Document 363x1MbK91g1yjrZzLRYEXqn

5M ComPpagaenly Date: May 16,2006 EpFiIdeNmiAolLogRy,E2P20O-3RW-T05 Med3iMcalCoDmeppaarntyment St. Paul, MN 55144 STiutrlvee:ilAlnancAenaPlryosgisroafmtahte32M00C0omFplauonryo'csheAmnitcwaelr(pPe(rBfellugoiruomo)c,taCnootattaeg,ePGFrOoAv)e Medical (Minnesota), and Decatur (Alabama) Facilities Geary W. Olsen, DVM, PhD! Larry R. Zobel, MD, MPH" 1. 3M Company, Medical Department, Mail Stop 220-6W-08, St. Paul, MN 55144-1000 3M ComPpagaeny2 Abstract `The 3M Company has published several periodic medical surveillance studiesofits fluorochemical production workers at its Antwerp (Belgium), Cottage Grove (Minnesota) and Decatur (Alabama) manufacturing facilities. These programs have compared clinical chemistry results in relation to serum measurementsofperfluorooctanesulfonate (PFOS) andlor perfluorooctanoate (PFOA). Toxicologically, PFOAis a peroxisome proliferator alpha receptor (PPAR) agonistandexerts morphological and biochemical effects characteristic ofPPARa agonists including beta-oxidation of fatty acids, increased CPY450-mediated reactions, and inhibitionofthe secretion ofvery low-density lipoproteins and cholesterol from the liver in rats and mice. In worker studies there have beennoconsistent associations with PFOA and various clinical chemistry measurements. A weak positive association between PFOA and serum total cholesterol was reported in `an analysis ofa the 2000 fluorochemical medical surveillance data for Antwerp and Decatur employees combined but it was considered a spurious finding based on the understanding thata hypolipidemic effect would be the hypothesis to be tested from the toxicological data. A subsequent unpublished brief letter from DuPont reported that total cholesterol, LDLandtriglycerides, but not HDL, were positively associated with their employees' serum PFOA concentrations but the authors urged caution as minimum variance was explainedintheiranalyses. In lightofthese findings, the purposeofthis studywas to reanalyze the 2000 fluorochemical medical surveillance male employee database (n = 552) which included three facilities: Antwerp, Cottage Grove and Decatur and to stratify the analyses by employees' self-reported cholesterol lowering medication 3M ComPpaagney status concentrating on those individuals (n = S06, 92 percent) who reported they were not taking cholesterol lowering medications. Using several typesofstatistical analyses (analysisofvariance, analysisofcovariance, logistic regression and multiple regression using multiplicative models), there was no evidence that PFOA [adjusted for age, body mass index (BMI) and alcohol] was statistically significantly (p <0) positively or negatively associated with serum total cholesterol or LDL (calculated by the Friedwald formula when serum triglycerides were less than 400 mg/dL) regardlessof cholesterol lowering medication status. Although HDL was negatively associated with PFOA for the combined three locations, only one percent of its variance was explained with PFOA and this association was not observed `when the data were stratified by location. Thus, the association observed with HDL and PFOA was likely due to residual confounding between the locations themselves rather than to any possible causal association with PFOA. For the combined locations, serum triglycerides were positively associated with PFOA but not consistently by location. Antwerp, the location with the lowest mean triglyceride and PFOA concentrations, showed a positive association whereas Cottage Grove, with the highest mean PFOAand triglyceride levels, showed no association between triglycerides and PFOA. A hypothesis is offered that the inconsistent associations observed for serum triglycerides and PFOA might bedueto non-adherence to fasting requirements by night shift production workers since all sample collections occurred in the moming for production `and non-production employee participants. The former would have higher serum PFOA. concentrations. Also, shift workers, in general, have been reported to have more elevated 3M ComPpaagneys serum triglycerides. To clarify an association, ifany, between PFOA concentrations and serum triglyceride levels, the methodological questions raised herein merit further inquiry. Adjusted for the potential confounding factorsofage, BMI or triglycerides, and alcohol, the three locations combined did not result in consistent associations with hepatic enzyme tests. Whereas the individual Antwerp and Cottage Grove locations showed no. statistically significant associations, weak positive associations were observed in the multiplicative models between alkaline phosphatase, ALT, and GGT with PFOA (adjusted for age, BMI or triglycerides, and alcohol) among the Decatur male employee `population which could be due to residual confounding as minimum variance was explained by PFOA in these models. Several epidemiologic research studiesofthe Decatur workforce have not found statistically significantly increased risks for hepatic disease (malignant or nonmalignant conditions) using a varietyofdata sources including episodesofcare, self-reports, and death certificates. Other worker populations and a `community-based PFOA-exposed population have also not reported positive associations. between liver enzyme test results and serum PFOA concentrations. Few individual thyroid-related hormone (TSH, T4, free T4, and T3) results were out-ofreference range. Analyses of the multiplicative models predicted these thyroid-related hormone measurements would be well within their normal reference ranges when PFOA concentrations varied between 0.005 g/mL (average general population concentration reported) and 100 pg/mL (high rangeofoccupational measurements historically 3M ComPapgea'nsy reported). The lackofan association with clinically relevant thyroid testresults has also been reported in other worker and community-based studies of PFOA whose average serum concentrations were at or below those reported in the present study. Introduction 3M ComPpaagneys `The 3M Companyhas published several periodic medical surveillance studies ofits fluorochemical production workers at its Antwerp (Belgium), Cottage Grove (Minnesota) `and Decatur (Alabama) manufacturing facilities (Ubel 1980; Gilliland and Mandel 1996; Olsen etal. 1998; 1999; 2000; 20033; 2003b). These programs have compared clinical chemistry results in relation to serum measurementsofperfluorooctanesulfonate (PFOS) and/or perfluorooctanoate (PFOA). Serum PFOS concentrations measured in these 3M employeeshavebeen routinely `compared with serum total cholesterol levels as adeclineinthe latter has been a consistent early reliable measureofclinical response reported in laboratory animal studies (3M Company 2003; Haughom and Spydevold 1992; Seacatetal. 2002; 2003). No association between PFOS and cholesterol has been observed in these fluorochemical medical surveillance programs, and is likely duteo thefactthat the serum PFOS. concentrations measured (average approximately 1 to 2 pg/mL, maximum < 14 ug/ml) were below those measured in laboratory animals where effects have been reported (Olsen ctal. 1999; 20033; 2003). Among cynomolgus monkeys fed PFOS in the diet for 6 months, statistically significant decreases in cholesterol from predose values occurred in the high dosegroup (0.75 mg/kg/day)thathad serum PFOS concentrations above 100 `ppm (Seacat et al. 2002). This serum PFOS concentration in the high dose group. corresponded to $0 mg/kg and 34 mg/kg cumulative doses inmalesand females, respectively. Seacat et al. (2002) considered this decrease in serum total cholesterol the 3M ComPpaagney? earliest reliable measureofclinical response to PFOS in these cynomolgus monkeys. Lowered HDL values were also reported inthe 0.75 mg/kg/day dose group. However, a lackofprestudy and interim HDL values ina lower dose group (0.15 mg/kg/day) made this interpretation more problematic. Ina sub-chronic dietary toxicity of PFOS in rats, `Seacatet al. (2003) did not report strong evidence for hepatocellular peroxisomal or cellular proliferation at the doses tested (0, 0.5 2.0 5.0 and 20 parts per million in dict) Lowered serum cholesterol was observed although decreased glucose among male animals at the high dose at 4 weeks (but not at 14 weeks or with females at 4 or 14 weeks) and elevated alanine transaminase levels (ALT) at 14 weeks in males (but not females) were also reported. The modeofactionforserum cholesterol reduction is uncertain but may be due, in part, to PFOS acting as a peroxisome proliferator alpha. receptor (PPARa) agonist. PFOA is a PPARa agonist and exerts morphological and biochemical effects characteristic of PPARa. agonists including beta-oxidationoffaty acids, increased CPY450-mediated reactions, and inhibition ofthe secretion of very low-density lipoproteins and cholesterol from the liver in rats and mice (Haughom and Spydevold 1992; DePierre 2002; Kennedy etal. 2003;Xieet al. 2003). These effects resulted in a reductionofserum cholesterol and triglycerides inratsand mice. Xie etal. (2003)have shown that severe adipose tissue atrophy occurs upon dietary treatment of mice with PFOA buts rapidly reversed after terminationoftreatment. Hepatomegaly, however, was much more persistent after dose termination. Unlike ratsandmice, there was no reduction in serum cholesterol in cynomolgus monkeys dosed with PFOA (ammonium 3M ComPapgaensy salt) fo6r months (Butenhoff etal. 2002). PFOA was statistically significantly positively associated with triglycerides in the high-dose (30/20 mg/kg) group whose steady state Serum PFOA concentration was 158 g/mL 10 g/mL (range 20to 467 pg/mL). This association between PFOA and serum triglycerides was observed in measurements taken after one monthofdosing at which time the group mean triglyceride level was significantlyhigherthan control valuesaswell as within group pretreatment values. At the endofthe study the mean iglyceride was elevated compared to time related controls but not 10 the animals" pretreatment values. However, only two primates were evaluated inthe high-dose group at end of study. Inspection of individual values for PFOA serum concentration and serum triglyceride values did not reveal a meaningful association between these two parameters (John Butenhoff, personal communication;seeOlsen et al. 2003a). In worker studies there have been no consistent associations with PFOS or PFOA and various clinical chemistry measurements (Gilliland and Mandel 1996; Olsenetal. 1998; 1999; 2000; 2003a; 20030). Gilliland and Mandel (1996) did report that PFOA (measured as the surrogate serum total organic fluorine) may negatively modulate the effect alcohol has on high-density lipoprotein (HDL) levels and exacerbate the effect that obesity has on hepatic enzyme tests. However, three subsequent analyses ofthis employee population that measured specifically for PFOA did not find that it modulated hepatic responses to either obesity or alcohol consumption (Olsen et al. 2000). 3M Company Page Previous analysesofthe most recently published fluorochemical medical surveillance data conducted at the 3M Antwerp and Decatur facilities (conducted in 2000) primarily concentrated on PFOS analyses although data were also presented for PFOA. In these PFOA-specific analyses, Olsen etal. (20033) reported a positive association with serum cholesterol and triglycerides with PFOA. They atributed this to be a spurious finding given the toxicological evidence would suggest a hypolipidemic (not hyperlipidemic) effect, along with the fact that minimum variation was explained in the statistical models (partial R* for PFOA in the models was <0.01 (cholesterol) and 0.03 (riglycerides), respectively) used in the analyses. In the same published paper, Olsen et al. (2003a) also provided a longitudinal analysis of 174 Antwerp and Decatur male employees who participated in at least twoofthree fluorochemical medical surveillance programs between 1994 and 2000 and found PFOA to also be positively associated with cholesterol as well as triglycerides. This association was primarily tributed to 21 Antwerp employees whose mean serum PFOA levels increased over a six-year period of time from 1.32 ppm 10 2.06 ppm. During the same time period, their mean cholesterol values increased from 208 mg/dL. to 229 mg/dL and their triglyceride levels increased from 85 mg/dL to 123 mg/dL. Their BMI, however, also increased from 23.4 to 24.3 during this same time period. No statistically significant associations were reported between PFOA and cither serum cholesterol or triglycerides for the Cottage Grove male employees who participated in the 2000 fluorochemical medical surveillance program (Olsen et al. 20034). Analysesofprior fluorochemical medical surveillance programs at Cottage Grove (1993, 1995 and 1997) showed mean serum triglyceride levels were highest `amongtheCottage Grove PFOA production workers with the highest (>= 10 ppm) PFOA 3M CoPmapgean1y0 serum concentrations although controlling for potential confounders did not provide consistent results as well as the fact that severalofthe observations in the high exposure. category were repeated measurements from the same individuals. Based on the possibility ofa positive association between PFOAand serum cholesterol `and/or triglycerides that was not supported by laboratory (toxicological) data and was therefore consideread spurious association (Olsen et al. 2003a), thers have since provided nonpublished data (Costa 2004; Leonard 2005). Costa (2004) provided cursory analysesofapproximately 35 Italian fluorochemical production workers and reported a "slight increaseoftotal cholesterol in workers exposed to PFOA." There was no increaseofother lipids, such as wriglycerides, but the fractionofnon-HDL cholesterol appeared elevated. Costa suggested his findings might be consistent with the hypothesis that PFOA might influence cholesteryl ester transfer protein (CETP). CETP is a plasma glycoprotein that facilitates the transferof cholesteryl esters from HDL (apolipoprotein A-containing lipoprotein)to apolipoprotein B-containing lipoproteins (e.g., LDL) (Brousseau etal. 2004). InhibitionofCETP has been shown to markedly elevate plasma levels of HDLand apolipoprotein Al. However, there did not appeatro be a decrease in HDL in the Costa analysis. Several additional caveats also existed in the Costa