Document ybkm71GeQkEGn7J7eGN7x5vYV

CorporatOeccupationMaeldicine 3M CenterB,uildin2g20-3W-05 St-PaulM,N 55144-1000 651 7374230-,elephone 651 7339066Fax 8. PlasmaCholecystokinaind HepaticEnzymes,Cholesteroalnd Lipoproteinisn Ammonium PerfluorooctanoaPtreoductioWnorkers. Toxicologryesearcshuggestedt,oa limiteedxtentt,hatheincidencoefpancreaascinar celladenomas inratsfed perfluorooctanoiaccidmay be the resultof a mild but sustained increaseincholecystokini(nCCK) asa consequence of hepaticcholestasis.To assess thishypothesis,plasma CCK levelswere measured by radioimmunoassay forthose Cottage Grove employees participatinignthe 1997 fluorochemicalsurveillance examinations(n = 84). The mean serum PFOA level,as measured by high performance liquidchromatography mass spectrometrymethods,was 6.8 ppm (median 1.3ppm, range 0.1 - 81.3 ppm). CCK valuesapproximated theassay'sreferencerange fora 12 hour fastand were negatively,not positivelya,ssociatedwith employees'serum PFOA levels. Inadditiontothe CCK analyses,medical surveillanceexaminationswere reviewed for the 1993, 1995 and 1997 time periodsto assessthe initiahlypothesisgeneratedfrom the 1990 fluorochemicalmedical surveillancdeatafrom thisCottageGrove workforce that PFOA (measured astotalorcanicfluorine)may modulate hepaticresponsestoobesity and alcoholconsumption (seestudies9 5 and 97). Inthesethreesubsequenttime periods,PFOA, assayed by mass spectrometry,did not appearto modulate hepatic responsesto eitherobesityor alcoholconsumption. Regardlessof the surveillancyeear, therewas no indicatioonf sic@gnificanctlinicahlepatictoxicityatthe PFOA levels observed inthisworkforce.The studyprotocolforthe assessmentofcholecystokinin levelst,he internal3M FinalReport and a manuscriptthathas been acceptedfor publicatio(nDrug and Chemical Toxicology)aresubmitted.[Note: thePFOA category levelsfortheunivariataenalysesinthepublicatiomnanuscript(0 -< I ppm, I - < 10 ppm, > 10 ppm) aredifferentthanthosereportedin the3M finalreport(0 - < I ppm, I < 10 ppm, 10 - < 30 ppm, > 30 ppm). This change inthe publicationmanuscriptwas done per a requestmade duringthepeer review process.This differenceincategoriesdid not affecttheresultsor conclusions.] 16 p-,-;es2s0001 Drug and Chemical Toxicology] PLASMA AND CHOLECYSTOKININ AND HEPATIC ENZYi*vIES, CHOLESTEROL LIPOPROTEINS l@NA.NINIONIUNI PERFLUOROOCTA.NOATE PRODUCTION %VORKERS Geziry'A'.Olsen*, Jean NI.Burris.,'vllcheNlie.Buriew, JeffreyH..'vlandel 'vledicalDepartment, 3.@vCiompany, 220-3W-05, St.Paul,'vIN 55 144- 1000 ABSTRACT Ammonium perf7uorooctanoate is a potent syntheticsurfactantused in industrial applications. It rapidly dissociatesin biologic media to per,fluorooctanoate [CF3(CF2)6CO2'1, which is the anion of perfluorooctanoic acid [PFOA, CFj(CF2)6COOH]. PFOA isa peroxisomeproliferatokrnown toincreasethe incidence of hepatic,pancreas and Leydig celladenomas in rats.The pancreas acinar cell adenomas may be the consequence ofa mildbut sustainedincreaseofcholecystokinians a resultof hepatic cholestasis.Although no significanctlinicalhepatic toxicitywas observed, PFOA was reported to have modulated hepaticresponses to obesityand alcoholconsumption among productionworkers. To furtherassessthesehypotheses,we examined medical surveillance data of male workers involved in ammonium perfluoroocranoateproduction in 1993 (n = 111), 1995 (n = 80) and 1997 (n = 74). Serum PFOA was measured by high-performance liquid chromatography mass spectrometry methods. Plasma cholecystokiniwnas measured (onlyin 1997) by the use of directradioimmunoassay. Serum biochemical testsincluded hepatic enzymes, . cholesteroland lipoprotenis. Serum PFOA levels,by year,were: 1993 (mean 5.0 ppm, SD 12.2,median ].Ippm, range 0.0 - 80.0 ppm), 1995 (mean 6.8 ppm, SD 16.0,median 1.2ppm, range 0.0 - 114.1 ppm); and 1997 (mean 6.4 ppm, SD 14.3,median 1.3ppm, range 0.1 - 81.3 ppm). CCK values(mean 28.5 pglml, SD 17.1, median 22.7 pglml, range 8.8-86.7pglml)approximated theassay'sreferencerange (up to80pglml)for a 12 hour fast and were negatively,not positivelya,ssociatedwith employees'serum PFOA levels.Ourfindings continue tosuggestthereisno significanctlinicahlepatictoxicity associated with PFOA levelsas measured in thisworkforce. Unlike a previously reporfed observation,PFOA did not appear to modulate hepaticresponses to either obesityor alcohol consumption. Limitationsof thesefindingsinclude:1) the crosssectionaldesign as only 17 subjectswere common for the three surveillanceyears; 2) the voluntaryparticipationthat ranged betw-een50 and 70 percent;and 3) the few sltbjecrwsith serum levels> 10 ppm. *Correspondin0a Author INTRODUCTION Ammonium perfluorooctano[aAtPeFO; CF3(CF-,)6CO-,-NH4'i]s a potent synthetic surfactantused in industriaalpplicationwshich rapidlydissociateisnbiolo2icmedia to perfluorooctanoat[eCf@l(CF,)6COl'w]h,ich is the anion of perfluorooctanoiaccid [PFOA, CF3(CF,)6COOH]. In laboratoraynimals,PFOA and itssaltsare:1)absorbed by in-estioni,nhalatioonr dermal contact;1'--3)d)istributepdrimarilyin the liverand 4 blood, 3) not biotransfortnecdo,njugated,incorporateidntolipidsor form coenzyme A conjugates;'-a'nd 4) eliminatedin thefemale ratata createrateof renalexcretionthan the male ratalthoughno gender differenceisnexcretionof PFOA have been seeninother laboratoryanimal species1.,4,9 In rats,administratioonf APFO resultsin peroxisome proliferationu,ncouplingof mitochondrialoxidativephosphorylatioannd alteredlipid metabolism."""' In a 90-day cravagestudyof rhesusmonkeys, allanimalsin the 100 m-lk-/day -roup and '@ofthe 4 animalsinthe30 mg/k--c-,/ady croup diedbeforetheend of studv.1,12 Histopatholooiecxaminationrevealedmarked diffuselipiddepletionin the adrenals,sli-htto moderate hypocellularitoyf bone marrow, moderate atrophyof lymphoid follicleisn thespleenand moderate atrophyof the lymphoid follicleosf the lymph nodes inthetwo highesttreatmentgroups. There were no histopathologicchanges in the 0, 3 and 10 mu kc,day dose oroups. In lifetimefeedin-bioassaysof rats1,3,14 APFO in thedietat ')00ppm (dailydose of 15 mc,/kc,/dayi)ncreasedthe incidenceof liverL.eydig celland pancreasacinarcelladenomas. The liverand testiculatrumors most likeivoccur via nongenotoxicmechanisms: oxidativestressand apoptosisin the development of thelivertumors;and enhanced hepaticaromatase activitwyhich results ina hormone-mediated mechanism (increasedestradiolf)orthe formationof Leydig cell 1-5-17 tumors. The pancreasacinarcelladenomas were hypothesizedto be a resultof a mild but sustainedincreasein cholecystokini(nCCK) levels secondary to hepatic cholestasis.1C8CK has been shown in animal models toproduce pancreatichypertrophy, hyperplasiaand neoplasia.91-23 Hepatic toxicityand hypolipidemiahave not been observed in ammonium perfluorooctanoatperoductionworkers.21.2' Gillilandand Mandel did reportthatPFOA may negativelymodulate the effectalcoholhas on hicah-densitlyipoprotein(HDL) levels and exacerbate the effectthatobesityhas on hepaticenzyme teStS.2,5 This workforce was not found to be atan increasedmortalityriskfor livercancer or liverdisease.25However, therewere four pancreaticcancer deathscompared to two expected deaths (Standardized MortalityRatio 2.0,95% Confidence Interval0.5-5.0).One of these four pancreatic cancer deaths had worked in the buildingwhere ammonium produced. perfluorooctanoatewas The purpose of this analysiswas to examine severaladditionalyears of medical surveillancedata at thisammonium perfluorooctanoatperoduction plant in order to deter-mine:1) whether CCK levelsare positivelyassociatedwith serum PFOA levels among, productionemployees; 2) whether PFOA resultsin clinicalhepatictoxicity;and 3) whether PFOA may modulate hepaticresponsestoobesityand alcohol. PFOA Production METHODS Ammonium perfluorooctanoateproduction at this3M plantbegan in 1947. Ammonium perfluorooctanoatea, white powder, isproduced via a five-staceprocess:electrochemical fluorinationi;solati@Dngand convertint:glthe chemical to a saltslurry;convertingthe slurry 'Loa saltcake; drying thecake; and packaging. The greatestlikelihoodforexposure has occprred in the drying area. Subject Selectionand Data Collection Voluntary medical surveillanceexaminations were offered to the fluorochemical production workers in 1993, 1995 and 1997. The totalnumber of male subjects,by year, who participateidnthesethreecross-sectionailnvesti-ationwsere: 1993 (n = I11);1995 (n = 80);and 1997 (n = 74). (There were too few female employees to includein thedata analysis.) Ellc,,ibvloeluntaryparticipatiornatesamong these productionworkers ranged from approximately50 (1997) to 70 (1993) percent.There were 68 subjectsin common for 1993 and 1995; 21 subjectsincommon between 1995 and 1997 (lo@k-enrumber due to employee turnoverand re-assc@ionments);and 17 subjectsin common forallthree ears. y Surveillanceactivitieisncludeda self-administereqduestionnaire,measurement of heicht, 1-1 wei-ht and pulmonary function,standardbiochemical and urinalysitsests,serum PFOA deter-minationand severalmale reproductivehormone assays. The hormone data were collectedin 1993 and 1995 and the findingshave been reportedelsewhere.27 Serum hepaticand lipoprotein-relatebdiochemicaltestsincluded:alkalinephosphatase (ITJ/L); gamma glutamyl transferase(GGT, rU/L); aspartateaminotransferase(AST, ITJ/L); alanineaminotransferas(eALT, IU/L);totalbilirubin(me./dl)d;irectbilirubi(nmg/dl); total cholesterol(mg/dl); low-densitylipoproteins(LDL, m,-/di);high-density lipoproteins(HDL, mg/dl);and triclyceride(smc,/dl).In 1997,emplovees@ plasma CCK33 (p_z/mi)levelswere determined. CCK existsin variousforms and len-ths,althoucrh 41sulfatedCCK-33 (i.e.a,33 amino-acid arrancrement)appears to be the predominant form. For purposes of brevity,we willreferto CCK-33 as CCK. Employees were requiredto have fastedfor 12 hours priorto theirvenipuncture.Because CCK analyseswere not a standardanalysisof the company's fluorochemicalmedical surveillanceprogram, a study protocolwas reviewed and approved by the company's human subjectscommittee and a sit@gned,informed consent was obtainedfrom each participant. Senim chemistriesand hematology were evaluatedat AllinaLaboratories(Minneapolis, Minnesota). Plasma CCK was measured by directradioimmunoassay by InterScience Institut(eInglewood,California).Serum PFOA (i.e.p,er-fluorooctanoatwea)s determined f by thermospray (199') and 1995) and electrospray(1997) hi-h-performance liquid chromatography/mass spectrometrymethods.28,29 Data Analysis Simple and stratifiedanalyses, analysis of variance (ANOVA), and multivariabie recg@ ressiontechniques were used to evaluatelinearand nonlinearassociationsbetween PFOA and the biochemical parameters with adjustment for potentialconfounding variables3.0 Various serum category levelswere used forthe stratifieadnalysiswith no significandtifferencesobserved based on cutpointsas high as > 30 ppm. However, the number of employees at> 30 ppm was 5 or fewer (basedon year). For purposes of this report,employees were stratifieidntothreePFOA cateeories(0 - <1 ppm; I - <10 ppm: and 10 ppm) in orderto providea oreaternumber of emplovees in the hiahestL> 10 ppm) category. For multivariablere-ressionanalyses,PFOA, ace, body mass index (BMI), alcoholuse, and cigaretteuse were examined as both categoricaland continuous variables.Alcohol use was analyzedas lessthan I drinkper da,,,>, I drink per day and non-response to thisquestionnaireitem (almost allsubjectsreportedbetween <1 - 3 drinks/day). Linear and nonlineartransformationsof PFOA were used to testfor associations.In particulart,hemultivariablmeodels employed by Gillilandand Mandel2' were used to determine whether PFOA has a modulating effecton obesityor alcohol consumption in reaardsto hepaticserum chemistries(ALT, AST and GGT) and HDL, respectively. RESULTS Serum PFOA levels,by year,were: 1993 (mean 5.0 ppm, SD 12.2,median 1.1 ppm, range 0.0 - 80.0 ppm); 1995 (mean 6.8ppm, SD 16.0,median 1.2ppm, ranae 0.0 - 114.1 ppm); and 1997 (mean 6.4 ppm, SD 14.3,median 1.3ppm, range 0.1 81.3 ppm). Pi-o'@.,idiendTable I are the mean, standard deviation,median and rancreof the employees' age, BMI and hepatic,cholesteroland lipoproteinserum chemistry data stratifiebdy PFOA leveland the year of the medical surveillancexamination.Depending upon the surveillanceyear,one to two ordersof maanitude of differencewere observed between the means (and medians) of the lowest and highestserum PFOA categories. There was no evidence for abnormal liverfunctiontests,hypolipidemiaor cholestasisassociated with increasingemployees' serum PFOA levels.Controllin-for potentialconfounders, multivariableregressionanalysesdid not suggest otherwise.Other measures including renalfunction,blood glucose and hematology were not associatedwith serum PFOA levels(datanot shown). The mean CCK value was 50 percentlower amon-c employees with serum PFOA values 2! 10 ppm than forthoseemployees with serum PFOA levels< I ppm (Table1), Figure I isa scatteprlotofthenaturallogofCCK andPFOA. Allbuttwo CCK valueswere withintheassay'sreferencerange(up to 80 pg/mi). These two CCK values(80.5pc,/ml and 86.7 pp/ml) were of employees with 0.6 ppm and 5.6 ppm serum PFOA levels, respectivelyA.djustinaforpotentiaclonfounding variablesw,e continuedto observe a negativeassociatiobnetween thenaturallog of CCK and serum PFOA levels(TableII) althoughminimum variatiownas explained(R2 =.08). Providedin Table IIIarethe multivariablree-ressionmodels (asoricrinalrleyportedby Gillilandand Mandel with the 1990 medical surveillancdeata25)which exarninedthe potentiamlodulatingeffectof PFOA on the associatiobnetween alcoholand HDL. The coefficientosf determination(R2,adjustedR 2) were not largefor any model. Based on thesei-nodelsT,able rV shows the change in HDL levelsassociatedwith a 10 ppm increasein serum PFOA levelsamong liahtand moderate drinkers@@ I drink/day) compared to lilyhdtrinkers(< I drink/day)[.Note:totalserum organicfluorine measurements,ratherthan serum PFOA, were used in 1990.] Unlike thefindingsfrom the 1990 data,theeffectof alcoholuse on increasingHDL levelswas not bluntedby a 10 ppm increasein PFOA in any of the subsequentmedical surveillanceexamination years. Presentedin Table V are multivariablreegressionmodels, includinct,hoseoriginally reportedusing the 1990 surveillancedata2,5 regardingthe potentialeffectof PFOA on hepaticenzyme responsesto obesityas measured by BMI. Again, the coefficientosf determination were not largefor any model. In the 1990 surveillancdeata,ALT increasedamong obese (BMI = 35 kg/m 2) but not non-obese(BMI = 25 kg/M2) workers t who had a 10 ppm chanct:el in serum PFOA (TableVI). However, thisinteractiownas not observedinthe 1993, 1995 or 1997 medical surveillanceexaminations.Likewise,we did notobserve associationwsith AST or GGT (datanot shown) as was reportedinthe 1990 medical sur@,,eillanecxeaminations.25 DISCUSSION We observead negativaessociatiboentweenserumPFOA andplasmaCCK among 74 workers engaged intheproductionof ammonium perfluorooctanoate.This observation isoppositethatproposed by Obourn et a] who questionedwhether chronicexposure to peroxisome proliferatorlsi,kePFOA, can cause pancreaticadenomas in the rat as the consequence of a mild but sustainedincreasein CCK levelssecondary to hepatic cholestasis17. We do notbelievethenegativeassociationobserved inour studyrepresents an entireldyifferenbtiologicarlelationshitphanwhat was originallpyostulatedbecause: 1)allCCK valuesobservedinthisstudywere withinthe assay'sreferenceexceptfortwo values(which were not associatedwith hi-h serum PFOA values);and 2) therewas no suggestionof cholestasiwshich was consideredthe underlyingreason for the elevated CCK levelsintherat. There areseveralexplanationsfor the lackof a positiveassociationbetween PFOA and CCK in our study. Firstand foremost,the serum PFOA measurements in these productionworkers may have been too low to cause an increasein CCK if such a mechanism existsin humans. Second, the mechanisticreason for the elevatedCCK levelsintheObourn etalstudv'7was not clearlyestablished.Obourn etalexamined the effectsof Wyeth 14,643,a more potentperoxisome proliferatotrhan PFOA: thus their findicn:-smay notbe directlryelatedto PFOA. The clinicaplatholo,cyrdataindicativoef cholestasiisncludedalterationisnbileflow and bileacidoutput.Absolute bileflow and flow relative-to-bowdeyight were marginallyincreasedand acinarcellproliferation, alt9oughnumericallyincreasedat 3 months, returnedto controllevelsat 6 months. Obourn et al alsoconducted in vitroexperimentsof both Wyeth-14,643 and PFOA which argued againstotherbiologicaplathways known to elevateplasma CCK levels includingCCK-A receptora-onism, trypsininhibitioannd increaseddietaryfatcontent. Third, theprimary setofbiochemicaland cellulaerventsidentifieidnrodentssusceptible to the hepatictumor effectsof peroxisome proliferatohrass not been identifieidn either liverbiopsiesfrom humans exposed toperox*lsomeproliferatoorrsininvitrostudieswith human hepatocytes;however, theperoxisome proliferator-activarteecdeptor(PPAR-a) is expressedat very low levelsinthe human liver1.7,31 Itshould also be noted thatunlike the pancreas of the rat.the human pancreas has no detectableCCK-A receptorsand little tono MRNA forthereceptor."