Document Y9pjBmZ3QqO8NY83Mk6BmqpL8

r--- ----------- - Corporate Health Physics Corporate Occupational Medicine Corporate Product Responsibility Corporate Toxicology 3M Medical Department 3M Center, 220-2E-02 PO Box 33220 St. Paul, MN 55133-3220 651 733 1110 r YX - 0 5 0 0 - /37 AR2V2ol6-H0L330 3M April 21, 2000 ( SkOCD00<p00lG) Attachments To April 21, 2000 Letter to C. Auer from W. Weppner (Medical Surveillance and Epidemiology) Final Reports 1. An Epidemiologic Investigation of Clinical Chemistries, Hematology and Hormones in Relation to Serum Levels of Perfluorooctanesulfonate in Male Fluorochemical Production Employees. [Included is a 3M report of medical surveillance data. Also included is a published paper from this 3M report.] 2. Fluorochemical Exposure (Serum) Assessment of (3M) Decatur Chemical and Film Plant Employees. [Included are a 3M study protocol and report.] 3. Mortality Study of Employees at 3M Plant in Decatur, Alabama [Included is a University of Minnesota final report.] 4. Determination of Serum Fluorochemical Levels in Sumitomo 3M Employees [Included are a 3M study protocol and final report.] 5. Analysis of Selected Decatur Employee Serum for Sulfonic and Carboxylic Fluorochemicals [Included is a 3M technical report.] 6. Fluorochemical Control Study [Included is a 3M report of medical surveillance data.] 7. Working Memorandum on Data Quality Assessment [Included is a Battelle Laboratory memorandum regarding the mean and range of perfluorooctanesulfonate sera sample data collected by 3M from current and historical human populations.) 001068 Corporate Health Physics Corporate Occupational Medicine Corporate Product Responsibility Corporate Toxicology 3M Medicai Department 3M Center, 220-2E-02 PO Box 33220 St. Paul, MN 55133-3220 651 733 1110 (D 3M An Epidemiologic Investigation of Clinical Chemistries, Hematology and Hormones in Relation to Serum Levels of Perfluorooctane Sulfonate in Male Fluorochemical Production Employees Following reports of the finding of organic fluorine in sera samples, a fluorochemical medical surveillance program began at 3M's Decatur manufacturing facility in the late 1970's. The surveillance program has generally consisted of annual or biannual tests of clinical chemistries, pulmonary function, blood counts and a biomonitor of fluorochemical exposure. A total organic fluorine measurement was routinely done until 1993. This measures the amount of fluorine that was covalently bound to carbon in the serum sample. When test data were available, a company physician reviewed each employee's results. These physicians did not, and have not, found abnormalities in individuals that they felt were related to fluorochemical exposure. That is, medical conditions, medications and lifestyle factors adequately explained the laboratory abnormalities (which one expects to find in this type of program.) Beginning in 1994, the 3M Decatur (Alabama) plant medical surveillance program incorporated a serum measurement of perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA). Total organic fluorine was not measured. A formal report was written of the aggregate analyses conducted of the medical surveillance clinical program data for the Decatur (Alabama) and Antwerp (Belgium) employees who voluntarily participated in 1994, 1995 and 1997. The findings from this aggregate 001069 analysis suggested that, among these participating Antwerp and Decatur male fluorochemical production employees, significant hematological, clinical chemistry and hormonal abnormalities were not associated with serum PFOS levels up to 6 ppm. It was not possible to derive inferences from the few employees with serum PFOS levels > 6 ppm. Limitations of this study include its cross-sectional design, the voluntary participation rates, the few subjects exposed at the highest levels, and the lower levels of serum PFOS measured among these employees compared to those estimated to cause effects in laboratory animals. Results of the hepatic and lipid clinical chemistry tests were published in the Journal o f Occupational and Environmental Medicine (1999;41:799-806). In the Spring of 2000, medical surveillance will again be offered to 3M fluorochemical production employees at the Antwerp and Decatur manufacturing sites. 0010*70 April 22, 1998 An Epidemiologic Investigation of Clinical Chemistries, Hematology and Hormones in Relation to Serum Levels of Periluorooctane Sulfonate in Male Fluorochemical Production Employees Geary W. Olsen, D.V.M., Ph.D. Jean M. Burris, R.N., M.P.H. Jeffrey H. Mandel, M.D., M.P.H. Larry R. Zobel, M.D., M.P.H. Medical Department, 3M Company, 220-3W-05, St. Paul, MN 55144 Page 1 001071 ABSTRACT 3M manufactures products which contain chemical compounds, either as intentional components or residual impurities, that have as a parent molecule, perfluorooctane sulfonyl fluoride. These chemicals include: perfluorooctane sulfonate (PFOS), N-ethyl perfluorooctanesulfonamide, N-ethyl perfluorooctanesulfonamido ethanol, N-methyl perfluoroctanesulfonamido ethanol and chemicals derived from it, and the mixture of mono-, di- and tri [N-ethyl perfluorooctane sulfonamidoethyl] phosphates. There may be other precursors in the workplace. These molecules enter a number of product applications (e.g., surfactants, food packaging additives, polymers). These compounds may be expected to transform metabolically, to an undetermined degree, to PFOS as an end-stage metabolite. Potassium perfluorooctane sulfonate (CgFi70S02K+) is, itself, a surfactant used as a wetting and foaming agent in industrial and commercial processes. Subchronic studies in rats and primates suggest there may be a potential for cumulative toxicity with PFOS over time with the primary effect related to metabolic wasting. Although the mechanism of toxicity is not fully understood, toxicity may be due to an effect on peroxisome proliferation, fatty acid metabolism, membrane function, protein synthesis and/or mitochondrial bioenergetics. Medical surveillance has been routinely performed on 3M fluorochemical production workers (in Decatur, Alabama and Antwerp, Belgium) with potential exposure to PFOS and/or to perfluorinated precursors that may metabolically degrade to PFOS. The purpose of this study was to provide an analysis of the hematology (hematocrit, hemoglobin, red blood cells, white blood cells and platelet count), clinical chemistries Page 2 001072 (alkaline phosphatase, gamma glutamyl transferase, aspartate aminotransferase, alanine aminotransferase, total and direct bilirubin, blood urea nitrogen, creatinine, glucose, cholesterol, low density lipoproteins, high density lipoproteins and triglycerides) and hormonal parameters (cortisol, dehydroepiandrosterone sulfate, estradiol, follicle stimulating hormone, 17-alpha hydroxyprogesterone, luteinizing hormone, prolactin, sex hormone binding globulin, free testosterone, bound testosterone, and thyroid stimulating hormone) in relation to serum PFOS as determined by high performance liquid chromatography mass spectrometry methods. These relationships were assessed in fluorochemical production employees from two time periods, 1995 (N = 178) and 1997 (N = 149). Descriptive simple and stratified analyses, Pearson correlation coefficients, analysis of variance and multivariable regression were used to evaluate for possible associations between PFOS and each hematological and clinical chemistry test and hormonal assay. Age, body mass index, current alcohol consumption (drinks per day) and cigarette use (cigarettes smoked per day) were potential confounding factors that were considered in the analyses. Multivariable regression models were fitted with PFOS analyzed as a continuous variable using linear as well as non-linear transformations in order to maximize the possibility of finding associations between PFOS and the parameters of interest. Four categorizations of serum PFOS levels were assessed in relation to the response variables: 0 - < 1 ppm; 1 - < 3 ppm; 3 - < 6 ppm; and > 6 ppm. In 1995, mean serum PFOS levels by category were 0.49 ppm, 1.82 ppm, 4.12 ppm and 8.17 ppm, respectively. In 1997, mean serum PFOS levels by category were 0.52 ppm, 1.78 ppm, Page 3 001073 3.87 ppm and 7.20 ppm, respectively. For both years, 95 percent of the employees' serum PFOS levels were below 6 ppm. Although the two plant populations differed by age, body mass index and alcohol consumption, no consistent associations, by both plant locations and year, were observed between the clinical chemistries, hematology and hormone parameters and the employees' serum PFOS levels. The findings from this study suggest that, among these Antwerp and Decatur male fluorochemical production employees, significant hematological, clinical chemistry and hormonal abnormalities are not associated with serum PFOS levels up to 6 ppm. It is not possible to derive inferences from the few employees with serum PFOS levels > 6 ppm. Limitations of this study include its cross-sectional design, the voluntary participation rates, the few subjects exposed at the highest levels, and the lower levels of serum PFOS measured among these employees compared to those that caused effects in laboratory animals. Page 4 001074 INTRODUCTION 3M manufactures products which contain chemical compounds, either as intentional components or residual impurities, that have as a parent molecule, perfluorooctane sulfonyl fluoride. These chemicals include: perfluorooctane sulfonate (PFOS), N-ethyl perfluorooctanesulfonamide, N-ethyl perfluorooctanesulfonamido ethanol, N-methyl perfluoroctanesulfonamido ethanol and chemicals derived from it, and the mixture of mono-, di- and tri [N-ethyl perfluorooctane sulfonamidoethyl] phosphates. There may be other precursors in the workplace. These molecules enter a number of product applications (e.g., surfactants, food packaging additives, polymers). These compounds can be expected to be transformed metabolically, to an undetermined degree, to PFOS as an end-stage metabolite [Gibson et al., 1983]. Potassium perfluorooctane sulfonate (CgFn0S02K+) is, itself, a surfactant used as a wetting and foaming agent in industrial and commercial processes. Potassium perfluorooctane sulfonate is readily absorbed by ingestion [Johnson and Ober, 1979; O'Malley and Ebbens, 1980]. Ninety five percent of a single oral dose of [14C] PFOS administered to male rats was absorbed within 24 hours [Johnson and Ober, 1979]. After a single, 24-hour occluded dermal exposure to PFOS at a dose of 5000 mg/kg, total serum organic fluorine concentrations were 10.3 and 0.9 ppm for male and female albino rabbits, respectively [O'Malley and Ebbens, 1980]. Twenty eight days after dosing, total serum organic fluorine concentrations had risen to 130.2 and 128.0 ppm for male and female albino rabbits, respectively. On the other hand, no quantifiable Page 5 001075 organic fluorine could be detected 28 days after a single, 24 hour occluded dermal exposure to a 0.06% solution of PFOS in water at doses of 0, 0.003, 0.06, and 0.3 mg PFOS solution/kg, respectively, to 3 male and 3 female albino rabbits per dose group [Glaza, 1995]. Once in the body, PFOS concentrates primarily in the liver of rats [Johnson et al., 1979]. Eighty-nine days after a single intravenous dose (mean 4.2 mg/kg) of radiolabeled PFOS, mean tissue concentrations (pg PFOS equivalent/g tissue) were: liver, 20.56; plasma, 2.21; kidney, 1.09; lung, 1.06; spleen, 0.51; bone marrow, 0.46; red blood cells, 0.45; adrenals, 0.41; testes, 0.36; skin, 0.35; muscle, 0.29; subcutaneous fat, 0.20; eye, 0.16; abdominal fat, < 0.08; and brain, < 0.05. Johnson et al. [1979] observed that 30.2 percent of the dose 89 days after administration had been excreted in the urine and 12.6 percent in the feces. Analyses of the urine, feces and tissues have suggested that PFOS is not metabolized [Johnson et al., 1984], The plasma half-life was calculated to be 7.5 days after a single oral dose of radiolabeled PFOS (mean dose, 4.2 mg/kg) in solution to three male rats [Johnson and Ober, 1979]. There appears to be significant enterohepatic circulation of PFOS with both urinary and fecal excretion [Johnson et al., 1979; 1980; 1984]. In male rats, cholestyramine administered in the