data analysais outlined by Kaplan (2005) including: 1) the Costa data set was a small and arbitrary collectionof subjects; 2) there were no pre-employmentbaseline lipid levels for historical reference; 3) concomitant exposure to other chemicals was unknown; and 4) inadequate adjustment for important confounding factors. In addition, manyofthese: 3M CoPmapgean1y1 individuals were the samesubjectsovertimeand no repeated measures analysis was conducted. Leonard et al. (2005) have provided to the U.S. EPA an unpublished one page report that briefly summarized a cross-sectional analysis of 782 male and 243 female (combined eligible population 1,863) DuPont workers with potential exposure to PFOA at the Parkersburg, West Virginia facility. Of the 62 clinical chemistryand hematology `endpoints measured, Leonard et al. found most were well within normal ranges and not associated with serum PFOA levels. Measured PFOA concentrations were reported as high as 10 g/mL (parts per million, ppm) butthe average was less than I ug/ml. (Leonard, personal communication). No statistically significant associations were observed for PFOA and serum liver enzymes or any hematology measures and PFOA. EKGsand C-reactive protein were not associated with PFOA. Statistically significant `positive associations were identified for total cholesterol, LDL and triglycerides with serum PFOA levels adjusted for BMI, alcohol and age in both males and females although the percent variation explainedwasgenerally low. The potential change in total cholesterol at the highest serum PFOA level was approximately 10 percent. No effect, however, was seen in the HDL fraction. Smal, but statistically significant increases in uric acid and iron were also reported with the highest PFOA blood levels. A suggested role for PFOA to enhance CETP activity, as discussed by Costa, however, was not supported by the much larger Leonard et al. analyses since the latter investigators showed no association between PFOA and HDL. Leonard et al. offered no hypotheses or suggested causal models (Hein et al. 2002)to explain their observations. Because 3M CoPmapgean1y2 PFOA has beenshownto bind to blood albumin in the rat, monkey and human (Han et al. 2003; Kerstner-Wood et al., 2003), a positive correlation, and not a causal association, between lipoproteins (binding) and PFOAis a possibility. Re-examination ofthe Olsen et al. analyses conductedofthe 2000 fluorochemical `medical surveillance programattheAntwerpand Decatur facilities (Olsen et al. 2001a; 2001b; 2001c; 2001; 20033)and the Cottage Grove facility (Olsen et al. 2003b) suggested some limitationsofthese data analyses as they may relattoe testing a hypothesis that PFOA is positively associated with total cholesterol and its fraction, LDL. First, all male subjects were included in the original analysis regardless of their cholesterol lowering medication status. A positive association may be masked by inclusionofsubjects whose serum cholesterol levels have been reduced by medication if suchan increase was, in part, associated with higher PFOA concentrations. Second, LDL `analyses were not restricted to those instances where serum triglycerideswere400mg/dL orless. The potential for bias steadily increases with higher triglyceride levels (Nakanishietal. 2000). Third, the Antwerp and Decatur facility study (Olsen etal. 20030) concentrated primarily on PFOS and not PFOA. Both PFOSand PFOA have been shown to result in hypolipidemia in rats at high concentrations and thus the causal `model hypothesized in these earlier analyses concentrated on testing for this effect, not hyperlipidemia. Olsen et al. (20033) did not show a causal association between PFOS and lowered cholesterol attheserum concentrations measured in these workers. Direct `comparison with PFOA from the Olsen et al. (2003a) report is not possible since the manyofthe tabular analyses (.g., Table 2 of Olsen et al. 2003a) were PFOS-specific. 3M CoPmapgean1y3 Although the employees' PFOS and PFOA concentrations were correlated, the: categorical analysesofPFOS did result in overlappingof employees' PFOA concentrations. Fourth, there wasnocombined analysis ofthe three manufacturing facilities, Antwerp, Cottage Grove and Decatur, as the Cottage Grove analysis has historically been reported separately (Ubel et al. 1980; Gilliland and Mandel 1996; Olsen etal. 2000; 2003a). Those sites that were combined (Antwerp and Decatur) may have resulted in unmeasured and/or residual confounding since the two locations were distinctly different with some potential confounding variables, most especially alcohol consumption and BMI (Olsen et al. 1999; 20033). `With the above limitations in mind, the purposeofthis study was to reanalyze the 2000 fluorochemical medical surveillance program data in order to examine the hypothesis that PFOA may be positively associated with increased cholesterol, LDL levels and. triglyceride levels. To address someofthe above limitations, the three manufacturing facilities were analyzed separately andjointly. Analyses concentrated on those male employees who self-reported that they were not taking cholesterol lowering medications in order to minimize any unexplained bias. Because PFOS concentrations were not previously associated with lowered serum cholesterol (Olsenetal. 2003), it (PFOS) was not considered an important potential confounding variable in these reanalyses for PFOA. Three covariates were: age, BMI and alcohol. Age is known to be positively associated with serum cholesterol. BMI is positively associated with serum triglyeerides (and to some degree cholesterol) and alcohol consumption (e.g., red wine) has been positively associated with increased HDL. Inaddition to the lipid variables analyzed, liver enzyme 3M CoPmapgean1y4 `and thyroid tests were also reanalyzed. Although liver enzymes, in particular gamma glutamyl transferase (GGT), aspartate aminotransferase (AST) and alanine aminotransferase (ALT) can be positively associated with heavy alcohol consumption, liver enzymes can also be due to obesity and dyslipidemia (Collantes et al. 2004; Mofrad and Sanyal 2003; Ruhl and Everhart 2003) as nonalcohol faty liver discase has, substantially increased in prevalence in the United States population, as indicated by the third National Health and Nutrition Examination Survey (NHANES) (Clark et al. 2003). `Whereas the majorityofelevated aminotransferase activity in NHANES could not be explained by alcohol consumption, viral hepatitis or hemochromatosis, unexplained aminotransferase enzyme elevation was significantly associated with higher BMI, waist circumference, triglycerides, fasting insulin, and lower HDL (Clark et al. 2003). Clark et al. (2004) concluded that unexplained aminotransferase elevation was strongly associated with adiposity and other featuresofthe `metabolic syndrome',andthus may represent nonalcoholic fatty liver disease. [Note: The metabolic syndrome includes abdominal obesity, atherogenic dyslipidemia (elevated triglyceride, small LDL particles, low HDL cholesterol), raised blood pressure, insulin resistance (with or without glucose intolerance), and prothrombotic and proinflammatory states (Expert Panel 2001). Clinical identificationofthe metabolic syndrome in men includes any threeofthe following: waist circumference >40 inches (surrogate BMI >= 30); triglycerides 150 mg/dL; HDL <40 mg/dL; blood pressure 130/85 mmHg; and fasting glucose 110 mg/dL] As they relate to the present study, the NHANES findings indicate BMI and serum triglycerides are important predictors of liver enzyme values and therefore should be controlled in any analyses that examinetherelationship between PFOA and liver 3M CoPmapgean1y5 enzyme tests (e.g, ALT, GGT), as was done previously with analyses with serum PFOS concentrations in this employee population (Olsen et al. 2003a). Methods "The following clinical chemistry variables were considered dependent variables for reanalysisofthe 2000 fluorochemical medical surveillance program data: alkaline phosphatase (IU/L), gamma glutamy! transferase (GGT, IU/L), aspartate aminotransferase (AST, IU/L), alanine aminotransferase (ALT, IU/L), total and direct bilirubin (mg/dL), blood glucose (mg/dL), cholesterol (mg/dL), low density cholesterol (LDL, mg/dL), high density cholesterol (HDL, mg/dL), and triglycerides (mg/dL). `These measurements were performed at Allina Laboratories (St. Paul, MN). LDL was an indirect calculation using the Friedwald formula [LDL = total cholesterol -- (wiglycerides/s)). Presented in this report are LDL values where serum triglycerides did not exceed 400 mg/dL. LDL was also calculated when serum triglycerides did not exceed 250 mg/dl as well as for all triglyceride levels. Results were not substantially different for LDL regardlessofformula used. Thyroid stimulating hormone (TSH; WIU/mL); serum thyroxine (T4; ug/dL); free thyroxine (free T4; ng/dL) and serum triiodothyronine (T3; pg/mL) were also re-examined. These thyroid related hormones were measured by LabCorp (Kansas City, MO). Thyroid hormone analyses excluded the five individuals who were already diagnosed with thyroid-related conditions and likely taking medication. Analyses did not substantially differ with these individuals included. 3M CoPmapgean1y6 Detailsofthe method validation including matrix extraction and analytical measurement `procedures used to determine serum PFOA and PFOS can be found elsewhere (see U.S. EPA docket AR226-1208, AR226-1209, AR226-1210). Theanalytical method consisted ofa liquid-liquid extraction procedure followedbyevaporation and reconstitutionofthe extract residue with 30:70 20 mM ammonium acetate in water:20 mM ammonium acetate in methanol (v/v). The samples were analyzed by liquid chromatography tandem mass spectrometry usinag PE Sciex API 3000. The instrumentwasoperated in the multiple reaction monitoring (MRM) mode under optimized conditions for detection of PFOS and PFOA for detection of negative ions formed by turbo ionspray ionization. Laboratory `analyses were conducted at Tandem Laboratories, SaltLakeCity, UT(formerly Northwest Bioanalytical Laboratory Inc.). All employee serum values (g/mL) for PFOA (and PFOS) were above the lower limitofquantitation (LOQ). Statistical analyses included both univariate and multivariable methods (analysis of variance, analysisof covariance, logistic regression, multiple regression) using IMP (Cary, NC) and Stata (College Station, TX) software. Age, BMI and alcohol (drinksper day) were considered as covariates in multivariable models. For analysis of hepatic enzyme tests in these models, triglycerides were also considered in place of BMI. Distributions were examined to assess nommality using nontransformed and transformed (log, square root, inverse) variables. In general, log transformations improved normality assumptionsofresponse and explanatory variables and thus multiplicative models were considered applicable (T. Church, personal communication). For the log transformation ofalcohol, 0.1 was added to drinks per day to prevent the logof0. Residual plots were 3M CoPmapgaenly? examined to detect model inadequacies including homoscedasticity. Goodnessoffit statistics were examined including R* and adjusted R'. Multicollinearity in the models was assessed by the inverse oftolerance (Variation Inflation Factor). Crude and adjusted odds ratios for PFOA categorized by deciles were determined via logistic regression analyses for reference range valuesofthe response variable. A revoifethwe 2000 CottageGrovedatabase used intheOlsen et al. (2003b) report found eight male employees had missing BMI and/or alcohol information. For the present study, a reviewofthe subjects' Cottage Grove medical records by the plant nurse provided these data and was generally within two yearsofthe 2000 Cottage Grove `medical surveillance program. Graphical analyses were performed to examine pattems ofassociation (see Appendix A) for PFOA. Graphs in Appendix A include nontransformed as well as transformed (natural log) variables. Density ellipsesarealso provided in Appendix A for PFOA. A density ellipsoidwasconsidered a good graphical indicatorofthe correlation between two variables. The ellipsoid collapses diagonallyasthe correlation between two. variables approaches either 1 or -1. The ellipsoid is more circular (less diagonally oriented)ifthe two variables were uncorrelated. Results A total of 196 (95%)ofthe 206 Antwerp male employee participants, 122 (93%)ofthe 131 Cottage Grove male employee participants and 188 (87%)ofthe 215 Decatur male 3M CoPmapgean1y8 employee participants were included in the primary reanalysesof the 2000 fluorochemical medical surveillance program (Table 1). These 506 (92%) participants represent those who did not self-report taking choleseterol lowering medications. The mean PFOA concentration was statistically significant(lpy < 05) higher among CottageGrovethan either Antwerp or Decatur male employee participants but the median PFOA concentration for Decatur was somewhat higher than either Antwerp or Cottage Grove participants (Table 2). [Note: Decatur also hadaslightly higher mean concentration for PFOS than Antwerp or Cottage Grove employee participants, with the `median value approximately twice that of Antwerp or Cottage Grove] Ashasbeen reported before (Olsen et al. 1999; 20033), Antwerp employees were statistically significantly younger, had lower BMIs, and consumed more alcohol than their `counterparts at Decatur, and inthepresent analysis, also Cottage Grove. Whereas mean cholesterol and LDL were not statistically significantly different between the three locations, mean HDL levels were statistically significantly higher among Antwerp employees whereas their triglyceride and blood glucose levels were significantly lower. Mean (arithmetic) liver enzyme tests were statistically significantly lower for Antwerp compared to Cottage Grove and Decatur employees except for total and direct bilirubin. Antwerp employees had a statistically significantly lower mean TSH than Decatur employeesandy higher mean T> than the Cottage Grove or Decatur employees. These `mean thyroid values, however, were well within the reference range. 