-"Finallyw,hetherCCK initiateosr promotespancreatic cancer isa controversiailssue3.5 Data from more thanseventylaboratoryanimal studies have variablysuacestedthatCCK has positivetrophiceffectsi,nhibitoryeffectso,r no involvement in pancreatictumor growth.36 CCK has promoted growth of human pancreaticcancersincellculture.s37 On the otherhand,fastincp,lasma concentrationosf CCK inunresectedpancreaticancerpatientdsid not differfrom healthycontrols38.Also acinarcellmali0gnanciesin ratsarerarein thehuman.39 Activationof the c-K-rasgC)ene is i frequentin both human and hamster pancreaticancer but is not found in azaserineinduced pancreaticcancer acinarcelladenoma models intherat4.0,41 NeitherGillilandand Mandel2.i nor ourselvesobserved si-nificanctlinicalhepatic toxicityassociatedwith theserum PFOA levelsmeasured inthisworkforce. In the three surveillancyeearswe examined (1993, 1995 and 1997),88 percent,81 percentand 85 percentof thoseemployees who volunteeredhad serum PFOA levelslessthan 10 ppm, respectively.In a laboratorystudy,no si-nificanhtepatictoxicitywas reportedforthe lowestdose (3 mg/k_-Jday)gIroup of rhesusmonkeys administeredAPFO by t-,aN'za:-l efor 90 days."12 One of four primatesin the next hic-hestdose (10 m(-z/kz@-/d-avc1*r1@oup) developed anorexiaand black stoolsdurinc,thecourseof the studv.There were no other abnormalitiesreportedfor thisccroup. Only one animal survivedin the30 m c_,kJc,/day group and none survivedin thehighest(100 mg/ka/day)dose croup. For the'3and 10 mar/kc,/dadyose -roups,theirmean serum totalorganicfluorinelevelswere 54 and 67 ppm, respectively. Sixty-ninepercentof themolecularweight of PFOA isorganic fluorinet;hereforethesetotalor-anicfluorinelevelsin the lower dose groups may correspondto serum PFOA levelsof 80 to 100 ppm. Totalorcanicfluorinelevelswere analyzed inthe liverfortwo animalsineach of the two lowesttwo dose groups and the means were 5 and 10 ppm, respectivel1y2. Livertotaloraanicfluorinelevels",erealso analyzedforthe30 mg/kg/day (n = 4, mean = 98 ppm) and the 100 mg/kg/day (n = 2, mean = 213 ppm) dose aroups;however,only the sole survivinganimal in the 30 m-/k- day croup was analyzedfor serum totalorcanicfluorine(145 ppm which approximate2s10 ppm of PFOA). The 30 mg/kg/daygroupdidhavehigherhepatic transaminasvealuesafte3r0 daysofcompound administrattihoannthecontrolcroup. Serum chemistrieswere not performed for the 100 m,-,/k,-,/gdraoyup afterthe onsetof compound administrationR.elative(% body weight)liverweightswere higheramong the 30 mg/ka,/day(3.84%) and 100 macAocr/da(y3.31%) treatmentcroups thanthe control (2.36%),3 mg/ka-/day(2.36%) and 10 mg/k-g/day(2.32%) dose groups. No abnormal liverfunctionresultswere observed amona4-- employees with the hiahestserum PFOA t@ levels;neverthelesstheseworkers have been restrictefdrom potentialhi-h exposure workplace areas. We were unable to replicate,inthreesubsequentmedical surveillancexaminations,an earlieirnvestiaationf'isnding thatPFOA may modulate hepaticresponsesto obesityand alcohol.25 Total serum orc-anic fluorinewas used as a surro-atevariablefor PFOA in the 1990 medical surveillanceexams. The use of totalserum orcanicfluorineconstitutes additionalpotentialexposure to perfluorocarbonsh;owever, data suggestthat PFOA would representthe greatestfractionof totalserum organic fluorinelevelsin this employee population2.4 Another explanationforthe disparatefindings,in particulaars relatedto BMI, isthattheremay have been measurement errorregardingBMI in the ori-inalstudv as well as in our investigationW.e have previouslynotedthelack of an expectedpositiveassociatiobnetween BMI and estradioiln the 1990 data2.7 Yet in the presentstudywe did notobservetheanticipatesdtrong,positivecorrelatiobnetween BMI and ALT exceptin 1997. Correlatiocnoefficient(sp valuein parenthesesw)ere: 1993,r .16,(p = .09);1995,r = .13,(p = .27);and 1997,r = .43(p = .0001).Self-reported alcoholdata collectedin theoccupationalsettingcan be questionedfor itsreliabilitays wellas validity.The lattewras notfeasibletoaddress;however, topartialleyxamine the issueofreliabilitwye, examined theanalysesof the 68 employees who participatebdoth in 1993 and 1995. The datashowed good correlatiobnetween thesetwo yearsforthe potentiaclonfoundingfactorsofBMI (r= .94,p = .0001),alcoholconsumption(r= .67,p .0001)and ci-arettsemokinz (r= .84,p = .0001). Severaladditionalissuesareworthy ofconsideration.The medical surveillancperogram isvoluiltar@O,v.erallparticipatiornatesdeclinedfrom approximately70 percentIn 1993 to 50 percentin 1997. Serum PFOA len,elcsould differbetween participantasnd nonparticipantsT.he highturnoverof employees between 1995 and 1997 detractedfrom the opportunityfora longitudinaalssessment.In thisre-ard,we did examine the chanae in severalparametersamong the 68 subjectsin common for 1993 and 1995. For example, theaveragedifferencein serum PFOA was + 0.1 ppm (Wilcoxinsigned-rank test -540.5,p test 50.0,p .0001),themean chance in ALT was +0.5 IU/L (Wilcoxinsl-,ned-rank 0.6)and the mean change in cholesterowlas -1.6mg/dL (Wilcoxin signed-ranktest -60.0,p = 0.4). The change inserum PFOA levelsdid notpredict,via regressionanalyses,the change in ALT or cholesterolB.esides the few employees in common acrossallthreeyearsof medicalsurveillancdeata (n = 17),anotherdifference thatoccurredin 1997 was the method of analysisof PFOA changed from thermospray (1993 and 1995) to electrospra(y1997)high-performanceliquidchromatography mass spectrometry. Also, the laboratoryreferencerange substantiallcyhanoed for ALT in 1997 (ascan be seen inTable I by examiningthelowermean valuesforALT in 1997). Finally,the issueremains thatthe lack of a clinicalhepatotoxiceffectreportedby Gillilandand Mandel25 and ourselvesdoes not ruleout the possibilittyhatPFOA may resultin a subclinicahlepaticeffectin thisproductionpopulationthatwe have yet to observe. Resultsfrom ongoing laboratoryanimal studiesi,ncludinga 6 month APFO gelatincapsulefeedingstudyof cynomolgus primates,may providefurtherinsightinto thedirectioonf medicalsurveillancaectivitiefsorthisworkforce. REFERENCES I GriffitFhD, Long JE. 1980. Animal toxicitsytudieswithammonium perfluorooctanoatAem. IndHyg Assoc J 41:576-583,1980. 2. Kennedy G. Dermal toxicityof ammonium perfluorooctanoate. Toxicol Appl Pharmacol 81:348-355, 1985. 3. Kennedy G, HallG, BrittelJl,iChen H. Inhalatiotnoxicitoyf ammonium perfluorooctanoatFed. Chem Toxicol24:1325-1329,1986. 4. Vanden Heuvel J,KuslikisB, Van Refelchem M, PertersonR. Tissuedistribution, metabolism and eliminationofperfluorooctanoiaccid.J Biochem Toxicol6:83-92, 1991. 5. PastooTrP,LeeKP, PerrMiA, GilliePsJ.Biochemicaalndmorphologicsatludies of ai-nmonium perfluorooetanoate-inducehdepatomecaly and peroxtsome proliferationE.xpMolPatliol47:98-109. 1987. 6. Ophaug R, Sincer L. Metabolic handlinz of perfluoi-ooctanoicacid in rats. Proc Soc Exp Biol Med 163:19-23, 1980. 7. Yllnen M, Koho A. HanhijarviH. PeuraP. Dispositioonfpertluorooctanoiaccidin the rataftersingleand subchronicadministrationB.ullEnviro?zCo?itamToricol 44:46-53,1990. 8. KuslikisBI,Vanden Heuvel JP.PetersonRE. Lack ofevidencefor perfluoroodecanoylo-r perfluorooctano-ycloenzyme A formationinmale and female rats.J Biochem Toxicol 7:25-29,1992. 9. HanhijarviH, Phaug R, SingerL. The sex-relateddifferenceinperfluorooetanoate excretionintherat.Proc Soc Exp BiolMed 171:51-55,1982. 10. Haughom B, Spydevold 0. The mechanism underlyingthe hypolipemiceffectof perfluorooctanoiaccid(PFOA), perfluorooctanseulphonicacid(PFOSA) and clofibriaccid.Biochim BiophysActa 1128:65-72,1992. 11. KellerB, Marsman D, Popp J,Thurman R. Severalnongenotoxiccarcinocens uncouple mitochondrialoxidativephosphorylation.BiochimBiophysActa 1102:237-244,1992. 12. GoidenthalEI,Jessup DC, GeilRG, Mehring JS. Ninety-day Subacute Rhesus Monkey ToxicityStudy. Study No. 137-090.InternationaRlesearch Development Corp, Mattawan (MI), November, 1978. 13. SibinskiIJ. Two-year oral(diett)oxicity/carcinogenicsittuydyoffluorochemical FC- 143 inratsR.ikerLaboratoriesS,t.Paul (MN), 1987. 14. Rao MS, Reddy JK. Hepatocarcinogenesis of peroxisome proliferators.Ann NY Acad Sci 804:573-587, 1997. 15. Cook JC, Murray SM, Frame SR, HurttME. 1992. Inductionof Leydig cell adenomas by ammonium perfluorooctanoatea:possibleendocrine-related mechanism. ToxicolAppIPharmacol 113:209-217,1992. 16. Cook JC, HurttME, Frame SR, BiecelLB. 1994. Mechanisms of extrahepatic tumor inductionby peroxisomeproliferatorisnCrl:CD BR (CD) rats.Toxicologist 14:301,1994. 17. BiegelLB, Liu RCM, HurttME, Cook JC. Effectsof ammonium perfluorooctanoatoen Leydig cellfunction:invitroi,nvivo,and ex vivo studies. ToxicolAppl Pharmacol 134:18-25,1995, 18. Obourn JD, Frame SR, BellRH, Lon-necker DS, ElliotGtS, Cook JC. Mechanisms forthepancreaticoncogenic effectsoftheperoxisome proliferator Wyeth- 14,643. ToxicolAppl Pharmacol 145:425-436,1997. 19. Pandol SJ. Pancreaticphysiologyand secretorytesting.(In)SleisengerM, FordtranJS (eds)Gastrointestinaanld LiverDiseases,Volume 1,W.B. Saunders Co.,Philadelphia,1998, pages771-782. 20. Longnecker DS. Experimentalmodels ofexocrinepancreatitcumors. (In)Go VLW (ed)The ExocrinePancreas:Biology,Pathologyand Diseases,Raven, New York, 1986,pages 443-458. 21. Lonanecker DS. Experimental pancreaticcancer:roleof species,sex and diet.Bull Cancer 77:27-37, 1990. 22. Pour PM, Runge RG. BirtD, GingellR, Lawson T,Nagel D, Walicave L, Salmasi S.Currentknowledge of pancreaticcarcinogenesiisnthehamsterand itsrelevance tothe human disease. Cancer 47:1573-1587,1981. 23. Pour PM, Lawson T, Hel2eson S,Donnelly T, Stepan K. Effectofcholecystokinin on pancreaticcarcino-enesiisnthehamstermodel. Carciizogetiesi9s:597-601, 1988. 24. Ubel F, Sorenson S, Roach D. Health statusofplantworkers exposed to fluorochemicalsA: preliminaryreport.Am Ind H,,,Agssoc 41:584-599,1980. 25. GillilandFD, Mandel JS. Serum perfluorooctanoiaccidand hepaticenzymes, lipoproteinasnd cholesterola:studyofoccupationalleyxposed men. Am J I?zd Med 129:560-568,1996. 26. GillilandFD, Mandel JS. Mortalityamona employees of a perfluorooctanolaecid productionplant.J Occup Med 35:950-954,1993. 27. Olsen GW, GillilanFdD, Buriew MM, BurrisJM, Mandel JS,Mandel JH. An epidemiologicinvestigatioonfreproductivehormones inmen withoccupational exposure toperfluorooctanoiaccid.J Occ Env Med 40:614-622,1998. 28. Johnson JD, Wolter JT, ColalzyGE, RethwillPA, Nelson RM. Quantificatioonf perfluorooctanoataend perfluorooctanesulfonaitnehuman serum usingion-pair extractionand highperformance liquidchromatocraphy-thermospraymass spectrometrywit@ automated sample preparation3.M Compati.v,St.Paitl(M.V), 1996. 29. Anderson DJ, Mulvana DE. Analyticalreportforthe determinationof perfluorooctanaoataend perfluorooetanesulfonaitnehuman serum bNlLC/NIS. Advanced BioatialyticaSlervices,Inc.,Ithaca(NY), 1997. 30, SAS InstituteI,nc. SAS Users Guide: Statistics.Version 6. SAS InstituteI,nc., Cary (NC), 1990. 31. CattleyRC, DeLuca J,Elcombe C, Fenner-CrispP,Lake BG, Marsman DS, PastoorTA, Popp JA, Robinson DE, Schwetz B, Tuawood J,Wahil W. Do peroxisome proliferatincco,mpounds pose a hepatocarcinocrenihcazard to humans? Reg ToxicolPlianizacol27:47-60,1998. 32. Wank SA, Pise(znaJR. deweerth A. Cholecystokinin receptorfamily. Anil NY Acad Sci 713:49-66, 1994, 33. Gavin CE, Martii NP, SchlosserMJ. Absence of specifiCcCK-A bindinz:glsiteson human pancreatic membranes. Toxicologist 30:334, 1996. 34.f Gavin CE, Malnoske JA, White J,SchlosserMJ. Speciesdifferenceisnexpression of pancreaticcholecystokinin-Areceptors.Tocicologis3t6:232 1997 35. Axelson J,Ihse1,Hakanson R. Pancreaticcancer: theroleofcholecystokinin? Scand J Gastroenterol1992;27:993-998.ToxicolAppl Phannacol 134:18-25,1995. 36. HerringtonMK. Adrian TE. On theroleofcholecystokiniinnpancreaticancer. It?tJ Pancreatol 17:121-138,1995. 37. Palmer-SmithJ.Krame ST, Solomon TE. CCK stimulategsrowth ofsixhuman pancreaticcancercelllinesinserum-freemedium. Reg Peptides22:341-349,1991. 38. Rehfeld JF.van Solinc,yWeW. The tumor biolocyof gastrinand cholecystokinin. Adv Cancer Res 63:295-347,1994. 39. Anderson ICE,PotterJD, Mack TM. Pancreaticancer.(In)SchottenfelDd, Fraumeni JF (eds)Cancer Epidemiologyand Prevention(2'dedition).Oxford UniversityPress,New York,1996,pages725-77 1. 40. van Kranen HJ, Vermeulen E, Schoren L, Bas J,Woutersen RA, van IerselP,van KreijlCF, SchererE. Activationofc-K-ras isfrequentinpancreaticcarcinomas of Syrianhamsters,butisabsentinpancreatictumors ofrats.Carcinogenesis12:1471482,1991. 41. Caldas C, Kem SE. K-ras mutation and pancreaticadenocarcinoma. IntJ Pancreatol 18:1-6,1995. Table 1. Mean, Median,StandardDeviation(S.D.)of Mean and Range of Demographic,fiepaticC,holesteroalnd Lipoprotein Categorieasnd MedicalSurveillancYeear (1993,1995,1997) PFOA* Category(ppm) Mean 1993 Median S.D. Range 0-<l 1-<10 > 10 0.48 3.38 30.88 F Value 0.48 2.50 19.50 0.27 2.17 25.12 68.3,p = .0001 0.00-0.99 1.03-8.92 11.90- 80.00 1995 Mean Median S.D. PFOA (ppm) 0.31 3.03 30.06 0.20 2.40 25.50 0.32 1.84 26.58 F Value = 39.1,p = .0001 Range 0.00-0.90 1.10-8.20 10.3-114.1 Mean 0.47 3.13 32.13 F Value 0-<] 1 -<10 > 10 0-<l 1-<10 > 10 43 39 39 F Value 44 9.2 38 7.8 38 6.6 3.7,p =.02 27-61 27-60 25-49 28.0 26.9 28.4 F Value 27.6 4.3 26.3 2.5 28.5 2.4 1.7,p =, 14 20.9-42.0 21.6-32.5 22.4-32.0 Age 42 41 8.3 41 40 8.6 43 45 8.4 F Value= 0.2,p =.85 29-60 24-58 27-55 BMI (kg/M2) 27.6 26.8 4.2 28.6 27.9 3.4 28.4 28.8 3.5 F Value = 0.7,p = .51 21.9-45.2 22.1-38.3 21.2-34.8 40 41 42 F Value 28.7 29.5 27.6 F Value 0-<l 1- <10 > 10 Not done in 1993 CCK (pg/ml) Not done in 1995 33.4 28.0 17.4 F value= 3. PFOA* Category(ppm) Mean 1993 Median S.D. 0-<l 1-<10 > 10 88 82 26 82 78 23 83 75 24 F Value 0.7,p =.52 1995 Range 37- 161 47- 151 58- 132 Mean Median S.D. AlkalinePhosphatase(IU/L) 78 76 18 80 76 25 89 76 31 F Value= 1.4,p = .25 Range 40-114 48- 165 55- 146 0-<l 1 -<10 > 10 33 50 37 F Value 30 19 30 70 33 17 1.5,p = .24 11-84 6-472 19-77 GGT (TU/L) 42 34 27 51 36 41 40 38 13 F Value= 0.9,p =.41 16-149 19-190 21-61 0-<l 1 -<10 > 10 23 22 7 26 24 11 24 24 5 F Value = 1.1,p =.33 11-60 12-83 16-35 AST (IUAL) 21 20 6 24 20 13 21 20 4 F Value= 0.8.p =.45 13-36 13-75 15-29 0-<l 1-<10 > 10 45 42 14 48 45 29 46 47 8 F Value = 0.2,p = .82 22-88 22-221 35-62 ALT (rU/L) 44 40 13 53 40 34 47 48 12 F Value= 1.2,p =.30 27-80 27- 175 28-71 Mean 79 87 80 F Value 34 36 30 F Value 26 25 25 F Value 31 33 35 F Value PFOA* Mean category(ppm) 1993 Median S.D. Range 0-<l 1-<Io 10 138 142 40 143 137 38 140 143 42 F Value 0.2,p =.84 27-227 72-223 60- 188 0-<l 1 -<10 > 10 171 205 221 F Value 145 124 129 408 223 159 0,3,p =.77 37-636 47-2845 41 -564 1995 Mean Median S.D, LDL (mg/di) 131 130 32 133 137 40 130 118 39 F Value = 0.1,p =.96 Range 31- 191 28-210 62-211 Triglyceride(smg/di) 170 175 254 F Value 152 93 123 144 183 154 2.7,p =.07 57-371 59-743 77-563 Mean 114 134 134 F Value 219 176 251 F Value StudyPopulationsizeby PFOA category(Ppm)and year SarnpleSize PFOA Category 1993 1995 1997 0< 1 52 39 I -< 10 46 26 29 T>ot1a0l 13 15 3141 80 74 TableU. MultipleRegressionModel ofFactorsPredictintghe NaturalLog ofPlasma CholecystokinininWorkers with Serum PFOA Levels Variable Intercept PFOA Age Alcohol BNU Ciizarettes B 3.02 -0.009 0.0001 -0.005 0.009 -0.009 SE(B) 0.48 0.004 0.007 0.086 0.015 0.007 p value .0001 .07 .98 .95 .53 .17 R@ =.08 Adj R'=.02 TableIfI. MultivariablReegressionModels-of FactorsPredictinHgigh DensityLipoproteiinnWorkers with Serum PFOA Level 1990 1993 1995 Variable B SE(B@ p value B SE(B) p value B SE(B) p value Intercept PFOA LightDrink Interaction" 65.00 -1.61 -9.92 1 @62 10,07 0.77 3@51 0,80 .0001 .04 006 .04 55.00 -0.14 -4.83 0.02 16,70 '001 0.33 .67 3.76 20 0.35 .96 52.10 -0.10 -5.11 0,02 11.92 0.08 2.61 0.13 .0001 .18 .05 .87 R'=.17 Adj R' = Not reported R'=.10 Adj R' = .02 R' = .30 Adj R 2 19 Adjustedforage,BMI, cigarettuese,and non-respondenttsoalcoholquestion(allfouryears)and testosteronleevel(1990,1993au The 1990 regressiomnodel used totalorganicfluorinaes thedependentvariable(seereference#25). Interacti=onPFOA x LightDrink TableIV.Change inHDL* from LightAlcoholDrinker(< I drink/dayt)o@1) a LightDrinkerwitha 10 ppm Change inSerum PFOA; 2) a Moderate Alcohol Drinker(2!1 drink/day)a;nd 3) a Moderate Alcoholwith a 10 ppm Change in Serum PFOA. LightDrinkerwitha 10 pm increaseinPFOA Moderate Drinker Moderate Drinkerwith 10 ppm increaseinPFOA Year Change inHDL (mz/dl) 19904 0.1 9.9 -6.2 1993 -1.3 4.8 3.4 1995 -0.8 5.1 4.1 1997 1.7 5.8 3,9 *Determined from multivariablmeodels (seetable'3)adjustedforage,body mass indexcigarettuese,and non-respondenttso alcoholquestion(alflouryears)and testosteron(e1990,1993 and 1995 only). Data in1990 analyzedfortotalserum organicfluorin(eseereferenc#e25). TableV. MultivariablReegressionModels'of FactorsPredictinAgLT inWorkers with SeruinPFOA Levels 1990 Intercept 58.13 PFOA -15.80 BMI 0.30 Interaction' 0.62 $E(B) 24.60 4.58 0.82 0.17 R' = .21 AdjR'= NotY"jicd # Adiusted for age, alcohol and 'Interacti=onPFOA x BMI cigarette - p value .02 .0008 .72 .0004 B 27,32 0,89 1.07 -0.03 R' = .06 AdjR=.Ol 1993 SEn 19.13 2.88 0.67 0.10 use. ne 1990 regression model used p value .16 .76 .11 .79 total organic -- B 40.93 0.81 1.08 -0.03 Rl=.il AdiR'=.Ol 1995 SE(B) _E_Kalue 22.90 .08 2.62 .75 0.74 .15 0.09 .76 fluorine as the dependent variable (see r TableVI. Change inAlanineAminotransferas(eALT)* Associatedwitha 10 ppm Change inSerum PFOA forThree Body Mass Indices BNU (kp,/m2) 25 30 35 Year Iggo;, 1993 1995 1997 Change inALT (W -3.0 28.0 2.2 0.9 1.5 0.1 5.3 0.8 59.0 -0.5 -1.2 -3.7 * Determined from multivariablreegressionmodel (seetable5) adjustedforage,alcoholand cigarettuese. 