feed decreased the retention of radiolabeled PFOS in liver, plasma, and red blood cells, 3.8, 7.7 and 6.0 fold, respectively, and increased its elimination via feces 9.5 fold after the rats were given intravenous radiolabeled PFOS (mean dose, 3.4 mg/kg) [Johnson et al., 1980; 1984]. There was a lower clearance rate of 14C in the urine compared to control animals because of the increased rate of fecal elimination. Cholestyramine is a bile acid sequestrant that acts by binding bile acids in Page 6 001076 the intestinal tract. This reduces bile acid resorption and its return to the liver. Decreased flow of bile acids in the enterohepatic circulation results in increased conversion of hepatic cholesterol into bile acids. This results in a decline in hepatic cholesterol concentration and the stimulation of low density lipoprotein (LDL) receptor synthesis, which subsequently produces a decline in serum LDL cholesterol levels. Upon acute exposure, PFOS was moderately toxic by oral administration [Gabriel, 1976; Dean et al., 1978; Rusch and Rinehart, 1979], but not dermal [O'Malley and Ebbens, 1980]. There have been two acute oral toxicity studies reported [Gabriel, 1976; Dean et al., 1978], In the more recent study PFOS was suspended in a 20% acetone/80% com oil mixture and administered orally by gavage levels to 5 male and 5 female rats per group at the following dosages: 100, 215, 464 and 1000 mg/kg [Dean et al., 1978], Animals were observed for 14 days. The acute oral LD50 values (95% confidence limits in parentheses) were male rats, 233 (160 - 339) mg/kg; female rats 271 (200 -369) mg/kg; combined male and female rats; 251 (199-318) mg/kg. Clinical signs included diarrhea, hypoactivity, decreased limb tone, ataxia, comeal opacity, ptosis, piloerection, prostration and tremors. In the previous study, PFOS was administered in water and the LD50 in the rat was determined to be 1.25 - 2.50 g/kg [Gabriel, 1976]. Gabriel's results appear to be inconsistent with subsequent toxicity studies. In an acute inhalation toxicity study [Rusch and Rinehart, 1979], a series of onehour inhalation exposures in rats at exposure concentrations of PFOS at 24.09, 7.05, 6.49, 4.88, 2.86, 1.89 and 0.0 mg/L produced 100 percent mortality at the highest level and partial mortality (10 - 80%) at all other PFOS levels. Observations included dyspnea, tremors, convulsions, hypersensitivity, hypoactivity, excessive salivation and lacrimation Page 7 001077 and general poor condition. The LC50 was determined to be 5.2 mg/L (95% Cl = 4.4 6.4 mg/L). PFOS has been shown to be a potent inducer of hepatic peroxisomes and fatty acid beta-oxidation in the rat [Ikeda et al., 1987] and mouse [Sohlenius et al., 1993]. After feeding male rats for two weeks with a powdered chow containing 0.02% PFOS, hepatic catalase, fatty acyl-CoA, carnitine acetyl transferase and carnitine palmitoyl transferase increased by 1.74, 4.90,6.84 and 1.69 fold, respectively, compared to control animals [Ikeda et al., 1987]. PFOS also induced cytochrome P-450 activity. Male mice administered periluorooctane sulfonic acid at a concentration of 0.05% weight/weight in the diet for 5 days resulted in weight loss and increased peroxisomal fatty acid betaoxidation, peroxisomal catalase activity, Q-hydroxylation of lauric acid, cytosolic epoxide hydrolase activity and cytosolic DT-diaphorase activity [Sohlenius et al., 1993]. Haughom and Spydevold [1992] fed 0.02% periluorooctane sulfonic acid in the diet for 7 -1 4 days to male Wistar rats which resulted in increased liver weight, liver triacylglycerol, liver free cholesterol and decreased liver cholesterol ester as well as decreased serum cholesterol and triacylglycerol levels. There was reduced cholesterol synthesis from acetate, pyruvate and hydroxymethyl glutarate but no reduction in synthesis from mevalonic acid in the hepatocytes from the treated rats. The activity of liver hydroxymethyl glutaric acid-Co-A reductase (HMG-CoA) and acyl-CoA cholesterol acyltransferase (ACAT) was reduced. Haughom and Spydevold [1992] suggested that the hypolipidemic effect of perfluorooctane sulfonic acid may be due to downregulation of HMG-CoA reductase and ACAT with enhanced fatty acid oxidation in the liver. This would subsequently reduce very low density lipoprotein (VLDL) production by the liver. Page 8 001078 Recently, Nabbefeld et al. [1998] tested the hypothesis that PFOS and other fluorocompounds may act as peroxisome proliferators by displacing fatty acids from liver fatty acid binding protein (L-FABP). 10 pM PFOS caused a 66 percent reduction of the fluorescently labeled fatty acid analog 1l-(5-dimethylaminonapthalenesulfphonyl)undecanoic acid from L-FABP in vitro. Comparable results were observed for bovine serum albumin. These findings demonstrated that PFOS has a high affinity for fatty acid carrier proteins and can displace the endogenous ligand. Results from three subchronic studies have been reported [Goldenthal et al., 1978a; 1978b; 1979]. PFOS was fed in the diet of Charles River CD rats at 0 (control), 30, 100, 300, 1,000 and 3,000 ppm for 90 days [Goldenthal et al., 1978a]. At the 300, 1,000 and 3,000 ppm dosage level, all rats died prior to scheduled termination. Toxicity signs included emaciation, convulsions, ocular and anogenital discharges, increased sensitivity to external stimuli and reduced motor activity. Histopathology showed compound-related lesions which included hepatic hypertrophy and necrosis, thymic and splenic follicular atrophy, bone marrow hypocellularity and atrophy of mesenteric lymph nodes, small intestinal villi and skeletal muscle. Among the 100 ppm dose group there was weight loss, elevated plasma creatinine phosphokinase, alkaline phosphatase, blood glucose and blood urea nitrogen, decreased hemoglobin, hematocrit, erythrocyte and leukocyte counts, hepatic enlargement and necrosis, and stomach discoloration and hemorrhage. Among the 30 ppm dose group there was weight loss, elevated plasma glutamate-pyruvate transaminase and plasma glutamate oxalacetate transaminase, and liver discoloration. Page 9 001079 In a subchronic (90 day) study [Goldenthal et al., 1979], two male and two female rhesus monkeys per group received 0 (control), 10, 30, 100 and 300 mg/kg/day of PFOS by oral gavage. All animals (except controls) died within 20 days and timing was related to dosage. Monkeys treated with 300 mg/kg/day died between the second and fourth day. Monkeys treated with 100 mg/kg/day died between the 3rd and 5thday. Monkeys treated at 30 mg/kg/day died between the 7th and 10thday and those treated at 10 mg/kg/day died between the 11th and 20th day of the study. Signs of toxicity at each dosage level were comparable and included anorexia, diarrhea, decreased activity, emesis, weight loss, marked weakness, prostration, and general body tremors. There were no consistent histopathologic changes with exposure. Adrenal changes, including congestion, hemorrhage and lipid depletion of the adrenal cortex, were observed in all dose groups. An additional subchronic rhesus monkey study was subsequently initiated at much lower dosages [Goldenthal, 1978b]. PFOS was administered by oral gavage to two male and two female monkeys at dosages of 0,0.5,1.5 or 4.5 mg/kg/day for 90 days. Animals treated at the 4.5 mg/kg/day dosage level died or were sacrificed in extremis by the seventh week with signs of gastrointestinal toxicity comparable to those observed in the previous rhesus monkey study by Goldenthal et al. [1978a]. Also, in the 4.5 mg/kg/day dose group, mean serum cholesterol levels declined from 183 mg/100 ml to 99 within 30 days. SGOT increased from 36 to 95 u/1 and alkaline phosphatase decreased from 1088 to 590 u/1. SGPT remained unchanged. Flistopathology showed compound-related marked diffuse lipid depletion of the adrenals as well as diffuse atrophy of the pancreatic exocrine cells. Animals in the 0.5 mg/kg/day and 1.5 mg/kg/day dosage groups survived to the end of the study. Occasional diarrhea, anorexia and emesis were observed. There Page 10 001080 was a decrease in serum alkaline phosphatase and inorganic phosphate in the 1.5 mg/kg/day group and a slight decrease in alkaline phosphatase in the 0.5 mg/kg/day group at the end of 90 days. Histopathology was unremarkable in both dosage groups. A no observable adverse effect level (NOAEL) was not identified from any of the above three subchronic (90 day) studies. The results from these three subchronic studies suggest there may be a potential for cumulative toxicity over time with the primary toxic effect related to metabolic wasting. This may be due to an effect on peroxisome proliferation, fatty acid metabolism, membrane function, protein synthesis and/or mitochondrial bioenergetics. To date, there are no data regarding the chronic toxicity and carcinogenicity of PFOS. PFOS was not observed to be mutagenic in several Salmonella typhimurium strains with or without metabolic activation [Jagannath and Brusic, 1978]. PFOS was negative in an in vivo mouse bone marrow micronucleus assay [Murli, 1996]. There have been two teratology studies conducted with PFOS. Oral administration, via com oil, of PFOS at doses of 0 ,1 ,5 and 10 mg/kg/day to pregnant rats during days 6 - 15 of gestation resulted in fetuses with what was initially reported as teratogenic changes in the eye. [Gortner et al., 1980]. These fetal lens abnormalities were subsequently interpreted to be artifacts of the tissue sectioning process. Maternal body weights in the high dose group were significantly reduced but no significant treatment-related teratogenic or embryotoxic effects were reported. In the second teratology study, PFOS was administered in com oil by oral gavage to groups of 25 pregnant rats on days 6 -15 of gestation at doses of 0, 1,5 and 10 mg/kg/day [Wetzel et ah, 1983]. Maternal body weights and food consumption at 5 and 10 mg/kg/day were Page 11 001081 significantly reduced compared to the control animals. Two female rats in the high dose group died before day 20. Clinical signs in surviving dams included hunching, thinness, alopecia, rough haircoat and anorexia. Treatment-related effects, primarily occurring in the high dose group, included increase resorptions and fetal death, decreased fetal body weight, delayed skeletal ossification, cleft palate, subcutaneous edema and cryptorchidism. During the past 15 to 20 years there have been several endeavors designed to ascertain the health and exposure status of workers involved with fluorochemical production at the company's Decatur, Alabama and Antwerp, Belgium plants. Medical surveillance has been routinely conducted of fluorochemical production workers at both plants. Medical surveillance activities analyzed for total serum organic fluorine levels until the mid-1990's when serum PFOS determination, quantifiable by liquid chromatography mass spectrometry, became incorporated in the biennial medical surveillance examinations. However, we are aware of one occasion in 1979 where the serum of 5 Decatur employees was measured for PFOS by electron capture gas chromatograph and microwave plasma detection methods [Central Analytical Laboratory, 1979]. Total serum organic fluorine levels for these five employees were 10.1, 5.7, 9.4, 11.8 and 4.1 ppm. The percent of PFOS found was 60%, 70%, 80%, 55% and 65% of the total serum organic fluorine levels, respectively. In 1981, selected clinical chemistries and hematology values of Decatur employees in the chemical plant were compared to those results of employees in the adjacent 3M film plant [Roach, 1982; Schuman, 1982]. There were no significant correlation coefficients between total serum organic fluorine and gamma glutamyl transferase, serum glutamic oxaloacetic Page 12 001082 transaminase, serum glutamic pyruvic transaminase, alkaline phosphatase, cholesterol, hemoglobin or red blood counts. However, this analysis was limited in scope (no dose response analysis), did not analyze specifically for PFOS, and did not account for several potential confounding factors. Another research initiative into the health status of Decatur employees was a retrospective cohort mortality study (1961 - 1991) conducted of former and current employees who had worked