3M CoPmapgean1y9 `The number and percentofemployees by reference points for demographic factors and clinical chemistry results are presented in Table 3. As expected based on the data presented in Table 2, there were substantially fewer Antwerp employees (5%) who would be categorizaesd obese (BMI 2 30) compared to Cottage Grove (44%) or Decatur (32%) employees. More than 40%of Antwerp and Cottage Grove employees reported drinking 1 alcohol beverage per day compared to only 1%ofDecatur employees. Statistically significant differences between the three locations were seen for blood glucose, HDL, triglycerides, ALT, and total bilirubin. There was only 1 percentof the Antwerp employees who were categorizedashavingthe `metabolic syndrome' compared t0 20% and 13% among the Cottage Grove and Decatur employees, respectively. The present study definitionofthe metabolic syndrome included any threeofthe following four: BMI 30; triglyceride>s 150 mg/dL; HDL <40 mg/dL; and fasting glucose > 110 mg/dL. Presented by PFOA quartile ranges arethedata for Antwerp (Table 4), Cottage Grove (Table 5) andDecatur(Table 6). Quartiles were divided to allow foar defined break in PFOA concentrations. Among Antwerp employees, the mean value in the first PFOA quartile was statistically significantly different than the fourth PFOA quartile for PFOA, PFOS, age, blood glucose, triglycerides and T3. Among Cottage Grove employees, the `mean value in the first PFOA quartile was statistically significantly different (higher) than the fourth PFOA quartile for PFOA and AST. Among Decatur employees, the mean value in the first quartile was statistically significantly different (lower) than the fourth PFOA quartile for PFOA, PFOS, ALT, AST/ALT ratio, total bilirubinand T3. The 3M CoPmapgean2y0 AST/ALT ratio is a useful index for distinguishing nonalcoholic fatty liver (steatohepatitis) disease (AST/ALT less than one) from alcoholic liver discase (AST/ALT greaterthan one). It should be noted that those locations (Cottage Grove and Decatur) with the highest obesity (as measured by BMI which is not necessarialsy good ofa `measurement as waiship circumference ratio) had the lowest AST/ALT ratio. `The 506 employees from the three locations were combined and categorized into PFOA decilesin Table 7. Mean PFOA decile concentrations ranged from decile 1 0f0.06 g/mL (range0.007 ~0.13 pg/mL) to decile 10 of 12.15 pg/mL (range 3.71 ~92.03 g/mL) with corresponding median values of 0.06 pg/mL and 4.94 pg/mL, respectively. "The distributionsofdemographic factors by PFOA decile are presented in Table 8. Therewas an uneven distributionofemployebeys location inthesedeciles. Ifthere was aproportional distribution across deciles bylocation,then there would be 39% Antwerp, 24% Cottage Grove, and 37% Decatur employees in each decile. Instead, as seen in Table 8, there are higher percentages ofAntwerp employees in the frst 3 deciles and lower percentages in the highest two deciles (9:and 10). Likewise, there are lower percentagesofDecatur employees in the first three deciles and higher percentages in the upper deciles (six through 10), in particular decile 9. Cottage Grove employees have higherthan expected percentages in decile 1 and decile 10. Becauseofthese uneven distributions, mean BMI values are greater in the upper deciles and alcohol consumption is lower, reflecting the demographic differences seen across the three facilities 3M Company Page2l "The mean and median PFOS concentrations by PFOA decilearepresented in Table 9. In general PFOS concentrations increased by decile but not consistently. Deciles 7 through 10areall statistically significantly higher than deciles 1 through 5. Presented in tables 10 through 13 are the analysesforcholesterol, LDL, HDL and triglycerides. Table 10 is the univariate analysis by PFOA decile. Table 11 presents the `mean decile values adjusted for age, BMI and alcohol using analysisofcovariance. Table 12A presents adjusted odds ratios and 95% confidence intervalsofthe deciles by reference range cutoff points listed in the table using decile 1 as the reference. Odds ratioswere adjusted for age, BMI and alcohol. Table 12B presents crude and adjusted (for location) odds ratios and 95% confidence intervals ofthe deciles by reference points for HDL and triglyceride. Table 13 presents non-adjustedandadjusted PFOA. coefficients with cholesterol, LDL, HDL and triglycerides adjusted for age, BMI and alcohol in the multiplicative models. Analyses included alllocationsas well as presented individually for each location by itself. Based on the analyses presented in Tables 10 through 13, cholesterol and LDL were not statistically significantly associated with PFOA whether across the combined three locations or each location analyzed separately. Adjusted mean HDL levels were lower in the highest (10) PFOA decile comparedtodeciles 1 throug4h (Table 11). The crude odds ratio was 3.0 (95%CI 1.2-7.5) for HDL < 40 mg/dL in decile 10 compared to decile 1 (Table 12B). However, thiswasreduced to an adjusted odds ratioof 1.5 (95% CI 0.64.0) when location was considered a covariate. Similar results were observed for 3M CoPmapgean2y2 triglycerides (Table 12B). The adjusted PFOA coefficient in the multiplicative model for HDL was statistically significant at p= 01 (Table 13) but only explained 1 percent of the HDL variance in the full model (R = 0.26), which was primarily attributable to BMI and alcohol. This finding with PFOA was also attributed to the decile disproportionate distribution of subjects as shown previously with the much greater proportionofCottage Grove and Decatur employees in the highest two PFOA deciles. Multiplicative models stratified for eachofthe three locations showed no statistically significant findings between PFOA and HDL (Table 13). Mean serum triglyceride levels were highestin decile 10 whether unadjusted or adjusted (Tables 12A and 12B). Odds ratios for triglyceride level>s 150 mg/dL were highest for deciles 8 through 10. The multiplicative model for the combined three locations indicated a statistically significant coefficient (p < 0001) for PFOA (Table 13) which explained approximately 4 percentofthe varianceofthe response variable. Stratified by location, however, showed only Antwerp to produce a statistically significant coefficient although Decatur had a marginal nonstatistically significant coefficient for PFOA (p= 07). No association (p= 38) was found between PFOA and triglycerides for Cottage Grove in the multiplicative models (Table 18). Cottage Grove had the highest PFOA concentrationsofthe three locations. As seen in Tables 2 and 3, serum triglyceride levels were considerably lower among Antwerp than Cottage Grove or Decatur employees. Presented in tables 14 through 18 are the analyses for alkaline phosphatase, AST, ALT, AST/ALT ratio, GGT, and total and direct bilirubin. Table 14 is the univariate analysis 3M CoPmapgean2y3 by PFOA decile. Table 15A presents themeandecile values adjusted for age, BMI and alcoholusinganalysisof covariance. Table 1B presents the mean decile values adjusted for age, triglyceriadneds alcohol using analysisofcovariance. Table 15C presents the adjusted AST/ALT ratios. Table 16presents the percentof employees by PFOA decile that were above reference range for these liver clinical chemistry tests. Tables 174and 17B present adjusted oddsratios and 95% confidence intervalsofthe deciles by different reference rangecutoffpoints for ALT and GGT provided in table 16 using decile 1 as the reference. These odds ratios were non-adjusted (crude), adjusted for age, BMI and alcohol, and adjusted for age, triglycerides and alcohol. Table 18 presents non-adjusted and adjusted PFOA coefficientsofthese liver enzymes using either age, BMI and alcohol or age, triglycerides and alcohol (natural logs) as covariates in the multiplicative models. Analyses in Table 18 included all locations as well as presented individually for each location. Overall, the mean non-adjusted ALT is statistically significantly higher in deciles 9 and 10compared to deciles | through 6 (Table 14) but this association is not evident upon adjustmentofthe mean values in Tables 15A or 15B. Odds ratiosfor values above the reference range are only calculated for ALT and GGT (Tables 17A and 17B) because: these were the only liver clinical chemistry tests that had more than a few employees outside the reference range (Table 16). There were no statistically significant odds ratios (non-adjusted or adjusted) for ALT or GGT when compared to decile 1 (Table 174). Analysis of a lowercutoffreference point (= 40 1U) did not substantially alter the findings (Table 17B). Analyses did not result in any statistically significant associations 3M CoPmapgaen2y4 when all locations were included in the multiplicative models, except for total bilirubin (negatively associated) (Table 18). Stratified by location,Decaturhad marginally statistically significant positive coefficients for alkaline phosphatase, ALT, GGT and a negative log coefficient for total bilirubin in the models. The amountofvarianceofthe hepatic response variables explained by PFOA in these models was minimal ranging from <1 to3 percent. These hepatic enzyme associations with PFOA were not observed for the individual Antwerp or Cottage Grove analyses. Presented in tables 19 through 22 are the PFOA analyses with TSH, T4, free T4 and T3. `Table 19 is the univariate thyroid analyses by PFOA decile. The mean TSH value in the 4" decile (3.41 uIU/mL) is influenced by one subject whose TSH value was 65.28 pIU/mL. If removed from analysis, the meoaf tnhe 4 decile became 2.15 uIU/mLand not statistically significant from any other decile. Table 20 presents the mean values adjusted for age, BMI and alcohol using analysisofcovariance which includes the outlier in decile 4. The adjusted decile 10 ofmean free T4 was statistically significantly lower than thatofthe first decile (Table 20). Table 21 presents the percentofemployees by PFOA decile that were above reference range for these thyroid tests. Adjusted odds ratios and 95% confidence intervalsare not presented ina separate table becauseofthe few subjects whosethyroidtests were above or below the reference range as shown in Table 21. Only twoofthe 18 subjects in Table 21, thathad TSH values above the upper reference range, had another thyroid related hormone result outofreference range (Table 22). One individual had aT4 value below reference range (with the highest TSH value in the analysis) and the other was above the reference range for T4 (Table 22). The former 3M CoPmapgean2ys was likely a clinically diagnosed hypothyroid individual (subject 45 in Table 22) who also had a low T4just abovethereference range. Table 23 presents non-adjusted and adjusted coefficients (natural log) of PFOA in the multiplicative models. For all locations combined, there were no statistically significant adjusted coefficients in the models except for free T4 (negative coefficient for PFOA) and TS (positive coefficient for PFOA). However, the full multiplicative models (4 independent variables) explained only and 2 percentof the varianceoffree T4 and T3, respectively. There were also positive PFOA coefficients for TS in the individual Antwerpand Decatur locations (Table 23). Percent variances explained in these models were9 and 7 percent, respectively. Because so few values were out-of-reference range (Table 21), the results presented in Table 23 are not suggestiveofclinically relevant findings as the predicted results from these models (Table 24) were well within normal reference ranges for these thyroid parameters. Predicted serum PFOA concentrations, up to 100 pg/mL in Table 24, were $ ordersofmagnitude greater than the average PFOA concentration reported in the general population (approximately 0.005 pg/mL). As shown in Table 24, these predicted PFOA concentrations resulted in TSH, T4, free T4, and TS all within their normal reference ranges. `There was no association observed between metabolic syndrome, as categorized in this study, and PFOA, as shown in Table 24 which presents the numberofpeople defined with metabolic syndrome and non-adjusted and adjusted (age) odds ratios and 95% confidence intervals by PFOA decile. 3M CoPmapgean2y6 In addition to the analyses that focused solely on those male subjects who self-reported they were not taking cholesterol lowering medications as shown in Tables 1 through 24, other analyses examined those subjects (n = 46) who self-reported cholesterol lowering medication usage (See Appendix B, Tables BI through B4). These 46 subjects were significantly older, and had tended to have higher mean BMI, serum glucose, triglycerides and liver enzyme tests than the S06 subjects who did not report cholesterol lowering medications. As seen previously, the Cottage Grove and Decatur employees were significantly more likely to have higher BMI values than Antwerp employees. Multiplicative models did not indicate, however, PFOA to be associated with the independent variables presented in Table B4, including cholesterol, LDL, HDL and triglycerides. Appendix C (Tables C1 through C9) contains the combine(dN = 552) analysesofthe 506 male employees who did not report cholesterol lowering medications (Table 1 through Table 24) and the 46 who did (Tables BI through B4 in Appendix B). Analyses were not substantially different whether they excluded (Tables 1-24) or included those employees who self-reported cholesterol lowering medications (Tables C1 through C8). For cach analysis presented in Appendix C for the 552 male employees, there is a comparable analysis for those employees who did not take cholesterol lowering medications. For example, Table C1 is the PFOA decile distribution analysis for all 552 `employees, regardlessof cholesterol lowering medication status. The 46 employees who took cholesterol lowering medication are included in the deciles usedforthose employees who did not take such medication. See Table 7 for the comparable analysis for only 3M CoPmapgean2y7 those employees (n = 506) who did notselfreport cholesterol lowering medications. Findings did not differ. Presented in Tables C2 and C3 are the odds ratios by decile for the lipid out-ofreference range values for all male employee participants (N = 552) regardlessof cholesterol lowering medication status. See Tables 124 and 12B for the 506 employees who did not self-report cholesterol lowering medications. Results were `comparable. Table C4 presents the odds ratios by PFOA decile for ALT and GGT values out-of-reference range for all male employee participants (N = $52) Comparable analyses for the more restricted subsetof 506 employees who did not take cholesterol lowering medicatioanres found in Table C7. Findings did not differ. Tables C4 through C6 contain both the non-prescribed (n = 506) and all male participants (n = 552) regression coefficientsfor the lipid (Table C4), liver enzyme (Table C6), and thyroid (Table C7) analyses. Analyses from the multiplicative models were not substantially different whether all male employee participants, or only those who did notselfreport cholesterol lowering medication usage, are included. The few employees among all `participants (n = 552) who had out-of-reference range thyroid tests are presented in Table C8 (see Table 21 for comparison purposes for the 506 employees who did not take cholesterol lowering medications). Table C9 (as was previously done in Table 24 for `Table 22) are the predicted thyroid test values for the regression model coefficients that were shown in Table C7. These results in Table C9 again showed that an increase in serum concentrations of PFOA, as high as 100 g/mL, would not result in out-ofreference range thyroid clinical chemistry tests. 