-M 0-*Lrv-3.02 0.41-4.= 0.39-4.00 1.10-9.11 1.03-2.10 0-3s-o.ll Fmi.72 284 TABLE A4.2.4 HORMONE RATIOS i3yTOTAL SERUM FLUORIC)E: ESTRADIOLILUTENIZING KORMONE (EA-H) ESTRADIOL/FOLLICLLE STIMULA*MNG HORMONE(E/FSH) FOLLICLLE STIMULATING HORMONE/LUTENIZING HORMONE (FSHILH) 1990 PERFLUOROCHEMICAL EFFECTS STUDY. 3M CHEMOUTE PLANT, COTTAGE GROVE, MINNESOTA N TOTAL FLUORIDE PPM ,41 22 2oul-3 63 Is 2010-15 6 *lS-26 5 TOTAL 112 23 *3-10 Is :*10-15 6 ,volS-26 5 TOTAL 113 -el ioul-3 *3-10 2,10-15 *,15-26 TOTAL 23 63 is 6 5 112 #univariaAtneova MEAN a." 6.85 724 721 ?." 7.34 1027 7.71 7.74 8.04 10.32 a.37 0.91 1.04 0.99 1.00 0.91 1.00 so EILM 4.73 4.01 2.74 zis 2-71 3.91 EIFSH GAS 5.90 326 4.24 6.60 6.41 FSHVLH 0.30 CA4 0.43 0.50 om OAI MEOTAN 2.28 6.19 7.SS 7.09 7.c3 7.01 9.62 6.10 6.89 827 7.03 S." 0.89 0.94.0.91 1.03 0-" a." RANGE TEST# C.9S-20.59 1.96-11m 4.39-IO.V 6.03-12.70 OAK-2o-s9 P-.45 1.36-33-12 1-110-29.60 3.71-IL42 3.WlS.30 S.AS-2i.SO 1-30-33.12 F.I.00 P-041 0.42-1.48 *.37-2.2$ 0.41-1.95 C-S3-1.78 A-CS-IAO 0.37-2.25 F..4i P..,73 285 TABLE A4.2.5 HORMONE FLAIIOS 13YTOTAL SERUM FLUORIDF-: FREE TESTOSTERONEfniyRolo STIMULATING HORMONE (TF/TSH) BOUND TESTOSTRONE/THYROID STIMULATING HORMONE OSrrSH) FREE TESTOSTERONEA.UTENIZING HORMONE (TFA-H) BOUND TESTOSTERONs./LLrrENIZING HORMONE (TSA-H) 1990 PERFLUOROCHEMICAL EFFECTS STuoy - 3M CHEMOUTE PLANT. COTTAGE GROVE. MINNEIOTA N TOTAL FLUORIDE PPM -el 23 a-al-3 *4.10 2,10-13 a@15-26 5 TOTAL 112 oti 23 a-ul-3 w3-10 :Olt)-15 6 a-15-26 5 TOTAL 113 .41 23 pul-3 64 *3-10 Is 31-It@-15 6 3015-26 5 TOTAL 113 4ti 23 a-mi-3 64 b3-lo is S,10-15 6 0,15-26. 5 TOTAL 113 #untvariateAnova MEAN 12.6 12-7 11.9 cz 71 12-1 456 479 416 334 314 451 &7 3.S 3.3 3.1 3.0 2A 127 121 115 lie 127 122.1 so MEDIAN RAN(3E TF_ST# TFITSM ILS ?.S 1.7 1.9 7.S TSITSM 330 363 270 296 49 323 TF/LH 2.3 1.2 1.0 1.33 04 1.7 TBfLm so 51 sa 21 67.3 9s 11.1 ICA 7.S 7.9 2.9 320 370 401 226 317 3S3 2.3 3.2 3.3 2-89 &4 3.2 12S 114 ics 105 12S lit 3.2-V2 17-C@-m7 4.&9.6 1.7-43.7 170-1367 51-2102 9s-ilas 57-900 247-370 51-2102 1.2-11.3 C.".l 12-" 1.4." li@-LO 0.6-11.3 31)-2N 24-286 52-234 61-al 122-149 24-3-"$ Fen.93 P-o.45 FmZ4 pa.70 Fu2S pmse Fm.13 pmm,93 286 TA13LE A4.2.6HORMONE RATIOS BY TOTAL SERUM FLUORIDE: THYROID STIMULATING HORMONEIFOLLICLE STIMULATING HORMONE (TSFi/F THYROID STIMULATING HORMONE/LUTENIZING HORMONE (TSKILH) 1990 PEFIFLUOROCHEMICAL EFFECTS STUDY, 3M CHEMOLRTE PLANT, COTTAGE GROVE. MINNESOTA N TOTAL FLUORIDE -clPPM 23 2-al-3 64 2.@ic Is 3.10-15 6 *15-25 5 TOTAL 113 ,cl *=1-3 so@ic M-10-15 1-15-25 TOTAL 23 64 113 .91 23 3,ul-3 P3-10 is 3.10-15 6 a-15-26 5 TOTAL 113 #univariateAnova MEAN 0.42 0.40 a." 0.49 0.49 0.42 CL." cis On 0.45 0.40 0.37 4.40 3.77 &78 &42 3J3 SD TSWFSH 0.24 0.38 ox am CL17 am TSKILH om 0= U? CLI$ CL37 OL26 TF/FSH 227 2.U 1." 1.93 zis MEDIAN 0.40 029 ex 0.49 0.45 0.35 0.33 028 oil 0.43 0.40 oil &7 3.1 3w2 3.9 &g RANGE TEST# 0.08-0.89 0.06-2.U 0.0&120 O.IMM 027-0.U O.Oir@,-2.34 F-OM P..,.92 O.Wi.00 F-OM p..22 0.22-0.70 0-1".45 0.04-1.70 I-C.-IS.6 .7-11.1 1.7-7J J-" li-Ls F-.34 P-45 287 TABLE 4.1.7aUNEAR MULTIVARIATE REGRESSION MODEL OF FACTORS PREDICTING THE HEMAGLOSIN AMONG 111 MALE WORKERS. 3M CHEMOLITE PLANT. COTTAGE GROVE, MINNESOTA Varlabi* 0 SE(S) --P!Valuo lnterr.opt TotalFluorin(eppm)* 14.51 -.o02 .67 .0009 .0001 .02 Alcohol# low (<Icz/day) .22 .20 .27 nonresponse(NR) .56 .33 .09 Age (years) .001 .009 as SMI (kg/M2) .01 .02 .65 Cigarettestday .01 .007 .20 Cigs/dayX Fluodne2** .0003 .0001 .0005 Estradio(lpgtml) .01 .006 .07 R2.M Isquamtranstom3oobfntoW IkioM@ #Refororceategmismderatedtinkewrhsoconv=w 1-3czothanouday. lmenictitoenrmbetweecnigarettpesrdaywW swam tmnsbnr6Mionftotaflluarkie