at least one year at the Decatur plant [Mandel and Johnson, 1995]. Vital status was determined for 99.7% of the 1,957 cohort members who had worked in the chemical and film plants. A total of 74 deaths were identified compared to 117.7 expected (U.S. rates). Among male employees who had worked only in the Decatur chemical plant, there were 32 deaths compared to 44.1 expected. There were no specific causes of death that had significantly elevated standardized mortality ratios. Because of its more recent construction in the 1970's, there has not been a retrospective cohort mortality study conducted of employees at the Antwerp plant. The purpose of this report is to provide an aggregate analysis of the hematological, clinical chemistry and hormonal parameters, as measured in the medical surveillance examinations of Antwerp and Decatur employees in two separate time periods, in relation to the workers' serum PFOS levels. Although female employees also participated in these medical surveillance examinations, their actual numbers were too few to provide meaningful statistical analysis. METHODS PFOS Production Page 13 001083 PFOS production began in Antwerp in 1976 and in Decatur in 1961. In general, perfluorinated chemicals are produced via an electrochemical process: a solution of organic substrate is electrolyzed in anhydrous hydrogen fluoride at a low voltage (Simons and Bryce, 1954). Basically, the products of this electrolysis cell reaction are highly fluorinated compounds with the end-product defined by the starting material. Products manufactured at these two plants include: ScotchgardTM brand fiber, leather and carpet protector; Light WaterTM brand aqueous film forming foam (AFFF); ScotchbanTM paper th m __ _ TTL/f treatment; Kel-F brand plastic and Fluorel brand elastomers. Subject Selection General medical surveillance occurs biennially for employees at both of these plants. Participation is voluntary with approximately 100 Antwerp and 250 Decatur employees eligible for surveillance. A total of 88 Antwerp employees participated in the medical surveillance examinations in the Spring, 1995 and 90 Decatur employees participated in the Fall, 1994. In the Fall of 1997, a total of 149 employees (Antwerp = 65; Decatur = 84) participated in medical surveillance examinations. For purposes of brevity, these time periods will be referred to as 1995 and 1997. Altogether, 61 employees participated in both examination years (1995 and 1997). This lower number was due to a large turnover of employees at both plant locations during 1996-1997. For each time period the surveillance consisted of a medical questionnaire, measurement of height, weight and blood pressure, standard clinical chemistry and hematology tests, and determination of serum PFOS levels. Page 14 0010S4 In 1995, several hormones were also analyzed for male employees who were judged a priori to have had likely PFOS exposure (i.e., those working in or in the immediate vicinity of the PFOS production area). Of the 88 Antwerp employees, 50 had hormone measurements. Of the 90 Decatur employees, 38 underwent hormone measurements. PFOS Analysis In 1995 the analysis for serum PFOS was conducted by 3M'S Environmental Technology Services in St. Paul, Minnesota. The method used tetrabutylammonium to ion-pair with PFOS in serum. The ion-pairs were then extracted with ethyl acetate. The abstraction product was then analyzed using high performance liquid chromatographythermospray mass spectrometry [Johnson et al., 1996]. In 1997 the serum samples were analyzed by TurboIonSpray liquid chromatography/mass spectrometry using selected ion monitoring in the negative ion mode by Advanced Bioanalytical Services, Inc. [Anderson et al., 1997a; 1997b]. The lower limit of quantitation was 0.1 pg/mL for PFOS. Laboratory Analyses For both time periods, the United Laboratory Services (St. Paul, Minnesota) performed the standard hematological and clinical chemistry tests. These included the following hematological tests: hematocrit (percent), hemoglobin (gm/dl), red blood cells (RBC, 1000/mm3), white blood cells (WBC, 1000/ mm3) and platelet count (1000/ mm3); and the following clinical chemistry tests: alkaline phosphatase (IU/L), gamma glutamyl transferase (GGT, IU/L), aspartate aminotransferase (AST, IU/L) formerly Page 15 0010S5 known as serum glutamic oxaloacetic acid (SGOT), alanine aminotransferase (ALT, IU/L) formerly known as serum glutamic oxaloacetic transaminase (SGPT), total and direct bilirubin (mg/dl), blood urea nitrogen (BUN, mg/dl), serum creatinine (mg/dl), glucose (mg/dl), cholesterol (mg/dl), high density cholesterol (HDL, mg/dl) and triglycerides (mg/dl). Low density lipoprotein (mg/dl) was calculated as the following: LDL = [cholesterol - HDL - (tryglycerides/5)]. Eleven hormones were assayed in 1995: cortisol, dihydroepiandrosterone sulfate (DHEAS), estradiol, follicle stimulating hormone (FSH), 17 alpha hydroxyprogesterone (17-HP), free testosterone, total testosterone, luteinizing hormone (LH), prolactin, thyroid stimulating hormone (TSH) and sex hormone binding globulin (SHBG). All but SHBG (Endocrine Science Reference Laboratory, Tarzana, CA) were analyzed at the University of Minnesota's Endocrinology Laboratory. Cortisol was assayed using a fluorescence polarization immunoassay (Abbott TDx). Radioimmunoassays (RIA) were used for DHEAS (Pantex), estradiol (modified Pantex), 17-HP (modified CIS) and total testosterone (Diagnostic Product Corp. Coat-A Count). Free testosterone was determined using equilibrium dialysis. LH, FSH and prolactin were assayed using a microparticle enzyme immunoassay (Abbott Imx). TSH was determined using a chemiluminescence immunometric assay (Nichols). SHBG was assessed via a radioimmunoassay after chromatographic sample purification (Endocrine Science Reference Laboratory). Bound testosterone was calculated as total testosterone less free testosterone. Data Analysis Page 16 0010S6 Descriptive simple and stratified analyses, Pearson correlation coefficients, ANOVA and ordinary multivariable regression were used to evaluate associations between PFOS and each hematological and clinical chemistry test and hormonal assay. Age, body mass index, current alcohol consumption (drinks per day) and cigarette use (cigarettes smoked per day) were potential confounding factors that were considered in the analyses. For stratified analyses, employees were divided into four PFOS categories: 0-1 ppm, 1 - < 3 ppm, 3 - < 6 ppm and > 6 ppm in order to determine if an effect existed at the highest serum PFOS levels. Other categorical cutoff points were also used which provided similar results. For multivariable regression analyses, PFOS and the potential confounders of age, body mass index (BMI), alcohol use and cigarette use were examined as continuous explanatory variables in the models. Multivariable regression models were fitted with PFOS analyzed as a continuous variable using linear as well as non-linear transformations (quadratic, square, square root and inverse) in order to maximize the possibility of finding associations between PFOS and the dependent variable of interest. Linear and nonlinear relationships were examined by residual diagnostics using studentized and Cook's distance values. Natural log transformations of the dependent variables were performed, when necessary, to normalize variables and to enhance model fit. Traditional stepwise selection procedures were also employed (selection in and out of model was set at p = 0.1) as well as taking into account other covariants that may be on the biologic pathway of effect [Greenland, 1989]. We did not examine changes in measured PFOS between the two time periods because the estimated half-life of PFOS is at least two years. Study results were analyzed using the SAS System [1990]. Page 17 001087 Analyses are presented by plant, year and the three major groups of participants: all employees who participated in each year (1995: N = 178; 1997: N = 149); only those employees (N = 61) who participated in both years; and only those employees (N = 88) who participated in the hormone measurements in 1995. RESULTS The distribution of employees, by serum PFOS exposure categorization, is presented in Table 1. Whereas 20 percent of the Decatur employees had exposures at > 3 ppm for both years, this proportion in Antwerp went from 25 percent in 1995 to 13 percent in 1997. For both years 95 percent of the measured serum PFOS levels were below 6 ppm. There were no PFOS measurements > 6 ppm in Antwerp in 1997. The overall mean values of PFOS, demographic, serum chemistry and hematological parameters for both locations, as well as each location separately, are presented in Tables 2 and 3, respectively. In particular, the Antwerp male employee population was significantly younger than Decatur, had lower body mass indices and higher self-reported daily consumption of alcohol. In addition, their clinical profiles were also different for several tests. The Antwerp employees had lower mean alkaline phosphatase, creatinine, glucose and triglyceride values and higher total bilirubin, HDL and hematocrit values. Presented in Table 4 are the Pearson correlation coefficients between PFOS and the selected parameters of interest by both locations combined, each location separately, and by year of examination. In 1995, variables that were significantly (p < .05) correlated with PFOS for both locations combined included total bilirubin, white blood Page 18 001088 cells and platelets. Although creatinine was not significantly correlated when both locations were examined, it was negatively correlated with PFOS in Antwerp but positively correlated in Decatur. In 1997, variables that were significantly correlated with PFOS for both locations combined were BMI, ALT, direct bilirubin, cholesterol, LDL and hematocrit. In addition, GGT and triglycerides were significantly positively correlated with PFOS among only Antwerp employees. Provided in Table 5 are the mean, median, standard deviation and range of the covariates and the clinical chemistries and hematological parameters by four levels of PFOS categorization (0 - < 1, 1 - < 3, 3 < 6 and > 6 ppm) for both years. Several observations are noteworthy. First, the mean for the > 6 ppm PFOS category was one order of magnitude higher than the lowest PFOS category (0 - < 1 ppm) for both years. Also, the means of the four PFOS categories were significantly different from each other. Second, the youngest employees had the lowest serum levels of PFOS. Third, there was only one variable, total bilirubin, which had significant (p < .05) F tests for differences in means in both years of analysis. Besides total bilirubin, the only other variable in which the mean of the higher levels of PFOS exposure (3 - < 6 ppm or > 6 ppm) was significantly different from the lowest category level of PFOS exposure (0 - < 1 ppm) was for WBC's in 1995. This was not observed in 1997. The lowest mean platelet count was observed at the highest PFOS exposure category in both years although the mean platelet counts by PFOS categories were not significantly different from each other. Provided in the next two tables are the mean, median, standard deviation and range of the covariates and the clinical chemistries and hematological parameters by the four levels of PFOS categorization for each plant for 1995 (Table 6) and 1997 (Table 7). Page 19 001089 In 1995 (Table 6) in Antwerp only, alcohol consumption was associated with higher PFOS levels. Mean serum creatinine levels declined in Antwerp but increased in Decatur employees. Antwerp employees in the 3 - < 6 ppm PFOS category smoked more cigarettes and had higher WBC levels. In 1997, Antwerp employees in the lowest PFOS exposure category were significantly younger than their counterparts. Antwerp employees in the higher PFOS category levels had higher mean alcohol consumption levels. Mean cholesterol, LDL and triglyceride levels trended upwards by PFOS exposure categories for Antwerp employees. Linear and nonlinear relationships between PFOS and the dependent variables of interest, taking into account the potential confounding affects of age, BMI, alcohol and cigarettes, resulted in numerous analyses. For