3M CoPmapgean2y8 PFOA and PFOS serum concentrations were highly correlated among these employee participantsofthe 2000 fluorochemical medical surveillance program (Appendix A). For purpoofbrseveitsy, PFOS-related graphs are presented in Appendix D for the 506 employees who self-reported they did nottakecholesterol lowering medications. Also, found in Appendix E arethecoefficientsfromthe multiplicative models examining the association for PFOA, as was done previously for PFOA for lipids (Table E1), liver enzymes (Table E2)and thyroidtests (Table E3) for the combined three facilities (Antwerp, Cottage Grove, and Decatura)s well as each facility separately. As seen with PFOA,therewas a positive association with PFOS and triglycerides which is inconsistent with the known toxicological effectsof PFOS. This association was likely due to the fact that higher PFOS concentrations were observed amongtheCottageGroveand Decatur employees than Antwerp employees, the latter had the lower serum triglyceride values. No association with PFOS and triglycewrasisdeeens when Cottage Grove and Decatur employees were analyzed separately as shownin Table EI. A positive association was observed with triglycerides and PFOS among Antwerp workers only, similar to what has already been reported for PFOA in this subgroup of workers (Olsen et al. 2003). Inconsistent findings were observed for ALT and PFOS as there was a statistically significant negative association with Antwerp employees and a positive association with Decatur employees (Table E2). Other liver enzyme tests provided no consistent pattems ofassociation. Thyroid related tests also provided no consistent measuresofassociation in the multiplicative models presented in Table E3. 3M CoPmapgean2y9 Discussion Based on an analysis of 506 male participantsofthe 2000 fluorochemical medical surveillance program offered at the Antwerp, Cottage Grove, and Decatur facilities who self-reported that they did not take cholesterol lowering medications, therewas no indication that these employees" serum PFOA concentrations were positiveolry negatively associated with serum total cholesterol or LDL. A weak negative association was observed with HDL that was likely due to uncontrolled (ie. residual) confounding, based on lower HDL values observed among the Cottage `Grove and Decatur workers than the Antwerp workers and their markedly different demographic factors (e.g., BMI). When the analyses were stratified by location, no statistically significant associations were observed between HDL and PFOA. In another occupational study, Leonard (2005) did not report an association between HDL and serum PFOA concentrations. Neither did Emmett (2004) in a community-based study. To further clarify any possible association with HDL, the A apolipoproteins, which form the major proteins found in HDL, couldbe measured although this was not part ofthe present study. Apolipoprotein A-l was not associated with serum PFOA of comparable concentrations in a small analysisof Italian production workers (Giovanni Costa, personal communication). Serum triglyceride levels were positively associated with PFOA but itis likely that tis association was also due, atleast partially, to residual confoundingas a consequence of the a disproportionate numberofCottage Grove and Decatur employees with slightly 3M CoPmapgean3y0 higher PFOA concentrations than the Antwerp employees. The same positive association was also observed between PFOS and serum triglycerides which are entirely inconsistent given the well-established hypolipidemia reported in PFOS-related toxicological studies whose concentrations were much higher than the present study (Seacat et al. 2002; 2003). Ofthe four lipid measurements used in this study, serum triglycerides have approximately 3 times more intraindividual biological variation than either total cholesterol, LDL, or HDL (Stein and Myers 1994). Serum triglycerides may also be influenced by obesity, alcohol intake, and inattention to fasting requirements for blood collection. We hypothesize that the associations observed for serum triglycerides and PFOA in these workers might be the consequenceofthe non-adherence to fasting requirements by some shift production workers and/or the effectofpostprandial metabolic responses in shift workers. Our explanation follows. First, production workers would be morelikelyto have higher PFOA concentrations than non-production workers (e.g, administrative, laboratory, engineers). This has been showntobe the situation at all three facilities: Antwerp (Olsenetal. 2001); Cottage Grove (Olsen et al. 2003b) and Decatur (Olsen ct al. 2001; 2003). ~ Second, only production workers engage in shift work. Third, required fasting blood collection occurred only during the morningforall participants. Fourth, itis conceivable that the night shift production workers, compared 10 day shift production workers as well as non-production workers, could have been less compliant with fasting requirements as well as consumed caffeinated beverages prior to their morning blood collection. Fifth, several studies have indicated postprandial serum triglyceride levels are higher among night shift workers (Al-Naimi et al. 2004; Morgan et 3M CoPmapgean31y al. 2003; 1998; Karlsson et al. 2001; Lund etal. 2001; Sopowski etal. 2001). Therefore ifa subsetofsubjects (production workers with higher PFOA serum concentrations) who worked night shift were less likely to adhere to the fasting requirements and/or have `postprandial metabolic profilesofnight shift workers, then a non-causal positive association between PFOA concentrations and serum triglycerides could be observed when all subjects are included in the analysis. Atthe timeofdata blood collection we did not obtain shift status information and therefore within this database we are unable to further address this methodological question. Countering this possible hypothesis is the fact that blood glucose, also requirinag fasting sample, was not associated with PFOA at any site. However, hyperglycemia has not been consistently associated with nigh shift workers (Al-Naimietal. 2004; Karlsson etal. 2003). Also, the association between serum triglycerides and PFOA was not observed atthe Cottage Grove site which had both production and non-production workers. Nevertheless,a positive association between serum triglycerides and PFOA was reportbeyd Leonardetal. (2004) whose study `population also consistedofproduction (shift workers) and nonproduction (nonshift workers) participants. To further clarify possible association,ifany, between workers" PFOA concentrations and serum triglyceride levels, adjustment for shift work becomes a methodological necessity in the data analyses. Adjusted for BMI and/or serum triglycerides, which are important potential confounders in the analysis of iver enzymesastheyareassociated with the increasingly prevalent nonalcohol fatty liver disease observed in the United States (Clark et al. 2003), there was. no indication that the measured liver enzymes (alkaline phosphatase, AST, ALT, GGT 3M CoPmapgean3y2 and total bilirubin) were causally associated with employees' serum PFOA concentrations. A weak positive association between alkaline phosphatase, ALTand `GGT with PFOA was observed among the Decatur male employes, however, several epidemiologic research studiesofthis workforce, past and present, have not reported `associations with hepatic disease (malignant or nonmalignant conditions) using a variety ofdata sources including death certificates (Alexander et al. 2003), episodesof care (Olsen et al. 2004) and self-reports (Alexander and Grice 2006). Neither have other `worker populations (Leonard et al. 2004) or a community-based PFOA-exposed `population (Emmett 2004) reported positive associations between liver enzyme test results, liver disease and serum PFOA concentrations measured at or below those reported in the present study. For example, in the Emmet study (a presentation is available athttp://www.lhweSstudy.org/index.him), the median serum PFOA concentration measured in this community-based studywas 0.34 pg/mL. Individuals older than 60 years had significantly higher levels of PFOA compared to all groups except those lessthan six yearsofage. There was no relationship between the blood levels ofPFOA and the results for cholesterol, liverfunctiontests (serum protein, albumin, bilirubin, serum alkaline phosphatase, AST, ALT and GGT), kidney function tests (BUN, creatinine) and TSH. Furthermore, there was no relationship between PFOA serum concentrations and being treated for or informed by a physician that a community participant had liver disease (cirrhosis, hepatitis, and any other liver condition). In the present study, there were no consistent associations with TSH, T4, free T4 or T3 with PFOA (or PFOS) across the individual facility locations. For the combined location 3M CoPmapgean3y3 analyses, adjusted PFOA coefficients that were statistically significant were the result of `models whose range of predictions for serum PFOA concentrations (up to 100 g/mL) `were well-within the normal reference ranges for these thyroid related tests. This was not `unexpected as few thyroid-related tests were out-of-reference range. The lackofthyroid related hormone associations is also consistent with the worker data reported by Leonard etal. (2004) and apopulation-based study by Emmett (2004) that included thyroidrelated conditions (hyperthyroidism, hypothyroidism, goiter). In rats the presence of PFOS or fatty acids suchasoleic acid in serum appeared to compete with free T4 for protein bindings, and their presence in serum resulted in anegative bias in analogfreeT4 `measurements (Tanaka etal. 2005). PFOS, however, did not reduce cither free T4 by equilibrium dialysis radioimmunoassay (ED-RIA) or the liver response to thyroid `hormones [e.g., malic enzyme and uridinediphosphate-glucuronsyliransferase (UGTIA)]. These observations therefore suggested that prior reports of reduced free T4 in the presence of PFOS may have been artifactsofthe analog methods. It is conceivable this negative biasofthe assay could also effect measurement ofPFOA. 3M CoPmapgean3y4 References A3cMidCoanmdpaItnsySa2l0t0s3.. USHeaElPtAhadnodckEentvAiRr-on2m2e6n-t1a4l86A.ssessment of Perfluorooctane Sulfonic Alexander BH. 2001. Mortality studyofworkers employed at the 3M Cottage Grove facility. Minneapolis 1030a018. (MN):University of Minnesota. US EPA docket AR-226- eAlmepxlaonydeeersoBfHa, pOelrsfelnuoGrWo,octBaunrersiuslJpMh,onMy!anfdleuolriJdHe,mMaannudfaecltJuSr.in2g00fa2c,iliMtyo.raOlcitcyupofEnviron Med 60:722-729. Alexander BH, Grice M. 2006. Self-reported medical conditions in opferMfilnunoersoootcat.aneUsuSlfEoPnyAl dfolcuokreitdeARm-a2n2u6f.acturing workers. Minneapolis (MN):University AlNaimi S, Hampton SM, Richard P, Tzung C, Morgan LM. Postprandial metabolic profiles following meals and snacks eaten during simulated night and day shift work. Chronobiol Int 21:937-947. JBPr,oRuasdseearuDJM.E,20S04c.haeEffefrecEtJs,oWfoalnfeinMhiLb,itBorloofedcohnolLeTst,erDyilgeesntieorAtrGa,nsCflerarpkroRteWi,n oMnanHcDusLo cholesterol. N EnglJMed 350:1505-1515. LBiuetdeernhPo, fOJfl,sCeonsGt,a TGh,oEmlfcoormdbeP.C,20F0a2r.rarToDx,icHiatnyosefnaKm,moIwnaiiuHm, JpuernfgluRo,rKooecntnaendoaytGe,in `male cynomolgus monkeys after oral dosing fo6r months. Toxicol Sci 69:244-257. aCmlianroktJrMan,sfBerraanscaetlievFeLl,s DiincthhleAUnMi.ted20S0t3a.tesT.heApmreJvGaalsentcreoeanntdereotliooglyog9y8o:f96e0l-e9v6a7t.ed Collantes R, Ong JP, Younossi ZM. 2004. Nonalcoholic faty liver disease and the epidemicofobesity. Cleveland Clinic J Med 71:657-664. Costa G. 2004. Letter. US EPA docket AR 226-1868. DePierre JW. 2002. Effects on RodentsofPerfluoro fatty Acids. (In) The Handbook of AEHnv)i.rVonomleunmteal3,Chpeamgiesstr2y0.3-2P4ar8t.N. Organofluorines. Berlin:Springer-Verlag (ed Neilson Emmett E. 2004. Community C8 Study. Little Hocking Water Service Area. Available at http//wwwJhweSstudy.org/indexhim. 3M CoPmapgean3y5 EExdpuecrattiPoannePlr.og2r0a01m.