0Data in1990 analyzedtotalserum organicfluorine (seereference#26). Figure 1. ScattePrlotofNauturalLog CCK (pgln-db)y Serum PFOA Level (ppm) (LinearRegressionModel: Ln CCK = 3.24-0.006PFOA; p value of PFOA coefficein=t 0.19;R@ = .021 f 5 4 3 L) 2 0- 0 10 20 30 40 50 60 70 80 90 100 PFOA Level(ppm) 3M Cornpany EPI-OW3 Page I of 42 PROTOCOL Epidemiology Medical Department 3M Company 220-3W-05 St.Paul, MN 55144 Date: September 3, 1997 Tide: An EpideniiologicInvestigatioonf Plasma Cholecystokininand Hepatic Function in PerfluoroocatanoicAcid ProductionWorkers S tudy StartDate: September 3, 1997 Estimated Date of FinalReport: April15,1998 IP,B Approval Date: ProtocolNumber: EPI-0003 IRB Approval Exempt Expedited x PrincipalInvestigator: Co-investigators: Geary W. Olsen,DVM, PhD' Jean Burris,RN, NTH' NficheleM. Burlew, MS' JeffreyH. Mandel, NM, NTH' Study Director: JeffreyH. Mandel, MD, NDH 1. OccupationalMedicine,3M Company, 220-3W-05, St.Paul, MN 55115 3M Company EPI-OW3 Page2 of42 ABSTRACT Two-year feedingstudiesin CT1:CD BR (CD) ratsata maximum amount of 300 ppm perfluorooctanoiaccid (PFOA) showed, in additiontoliveradenomas and Leydig cell adenomas, an increasedincidence of pancreas acinarcelladenomas. However, PFOA was not found to be mutagenic; thusthe induction of thesetumors likelyoccurs Nia nongenotoxic mechanisms. Recent researchhas suggested thatthepancreas adenomas are a secondary effectfrom elevatedcholecystokinin(CCK) levelsdue to hepaticcholestasis. Increasedplasma CCK levelshas produced pancreatichypertrophy,hyperplasiaand neoplasiain some, but not all,animal models. A cynomologus monkey iscurrentlybeing used in a PFOA feeding study tofurtherassess therelationwith CCK because the rat pancreas CCK receptormay be quitedissimilartothatof the human and monkey. Because the roleof CCK in thepromotion of pancreaticcancerremains quite controversialin laboratoryanimals,additionalinsightiswarranted intoany possible associationbetween occupationalexposure to PFOA and plasma CCK levels.Therefore, the purpose of thiscrosssectionalepidemiologicalstudydesign istodetermine whether thereisan associationbetween plasma CCK-33 levels,as measured by tadioimmunoassay, and serum PFOA levelsamong 3M Cottage Grove fluorochemical productionworkers.We willalsoexamine hepaticenzymes, bihmbin and lipoproteinisn relationto the employees' serum PFOA levelsdue to the toxicologicailssueregarding hepatic cholestasis. U,;TRODUCTION 3M Company EPI-"3 Page3of42 Fluorocarbons are compounds of fluorine,carbon and otherelements such as oxygen, nitrogenand sulfur.Perfluorocarbonsarestructurallaynalogous to hydrocarbons, except thehydrogens arereplacedby fluorine[Bryce,19641. In general, perfluorocarbonsare inertand heat stable;thusthey are oftenused in high temperature applicationsand make excellentinsulatorsand surfactants.Synthesishas been accomplished by electrochemicalfluorinationd,irectfluorinationt,eleomerizationa,nd catalyticmethods using high valence metals. Although fluoride(inorganicionicfluoridew)as identifieidn human blood 140 yearsago [Nickles,1856], the presence of fluorinein a freeior-dcstateas weu as a covalendy bound organicstatewas firsrteportedin 1968 [Taves, 1968a; 1968b]. Guy [1972] subsequentlyidentifiepderfluorooctanoicacid(PFOA, C7Fl5CO-,H) as a major component of theserum organicfluorinefraction.Ammonium perfluorooctanoatea, potent syntheticsurfactantused in industriaalpplicationsr,apidlydissociatesinaqueous solutiontoPFOA. SinceTave's and Guy's observations[Taves,1968a;1968b;Guy 1972),PFOA has been thesudjectof severaltoxicologicstudies. f In laboratoryanimals,PFOA acid or itssaltsisabsorbed by ingestion,inhalationor dermal exposure [Grfffitahnd Long, 1980; Kennedy 1985; Kennedy etal.,1986]. PFOA isnot metabolized [Ophaug and Singer,1980; Ylinen et al.,1990; Vanden Heuvel etal., 1991;Kuslikisetal.,1992].PFOA isdistributepdrimarilyin theplasma and liverof ni0e ratsand the liver,plasma and kidney in female [Vanden Heuvel etal.,199 11.Tllemajor route of elin-dnatioinn the male ratisvia urineand feceswhereas in the female ratthereis 3M Cofnpany EPI-OW3 Page 4 of42 a I0-fold greaterratein renalexcretion [Vanden Heuvel et al.,199 1; Hanhijarvi et al., 1982; 1987]. Castratedmale ratstreatedwith estradiolhave PFOA urinaryexcretion ratessimilarto female rats[Ylinen et al.,1990; Vanden Heuvel et al.,199 11. Peroxisome proliferatorsl,ikePFOA, are a diverse class of chemicals that cause hepaticperoxisome proliferationand enzyme induction,liverhyperplasiaand, in some instances,hepatocarcinogenesis in ratsand mice [Ikeda et al.,1985; Pastoor et al.,1987; Sibinski 1987; Cook etal.,1992; Biegel et al.,1995; Liu etal.,1995; Lemberger etal., )J 1996]. Peroxisome proliferatorbsind to and activateperoxisome proliferatora-ctivated receptors (PPAR) belonging to the superfamily of nuclear hormone receptors.Upon binding of a peroxisomal proliferatoror other ligands,such as fattyacids,PPAR interacts with RXR, another nuclearhormone receptoractivatedby 9-CIS retinoicacid. This heterodimer binds to specifichormone recognitionelements called peroxisome proliferator response elements (PPRE) in the promoter of targetgenes, resultingin the coordinated transactivationof a setof genes in peroxisomal, mitochondrial, niierosomal and cytosolic cellcompartments involved in lipidhomeostasis. Two-year feeding studiesin Crl:CD BR (CD) ratsata maximum amount of 300 ppm PFOA showed, in additionto liveradenomas and Leydig cell adenomas [Sibinsld 1987; Cook et al.,1994],an increased incidence of pancreas acinarcelladenomas [Cook et al.,19941. PFOA was not found to be mutagenic [Griffithand Long, 19801; thus the induction of these tumors most likelyoccurs via nongenotoxic mechar@sms [Biegel et al., 1995]. Evidence stronglyimplicatestherole of oxidative stressin livertw-nor development for peroxisome proliferators[Rao and Reddy, 1996]. Cook et al[19941 showed 3M Company EPI-0003 Page 5 of42 thatthe Leydig celltumors arelikelytheresultof increased estradiolevelsdue to inductionof hepaticaromatase activity. The pancreas tumors have been hypothesized tobe a secondary consequence of PFOA's effecton theratlivervia cholestasis-induceidncreasedplasma cholecystokinin (CCK) concentrations[Oboum etal.,1997a; 1997b). SpecificallyO,boum etal [1997a;1997b] exan-dnedthe possiblemechanisms for the pancreaticoneogenetic effects of two potentperoxisome proliferatorsW:yeth 14,643 and ammonium perflurooctanoate (C8).Both compounds in vitrofailedto: 1)bind to the CCK-A receptorin a competition binding assay;and 2) inhibitwsin in a continuousspectrophometricassay. Rats fed 100 ppm Wyeth 14,643 for60 days were found to have no pancreaticweight effects,increases in plasma CCK, acinarcellproliferatioonr increasedfecalfat. However, ratsfed 100 ppm Wyeth 14,643 for3 and 6 months had increasedpancreaticweight (6% and 17% above control,respectively)i,ncreasedCCK plasma levelsand acinarcellproliferation. Increased serum concentrationsof serum bileacids,alkalinephosphatase and buirubin suggested choleostasiasnd thiswas confirmed by measuring bileflow.Relativebileflow (relativteo liverweight)was decreased at 6 months to 72% of thecontrolgroup. Oboum et alconcluded thatWyeth 14,643 (and alsoinferredforPFOA) causes liver effects,includingcholeostasist,hatresultin increasedplasma CCK levels.IncreasedCCK levelshave been shown in otheranimal models to produce pancreatichypertrophy, hyperplasiaand neoplasia[Longnecker, 1986; 1990; Pour etal,198 1]. Appendix A providesa literaturreeview which offersa more detailedappreciation of theroleCCK playsinpancreaticphysiolooy.hypertrophy and neoplasia.Included in Appendix A arereviews of the followingtopics:1)an overview of theanatomy and 3M Company EPI-0003 Page6of42 histologoyfthepancreas2;)pancreatiecxocrinephysiologyasitrelatepsrimariltyoCCK; 3)theepiden-dologoyfpancreaticcancer,4)animalmodelsofpancreaticcancer,5)CCK receptorsand thepancreas;and 6) theroleof CCK and pancreaticcancer. A very brief review of theinfon-natiopnrovided in Appendix A issununarizedbelow. The principalinorganiccomponents of exocrinepancreaticsecretionsare water, sodium, potassium,chlorideand bicarbonate[Pandol,1993]. The principalorganic constituentsare proteins,primarilydigestiveenzymes, produced from the acinarcells.Thd enzymes are secretedintothe pancreaticductulesin an inactiveform with activation occurringin theintestinallumen primarilyby cleavagewith trypsin.Regulationof enzyme secretionsiscontroeed by both neuro- and humoral stimulationof the acinarcees. Receptors locatedon the basolateralsurfaceof the acinarcellshave been reportedfor CCK ,acetylcholineb,ombesin, substanceP and vasoactiveintestinapleptide(VIP) in the guinea pig,ratand mouse. However, CCK-A receptorswere not found in human, cynomologus and rhesusmonkeys [Gavin et al.,1997a;1997b]. The intestinaplhase representsthe most important aspectof pancreaticenzyme secretionwhich ismediated by both enteropancreaticvagovagal reflexesand hormones. Secretinisthe major humoral mediator of ductalbicarbonateand water secretionwhereas CCK isthe major humoral mediator of meal-stimulatedenzyme secretion.CCK isa potentregulatorpeptidethat alsostimulatesgallbladder contractionp,otentiatessecretin-inducepdancreatic bicarbonatesecretionand slows gastricemptying time. CCK isalsoa major neurotransn-dtteirn the brain. CCK existsin variousforms and lengthsalthoughsulfated CCK-33 appear to be the predominant form. The ocm peptideretainsfun activityof the 33 peptidemolecule.CCK in the duodenum isinhibitedby a negativefeedbackmechanism involvinpgrimariltyrypsinW.hether CCK 3M Company EPI-0003 Page7 of42 initiatoerspromotespancreaticcancerremains highlycontroversial.Itisclearthatadministrationof exogenous CCK or itsanalog ceruleindoes induce pancreatichypertrophy and hyperplasiain severalspeciesas measured by increasedDDN synthesis,DNA content,RNA content and P-landulawreight. Various methodologic approaches have examined the issueof whether CCK causes pancreaticneoplasia.These endeavors have includedthe administrationof exogenous CCK, manipulationof endogenous CCK, administrationof CCK receptorantagonists, measurement of CCK receptorbindingactivityand detectionof CCK receptorMRNA. Data from more than seventy laboratoryanimal studieshave provided mixed results [Herringtonand Adrian, 1995]. CCK-8, CCK-9 and CCK-39 have promoted growth of human pancreaticcancers incellculturein serum-freemedia [Palmer-SrrLitehtal.,199 1]. However, fastingplasma concentrationsof CCK in unresectedpancreaticcancer patients did not differfrom healthycontrols[Rehfeldet al.,1994;Adrian et al.,1994). Because theCCK receptoractivitoyf theratmay be quitedissimilarto thehuman [Gavin et al.,1997 a;1997b], a cynomologus monkey may be a more appropriateanimal model touse for a pancreaticpathophysiology consequence of exposure to PFOA in the human. In thisregard,3M, in conjunctionwith the APME, has initiateda minimum 3 month feedingstudyof PFOA to cynomologus monkeys. A primary study hypothesisis whether increasedCCK levelsand pancreatichypertrophy are associatedwith serum PFOA levelsinthesecynomologus monkeys. Because thepancreaticeffects(e.g., increasedCCK levels,increasedpancreasweights)in ratswere not observed by Oboum et al [1997a;1997b] untilthe sixthstudy month, itisanticipatedthatatleastthe same 3M Company F-PI-0003 Page9 of42 comparabletimeperiod,ifnotmuch more,willbe neededtodeterminewhethersuchan effecetxistsi,fatall,incynomologusmonkeys. Because the roleof CCK with pancreaticcancer remains quitecontroversialin laboratoryanimals,furtherinsightintoany roleexposure to PFOA may have in relationto CCK in humans should be considered.In thisregard,we propose toexamine, via a cross sectionalepidemiologicalstudy design,theplasma CCK levelsof 3M Cottage Grove fluorochemicalproductionworkers in relationto theirserum PFOA levels.Due to the hypothesisthatelevatedlevelsof CCK in theratare a resultof choleostasisw,e willalso examine hepaticenzyme function,buirubinand lipoproteins. METHODS Study Design and Population 'Me studyisa cross-sectionadlesign.Cottage Grove production have biennial medical surveillanceexaminations as specifiedin the 3M Medical SurveillanceProtocolfor fluorochen-dcals.Eligiblemployeees areallbuilding15 employees (departmentnumbers 3020 and 3060),building6 employees (departmentnumber 3036) and building3 employees (theywillautomaticallybe includedas they are staffedfrom buildings6 and 15). In addition,allplantengineeringemployees and etaftworkerswho sevicebuildings6 and 15 areeligibleforthestudy (departments0698 and 0690). Itisanticipatedthatthere willbe a maximum of 100 employees eligiblfeorthemedical surveillancexamination. Data Collection As specifiebdy the3M 3M CoFnpany EPI-OW3 Page9 of42 MedicalSurveillancPerotocolforfluorochemicaltsh,e medical history,testsand physicalexan-dnationrequirementsinclude:a medical questionnairealong with a self-administeresdpecialquestionnairedesigned forthe fluorochemicalproductionarea (Appendix B), spirometry,serum chemistriesh,ematology, urinalysiss,erum PFOA measurements, and the measurement of the employee's height, weight,blood pressureand pulse. These medical surveillancexan-dnationsare scheduled to take placebetween September 15, 1997 and October 31, 1997. Clinicallaboratory evaluationswilloccur atUnited Hospitals(SL Paul,Minnesota). Serum PFOA levels willbe evaluatedby Dr. Jack Henion attheAdvanced BioanalyticalServicesInc.(Ithaca, New York). In additiontotheabove medical surveillancpearameters,we willalsodraw 10 rrd of blood forplasma CCK-33 measurements. CCK-33 willbe measured by direct radioimmunoassay by InterScience Institut(eInglewood,California)T.he InterScience Institute'tsechnicalcoordinatorforthisprojectwillbe Alan Kacena. T"nerequirements for plasma CCK determinationare a 10-12 hour fastpriorto collectionof the specimen. Antacid medication or medications affectingintestinamlotilityshould alsobe discontinued, ifpossible,foratleast48 hours priorto collection.The InterScience Institutweillsupply 3M a 10 rrdEDTA collectiotnube which containsthespecialG.I.preservativeT.rasylol. Plasma should be separatedfrom the ceus inunediatelyaftercollectionand then frozenin a plasticvial.Specimens should be shipped on dry iceto the laboratory.Specimens can be shipped in a batch sample. The expectedupperreferencelevelofCCK 3M Company EPI-OW3 Page10of42 atthislaboratoriys80 pg/ml.Ile laboratoryhas provided the following dataregardingqualityassuranceaspectsof itsCCK- 33 radioimmunoassay. SRgcificii-yCross reactivitwyas determined atthe50% inhibitioonf bindinglevel. Compound CCK Gastrin Secretin Glucagon Insulin Vasoactive IntestinaPlolypeptide GastricInhibitoryPolypeptide PancreaticPolypeptide Modhn Other compounds tested Cross-reactivity 1.00 0.04 0.02 <0.01 <0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 Recovery - Specimens were spiked with a known quantityof CCK and measured to determine theamount of recovery. Specimen (p-p-/n-d) 17 18 92 Amount added 20 50 50 Amount measured 34 66 131 Amount expected 37 68 142 Recovery (%) 91.9 97.0 92.2 Intra-assaVyariabilit-y ne mean, standarddeviationand coefficienotf variation for threecontrolsassayed in one run are listedbelow. Mean (DP-/rrd) 16 94 159 Standard Deviation (Rp-lnll) 2.3 7.4 13.0 Coefficientof Variation (%) 14.4 7.9 8.2 3M Company EPI-OW3 PageIIof42 InteT-assgLXgdati-oTnhe mean,standarddeviationa,nd coefficieonftvariation for threecontrolasssayedindifferenrtunsarelistebdelow. Mean (Dg/n-A) 18 91 163 Standard Deviation (R&LrLil) 2.7 8.1 15.2 Coefficientof Variation (%) 15.0 8.9 9.3 A consentform willbe signedby allfluorochernicaplroductionemployees who Ji participatein the detemiination of theirplasma CCK levels(Appendix C). In essence, the only additionalitem thatwillbe collectedbeyond that already specifiedin the Medical Surveillance Protocol Manual for fluorochen-dcal workers isthe 10 ml of blood necessary for plasma CCK radioinununoassay analysis. Data Analysis After consideration of normality of the data, stratifiedanalyses, ANOVA, Pearson correlation coefficients,and linearmultivariate regressions will be used to evaluate for positive or negative associations between PFOA and CCK-33, hepatic enzymes and lipoproteins. Itis anticipatedthatthe following PFOA levels wW be used for the (categoricalanalyses: 0 - < 1 ppm, 1 - < 10 ppm, 10 - < 30 ppm and > 30 ppm. These levels were used in a previous study of reproductive hormone levels and PFOA among the same Cottage Grove fluorochen-dcalproduction workforce [Olsen et al.,1997]. Four potential confounders will be considered in the analyses: alcohol consumption, cigarette smoking, body mass index and the employee's age. Study resultswill be tabulated and analyzed by packaged procedures in the SAS System [SAS, 1990]. 