purposes of brevity, linear regression models are presented in Table 8 which show the effect that the parameter of interest, PFOS, has on the various dependent variables, adjusted for age, body mass index, alcohol and cigarette use. These covariates were analyzed as continuous variables. In the case of serum creatinine and total bilirubin, a quadratic (PFOS + PFOS2) analysis provided the best statistical model of the data adjusted for the four potential confounders. The natural log transformation of total bilirubin, GGT and glucose provided the best fit for these response variables. PFOS was significantly associated ( p < .10) in both years for only one clinical parameter: total bilirubin. PFOS was associated in one of the two years for the following variables: direct bilirubin, creatinine, cholesterol, LDL, HDL, hematocrit, hemoglobin and platelet count. Those variables that were observed to be associated in at least one year in the regression models in Table 8 are separated by plant location and year in Table 9. After Page 20 001090 separate analyses by employee population, only two variables, total bilirubin and HDL, remained significantly (negatively) associated with PFOS for at least one plant location for both time periods. Total bilirubin showed a significant negative association with PFOS (quadratic relation) for employees at the Decatur plant in both years. There were no significant associations among the Antwerp population between PFOS and total bilirubin. HDL was significantly negatively associated with PFOS in Antwerp in both 1995 and 1997 but was not significantly associated with PFOS in Decatur in either year. As for inconsistent associations observed in Table 8, direct bilirubin was not significantly associated with PFOS in either plant location (Table 9). The quadratic association for PFOS with creatinine was observed in Antwerp in 1995 and Decatur in 1997 but not in Antwerp in 1997 or Decatur in 1995. Cholesterol (and LDL) was observed to be positively associated with PFOS only in Decatur in 1997. Hematocrit and hemoglobin were associated with PFOS only in Decatur in 1997. Platelet counts were observed to be significantly negatively associated with PFOS only in Decatur in 1995. Traditional stepwise regression modeling techniques were also used as well as testing models with variables that would be considered on the biological pathway of effect for any dependent variable. The associations (or lack thereof) from these analyses were similar to what has been presented in Tables 8 and 9. For purposes of brevity these analyses are not shown. To further understand the association between total bilirubin and PFOS, scatter plots are presented for both time periods and by location in Appendix A. In addition, unconjugated bilirubin was also calculated (total bilirubin - direct) and these scatter plots are presented in Appendix B. Table 10 is a summary of these scatter plots from both Page 21 001091 Appendices. The strongest associations appeared to be quadratic in nature primarily for the Decatur location and the percent of variability explained ranged between 3 (1995 data) and 7 percent (1997 data). The linear component of the quadratic was negative in direction. The upward trend appeared to occur around 6 ppm PFOS where the data are sparse. These simple linear and quadratic models were not influenced by any one employee according to residual diagnostics. To further understand the possible association between HDL and PFOS, scatter plots are presented for both times and by location in Appendix C. Both locations combined resulted in significant negative linear and nonlinear (quadratic) associations in 1995 although the percent of variability explained in these models ranged between 3 and 5 percent. No significant associations were observed for each plant location in 1995. There were no significant negative associations between HDL and PFOS in 1997 for either the combined locations or each separate plant site. Provided in Tables 11 through 14 are the analyses restricted to the 61 employees who participated in surveillance in both years. Table 11 provides the mean values for each parameter for the employees who participated in both exams compared to those who participated in only one of the two years. Overall, there were few differences. The mean age of the 61 employees was lower than that of the 1995 employees who didn't participate in 1997. Conversely, the mean age of the 61 employees was higher than that of the 1997 employees who didn't participate in 1995. Cholesterol and LDL were significantly higher in the 61 participants in 1997. Tables 12 and 13 present the mean values by plant location for 1995 and 1997, respectively. Of these 61 employees, 27 were from Antwerp and 34 from Decatur. Of noteworthy importance are the differences Page 22 001092 between the 27 Antwerp employees and their fellow employees in 1997 (Table 13). The 27 Antwerp employees had significantly higher mean PFOS exposures, were significantly older, had greater BMI's and higher cholesterol values. Multivariable regression analyses for the 61 employees are presented in Table 14. The only significant association with PFOS appeared to be with serum creatinine (quadratic) in 1995. Regardless of plant location, mean PFOS levels were higher for those employees who were selected for hormone measurements in 1995 (Table 15). This was expected as these employees were selected with the a priori belief that their serum measurements would be higher due to their workplace experience. For example, of the 42 employees in 1995 whose serum PFOS levels were > 3 ppm, 76% had hormone measurements. Presented in Table 16 are the mean values for PFOS, demographic, serum chemistries and hematology for those employees who had hormone measurements compared to those employees who did not in 1995. Those employees who had hormone measurements were younger, higher users of alcohol (Antwerp only) and cigarettes (both locations), and had lower serum creatinine (Antwerp only) and higher WBC levels (both locations). The latter observation is confounded by cigarette smoking as among nonsmokers, those selected for hormone measurements had a mean WBC of 6.24 compared to 6.03 for non-selected employees (p = .36). Among smokers, those selected had a mean WBC of 8.69 compared to 8.06 for non-selected employees (p = .23). All other clinical parameters were comparable, by PFOS exposure categories, between subjects who had hormone measurements and those who did not in 1995 (Tables 17 and 18). The Pearson correlation coefficients between PFOS and the hormones tested among the 88 employees were the following: cortisol (.07), DHEAS (-.13), estradiol Page 23 001093 (.09), FSH (.06), 17-hydroxyprogesterone (-.04), LH (.03), prolactin (.06), SHBG (.11), free testosterone (-.06), bound testosterone (.06), TSH (.01). None were statistically significant. Presented in Table 19 are the mean, median, standard deviation and range of the various hormones by the four PFOS categories: 0- <1 ppm, 1 - < 3 ppm, 3 - < 6 ppm and > 6 ppm. Several observations are noteworthy. First, the mean age of the lowest PFOS exposure category was 10 years less than that of the highest exposure category. Therefore it was not unexpected to observe that the mean DHEAS, 17-HP, free testosterone and bound testosterone levels of this lowest exposure category were greater than the means of the higher PFOS exposure categorizations. Adjusting for the differences in age (as well as the other three potential confounders) in the regression models (Table 20) resulted in no significant associations between PFOS and the hormones analyzed, except for estradiol. With estradiol, a quadratic model provided the best fit of the data and both PFOS terms were significant. Upon residual diagnostics it was determined that this estradiol model was influenced by one specific employee (employee A). The influence of employee A is best seen in Figures 1 and 2 which are simple scatter plots of both the linear and quadratic fits of estradiol and PFOS, with and without employee A, respectively. Employee A had a 12.83 ppm serum level of PFOS which was the highest value recorded in 1995. His estradiol value was 92 pg/dl (see upper right hand comer of Figure 1). Employee's A estradiol value was also influenced by the fact that his body mass index was 33 kg/m2. Exclusion of this employee resulted in a nonsignificant quadratic equation. The variability (R2) of the data explained went from 7.6 percent to 2.1 percent upon exclusion of this employee. The slope of the linear equation changed from Page 24 001094 positive to negative although it was nonsignificant in both Figures 1 and 2. Finally, it should be noted that the estradiol models in Table 20, with and without employee A, did predict the known positive assocation between estradiol and body mass index. DISCUSSION We conducted two cross-sectional analyses of surveillance data to examine the associations between serum PFOS levels and several hematological, clinical chemistry and hormonal parameters in male fluorochemical production employees. For both years, 95 percent of the measured serum PFOS levels were below 6 ppm. Because the Antwerp and Decatur employees were dissimilar by age, body mass indices and self-reported alcohol use, we conducted combined as well as separate analyses by plant location. These three demographic differences likely explain why the Antwerp employees had lower mean serum levels of alkaline phosphatase, HDL, triglycerides and blood glucose [Davem and Scharschmidt, 1993; Lewis, 1994; Friedman, 1998; Fu, 1998; Wolf, 1998], In the present study, alkaline phosphatase, GGT, AST and ALT values were not significantly associated with the measured serum PFOS levels. This was an a priori question due to the fact that PFOS: 1) is a peroxisome proliferator in the rat [Ikeda et al., 1987; Sohlenius et al., 1993]; 2) resulted in slight to marked increases in plasma glutamic oxalacetic and pryuvic transaminase levels in a 90 day study of rats fed diets which contained PFOS at 100 ppm along with hypertrophy and liver necrosis observed at histopathology [Goldenthal, 1978a]; and 3) increased SGOT and decreased alkaline phosphatase levels in monkeys after administration, by oral gavage, for 30 days of doses of 4.5 mg/kg/day of PFOS [Goldenthal, 1978b]. On the other hand, SGPT values Page 25 001095 remained constant and no histopathologic abnormalities were noted in the livers of these monkeys which died by the 7thweek of the study. No significant liver enzymatic or histopathology changes occurred in monkeys in the 0.5 and 1.5 mg/kg/day dose groups. We did observe a quadratic association with total bilirubin among only the Decatur employees. We do not suspect this is a biological association because the bilirubin levels were within the normal reference range. Also, the percent variability explained of total bilirubin by PFOS in the regression models was low. The Antwerp employees' total bilirubin levels were significantly higher than the Decatur employees'levels. We offer several possible explanations for this observation. First, we suspect there may be a greater prevalence of Gilbert's syndrome [Lidofsky and Scharschmidt, 1993; Friedman, 1998] among the Antwerp employees. In 1995, 15 (17%) Antwerp employees had total bilirubin values >1.2 mg/dl compared to 3 (3%) Decatur employees' levels. In 1997, there were 9 (15%) Antwerp and 2 (2%) Decatur employees with total bilirubin values >1.2 mg/dl. However, there was not a concomitant decline in bilirubin conjugation (i.e., direct bilirubin levels) as might be expected among individuals diagnosed with Gilbert's syndrome. Nevertheless, exclusion of these possible Gilbert's syndrome employees still resulted in higher mean total bilirubin values among the Antwerp employees in both years. Secondly, there were four Antwerp employees who self-reported hepatitis A histories and one employee selfreported a history of Hepatitis B. Fish and shellfish consumption is likely much greater in Antwerp than Decatur due to its vicinity near the North Atlantic. Third, bilirubin is a tetrapyrrole that is an end-product of heme degradation [Lidofsky and Scharschmidt, 1993]. Bilirubin levels may be increased due to