(CENxeCcEuPti)vEexSpuermtmPaarnyeolfotnheDeTthecitridonR,epEovratlouafttihoen,NaatnidonTarleaCthmoelnesttoefrol High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 285:2486-2497. Ford ES, Giles WH, Dietz WH. 2002. Prevalence ofthe metabolic syndrome among US adults. JAMA 287:356-359. Gilliland FD, Mandel JS. 1993. production plant. J Occup Med Mortality among 35:950-954. employeesof a perfluorooctanoic acid lGiilpliolparndotFeDa,inMdsancdhoellesJtSe.rol1:99a6.stuSdeyroufmocpceurpfaltuiooronoacltlayneoixcpoasceidd amnedn.hepAamticJ eInndzyMmeesd, 129:560-568. Grundy SM, Brewer HB, Cleeman JI, Smith SC, Lenfant C. 2003. Definition of `metabolic syndrome. Circulation 109:433-438. Han X, rat and Snow TA, Kemper RA, human plasma proteins. Jepson Chem GW. 2003. Res Toxicol 1B6i:n7d7i5n-g7o8f1.perfluorooctanoic acid to pHearufgluhoormoocBt,aSnopiycdeavcoidld(POF.OA1)9,92p.erTfhlueormoeocchtaanneissmulupnhdoenrilcyiacnigdt(hePFhOypSoAl)ipaenmdiccleiffifbercitcof acid. Biochm Biophys Acta 1128:65-72. aHepmrierenquMiAsi,teHfeomrcnodnefzo-unDdiianzgSe,vWaleuralteironM:Ma,n aMpiptlcihcealtlioAnAt.o b2i0rt0h2.defCeacutssaelpkindoewmlioeldoggeyas Am J Epidemiol 155:176-184. aKlaspolUanSAEMP.A 2d0o0c4k.etAAmRm22o6n-i1u8m67PearnfdluAoRr2oo2c6t-a1n8oa6t8e.. US EPA SEHQ-0904-00373. See Karlsson B, Knutsson A, Lindahl B. 2001. Is thereanassociation between shift work `paenodplhea.viOcnacgumpetEanbvoilriocnsMyenddrSo8m;e7?47-Re7s5u2l.ts from a population based study of 27,485 iKnarmlaslseonwoBrHk,erKsnuwtitshsornotAatKi,ngLtihnrdeaeh-lshiBfOt,woArlkf.redRsessounltLsSo.ft2h0e03W.OMLeFtabsotuldiyc. diIsnttuArrbcahnces Occup Environ Health 76:424-430. LKBe,nnMeudrypGhLy,SBR,utFaernrahroDJfGLf.,OTlhseentoGxiWc,olOo"gyCoofmpneorrflIuCo,roSoecatcaantoAatMe,. PCerriktiicnaslRRGe,vBTioexgieclol 34:351-384, sKuelrfsotnnateer,-Wpoeorfdl,uCo,roChoexwaanredsLu,lfGonoartme,anpeGrf.l2u0o0r3o.octParnoetesiunlfboinnadtienagonfdppeerrfflluuorrooboucttaanneoate to 3M CoPmapgean3y6 Sploaustmhaer(hRuemsaena,rcraht,Inasntditumtoen.kSetyu)d,ya9n9d21v.a7r.ioUusShEuPmaAn-ddoecrkievteAdRp-l2a2s6m-a13p5r4ot.ein fractions. Lakka HM, Salonen JT. L2a0a0k2s.onTehneDmEe,taLbaoklkica sTyAn,dNriosmkeanaennd LtoKta,l KaunmdcpaursdailoovaEs,cuTluaormdiilseehatsoe J, mortality in middle aged men. pLeerofnlaurodroRocCt,anGoraetee:n cJrWo,ssK-sreecctimoannalKsHu,rvReielylnaonlcdeosfJc,liilniiceaslPmSe.as2u0r0e5.sofAgmenmeornailuhmealth status related t0 a serum biomarkerofexposure. Newark (DE):DuPont Haskell Laboratory for Health and Environmental Sciences. May 31,2005. Lund J, Arendt J, Hampton SM, English J, Morgan LM. 2001. Postprandial hormone and metabolic responses amongst shift workers in Antarctica. J Endocrinol 171:557-564. Mofrad PS, Sanyal AJ. 13. 2003. Nonalcoholic fattyliverdisease. Medscape Gen Med 5:1- TMaoyrlgoarnK.L, 1A9r9e8n.dtEJf,feOcwtseonfsthD,e eFonldkoagrednSo,usHacmlopctkonanSd,sDleeeapcotinmSe,oEnngmleilsahtoJn,iRni,beiinsruoliDn,, glucose, and lipid metabolism. J Endocrinol 157:443-451. Mukai M, Ozasa K, Hayashi K, Kawai K. disorders and related physical conditions anVdarliifoeu-sstySl-e.GODTi/gSD-iGsPSTeira4t7i:o5s4i9n-5n5o5n.viral liver tNhaekacnoinsvheintNi,onMaaltisnudoireYc,tYmoentehdoadsHi,nNclaukdaimnugrFariKe,deSwuazludkimeKt,hToadtafroar dKe.te2r0m0i0n.inVgalsiedirtuymof elonwz-ydmeantsiictyanlailpyospirso.teJinOcchcolHeesatletrhol42l:ev1e3l0:-c1o3m7p.arison with the direct homogeneous eOplisdeenmGioWl,ogGiiclliinlvaensdtiFgDa,tioBnurolferwepMroMd,ucBtuirvreihsoJrMm,onMeasndienlmeJSn, wMiatnhdoeclcuJpHa.ti1o9n9a8l.eAxnposure: 0 perfluorooctanoic acid. JOEM 40:614-622. aOlnsdehnepGaWti,c BaunrdrilsipJiMd,clMinaincdalelchJeHm,isZtorbyeltesLtRs.in1f9l9u9o.roScheermuimcaplerpfrloudourcotoicotnaneempsluolyfeoneast.e JOEM 42:799-806. hOelpsaetnicGeWn,zyBmuersr,iscJhoMl,esBtuerrollewanMdMl,ipoMparnotdeeilnsJHin.a2m0m00o.niPulmaspmearfclhuoolreocoycsttaonkoiantien and production workers. Drug Chem Toxicol 23:603-620. sOlesreunm GpeWr,flBuourroloecwtaMneMs,ulfBounrarties (IPMF,OMSa)nadnedlpJeH.rfl2u0o0r1oao.ctaAnocartoess(-PseFcOtAio)nailn raenlaaltyisoinstoof clinical chemistry, thyroid hormone, hematology and urinalysis results from male and 3M CoPmapgean3y7 sfuermvaelielleamnpcelopyreogerpaamr.ticSitp.aPntasuolf(tMhNe)2:030M0 CAnotmwpearnpy.anUd.DSecEaPtuArdfolcukoertocAhRe-mi2c2a6l-1m0e4d7i.cal Olsen GW, Burlew MM, Burris JM, Mandel JH. 2001b. A longitudinal analysis of tsoerliupmidpearnfdluhoerpoaoticctacnleisnuiclafloncahteemi(sPtFrOySt)esatnrdespuelrtfslfuroormoomcatalneoeatmepl(oPyFeOeAp)arlteivceilpsainntsoreflatthioen (1M99N4)/9:53,M1C9o9m7paanndy.200U0.S.fluEoProAchdeomcikceatlAmRe-d2i2c6a-l10su4r8v,eillance program. St. Paul JOKl,seMnaGndWe,lSJcHh.mi2c0k0l1ecr. MDNe,scrTiipetrievnes sJMu,mmLaorgyaonfsPeWr,uBmurfrliusoJroMc,heBmuirclaelwlMevMel,s aLmunodnbgerg epmropglroayme.e pSat.rtPiacuilpa(ntMsoN)f:t3heM YCeoamrpa2n0y0.0 AUn.tSw.eErPpAflduoocrkoectheAmRi-c2a2l6m-e1d0i3c0aal02su0rbv.eillance Olsen GW, Logan PW, Simpson CA, Burris JM, Burlew MM, Lundberg JK, Mandel JH. p2a0r0t1idc.ipaDnetsscofritphteiveYesaurm2m0a0r0yDoefcsaetrurumflfulouroorcohcehmeimciaclalmeldeiveclaslamsuornvegilelmapnlceoyperoegram. St. Paul (MN):3M Company. U.S. EPA docket AR-226-10302020a. Olsen GW, Burris JM, Burlew MM, Mandel JH. 20033. Epidemiologic assessment of c`woonrckeenrtrsaetriuomnspaenrdflmueodroioccatlanseusrvueliflolnaantcee(ePxFaOmSi)naatnidonsp.erfJlOuoErMooc4t5a:n2o6a0t-e27(0P.FOA) Olsen GW,Butenhoff JL, Mandel JH. 2003b. Assessment of lipid, hepatic and thyroid fPuanuclt(ioMnNi)n:r3eMlatCioomnptaonayn.ocUcuSpaEtPioAnadlocbkieoltoAgiRc-2l2im6i-t1v3a5l1u.eforperfluorooctanoate. St. OVlesneknatGeWsw,arLlougPa,nSPcWh,umHpaenrsteJnC,KJM,aSnidmeplsJoHn.C2A0,03Bcu.rriAsnJoMc,cuBpuartlieownaMlMex,poVsourraerath PP, Jasosuensaslme6n4t:6o31fa-6p5e9r.fluorooctanesulfonyl fluoride production site: biomonitoring. AIHA. Olsen GW, Burlew MM, Marshall JC, Burris JM, Mandel JH. 2004. Analysis of e8p4i6.sodesof care in aperfluorooctanesulfonyl fluoride production facility. JOEM 46:837Rickard RW. Re: PFOA Risk Assessment Science Advisory Board. (Letter to Dr. Suhair EShnavlilraolnomfenUtSalEPSAci.e)ncNese.waJrukne(D29E,):2D00u5P.ont Haskell Laboratory for Health and Ruhl CE, Everhart JE. 2003. Determinantsofthe associationofoverweight with e1l2e4v:a7t1e.d79s.erum alanine aminotransferase activity in the United States. Gastroenterology 3M CoPmagpean3y8 SSeuabccahtroAnMic,tTohxiocmiftoyrsdtudPJi,esHoannspeernfKlJu,orOoloscetannGesWu,lfoCnaastee MpoTt,asBsuitumensahltoJfinLf.cy2n0o0m2o.lgus monkeys. Toxicol Sci 68:249-264. JSLe.aca2t00A3M,SuTbh-ocmhrfoonridc PdJi,etHaarynsteoxnicKiJt,yoCflpeomteanssLiAu,mEpledrrfildugoerSooRc,taEnlecsoulmfboenaCtRe,inBurattes.nhoff Toxicology 183:117-131. tSroipaocywlsgkliyceMrJo,l HreasmppotnsoensSiMn,siRmiubleaitreodDnCig,htMoarngdadnayL,shAifrte:ndgtenJd.er20d0i1ff.erPeoncsetsp.ranJdiBailol Rhythms 16:272-276. `STteeixntbEoAo,koMfyCelrisniGcLa.l Ch1e9m9i4s.trLyi.pid(se,dsliBpuorptriosteCiAns,,AasndhwaopooldipEoRp)r.oteiPnhsi.lad(eIlnp)hTiiae(tPzA): W.B. Saunders Co. pp 1002-1093. `MTeaansaukraeSm,eTnhtiobfotdheyaruoxidJ,hEoarsmtovonledsMi,n rBajtosrekraJ,coEnhtraeisnminagnpDe,rfBluuotreonochtoafnJfe.su2l0f0o5n.ate (PFOS). Toxicologist 8485-1) 423 (abstract 2066). UflbueolroFcAh,emSiocraelnss-onaSpDr,elRimoiancahryDEre.por1t9.80A.mHeIanldtHhystgatAusssoofcplJan4t1:w5o8r4k-e5r8s9.exposed to XpieeroYx,isYoamnegpQro,liNfeelrsatoinonBaDn,dDaedPiipeorsreetJiWss.ue2a0t0r3o.phTyhiendruelcaetdiobnyshpieprboextiwseoemne lpirvoelriferators exposure and withdrawal in mice. Biochem Pharmacol 66:749-756. `Table 1. Number (percent)ofMale Employees Who Self-Reported Cholesterol Lowering Medications, By Location, 2000 Fluorochemical Medical Surveillance Location Antwerp Cottage Grove Decatur Total Choleseterol Lowering Medication Yes (%) No (%) Total 1005) 196 (95) 206 97) 122(93) 131 2703) 88 (87 215 468) 50692) 552 "Tabl2e. Mesn, StandardDeviatBiyonL,ocMaetdiiosn,na2n0d0R0aFnlugoeroofchePmFiOcaAl,MPeFdOiSca,lDSeurrnvoegirlalpahnicce FParcotgorrasmnd Clinical Chemistry Results, en he POA LOZ 106 065 001704 POS 095 097 055 004624 I Ws 3% as BMI 27130 25 1537 Aeohol LIP LI 09 0064 Ghose ST 178 3LBL Cholesterol 207 41 216 105331 LoL new ans HL sas mo 2% Tigheides 120) $3 100 3231 AkPhos @0F 14 6 2m As 26 2 ms ar 2 0 a sm ASUALT 108) 012 12 081447 en 46M 1253 095 086 098 045 a 9 a0 29% 48 95 07 07 05 100 23 9s 210 3 206 0 Bs a nas 187 139 12 6 as 6 a8 4 Woon 098 009 092 re 0019208 003479 ug 197520 0030 seas B03 32s wm ms si lest pes 070425 ven" 189 161 129 092 @9 286 44 01 03 9M 14 24 4 Be 3% a 10 182 10 BY 0 68 M16 095 011 iin ee 1S 008270 100 006417 8 2%6 25 172501 00 0020 91 Ts 21 121319 1m ans 2 ue let 327% 26160 5 D6 0 en 093 075.18 ManAn(StD)e(NM=e1d96) Range Table? (Continued) MenCota(gSeD)GrovMeo(lN=122)Rage aor TowlBil DirtBil Si Ek Fees n 2 ns 10 03 10 OF 005 01 20 16 17 52 14 82 LP 02 LI BP 0 em 0423 0004 00M94 42120 0616 ss Wom wm mu 09 03 08 0423 01 002 01 0002 24 14 20 0094 280 11 79 49AL6 LI 01 LI 0818 Bs 0m maw 21.. SSutattiissttiiccaallllyyssiiggnniiffiiccaannttllyy((p<<0055)) ddififefrntftthehaannrACnoettwnaegrtepGrove: 3. Statisticallysignificantly (p < 05) difethranDeecnattur Mem [ (SD) a Med Range Wn ow am 07 02 07 03s or 007 01 0007 28 52 19 00363 8 14 8S 46d LI 02 LI 0615 ns 2 20 se1% Te NCorbetrsnPCronmtofERpsos00b8yFLocataioncReeferlsPSoenmseolfDeemFuogggainc crs Arey ConsGone Der fone Be B00 pune wz sw sw a we Jry---- 2 ow oe ow rw ew [--_-- 0 o 2 a 6 oo Cuiesetz200ml wow ee wow om Cnewozzoma @ oO) nan 0 oo WoL<n 6 % oo noo mo LoL omg we sw w ew Tevez 6 @ on 6) me om Aor mor I) oo soo a Astasomn oo "oo rw om aso rw ao 2 a ee carson soo sw w = Toatbi> 13mg soo oo oo ma Dist > 0 mpl oo oo row ow TM"oss um Co I Ca ow sss so sw soe oe ig <4Sug nou FeStt a iso army. Sop oo 1 [0] oe TY I) ow 0 ow oo Ce I soo 0 46 oom FERC 0 ow oo ow rw ow Metabolic Syndrome" 2 O + ieorrs-- tat stone 2 20) 28 a3) 0001 Tc. Men,SdONDti,HM o hnedd RFar eof nPOA,o10Sl,DeFopgcrsrntdyCi1l1l9C0his Re, E S530 et iea v3i p.Nor ir hn re ET Me Wn ow omm ow 7% ss Ws moan x 7 ou nw ma mes ewsur 22 30 Me Name MP 3 MUSES HS 2 M2 9S Nebo 10 0 03 03 10 12 a7 esas 12 Mw ees 12 uu ess FE Gu moo usw am wom mm Ww bow wsw Bs 6m we WLW ow wan mum was me we meas ww nw sae EP Weow @ mows om mw sem ms mam we won ses FE FE MT ul wwe moron we wx wos Wn 2 oem ASAT LG M01 0SZ LB OIL GAN M6 0M 10OEAR 104 OM IN ome GW 2 wows x mom oem om mn % on ow Tim to 03 10 0s: Go od lo 6s3 le 6s 10 0s: 0 03 10 ose Dili 01 om 01 002 01 omar 600s 01 aban eos 1 om oi 0002 Tu un a7 esse 1s os te omas 1s wos ase 4 21 1s ois (CTonaieno.ct) Q Nom uSD MedlRange NQwuuSDaM(otNo Rdin_nge M_eQm_uuSD.kMiotMi uRannge_ _NiQe_nuuSD.kMiinniRinmge TBA 0S 8S SONS BI 14 79 4220 So FTL 02 1 0s1s 12 02 12 094s AF Do nS Deol Lo lenwm am La fo sed 83 1s 8s 4700 02 Loe LI 01 LI osie moms nem1 eau 251 SSSiaosttaalayssgincgiainnciatfnyyy ((pp<0<5008))9ddifirlocrnts hhhnnn213qqquauurtie Subclp<i09n)diiethcn a* qulate Table. Morhn,ySTuROdNQuDaenin3,00M0PdocshdeRar MofedPFaOSA,ePOaStesPgegn 1Faocttor anGdrCovleic1-C1h5e9ssyRess, EC Fe i Ra ie Woetp a vib ew ae T New 3p. T denRe HOS 06 0% 01s0mae LN IR 08 024 0m 0 055 0oM2 0 am 0st ona Me Ws moma woe ow ma wis uw ne wo a ms BO 34s Me wAen 1 47 32 Dems Bo Se 31 Dest doo 36 ds mean Aki 05 0S 05 0020 0s ox 0s aslo 01 0s os si 07 06 03 0020 Game 00% mam ow ow ow mam wm nw was wu es sa GoW me Mem au 0 0 eam we wo ms isa a Ms tea Wem on mss mon mowm mono pa pe 3 ns wm Won we mm wn ow me 6 nw se 8 0 a we Tew ms moe woe ss ssa me wm ose ems wm mom ss Mies ne un 6 6 Bm & 0 6 wm @ 1s 6 we ATW was owe ow 7 x sw 3 0 3 ws ws nu AT 3m mous wom 0 ws nm wom nm mu iss ASTALT 09% 000 035 084135 031 009 033 075000 03 0m 091 0013 02 00 03 OFAN Grow os omoam mom om ww mw um x ww ww TwiB 09 02 08 0415 09 03 69 0623 0s 02 03 o4i4 03 03 03 oais DiaBi0l1 0m 01 002 01 0m a1 orez a1 oo 01 oil 01 om 61 600s T2410220850 2312 20 esee 22 131 vers 2s 1s20 emas (Cboentse) E New e SD violfonpe Vo iew Sbw Mwe forge_ No ew SDwon n Forge_ i_Qc_ uSDaMeki adne TB BeTOU Dm 4 m2 M6 02 10M um wemm S012 79 4sam0 79 08 78 e297 7 1a 7 seas LL 01 Lene 1 01 Wes Lol b emid wow wm osu wow one som om us mae 231. SSSaicaakaiatlllysigggnaiafitacaanntly (7(0p<<<00059))ddicierttfhhhanaon321"eqquruatttiee Sataysignin < 05)lono 1 quate ale. Men, SdONDSeaii,oMmPeondiReaentof PeFtOeAvPROSe, esmFegnFSanrdrC8lsi1c50Chay Ress, Fi 0 inoS.ih Ri WSE.Eone EiR s 3bE . od Re MOS 0SiTMO 04s 0061s ORC 0510 GIS 12S 035 LIS 03S 25 03820 064M Me Won wma wx ow ue wos um we we WU me 41 moures wa 49 we Mbml Ws Ss ma naam Ms 49 m3 waa J Tr i Guoe mW mosis wm wo wm % mow ms ow now sm Cosine? 0 aes Be @ 2 wes 4 maw wa saw Wor os mews wr wow mm mw we a bs om eas Wow on oaome 8 sok uw 4 wow de @ sw wn me WR mow Ww Mes mes m0 ns wm mus we ew MP WB em HW mos ow 7 Wn we noe wn AT ws moss m7 ou mm 3 6% wk nwo ise ATW us mone a wom ma 3 BoM ws 00 8% ew ASTALT 0 010 03 0madn 035 om 0% 0750 0% 0m 0s 0713 0m 010m omim Gr wow mow W Boa um ww mow nw nw ew Tem or 02 0s ovis 0s 0207 oez 7 0207 oma 07 02 01 oa Diallo 0m 01 002 01 omar 0802 a1 001 0007 01 OWOI 0006 TH 3S a31somesa 26 2619orms2 16is ode 21 3122lena (CTonatees) C Mow u SD Mol k Range TRA 0S 87 46ND BTU a1 0613 = io : Mo em SD Mw eow Range BA LA 83 SONA LI 02 Los EEL G New Su D Mell Farge N_ew omSuDkiV__Rone 82 16 86 IMI 86 Le 8S Sea LL 02 LL oes LL 02 11 os3 TR WI a O 231 SSSmititalallsiignniifnfiiccaanttly0(ip<<< 0009)idirrieatfthiomn2c3*qqiuuaarlte 4 Stisialy gnicanty ( 05)diomr1 quear Table 7. NumberofEmployees, Mean, 95% Confidence Interval, Median and Range by PFOA Decile Distribution , 2000 Fluorochemical Medical Surveillance Program for Antwerp, Cottage Grove andDecatur(N = 506) PFOA Decile N 1 st 9) s1 3 st 4 50 5 st 6 49 7 50 8 54 9 49 1050 Mean 006+ 020 036" 0.551 091" 126 164 21m 300 12.15" 95% CL PFOA 005-007 019-021 035-037 053-057 087-094 123-128 1.60-1.67 212223 2903.10 721-1710 Median 006 019 036 054 091 125 1.63 2.18 296 494 Range 0.007-0.13 013-029 0300.44 044-071 072-110 111-140 142-185 1.86-2.50 2513.69 3.71-92.03 1 Statistically significantly (p <.05) different than PFOA decile(s)1,2,3 . . and/or10 "Table. DistbutionofDemographic Factors by PFOA De200c 0Fuorl ochem,ical Medical SurveillanceProgram(N =506) PFDeOciAle Awerp CotctaagteioGnrove Decatur MAegaen(SD) MeBmMnL(SD) BNMI0>=030 MeDimn(kSdDd)y >DN=i_i(n9k)iday '! Tose ney 609 Man 2aEy Bey 071 ON 2A Mn 2 BM) BE BEY He 076) 70) 0709 17 GY) 3OMeH NE) Re) WE 26409 8 (9 07 00 15 09) 4 0a BE EH M00) 28H BEH 07 AF 13 Qo) ; 5 o2Ey NE BEY W000 2260 8 aD 0740 16 G2) : 6 AEH Tay NE) WE 24EH Re) 0740 14 @) 7 En) 9 ay Ey WE 26909 10@) 0709 14 ey 5 BE) NQ) 2566 4140) 27564 170) 06 09 16 GO) 9 9a Tan WED WO BINT A@) 03 09 5 a0 00S an BEY 26H DEO WEY 156) 050) 12 Qh Sttisically significpanetloyo(po<"05) diferent than PFOA decile 1,2,3.. 10 pent pea" 10valueassociatedwithhi square estfolocaton, BMI disebutanodn> | rink/day, respectively. : Table 9. Mean, 95% Confidence Interval, Median and RangeofPFOS Serum Concentrations by PFOA Decile Distribution, 2000 Fluorochemical Medical Surveillance Program (N = 506) PFOA Decile N Mean PEOS 95% CL Median Range 1 s1 030% 0.15-046 013 0.02264 2s 061%" 0380.84 035 0,025.19 30 81 086 059-113 052 0.13479 4 50 078 062-094 058 012-341 ss) 089-710 0.74-1.04 082 006-221 6 49 paghEa 090-133 0.98 0024.22 7. 