3M Cornpany EPI-"3 Page12of42 DISCUSSION The purpose of thisepidemiologicinvestigatioinsto determine whether plasma CCK-33 levelsareassociatedwith serum PFOA levelsamong fluorochemicalproduction employees at the 3M Cottage Grove plant.There are severalstrengthsto the proposed study. First,althoughitishypothesizedthatPFOA may cause CCK-induced pancreatic acinaradenomas inthe rat,thismodel may be irrelevantas the histologyand CCK-A receptoractivityof the ratappears to be quitedifferenthan thatobserved forhumans. Second, thisstudywillmeasure plasma CCK and serum PFOA levels.This levelof specificitoyf exposureisoftenmissing in occupationalepiden-dologyinvestigations. Qualitative/subjectidvaeta willonly be used in the analysisof alcoholconsumption and cigarettesmoking. Third,the resultsof thisstudy may have considerableimportance in the interpretatioonf CCK-related dataobtainedfrom the cynomologus monkey study. Although an infrequentoccurrence,resultsfrom thisepidemiologicinvestigatiomnay directfurthertoxicologicasltudies.Fourth,in additionto plasma CCK-33 levels,we will alsoexamine hepaticenzymes, bwmbin and lipoproteilnevelsin relationto serum PFOA levelsbecause choleostasiisstheproposed mechanism forenhanced CCK levelsin theraL Severalmethodologicalissueswillneed to be consideredin evaluatingtheresults from thisstudy.First,thecross-sectionadlesign does not allow fora directanalysisof the temporalityof an association.Given thelong-halflifeof PFOA, itisconceivablethat theremay be some biologicalaccommodation to the effectsof PFOA, as suggested by Biegel et al[1995]. Second, although theremay be upwards of 80 employees eligiblefor CCK-33 determination,not allsubjectsmay participateA.ny refusalswilldecreasethe 3M Company EPI-0003 Page13of42 statistpiocwaelrofthestudyandresulitnnonresponbsieas.Thirdt,hereexistpsotential measurement errorfortheconfoundersofinterestC.urrentlya,lcoholconsumptionand cigarettuesage arepoliticallsyensitiviessuesand thustheaccuracyoftheseself-reported datamust be questioned.Fourth,itisrecorrkmendedthatCCK measurements be done aftera 10-12 hour fast.Employees willbe reminded of thisfact-h,owever,whetherthey adhere to thisremains-to-bes-een. 'ne protocola,ny addenda totheprotocold,ataanalysesa,nd a copy of thefinal reportwillundergoa QualityAssuranceaudit.Permanentrecordsofallotherdata generatedduringthecourseofthisstudyaresubjecttoprivacyand confidentiality considerationsA.lldatagatheredorgeneratedincludingprotocoladdendum and thefinal reportwillbe a-rchivebdy theMedicalDepartment,3M Company, St.Paul,Minnesota. Costsforroutinemedicalsurveillancaend theepidemiologicanalyseswillbe bome by theMedicalDepartmenL CostsforCCK-33 analysiwsillbe coveredby theSpecialty ChemicalsDivision.InterScienceLaboratoryhasquoted$120 persampleforCCK-33 radioimrnunoassaybased on an 100 personbatchsample. Thus, costsshouldnotexceed $12,000fortheCCK-33 analyses. Itisestimatedthatdatacollectiownillbe completedby November 15,1997. Sample analysesshouldbe completed by January15,1998. Data analysiasnd report writingwilltakeapproximately3 months. Thus, a draftmanuscriptreadyforreviewand approvalshouldbe availablbey April1,1998. Upon completionofthisstudy,resultwsillbe conununicatedtomanagement, fluorochen-dcparloductionemployeesand thetechnicaclommunity. A manuscriptfor publicatiocnonsideratioinna scientifjiocurnalwM 3M Coinpany EPI-OW3 Page 14 of 42 be preparedand submittedforPeer- review. 4 f .I 3M Coffnpany EPI-OW3 Page 15 of 42 APPENDICK A ii f Pancreas Anatomy and Histg!M 3M Company EPI-0003 Page 16 of42 The pancreasfirstappears embryologicallyatthe fourthweek of gestation[Ermak and GrendeU, 19931.'Me pancreas arisesfrom a dorsaland ventraloutpouching. ne tail, body and partof the head of the pancreas are formed by the dorsalcomponent. The remainder of the head of the pancreas develops from the ventraloutpouching.In the adult, the pancreas measures 12 to 15 cm inlength and weighs between 70 and 100 grrlThe i circulationof thepancreas isderivedfrom branches of theciliacand superiorn-iesenteric arteries.The venous drainage flows intothe portalsystem. The visceralefferent innervationof the pancreas isthrough the vagi and splanchnicnerves via thehepatic and celiacplexuses,respectively. The pancreasfunctionsas both an exocrine and endocrine organ. 7he endocrine pancreasislocatedin the isletsof Langerhans. Glucagon, insulin,somatostatinand pancreaticpolypeptideare produced in the A, B, D and PP ceus of theisletsof Langerhans, respectively.B cellsare the most abundant representing50 to 80 percent of islevtolume. The exocrinepancreas consistsof countlessaciniwith theiraccompanying ductules.The isletsareintermixedwith theacini. Acinar cellssurroundingtheisletsof Langerhans appear tobe morphologicallyand biochen-dcauydifferenftrom acinifurther removed from theisletsA.n acinusisa network of adjoiningacinarceus which can take on a varietyof shapes(e.g.,sphericalt,ubularor irregular)T.he intercellulacronnection between acinarcellsisthe gap junctionwhich functionsas a pore to allow smallmolecules (500 to 1000 daltons)topass between thecells.The acinarcellssynthesize,storeand secretedigestiveenzymes intothe lumen of the acinus. On an acinarcell'sbasolateral membrane are receptorsfor hormones and neurotransrnittertshatstimulateenzyme 3M Company EPI-OW3 Page 17 of42 secretion.The basalregion(ftu-thesatway from thelumen of the acinus)of theacinarcell containsrough endoplasmic reticulumforproteinsynthesisand comprises about 20 percentof thecellvolume. The apicalregion(nearestthelumen of theacinus)of the acinarcellcontainszymogen granuleswhich hold the digestiveenzymes. TlleGolgi complex islocatedbetween the nucleus and zymogen granules.ne lumen of the acinus is the originof the secretoryduct and containscentroacinarcells. The centroacinarceus functionsimilarlyto theductuleepithehalcellsby secretingionsand water.The lumen o? the acinusleadsto intralobuladructs which eventuallyanastomose to createinterlobular ducts. The interlobuladructs anastomose to form themain pancreaticduct. The primary functionof the ductalsystem isto secretean alkalinesolutionwhich aidsinthe transport of digestiveenzymes tothe lumen of the intestine.This alkalinitaylsoraisesthe intralumenpH of the intestinneecessaryfordigestiveenzyme action. Pancreas Physioloev The principalinorganiccomponents of exocrinepancreaticsecretionsare water, sodium, potassium chlorideand bicarbonate[Pandol,1993]. Smaller quantitieosf calcium,magnesium, zinc,phosphate and sulfateare alsosecreted. The purpose of these secretionsisto aid inthe deliveryof digestiveenzymes to the intestinallumen and neutralizegastricacidemptied intothe duodenum. These inorganicpancreaticsecretions, produced primarilyfrom the ductalsystem as the resultof secretinstimulationin the upper intestinamlucosa, vary from 0.2 ml/min in the restingstateto 4.0 n-d/rnidnuring stimulation,Aiththetotaldailyvolume equal to 2.5 litersT.he bicarbonateconcentrate variesbetween 25 nEq/L (low flow) to 120 mEqL (highflow).Chlorideconcentrations 3M Cornpany EPI-0003 Page18of42 varyinversely.Secretisntimulatessecretiobny activatinagdenylatceyclaseand increasingcyclicadenosinemonophosphate (CANT) which ultimatelyresultisn an exchange of chloridefor bicarbonateionsin the lumen of thepancreaticductule. The principalorganicconstituentsof the pancreas are proteins,primarilydigestive enzymes, produced from theacinarcells[Pandol,1993]. The major enzymes functionin thefollowingmanner: digeststarchand glycogen (wnylase);hydrolyzetriglyceride(slipase and phospholipase A2) as well as cholesterolesters,lipidsolublevitamin esters, diglycerideand monoglyceride (carboxylesterasea)n;d cleavepeptidebonds (tr)Tsin, chymotripsin and elastase). The concentrationof proteinin thepancreaticsecretionsdepends on the output ratefrom the acinarcells.The enzymes are secretedintothe pancreaticduct in an inactive precursorform because theycould potentiallydigestthe pancreas. Digestiveenzymes in the acinarcellinclude:proteolyticenzymes (u-@Tsinogen,chymou-ypsinogen, proelastase, procarboxypeptidaseA, procarboxypeptidaseB); amyolyticenzyme (alphaamylase); lipolytiecnzymes (lipasep,ro-phospholipaseA2, carboxylesteraselipase)n,ucleases (deoxyribonucleaser,ibonuclease)and otherenzymes (pro-colipaseand nypsin inhibitor). Activationof theprecursorform to the finalenzyme occursin the intestinallumen. Trypsinogen isactivatedto n-ypsinviatheactionof enteroldnase.ne uwsin then,in turn,activatesothertypes of proenzymes to theirfinalenzyme form (e.g., chymoawsinogen, proelastasep,rocarboxypeptidasesA and B, prophospholipase A2 to chymouwsin, elastasec,arboxypeptidasesA and B, and phospholipase A2,respectively). In additionto the proenzyme activationof theseotherenzymes, t-ypsinalsoactivatesthe conversion of u-@Tsinogen toawsin. 3M Ccxnpany EPI-"3 Page19of42 T'hesynthesiosfthesedigestiveenzymes occursintheroughendoplasniic reticulumof the acinarcells.Once synthesized,theseproteinsare transportedto the acinarcell'sGolgi complex where furtherglycosylationand concentrationoccur. Finally the enzymes are transportedto the zymogen granulesvia vesiclesthatcycleback and forth between the Golgi complex and the granules.Itappears thateach zymogen granule containsthe entirecomplement of secretoryenzymes althoughtheirconcentrationsmay differbetween granules.The percentceU volume thatthe zymogen granulesrepresents can vary from 15 to 20 percent(fastedadult)to lessthan one percent afterstimulation with a cholinergicdrug or the gastrointestinhaolrmone, cholecystoldnin.The enzymes are believedto be releasedfrom the zymogen granulesby exocytosis(thefusionof the granule membrane with the apicalceU membrane and subsequent releaseof the granule content intotheductulelumen). Itisthoughtthatdifferentstimuliresultin theselectivesecretion of specificenzymes from theacinatcees. Furthermore,theregulationof protein (enzyme) synthesiscan be alteredby dietaryintake. Increasedconcentrationsof amylase and decreased levelsof chymotrypsinogen have been reportedwith carbohydrate-rich diets. Studies on theregulationof enzyme secretionin humans are lin-dte[dLu et al., 1989; Chey, 1991; Pandol,19931. The majorityof informationknown, to date,isfrom animal models. Regulationof enzyme secretioniscontrolledby neuro-and humoral stimulationof theacinarcells. Receptors,locatedon thebasolateralsurfaceof theacinar cell,have been reportedforcholecystokinin(CCK), acetylcholineb,ombesin, substanceP, vasoactiveintestinapleptide(VIP) in theguinea pig,rat,and mouse. Depending on their mode of stimulus-secretioncoupling,thereceptorshave been divided intotwo categories. 3M Company EPI-"3 @age20of42 VEPandsecretiinncreacseellulCaArMP viaactivatoifoandenylactyeclasAen.increaisne CAMP mediatesthesecretorryesponse.The acinarcellalsocontainsreceptorfsorCCK. acetylcholinbeo,mbesin and substanceP. RatherthanstimulatincgAkT, these compounds increasecellulamretabolismofmembrane phospinositideasnd calcium. A majoreffectofCCY, acetylcholinbeo,mbesin and substanceP isthemobilizatioannd releaseofintracellulsatroresofcalcium. These agonistsalsoincreasecellulacryclic guanosinemonophosphate,arachidonatereleasefrom membrane phospholipidsa,nd electricamlembrane potentiaclhanges.ne exactmechanisms by which thesestepslead toenzyme secretiornemaintobe fullyelucidated.There does appeartobe a greaterthan additivreesponsetothejointeffectsofagonistswhich alteCrAMP (e.2.V,EP and secretin) and calcium(e.g.C,CK and acetylcholine). Exocrinepancreatiscecretioonccursduringfasting(interdigestiavsew)ellas after ingestioonf a meal (digestive)T.he digestivsetateisdividedintothreephases:cephalic, gastricand intestinalT.he cephalicand gastricphases areneuro-regulatewdhereas the intestinpahlase isregulatedby neurologicaalnd hormonal factors. ln thecephalicphase,increasedpancreaticsecretionosf bicarbonataend digestive enzymes willoccuraftersmellingt,astingc,he*ing and swallowingfoodwithor without duodenalchyme or acidificatioTnh.isisdue toincreasedvagaltonewhich increases intrapancreatpiocstganglionicchohnergicinputwhich subsequentlystimulatetsherelease of thebicarbonateand enzymes. These neuronsinthepancreasareactivatebdy central inputduringthecephalicphase and by vagovagalreflexesinitiatebdy thestimulation duringgastriacnd intestinpahlases.Besidesacetylcholinteh,ereareother neurotransmitteirnsthepancreasthatcontainpeptides,VIP, gastrin-releasipnegptide, 3M Compmy EPI-OW3 Page21of42 CCK, neuropeptidYe, neuortensins,ubstanceP,enkephalinsc,alcitonignenerelated peptide and galanin.T"nesepeptidesmay coexistwith nonpeptide transmitterisn autonomic nerves and thus appear toplay significanrtolesinthe regulationof the exocrine pancreas. 'Me gastricphase of pancreaticsecretionresultsfrom food stimuliinthe stomach. The gastricstimulicause primarilyenzymatic releasewith rrdnimum secretionof water and bicarbonate. The intestinaplhase representsthe most important aspect of pancreaticenzyme secretion.The intestinaplhase begins upon entryof chyme intothe duodenum. The intestinaplhase ismediated by both enteropancreaticvagovagal reflexesand hortnones. Ilere aretwo major mediators of these pancreaticenzymatic secretionss:ecretinand CCK. The type ofpancreaticsecretiondepends on whether itisreleasedby secretinor CCK. Secretinisthemajor mediator of ductalbicarbonateand water secretion.CCK is the major mediator ofpancreaticenzyme secretion. T'he releaseof secretinfrom the duodenal mucosa depends upon the acid load (minimum pH of 4.5)thatisdeliveredto theduodenum. Fatty acidsgreaterthan eight carbons in lengthand bileacidsare likelysecondary stimulantsforsecretinrelease. Secretionof bicarbonatesalsooccursvia cholinergicinpul Although CCK by itselfdoes not invoke bicarbonaterelease,itcan augment secretin-inducedbicarbonatesecretion. CCK isthemajor humoral mediator of meal-stimulatedenzyme secretion.CCK is releasedfrom the'1'(orsometimes ca.Ued'M') cellsof theupper smallintestinamlucosa by productsfrom fatand proteindigestion. Phenylalanine,valine,methionineand tryptophan are the most potentamino acidsfor CCK release(and subsequentpancreas 3M Cornpany EPI-0003 Page22of42 enzyme secretion)A.low levelofintestincaolntentmsediatepancreatiecnzyme secretion by an enteropancreatinceuralreflexwhereas high loads of intestinaclontentsmediate enzyme secretionpredominantly by thehormonal effectsof CCK. The releaseof secretin and CCK ismediated by releasingpeptidessecretedfrom the intestinamlucos& These releasingpeptidesare,in turn,inactivatedin the upper smallintestinallumen by pancreatic proteases. Specificallyt,heregulationof CCK occursin thefollowingmanner: 1)chyme entersthe duodenum which resultsin thereleasingpeptidessecretedfrom the intestinal mucosa which subsequentlyreleaseCCK from the 'I'ceus;2) intestinaClCK isabsorbed and travels,via the circulatorysystem, to the pancreas where itbinds to receptors;3) the activatedreceptors causes the releaseof pancreaticprecursorenzymes from the acinar ceus' zymogen granulesintothe pancreaticductuleswhich subsequentlydraininto pancreaticductand then empties intothe intestinallumen; 4) the precursorenzymes are activatedin theintestinallumen where one of theseenzymes, t-ypsin,assistsin the digestionof protein;5) upon cessationof eating,theconcentrationof freenwsin increasesin theintestinallumen with thedecreasedlumen contents;6) thisfreetr)Tsin inhibitstwo substancesthatactas stimulifor CCK release(monitor peptideand CCKreleasingfactor)7;) thisinhibitioonf monitor pepdde and CCK-releasing factor subsequentlyresultsinthe cessationof CCK releasefrom the duodenal 'I'cens; and finally 8) a decreasedCCK circulatorcyoncentrationresultsinlessCCK bindingactivityatthe acinarmembrane which ultimatelydecreasespancreaticpro-enzyme secretionsfrom the zymogen granules. Of course,thisnegativefeedback cyclebegins again with the next stimulienteringtheduodenum (i.e.t,henext meal). Itappears,albeitto a much lesser 3M Company EPI-OW3 Page23of42 extent,thatcirculatiinngsulins,tarchg,lucoseand calciummay alsoproducean pancreaticenzyme secretoryresponse. The evaluationof exocrinepancreaticfunctionisaccomplished by both directand indirecttests. Because of itslargefunctionalreserve,malabsorptiondoes not occur until CCK-stimulated digestiveenzyme secretionisreduced to 10 percentof nontal; thus most testshave low sensitivittyo detectmild to moderate degrees of pancreaticinsufficiency. ne directfunctiontestsare based on the measurement of enzyme and bicarbonate secretionupon duodenal intubationand stimulationof secretinand/or CCK. A varietyof indirecttestsareavailablewhich exan-dnesurrogatesof eitherstimuliand/or the actual digestiveenzyme secretions[Pandol,1993]. Ej2iden-dologyofPancreaticCancer Briefly,cancer of the pancreasisthe ninthleadingcause of cancer and the fourth leading cause of cancer death formen and women in the United States[Anderson et al., 19961. Five-yearsurvivorshipislessthan 5 percenl Pancreaticcancer occursfifty percentmore frequentlyin males thanfemales and blacks than whites. Adenocarcinoma of thepancreaticductulesistheprimary histologictype. 