disorders of bilirubin metabolism, liver Page 26 001096 disease and obstruction of the bile ducts. Other hematological and clinical chemistry results did not suggest these conditions existed among the Decatur employees. Fourth, post-collection procedures may result in error. Total bilirubin determination may be falsely depressed if hemolysis is present because of increased absorbence in the blank [Kaplan and Pesce, 1984]. Bilirubin is also sensitive to and destroyed by light and heat. We are uncertain whether these factors could have contributed to the lower total bilirubin levels in the Decatur samples in both years. Finally, the linear component of the quadratic association observed among Decatur employees is negative in direction in relation with their measured PFOS levels. That is, total bilirubin levels declined with increasing PFOS levels. We would expect a positive association if PFOS impaired bilirubin conjugation. The trend upwards in the quadratic appears to occur at levels 6 ppm and higher where the data are sparse. We conclude that the association observed among only the Decatur employees is unlikely to be related to serum levels of PFOS. We did observe a positive association between serum PFOS and serum cholesterol levels in the 1997 time period for Decatur employees. This result is unlikely to have a biological explanation as PFOS is a known peroxisome proliferator in the rat and was shown to have hypolipidemic properties in rhesus monkeys [Ikeda et al., 1987; Sohlenius et al., 1993; Goldenthal 1978b; 1979]. Rhesus monkeys fed PFOS at 4.5 mg/kg in their chow had serum cholesterol values reduced from 183 mg/L to 99 mg/L within 30 days. Rhesus monkeys fed 1.5 mg/kg in the chow had cholesterol levels reduced from 195 mg/kg to 111 mg/kg within 90 days [Goldenthal 1978b]. As for HDL, although the multivariable analyses were suggestive of a negative association between HDL and PFOS Page 27 001097 in Antwerp (but not Decatur), the scatter plots presented in Appendix C do not support the notion of a biological association between PFOS and HDL. It should be noted that total organic fluorine levels, primarily consisting of perfluorooctanoic acid (PFOA, C7F15COO ), a seven carbon perfluorinated carboxylic acid, were reported to reduce the effect that alcohol has on HDL levels among higher exposed male PFOA production workers in Cottage Grove, Minnesota [Gilliland and Mandel, 1995]. However, this finding was not observed in subsequent analyses of these employees (Olsen et al., unpublished findings). This observation by Gilliland and Mandel was testable in the present study as both Antwerp and Decatur employees had measurable quantities of PFOA. We did not observe a significant negative modulation of the effect of alcohol consumption on HDL levels among Antwerp and Decatur employees with higher serum PFOA levels although their serum levels were approximately 3 to 5fold less, on average, than that reported in Cottage Grove employees [Olsen et al., 1998]. The Antwerp and Decatur employees were exposed to PFOA, not in its actual production, but rather in its use as a surfactant in the production of fluoropolymers. In 1995 the mean serum PFOA level among the Antwerp and Decatur employees combined was 1.46 ppm (range 0 -13.20 ppm) and in 1997 the mean serum level was 1.57 ppm (range 0.11 11.10). Stratified by plant location, the 1995 and 1997 mean serum PFOA levels were 1.19 and 1.78 ppm in the Antwerp employees and 1.72 and 1.40 ppm in the Decatur employees, respectively. The multivariable regression models showed a negative association between PFOS and platelet counts at PFOS levels above 6 ppm in 1995 and this trend was also apparent, although to a lesser extent, in 1997. Nevertheless, platelet levels were well Page 28 001098 within the normal reference range in both time periods. This association is not supported by a 90 day subchronic toxicity study which showed no decline in platelet counts for monkeys fed 0.5, 1.5 or 4.5 mg/kg for up to 90 days [Goldenthal, 1978b]. Mean platelet counts among the 1.5 mg/kg/day and 0.5 mg/kg/day dose groups were 226 and 231 (103/cmm) compared to 218 in the control group [Goldenthal, 1978b]. There were no platelet counts reported in the 90 day rat study although at the end of 3 months of study there were slight to moderate decreases in hemoglobin, hematocrit and erythrocyte counts observed for male and female rats in the 100 ppm dose group [Goldenthal, 1978a], No consistent associations were observed between PFOS and hemoglobin, hematocrit or RBC values in the present epidemiologic investigation. In a prior subchronic rhesus monkey study that was aborted early due to all animals died by the 20thday, mean platelet counts were 203, 219, 136, 172 and 185 for the 300 mg/kg/day, 100 mg/kg/day, 30 mg/kg/day, 10 mg/kg/day and control groups, respectively [Goldenthal, 1979]. After controlling for age, a confounder for male testosterone hormone levels [Dali et al., 1981; Griffin and Wilson, 1994], we observed no significant associations with serum PFOS measurements. We did observe a quadratic association between estradiol and PFOS. Upon further examination, this finding was influenced by one particular employee who had the highest PFOS level but was confounded by the individual's large body mass index. Exclusion of this employee resulted in nonsignificant findings. Thus, any interpretation with estradiol is difficult because of the influence this one employee has on the statistical analyses. It should be noted that perfluorooctanoic acid (PFOA), at approximately 50 - 100 ppm levels in serum, enhances the aromatase conversion of testosterone to estradiol in the Page 29 001099 rat [Cook et al., 1992; Biegel et al., 1995]. However, PFOA production workers in Cottage Grove with serum levels up to 30 ppm appeared not to have altered serum estradiol levels [Olsen et al., 1998]. Again, like HDL, this was a testable hypothesis among the Antwerp and Decatur employees although their serum PFOA levels were lower than Cottage Grove employees. We did not observe any significant positive association between estradiol and serum PFOA levels in these Antwerp and Decatur employees. Several methodological issues should be considered in evaluating the results from this study. First, the cross-sectional design does not allow for a direct analysis of the temporality of an association. Second, the voluntary participation rates in medical surveillance were not ideal as among eligible employees we had 88 and 65 percent participation in Antwerp for 1995 and 1997, respectively, but only 35 to 40 percent in Decatur for both years. Third, given the suspected long half-life of PFOS (at least two years), it may be conceivable that there may be some biological accommodation to the effects of PFOS which would minimize the possibility of finding an association. Fourth, it is known in laboratory animals that PFOS concentrates primarily in the liver. Serum measurements of PFOS may not adequately reflect body burden. Fifth, the two crosssectional analyses cannot be viewed as independent populations as 61 employees were studied in both years. This was due, in part, to a large turnover of employees at both plants between examinations. Sixth, there could be measurement error in important confounding variables. Analysis of the data of the 61 subjects who participated in both years showed that there was excellent correlation for the confounding factors of BMI (r = .92, p = .0001), self-reported aspects of alcohol consumption (r = .88, p = .0001) and Page 30 O O llO O cigarette smoking (r = .79, p = .0001). As expected, these 61 employees' serum PFOS levels for the two years were highly correlated (r = .92, p = .0001). Seventh, the quality of medical surveillance data, prior to its use for studying an a priori hypothesis, can often be evaluated by whether known positive associations are observed. In this regard, we observed various expected associations including cigarette smoking and elevated white blood cell counts and large body mass indices associated with elevated liver transaminase levels [Olsen et al., 1991; Bums et al., 1997]. Finally, the pulsatile nature of some of the hormones studied (e.g., FSH, LH, testosterone) has resulted in prior recommendations that mean hormone measurements should be the result of pooled blood from multiple samples taken at short intervals [Goldzieher et al., 1976]. In our study multiple samples were not feasible because of the low probability of employees voluntarily giving three serum samples over a 45 -60 minute period of time. In summary, we conducted two cross-sectional analyses and did not observe consistent associations by plant location or time for several hematological parameters, serum chemistries and reproductive hormones with measured serum PFOS levels in male fluorochemical production employees. Ninety-five percent of the employees had serum PFOS levels below 6 ppm. Our findings suggest that, among these Antwerp and Decatur male fluorochemical production employees, significant hematological, clinical chemistry and hormonal abnormalities were not associated with serum PFOS concentrations less than 6 ppm. Any inferences derived from the few employees with serum PFOS levels > 6 ppm would be tenuous, at best. Limitations of this study include its cross-sectional design, the voluntary participation rates, the few subjects exposed at the highest levels, Page 31 001101 and the lower levels of serum PFOS measured among these employees compared to those that caused effects in two species of laboratory animals. 001102 Acknowledgements The authors gratefully acknowledge the assistance of Michele Burlew, Martha McGough, Jane Quarfoth and the helpful comments of Drs. John Butenhoff, Marvin Case and Andrew Seacat. Page 33 001103 References Anderson DJ, Mulvana DE (1997a). Analytical report for the determination of perfluorooctanoate and perfluorooctanesulfonate in human serum by LC/MS. Ithaca:Advanced Bioanalytical Services, Inc., August 25, 1997. Anderson DJ, Mulvana DE (1997b). Analytical report for the determination of perfluorooctanoate and perfluorooctanesulfonate in human serum by LC/MS. Ithaca:Advanced Bioanalytical Services, Inc., September 22, 1997. Biegel LB, Liu RCM, Hurtt ME, Cook JC (1995). Effects of ammonium perfluorooctanoate on Leydig cell function: in vitro, in vivo, and ex vivo studies. Toxicol Appl Pharmacol 134:18-25. Bums CJ, Boswell JM, Olsen GW (1997). Liver enzyme activity and body mass index. J Occup Env Med 38:1248-1252. Central Analytical Laboratory (1979). Characterization of Decatur RF Values. 3M Project Number 91721100, Report No. 7230, October 4, 1979. Cook JC, Murray SM, Frame SR, Hurtt ME. Induction of Leydig cell adenomas by ammonium perfluorooctanoate: a possible endocrine-related mechanism. Toxicol Appl Pharmacol 113:209-217. Davem TJ, Scharschmidt BF (1993). (In) Feldman M, Sleisenger MH, Scharschmidt BF (eds), Gastrointestinal and Liver Disease, 6thEdition. Philadelphia:W.B. Saunders Co., pages 1112 - 1122. Dali WS, Kuller LH, LaPorte RE, Gutai JP, Falvo-Gerard L, Caggiula A (1981). The epidemiology of plasma testosterone levels in middle-aged men. Am J Epidemiol 114:804-816. Friedman LS (1998). Liver, biliary tract and pancreas. (In) Tierney LM, McPhee SJ, Papadakis WA (eds), Medical Diagnosis and Treatment, 37thEdition. Stamford (CT):Appleton and Lange pages 628-630. Fu, PC (1991). Lipids, lipoproteins, and apolipoproteins. (In) Howanitz JH, Howanitz PJ (eds), Laboratory medicine: test selection and interpretation. New York:Churchill Livingstone Inc., pages 173-190. Gabriel KL (1976). Acute oral toxicity in rats of T-1389. Biosearch, Inc., Philadelphia, PA. Page 34 001104 Gibson SJ, Johnson JD, Ober RE (1983). Absorption and biotransformation of N-ethyl FOSE and tissue distribution and elimination of carbon-14 after administration of N-ethyl FOSE-14C in feed. St. PauFRiker Laboratories, Inc. Gilliland FD, Mandel JS (1996). Serum perfluorooctanoic acid and hepatic enzymes, lipoproteins, and cholesterol: a study of occupationally exposed men. A J Ind Med 29:560-568. Glaza SM (1995). Single-dose dermal absorption/toxicity study of T-6049 in rabbits. HWI 6329-130, Hazelton Laboratories, Madison, WI. Goldenthal El, Jessup DC, Geil RG, Jefferson ND, Arceo RJ (1978a). Ninety day subacute rat toxicity study. Study No. 137-085, International Research Development Corp., Mattawan, MI. Goldenthal El, Jessup DC, Geil RG, Jefferson ND (1978b). Ninety-day subacute rhesus monkey toxicity study. Study No 137-085, International Research Development Corp., Mattawan, MI. Goldenthal El, Jessup DC, Geil RG, Jefferson ND (1979). 