127% 1.07-1.48 121 0183.25 gis FAY 1101.73 1.03 020-624 9 49 1.74% 143-206 183 0.124.386 1050 1.58% 123-104 1.06 014-431 11 Statistically significant(lpy < .05)different than PFOA decile(s) 1,2, 3 . . . and/or 10 Table 10. Mean and Standard Deviation of Serum Lipid Clinical Chemistry Results by PFOA Decile, 2000 Fluorochemical Medical Surveillance Program PFOA Cholesterol LDL HDL Del Ry SROD ee ED) 1 21543) 13739) 5013)" 2 aw) 13733) 203" 3 20837) 12831) 52014)" 4 210 (41) 5 2869) 6 296) 13234) 14135) 14141) 50(19) 4902" 4902" 7 214 (40) 133 37) 5011)" 8 216 (43) 9 2860) 10 2142) 136 (40) 13831) 13235) 47(13) Asan?' 8097 Triglycerides SED 1492101)" 12066) 144 (121)" 147 (101)" 16117)" 161 (104)" 154.(106)' 175 (109 17493 214168)" 11 Statistically significantly (p < 05) different than PFOA decile(s) 1,2,3 . .. and/or 10 TabfolreL1i.pidAdCljiunsitceadl"ChMeemainsarnydR9e5s%sCobnyfiPdFenOcAe DInetceirlveals PDFecOilAe Mean Ch9o5le%siCelro, New DLosc Men Hp) oR ' 21 was 13 naa 0 653 2 a 922 ns sus si nse 3 0m 220 us ngs si ast 4 20 wm nas so 65 5 a 2628 10 Boise a asst 6 28 0629 1a Bost a asst 7 a mas im nan 50 53 8 as m6 136 ness 41 450 9 om a2 0 Boise a "2 0 216 wm Pa WY aw A19djSautsisttfcaolrlaygse,igBnMiIfiacnadn(atlplcoyho<l.05u)sdiinfgfarneanltytshisanofPcoFvOaAridaencciele(s 1,23... andlor 10 Nem Triglycersidevs c 14s nein na esas 153% na182 145 16175 133191 160 131190 158 128.187 mm 144201 165 135194 aw eas "Table 124. Adjust OsRabtiyosPF(O0.A D )ande9250%0C0s oFnlfaiodi reoncchee, nIinctaelrvaMle(d9ic5a%lCS1uiolrlaLnicpeidPCrloignircaaml Chemise Reference Poi, . DPuFiOlAe ChoOlR>T-290%0mCgLd,OChRol>=2040%mCgiIdl, roe - wo. 2 04 0210 13 osu 3 09 0420 09 0322 409 0421 11 oar S17 0740 16 0739 6 10 0422 11 os2s 708 oats 12 0520 510 0422 12 0s29 0 14 0633 1s 0637 1 0526 10 42s 1. Adjusted forage, BMIand alcohol LODRL' 1030SmCyLdl, wo07 0316 07 0316 08 041s 140631 09 0420 07 034s 10s 12 0527 12 0528 ORHTDLo-O40MmCyiLdl, oo. 10432 05 ots 20 0ns4 10 0420 17 0747 06 0218 14 0537 09 0324 26 1068 TORr!ghv9e9%gcr1li5d0emsgd>l=. wo. 07 031s 10 0424 13 0531 12 0530 17 0740 09 0422 27 1265 24 1059 24 1058 3 Te a T a p; ieae ee. | : Wo . Wows Wae e wa min wan oan ow wm lows now ow we we aum aown Table 13. Non-adjusted and Adjusted Ln PFOA Coeflicients for Ln Lipid Clinical Chemistry Result, 2000 Fluorochemical Medical Surveillance Program Ln PFOA LnPFOA NCooenf-liacdijenuts_t_ed SE p value ACodejfufsitceide'nt SE p value n Cholesterol All Locations 0.0059 00060 32 00076 00059 20 Antwerp 00051 00106 63 00130 0009 18 Cottage Grove 0.0034 0.0089 70 00021 00100 83 Decatur 0021 0039 a1 00266 00141 06 LnLDL All Locations 00012 0.0089 89 00021 0009 81 Antwerp 00037 00157 81 00106 00147 47 Cottage Grove 0.0022 00139 87 00049 00145 73 Decatur 00258 0019 20 00302 00200 13 LaHDL All Locations -00307 0.0079 0001 00183 00069 01 Antwerp 00057 00136 68 00095 00131 47 Cottage Grove 0.0153 00122 21 00192 00120 1 Decatur 00256 00149 09 Ti 00207 00141 14 All Locations 0.0892 Antwerp. 00840 Cottage Grove 00343 TDeTcatuSreedymods007A1g5ee 0.0185 0001 00288 004 0.0316 28 Tn0A0g4e00LBM 0a8A 00711 00169 0001 00980 0.0270 0004 00280 00314 38 00689 0.0376 07 Tableoy14.PFMOA DeeancdlaS,ta2n0nd0a0rMd DeevicstiFolntoofcHheempiaclCSluirvcelilClahnecemiPrsoegrRaemsult, Tou Dirt . PDeFlOeA AMlekmP(oSsD) ASMeTm(SD) ALMeTw(SD) AMSeUmnA(LSDT)~~ GMeGaT (sD) BiMleiarnu(bSinD) MBielairnubi(n5D) 1 6503) 28) 205" 0990.0" 265 09 3) 01 (004) 2 905 25) 205" 10001" 25018) 09" (3) or (004) 3 eae? %@t mat een wen 10 0) 01 06 4 6515) seat 248) ne 205" 20st 0980.1)" 2820) omop way 09" (03) 0 y 01 (004) of oo 6 6619) 26M ne oEn* 2001 09 03) 01 (006) 4 1 map wa wast omon san 090 ) 01 oo) 5 ean ue mae wep wey of 02 or om 5 met WEF 609% 0;mEDU Hawt 07 on ort eon 10 6820 240) sant 093002)" 30016) 0s (03) 01 (008) 14 Statistically significantly (p < 05)different than PFOA decile(s) 1,2, 3.. and/or 10. for"THaebplaeti5cA.ClinAidcjaulsCtheedmiMsetaeynRaensudl9ts5b%yCoPnFfiOdAenDceeciIlneteArdvjaulssted PIOA Al Phos AST Au gr. BTioluilnbin_____DDiilriemctbin__ ' Decile Mean OCT Mean 95%Cl Mean OS%CI Mean 9S%CI Mean 99%CI Mam 95%CI ' 6 elm 2 um 2 253 R23 09 0612 01 009012 . 2 SU sie 25 Ba 2B 25 93 09 0800 01 009012 ' 3 oF see 26% M2 0 263 2 2B 10700R106 01 009012 : 4 6 60 4 ma WM um 3 2M 09 0R0 01 00s 5 oF eer 2 25 2 na 20 1830 09 079095 01' 010013 Fl 6 en 2 a 29% ax 2 1527 09 078095 01 0080 7 WU ma ua NC was 3 23 09 07094 01 0osol 5 6 em 26% wm 2 263m 2 235 09 0709 0 010013 9 M0 e6Ts 26% 228 MC 3038 WY 239 0808s 01% 007010 0 emu 226 03 3 A308 07509 01 00801 ' AdSutdstifcoarlAysgigeniBfiMcIantly(Alpc<oh0o5l)dwiifnfegraenatbtshiasnoPfFcOoAvadrieainlce(s 1,2, 3. andlor 10 Table 1f5o8r.HeApdatjiucstCleidnMicealanCahenmdis9t5y% RCeosnuflitdsebnycePInFtOeArvaDlesl(e95%CI 708 n ATL ol oToul I Direct Decile Men OCT Mean 9%CI Mean 95%Cl Mean 05%Cl Mean 95%Ci Mean 95%C1 ' 6 67 26% 2438 30 234 WF 230 2 W0 Ss6s 2 2327 30 234 26 02 3 oF see 26 228 2 253 2 23 ] 6 67 24 2a 2 2533 234 5 O66 27% 2024 2 33 27 1830 6 66 GT 22024 2900 2128 A 1527 7 WG 260 428 NC ;3s 28 B34 8 6 6 26 228 29 253 2% 234 9 W667 26 2438 36H N39 3H M9 0 Qn au 226 WP 23 2 23 09 08101 01 009012 09 081099 01 009012 10 0%0106 01 0090 09 08209 01 00801 09 08009 01 010013 09 078095 01 0080 090 070% 01 008011 09 07609 01' 010013 08 068085 008 0070.10 09 007909 01 0m012 4 Statisticallysignificantly (p<05) differentthan PFOA decile) 1,23,... andlor10 Table 15C. Adjusted Means and 95% Confidence Intervals (C1) for AST/ALT Ratios by PFOA Decile Peon vent scr vest scr TT 0s osero2 os ossiol 2 098" 0961.02 098 095-102 3 1.00" 0.96-1.03 1.00 097-103 4 098 095-101 098 094-100 5 097 094-101 095 0.94-1.00 6 1.001% 097-1.04 LoTM 097-104 7 098 095-101 098 095-101 8 1.00 097-103 1.00% 097-104 9 095% 092-098 095% 0910.98 10 0.943% 091-098 095% 0092-098 aAdjusted forAge,BMIandAlcoholusinganalysisof covariance b.Adjusted for Age, Triglycerides and Alcohol using analysisofcovariance "1 Statistically significantly(p < 05)differentthan PFOA decile(s) 1,2, 3... and/or 10 abe 16. Numberand Percent ofHepatic Clinical Chemistry ResbyuRelfertensce Pints, by PFOA Decl, 2000 Medical luorochemical Surveillance Program v PDeEciOleA NAGKP)hos>=120 NAeSoT>=S0 NAeLoT>=S0 NGeGoT>=50 TBoiullirubin >15 BDiirelcit ru>bi0.n4 ' om 26) 20 0 san 62 4 26) 2 woo CER) 30 0 am 500) im oo 40 oo 4 S00) 2 @ ow ' s ro 10 30 26) rte oo ' 6 10 oo 7 lc) lc) 80 24) oo 26) 4 500) 50) 50) te oo re oo oo 1 1 0 noe 00 704 806) 709 am re vo ro oe pualuet 81 2 3s a6 8 56 "chi squaretetofsignificance "Table 17.OddsRatioTsy (PFOOA.aDRnedc9i.l5e)%20C0o0nfFiudoernoccehIenmtiecravallsM(e9d5i%ca SCu.f1reoirlHaenpcaetiPcroClgirnaimcal Chemistry Results, ' pop mS m B ATsI son R Duk OR. 9%CL OR. 9%CL OR. 0%CL Wl. We. wo. ol 2 06 0224 0s 0230 0s 0232 506 0224 07 0229 06 0122 4 07 0224 0 m23 06 0124 Sos ome 04 ome 04 omar 6 02 oso or 000s 01 0009 7 07 e224 os 0230 0s 0123 50s 0227 07 eas 07 022s 9 13 0442 10 0334 10 0336 012 oar Los 07 e226 21.. ANdotiasdiutfoedage, BMI andalcool 3. Adslforag,vihyeeanrdsdkeoosl rGE arson it OR. 9mCl OR. 99%CL OR 9sWCL 48) o- wo. wo. 03 ose 03 oss 03 0si6 os 0220 09 0231 07 22s 0s 0230 0s 0227 0s 0229 03 0si4 03 0013 03 003 03 aes 03 013 03 oss 0s 0230 09 0231 0s 0229 0s 0227 07 0226 06 0224 1s sas 17 0sse 17 0sss 07 0224 07 0225 04 or7 "Table 178. Odds RatiToys (POF.RO.A) aDnedc9il5,%2C0o0n0fFildveonrcoechIenmtiercvaallsM(e9d5i%caClLS)urfvoeirllHaenpcaetiPcrColgirnaimca Chemistry Rests, i DPOiAe OR. 0CL row 2 es m4 5 es maar 4 09 032s S03 ema 6 03 oles 713 esa 514 0637 9 22 09s6 014 063s ATz 00 OR. osci wo. woos os 032s 0s 0324 03 od 02 019 as osu 1s 063 17 074s 12 0s34 21 ANdjoutsstdeudsfeodrae, BM andskool 3. Adjusted forage,viglceanrdsdkeoosl OR. 9%CL i. Loan 07 0220 09 0324 03 ono 02 0109 12 0saa 13 0s 17 014s 09 03a eGe ordonA EUAN OR. 9%CL OR. 99%CL OR 9s%Cl wo 10 wo. 06 029 06 0219 07 0223 07 0222 0s 0324 07 0222 10 0420 10 oa2s 10 0330 09 032s 0s 0324 08 ons 05 0217 0s ome 1 0433 12 eaa2 12 esas 06 021 07 0220 07 0220 16 0647 1506 16 064s 12 0add 1606 17 064s 13 036 `Table 18. Non-adjusted and Adjusted Ln PFOA Coeflicients for Ln Hepatic Clinical Chemistry, 2000 Fluorochemical Medical Surveillance Program Non-adjusted Adjusted"? Ln PFOA Ln PFOA Coefficient _SE p value Coeflicient SE p value Ln Alkaline Phosphatase All Locations 0.0155 Antwerp 00025 Cottage Grove 0.0141 Decatur 00394 0.0082 06 00093 0.0081 25 00037 0.0081 65 00137 85 00060 00139 67 00170 00140 22 00113 21 00127 0017 28 0080 00117 24 00191 04 00460 00192 02 00394 00192 04 LnAST AllLocations -00018 Antwerp 00048 Cottage Grove 0.0281 Decatur 00205 00086 83 0001 0008 55 00089 00087 31 00137 73 00029 0038 83 00066 00142 64 00141 05 00258 00146 08 00271 00145 07 00200 31 00114 00203 57 00062 00203 76 nALT All Locations ~~ 0.0402 00143 005 00249 00132 06 00115 00136 40 Antwerp 0012 0020 58 0000028953 0022 00222 70 19 "Table 18. (Continued) Non-adjusted Ln PFOA Coefficient SE p value Adjusted'? Ln PFOA Coefficient SE LnALT(Cont'd) p value Cottage Grove - 0.0131 0.0215 54 0.0096 0.0209 65 -0.0008 0.0208 69 Decatur 0.0954 0.0300 002 0.0704 0.0287 02 0.0581 0.0287 04 0 GGT All Locations 0.0409 Antwerp 00170 Cottage Grove -00088 Decatur 0.0754 0.0174 02 00307 58 0.0292 76 0.0326 0.0166 05 0.0097 0.0163 55 00269 00294 36 -0.0047 0.0295 87 00198 00286 49 00233 00270 39 0.0344 03 0.0800 0.0344 02 0.0599 0.0329 07 n Total Bilirubin All Locations ~~ -0.0406 0.0101 0001 -0.0325 0.0099 001 0.0267 0.0101 01 Antwerp -0.0117 0.0178 51 0.0122 0.0182 50 -0.0093 0.0188 62 Cottage Grove -0.0060 0.0138 66 00.00006978 00.00114412 4694 Decatur -0.0528 0.0203 01 -0.0537 0.0209 01 0.0462 0.0206 03 T. See study methods. Adjusted for Ln Age, La BMI, Ln Alcohol 2. Seestudy methods. Adjusted for Ln Age, Ln Triglycerides, Ln Alcohol `Table 19. Mean and Standard Deviation (SD)ofThyroid-Related Clinical Chemistry Result, by PFOA Decile, 2000 Fluorochemical Medical Surveillance Program PFOA TSH T FreeT4 Delle Mean(SD) Mean (SD) Mean (SD) hs] Mean SD 1 206 (117% 828 (151) LIS (0.157 124 (28) 2 203 (109) 841 (140) LI (015) 24 Qn 3 184 (073) 812 131) LIS (014 125 (19) 4 341 (907% 804 (151) 112 (0.19) 127 (19) s 260 (2.63) 79 (137) 1.07 (019) 126 (7) 6 210 (109) 849 (133) L10 (0.13) 28 3) 7 219 (123) 808 (129) 109 (013) 9 @y 8 243 9) 837 (29) 109 O17) 130 @) 9 28 (40) 849 (130) LI0 (0.14) 130 0) 10 241 (2m 797 (133) 1.07 (014% 130 (1) 1 Statistically significantly (p < 05) different than PFOA decile(s) 1,2,3 . and/or 10 TfaobrlTeh2y0r.oiAddjCulisntiecda"l MCheeamnisainrdy9R5es%ulCtosnfbiydPenFcOeAInDteecrivlales PDeOlAe Mem Ts 9SACI Mam 7 OSGI Mam FreeTd9%Cl Men -- 95% ' 207 ae 82 791866 LIS Ln 12 nga 2 200 104206 84s sores Ln 1071s ns ngs 5 13 094285 0% 20846 LEY Les 12 ns 4 3 ama soe 28a Ln 108106 126 120133 5 260 esas 798 260835 L070 Lean 126 1204133 6 207" a0 8S sIss9 LI 10114 18 21134 7 a 126306 80 708M 109 os 129 3135 8 239 Les 840 803876 110 10611 130 124136 9 284 1838 sa s0n880 Ln ors 13 13136 0 201 Las 79 256832 LY lean 19 123136 *Adjusted for Age, BME and Alcoholusinganalysisofcovariance 1 Statistically signi(pf<.i05c)daiffneretntlthyanPFOA deciles 1,2,3 andlor 10 `Table 21. Number and Percent Subjects Above or Below Reference Points for Thyroid-Related Clinical Chemistry Results, by PFOA Decile, 2000 Fluorochemical Medical Surveillance Program PFOA SI0S3H5 I>U5/5mL) _5Td(ug2/1d2l0) F0re0eTd(n2g1/8d3L) m S60gd21) 81 Decile N(%) N(%) N(%) N(%) NC) NCO NE) N%) 1 00 00) 0 00) 0 10 0 10) 2 12) 000) 00 000) 01 00 3 00) 000) 12 00 0 00 0 00) 4 12 24 0) 00) 12 1 00) 0(0) 5 00) 4@) 0 0 1 00 00) 000) 6 0 1@ 00 00 00 0) 00) 24) 7 00) 24) 00) 00) 00) 0 00 1 3 00) 4 0@ 00 12 26 00 2@) 9 00) 4) 0 00) 00 0@ 00) 00) 01 1@ 00 00) 00 00 10 paler 63.10 a" | ---- 65 46 - 46 eee "chi square testofsignificance oTradbelre)2a2n.d TchoyrrroeisdpornedliantgedPhFoOmAmo(nge/tmesLts)rceosnulctesntbryatsiuobnj.ecRtewsuhletnsTouStH-of-r4e.f0eIreUnc/emrLan(gaes(cseeneding Table 21) are shaded. Sued1 I4SmH 23 a4026 45 4a1r5m 67 442211 s8 4422 n1o0 44235 123 4axs 1"5 a40st 76 44%71 118 44sstt 2210 4s%s 2 2 5s2z0 24% 553%% 227% ssed1 220 558808 3 S857 2 3 85081 3#% 80B508 3%7 B781 33% B7S8 4 sa 4 MB 443 11983798 4454 G2532084 6114 Fre1e1T3d 7632 o11s3 5812 o11m7 880s 0180 180s8 110233 e64 211225 7627 0% 091 8768 110113 8812 0s 1m 6738 017188 8708 1123 8842 00%8 s7a6 0s 1m 8880 111090 7s0s 08 100 8814 tor 109 6777 oomm 88s8 111250 6s7s o0ss 7ss 1M0B4 50 123 54 107 4874 008854 465s O08815 10T33 P1F6O1As 11076 6a02n2s5 108 03505740 11079 o16o0m6 1186s 0o08o1 183 0604 112s5 00821m7 11132 2110403 tMae 02218s 1t0a9 11248125 11109 0t1r8257 128 21% 110129 01136808 157 1269 11076 40115%7 11424 2S517849 11325 11868285 11168 017130 10 0718 11436 0158275 11344 24078108 110020 23364730 13 164 36% 0576 110223 2074235 112210 0274175 Table 23. Non-adjuste2d00a0ndFlAdujoursotcehde'miLcnalPMFeOdiAcaColefSfuircvieeinltlsanfcoer PLrnoTghryarmoid-Related Hormone, NLonn-PaFdOjuAsted Coefficient SE All Locations 0.0395 0.0204 Antwerp 00509 00329 patie Lo TSH 05 12 ALdnjuPsFteOdA" Coefficient __SE___pvalue 00360 00207 08 00391 00333 24 Cotage Grove 0.0016 ~~ 0.0310 96 Decatur 00343 00497 49 0011 00322 73 00365 00513 48 LaTd All Locations 0.0037 00054 50 Antwerp 0002 0009 83 Cottage Grove -00124 0.0072 09 0.0057 00054 29 00041 0009 68 00093 00072 20 Decatur 00012 00126 92 00083 00127 51 Ln Free T4 AllLocations -00138 00044 002 Antwerp 00108 00078 17 Cottage Grove -0.0093 0.0058 11 Decatur 00138 00103 18 LnT3 All Locations 00107 00052 04 Antwerp 00222 00077 005 Cottage Grove 0.0026 0.0096 79 Decatur 00317 00117 008 1. Seestudy methods. Adjusted for Ln Age, La BLMnAlcIohol 0017 00043 01 00140 00078 07 00071 00059 23 00184 00105 08 00105 00053 05 00216 00077 006 00006 0009 95 00 oon 0 Table 24. Predicted Thyroid-Related Clinical Chemistry Results Based on Multiple Regression Models for 40 Year Old Male with BMI =28 `and Drinks 0.5 Alcohol BeverpaergDeays. Reference Range PFOA Serum Concentration (ug/ml) 0.005 001 0.10 050 1.00 5.00 10.00 50.00 00.00 Predicted Predicted T(S0H25(-uU5/5m)L) T(445(-u1g2/0L)) 1.58 2 162 8.19 176 308 187 801 191 798 203 7.90 207 787 220 7.80 226 mm Predicted Predicted Free T4 (ng/dL) T3 (ng/dL) (070-153) (60-181) 115 ne 11s 19 11 121 1.09 124 1.09 125 1.06 128 1.06 128 104 131 03 32 `Table 25. Number (percent), Odds Ratios (O.R.) and 95% Confidence Intervals (95% C.L) with Metabolic Syndrome, by PFOA Decile, 2000 Fluorochemical Medical Surveillance Program DePcFiOlAe NG) 1 612) 2 4 3 1@ 4 8316) 5 5310 6 612 7 360 8 6312 9 6312) 10 6012) 1. Adjustedfor age NOonR-adj9u5s%teCdlL 10 06 0224 02 0009 14 0547 08 0229 10 0336 05 0119 09 0332 10 0336 10 0335 OR.Adju9s5te%dC"lL 0) ox 06 0122 02 0010 15 0549 08 0229 10 0335 05 0120 09 0330 110336 10 0336 . AppendiAx "Table AL. Coreaton Coeficients (saasTmabele DI) Ln Variable La FO) a(PEOS) PPRFOOSA 01s0seene 10 ABgMeI o00n3e 00006 AlCchoolheoslterol 0014200 "000.1800 HDIDLL "000oe 000067 TGlrugchooseerides 002011000 oorre AalsktPhos 000081 000080 AALSTTALT oainsee 00.00800 aTooutl Bilimkin o0l.o1r8400 0a030 SDirHect Biibin ~~ 00.0096 000.403 FTreTd 0o0r3e 00.00062 n 00 oo; ppossn pd<o.o0r0 Latage) 012000eer "Dissetre "o0s1e30 o02g60e0eee 000080 000035 "o0i0s8eer 0"00031 ""0s09ee 100 BMD. 01206meer 000002 oodaa3veeeseee o0n2seer oasreerweee 003310emneees ""a0i0s3e 000085 D00a2gens (Alcohol o10or o03o7l0000 "Daorvees ""0o0a4ses 0167tn 0"200020wee 0"0020 "00o8ie "00 Age by PFOA (ng/mL) (| 7o2F.. twEg:.