'Me proportionof pancreatic f cancer from acinaroriginisestimatedat I to 15 percent.Numerous pancreaticcancer case-controlstudiespublished over the last15 years have failedto show any strong associations.The most consistentassociationis,on average,a two-fold increasedrisk with cigarettesmoking. Severallinesof epiderr@ologicevidence suggest dietshigh in animal fatand low invegetableand fruitsare weakly associatedwith pancreaticcancer. 3M Company EPI-OW3 Page24of42 No consistenatssociatiohnasvebeenreportedforindustryo,ccupationo,rspecific chemical exposures. The most important discoveryduring thepast 15 years has been the accumulation of datawhich shows thatmutations in cellularptoto-oncogenes and tumor suppressor genes are importantevents inpancreaticcarcinogenesis[Caldas and Kem, 19951. K-ras mutations are a frequentfindingin adenocarcinomas of the pancreasin humans with the greatmajorityof thesemutationsfound incodon 12 of c-Kirstenras. In fact,pancreatic canceristhehuman tumor with thehighestincidenceof ras mutations. 7ne ras gene familyencodes proteinsinvolved in cellgrowth and differentiationT.hus, pointmutations inras genes may play an importantrolein earlypancreaticcarcinogenesis.Other evidence suggestrolesformutationsin otheroncogenes (myc, erbB-2) and tumor suppressorgenes (apc,p53). These findingssupporta geneticmodel of pancreatic tumorigenesis:ductalcellsdrivenby almost universalmutation of a dominant oncogene, K-ras and deregulationof cell-cyclceontrol.YIhethermultiplecarcinogensmay cause pancreaticcancerremains a plausiblehypothesis However, thedegree of consistencyin themutations,transversionsand transitionosf nucleotidesfrom G to T or G to A in codon 12 of K-ras may eventuallypointto only a few specificcauses. t Animal Models of PancreaticCancer There aretwo animal models of pancreaticcancer [Longnecker, 1990]. A number of nitrosocompounds have been used to produce ductalpancreaticcancer in hamsters includingN-nitrosobis(2-oxopropyl)arni(nBeOP), N-nitrosobis(2-hydroxypropyl)an-Line (BHP), and N-nitroso(2-hy&oxypropyl)(2-oxopropyl)ami(nHePOP) [Pouretal., 3M Company EPI-OW3 Page 25 of 42 1975;1981].In therat,azaserin(eo-diazoacetyl-L-serinset)hemost conunonlyused carcinogen[Longnecker,19861.However thehamsterand ratmodels areofdifferent cellulaorrigin.Carcinogensinratsinduceacinarcellmalignancieswhich arerareinthe human. Hamster models areofductalcelloriginand thusresemblehuman pancreatic cancer.Furthermore,thepatternof geneticmutationsfound intumorsfrom BOP-treated hamstersparalleltshatof human cancers.Activationofthec-K-rasgene isfrequentin both human and hamsterpancreaticancerbutisnotfound inazaserine-induced pancreaticancerin therat[vanKranen etal.,1991]. CCK Receptor-san.dthePancreas 3M Cornpany EPI-M3 Page26of42 As discussedpreviously,CCK isa potentregulatorypeptide thatisthe major stimulusfor pancreaticenzyme secretionC.CK-33 refersto its33 amino acid arrangement: Lys-Ala-Pro-Ser-Gly-Arg-Met-Ser-Ee-ValLys-Asn-L-eu-Gln-Asn-Leu-Asp-Pro-Ser-HisArg-Ile-Ser-Asp-Arg-Asp-Tyr-Met-Gly-TrpMet-Asp-Phe where Ala alanine Arg arginine Asn asparagine Asp asparticacid Gln glutamine Gly glycine His histidine Ile= isoleucine Leu = leucine Lys = lysine Met = methioriine Phe = phenylalanine Pro = proline Ser = serine Trp = ayptophan Tyr = tyrosine Val = valine CCK stimulatesgallbladdercontractionp,otentiatessecretin-inducedpancreatic bicarbonatesecretiona,nd slows gastricempriing. CCK alsoexistsinvariousforms and lengthsfrom metabolitesCCK-4 to a pro-CCK peptide,CCK-58. The octapeptide retainsfullactivitoyf the 33 peptidemolecule. Sulphated CCK-8 and CCK-33 appear to be the predominant forms. CCK actsthrough more than one type of receptor.In animals, CCK-A receptorsarefound on pancreaticacinarceus, smooth muscles, vagal afferent fibersand some centralneurons. CCK-A receptorshave a high affinitfyorsulphated CCK-8 andrecognizegastripnoorly.On theotherhand,CCK-B 3M Company EPI-0003 Page27of42 receptorasreabundantly found in the centralnervous system and on gastricglands and recognizesboth sulfatedand nonsulfatedCCK-8 and gastrin.Because circulatinlgevelsof gastrinareappreciably higher thanthose of CCK, the CCK-B receptorsare usuallyactivatedby gastrinleaving the CCK-A receptorsas thepredominant form forCCK binding. CCK-A and CCK-B receptorsarenot consistentlyfound in allspecies.Recent studiesindicate,thatunlikeratand dog pancreas, the human pancreas has no detectable CCK-A receptorsand littlteono MRNA for the receptor [Wank et al.,1994]. In particularh,uman, cynomolgus and rhesus monkeys lack specificbinding sitesfor the selectivCeCK-A ligand'[H]L-364,718 [Gavin et al.,19973. In contrast,the baboon pancreas exhibiteda significanntumber of specificCCK-A binding& sitesalthough not as much as therator guinea pig pancreas. As alsostatedpreviously,CCK isinhibitedby a negativefeedback mechanism involvingprimarilytrypsinl:ackof trypsinin theduodenum and jejunum resultsin CCK release,whereas the presenceof freeti-)TsiinnhibitsCCK release.Thus, any interference with thisfeedbackmechanism by orallyadministeredansin inhibitorwsillresultin continuedpancreaticsecretionviathereleaseof CCK Tllesetrophiceffectsof CCK have been demonstrated in both ratsand hamsters [Pour et al.,198 1; Longnecker, 1990]. ne Role ofCCK and PancreatiCcancer 3M C(xnpwiy EPI-0003 Page28of42 Whether CCK initiateosr promotes pancreaticcancer isnot a new questionand the researchpublished,to date,remains higmy controversiawlith elusiveanswers [Axelson et al.,1992; Herringtonand Adrian, 19951. What isclearisthatadministratioonf exogenous CCK or itsanalog ceruleindoes induce pancreatichypertrophyand hyperplasia in severalspeciesas measured by increasedDNA synthesis,DNA content,RNA content@ totalproteincontent and glandularweight [Mainz etal.,1973; Petersenet al.,1978; Zucker et al.,1989).Increasedendogenous plasma CCK levels,via pancreaticobiliary diversion(thoughtto inhibithenegativefeedback regulationof CCK release)h,ave also resultedin pancreatichypertrophy and hyperplasia[Gasslandereta].,1990]. Furthermore, administrationof a CCK-A receptorantagonistwillabolishor markedly reduce the pancreatictrophicresponse of CCK. Feeding ratsor hamsters a high-fator high-protein dietinduces pancreatichypertrophy and hyperplasia[Longnecker etal.,1990].. Various methodologic approaches have examined the issueof whether CCK causes pancreaticneoplasia,not justhypertrophy and hyperplasia These research endeavors have includedthe administrationof exogenous CCK, manipulationof endogenous CCK, administrationof CCK receptorantagonistsm,easurement of CCK receptorbinding activitya,nd detectionof CCK receptorMRNA. Herringtonand Adrian [1995] have summarized thisbody of researchby stating,"in spiteof extensiveresearch, theroleof CCK inpancreaticcancer isstilvlery unclear,and the subjectishighly controversial.Data from more than seventystudieshave variablysuggestedthatCCK has no positivetrophiceffectsi,nhibitoryeffectso,r no involvement inpancreatictumor growth." 3M Ccffnpany EPI-0003 Page29 of42 Providebdelowforillustrafipvuerposesarefiveexamples offeredby Herrington and Adrian [1995] re2ardinp-the controversialissues sur-roundingC-CK and pancreatic cancer. 1.) Several in vivo experiments with exogenous CCK or cerulein (a CCK analog derived from frog sldn)enhanced the development of pancreaticcancer in hamsters t:reatewdith carcinogens [Howatson and Carter, 1985; Satake et al.,1986; Pour et al., 19881 yet other hamster studiesshowed CCK reduced the incidence of pancreaticcarcinomas in hamsters or produced no effect[Pour et al., 1988, Johnson et al.,19883. 2.) Endogenous CCK levelshave been associatedwith pancreatictumors. For example, pancreaticobiliarydiversionenhanced azaserine-treatedpancreaticcancer development in rats [Stewart et al.,1991]. Long-term feeding of raw soya flourto rats.which contains protease inhibitors,produced pancreas acinar adenomas and adenocareinomas [McGuinness etal.,1980]. However, feeding raw soya flouror syntheticprotease inhibitorsto hamsters did not resultin increased pancreatic tumors even thoup-hplasma CCK levelswere increased [Herrington etal.,1994]. 3.) CCK receptor antagonistshave been used to investigatethe role of CCK, exogenously or endogenously, in pancreaticcancer. These antagonistsinclude glutararnicacid derivatives(proglumide, lorglun-dde,loxiglumide) and nonpeptide compounds (asperlicin, devazepide, L365,260). The resultsof these studiesvaried considerably depending upon animal model, the timing of the administrationof the antagonistinrelationto administrationof the carcinogen, dosage administered and culturedcells.Studies of the CCK receptor antagonistdevazepide in patientswith pancreaticcancer have not shown significanteffectson sur-vival[Abbruzzese et al.,1992]. 4.) CCK binding was consistentlydemonstrated in whole ceus and membranes from acinar (celltumors in the ratazaserine model. However, specificCCK receptorshave not been demonstrated on intact,culturedhuman tumor cellsof ductalcellorigin[Adrian et al., 19941. Nevertheless,CCK binding has been demonstrated in membrane fractionsfrom severaldifferentpancreaticcancer celllines[Singh et al 1986; Sn-dthet al., 1994] but this significanceisuncertain when receptors cannot be demonstrated on the cellsurface [Herrington and Adrian, 1995]. 5.) CCK (CCK-8, CCK-9 and CCK-39) promoted growth of pancreaticcancer in ceU culturein serum-free media [Palmer- Sn-dthet al.,199 1). On the otherhand, fasting plasma concentrationsof CCK in unresected pancreaticcancer pafientsdid not differfrom levelsin healthy controlsor in patientswith colon or lung cancer [Rehfeld etal., 3M Company EPI-OW3 Page 30 of42 1994;Adrian etal.,1994]. CCK levelsdiddeclinein pancreaticcancerpatientsafter pancreatoduodenectomy due to thelossof CCK-producing I ceus locatedin the duodenum [Adrian etal.,1994]. In conclusion,CCK likelypromotes pancreatictumors in some animals but thereis no evidence thatitisan inducerof pancreatictumors. Although CCK receptorshave not been extensivelyinvestigatedin thenormal human pancreas,theremay well be a different patternof receptorexpressionin humans than otherspecies.Iftrue,then resultsfrom laboratoryanimals,inparticulatrhe azaserine-raatnd nitrosamine-hamstermodels, may be of nib-dmum applicabilityo understandingCCK and itsrelationto abnormal pancreatic funcdon in humans. APPENDIX B I 3M Ccimpany EPI-"3 Page 31 of42 JJ t Last Name 3M Health Questionnaire FirstName Middle Name Employee Number Today's Date 3M Company EPI-0003 Page 32 of 42 Oats of Hire Directions:There are three partstothisquestionnaire.Read the directiontsoeach partofthe questionnaire.Answer the questionsinsequence startinwgith Part 1.CircJeyour response on the questionnaire.For questions thatrequirea filIlnthe blank response,please provideyour answer on the space providedon the questionnaire.Erase cleanlyany answer you wish to change. We greatly appreciateyour assistanceIncompletingthisquestionnaire. PartI GeneralWork History 1. Have you ever worked inthe Chemical Division? A. Yes If'yes',how many years did you work there? S. No What year didyou startworking inthe Chemical Division? Part 11 Tobacco Smoking 2. Do you smoke cigarettesnow? (as ofone month ago) A. Yes B. No 3. Have you smoked at least100 cigarettesduringyour entirelife? A. Yes B. No Of 'no',please go toquestion25) 4. How oldwere you when you firssttartedregularcigarettesmoking: 5. How many years have you smoked rigarettes?(Do not countthe timewhen you periodically stopped smoking) S. On average,how many cigarettesdo you smoke per day now? 7. On average,theentiretime you smoked, how many cigarettedsidyou smoke per day? 8. Ifyou stopped smoking cigarettescompletely,how oldwere you when you stopped? t Alcohol 9. Do you now, ordid you ever drinkatleast12 alcoholicbeverages ina one year periodof time? (1 alcoholbeverage = I beer,I glassofwine;or 1 shot ofhard liquor) A. Yes B. No Of no, please go to PartVI) 10. Age when you firstbegan drinkingalcoholicbeverages: 11. Do you rurrentlydrinkalcoholicbeverages? A. Yes B. No If'no',how longago did you stop (months,years) 12. When you were drinking alcoholicbeverages how drinksper week didyou drink? 13. On the average. how many alcoholicbeverages per week do you currentlydrink.? 3M Company EPI-0003 Page 33 of 42 Part III Directions: The next setof questionsare about your health.The informationyou providewillbe kept strictcloynfidentialW.e would liketo know ifa doctoror otherhealthcare professionalhas ever toldyou thatyou have any of the medical problems listedbelow. We need to know the diagnosis.the year you were firsdtiagnosed and the name of the doctorwho saw you forthismedical problem. Please mark each question. A. Cancer Type ofCancer Bone Yes No IfYes, year you were firsdtiaonosed Name and address of Doctor Thyroid Lymphoma (NonHodgkiffs) Hodgkins Leukemia Muftip(eMyeloma Lung Colon Pancreas Liver Kidney Bladder Testicular Pmstate Bri3ast Ovarian Uterine Other Cancers Identiftyhe cancer types B. LiverDisease Diagnosis Yes No HepatitisA (Infectiouhsepatitis) HepatitisB (Serum hepatitis) Hepatitiscaused by medications Cirrhosisofthe liver if.Yes, year you were firsdtiagnosed AbnorTnallivertests Other liverconditions Identifoytherliver conditions C. Problems with lmmu Diagnosis Low gamma globulin 0 syswm or Endocrine (Mormons) 3 Yes No IfYes, year you were firstdiagnosed Autoimmune disorders (Rheumatoid arthritis, lupus,scieroderma Hyperthmidism (overactivtehyroid) Hypothyroidism (underactivtehyroid) Benign pmstatic hypertrophy(prostatic enlargemen Osteopamsis Diat>etes Other conditions lde(itioftyherconditions 3M Company EPI-0003 Page 34 of 42 Name and address of Doctor who diaonosed you withthiscondition m Name and address ofdoctorwho diagnosed you withthiscondition For Nursing use only: Height in inches, without shoes Weight in pounds Blood Pressure Please staplepulmonary functiontestresultstothe back of thispage. Urinalysis Blood Sugar Albumin APPENDIKC i 3M Coinpany EPI-"3 Page 35 of 42 i f CONSENT FORM 3M Company EPI-OW3 Page 36 of42 Titleof Studv Protocol An EpidemiologicInvestigatioonf Cholecystokiriiand HepaticFunction amonc,PerfluorooctanoiAccid ProductionWorkers Introduction You and your colleaguesinvolvedinfluorochemicalproductionatConage Grove arebeing invited toparticipatiena researchstudy.Pleasereview thisconsentform carefuuyand be sureyour questionsare answered beforeyou make a decisionto participateT.ne nurse and/orphysiciancan provideyou additionailnformationatthetimeof yourmedicalsurveillanceexaminatiorl Puri)oseof Studv The purposeof thisstudyistoconduct theroutinemedicalsurveillanceexam d= isofferedtoyou every two years.In additiontothisexam we willalsocheckyour blood cholecystokini(nCCK') and FC-143 levels.CCK isa hormone thatisnecessaryinpancreaticenzyme production.The informationgL-ainedfrom thisstudy willfurtherhelpus understandany human exposuretoFC-143. Studv Procedures Ifyou choose toparticipatwee willbe askingthatyou filoluttheHealthQuestionnaireand provide blood samples forlaboratorytestiiiszT.he exam willconsistof a medicalsurveiuanceform,blood pressure,height.weight and lung functiontesting.Laboratorytestingwillincludehematology: complete blood count and differentiapll,atelectount-a,ndblood chemistriest:otalH,DL and LDL, cholesterolt,riglyceridegsl,ucose (bloodsugar),liverand kidney functiontestsand a pancrea.