90-day subacute rhesus monkey toxicity study. Study No. 137-087, International Research Development Corp., Mattawan, MI. Goldzieher JW, Dozier TS, Smith KD, Steinberger E. Improving the diagnostic reliability of rapidly fluctuating plasma hormone levels by optimized multiple-sampling techniques. J Clin Endocrinol Metab 43:824-830. Gortner EG, Lamprecht EG, Case MT (1980). Oral teratology study of FC-95 in rats. Expt No. 060TR0008, Riker Laboratories, Inc., St. Paul, MN. Greenland, S (1989). Modeling and variable selection in epidemiologic analysis. Am J Public Health 79:340-349. Griffin JE, Wilson JD (1994). Disorders of the testes. (In) Isselbacker KJ, Braunwald E, Wilson JD, Martini JB, Fauci AS, Karper DL (eds): ) Harrison's Principles of Internal Medicine. New York:McGraw-Hill, 1994:2006-2017. Haughom B, Spydevold O (1992). The mechanism underlying the hypolipmie effect of perfluorooctanoic acid (PFOA), perfluoroctane sulphonic acid (PFOSA) and clofibric acid. Biochemica et Biophysica Acta 1128:65-72. Ikeda T, Fukuda K, Mori I, Enomoto M, Komai T, Suga T (1987). Induction of cytochrome P-450 and peroxisome proliferation in rat liver by perfluorinated octanesulfonic acid. In: Peroxisomes in Biology and Medicine, Fahimi HD, Sies H (eds), Springer Verlag, New York, pages 304-308. Page 35 001105 Jagannath DR, Brusic D (1978). Mutagenicity evaluation of T-2014CoC in the Ames Salmonella/microsome plate test. Litton Bionetics, Inc., Kensington, MD. Johnson JD, Gibson SJ, Ober RE (1979). Extent and route of excretion and tissue distribution of total carbon-14 in rats after a single i.v. dose of FC-95-14C. Project No. 8900310200, Riker Laboratories, Inc., St. Paul, MN. Johnson JD, Ober RE (1979). Absorption of FC-95-14C in rats after a single oral dose. Project No. 8900310200, Riker Laboratories, Inc., St. Paul, MN. Johnson JD, Gibson SJ, Ober RE (1980). Enhanced elimination of FC-95-l4C and FC143-14C in rats with cholestyramine treatment. Project No. 8900310200, Riker Laboratories, Inc., St. Paul, MN. Johnson JD, Gibson SJ, Ober RE (1984). Cholestyramine-enhanced fecal elimination of carbon-14 in rats after administration of ammonium [14C] perfluorooctanoate or potassium [14C] perfluorooctanesulfonate. Fund Appl Toxicol 4:972-976. Johnson JD, Wolter JT, Colaizy GE, Rethwill PA, Nelson RM (1996). Quantification of perilurooctanoate and perfluorooctanesulfonate in human serum using ion-pair extraction and high performance liquid chromatography - thermospray mass spectrometry with automated sample preparation. J Chromatography Biomed Appl (submitted for publication). Kaplan and Pesce (1984). Clincal Chemistry. New YorkrC.V. Mosby Co., page 1238. Lewis JA (1994). Illustrated Guide to Diagnostic Tests. Springhouse (PA): Springhouse Corp., pages 100 - 107, 194-207. Lidofsky S, Scharschmidt BF (1993). (In) Feldman M, Sleisenger MH, Scharschmidt BF (eds), Gastrointestinal and Liver Disease, 6th Edition. Philadelphia:W.B. Saunders Co., pages 220-224. Mandel JS, Johnson RA (1995). Mortality study of employees at 3M plant in Decatur, Alabama. Division of Environmental and Occupational Health, University of Minnesota, March 13, 1995. Murli H (1996). Mutagenicity test on T-6295 in an in vivo mouse micronucleus assay. CHV study No. 17403-0-455, Coming Hazelton Inc., (CHV), Vienna, VA. Nabbefeld J, Butenhoff, J, Bass N, Seacat N (1998). Displacement of a fluorescently labeled fatty acid analogue from fatty acid carrier proteins by wyeth-14,643, ammonium perfluorooctanoate, potassium perfluorooctane sulfonate and other known peroxisome proliferators. Toxicologist (in press). Page 36 001106 Olsen GW, Kusch GD, Stafford BA, Gudmundsen SL, Currier MF (1991). The positive known association design: a quality assurance method for occupational health surveillance data. J Occup Med 33:998-1000. Olsen GW, Gilliland FD, Burlew MM, Burris JM, Mandel JS, Mandel JH (1998). An epidemiologic investigation of reproductive hormones in men with occupational exposure to perfluorooctanoic acid. J Occup Env Med (in press). O'Malley KD, Ebbens KL (1980). Twenty-eight day percutaneous absorption study with FC-95 in albino rabbits. Expt No. 0979AB0632, Riker Laboratories, Inc., St. Paul, MN. Roach DE (1982). Fluorochemical Control Study. 3M unpublished report. May 25, 1982. Rusch GM, Rinehart WE (1979). An acute inhalation toxicity study of T-2306CoC in the rat. Bio/dynamics, Inc. SAS Institute, Inc. (1990). SAS Users Guide: Statistics. Version 6, SAS Institute, Inc., Cary, NC. Schuman LM (1982). Letter to Dr. Frank Ubel (3M Medical Director) concerning Fluorochemical Control Study. April 29,1982. Sohlenius AK, Eriksson AM, Hogstrom C, Kimland M, DePierre JW (1993). Periluorooctane sulfonic acid is a potent inducer of peroxisomal fatty acid beta-oxidation and other activities known to be affected by peroxisome proliferators in mouse liver. Pharmacol Toxicol 72:90-93. Wetzel LT, Burdock GA, Durloo RS, Colpean BR (1983). Rat teratology study. Project No. 154-160, Hazelton Laboratories America, Inc., Vienna, VA. Wolf PL (1991). Liver function. (In) Howanitz JH, Howanitz PJ (eds), Laboratory medicine: test selection and interpretation. New York:Churchill Livingstone Inc., pages 173-190. Page 37 001107 Table 1 Distribution of Employees by Year, Location and PFOS Exposure Level (ppm) All Employees PFOS N (%)_________ N 0 - < 1 ppm 45 25 1 - < 3 ppm 91 51 3 - < 6 ppm 35 20 >= 6 nom 74 178 (100) 1995 Data Antwerp (%)_______ 34 39 32 36 19 22 33 88 (100) N Decatur (%) N 11 12 59 66 16 18 44 90 (100) All Employees (%) N 60 40 63 43 21 14 53 149 (100) 1997 Data Antwerp (%) Nr 31 48 25 38 9 14 00 65 (100) Decatur (%) 29 35 38 45 12 14 56 84 (100) 001108 Table 2 Mean, Median, Standard Deviation and Range of Study Parameters, Both Locations, 1995 and 1997 Variable PFOS Age A lcoh ol BMI Cigarettes Aik Phos GGT AST ALT T. Bilirubin D. Bilirubin Mean 2.19 40.6 0.7 26.6 6.3 86.2 44.4 27.6 45.9 0.72 0.21 1995 Data____________ M edian 1.70 40.0 0.1 25.8 0.0 85.0 36.0 26.0 42.0 0.70 0.20 Std D ev 1.87 8.76 1.1 5.1 10.7 24.6 32.1 10.9 18.6 0.37 0.05 Ranee 0 -1 2 .8 3 2 1 -6 0 0 -6 .0 17.9 - 60.7 0 -4 0 31 - 191 2 -2 9 3 1 3 -9 6 18-183 0.20 - 2.90 0 .1 0 -0 .4 0 __________________ 1997 Data Mean 1.75 39.3 0.5 27.1 6.1 79.8 31.6 26.2 32.6 0.67 0.13 M edian Std. D ev. 1.30 1.60 40.0 9.7 0.1 1.0 26.0 5.2 0.0 10.3 77.0 20.8 .24.0 25.9 25.0 7.0 30.0 14.4 0.60 0.32 0.10 0.06 Ranee 0 .1 0 -9 .9 3 2 1 -6 3 0 -7 .1 1 8 .1 -4 8 .5 0 -4 0 29 - 163 10-179 13-56 1 0 -8 9 0.30 - 2.30 0 .1 0 -0 .4 0 001109 001110 Variable BUN Creatinine G lu cose Cholesterol LDL HDL Triglycerides Hematocrit H em oglobin RBC MCH MCHC MCV WBC Platelets Mean 15.9 1.0 86.7 215.9 136.6 48.5 151.0 46.3 15.3 4.9 31.1 33.1 93.7 6.9 226.4 M edian 15.0 1.0 85.0 214.0 134.0 48.0 126.0 47.0 15.3 4.9 31.0 33.0 94.0 6.6 223.0 Std D ev 3.8 0.2 18.8 42.5 39.1 12.7 99.9 2.7 0.9 0.3 1.5 0.7 4.9 1.9 45.4 Ranee 8 -2 6 0 .6 - 1.6 60 - 260 100 - 340 28 - 261 2 3 -9 4 34-651 38.3 - 52.0 1 3 -1 7 .4 4.0 - 5.74 26.0 - 36.9 3 1 .3-34.7 7 9 .7-114.6 3 .6-15.5 122 - 367 Table 2 (Continued) M ean 14.4 0.9 90.7 211.2 135.3 45.5 156.5 45.8 15.4 5.1 30.5 33.6 90.7 6.5 225.7 Median Std. D ev. 14.0 3.0 0.9 0.1 86.0 29.9 213.0 40.7 135.0 35.0 45.0 10.6 122.0 130.0 46.0 - 2.6 15.5 0.9 4.8 0.3 31.0 1.5 34.0 0.5 91.0 .42 6.2 1.7 230.0 50.4 Ranee 6 -2 6 0 .6 - 1.4 58 - 303 110-365 50 - 290 19-74 38 - 1209 39 - 52.9 12.5 - 17.3 4.1 -5 .9 2 6 .2 -3 4 .1 3 1 .8 -34.9 8 0 .0 - 101.0 3 .8 -1 3 .2 106 - 406 Table 3 Mean Values of PFOS, Demographic, Serum Chemistry and Hematologic Parameters for Antwerp and Decatur, 1995 and 1997 Examinations Variable PFOS (ppm) Age BMI _____1995 Data Antwerp Decatur 1.93 2.44 37*** 45 23 9*** 29.2 _____ 1997 Data Antwerp Decatur 1.48 1.96 33*** 44 23.5*** 30.0 Cigarettes Alcohol Aik Phosphatase 4.7* ^ 3*** 75*** 7.9 0.2 97 5.5 l 1* * * 70*** 6.6 0.1 87 GGT 41 48 26* 36 AST 26* 29 27 26 ALT 44 47 31 34 Total bilirubin 0.86*** 0.58 0.80*** 0.58 Direct bilirubin BUN Creatinine 0.22 27 o*** Q9*** 0.21 14.8 1.1 0.15*** 14.9 0 9** 0.12 14.1 1.0 Glucose g l*** 92 g l*** 98 Cholesterol 214 218 206 215 LDL HDL Triglycerides 138 54*** 115*** 136 43 187 134 50*** 2j |*** 137 42 192 001111 Table 3 (continued) Variable Hematocrit 1995 Data Antwerp Decatur 47*** 46 Hemoglobin 15.4 15.2 RBC MCH 4.9 4*** 5.0 30.7 MCHC MVC WBC 32 9*** 95 7*** g 4*** 33.4 91.8 7.5 Platelets 224 229 * p < .0 5 ; * * p < .0 1 ; ***p < .001 _____ 1997 Data Antwerp Decatur 46* 45 15.4 15.3 5.1 5.1 30.5 30.4 33.3*** 33.7 91.6* 90.0 6.5 6.4 237* 217 001112 Table 4 Correlation Coefficients Between PFOS and Selected Variables by Location and Year of Examination Variable Age 1995 Data______________ Both Locations Antwerp Decatur .12 .10 .04 ______________ 1997 Data Both Locations Antwerp .20* Decatur -.01 Alcohol .09 .25* .06 .10 .40*** -.07 BMI .01 -.14 -.05 .18* .22 .09 Cigarettes .14 .19 .08 -.03 .08 -.08 BUN Creatinine -.05 -.04 .01 *** .05 -.37 .28** .01 .13 -.01 .08 -.12 .12 Glucose -.02 -.17 -.02 .10 .04 .06 Alkaline Phosphatase .06 .07 -.06 .07 .09 -.02 GGT .01 .02 -.04 .07 .26* -.06 AST O h* co .06 .06 .0002 .003 .03 .001 Variable ALT 1995 Data Both Locations Antwerp -.01 -.03 Total Bilirubin -.15* -.13 Direct Bilirubin -.03 -.13 Cholesterol -.02 .02 LDL .01 .02 HDL -.17* -.14 Triglycerides .04 .11 Hematocrit .002 .10 Hemoglobin -.01 .05 001114 Red Blood Cells MCH -.03 .04 -.11 * .21 Table 4 (continued) 1997 Data Decatur_____________ Both Locations Antwerp_____ Decatur -.03 .16* .11 .16 -.07 -.13 -.12 -.07 .09 -.18* -.13 -.17 -.08 .25** .34** .19 .02 .22** .30* .17 -.12 -.07 -.05 .0005 -.10 .10 .39*** -.01 -.0003 -.16* -.05 -.20 -.02 -.15 -.05 -.20 .01 -.11 -.15 -.10 -.04 -.04 .17 -.10 Variable MCHC MCV 1995 Data Both Locations Antwerp -.03 -.11 * .05 .25 White Blood Cells .18* .27 Platelets -.15* *** p < .05; p c . 01; p c . 001 -.06 Decatur -.09 -.01 .04 -.25* Table 4 (continued) -- 1997 Data Both Locations Antwerp .01 -.03 -.03 .20 .01 .18 -.11 -.06 Decatur -.06 -.09 -.09 -.10 Table 5 Mean, Median, Standard Deviation (SD) of Mean and Range of PFOS, Demographic, Serum Chemistries and Hematological for Antwerp and Decatur Employees Combined, 1995 (N=178) and 1997 (N=147) PFOS* (ppm) Mean 1995 Data Median SD Range Mean 1997 Data Median SD Range 0 - < lppm 1 - < 3ppm 3 - < 6ppm > 6ppm 0.491 0.5 1.82' 1.77 4.121 3.97 8.17' 7.73 F value = 321.9, p < .0001 0.27 0.58 0.81 2.52 PFOS (ppm) 000 - 0.90 1.00-2.91 3.00 - 5.80 6.06 -12.83 0.521 0.52 1.78' 1.64 3.87' 3.59 7.201 6.68 F value = 367.6, p < .0001 0.27 0.56 0.70 1.59 0.10-0.97 1.02-2.89 3.09 - 5.30 6.05 - 9.3 0 - < 1 ppm l - < 3 ppm 3 - < 6 ppm > 6 ppm 37 36 422 41 40 40 45 43 F value = 3.7, p = .02 Age (yrs) 8 21-58 36 34 11 9 25-60 422 41 9 7 26-55 41 42 5 7 37-56 42 45 9 F value = 5.1, p = .002 21-62 24-63 32-54 29-52 001116 PFOS (ppm) Mean 1995 Data Median SD 0 - < lppm 1 - < 3 ppm 3 - < 6 ppm > 6ppm 0.8 0.6 0.5 0.1 1.2 0.3 0.7 0.0 F value = 4.0, p = .009 0.9 0.7 1.9 1.1 0 - < 1 ppm 1 - < 3 ppm 3 - < 6 ppm > 6 ppm 25.5 24.8 27.7 26.3 24.93 25.0 27.7 29.4 F value = 3.7, p = .02 4.2 5.8 3.8 4.2 0 - < 1 ppm 1 - < 3 ppm 3 - < 6 ppm > 6 ppm 2.6 0.0 6.8 0.0 10.6 8.0 0.4 0.0 F value ==4.8, p = .003 6.4 11.3 12.4 1.1 Table 5 (continued) 1997 Data Range________ Mean_________ Median__________ SD__________ Range Alcohol (drinks/dav) 0.0 - 3.6 0.0 - 3.6 0.0 - 6.0 0.0 - 2.9 0.5 0.1 0.5 0.1 1.0 0.1 0.2 0.1 F value = 1.8, p = .15 0.8 0 -4 .3 0.8 0 -5 .0 1.8 0-7.1 0.2 0.1 -0.8 BMI (kg/m2) 17.9 - 38.7 19.6 - 60.7 17.9 - 32.5 20.6 - 33.0 26.0 24.9 27.7 26.4 27.3 27.9 30.8 29.7 F value = 2.1, p = .10 4.9 20.1-41.7 5.7 18.1-48.5 4.4 19.1-36.0 4.0 26.1 -36.2 Cigarettes (per day) 0.0 - 25.0 0.0 - 40.0 0.0 - 40.0 0.0 - 3.0 4.7 0.0 8.2 0.0 4.1 0.0 6.0 0.0 F value = 1.5, p = .23 9.4 0 -4 0 11.3 0 -4 0 8.3 0 -3 0 13.4 0 -3 0 PFOS (ppm) Mean 1995 Data Median