=. | =2 SEt ErEm . : i Pron ELX) LN Age by LN PFOA (ug/mL) | ] a ii | deoomua atse wl Ra | a= Comma aafo tT ERR Weprop e peepee nro TLNeAge by LN PFOA oo (g/mL) ok: a {of "DIF V 0% SererrE BMI by PFOA (ng/mL) GE wl. | oly. Eels & all gL Lc || EC . || 20] B0d10g 20 e %0 4% 6 000 Pron LN BMI by LN PFOA (ng/mL) pe N3a87]e],el Feitn me-onry 35 * oie o5y Ll R BREE E d| saq1 I.ITEEREIAE. . -)1 o LTE | 2Bs]lPEE A EEREREE npron LN BMI by LN PFOA (ug/mL) 26+] mmm = | 284 nn: a we. rEs ES 5 LAE Ne | Bi aty, L IE oE nD aS) 1|] 3a27 JI nnN pSeE e Loc 1 2267b lie sre 54324012345 InPFOA Alcohol (Drinks/Day) by PFOA( pg/mL) [| ooFf: E 1 |lewelk., - EE=F . || | =Thee ha enr e prox ON LN Alcohol (Drinks/Day) by Ln PFOA( ug/mL) i | a Eon | Be8 oi pis oo =BAimTEgEE | 2a EL 4 | dr SrErI ErL S propnES rEee em LNDrinbkyPsFO/A(dg/amLy) = fr itd ii EEE Bil mrw or g PFOS(ugbL y/ PFOm A(uL g/m) l) _ 1 By 1di.% - | AE : {AE | o ihe m Rs ese hee: | PFOA a LN PFOS sl LN PFOA /mL) |e sg | 12 i ER Br |2 al, [Fa] aR - ffi || tiemenron e] (Tr LrNePFOS(g/E m)byLN Px FOA o JER. 18. [of ei 131 na 2] AERIS .: a REE. ao | TETIEprIonITIT y-- Glucose (mg/dL) by PFOA (g/mL) gB= o4bo... 2 | ET: E LN Glucose (mg/dL) by LN PFOA (ng/mL) a v" L Dr chanae il -- TIE eT Pp-- LN Glucose (mg/dL) by LN PFOA (ug/mL) [3 Ne rr = | r Trr e1vr or Cholesterol (mg/dL) by PFOA (ng/mL) _ 311 EE: . - tres TATEETT: | LN Cholesterol (mg/dL) by LN PFOA (ng/mL) od upg. | attr -- LN Cholesterol (mg/dL) by LN PFOA (ug/mL) faa.J, 0 |g > Lag fy | L(l ar lSrri I rrrr rTrT i: TTe | err | LDL (mg/dL) by PFOA (ng/mL) 2008 ;| ` CRN en ronlnn LN LDL (mg/dL) by LN PFOA (ug/mL) J55DhnoGinrmiEaee {| 1 3 | ses Zineoon zs ss LN LDL (mg/dL) by LN PFOA (ug/mL) Raat <3 REE 3I 1 ehR s PE AE) || Ta aa0npron sss | a HDL (mg/dL) by PFOA (ug/mL) = 7] i=- | = 1 |*= . || 3 of . Lo") 1 RofyT ThE ET RrEon eT ee | S LNHDLb (mg/dL)bya LN PFOAa (g/mL) | en |lG g n ri Eg [FETE j= eprrrptp TR S A anei on | LN HDL (mg/dL) by LN PFOA (ug/mL) ] rn Fret | rrror TITY Triglycerides(mg/dL) by PFOA (ug/L) PETE---- 7007 & | w mTy : | (gor. } ods [2 : } | ThE C| m Rhee nsw --P-- FOA EY LN Triglycerides (mg/dL) by LN PFOA (ug/mL) od Lolrg l|Bgol:]. Co ELdlEaeRY | =3 enwiRm. r Tisr aasr iis { pron LN Triglycerides (mg/dL) by LN PFOA (ug/mL) (r-------- |e SEE NL E[ool LJiragidia 1m7el3 G LEoFiT, || STprI ox S Alkaline Phosphatase (IU/L) by PFOA (ug/mL) pas IIE | I 1207, + || Fi 12 wf. : | of. % oF" . | WEhEeR reIpLrgYrg)red | i ron LN Alkalr ine PhophatT ase (IU/L) by LN PFOA (ug/mL) Ve 3 Do000G trary [#B1g T aE ne . J ete | EIE pronIO LN Alkaline Phosphatase (IU/L) by LN PFOA (ng/mL) {= os a | rn I4 Clea. 2 TEER ge, a | TTT 1 PFOA AST (IU/L) by PFOA (ng/mL) | | of | Br 12% | Bee =f i EEE bI ron EEE I rE : LN AST (IU/L) by LN PFOA (ug/mL) I1R s% E CinE niIgS EeElE. | E | PL EEprRonE ERE) LN AST (IU/L) by LN PFOA (ng/mL) PR om i bronrwioen rrrs Pp ---- ALT (IU/L) by PFOA (ug/mL) | wd" = on || = iwg f" . --. :2 . [I ey | TRE ras ene | PFOA LN ALT(IU/L) by LN PFOA (g/mL) lp aE 12 rial bb e = omex n | won LN ALT (IU/L) by LN PFOA (ug/mL) fr | +. md x Sa giggle d SIr Ic VEr noIvosIr r TTIcad | AST/ALT by PFOA (ng/mL) EEE 1.2% |2Z;.% : { os ge. . : | a ot EEE EXEL) pron [So EEE AST/ALT by LN PFOA (ug/mL) I ---- o0s4: = |,os os 1[B3o0d43- mw onSiridii dn: 1307 ER a || <=2 RISrh E. | "a TIT TIe aTTTY | PFOA LN AST/ALT by LN PFOA (g/mL) Suad . os - Lo my,a [Bal wid of ARN aay . |V | Saasy I TRTI REREI : | or GGT (IU/L) by PFOA (g/mL) Pa ||| E $m. {-n.. VinBt "oni a ok 0 1020 30 40 5 60 70 80 LN GGT (IU/L) by LN PFOA (ug/mL) ee by nt, Le 7" abil. Lr | nPFOA LapN GGT (IU/L) by LN PFOA (g/mL) Ve SU be viTdEiS LYr Ctedma P | en nE ore T Total Bilirubin (mg/dL) by PFOA (ug/mL) ir | Hf EEE EEE EEE) LN Total Bilirubin (mg/dL)by LN PFOA (ug/mL) [3] roimntin. 0. mma I | AAA a:EAS ETE Tim LN Total Bilirubin (mg/dL) by LN PFOA (ug/mL) HEE Ped. IGE || 48 TrITITITYT | | or = Direct Bilirubin (mg/dL) by PFOA (g/mL) os] - = | gos] ol | S02] = [odes oo 3 d= = || ThanaPrnon mnee LN Direct Bilirubin (mg/dL)by LN PFOA (ug/ml) I| E - ! 2 EO | STI proTT LN Direct Bilirubin (mg/dL) by LN PFOA (g/mL) 2 Bl es . { Br | 1 InPFOA | FeTSH (uIU/mL) by PFOA (ng/mL) LO||B40 | | 2dr i RE . i em erdeos Te | LN TSH (W[U/mL) by LN PFOA (ug/mL) - ae 2 mL [300s eR Soal3 vepe m | 21 IEE nee ror ss | LNTSH(ug/mL) byLN PFOA(g/mL) I ITT I 15 o EEy , nitions | : Tp TITp oIrcIa Ts TTT OE T4 (ug/dL) by PFOA (ug/ml) r Pr e | ng of. EA EER i f | 3 = 5 : Ed | - SHHe0ro N00 LN T4( g/dL) by PFOA (g/mL) LEE [adr oeSRR od EET [547 Ae | 1s } | rrr {= . | TSITITmeIronTTTTY | LN T4 (ug/dL) by LN PFOA (g/mL) [ee aad ERLATEI 21] Lig An |lz ~Ne VFE ARA+oY fos" .- | 547 a 15: 2 1 | | oe TIIpIronITIT | Free T4 (ng/dL) by PFOA (ng/mL) | ---- | 158 ilozkaf=-. - | orf || oolTomtomgfeneo | Pron LNFree T4 (ngldL) by LN PFOA(ng/mL) Fil a an :11] -T % eweg.sg | + mn || $2353prTonTivs LN Free T4 (ng/dL) by LN PFOA (ug/mL) [= I - | | STI wrronTT T3 (ng/dL) by PFOA (g/mL) f oe NN = | gaol || Thanarsoxnnew LN T3 (ng/dL)byLN PFOA(pg/mL) [of EF 5o5s- o: 91 -LnikEe | eT lead So of. ARR 4s: BeX ow oFSTITTpnIereoIr rIm TTeY LNT3 (ng/dL) by LN PFOA (ug/ml) ar | ss | {= = [fof Lod So) EERE: og CITnI proA ES Appendix B `TaMbelediBIa.naDnedmoRgarnagpeh)iocfaEnmdpClloiyneicesChWehmiostSrelyfC-oRmeppoarrtiesdonPrses(cMreiabne,d SCthaonldeasrtderDoelv-iLaotwieorni(nSgD), Medi2ca0t0i0onFslu(oNro=ch4e6r)iVcaelrsMuesdTichaolseSuWrhveoilDliadncNeoPtr(ogNr=am506), Prescribed Medication (N=46) Not Prescribed (N=506) Mean (SD) Median Range Mean (SD) Median Range PFOA 198 212 121 0141030 PROS 169 172 127 0111006 Age 490 7 50 3160 BMI 28 45 284 199393 Alcohol 04 06 00 0020 %Anwep 210 - - - %Cotiage Grove 20* ~~ - % Decatur sv... Glucose 103 38 99 5431 Cholesterol ~~ 221 45 217 144-384 LDL 134 40 130 59222 HDL a7 14 4 3190 Triglycerides 226% 160 194 35792 Alk Phos TH T 3026 AST ms 2s 748 ALT 6 20 2 99 GGT 36+ 24 28 10-144 Towl Bilibin 08 02 08 041s 221 640 110 105 097 072 0 9 39 274 46 266 06 09 03 EI - %- : 7- - 91 19 91 24 41 2 136 36 13 4913 46 159 12 128 6 18 64 3 7 24 0015 26 32 2 09 03 08 001-9203 0.02624 206 172521 0064 v 31251 105331 37235 182 2479 21160 1069 6103 63M 0323 Table BI. (continued) LipiPrde-sLcoriwbeerding Medication (N=46) Not Prescribed (N=506) Mean (SD) Median Range Mean (SD) Median Range Direct Bilirubin 0.1 006 0.1 00-03 AST/ALT 089 042 082 022222 TSH 28 31 21 04215 T4 8.1 14 79 58-129 Free T4 Ll 02 11 0815 3 125 19 23 93-190 +p <0 (prescribed vs. non-prescribed) 0.1 0.06 0.1 098 0.12 097 24 34 19 82 14 82 11 02 Ll 27.2% 125 0.0-0.7 0.70-1.82 0.03-65.3 42-120 06-18 78-300 Table B2. MeanandResSutlatnsdabrydDLeovcaitaitoinonby(SCDh)oloefsPtFroOlA-,LoPwFeOriSn,gDMeedimcaotigonrSaatnpadshCliincicsal Cheristry Aner CotageGrove Decatur MaYnes(SD) MenNo(SD) MaYnes(SD) MaNno(SD) MeYmes(SD) ManNo(SD) PRON PROS TM Bi Aleobol Glos Cholesterol LoL HL Trigheerides AkPhos LI6 (SH) 125 051) ooo 29 09 07 08 9% GO 232 (8) 2a 2a 195 (129) SQ) 102 (06 095 97) 9 27 GO LI (LD) 8 aD) 218 (41) 1 en san 120 (83) 60 (4) 291 (69 076 (79 oo 25 (2 07 0) 0a) woe ne @n 2 oan moo@) moan 46 (25) 08 (09%) a 29 GH 01 en 10 ey 2G) Bo 4% ay wow es ay 197 (5) 18 (6D) 208 QO) 129 (09) woo 2 oo WT G6 BE (6 02 01 03 le we B04 2 G0) 2 @) Bo G9 Be GO Mayu ao se am) 1m (10) moan Ban Antwerp ChoYleesuterol.LoweringNMoed Meh (SD) Man (D) TableB2 (continu) CoteGone ChYolesssenLowsriNngoMed Men (SD) Mem (6D) Desaur ChYaeleut owsringMNoed Men (6D) Mem (SD) ss BD nom os Toe ar 2 3 a soe Moan oa) ASTAT 1306 12 0H 08 ) 08 0) 08 ) Gr x ay ap Zo Noes Way TaBlikin 10 03) 10 03) 09 09 03 07 ) Dibillab0in1 00) 01 (0 01 (0) 01 ) 01 ) Ts oe) 2 a6 25 09 24 aH 32 Go TM san 82 oan 72 02 9 an 82 0) Fem 12 0) 1 0 woen uoen ue TM usoay a9 woe ps Gy woe) = 05cholesrolloweringmdcaion(es venowihmeach locaton ws Mu ae 09 Bo 07 ) 01 (0F 28 6) 89 ue ns @) TBayblLeoc3a.tioMne,a2n0,0S0taFunodraorcdhDeemvsiiactailonM,eMdeidciaalnSuarnvdeilRlaanngcee PorfogPrFaOmA,,FoPrFOESm,plDoeymeoegsrParpehsiccriFbedaChcaaltnesdtColrionrli-cLsaolweCrhienmgisMterdyicRaetsiulotnss, Mem A(nStDw)erp (NM1e0di)an Range MemCott(aSgDe)GrovMeed(iN-a9n) Range MewDec(aStDu)r(4M-e2d7i)an Range PFOA LI6 LSS 06 014531 205 67 LSS 0290000 197 152 176 015497 POS 125 091 095 02277 076 07 036 01210 218 201 15% 0151006 Age 0s sass os 7 2 0 48 7 4 6 BM 2390 34 247 2aM2 BS 42 309 Bass 297 46 4 199393 Aeohol 07 08 05 0020 07 07 10 0020 02 04 00 0016 Glew 9% 30 8 SLI 15 13 104 202 106 46 99 031 Chole! 202 38 243 19280 24 38 206 ISL 20 50 26 144384 or. 12M ie 822 16 47 4 HAI 10 35 19 6626 OL 2 as 4s ww on 0 aw wa " 0 oe un Trgherides 195 124 179 35463 180 8 193 @294 23 18 20 299 AkPhos SE 20 ss des 7 Bm doa0s 7 10 Wa Ast. 5 6 wu uw ow 5 2 wa n 9 as 1a ar 20 9 nea EE I ASUALT 13 06 12 0622 08 03 08 0413 08 03 07 0212 (CToanbtlien3ued) MenA(nSD)e(M-o9)d Range MomCott(aSgDe)GrovMee(dN=0) Range aor TowlBil DictBil SH TM FeeT4 2% Bw ae 10 03 09 0745 01 003 01 0102 19 07 18 0231 3 17 83 sens 12 02 12 0945 DSI ms ee sm? ow 4 Mas 09 02 09 0612 09 003 01 0001 25 06 21 1735 720 12 74 S891 11 01 10 0813 17 16 7 SMI 21.. SStuatiisstiicaalllyyssiginginfiicfanitclyan((tpl<<y 0055)) ddiiffffeerreenntt tthhaann AContctwaegrepGrove 3. Sutisicallysignificantly(p< 09) different than Decatur ManDe(cSaDt)ur(Me=2d7) Range 3 18 B DBE 07 02 07 04d1 01 007 01 0003 32 40 25 0421S $2 13 80 62290 LI 02 LI 0814 127 20 13 9% Table B4. Non-adjusted and Adjusted* Regression PFOA Coefficients for Lipid Clinical Chemistry Results, 2000 Fluorochemical Results Restricted to Employee `Taking Cholesterol-Lowering Medications (N = 46) Non-adjusted Ln of Response Ln PFOA Variable Coefficient SE Cholesterol ~~ 0.0453 0.0263 LDL 00253 00494 HDL 00003 00364 Triglycerides 00681 0.0968 Alk Phosphatase 0.0612 0.0420 AST 00305 00437 ALT 01040 00672 GGT 00591 00734 Tot Bilirubin -0.0818 0.0356 TSH T+ Free T4 i} 00928 00904 0023 0026 00219 00197 00007 0019 p value 09 61 9 49 IS 49 13 43 03 31 33 27 72 Adjusted Ln PFOA Coefficient SE 00501 00292 00303 00536 00140 00346 01015 00972 00533 00420 0.0347 0.0428 00434 0.0475 0.0340 0.0469 00862 00674 0.054% 0.0669 00653 00812 0.0404 00746 00649 00369 0.0623 00361 01803 00918 00415 00236 00307 00213 00041 00207 p value 09 58 69 30 21 42 37 47 21 42 43 59 09 09 06 09 16 84 * Ln ofPFOA coefficient adjusted for Ln Age, Ln BMI and Ln Alcohol unlessotherwise (**) noted ** Ln PFOA coefficient adjusted for Ln Age, Ln Triglyceriadneds Ln Alcohol Appendix C Table C1 Number of Employees, Mean, 95% Confidence Interval, Median and Range by PFOA Decile Distribution , 2000 Fluorochemical Medical Surveillance Program for Antwerp, CottageGroveandDecatur(N = 552) PFOA Decile N 1 st 28 3 54 4 ss s 59 6 50 7 54 8s 60 93 0 ss Mean 0.06" 020% 036" 055 0.90% 1.26" 164 2175 3g Ha 95% CL PFOA 005-0.07 0.19021 035-037 0530.55 087-094 123-128 1.60-1.67 212223 2903.10 698-1556 Median 006 0.19 036 055 038 126 1.63 216 296 494 Range 0007-0.13 013-029 030-044 044-071 071-110 111-140 1.42-1.85 1862.50 2513.69 3.719203 TSTtatistically significant(lpy <.05) different than PFOA decile(s) 1,2,3 .... and/or10 TableC2. Adjusted OdbdyRPaFtiOoAsD(eOc.iRl.e),a2n0d0905%loCrooncfhicdemniccealIMvedaica(l95S%urvCei1llafnocrePLirpoigdrCalmi(niNcal-C5h5e2m)istry Reference Pos, DPaFiOlA ChoOlRT 290%0CmLg] OCiRnTl>=2O0%0mCuLidl roe oo 2 04 0200 11 042s 3 os oss 09 0322 4 10 0422 12 0s30 S17 0839 16 07as 6 10 0423 12 0s30 7 09 041s 10 042s 801 0s24 1s 0634 9 1s 0735 14 063s 11 0525 10 oa Adjusted forage, BMIandalcohol OLRD'L= 91M0WmCgL] 0 06 0313 07 03s 08 0417 ess 08 0417 07 031s 10 0522 12 0526 09 041s ORHIDL=4S0AmgCiLdl wo. 140536 04 0113 19 0749 1040 17 0646 06 0218 12 0532 09 0323 29 1275 TORr! a99>d%= C1e50mlsdl wo. 08 0420 10 042s 140633 Loess 17 0742 10 042s 27 1264 24 1057 27 1263 Tl.TrNoeretsesd1A5f0i1OpsPRaONtDs0i)0ndr5CeomnnMenserosSvSe91a)foc rDL(03<-055g)il) ; - Co o dcemapT oa TietSnReLn, 2 noes ue wa 0 ous Moa uss wens Notts ed ortn aowe ua "Table C4. O4dsRatios (OR) and 95% ConfidenceIntervals(95%C.1 for Hepatic Clinical Chemistry Results, by PFOA Decile, 2000 Fuorochemnical Medical Surveillance Program (N = 552) mDouri OR. 9%CL ve 2 07 0225 3 0s a2 4 os 0227 S04 ome 6 02 0009 70s 0227 5 07 0224 9 1 043s us 0ess ATson OR. 9%CL oo. ox 0231 07 0220 07 0227 03 os 01 0109 oo eas 06 0124 09 0333 17 ess 21. ANodtajdujussootrdaeed,BM andskool 5. Adjusftoaegde igycearndisdcoes OR. 9%Cl wo 09 aaa 06 0122 ox 0227 04 omar or 0009 7 0227 0s 0120 10 o3as uo o4ar OR. 9%CL wo. 03 0012 03 0227 L036 04 ole 03 ola 09 0332 07 0224 17 06ss 07 022s arsonn.S aan? OR. 9%Cl OR svc. wo. oo. 02 ool 03 ooi2 03 0230 07 0227 10 033s 1 ose 04 oe 04 ons 03 0s 03 oss 10 0334 09 023) 06 0223 05 ome 20 0764 20 0766 07 0226 04 or7 `Table CS. Non-adiusted and Adjusted Regression PFOA Coefficients for Lipid Clinical Chemistry Results, 2000 Fluorochemical Medical Surveillance Program for Participants Who Did Not Self-Report Taking Cholesterol Lowering Medications (N = 506)and `All Participants (N = 552) Ln PFOA Non-adjusted LnPFOA Adjusted' All Locations NotPrescribed 00059 All Participants 00083 ANotPnreswcriebedp00051 All Participants 00050 LCnholesterol 0.0060 32 0.008 15 00076 0.0059 20 00099 0008 09 0.0106 63 0.0104 63 00130 00096 18 00121 00094 20 CoNotttPargeesGcrriobveed 00034 All Participants 0.0045 0.0089 70 0.0087 61 00021 00100 8 0001 00091 58 NDotePrecscraibted w0.0221 All Participants 0.0310 00139 11 0.0131 02 00266 00141 06 00348 00133 01 All Locations NotPrescribed 00005 All Participants 0.0018 