@,, functiontestI-n additionwe willmeasure forFC-143 inyourblood. The blood can be drawn with one needlestickand willrequireapproximately4 tubes.Please remernbertofastfor at least 12 hours prior totherest.Also,please do not use,ifatallpossible,a" antacidmedicationsfor 48 hours prior totherest.Schedulingwillbe coordinatedby theMedicalDepar=ent throughthe coordinatorforeach division. Inthe questionnairyeou willbe asked ifyou have specifimcedical diseases.For certainproblems, we need tobe absolutelysurethatyour hstedconditionactuallyexists.For thatmason we may be gettingintouchwith you toask foryour permissiontocontactyour physician.This would be ONLY about themedical conditionof interesttous. ,PotentiaRlisks/Discomforts The onlydiscomfortyou may feelisfrom theneedlestick.You may alsohave some temporary rednessand/orslightswellingin thisareaafterthe blood collectiorl Benerits There willbe no directbenefitfrom your participatiointhisstudy.However, theinformation gainedfrom thisstudywillfartherhelp us understandany human exposuretoFC-143. Your individualresultsaswellasthe overau resultosfthisstudywillbe communicated toyou. 3M Cornpany EPI-0003 Page 37 of 42 Comr)ensation/Medical Treatment From Related Illnessor Iniurv Ifyou sufferinjuryor a medical conditionthatappears to be the resultof participarinein this study,you willbe directetdo the3M Medical Deparanent forobservationand dia.9-nosiAst.your requestorattherecommendation of the3M Medical Department,you willbe referretdo another healthcare professionaaltno costto you. In the eventof a researchrelatedinjur@,c.ompensation wW be determinedon a case by casebasisby 3M. The contactformedical compensation isJeff Mandel, M.D., 733-8670 of the3M Medical Deparanent. Confidentialitv The informationyou provideon thequestionnaireo,rany informationprovidedtous.willbe kept strictlcyonfidentiaalnd willbe used forgroup analysisonly.'Me datacollectedinthisstudymay be used inpublicationosrpublicpresentationsY.ou name wiU notbe revealedinany publication orotherdocuments intendedforpublicatioenxaminatiorl a You willbe givenfeedbackabout your individualresultasshas been done inthepasl The periodichealthexaminationisnot a substitutfeoran examinationby your doctor.We encourage you tosharetheresultosfyour testswithyour privatedoctor. SubjectRi2hts/AvailabiliotfvInformation Ifyou have any questionsaboutthestudynow, or latero,r intheeventof a researchrelatedinjury or emergency, contactDr.JeffMandel (733-8670)orJean Burris,R.N. (737-7867).For answers to questionsaboutyour rightsinregardtothisresearch,you may contactDr.Larry R. Zobel, Chair,3M InstitutionRaelview Board at733-5181. VoluntaryParticipatioannd Withdrawal Participatioinnthisstudyisvoluntary.Refusaltoparticipatweillinvolveno penaltyor lossof benefitstowhich you areotherwiseentitledY.ou arefreeto withdraw atany time forany reasorl Your decisiontoparticipatoer withdraw from thisstudywillnot affectyour work statusor performanceappraisalsT.he investigatomray stopyour participatioin thisstudyatany time ifit isdeterminedthatyourcontinuedparticipatiownould be detrimentaltoyour healthor ifthestudy objectivesarechanged. Subject Consent By sigriintghisconsentform,IcertiftyhatIam atleast18 yearsold. Iconfirm thatIhave read @Ws consentform, and thatI have been given adequate oppomuiity to ask any questionsI may have aboutthisconsentform or abouttheStudy. I alsoconfum thatIunderstandthescope of my participatioinntl-@Sstudy,and thatallof my questionshave been answered to my satisfactionI. am signingthisconsentform voluntarilya,nd IdesiretoparticipatienthisStudy. Iunderstandthat Iam notwaivina orreleasin2any ofmy legalrightsby signingthisconsentform, orby participatioin thisStudy.I understandthatI willreceivea copy of thissignedconsentform. Signatureand Date 3M Emi)loveeNumtxr REFERENCES 3M Company EPI-0003 Page 38 of42 AbbnlzzeseJL,GholsonCF, DaughertyK, LarsonE, DuBrow R, BerlinR, Lp-vinB. A pilotclinicatlriaolf thecholecystokinirneceptorantagonisMtK-329 inpatients,%;ith advanced pancreaticancer.Pancreas7:165-171. AdrianTE, PermetJ,Wang Q, HerringtonMK, TakahashiT, LarssonJ,Pour PM (1994). IsCCK involvedinpancreatidcuctalcellcancer?Tenth InternationSaylmposium on GastrointestinHaolrmones, SantaBarbara,CA. AndersonKE, PotterJD,Mack TM (1996).Pancreaticancer.(In)SchottenfelDd, FraumeniJF (eds)"CancerEpidemiologyand Prevention(Z'd edition).N"ew York: Oxford UniversitPyress,pp 725-771. Axelson J,Ihse1,Hakanson R (1992).Pancreaticancer:theroleofcholecystokinin? Scand J Gastroentero2l7:993-998. BiegelLB, Liu RCM, HurttME, Cook JC. Effectsofammonium perfluorooctanoaotne Leydigcellfunctioni:nvitroi,nvivo,and ex vivostudies.ToxicolAppl Pharmacol 1995;134:18-25. Bryce H. Industrialnd UtilitariAasnpectsof FluorineChemistry.In:Simons J,ed. FluorineChen-dstryN.ew York:Academic Press,1964:297-492. CaldasC, Kem SE (1995).K-rasmutationand pancreatiacdenocarcinoma.IntJ Pancreatol18:1-6. Chey V*rY(1991).Regulationofpancreatiecxocrinesecretion.IntJ Pancreatol9:7-20. Cook JC,Murray SM, Frame SR, HurttME. Inductionof Leydig celladenomas by anunonium perfluorooctanoataep:ossiblendocrine-relatmeedchanism. ToxicolAppl Pharmacol 1992;113:209-217. took JC,HurttME, Frame SR, BiegelLB. Mechanisms of extrahepatitcumor induction by peroxisomeproliferatoirnsCrl:CD BR (CD) ratsT.oxicologis1t994;14:301(abstract). Ermak TH, GrendeB JH (19??).The pancreas:anatomy,histologye,mbryology,and developmentalanomalies.(In)SleisengeMr, FordtranJS (eds)GastrointestinDailseases, Philadelphia:W.BS.aundersCo.,pp 1573-1584. GasslanderT, AxelsonJ,Hakanson R, IhseI,LiliaI,RehfeldJF. Cholecystokiniins responsiblfeorgrowthof thepancreasafterpancreaticobiliadriyversioninrats.Scand J Gastroentero2l5:1060-1065. Gavin CE, MartinNP, SchlosserMJ (1996).AbsenceofspecifiCcCK-A human pancreaticmembranes. Toxicologist30-334. 3M Cornpany EPI-0003 Page39of42 bindingsiteson Gavin CE, Malnoske JA, W'hiteJ,SchlosserMJ (1997). Speciesdifferencesin expression of pancreaticchloecystokinin-Areceptors.Toxicologist36:1180 (abstract). GrffithFD, Long JE. Animal toxicitystudieswith ammonium perfluorooctanoateA.m Ind Hyg Assoc J 1980;41:576-583. Guy W. Fluorocompounds of human plasma: analysis,prevalence,purificationa,nd characterizationR.ochester,NY: Universityof Rochester,(Ph.D.disseration)1,972. HanhijarviH, Phaug R, SingerL. The sex-relatedifferenceinperfluorooctanoate excretionin therat Proc Soc Exp Biol Med 1982;171:51-55. HanhijarviH, Ylinen M, Kojo A, Kosma V34. Eliminationand toxicitoyf perfluorooctanoicacidduringsubchronic administrationin theWistar raL Pharrnacol Toxicol 1987;61:66-68. Howatson AG, CarterDC. Pancreaticcarcinogenesis- enhancement of cholecystokininin thehamster-nitrosaminemodel. Br J Cancer 51:1-7-114. HerringtonMK, Permert J,Kazakoff KR, Zucker KA, BilchikAJ, Pour PM, Adrian TE. Effectsof raw soya dietand cholecystold@ receptorblockade on pancreaticgrowth and tumor initiatioinn thehamster. Cacner Lett82:7-16. HerringtonMK, Adrian TE (1995). On theroleof cholecystoldnininpancreaticcancer. IntJ Pancreatol 17:121-138. Ikeda T, Aiba K, Fukuda K, Tanaka M. The inductionof peroxisome proliferatioin rat liverby perfluorinatefdattyacids,metabolicallyinertderivativesof fattyacids. I Biochem 1985;98-475-482. qohnson FE, LaRegina MC, Martin SA, BshitiHM (1983).Cholecystokinininhibits pancreaticand hepaticcareinogenesis.Cancer Detect Prev 6:389-402. Kennedy G. Dermal toxicityof anunonium perfluorooctanoateT.oxicol Appl Pharmacol 1985;81:348-355. Kennedy G, Hall G, BritLellJi,Chen H.. Inhalationtoxicityof an-imonium perfluorooctanoate.Fd Chem Toxical 1986-,24:1325-1329. KuslikisBI, Vanden Heuvel JP,Peterson RE. Lack of evidence forperfluorodecanoyl-or perfluorooctanoyl-coenzymeA formationin male and female rats. J Biochem Toxicol 1992;7:25-29. 3M Cornpany EPI-OW3 Page 40 of42 L.ernbergerT, Desvergne B, Wahli W. Peroxisome proliferator-activatredceptors:a nuclearreceptorsignalingpathway in lipidphysiology. Ann Rev CellDev Biol 1996;12:3^:@5-363. Liu RCM, Hurtt ME, Cook JC, Biegel LB. Effect of the peroxisome proliferator, arnmoniurn perfluorooctanoate (C8) on hepatic aromatase activityin adult male CRI:CD BR (CD) rats. Fund Appl Toxicol 1996;30:220-228. Longnecker DS (1986).Experimentalmodels ofexocrinepancreatictamors. (In)Go VLW (ed)The Exocrine Pancreas:Biology,Pathology and Diseases,New York-:Raven, pp 443-458. Longnecker DS (1990). Experimental pancreaticcancer:roleof species,sex and diet. Bull Cancer 77:27-37. Longneeker DS (1991). Hormones and pancreaticcancer. IntJ Pancreatol9:81-96. Lu L, Louie D, Owyang C (1989). A cholecystokininreleasingpeptidemediates feedback regulationof pancreaticsecretion.Am J Physiol256:G430-G435. Mainz DL, Black 0, Webster PD (1973). Hormonal controlof pancreaticgrowth. J Clin Invest52:2300-2306. McGuinness EE, Morgan GH, Levison DA, Frape DL, Hopwood D, Worrnsley KG (1980). The effectsof long-termfeedingof soya flouron theratpancreas.Scand J Gastroenterol15:407-502. Nickles M. Presence du fluordans lasang. Compt Rend 1856;43:885. Oboum JD, Frame SR, ElliotGS, Cook JC (1997a). Pancreaticoncogenic effectsof Wyeth 14,643. Toxicologist36:1181 (abstract). (Oboum JD, Frame SR, ElliotGS, Cook JC (1997b). Pancreaticoncogenic effectsof Wyeth 14,643. Haskell Laboratory forToxicology and IndustriaMledicine, E.I.du Pont de Nemours & Company, (unpublishedmanuscript). Olsen GW, GillilandFD, Burlew MM, BurrisJM, Mandel JS,Mandel JH. An epidemiologic investigatioonf reproductivehormones in men with occupa6onal exposure to perfluorooctanoiaccid.(Submittedforpublication.) Ophaug R, SingerL. Metabolic handlingof perfluorooctanoicacidin rats. Proc Soc ExpBiolMed 1980;163:19-23. 3M Cornpany EPI-0003 Page41of42 Palmer-SmiJt,KhramerST,SolomonTE (1991)C.CK stimulagtreoswthofsixhuman pancreaticcancer celllinesin serum-freemedium. Regulatory Peptides32:341-349. Pandol SJ. Pancreatipchysiology. (In)SleisengerM, Ford= Diseases, Philadelphia:W.B.Saunders Co.,pp 1585-1600. JS (eds)Gastrointestinal Pastoor T?, Lee KP, PerriMA, GiRies PJ. Biochemical and morphological studiesof anunonium perfluorooctanoate-inducehdepatomegaly and peroxisome proliferationE.xp Mol Pathol 1987;47:98-109. PetersenH, Solomon TE, Grossman MI. Effectof chronicpetnagastrinc,holecystokinin and secretinon pancreas of rats.Am J Physiol 234:E283-286. Poston GJ, GillspieJ,GuiUou PJ (1991). Biology of pancreaticcancer. Gut 32:800-812. Pour P, Kruger FW, AlthoffJ,Cardesa A, Mohr U (1975).A new approach for induction of pancreaticneoplasms. Cancer Res 35:2259-2268. Pour PM, Runge RG, BirtD, GingellR, Lawson T, Nagel D, Wocave L, Salmasi S (198 1).Current knowledge of pancreaticcarcinogenesisin the hamster and itsrelevance to the human diseaseCancer 47:1573-1587. Pour PM, Lawson T, Helgeson S, Donnelly T, Stepan K (1988). Effectof cholecystokinin on pancreaticcarcinogenesisin the hamster model. Carcinogenesis 9:597-601. Rao MS, Reddy JK. Hepatocarcinogenesisof peroxisome proliferators.Ann NY Acad Sci 1996;804:573-587. Rehfeld JF,van SolingeWW (1994). The tumor biology of gastrinand cholecystokinin. Adv Cancer Res 63:295-347. SAS InstituteI,nc. SAS Users Guide: StatisticsV.ersion 6. Cary, NC: SAS Institute, Inc.,1990. f Satake K, Mukai R, Kato Y, Umeyama K. Effectsof ceruleinon thenormal pancreas and on experimentalpancreaticcarcinoma in the Syrian golden hamster. Pancreas 1:246-253. SibinsldLJ. Two-year oral(diet)toxicity/carcinogenicistyudy of fluorochen-dcaFlC-143 in rats. St.Paul,MN:Riker Laboratories,1987. Singh P, Townsend CM, Upp J,LaridjaniA, Thompson JC. Characterizationof cholecystoldninreceptors(CCK-r) in normal and cancerous human pancreas. Fed Proc 45:291. 3M Cornpany F-PI-0003 Page42of42 SrrdthJP,LiuG, SoundararajaVn, McLaughlinPi,Zagon IS. Identificaitaonnd characterizatioonf CCK-B/gastin receptorsin human pancreaticcancerceU lines.Am J Physiol 266:R277-R283. StewartID, Flaks B, Watanapa P, Davies PW, Wflliamson RC. Pancreatobdiarydiversion enhances experimentalpancreaticcarcinogenesis.Br J Cancer 63:63-66. Tak iad VD, Fortune KP, PolloDA, Shah GN, Wank WA, Gardner JD (1994). Direct demonstration of threedifferentstatesof thepancreaticcholecystokininreceptor. Proc Natl Acad Sci 91:1868-1872. Taves D. Evidence thatthereare two fon-nsof fluoridein human serum. Nature 1968:217:1050-1051. Taves D. Electrophoretimcobilityof serum fluoride.Nature 1968;220:582-583. Vanden Heuvel J,Kushlds B, Van Refelghem M, PetersonR. Tissue distribution, metabolism and eliminationof perfluorooctanoicacid. J Biochem Toxicol 199 1;6:83-92. van Kranen HJ, Vermeulen E, Schoren L, Bas J,Woutersen RA, van IerselP, van Kreijl CF, SchererE (1991). Activationof c-K-ras isfi-equenitn pancreaticcarcinomas of Syrian hamsters,but isabsentin pancreatictumors of rats.Carcinogenesis 12:147-1482. Wank SA, PisegnaJR, deweerth A (1994).Cholecystokininreceptorfamily.Ann NY Acad Sci 713:49-66. Ylinen M, Koho A, HanhijarviH, Peura P. Dispositionof perfluorooctanoicacidin the rataftersingleand subchronicadministration.BuU Environ Contam Toxicol 1990;44:4653. Zucker KA, Adrian TE, BilchikAJ, Modlin IM. Effectsof the CCK receptorantagonist L364,718 on pancreaticgrowth in adultand developing animals. Am J Physiol 257:G51 1-G516. 3M Company EPl-M3 PageIof25 FINAL REPORT Epidemiology MedicalDepartment 3M Company 220-3W-05 St.Paul,MN 55144 Date: September4,1998 Title:An Epidem*,.-)IoIgnivcestigationf Plasma Cholecystokiaiannd HepaticFunctioniti PerfluorooctanoiAccid ProductionWorkers Study StartDate: September 3, 1997 IRB ApprovalDate: September3, 1997 ProtocolNumber: EPI-0003 IRB Approval Exempt Expedited x PrincipaIlnvestigator: Co-investigators: Geary W. Olsen,DVK PhD' JeanBurris,RN, MPH' Mchele M. Burlew@MS' JeffreHy. Mandel,MD, MPH' Study Director: JeffreHy. Mandel,MD, MPH' 1. OccupationalMedicine,3M Company, 220-3W-05, St.Paul,MN 55115 ABSTP,ACT 3M Company EPI-0003 Page2of25 Perfluorooctanoicacid(PFOA) isa peroxisome proliferatowrhich increasedthe incidenceof pancreas acinarcelladenomas inrats.Recent researchsuggested thatthese tumors may be the consequence of a mild but sustainedincreasein cholecystokinin(CCK) as a consequence of hepaticcholestasisI.n addition,an epideri,.ioloignivcestigationhad suggested thatPFOA may modulate hepaticresponses to obesity and alcoholconsumptiop in theseproduction workers. To furtherassessthesehypotheses,we conducted three cross-sectionalanalysesof the employees' serum PFOA levelsand medical surveillancedata collectedin 1993 (n = 111),1995 (n= 80) and 1997 (n= 74). Plasma CCK was onlymeasured in1997. Serum PFOA was measured by mass spectrophotometry methods and plasma CCK was assayed by radioimmunoassay. Mean serum PFOA levels,by year,were: 1993, mean = 5.0 ppm (range 0.0 - 80.0 ppm); 1995, mean 6.8ppm (range0.0 - 114.1 ppm); and 1997, mean = 6.4ppm (range0.1 - 81.3 ppm). CCK values(mean = 28.5 pglml, range 8.8-86.7pg/nil) approximated the assay'sreferencerange (up to 80 pg/ml) fora 12 hour fast. Employees' serum PFOA levelswere not positivelayssociatedwith eitherclinicahlepatictoxicityas measured by various serum liverenzyme tests,cholestasisor elevatedplasma CCK levels. Nor did serum PFOA levelsmodulate hepaticresponses(e.g.,liverenzymes and high densitylipoprotein)to obesityand alcohol,respectively. RNTRODUCTION 3M Company EPl-M3 Page3 of25 Perfluorocarbons are structurallaynalogous to hydrocarbons, except the hydrogens are replacedby fluorine[Bryce, 1964) and may containother elements such as oxygen, nitrogenand sunr. Ammonium perfluorooctanoateisa potentsyntheticsurfactantused in industriaalpplicationswhich rapidlydissociatesin aqueous solutionto perfluorooctanoic acid (PFOA, C-IFIsCO2H). i In laboratoryanimals,PFOA and itssaltsare:1) absorbed by ingestion,inhalation or dermal;2) not metabolized;3) distributepdrimarilyinthe plasma and liverof male rats and theEver,plasma and kidney infemale rats;and 4) eliminatedinthe male rat viafeces and urinewhereas inthe female ratthereisa greaterrateinrenalexcretion[Griffitahnd Long, 1980; Ophaug and Singer,1980; Hanhijarviet al.,1982; 1987; Justet al.,1989; Kennedy 1985; Kennedy et al.,1986; Ylinen etal.,1990; Vanden Heuvel et al.,199 1]. In rats,PFOA resultsinperoxisome proliferationu,ncoupling of mitochondrialoxidative phosphorylation,alteredlipidmetabofisrr@hypolipidemiaand an increasedincidenceof liver,Leydig celland pancreas acinarcelladenomas [Griffitahnd Long 1980; Kennedy, @985; Kennedy et al.,1986; Sibinski1987; Haughom and Spydevold, 