SD 0 - < lppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 16.5 15 15.4 15.0 16.4 16.0 15.1 14.0 F value = 1.1, p = .36 3.5 4.0 3.7 4.2 0 - < lppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 1.0 1.0 1.0 1.0 0.9 0.9 1.1 1.2 F value = 2.3, p = .08 0.2 0.2 0.2 0.3 0 - < 1 ppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 85 85 86 86 84 83 84 83 F value = 0.9, p = .44 15 22 12 14 Table 5 (continued) 1997 Data Range Mean Median SD BUN 11 -26 8.0 - 26.0 10.0 - 23.0 10.0-21.0 14.5 14.0 14.2 14.0 15.0 15.0 13.8 12.0 F value = 0.5, p = 0.67 2.8 3.2 2.9 4.1 Creatinine 0.7-1.6 0.7-1.6 0.7- 1.2 0.6-1.6 0.9 0,9 0.9 0.9 0.9 0.9 1.0 0.9 F value = 0.4, p = 0.78 0.1 0.1 0.1 0.2 Glucose 66 - 170 60 - 260 66-114 71 -105 87 85 93 84 95 89 89 88 F value = 0.6, p = .59 17 39 25 7 Range 9.0-21.0 6.0 - 26.0 9.0 - 20.0 9.0 -19.0 0.6-1.2 0.7 - 1.3 0.7- 1.1 0.8-1.4 58 - 174 65 - 303 74 -192 80-97 1995 Data PFOS (ppm)________ Mean________ Median_________ SD 0 - < 1 ppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 80 78 89 89 86 85 88 85 F value = 1.3, p = .28 22 27 21 24 0 - < 1 ppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 43 31 47 36 40 39 43 33 F value = 0.5, p = .71 28 39 15 18 0 - < 1 ppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 27 25 29 26 25 24 33 33 PF valueu q I i i p -..11.88, p == .11A4 13 12 5 6 6TTT00 Table 5 (continued) 1997 Data Range________ Mean________ Median_________ SD_________ Range Alkaline Phosphatase 31 -158 49 -191 32-124 63 -136 77 73 83 79 76 74 88 84 F value = 1.17, p = .32 17 49 -132 23 41 - 163 22 29 - 120 18 65-114 GGT 16-155 2-293 21-80 28-79 28 22 36 25 28 27 33 37 F value = 1.1, p = .34 20 10 - 142 33 10- 179 14 13-71 12 17-48 AST 15-96 14-90 13-37 26-43 27 25 26 25 25 23 29 28 F value = 0.5, p = .67 7 13-53 7 15-56 7 14-43 3 26-34 1995 Data PFOS (ppm)________ Mean________ Median_________ SD Table 5 (continued) 1997 Data Range________ Mean________ Median_________ SD_________ Range 0 - < 1 ppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 48 43 46 42 42 41 51 49 F value = 1.0, p = .38 ALT 20 27-118 31 30 21 18-183 33 29 7 30-59 34 31 17 29-82 41 45 F value = 0.9, p = .46 11 13-60 16 10-89 18 14-82 10 25-49 0 - < 1 ppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 0.88 0.70 0.662 0.60 0.642 0.60 0.76 0.70 F value = 4.4, p = .005 Total Bilirubin 0.50 0.40 - 2.90 0.77 0.60 0.30 0.20-1.50 0.612 0.60 0.28 0.20-1.40 0.63 0.50 0.23 0.50-1.20 0.58 0.50 F value = 2.9, p = .04 0.40 0.30 - 2.30 0.21 0.30-1.30 0.31 0.40- 1.30 0.24 0.40- 1.00 0 - < 1 ppm 1 - <3 ppm 3 - < 6 ppm > 6 ppm 0.22 0.20 0.21 0.20 0.21 0.20 0.20 0.20 F value = 0.6, p = .58 Direct Bilirubin 0.05 0.02 - 0.40 0.15 0.10 0.06 0.10-0.40 0.122 0.10 0.04 0.20 - 0.30 0.12 0.10 0.02 0.10 0.30 0.10 0.10 F value = 3.5, p = .02 0.07 0.10-0.40 0.04 0.10-0.20 0.04 0.10-0.20 0.00 0.10-0.10 001120 1995 Data PFOS (ppm)________ Mean________ Median_________ SD 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 219 215 216 213 214 214 213 221 F value = 0.1, p =.96 47 43 35 36 0 - <1 ppm 1 - < 3 ppm 3 - <6 ppm > 6 ppm 0 - <1 ppm 1- <3 ppm 3 - <6 ppm > 6 ppm 140 137 134 134 137 135 142 136 F value = 0.2 , P =-87 53 53 48 47 45 46 45 46 F value = 2.9, p =.04 0 - <1 ppm 1- <3 ppm 3 - <6 ppm > 6 ppm 129 96 161 133 158 142 132 151 F value = 1.1, p =.35 43 40 34 32 13 13 11 9 98 107 88 45 001121 Table 5 (continued) 1997 Data Range Mean Median Cholesterol 100 - 340 118-315 128 - 278 160-251 198 197 216 219 2292 224 229 238 F value = 4.3, p =.006 SD 40 42 29 26 LDL 29 - 261 44 - 234 65 -190 95 -178 124 141 1482 145 F value = 3.7, p =.01 128 134 142 156 31-94 26 -94 23-74 34-61 HDL 46 44 48 40 F value = 1.1, p =-34 48 45 47 38 Triglycerides 41 - 622 41-651 34-413 64-187 148 107 156 122 166 158 220 191 F value = 0.5, p =.67 34 38 24 26 11 10 10 4 162 108 92 83 Range 110 - 280 116-365 192-321 186-250 50 - 205 61 - 290 111 - 196 103 - 164 19-74 28-69 32-69 37-45 38 - 1209 44-534 45 - 394 149 - 352 PFOS (ppm) Mean 1995 Data Median SD 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 47 47 46 46 47 47 47 48 F value = 2.4, p =.07 2 3 2 2 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 15.5 15.5 15.2 15.2 15.5 15.5 15.5 15.4 F value = 2.2, p =.10 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 5.0 5.0 4.9 4.9 5.0 5.0 5.0 4.9 F value = 0.4, p = .75 0.7 1.0 0.8 0.7 0.3 0.3 0.2 0.5 001122 Table 5 (continued) 1997 Data Range Mean Median SD Hematocrit 43-51 38-52 41-52 44-49 46 46 45 46 46 44 45 44 F value = 2.1, p = .11 2 3 3 2 Hemoglobin 14.0- 16.7 13.0-17.4 13.6- 17.4 14.7-16.2 15.5 15.5 15.4 15.5 15.0 14.7 15.1 15.0 F value =:1.8,p =.15 RBC 4.3-5.7 4.3--5.7 4.6--5.5 4.0-5.7 5.1 5.2 5.0 5.1 5.0 5.0 5.0 4.9 F value == 1-4, p = .25 0.8 0.9 1.0 0.7 0.3 0.3 0.3 0.3 Range 40-52 39-53 39-50 42-47 13.5- 17.0 13.3 - 17.3 12.5- 16.7 14.1 - 16.2 4.3 -5.9 4.1 -5.5 4.4 -5.7 4.7 -5.5 PFOS (ppm) 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm Mean 1995 Data Median SD 31.3 31.2 30.9 30.9 31.2 31.3 31.2 30.5 F value = 0.9 , p =.45 1.6 1.4 1.5 2.8 Table 5 (continued) 1997 Data Range Mean Median SD MCH 27.7 - 34.3 26.4 - 34.3 26.0 - 33.6 28.2 - 36.9 30.4 30.3 30.6 30.5 30.2 31.2 30.5 30.5 F value = 0.6, p = .65 1.5 1.2 1.9 2.1 33.1 33.2 33.2 33.1 33.1 32.6 33.1 33.3 F value = 0.2, p - .90 94 95 93 94 94 94 94 92 F value - 1.1, p = .35 MCHC 0.7 31.9-34.7 33.6 33.6 0.6 31.7-34.5 33.6 33.6 0.7 31.3-34.3 33.5 33.5 0.6 32.2 - 34.0 33.9 33.9 F value := 0.7, p = .56 MCV 5.2 85-106 90 90 4.3 80-104 91 91 4.8 81-104 90 91 9.7 85-115 90 91 F value = 0.6, p = .59 0.5 0.5 0.7 0.5 4.4 3.6 5.2 5.7 Range 27.6-34.1 26.7 - 33.8 26.2 - 32.9 27.5 - 33.4 32.2 - 34.9 31.8-34.6 32.0 - 34.4 33.4 - 34.6 83 - 101 81-99 80-97 81-96 001124 PFOS (PPm) Mean 1995 Data Median SD Table 5 (continued) Range Mean 1997 Data Median SD 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 6.1 6.0 7.0213 6.8 7.62 6.9 7.0 6.9 F value = 4.3, p =.006 WBC 1.3 4.1 -9.4 2.0 3.6-15.5 2.2 4.1 -13.3 0.6 6.4 - 7.8 6.1 5.8 6.9 6.6 6.2 6.1 6.2 7.1 F value = 2.2, p =.09 1.4 1.9 1.5 1.5 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 226 224 229 224 228 226 185 182 F value ==2.1, p =.10 Platelets 40 159 - 309 47 122 - 367 45 132 - 344 50 132 - 277 227 220 229 223 220 221 199 191 F value = 0.6, p = 0.60 53 49 45 58 1. Mean is significantly different (p < .05, Bonferroni (Dunn) test) than the mean of the other PFOS categories. 2. Mean is significantly different (p < .05, Bonferroni (Dunn) test) than the mean of 0 - < 1 ppm PFOS category. 3. Mean is significantly different (p < .05, Bonferroni (Dunn) test) than the mean of 1 - < 3 ppm category. * Sample Size 0 - <1 ppm 1 - <3 ppm 3 - <6 ppm > 6 ppm 1995 Data 45 91 35 7 1997 Data 60 63 21 5 Range 3.8 -10.3 3.8-13.2 4.0 - 10.0 4.2 - 7.4 106 - 406 124 - 359 147-316 146 - 295 Table 6 Mean, Median (Med), Standard Deviation (SD) of Mean and Range of PFOS, Demographic, Serum Chemistries and Hematological Values by Plant Location, 1995 PFOS* (ppm) Antwerp Mean Med SD 1995 Data ` Decatur Ranee Mean Med SD Ranee 0 <1 1 <3 3 <6 >6 PFOS 0.461 0.45 0.29 0.00 - 0.90 0.601 0.63 0.19 0.25 - 0.88 1.691 1.50 0.57 1.00-2.90 1.891 1.89 0.58 1.00-2.91 3.961 3.70 0.79 3.00 - 5.60 4.311 4.23 0.82 3.11 -5.80 8.171 8.50 1.92 6.10-9.90 8.171 6.90 3.20 6.06 - 12.83 F value = 241.0, p = .0001 F value = 114.5, p = .0001 0 <1 1 <3 3 <6 >6 Aee 36 36 7 21-52 40 39 11 29-58 36 34 9 25-60 46 46 7 30-58 37 37 7 28-51 44 44 7 26-55 40 37 6 37-47 48 47 7 42-56 F value = 0.5, p = .71 F value = 2.0, p = .12 0 <1 1 <3 3 <6 >6 Alcohol 1.0 0.7 1.0 0.0 - 3.6 0.1 0.0 0.3 0.0 - 0.9 1.0 0.8 0.9 0.0 - 3.6 0.2 0.0 0.4 0.0 - 2.0 2.0 1.3 2.2 0.0 - 6.0 0.4 0.0 0.9 0.0 - 3.4 1.7 1.4 1.1 0.7 - 2.9 0.0 0.0 0.0 0.0-0.1 F value = 2.9, p = .04 F value = 1-1, P = .37 0 <1 1 <3 3 <6 >6 BMI 24.1 24.3 2.4 17.9-28.1 30.1 28.0 5.3 22.8 - 38.7 24.3 23.8 2.4 19.6-31.6 29.6 28.3 6.2 22.3 - 60.7 23.1 23.0 3.3 17.9-31.4 27.0 27.5 3.4 19.1 -32.5 23.7 24.7 2.8 20.6 - 25.8 30.7 30.2 1.6 29.4 - 33.0 F value = 0.9, p = .47 F value = 1.1, p = .35 001125 PFOS (ppm) 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 Table 6 (continued) ________ Antwerp__________ Mean Med SD Ranee _________ Decatur Mean Med SD Ranee 3.0 0.0 7.0 4.1 0.0 7.0 9.13 7.5 9.0 1.0 0.0 1.7 F = 3.1, p = .03 Cigarettes 0-25 ' 1.4 0.0 3.8 0 -1 3 0-23 8.3 0.0 12.9 0 -4 0 0-25 12.5 0.0 15.7 0 -4 0 0 -3 0 0.0 0 0 - 0 F value = 2.3, p = .09 Alkaline Phosphatase 75 75 16 31 - 104 96 99 29 47 - 158 73 66 19 49 - 108 98 95 27 49-191 79 78 20 32-12195 95 96 18 58 - 124. 74 74 11 63-85 98 91 27 73- 136 F value = 0.5, p = .69 F value = 0.1, p = .98 GGT 39 29 22 16-111 55 41 41 21 - 155 48 30 58 12-293 46 40 24 2-118 34 34 11 21-55 47 48 16 21 -80 31 32 3 28-33 52 49 20 30-79 F value = 0.7, p = .55 F value = 0.5, p = .71 AST 26 23 13 15-96 27 26 14 14-90 23 22 5 13-35 33 35 6 26-37 F value = 0.9, p = .45 32 31 9 29 27 10 27 26 5 34 32 7 F value = 0.9, p = .44 19-52 17-85 21-37 28-43 001126 PFOS (ppm) 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 Table 6 (continued) Antwerp Mean Med SD Range Decatur Mean Med SD Range ALT 46 42 18 27 - 108 55 47 24 37- 118 45 40 18 26 - 122 47 42 22 18-183 40 39 6 30-52 43 42 9 30-59 50 38 28 29-82 52 53 8 41 -59 F value = 0.7, p = .59 F value = 0.8, p = .49 Total Bilirubin 0.96 0.80 0.55 0.40 - 2.90 0.65 0.60 0.16 0.40 - 0.90 0.83 0.80 0.26 0.40-1.30 0.57 0.50 0.28 0.20-1.50 0.75 0.70 0.30 0.30-1.40 0.51 0.50 0.18 0.20-1.00 0.93 0.90 0.25 0.70-1.20 0.63 0.65 0.10 0.50 - 0.70 F value = 1.2, p = .31 F value = 0.7, p = .54 0.23 0.20 0.22 0.20 0.21 0.20 0.20 0.20 F value ii o 3-j Direct Bilirubin 0.06 0.20 - 0.40 0.20 0.20 0.00 0.20 - 0.20 0.04 0.20 - 0.30 0.20 0.20 0.06 0.10-0.40 0.03 0.20 - 0.30 0.20 0.20 0.04 0.20 - 0.30 0.00 0.20 - 0.20 0.20 0.20 0.08 0.10-0.30 p = .55 F value = 0.4, p = .74 17.0 17.0 3.6 16.7 16.0 3.8 17.3 17.0 3.8 17.3 21.0 6.4 F value = 0.1, p = .95 BUN 12.0 - 26.0 14.8 15.0 2.7 11.0-26.0 14.7 14.0 4.0 10.0 - 23.0 15.2 15.0 3.1 10.0-21.0 13.5 13.5 0.06 F value = 0.2, p = .87 11.0-21.0 8.0 - 24.0 10.0 - 23.0 13.0-14.0 Creatinine 1.0 0.9 0.2 0.8- 1.6 1.1 1.1 0.2 0.7- 1.3 0.9 0.9 0.1 0.7- 1.1 1.1 1.1 0.2 0.9- 1.6 0.93 0.9 0.1 0.7- 1.0 1.1 1.1 0.1 0.8 - 1.2 0.8 0.8 0.2 0.6 - 0.9 1.3 1.3 0.2 1.2- 1.6 F value = 4.7, p = .004 F value = 3.8, p = .01 001127 PFOS (PPm) 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 > '6 0 <1 1 <3 3 <6 > i6 0 <1 1 <3 3 <6 >6 Table 6 (continued) _________Antwerp__________ Mean Med SD Ranee _________ Decatur Mean Med SD Ranee Glucose 83 83 8 66 -103 90 86 28 66-170 82 81 13 60-12 6 93 89 25 67 - 260 80 82 9 66 - 101 89 89 14 66-114 72 72 2 71-74 93 92 12 83 - 105 F value = 1.3, p = .29 F value = 0.1, p = .94 Cholesterol 220 219 50 100-340 215 208 39 154-276 206 211 49 118-315 221 218 39 132-300 217 215 30 178-266 209 213 42 128-278 223 221 16 208 - 240 206 206 47 160-251 F value = 0.6, p = .61 F value = 0.5, p = .69 LDL 140 138 45 29- 261 139 130 38 79-- 192 131 124 46 44- 220 136 136 36 62--234 143 139 27 99- 189 131 128 41 65-- 190 144 136 20 130 - 168 139 142 42 95-- 178 F value = 0.4, p = .76 F value = 0.1, p = .94 HDL 56 57 13 31 -94 43 41 9 53 51 13 33--79 45 44 12 50 49 11 31 -74 39 39 9 53 49 7 48--61 39 39 5 F value = 1.1, P = .35 F value = 1.4, p = .25 31 -59 26--94 23--51 34--46 Trielvcerides 117 93 98 41 - 622 167 151 94 62 - 307 105 75 65 41 - 368 191 146 114 61-651 126 112 64 34 - 278 199 198 99 78-413 129 116 52 85 - 187 135 153 48 64- 168 F value = 0.3, p = .81 F value = 0.5, p = .67 001128 PFOS loom) 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 0 <1 1 <3 3 <6 >6 Table 6 (continued) ________ Antwerp__________ Mean Med SD Ranee __________Decatur Mean Med SD Ranee Hematocrit 47 47 2 43-51 46 45 2 46 47 2 42-51 45 45 3 48 47 2 44-51 46 46 3 47 48 1 46-48 46 46 2 F value = 1-5, P = 23. F value = 0.3, p = .83 43-50 38-52 41-52 44-49 Hemoelobin 15.5 15.5 0.8 14.0-16.7 15.5 15.4 0.7 14.5 - 16.4 15.2 15.3 0.7 13.8 -16.6 15.1 15.1 1.1 13.0-17.4 15.6 15.6 0.5 14.6 - 16.7 15.3 15.3 1.0 13.6- 17.4 15.4 15.4 0.6 14.8 - 16.0 15.5 15.6 0.8 14.7 - 16.2 F value = 1.4, p = .26 F value = 0.7, p = .58 RBC 4.9 5.0 0.3 4.3--5.7 5.1 5.1 0.2 4.9 4.9 0.2 4.3 -5.4 4.9 4.9 0.3 5.0 5.0 0.2 4.7--5.3 5.0 4.9 0.3 4.7 4.9 0.6 4.0--5.2 5.2 5.1 0.4 F value = 0.9, P = .46 F value = 1.1, p = .35 4.8 -5.5 4.3 -5.7 4.6 -5.5 4.8 -5.7 MCH 31.5 31.3 1.5 27.7-34.3 30.7 30.7 1.6 28.7 - 33.5 31.2 31.0 1.2 29.5-34.3 30.7 30.9 1.5 26.4-33.1 31.5 31.9 1.2 29.1 -33.3 30.9 31.2 1.7 26.0 - 33.6 33.0 32.5 3.7 29.6 - 36.9 29.9 30.4 1.2 28.2 - 30.7 F value = 1.4, p = .25 F value = 0.4, p = .74 32.9 33.0 32.9 32.8 32.8 32.7 32.6 32.4 F value il o MCHC 0.6 31.9-34.5 33.8 33.9 0.6 32.9 - 34.7 0.6 31.7-33.8 33.3 33.3 0.5 31.9-34.5 0.7 31.3-34.2 33.4 33.4 0.6 32.1 -34.3 0.5 32.2 -33.1 33.5 33.4 0.3 33.3 - 34.0 p = .78 F value = 2.2, p = .10 001129 PFOS (ppm) 0 <1 1 <3 3 <6 >6 Table 6 (continued) Antwero Mean Med SD Range Decatur Mean Med SD Range MCV 96 96 5 85 -106 91 90 95 95 3 90 - 104' 92 93 96 97 4 88 - 104 92 92 101 98 12 91 - 115 89 90 F value = 1.8, p = .16 F value 4 85-99 4 80 - 100 5 81 - 101 3 85-92 p = .50 ii it i--k oo o ll 4s.O 0 <1 1 <3 3 <6 >6 5.9 5.7 6.1 6.1 7.5123 6.9 6.5 6.4 F value ii a >n ll WBC 1.3 4.1 -9.4 6.8 6.9 1.1 1.3 3.8-8.8 7.5 7.0 2.2 2.5 4.1 - 13.3 7.7 7.4 1.9 0.2 6.4 - 6.7 7.4 7.4 0.4 .006 F value = 0.5, p = .72 5.0 - 9.0 3.6 -15.5 4.9-11.5 6.9-7.8 0 <1 1 <3 3 <6 >6 Platelets 224 222 42 159-309 233 246 222 220 41 153-318 233 225 2374 234 52 132-344 218 214 162 151 36 132 - 202 202 194 F value = 2.6. d = .06 F value 35 50 34 56 162-271 122 - 367 172 - 287 143 - 277 ^Sample sizes: FC95 Level Antwerp 0 < 1 ppm 34 1 <3 ppm 32 3 < 6 ppm 19 > 6 ppm 3 88 Decatur 11 59 16 4 90 1. Significantly different (p < .05, Bonferroni (Dunn) t-test) than the remaining three PFOS exposure categories. 