nLDI 0.0089 96 0.0088 84 00021 00091 81 00029 00089 75 ANotPnreswcriebedp-00037 00157 81 All Participants 0.0041 00156 79 00106 00147 47 00092 00146 53 CottageGrove NotPrescribed -0.0036 00138 79 All Participants -0.0001 ~~ 0.0140 99 00049 00071 0000114465 7633 DNotePrcescaritbewd r00258 All Participants 00276 0.0199 20 0.0189 15 00302 00200 13 00204 00191 12 Table C5 (continued) LnHD) ALNlolotcPraestcriiboedns0.0307 ~~ 00079 All Participants 0.0295 0.0077 ANotuPrewscreibred p0.007 00136 All Participants -0.0030 ~~ 0.0137 CottageGrove NotPrescribed 0.0153 0012 Alparicipanss ~~ -00154 00121 DNotePrcescaritbewd r0.0256 00149 All Participants 0.0200 00140 0001 00183 00069 01 0001 00167 00067 01 68 00095 00131 47 83 00078 00130 55 21 00192 00120 11 20 00199 001170 09 09 00207 00141 14 16 00159 00133 23 LTnriglycerides ALNlolotPcreasctriibeodns0.0892 0.0185 0001 All Participants 0.0917 ~~ 00185 0001 ANotPnreswcriebedp0.0840 All Participants 0.0734 00288 004 0.0294 01 00711 00169 0001 00738 00168 0001 00980 00270 0004 00920 00274 0001 CoNotttPargeesGcrriobveed 0.0343 All Participants 0.0317 00316 28 0.0306 30 00280 00314 28 00269 00300 37 DNeoctaPrteusrcr_ib_ed__ 0.071 Al Participants 0.0983 0.0400 08 00394 01 00689 00376 07 00106 00373 01 1. MileRegressionModelLN (pdclinicalchemisy) = nrc +LN(8) +LN (BM) + LN (akoho) +LN (PFOA) for Hepati`cTCalbilneiCc6al. CNhoenm-iasdtjruystReesdulatnsd, A2d0j0u0stFeldu"o?roRcehgermeiscsailonMPedFiOcaAlCSouerfvfeiiclileanntcse Program (N=552) NLonnP-aFdOjuAsted ALdnjuPsFteOdA"? Coefficient _SE pvalue Coefficient SE AllLocations 00155 Antwerp 00025 CottageGrove 0.0141 Decatur 00394 LnAlkalinePhosphatase NPotrescr(iNb=e50d6) 00082 06 00093 00081 00037 0.0081 00137 85 00060 00139 00170 00140 00113 21 00127 0ou7 00140 00117 0091 04 00460 0012 00394 00192 All Participants (N = 552) p value 25 65 67 2 28 24 02 04 AllLocations 00189 Antwerp 00082 Cottage Grove -0.0091 Decatur 00445 00082 02 00139 56 0.0115 43 00176 01 00117 00080 14 00006 0.0080 45 00109 00140 44 00218 00139 12 0000100908 00119 00120 41 40 00492 00179 01 00429 00181 02 an AST NoPtrescr(iNb=e50d6) All Locations 0.0018 0.0086 83 0.0051 0.0086 55 0.0089 0.0087 31 Antwerp 0.0048 0.0137 7 00.00006269 00.0011328 8643 Cottage Grove ~~ -0.0281 0.0141 05 0.0258 0.0146 08 -0.0271 0.0145 07 Decatur 0.0205 0.0200 31 0.0114 0.0203 57 0.0062 0.0203 76 _ All Locations 0.0001 0.0085 92 Antwerp 0.0050 0.0135 Tn Cottage Grove ~~ -0.0254 0.0139 07 Decatur 0.0239 0.0194 2 0.0019 0.0085 82 0.0059 0.0087 50 -0.0034 0.0135 80 -0.0069 0.0137 62 -0.0240 0.0144 10 0.0250 0.0145 09 0.0179 0.0198 37 0.0104 0.0200 60 LoALT NPotrescri(Nb=e50d6) All Locations ~~ 0.0402 00143 005 00249 00132 06 00115 00136 40 Antwerp 00122 0020 58 00085 0022 70 00293 00222 19 CotiageGrove 00131 00215 54 00096 00209 65 00008 00208 69 Decatur 00954 00300 002 00704 00287 02 00s81 00287 04 All Participan(tNs = 552) All Locations 0.0457 00141 001 Antwerp 00151 00218 49 Cottage Grove 0.0143 00215 50 Decatur 01081 00278 0001 00316 00131 02 00170 00135 21 00118 00219 59 00311 0028 15 0000103849 00207 00215 67 53 00898 00270 001 00691 00274 01 Table C6 (continued) LnGGT NPotrescr(iN=b5e06d) All Locations 0.0409 Antwerp 00170 Cottage Grove -0.0088 Decatur 00754 00174 02 00307 58 0.0292 76 00344 03 00326 00166 05 00097 00163 55 00269 00294 36 00047 00295 87 00198 00286 49 00233 0070 39 00300 00344 02 00599 00329 07 NPort escri(Nb=5e5d2) AllLocations 00454 00170 01 Antwerp ood 00302 71 Cottage Grove -00062 00295 83 Decatur 0.0851 00319 01 21.. AAdjusdtedjffoorrLLannAAsgge,,LLennBBdMMI, LLnaAAlkccobhooll 00380 00163 02 00135 00158 39 00214 00290 46 00097 00286 73 0018 00289 53 00290 00278 30 00974 00320 003 00673 00308 03 `Table C7. Non-adjusted and Adjusted" Regression PFOA Coefficients for Thyroid-Related Clinical Chemistry Results, 2000 Fluorochemical Medical Surveillance Program (N = 552) NLonn-PaFdjOuAsted -- (Coficiem SE ALdnjPusFtOeAd pvc___ Codfficiem SF pve LnTSH ALNlolot cPraestcriiboedns0.0395 All Participants 0.0439 0.0204 05 0019 03 Antwerp. NotPrescribed 00509 00329 12 All Participants 0.0568 ~~ 00319 08 Cottage Grove NotPrescribed 00016 0.0310 96 All Participants 0.0031 ~~ 00295 92 Decatur NotPrescribed 00343 All Participants 0.0319 0.0497 49 0.0477 50 00360 00207 08 00400 00201 05 00391 00333 24 00463 0032 IS oul 032 73 00066 00306 83 00365 00513 48 00328 00490 44 LoT4 NAloltPLroesccaritbiedo_n_s0.0037 00054 50 All Participants 0.0049 00053 35 ANotnPrtesewrieberd p0.0022 0009 83 All Participants 0.0035 00098 72 Cottage Grove NotPrescribed 00124 00072 09 All Participants 0.0149 00072 04 Decatur NotPrescribed 0.0012 00126 92 All Participants _0.0015__00150__90 00057 0004 29 00072 0002 17 00041 00099 68 00047 00098 63 00093 0002 20 00121 0002 12 00083 00127 51 -00057__ 0011663 Table C7 (continued) LnFreeT4 ALNlolotPcreasctriiboedns0.0138 0.0044 002 All Participants 0.0144 00042 001 Antwerp. NotPrescribed 0.0108 ~~ 00078 17 All Participants 0.0113 00077 1S 00117 00043 01 00124 00042 003 00140 00078 07 00139 00078 07 CNootttPargeescGrriobvede -0.0093 00058 11 All Participants 0.0012 00057 04 DNotePcresacrtibeud r00138 00103 18 All Participants 0.0100 ~~ 0009 30 00071 0009 23 00091 00057 12 00184 00105 08 00146 00098 14 ALNlolotPcreasctriiboedns00107 00052 All Participants 0.0103 00050 ANotPmresweriebedp0.0222 0.0077 All Participants 0.0207 0.0076 Cottage Grove Not Prescribed 0.0026 0009 All Participants 0.0012 00092 DNotePrecseraibted w0.0317 0.0117 Al Participants 0.0331 __00108 1. Adjusted fo La Age, Ln BMI Ln Alcohol LaT3 04 04 005 007 79 90 008 003 00105 00053 05 0009 00051 05 00216 00077 01 00204 00076 01 00006 0009 95 00001 00095 99 007 ons 02 00203 00109 01 `Table C8. Number and Percent Subjects Above or Below Reference Points for Thyroid-Related Clinical Chemistry Results, by PFOA Decile, 2000 Fluorochemical Medical SurveillanceProgram(N = 552) PFOA 0T2S5H (>u5lU5/mL) Dele N(o No 1 00) 0(0) 2 1@ 00 3 0 00 4 1@ 24) 5 0 4M 6 00) 12 7 00 26) 3 00) 610) 9 00) 4@) 01 1 pvalue 67 03 T54 (g2/d1L)20 Neo) Neo) 00) 00) 00 00 1 00 00 00 00 10 00) 0(0) 0@ 00 0 00) 0 00) 00 00 a1 49 0Fr7e0e T4 (2ng/8dL) Neo) Neo 0(0) 12 1 0 0 21 10 0 0(0) 0(0) 00 1 20 00) 00) 00 68 47 60T3 (ng/dL) 281 Neo) NCO 00) 1) 00 00 00 00 0 00 00 00 00) 24 00 1@ 00 20 00) 000) 00) 20) ui 38 Table C9. Predicted Thyroid-Related Clinical Chemistry Results Based on Multiple Regression Models for 40 Year Old Male with BMI = 28 and Drinks 0.5 Alcohol Beverages per Day (N = 552) Reference Range PFOA Serum Concentration (ug/ml) 0.005 001 0.10 050 1.00 5.00 10.00 50.00 10000 Predicted Predicted TSH (WIU/mL) T4 (ugidL) (0255.5) (4512.0) 157 828 161 824 177 8.10 189 8.00 1.94 797 207 788 213 7.84 227 774 233 71 Predicted Free T4 (ngidL) 020-153) 116 Lis 12 1.09 1.08 1.06 105 1.03 102 Predicted T3 (ng/dL) (60-181) 19 120 123 125 125 127 128 130 3 AppendiDx "Table D1. CorrelationCoeficints (sameasTable Al) Lnarable) LoFOA)_ LaPFOS) PPFROOAS os10seens AgBMeI o00r3 AlChoolheostlerol 0o04e HLoDLL "0g0e0ese GTrliugchoeseerides 00201100040 AAlsktPhos 000081 AaSTrALT aoinsee TGoaurl Blinbin ~~ -a0108%+++ SDiHrec Bilinbin ~~ 00.0096 FrTMeed a00s3e nm eee 01006 000.1000 000068 o0t0e7 "0001830 000080 0010000 20013 3 o004o 006 es "epoosl wleped oonor InAs) 0120 00s oaissee ionee 0ol2s60s 000080 000035 "O00s8e "0"00031 "0"0a9pes a LaBMD__Ln(Akohol) "10026000 00000 S0oaaqzreeerree o02nseeeee oodstveese P0e3rge0ere 00g3een "000088 "o170ee aw o1l0or 0o3o7l0000 o01g2r0e 0p0a1se ooi176eteeee 002000s 0"000 000187000 em Age (years) by PFOS (ng/mL) LEI| | 1 2 o3 s a4 5 8 LNAge(years)byLN PFOS(ug/mL) |g wl TA[eai e RBaaR e l $0 [ETERS 0]. iE R J on BMI by PFOS (ng/mL) w Tl e _ 1| oT i oR . || sg Ge ofBRTLE |I| weep eLN Be MI by LN Pe FOS (ug/e mL) aull:l TP ET N 3057;1 (TnoS l E Se EeNil 334 MY. < ey | E rNEl Te ERTagEerel Alcohol (drinks/day) by PFOS (ug/mL) fT it Eeoe py | 1 2 3 4 8 LN Alcohol (drinks/day) by LN PFOS (g/mL) |[ar --TM-- Li d |He frie IlEELT 5 i l(irhaeclieeidiiEnse | ros Glucose (mg/dL) by PFOS (ug/mL) [R=1 | gl a I | Sot LN Glucose (mg/dL) by LN PFOS (ng/mL) 18 1: EE orcs Cholesterol (mg/dL) by PFOS (ug/mL) Cy Volz, 7 | ] a0 Sir a || | RE | aliret o 1 z p3ros + 5 5 | | LN Cholesterol (mg/dL) by LN PFOS (g/mL) We| sed fost ial | 2geo3e5XETEIR ARE ER) G Tt al g1 TER prosT} LDL (mg/dL) by PFOS (ug/mL) | 3 PaEtinLa 1| fic || A CE rEIosT--I w ---- || & 1 LN LDL (mg/dL) by LN PFOS (ng/mL) JT F 1 oirweeioa | | gGriDfl(FBE rEiE mRavRN er. || Fl NRHA | 1 S| 1. ete" J "52 4ner8 o 1 3 | Ss | =F---- | HDL (mg/dL)by PFOS (ug/mL) wi. wd, . 1 : | LEA | SEE. piStEEe a I ft a a Ei LN HDL (mg/dL) by LN PFOS (ng/mL) i ie $7 od, LloTTEERLREEARL ? iy < 3% eraos 5 1 2 Triglycerides (mg/dL) by PFOS (g/mL) ee | 0d -- = | so BariyF 5a | Eon yo mnoll ie, : IEETTI TrTeTed | fn can! LN Triglycerides (mg/dL) by LN PFOS (g/mL) [ee] Tr ss]:vcEERRSLTE E a 2 1. EER a=e EAR e SRITE E 4 DN i FE a i praos a il | Alkaline Phosphatase (IU/L)by PFOS (ug/mL) a wl hEoSd nR 'tI Ra. | | -. wf] Etieee emer) LN Alkaline Phosphatase (IU/L) by LN PFOS (ng/mL) ee Lia Hl Yvad a AST (IU/L) by PFOS (ng/mL) Js]I + | 5% eo a a | edna BEER | || | wf: a* eros | LN AST (IU/L) by LN PFOS (ug/mL) E5 Rir RLf r Em r ; 3 2 er4os 6 1 2 ALT (IU/L) by PFOS (ng/mL) || e 2oe e ; ny. - | || | B oflfaiin Ea el : A| | oo 1 27 488 | LN ALT (IU/L) by LN PFOS (ng/mL) TR iil C[0 n aAeA r | w OEE | |= : ? AST/ALT by PFOS (ng/mL) TI 17- | 1s Foleo. Legg SE]I Ere eros | LN AST/ALT by Le N PFe OS (ug/ml) ooes] 2 Lo0s] TEe ETEn R 2 Ceaanz t Tl EFCAL id RnAdm TaE | T MTE TEY E Ee | pros GGT (IU/L) by PFOS (ng/mL) [wl | | 250- 1 RBel . I | edi iE | 9 Ld LN GGT (IU/L) by LN PFOS (ug/mL) Id ; 2d [Leeder :i x hE 4 2 REE [ = 3 2 "a 0 1 2 Total Bilirubin (mg/dL) by PFOS (g/mL) A : | i a an SEE 0 "FE REE ; ros LN Total Bilirubin (mg/dL) by LN PFOS (g/mL) (3:4: ordmmnin 13 i Eo oxen mtramexa=. l= J. NESTE | orcs a Direct Bilirubin (mg/dL) by PFOS (ug/mL) | gm |BJooisl....- Tn LN Direct Bilirubin (mg/dL) by LN PFOS (ng/mL) 1 - . | i oo 2 | 2: 1 -- | J." 3 2 4 @& 1 2 (TEE TSH (uIU/mL) by PFOS (g/mL) -hy E 1 Bao | 2d T. .| CE | Tis#rov s ss LN TSH (WU/mL) by LN PFOS (ug/mL) 7 .| gi raed | I23R eUlRS SERRE 2 -A rr | TE Eaeros 61: : ia T4 (ug/dL) by PFOS (ng/mL) ar I | IX? . | er ai w esc- rlperorrar) ros LN T4 (ug/dL) by LN PFOS (ug/mL) Ir Pepin. nl Riese| Tal od PUREE 1 Ai he - worn a | Free T4 (ng/dL) by PFOS (ug/ml) || owliin CoC5 |I |; h 2 g Ae oTE ET Ba 2 ord HH | 0 T 8 1 3 3 m + 5g8 pros LN Free T4 (ng/dL) by LN PFOS (ug/mL) 0s] Na wl rrFTa. He HIRE gee ol" TC REE o S 23] | r TEr E r a 6 ] qe pros T3 (ng/dL) by PFOS (ng/mL) 1 | =f - | 1B p00] LE | om to In SeFE | e TTRrbTTror s TTr Y| ST LN T3 (ng/dL) by LN PFOS (ug/mL) o5s]]1] ' | r.e Pe1o rly Gi ady i ". 41w[R freaodi RiE a ow. NTR CEEeroT s T Appendix E Table EI. Non-adjusted and Adjusted Regression PFOS Coeflicients for Lipid Clinical Chemistry Results, 2000 Fluorochemical Medical Surveillance Program ON =506) Non-adjusted Ln PFOS Coefficient SE Adjusted La PFOS p value Coefficient' SE p value All Locations 0.0151 Antwerp 0.0091 Cottage Grove 0.0151 Decatur 00247 All Locations Antwerp Cottage Grove Decatur 0.0153 00068 0.0115 0.0457 All Locations -0.0161 Antwerp 0.0014 Cottage Grove -0.0018 Decatur 0.0242 LCnholesterol 00081 06 00157 0044 83 0.0049 00130 25 0.0141 00173 1s 0.0235 LaLD) 00120 20 0021 75 00198 56 0029 07 00135 00137 0.0187 0.0421 LnHDL 00108 14 00093 00185 94 0.0020 00m 92 00114 00186 20 00166 00081 05 00128 70 00134 30 003 18 00120 26 00194 48 00203 36 0025 09 00094 33 0017s 91 0 52 00172 34 All Locations Antwerp Cottage Grove Decatur 0.0778 0.1066 0.0453 0.0416 Table El. (continued) LTnriglycerides 00255 002 0.0752 00381 006 0.1063 00463 33 0.0393 00501 41 0.0304 1. Adjusted for Ln Age, Ln BMI, Ln Alcohol 00230 001 00357 003 00455 39 00463 51 Table E2. Non-adjusted and Adjusted" Regression PFOS Coefficients for Hepatic Clinical Chemistry Results, 2000 Fluorochemical Medical Surveillance Program N=506) Non-adjusted Adjusted Ln PFOS Ln PFOS Coefficient SE pvalie Coefficient SE pvalue Ln Alkaline Phosphatase: All Locations 0.0189 Antwerp 0.0051 Cottage Grove 0.0137 Decatur 0.0405 002 09 00130! 0.0074 00110 24 0.0109 50 0.0185 78 00028! 00186 88 00140 00185 45 00165 41 00009" 0.0091 00170 59 00171 59 00238 09 00381! 0.0346 00237 11 00234 14 Ln AST All Locations -0.0006 oon7 96 Antwerp 0.0356 00183 05 Cottage Grove 0.0081 0.0210 70 Decatur 00111 00250 66 0.0008" 0.0063 0.0370! 004200 0.0140! 0.0123 00103! 0.0082* 00116 97 00117 59 00182 04 00184 02 0.0214 51 00216 57 00247 68 00245 74 All Locations 0.0354 Antwerp 0.0490 Cottage Grove 0.0510 Decatur 0.0927 All Locations 0.0534 Antwerp 0.0287 Cottage Grove 0.0417 Decatur 0.0583 Table E2. (continued) LaALT 00195 07 00295 10 0.0311 10 00376 01 003430 0.0106 00499) 00734 0.0490! 0.0409 00826! 0.0780 GGT 00237 02 00416 49 00425 33 00433 18 0.0558! 0.0251 00260! 0.0088 001! 00152 00612! 0.0524 00179 06 00183 58 00293 09 0.0289 01 00299 10 00310 19 00351 02 00347 03 00225 01 00217 25 00291 S51 00389 82 00414 44 00d01 71 00424 1s 00401 19 Table E2. (continued) Total Bilirubi All Locations -0.0415 Antwerp 0.0344 Cottage Grove -0.0142 Decatur 0.0531 0.0138 003 00240 15 00201 48 0.0254 04 21 AAddjjuussteedd oforr LLnn AAggee, LLni BTriMghyeLrnidAese,ohLanlAlcohol 0.0356! 200278 0.0354! 0.0324 00107" 0.0142 00565! 0.0525 00136 01 00135 04 00241 14 00247 19 00206 61 0.0205 49 00255 03 00250 04 | for Thyroid Cli`nTiacballe CEh3.emiNsotnr-yadRjeussutletsd,aZndRAdFjustled"eRegaressiMoendiCcoaeflfiScuirevnetisllance Program Non-adjusted Ln PFOS Adjusted" Ln PFOS LaTsH All Locations 0.0227 0.0275 41 Antwerp 00306 0047 49 0.0232 0.0228 40 00369 00446 41 Cottage Grove ~~ 0.0224 0.0442 S51 0.0094 0.0451 83 Decatur -0.0132 0.0620 83 0.0125 0.0626 84 LaT4 All Locations 0.0029 0.0073 69 0.0004 0.0072 95 Antwerp -0.0183 Cottage Grove 0.0010 0.0133 17 00104 92 0.0179 0.0131 18 00067 00105 51 Decatur 0.0101 0.0156 52 0.0062 0.0155 69 LnFreeT4 All Locations 0.0080 00059 18 00064 00058 27 | Antwerp 00207 00105 04 00208 00104 05 | Cottage Grove ~~ 0.0108 0.0083 20 0.0126 0.0082 13 | Decatur -0.0015 0.0129 91 -0.0019 0.0129 88 Table E3. (continued) | LoT3 | All Locations 00053 00071 46 00061 00071 39 | Antwerp 00173 00106 10 00186 00104 08 | CottageGrove -0.0200 00135 14 00190 00138 17 Decatur 00325 00147 03 00304 00146 04 1. AdjusftoreLdn Age, Ln BMLnAIlco,hol