1992; Kelleret al., 1992; Cook etal.,1992; 1994].The inductionof these tumors most likelyoccurs via nongenotoxic mechanisms because PFOA isnot mutagenic [Griffitahnd Long, 1980; Biegel et al.,1995). Causal mechanisms may includethe roleof oxidativestressin the livertumors and increasedestradiolevels,via inductionof hepaticaromatase activityi,n the development of Leydig celltumors [Cook et al,1992, 1994-,Rao and Reddy, 1996]. The pancreas acinaradenomas were hypothesized to be a resultof a mild but sustained 3M Company EPI-0003 Page4 of25 increaisnecholecystok(iCnCiKn)levelssecondartyohepaticcholesta(sOibsoumetal., 1997]. CCK isreleasedftom the"Nf'cellsintheduodenalmucosa inresponseto the presenceof food,bindsto receptorson thepancreasacinarcellsand subsequently stimulatetshereleaseof pancreaticenzymes intotheduodenum [Pandol,1998]. Itis controllebdy a negativefeedbackcycleinvolvingmonitorpeptideand trypsin.CCK has been shown, insome animalmodels,toproduce pancreatichypertrophy,hyperplasiand neoplasia[Longnecker,1986; 1990;1991Pour etal,1981;1988]. Hepatictoxicityh,ypofipidernainad abnormal hormone levelshave notbeen observedinPFOA productionworkers [Ubel etal,1980;GilMand and Mandel, 1996 Olsen etal.,1998]. GilWand and Mandel [1996]didreportthatPFOA may negatively modulatetheeffecatlcoholhason highdensitylipoprotei(nHDL) levelsand exacerbate the effecthatobesityhas on liverenzyme testsH.owever, thisworkforcewas notfound to be atan increasedmortalityriskforEver canceror liverdisease[GUWand and Mandel, 1993]. There were 4 pancreaticancerdeathscompared to2 expected (Standardized MortalityRatio1.96,95% ConfidenceInterva0l.53-5.01).One ofthesefourpancreatic cancerdeathshad worked inthebuddingwhere PFOA isproduced atthischemicalplant. The purposeof thisepiden-dologincvestigatiwoans to re-examinetheworkforce inthisPFOA productionplantinorderto detern-dne1:)whether CCK levelsarepositively associatedwithserum PFOA levelsamong productionemployees;and 2) whetherPFOA may modulatehepaticresponsesto obesityand alcohol. NIETHODS 3M Company EPI-OW3 Page5of25 PFOA Production PFOA productionat tWs 3M plantbegan in 1947. PFOA, a white powder, is produced by an electrochemicalprocess [Bryce, 1954).Production involvesa four-stage process:isolatingand convertingthechemical to a saltslurry,convertingthe slurryto a saltcake,dryingthe cake, and packaging.The greatestlikelihoodforexposure to PFOA occurred inthe dryingarea althoughjob historywas not predictiveof totalserum tl@iorinleevels(a surrogatefor ser-umPFOA) [CyilWandand Mandel, 1996]. Subject Selectionand Data Coflection Voluntary medical surveillanceexaminationswere offeredbienniauy(1993, 1995 and 1997) to the fluorochemicalproductionworkers. The totalnumber of subjects,by year,who participateidn thesethreecross-sectionailnvestigationwsere: 1993 (n = 111); 1995 (n = 80);and 1997 (n = 74). Eligiblveoluntaryparticipatiornatesamong these production workers approximated 70 percent.There were 68 subjectsin common for 1993 and 1995; 20 subjectsin common between 1993 and 1997 Oower number due to employee turnover and re-assigninentsa)n,d 17 subjectsin common for allthree years. Surveillanceactivitieisncludeda self-administereqduestionnaire,measurement of height, weight and pulmonary function,standardbiochemicaland urinalysitsests,PFOA determination and severalmale reproductivehormone assays. The hormone data were collectedonly in 1993 and 1995 and resultshave been reportedelsewhere [Olsen et al., 1998]. Serum biochemicaltestsincluded:alkalinephosphatase,ganuna glutamyl 3M Company EPI-OW3 Page6 of25 transfer(aGsGeT),serumglutamoyxlaloacettriacnsamin(aSsGeOT),serumglutamyl pyruvictransaminas(eSGPT), totalbilirubidni,rectbilirubicnh,olesterolo,w-density lipoprotein(sLDL),high-densitlyipoprotei(nHsDL), triglyceridbelso,odureanitrogen (BUN), creatininaend glucose.Hematology testsincludedh:ematocrith,emoglobir@red blood ceus(RBC), plateletasnd whiteblood cell(sVrBC). In 1997,employees'plasma CCK-33 levelswere determined.CCK existsinvariousforms and lengthsalthough sulfateCdCK-33 Ci.e.a,33 amino acidarrangement)appearsto be thepredominantform. Employees were requiredto have fastedfor 12 hourspriortotheirvenipuncture.One employee serlf-reporttehdathe didnotfastand thushe was excludedfrom thestudy.His CCK levelwas 123 pg/dI.Thisexclusionleft74 employeesavailablfeoranalysiisn 1997. Serum chemistrieasnd hematologywere evaluatedatUnited Hospitals(St.Paul, NEimesota).Plasma CCK-33 was measured by directradioimmunoassayby InterScience Institut(eInglewood,California)S.erum PFOA was deten-ninebdy thertnospra(y1993 and 1995) and electrospra(y1997)highperformanceliquicdhromatographymass spectrometrymethods (Johnsonetal.,1996;Advanced BioanalyticaSlervicesInc.,1997]. fdata Analysis Simple and stratifiaendalysisP,earsoncorrelatiocnoefficientasn,alysisofvariance (ANOVA), and ordinarymultivariartegressiownere usedto evaluatelineaarnd nonlinear associationbsetween PFOA and thebiochemicalparameterswith adjustmentforpotential confoundingvariable[sSAS, 1990].For stratifiaendalysese,mployeeswere dividedinto fourPFOA categories0:-<1 ppm, I-<10 ppni,10 -<30 pprria,nd >30 ppm inorderto determineifan effectexistedatthehighestserum levels.These categoriehsad been 3M Company EPI-0003 Page7of25 previousluysedtoexamineassociatiobnestweenmalereproductivheormonesand PFOA among these workers in 1993 and 1995 [Olsen et al.,1998]. For multivariablreegression evaluationP,FOA, age, body mass index (BNH), alcoholuse,and cigaretteuse were examined as both categoricaland continuous variables.Alcohol use was analyzed as less than I drinkper day, @:ldhnk per day (with almost allsubjectsbetween 1-3 dfinks/day), and non-response to the questionnaireitem.Linear and nonlineartransformationsof PFOA were used to testforassociations.In particulart,he multivatiablmeodels empILyed by GiUilandand Mandel [1996] were re-examined to determinewhether PFOA has a modulating effecton obesityand alcoholconsumption inregardsto hepaticserum chemistries(SGOT and SGPT) and HDL, respectively. RESULTS Mean serum PFOA levels,by year,were: 1993, mean = 5,0 ppm (SD = 12.3, range 0.0 - 80.0 ppm); 1995, mean 6.8ppm (SD = 16.0,range 0.0 - 114.1 ppm); and 1997,mean = 6.4 ppm (SD = 14.3,range 0.1 - 81.3 ppm). In 1997, the mean CCK value was 28.5 pg/n-d(SD = 17.1 pg/ml,range 8.8-86.7pg/rnl).Allbut two CCK valueswere ,ia,ithtihne assay'sreferencerange (up to 80 pg/ml). These two CCK values(80.5 pg/n-d and 86.7 pg/ml) were from employees with 0.6 ppm and 5.6ppm serum PFOA levels, respectively. Serum PFOA levelswere not consistentlycorrelatedwith any of the potential confounding variables,serum chemistriesor hematologicalparameters. The Pearson correlationcoefficient(sinparentheses)between PFOA and the variablesfor 1993, 1995 and 1997 respectivelyw,ere: age ( -.22,-.14,.02);alcohol(.10, .18,.01),BNE (.10,.10, -,01),cigarette(s.07,.Il, -.02)a,lkaline phosphatase (.11, .14, -.07), SGOT 3M Company EPI-0003 Page 8 of 25 (,12, -.01, .02), SGPT (.10, .04, .14), GGT (.07, -.01, -.05,),totalbillrubi(n-.0'--!.1,4, -.08),direct bilirub(i.n01,-.32,-.04c)h,olester(o.l15,.14,.18),LDL (-.01-,.07,.11),HDL (-.11, -.19,.03)triglycerid(e.1s7,.37,.11),glucose(-.08,.04,-.04),BLTN (-.12,-.11,.05), creatinin(e.07,.17,-.01),hematocri(t.22,.08,-.10),hemoglobin(.22,.11,-.11)R,BC (.09,-.01-,.19)p,latelet(s-.10.,04,.11)and WBC (-.01.,06,-.01).In 1997,thePearson correlatiocnoefficiefnotrPFOA and CCK was -.20(p=.09). TableIprovidesthemear4standardeviatioannd rangeofthepotential confounders,serum chen-@striaensd hematologiesby fourlevelsofPFOA categorization (0-<l,1-<10,10-<30,and >30 ppm) forthethreeyears(1993,1995 and 1997) ofmedical surveillanceexaminationsT.he mean of thePFOA ppm categoriedsfferedsignificantly with eachotherand therewere two ordersof magnitudedifferencbeetween thelowest and highestPFOA categorieisneachyear.There were no statisticasniygnifica(npt< .05) F valuesforany clinicaclhen-dstrtyestor hematologicalparameterexamined forany of thethreesurveillancyears.Itshouldbe noted thatthemean CCK valueswere 50 percent lower arnongemployeeswith serum PFOA values> 10 ppm. FigureI isa scatterplotf therelatiobnetween CCK (transformevdianaturalog) and PFOA. The linearegressioenquationwas: InCCK = 3.3-0.008PFOA (p valueof PFOA coefficie=nt.07;? of model = .04).We didnot observeany significadnitfferences inmean serum chemistryvaluesforthoseemployees withhighCCK values(e.g.@,!40) compared to thosewithlower CCK valuesor forthosesubjects.kithighPFOA values (e.g.@, 10 ppm) compared to thosewithlowervalues.Use ofmultivariabrlegression 3M Company EPI-0003 Page9 of25 models(datanotshown) continuedtoindicataeweak negativeassociatiobnetweenCCK and PFOA adjustingforpotentialconfounding variables(e.g.,age, body mass index, alcohol,cigarettesand clinicaclhemistrymeasures of hepaticfunction). Based on the multivariablmeodel used by Gill@andand Mandel [1996], Table II provides the change in HDL levelsassociatedwith a 10 ppm increaseinserum PFOA levelsamong moderate drinkers(> 1 drirwday) compared to Ught drinkers(< I drink-/day).Included inTable H isthechange originallryeportedby GiUilandand Mandel [1996] inthisworkforce with their1990 surveiuancedata. It shouldbe noted,however, thattheir1990 model was based on totalserum organic fluorinemeasurement-; ratherthan serum PFOA levels. Unlike 1990, therewas not a substantiamlodulation inHDL levels with increasedPFOA serum levelsamong moderate drinkers. Likewise, Table IIIpresentstheresultsof multivariablaenalyses,includingthose originallryeportedusing the 1990 surveillancdeata [GilWand and Mandel, 1996], regardingthe potentialmodulating effectof PFOA on hepaticresponsesto obesityin the three subsequent surveillanceyears.Whereas SGPT levelsincreasedconsiderablywith a ,10 ppm change in totalserum organicfluorinewhen theBNff was > 30, this association was not observed in 1993, 1995 or 1997. DISCUSSIO'.N' We observed a weak negativeassociationbetween serum PFOA and plasma CCK among 74 workers engaged inthe productionof ammonium perfluorooctanoate.This findingwas opposite thathypothesizedbased on the toxicologicaflindingsof Oboum et al [1997] who feddietsto ratscontainingeither0 or 100 ppm of Wyeth- 14,643,a potent 3M Company EPI-NO3 Page10of25 peroxisomeproliferatowrh,ichcausesthesame triadoftumors,includinpgancreasacinar celladenomas, as PFOA. Aftersixmonths, themean pancreatic'A,eightosf the treated ratswere 17 percent above controlanimals(p <.05), mean plasma CCK levelswere 44 percenthigher(p <.05) and markers of cholestasis(totalbileacids,alkalinephosphatase and bdirubin)were also significantellyevated, The clinicaplathology data indicativeof cholestasiswere associatedwith alterationisnbileflow and b@e acidoutput. Oboum et al 1997] had alsoconducted in vitroexperimentsof both Wyeth- 14,643 and PFOA which argued againstother biologicalpathways known to elevateplasma CCK levelsincluding CCKA receptoragonisrn,trypsininhibitioannd increaseddietaryfatcontent. Oboum et al (1997] concluded thatchronicexposure to Wyeth- 14,643 may induce pancreatic adenomas via a mild but sustainedincreasein CCK levelssecondary to hepatic cholestasis. We offerseveralexplanationsforthe lackof a positiveassociationbetween PFOA and CCK inour study. First,theprimary setof biochemicaland ceuulareventsidentified in rodents susceptibleto the hepatocarcinogeniceffectsof peroxsisome proliferatorhsave not been identifieidn eitherliverbiopsiesfrom humans exposed to peroxisome prohferatorsor in in vitrostudieswith human hepatocytes;however, the peroxisome fproliferator-activarteecdeptor(PPAR-a) isexpressedat very low levelsin the human liver [Oboum et a].,1997; Cattleyet al.,1998]. Consequently,an expertpanel has recently opined thatitisunlikelythatperoxisome prodferatorsarecarcinogenicto humans under anticipatedconditionsand levelsof exposure,however, theircarcinogenicpotential cannot be ruledout under extreme conditionsof exposure [Cattleyetat.,1998]. Second, even ifthe mechanism existedinhumans, the serum measurements inthese production workers may have been too low to cause an effect. Third,CCK receptorsappear to be 3M Company EPI-M3 Page11of25 differbeentweetnheratandhuman.Recentstudieisndicathea,utnliktehepancreaosf theratand dog, thehuman pancreashasno detectablCeCKA receptorsand littlteono MRNA forthereceptor[Wank etal.,1994]. Human, cynomologusand rhesusmonkeys lackspecifibcindingsitesfortheselectivCeCKAligand 3[HIL-364,718(Gavinetal., 1996,1997].Because theCCK receptoractivitoyf theratmay be quitedissirr@ltaorthe human, a cynomologus monkey may be a more appropriateanimalmodel to studythe pancreatipcathophysiologyconsequenceof exposureto PFOA inthe-@uman. Fourth, whetherCCK irdtiateosr promotes pancreaticancerisnot a new questionand the researchpublished,to date,remainscontroversia[lAxelsonetal.,1992].Data from more than seventylaboratoryanimalstudieshave variablysuggestedthatCCK has positive trophiceffectsi,nhibitoreyffectso,r no involvementinpancreatitcumor growth [Herringtoannd Adrian,1995]. CCK has promoted growth of human pancreaticcancers incellcultures[Paimer-Smitehtal.,1991]. On theotherhand,fastinpglasma concentrationosf CCK inunresectedpancreaticancerpatientdsidnotdfferfrom healthy controls[Rehfeldetal.,1994). Fiftht,heratmay be an inappropriatmeodel inthestudy of human pancreascarcinogenesisC.arcinogensinratsinduceacinarcellmalignancies (whicharerareinthehuman [Andersonetal.,1996]. Hamster pancreascancermodels are of ductalcelloriginwhich resemblehuman pancreaticancer(adenocarcinomasof the ductules)[Pouret al.,198 1]. Activationofthec-K-rasgene isfrequentinbothhuman and hamsterpancreaticancerbutisnotfound inazaserine-inducepdancreaticancer models intherat(van Kranen etal.,1991-1Caldasand Kem, 1995]. Finallyw,e must ask whethertheweak negativeassociatioonbservedinour studyrepresentasn entirely differenbtiologicarlelationshitphanwhat was originallpyostulatedbasedon thefindings by Oboum etal[1997]. We do notbelieveso because:1)allCCK 3M Company EPI-M3 Page12of25 valuesobservedin thisstudy were withinthe assay'sreferenceexcept fortwo values(which were not associatedwith high serum PFOA values);and 2) therewas no suggestionof cholestasis which was considered the underlyingreason for the elevatedCCK levelsin the rat. We were unable to replicateinthreeseparateyears the originalsuggestionthat PFOA may modulate hepaticresponsesto obesityand alcohol. Severalexplanationsfor the disparatefindingsexist. First,theremay be an associationthatwas not observed by us. In the originalreport[GilMand and Mandel, 1990], totalserum organic fluorinewas used as a surtogatevariablefor PFOA exposure because the assay was lessexpensive and technicauyeasierto perform at thetime.The use of a totalserum organicfluorinemay representother perfluorocarbons,which could be peroxisome profferatorsh;owever, data suggest thatPFOA would representthe greatestfractionof totalserum orgardc fluorine levelsinthisemployee populationRJbel,1980]. Another explanationforthe disparate BNH findingsisthattheremay have been measurement errorregarding body mass index in the originalstudy or in our study. We have previouslynoted thelack of an expected positiveassociationbetween BNff and estradiolin the 1990 data [Olsenet al.,1998]. In fonly one of theyears was therethe expected [Bums et al.,1997] strong positive correlationbetween BNE and SGPT values(1990, r = .20,p = .02;1993, r = .16,p = .09; 1995, r = .13,p = .27; 1997, r= .43,p = .0001). Self-reporteadlcoholdata collectedin the occupationalsettingshould alsobe questionedforitsreliabiliatsywellas validityT.o partiallayddress the issueof reliabilitwye, examined the analysesof the 68 employees who participatebdoth in 1993 and 1995. The data showed good correlationforthe confounding factorsof BNff (r= .94,p = .0001),alcoholconsumption (r= .67,p = .0001) and cigarettsemoking(r=.84,.0001).The few employeesincommon 3M Company EPI-0003 Page13of25 forallthreeyears (n = 17) prevent any conclusionsregardingthe reliabiliotfyself-reporteddata across all threeyears. The trend in the correlationwsas comparable forthe 1993 and 1995 analyses (e.g.,1995/1997 correlationswere BNE: r =.91, p < .0001,alcohol.37,r = .37,p < .15) cigarettesr,= .99,p < .0001). A few additionalissuesneed to be considered in evaluatingthe resultsfrom this study.The cross-sectionadlesigndoes not allow fora directanalysisof thetemporalityof an association.Given thatthehalf-fifoef PFOA isestimatedto be 18 to 24 months [Ubel et al.,1980],itisconceivablethattheremay be some biologicalaccommodation to the effectsofPFOAas suggestedbybiegelet al[1995]. Also,therewere fewer employees analyzed in 1995 and 1997 reducingthe statisticpaolwer of the study although the number of subjectswith serum PFOA measurements @:10 ppm remained comparable. Finally,the issueremains thatthe lackof a clinicahlepatotoxiceffectobserved by GiOand and Mandel [1996] and ourselvesdoes not negate the possibilittyhatPFOA may have a subcmcal effectin thisproductionpopulationthathas yetto be observed. Results from additionallaboratoryanimal studiesmay provide furtherinsight. In conclusion,our data do not suggestthat,at theserum levelsmeasured, PFOA isassociatedwith a n-dldincreaseinplasma CCK levels.Neither the original epidemiologicalfindings[Gillilanadnd Mandel, 1996] or our findingssuggest clinical hepatictoxicityat the PFOA levelsobserved.The factthatwe were unable to demonstrate, on threeseparateoccasions,PFOA modulation of hepaticresponses to obesityand alcohol,leadsus to believethatthis,too, isunlikelyat the serum PFOA levels measured inthisstudy. F,EFERE,NCES 3M Company EPI-0003 Page14of25 Anderson KF-,PotterJD, Mack TM (1996). Pancreaticcancer. (In)SchotterifelDd, Fraumeni JF (eds) "Cancer Epidemiology and Prevention(2 d edition)."New York.. Oxford UniversityPress,pp 725-771. Axelson J,Ihse 1,Hakanson R (1992). Pancreaticcancer: the roleof cholecystokinin? Scand J Gastroenterol27:993-998. BiegelLB, Liu RCK HurttME, Cook JC (1995). Effectsofanimonium perfluorooctanoateon Leydig cellfunction:in vitro,invivo,and ex vivo studies. Toxicol Appl Pharmacol 1995-134:18-25. 1 Bryce H (1964). Industriaalnd UtilitariaknspectsofFluorineChemistry. In:Simons J, ed. FluorineChemistry. New York:Acadeniic Press,pages 297-492. Bums CJ, Boswell JK Olsen GW (1996). Liverenzyme activitaynd body mass index. J Occ Env Med 38-.1248-1252. Caldas C, Kem SE (1995). K-ras mutation and pancreaticadenocarcinoma. IntJ Pancreatol 18:1-6. CattleyRC, DeLuca J,Elcombe C, Fenner-CrispP, Lake BG, Marsman DS, Pastoor TA, Popp JA, Robinson DE, Schwetz B, Tugwood J,WahH W (1998). Do peroxisome proliferatincgompounds pose a hepatocarcinoger@chazard to humans? Reg Toxicol Pharmacol 27:47-60. Cook JC, Murray SK Frame SR, HurttME (1992). Inductionof Leydig c-eUadenomas by anunonium perfluorooctanoate:a possibleendocrine-relatemdechanism. Toxicol Appi Pharmacol 1992;113:209-217. Cook JC, Hurtt NE, Frame SR, BiegetLB (1994). Mechanisms of extrahepatitcumor (inductionby peroxisome proliferatorisnCrl:CD BR (CD) rats.Toxicologist14:301 (abstract). Gavin CE, Martin.,NP,SchlosserNU (1996). Absence of specificCCK-A bindingsiteson human pancreaticmembranes. Toxicologist30:334. Gavin CE, Malnoske JA, V@Mte J,SchlosserMJ (1997). Species differencesinexpression of pancreaticchloecystokinin-Areceptors.Toxicologist36:1180 (abstract). GillilanFdD, Mandel JS (1993). Mortalityamong employees of a perfluorooctanoicacid productionplant.JOM 35:950-954. 3M Company EPI-OW3 Page15of25 GillilanFdD, Mandel JS (1996).Serum perfluorooctanoaiccidand hepaticenzymes, lipoproteinsand cholesterol-a.study of occupationallyexposed men, Am J Ind Med 129-.560-568. GriffithFD, Long JE (1980). Animal toxicitystudieswith animonium perfluorooctanoate. Am Ind Hyg Assoc J 41:576-583. HanhijarviK Phaug R, Singer L (1982). The sex-relatedifferencein perfluorooctanoate excretionin therat.Proc Soc Exp Biot Med; 171:51-55. HanhijarviI-LYlinenNLKojoA,KosmaVM(1987). Elin-dnatioand toxicitoyf perfluorooctanoicacidduring subchronicadministrationintheWistar rat. Pharmacol Toxicol 61:66-68. Herrington NIFC,Adrian TF, (1995). On the roleof cholecystokininin pancreaticcancer. Int J Pancreatol 17:121-13 8., Haughom B, Spydevold 0 (1992). The mechanism underlyingthehypolipemiceffectof perfluoocatnoicacid (PFOA), perfluorooctanesulphonicacid (PFOSA) and clofibriaccid. Biochim Biophys Acta 1128:65-72. Johnson JD, Wolter JT, ColaizyGE, RethwiU PA, Nelson RM (1996). Quantificoaftion perfluorooctanoataend perfluorooctanesulfonaitnehuman serum usingion-paierxtraction and highperformanceliquidchromatography-thermospramyass spectrometrywith authomatedsamplepreparation.3M EnvironmentalLaboartoryReport.St.Paul:3M Company. JustWW, Gorgas I,,H,artlFU, Heinemann P, SalzerK SchimassekH (1989). Biochemicaleffectasnd zonalheterogeneitoyf peroxisomeproliferatiionnducedby perfluorocarboxyhaccidsinratEver.Hepatology 9:570-581. KeflerB, Marsman D, Popp J,Thurman R (1992).Severalnongenotoxiccarcinogens /uncouplemitochnodriaol)ddativpehosphorylationB.iochimBiophysActall02:237244. Kennedy G. Dermal toxicitoyf ammonium perfluorooctanoatTeo.xicolAppl Pharrnacol 1985;81:348-355. Kennedy G, HallG, BrittelJh,Chen H. (1986).Inhalatiotnoxicitoyfammonium perfluorooctanoateF.d Chem Toxicol24:1325-1329. LongneckerDS (1986).Experimentamlodelsofexocrinepancreatitcumors.(In)Go VLW (ed)The ExocrinePancreas:Biology,Pathologyand DiseasesN,ew York: Raven, pp 443-458. 3M Company EPI-0003 Page16of25 Longnecker DS 1990). Experimentalpancreaticancer:roleof speciess,exand diet. BullCancer 77.27-37. Longnecker DS (1991).Hormones and pancreaticancer.IntJ Pancreatol9-.81-86. Oboum JD,Frame SR, BellW LongneckerDS, ElliotGtS, Cook JC (1997). Mechanisms forthepancreaticoncogeniceffectsoftheperoxisomeproliferatWoyreth14,643.ToxicolAppl Pharmacol 145:425-436. OlsenGW, GillilanFdD, BuriewNIK BurrisJK Mandel JS,Mandel JH (1998).An epidemiologiicnvestigationf reproductivheormones inmen withoccupationalexposure to perfluorooctanoiacid.J Occ Env Med 1998;40.614-622 1 Ophaug R, SingerL (1980). Metabolichandlingofperfluorooctanoaiccidinrats.Proc Soc Exp BiolMed 163:19-23. Palmer-SmithJ,Krame ST, Solomon TF-(1991). CCK stimulategsrowth ofsixhuman pancreaticancerceU linesinserum-freemedium. Reg Peptides22:341-349. PandolSJ (1998).Pancreatipchysiologyand secretortyesting.(In)Sleisenger FordtranJS (eds)Gastrointestinaanld LiverDiseasesV,olume 1, Phfladelphia:W.B. Saunders Co.,pp 771-782. PastoorTP, Lee KP, PerriMA, GiUiesPJ (1987). Biochemicaland morphologicalstudies of ammonium perfluorooctanoate-inducheedpatomegalyand peroxisomeproliferation. Exp Mol Pathot47:98-109. Pour PK Runge RG, BirtD, GingeU R, Lawson T,Nagel D, WaUcave L, SaimasiS (1981).Currentknowledge of pancreaticcarcinogenesiisnthehamsterand itsrelevance to thehuman diseaseCancer 47:1573-1587. Pour PK Lawson T, Heigeson S,DonneUy T, Stepan K (1988).Effectof cholecystokinin fonpancreaticarcinogenesiisnthehamstermodel. Carcinogenesis9:597-601. Rao MS, Reddy JK (1996).HepatocarcinogenesoifsperoxisomeproliferatorsA.nn NY Acad Sci 804:573-587. RehfeldJF,van SofingeWW (1994).The tumor biologyof gastriannd cholecystokinin. Adv CancerRes 63:295-347. SAS InstitutIen,c.(1990). SAS UsersGuide: StatisticVse.rsion6. Cary,NC: SAS InstitutIen,c. SibinskLiJ (1987).Two-year oral(diett)oxicity/carcinogensitcuidtyoffluorochemical FC-143 inrats.St.Paul,NIN:RikerLaboratories. 3M Company EPI-M3 Page17of25 UbelF,SorensonS,Roach D (1980).Healthstatusofplantworkersexposedto fluorochemicalsA: preliminaryreport.Am Ind Hyg Assoc 41:584-589. Vanden Heuvel J,KusWds B, Van Refelghem K PetersonR (1991). Tissue distributior@ metabousm and eumination of perfluorooctanoicacid. J Biochem Toxicol 6:83-92. van Kranen FU, Vermeulen E, Schoren L, Bas J,Woutersen RA, van IerselP, van Kreijl CF, Scherer E (1991). Activationof c-K-ras isfrequentin pancreaticcarcinomas of Syrianhamsters,but isabsentin pancreatictumors of rats.Carcinogenesis 12:147-1482. Wank SA, Pisegna JR, deweerth A (1994). Cholecystokininreceptorfamily.Ann NY i Acad Sci 713:49-66. YlinenKKohoA,HanhijarviKPeuraP(1990). Dispositioonfperfluorooctanoiaccid inthe rataftersingleand subehronicadministrationB.ull Environ Contam To)dcol 44:46- 53. 3M Company EPI-0003 Page 18 of 25 TableL MeanStandaDredviatioofnMean(SD)andRangeofPerfluoroctAacniod(iPcFOA)D.emographiCcl,inicCahlernistar-i Hcmatology,by SenLm PFOA Levels,and Year ofDataCollection PFOA 1993 Data 1995 Data 1997 Data pt)M Mean SD Range Mean SD Ranize Mean SD Range PFOA 0 -<1 ppm* 1-<10 ppm 10 -<30 ppm 2:30ppm 0.48' 0.2-7 0.00-0.99 3.38' 2.17 1.03-8.92 16.26' 3.39 11.90-21.00 60.13' 24.01 31.60-80.0 F value= 248.5,p 0001 0.312 0.32 0.00-0.90 3.032 1.84 1.10-8.20 17.11' 6.90 10,30-28.20 55.96' 33.29 34.20-114.10 F value = 77.6,p 0001 0.472 0.26 0.05-0.92 3.13'- 2.12 1.05-7.66 17.27' 5.19 10.50-23.13 58.14' 24.21 37.11-81.35 F value= 145.5,p 0001 Age 0 -<1 ppm 1 -<10 ppm 10-<30 ppm 30 ppm 43 9.2 27-61 39 7.8 27-60 41 5.0 34-49 33 7.4 25-43 F value= 3.3,p 02 42 8.3 29-60 41 8.6 24-59 45 7.4 30-55 38 9.2 27-50 Fvalue = 0.9,p =.46 40 9.1 41 8.7 44 11.3 40 11.2 F value- 0.3, 25-61 26-58 28-57 29-52 .84 B@E 0 -<1 ppm 1 -<10 ppm 10 -<30 ppm 2!30 ppm 28.0 4.3 20.9-42.0 26.9 2.5 21.6-32.5 28.3 2.8 22.4-32.0 28.5 1.6 26.9-30.2 Fvalue = 1.1,p =.33 27.6 4.2 21.945.2 28.6 3.4 22.1-38.3 27.8 4.0 21.2-34.8 29.8 1.8 28.2-32.6 F value = 0.8,p =.52 28.7 3.9 21.5-35.0 29.5 5.3 21.9-46.8 25.9 3.0 22.0-29.2 30.6 2.0 28.2-33.0 F value= 1.4,p =.24 Alcohol 0 -<1 ppm 1 -<10 ppm 10 -<30 ppm 30 ppm 0.4 0.5 0.0-1.9 0.7 0.7 0.0-3.4 0.8 0.6 0.4-2.1 0.9 0.8 0.0-2.0 F value- 2.9.p = .04 0.5 0.7 0.0-2.9 0.5 0.5 0.0-1.9 0.8 0.7 0.0-2.1 0.5 0.6 0.0-1.4 F value- 0.9,p =.43 0.7 1.0 0.0-3.4 0.5 0.6 0.0-2.6 0.4 0.6 0.0-1.4 0.7 0.6 0.1-1.6 F value= 0.7,p =.58 ci 0 -<1 ppm 1-<10 ppm 10-<30 ppm a 30 ppm 2 6 6 5 Fvalue 7.2 0-30 10.1 0-40 11.3 0-30 10.0 0-20 - 1.3,p =.26 4 9.4 040 3 6.0 0-20 9 15.2 0-40 6 8.9 0-20 F value = 1.3,p =.29 3 7.7 0-30 7 10.4 0-30 9 16.4 0-40 0 - - F value = 1.7,p =. 18 CCK 0 -<1 ppm 1-<10 ppm 10-<30 ppm 2:30ppm Not done Not done 33.4 15.7 13.4-80.5 28.0 19.2 8.8-96.7 15.7 4.1 11.4-23.0 20.6 7.2 12.9-29.9 Fvalue = 2.5,p =.06 TableL continued PFOA I)t)M 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm > 30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm > 30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm > 30 ppm 0-<Ippm 1 - <10 ppm 10 - <30 ppm 30 ppm 0-<Ippm 1 - <10 ppm 10 - <30 ppm 30 ppm 0 - <1 ppm 1 -<10 ppm 10 - <30 ppm 2:30ppm 3m Congany EPI-0003 Page 19 of 25 1993 Data Mean SD Range 88 26 37-161 82 23 47-151 75 14 58-107 103 33 74-132 F value = 1.6.p = .19 33 19 11-94 50 70 &472 36 14 19@-59 40 26 19-77 F value = 1.0,p = .41 23 7 11-60 26 11 12-83 23 4 16-28 28 6 23-35 F value = 1.1,p = .36 45 14 22-88 49 29 22-221 43 6 35-52 54 5 51-62 F value- 0.4.p - .75 0.63 0.28 0.20-1.30 0.58 0.22 0.20-1.30 0.53 0.17 0.20-0.80 0.65 0.13 0.50-0.80 F value= a7, p =.55 0.18 0.07 0.10-0.40 0.17 0.07 0.10-0.30 0.18 0.07 0.10-0.30 0.18 0.10 0.10-0.30 F value = 0.1, p = .94 1995 Data Mean SD Ranpe AUmbne Pholphatase 78 18 40-114 80 25 48-165 88 29 59-146 93 38 55-136 F value= 1.0,p - .39 GGT 42 27 16-149 51 41 19-190 39 13 21-61 42 15 23-58 F value= a6, p = .61 SGOT 21 6 13-36 24 13 13-75 21 4 15-29 20 4 15-25 F value = a5, p = .66 SGPT 44 13 27-80 53 34 27-175 45 12 28-70 51 13 39-71 F value= 0.9,p = .46 TotalBihrub 0.84 0.32 0.40-2.20 0,75 0.26 0.40-1.30 0.67 0.21 0.40-1.00 0,64 0.17 0.50-0.90 F value= ].5,p = .22 DirectBilirubi 0.22 0.04 0.20-0.30 0.22 0.05 0.10-0.30 0.19 0.06 0.10-0.30 0.18 0.04 0.10-0.20 F value = 2.2.p = -10 1997 Data Mean SD Range 79 19 27-122 87 23 47-164 81 28 61-142 78 16 64-100 F value= 0.9,p = .45 i 34 27 15-130 36 25 14-162 28 11 15-50 32 11 16-40 F value = 0.3,p = .91 26 25 24 27 Fvalue 7 7 2 5 = 0.3,p 13-41 14-48 22-28 22-34 =.84 31 10 13-59 33 15 14-80 30 11 18-51 43 13 27-57 F value = 1.0,p = .42 0.84 0.48 0.40-2.30 0.79 0.42 0.30-2.40 0.66 0.33 0.30-1.20 0.73 0.22 0.40-0.90 F value = 0.4.p = .76 o.lo 0.10 0.00-0.60 0.09 0.05 0.00-0-20 0.09 0.04 0.00-0-10 0.08 0.05 0.00-0-10 F value = 0.2.p = .89 TableL continued PFOA PI)m 0 - <1 ppm 1 -<10 ppm 10 - <30 ppm 2!30 ppm 0-<lppm 1 - <10 ppm 10 - <30 ppm 30 ppm 0-<lppm 1 - <10 ppm 10 - <30 ppm 2:30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm 2:30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm 30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm 30 ppm 3m Company ePI-OW3 Page 20 of 25 1993 Data Mean SD Range 215 44 136-297 219 39 155-309 230 38 171-305 236 42 175-268 Fvalue - 0.6,p =.65 43 47 52 37 Fvalue 11 13 23 10 2.0,p 22-83 29-90 34-97 27-47 139 40 27-227 143 38 72-223 144 37 65-185 131 57 60-188 F value - 0.2,p = .91 171 205 168 341 Fvalue 124 408 110 204 - a5, p 37-636 47-2845 41-362 70-564 -.67 15 4 9-25 14 4 5-22 14 4 7-17 13 1 11-14 F value = 0.9,p =.44 0.9 0.2 0.6-1.4 1.0 0.1 0.7-1.3 0.9 0.1 0.9-1.1 1.0 0.1 0.9-1.1 F value= 0.5, p = .68 1995 Data Mean SD Rang@ Cholestero 207 212 225 214 Fvalue 37 36 45 36 = 0.6,p 115-284 143-284 132-288 181-257 =.63 HDL 42 8 22-59 43 12 22-67 43 8 28-54 36 9 26-46 Fvalue = 0.9,p =.46 LDL 131 32 31-191 133 40 28-210 134 46 62-211 121 20 107-157 F value = 0.2,p =.92 TTiglycerid 170 175 239 296 Fvalue 93 57-371 144 59-743 147 77-539 iso 145-563 - 2.0,p -.13 BLTN 15 4 8-24 15 4 6-26 16 4 6-20 13 2 12-16 F value = 0.4,p =. 75 Creatinin 1.0 0.1 0.8-1.3 0.9 0.1 0.6-1.2 1.0 0.1 0.9-1.2 1.0 0.1 0.9-1.1 Fvalue - 1.4.p =.25 1997 Data Mean SD Ranize 199 30 129-257 213 41 121-315 228 57 164-334 230 21 212-258 F value = 1.8,p = @16 41 9 26-61 44 12 23-94 45 11 24-60 45 10 30-52 F value = 0.6,p =. 62 114 28 40-158 134 44 26-253 135 46 79-206 132 24 110-166 F value - 1.5,p =.22 219 176 241 269 Fvalue 116 53-445 85 44-360 241 63-718 79 198-360 - 1.4,p -.25 16 3 9-22 is 4 7-24 17 4 11-24 16 6 9-23 F value = 0.9,p =.44 1.0 0.1 0.8-1.2 1.0 0.1 0.7-1.3 1.0 0.2 0.8-1.4 1.0 0.1 0.9-1.1 F value = 1.1.p = .37 Table L continued PFOA RRM 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm 30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm 30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm 2!30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm 30 ppm 0 - <1 ppm 1 - <10 ppm 10 - <30 ppm 2:30ppm 3M Company EPI-0003 Page 21 of 25 1993 Data Mean SD RanLe 99 & 75-115 89 32 66-289 94 10 67-97 83 9 71-90 F value = 0.2.p = 87 45 2 40r5l 45 2 42-50 45 2 41-48 48 5 44-56 Fvalue = 1.6,p =.20 15.6 0.8 13.6-17.2 15.6 0.8 14.2-17.4 15.4 0.8 14.2-16.9 16.6 1.9 15.0-19.3 F value = 1.9,p = .14 5.1 0.3 4.6-6.3 5.0 0.3 4.5-5.9 4.9 0.3 4.4-5.4 5.4 0.5 4.8-6.0 F value- 2.3,p - -08 240 51 141-374 241 48 156-370 249 85 185-466 213 30 189-255 F value= 0.5,p = .72 1995 Data Mean SD Ranize Gluco 91 14 75-148 91 11 74-128 88 12 77-116 97 10 86-111 F value = 0.7,p = .59 Hematocri 44 2 41-51 44 2 41-49 44 2 40-48 45 5 41-54 F value = 0.6,p = .63 HCMORlob 15.1 14.9 15.2 15.7 Fvalue 0.8 13.7-17.1 0.8 13.2-16.7 0.8 14.0-16.7 2.0 14.2-19.1 = 1.0,p - 0.40 RBC 4.9 0.3 4.5-5.4 4.9 0.3 4.5-6.0 4.8 0.3 4.4-5.2 5.0 0.5 4.5-5.6 F value- 0.4,p - .78 Platelets 223 49 162-327 236 46 163-337 230 48 190-339 217 40 179-267 F value = 0.5,p =. 72 1997 Data Mean SD Range- 92 16 63-153 1()0 38 71-255 94 5 76-90 90 12 72-97 F value = 0.9,p =.47 45 45 45 44 Fvalue 3 2 2 3 = 0.5,p 38-52 42-49 40-47 41-46 =.66 15.3 0.9 13.2-18.0 15.4 0.8 13.7-17.0 15.2 0.6 13.9-15.8 15.0 0.9 14.2-16.0 Fvalue - 0.4,p -. 79 5.1 5.1 4.9 4.8 Fvalue 0.5 0.4 0.3 0.3 = 1.4.p 4.1-7.0 4.3-6.1 4.3-5.2 4.5-5.1 -.25 234 32 162-289 250 55 129-378 237 44 172-308 260 36 215-300 F value = 1.0,p =. 42 TableL continued PFOA Rpm 0 -<1 ppm 1 -<10 ppm 10 - <30 ppm 2:30 ppm 3M ConpaLny EPI-0003 Page 22 of 25 1993 Data Mean SD Ranize 6.5 2.1 3.3-15.0 6.5 1.9 3.9-12.7 6.6 2.0 4.6-11.4 6.3 1.9 4.4-8.7 F value- 0.1.p - .99 1995 Data Mean SD Ranie@- WBC 6.0 1.8 3.4-11.8 6.3 1.8 4.1-10.8 6.3 1.2 4.9-8.6 6.5 1.2 5.1-8.3 F value - 0.3.g - .82 1997 Data Mean SD Ranne- 6.3 1.1 4.3-8.9 7.1 2.3 4.2-16.5 6.8 2.2 5.0-10.9 6.1 0.9 5,0-7.2 Fvalue - 1.0.p =.739 1. Mean significantdliyfferen(tBonferronit-testp,< .05)than each of the otherPFOA ppm categories. 2. M= significantdliyffertn(tBonfermnit-tespt,<.05) thanthe10-30ppm and 30 ppm PFOA categories. *StudyPopulationby PFOA (ppm) Categoryand Year PFOA Category 1993 1995 1997 0 -<1 ppm 52 39 29 1 - <10 ppm 46 26 34 10 - <30 ppm 9 10 7 a 30 ppm .4 5 4 Figure I LN(CCK) vs.PFOA 4.5-- 4 3.6 3 2.6 2-1.6-- 0 10 20 30 40 so 80 70 PFOA (ppM) 3M Company EPI-0003 Page 24 of 25 Table II.Change inHDL* from LightAlcoholDrinker(< I drink/dayt)o Moderate AlcoholDrinker(> 1 dfink/dayA)ssociatedwith a 10 ppm Cha.ngeinSerum PFOA Level Moderate Moderate Drinker with Year Drinker 10 ppm increasein PFOA 19900 +9.9 -6.2 1993 +4.8 +3.4 1995 +5.1 +4.1 1997 +5.7 +3.8 'Detemiinedfrom multivariabmloedel adjustedforage,body mass index and smoldng (alflouryears)and testosteron(e1990,1993 and 1995 only). 0Data in 1990 analyzedtoalserum organicfluorine(seeGUWand and Mandel, 1996). 3M Company EPI-0003 Page 25 of 25 Table III.Change inSerum GlutarrdcOxaloacetiTcransaniinaseand Serum Glutamic PyruvicTransaminaseAssociatedwith a 10 ppm Change in Serum PFOA BAG (kg/m2) 25 30 35 SGOT 19900 1993 1995 1997 -2.4 3.7 9.7 1.3 0.5 -2.5 -0.3 0.1 0.5 2.1 0.1 -1.9 SGPT 19901, 1993 1995 1997 -3.0 28.0 59.0 2.5 1.5 0.5 0.8 -1.0 -2.1 5.3 0.7 -3.9 'Determinedfrom multivariablregressiomnodel adjustedforage,body nms indexand smoldng. 1Data in 1990 analyzedtotalserum organicfluorinleevel(see.GiUilanadnd Mandel, 1996).