2. Significantly different (p < .05, Bonferroni (Dunn) t-test) than the 0 - < 1 ppm and the 1 < 3 ppm categories. 3. Significantly different (p < .05, Bonferroni (Dunn) t-test) than the 0 - < 1 ppm category. 4. Significantly different (p < .05, Bonferroni (Dunn) t-test) than the 3 - < 6 ppm category. 001130 Table 7 Mean, Median (Med), Standard Deviation (SD) of Mean and Range of PFOS, Demographic, Serum Chemistries and Hematological Values by Plant Location, 1997 PFOS* (ppm) Antwerp Mean Med SD 1997 Data Decatur Ranee Mean Med SD Ranee 0-< 1 1 -< 3 3-<6 >6 PFOS 0.461 0.37 0.27 0.10-0.94 0.601 0.59 0.26 0.10-0.97 1.891 1.79 0.61 1.02 - 2.89 1.71' 1.53 0.52 1.04-2.85 3.871 3.66 0.59 3.22 - 4.83 3.871 3.51 0.81 3.09 - 5.30 --- - 7.201 6.68 1.59 6.05 - 9.93 F value:= 195.3, p = .0001 F value: = 218.1, p = .0001 0-< l 1-<3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 Age 29 28 6 2 1 -5 0 43 44 9 23-62 373 37 9 24-63 45 45 8 3 1 -6 2 373 37 3 3 2 -4 0 45 44 4 36-54 --- - 42 45 9 2 9 -5 2 F value 10.4, p = .0001 F value: = 0.5, p = .69 Alcohol 0.8 0.5 1.0 0.0-4.3 0.2 0.1 0.4 0.1 -2.0 1.0 0.7 1.1 0.0-5.0 0.1 0.1 0.1 0.1 -0.3 2.23 1.4 2.3 0.0-7.1 0.1 0.1 0.1 0.1 -0.3 0.2 0.1 0.3 0.1 -0.8 F value: = 4.3, p = .02 F value = 0.8, p = .49 BMI 22.9 21.9 2.1 20.2-28.3 29.3 28.9 4.9 22.4-41.7 24.2 23.9 2.8 18.1 -30.4 30.0 29.0 6.0 20.2 - 48.5 23.6 24.9 3.1 19.2-28.3 30.0 29.3 3.0 25.4 - 36.0 30.8 29.7 4.0 26.1 -36.2 F value:= 2.0, p = .15 F value; = 0.2, p = .91 001131 PFOS (ppm) 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0--< 1 1 -< 3 3- <6 > i5 0-< 1 1-<3 3-<6 >6 Table 7 (continued) Antwerp Mean Med SD Ranee Decatur Mean Med SD Ranee 4.2 0.0 6.7 7.3 2.0 8.1 5 0.0 7.8 -F = 1.3, p = .28 Ciearettes 0 - 20 5.3 0.0 11.6 0 -4 0 0-20 8.7 0.0 13.0 0 -4 0 0-20 3.3 0.0 9.0 0 -3 0 - 6.0 0.0 13.4 0 -3 0 F value = 0.8, p = .51 Alkaline Phosphatase 69 68 14 4 9 - 110 86 84 16 55 - 132 74 74 16 41 - 113 88 83 26 41 - 163 64 59 24 29 - 120 85 83 15 61 - 109 -. 88 84 18 65 - 114 F value = 1.5, p = .22 F value = 0.1,p = .96 GGT 21 17 10 10--50 36 32 25 13 -- 142 34 24 34 10- 144 37 27 33 13-- 179 25 22 10 14-43 31 28 15 13 --71 - -- 33 37 12 17 --48 F value = 2.5, p = .09 F value: = 0.2, p = .91 AST 27 26 7 17-53 26 25 7 13-48 27 25 7 15-48 26 25 7 18-56 25 24 4 19-30 25 23 9 14-43 -- - 29 28 3 2 6 -3 4 F value: = 0.2, p = .80 F values= 0.4, p = .77 001132 PFOS (ppm) 0-< 1 1-<3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1-<3 3 - <6 >6 Table 7 (continued) Antwerp Mean Med SD Range Decatur Mean Med SD Range ALT 30 25 12 13-60 33 31 16 1 3 -8 7 ' 28 23 10 14-46 - -- - F value: = 0.6, p = .58 33 31 11 17-57 33 28 16 10-89 38 33 21 17-82 41 45 10 2 5 -4 9 F value = 0.9, p = .45 Total Bilirubin 0.90 0.80 0.46 0.40 - 2.30 0.63 0.60 0.30 0.30-1.40 0.68 0.70 0.23 0.30-1.30 0.56 0.50 0.18 0.30-1.00 0.79 0.70 0.40 0.30-1.30 0.51 0.50 0.16 0.30 - 0.90 - -- - 0.58 0.50 0.24 0.40- 1.00 F value = 2.3, p = .11 F value = 1.0, p = .41 Direct Bilirubin 0.16 0.20 0.08 0.10-0.40 0.13 0.10 0.06 0.10-0.30 0.13 0.10 0.05 0.10-0.20 0.11 0.10 0.03 0.10-0.20 0.14 0.10 0.05 0.10-0.20 0.11 0.10 0.03 0.10-0.20 - -- - 0.10 0.10 0.00 0.10-0.10 F value: = 2.2, p = .12 F value:= 1.3, p = .28 BUN 14.3 14.0 2.1 11.0-19.0 14.7 14.0 3.3 9.0-21.0 15.2 15.0 2.6 10.0 - 20.0 13.5 13.5 3.5 6.0 - 26.0 16.0 16.0 2.7 12.0 - 20.0 14.3 14.0 3.0 9.0 - 19.0 - -. 13.8 12.0 4.1 9.0-19.0 F value = 2.1, p = .13 F value = 0.7, p = .56 Creatinine 0.9 0.9 0.1 0.7-1.2 1.0 1.0 0.1 0.6- 1.2 0.9 0.9 0.1 0.7- 1.3 1.0 1.0 0.1 0.7- 1.3 0.9 0.9 0.1 0.8- 1.1 1.0 1.0 0.1 0.7 - 1.1 - -- - 1.0 0.9 0.2 0.8 - 1.4 F value: = 0.3, p = .76 F value = 0.2, p = .89 001133 PFOS (ppm) 0-< 1 1 -< 3 3-<6 >6 0-< 1 1-<3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0--< 1 1 -< 3 3 -< 6 > i5 0-< 1 1-<3 3-<6 >6 Table 7 (continued) Antwerp Mean Med SD Ranee Decatur Mean Med SD Range Glucose 81 82 10 63- 114 93 91 21 58 - 174 79 78 9 65 -102 102 89 47 75 - 303 84 85 8 74-96 103 93 30 79 - 192 - -- - 89 88 7 8 0 -9 7 F value: = 0.8, p = .47 F values= 0.5, p = .67 Cholesterol 193 190 41 110-277 204 208 38 145 - 277 213 205 48 116-365 218 226 39 152 - 290 228 223 38 192-321 230 230 23 197 - 280 229 238 26 186 - 250 F value = 2.9, p = .07 F value = 2.0, p = .13 LDL 122 114 35 57 - 205 127 133 33 50 - 178 143 134 42 67 - 290 139 135 35 61 - 196 147 141 24 111 -195 149 144 25 113 -196 - 145 156 26 103 - 164 F value = 3.0, p = .06 F value = 1.6, p = .19 HDL 51 50 12 19--7 4 48 46 10 34--6 8 51 50 10 39--6 9 - -F value: = 0.8, P = .47 42 41 9 2 6 --5 9 42 41 10 2 8 --6 9 45 45 10 3 2 --6 2 40 38 4 37--45 F value! = l0.5, p = .67 Triglycerides 99 92 41 38- 175 200 128 219 46 - 1209 112 95 58 44 - 290 185 147 124 63 - 534 150 122 88 65 - 362 179 183 98 45 - 394 -- 220 191 83 149 - 352 F value = 2.9, p = .06 F value = 0.1, p = .95 001134 PFOS (ppm) 0-< 1 1 -< 3 3-<6 >6 0-< l 1 -< 3 3-<6 >6 0- < 1 1- <3 3- <6 >6 0-< 1 1-<3 3-<6 >6 0-< 1 1-<3 3-<6 >6 Table 7 (continued) Antwerp Mean Med SD Ranee Decatur Mean Med SD Ranee Hematocrit 47 47 3 4 2 -5 2 46 45 2 4 0 -5 0 46 47 3 4 0 -5 3 46 46 3 39-51 46 45 3 4 2 -5 0 44 44 2 39-48 - -- - 45 44 2 4 2 -4 7 F value: = 0.7, P = .51 F value = 1.6, p = .18 Hemoelobin 15.6 15.5 0.8 14.2 -17.0 15.4 15.4 0.7 13.5 - 16.9 15.4 15.6 1.0 13.3 -17.3 15.4 15.5 0.9 13.5 - 17.3 15.2 14.7 1.0 14.0 -16.7 14.8 14.7 1.0 12.5 - 16.5 -- - 15.1 15.0 0.8 14.1 - 16.2 F value = 0.7, p = .48 F value = 1-7, p = .18 RBC 5.1 5.2 0.4 4.6--5.9 5.1 5.1 0.3 4.3 -5.7 5.0 5.1 0.3 4.1 --5.5 5.1 5.1 0.3 4.2--5.5 4.9 4.8 0.4 4.4--5.7 5.0 5.0 0.3 4.5 --5.4 - -- 5.0 4.9 0.3 4.7--5.5 F value=1.7, P = .19 F value = 0.4, P = .78 MCH 30.4 30.4 1.3 27.9-33.5 30.3 30.2 1.8 27.6 - 34.1 30.6 30.5 0.9 29.2 - 32.8 30.6 30.6 1.4 26.7 - 33.8 30.9 31.4 1.4 28.8-32.7 29.7 30.5 2.2 26.2 - 32.9 - -- 30.5 30.5 2.1 27.5 - 33.4 F value:= 0.9, p = .40 F value=1.0, p = .41 MCHC 33.4 33.4 0.5 32.2 - 34.0 33.8 33.3 33.2 0.5 31.8-34.1 33.8 33.4 33.5 0.5 32.4 - 34.3 33.6 - 33.9 F value = 0.6, p = .54 33.7 0.5 32.8 - 34.9 33.8 0.5 32.7 - 34.6 33.9 0.8 32.0 - 34.4 33.9 0.5 33.4 - 34.6 F value: = 0.6, p = .62 001135 PFOS (Dom) 0-< 1 1-<3 3-<6 >6 Table 7 (continued) Antwerp Mean Med SD Ranee Decatur Mean Med SD Ranee MCV 91 91 4 84-101 90 89 5 83 -101 92 92 3 87-99 91 90 4 81-99 93 95 5 86-97 88 89 5 8 0 -9 6 -- - 90 91 6 81-96 F value = 1.3, p = .29 F value = 0.9, p = .45 0-< 1 1-<3 3-<6 >6 WBC 6.0 5.7 1.3 3.8-8.8 6.3 5.9 1.6 4.0 - 10.3 7.1 6.8 2.2 4.4 - 13.2 6.7 6.6 1.6 3.8-10.1 6.4 6.0 1.8 4.2 -10.0 6.1 6.1 1.3 4.0 - 8.9 -- - 6.2 7.1 1.5 4 .2 -7 .4 F value: = 2.6, p = .08 F value: = 0.8, p = .51 0-< 1 1 -< 3 3-<6 >6 Platelets 237 232 55 126-406 215 207 50 106 - 363 243 232 48 151-359 219 210 48 124-323 215 225 41 147 - 263 224 219 50 159-316 - - - _ 199 191 58 146 - 295 F value = 1.1, p = .35_____________F value = 0.3. p = .80 *Sample sizes: FC95 Level Antwerp 0 - < 1 ppm 31 1 - < 3 ppm 25 3 - < 6 ppm 9 > 6 ppm 0 65 Decatur 29 38 12 5 84 1. Significantly different (p < .05, Bonferroni (Dunn) t-test) than the remaining three PFOS exposure categories. 2. Significantly different (p < .05, Bonferroni (Dunn) t-test) than the 0 - < 1 ppm and the 1 - < 3 ppm categories. 3. Significantly different (p < .05, Bonferroni (Dunn) t-test) than the 0 - < 1 ppm category. 001136 Table 8 Multivariable Regression of Serum Chemistries and Hematological Parameters - Examination of the Effect of PFOS Adjusting for Age, Alcohol, BMI and Cigarettes, Antwerp and Decatur Data Combined, 1995 and 1997 Examinations Intercept PFOS Age Alcohol BMI Cigarettes Parameter 52.99 0.32 0.08 -5.43 1.04 0.85 R2 = .26 1995 Data SE 11.14 0.89 0.20 1.53 0.34 0.15 Adj R2= .24 Alkaline Phosphatase p value .0001 .72 .68 .0005 .002 .0001 Parameter 40.87 0.19 0.45 -2.09 0.69 0.49 R2= .19 1997 Data SE 9.60 1.03 0.18 1.61 0.34 0.15 Adj R2= .16 p value .0001 .86 .01 .20 .04 .002 Intercept PFOS Age Alcohol BMI Cigarettes 7* II Parameter 2.3816 0.0032 0.0040 0.0605 0.0372 0.0076 1995 Data SE 0.2711 0.0217 0.0048 0.0373 0.0082 0.0038 Adj R2= .12 InGGT p value .0001 .88 .41 .11 .0001 .05 Parameter 2.0471 0.0178 0.0036 0.0747 0.0365 0.0030 R2= .15 1997 Data SE 0.2592 0.0280 0.0049 0.0438 0.0091 0.0042 Adj R2= .12 p value .0001 .53 .47 .09 .0001 .48 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 12.62 0.42 0.0005 0.70 0.54 -0.11 R2= .08 1995 Data SE 5.52 0.44 0.10 0.76 0.17 0.08 Adj R2= .05 AST p value .02 .35 .99 .36 .002 .14 Parameter 23.25 -0.11 -0.07 0.40 0.23 -0.07 R2= .04 1997 Data SE 3.50 0.38 0.07 0.59 0.12 0.06 Adj R2 = .00 p value .0001 .77 .29 .50 .07 .20 001137 Table 8 (continued) Intercept PFOS Age Alcohol BMI Cigarettes Parameter 28.54 0.09 -0.18 -0.16 0.96 -0.14 R2 = .08 ALT 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE_________ p value 9.46 .003 11.60 6.78 .09 0.76 .91 0.97 0.73 .19 0.17 .28 -0.19 0.13 .14 1.30 .90 0.31 1.14 .79 0.29 .001 1.00 0.24 .0001 0.13 .28 -0.09 0.11 .40 Adj R2 = .05 R2= .13 Adj R2= .10 Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter 0.2742 -0.0984 0.0086 0.0052 0.0738 -0.0273 -0.0158 R2= .32 1995 Data SE 0.1958 0.0368 0.0039 0.0035 0.0268 0.0059 0.0028 Adj R2= .30 In Total Bilirubin p value .16 .008 .03 .14 .007 .0001 .0001 Parameter -0.1395 -0.1620 0.0188 0.0015 0.1220 -0.0097 -0.0054 R2 = .18 1997 Data SE 0.1945 0.0515 0.0070 0.0037 0.0327 0.0068 0.0031 Adj R2= .14 p value .47 .002 .009 .70 .0003 .16 .09 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 0.2196 -0.0009 0.0006 0.0074 -0.0012 -0.0007 R2 = .06 1995 Data SE 0.0259 0.0021 0.0005 0.0036 0.0008 0.0004 Adj R2= .03 Direct Bilirubin p value .0001 .68 .17 .04 .13 .06 Parameter 0.1929 -0.0061 -0.0002 0.0106 -0.0017 -0.0009 R2 = .14 1997 Data SE 0.0264 0.0029 0.0005 0.0045 0.0009 0.0004 Adj R2= .11 p value .0001 .03 .69 .02 .07 .05 001138 Table 8 (continued) Intercept PFOS Age Alcohol BMI Cigarettes Parameter 18.64 -0.02 -0.03 0.37 -0.05 -0.08 R2 = .08 1995 Data SE 1.94 0.15 0.03 0.27 0.06 .03 Adj R2= .05 BUN Dvalue .0001 .91 .39 .17 .42 .004 Parameter 13.22 -0.08 0.08 0.50 -0.07 -0.05 R2 = .09 1997 Data SE 1.47 0.16 0.03 0.25 0.05 0.02 Adj R2= .06 Dvalue .0001 .60 .005 .05 .20 .05 Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter 0.5744 -0.0223 0.0033 0.0059 -0.0260 0.0086 -0.0005 R2 = .29 1995 Data SE 0.0779 0.0146 0.0016 0.0013 0.0106 0.0023 0.0011 Adj R2= .27 Creatinine D value .0001 .13 .04 .0001 .02 .0003 .65 Parameter 0.9478 -0.0302 0.0055 0.0027 -0.0242 -0.0022 -0.0019 R2= .16 1997 Data SE 0.0602 0.0159 0.0022 0.0012 0.0101 0.0021 0.0010 Adj R2= .13 D value .0001 .06 .01 0.02 .02 .30 .05 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 3.8943 -0.0039 0.0045 -0.0026 0.0143 -0.0003 R2= .31 1995 Data SE 0.0732 0.0058 0.0013 0.0101 0.0022 0.0010 Adj R2= .29 In Glucose Dvalue .0001 .51 .0006 .79 .0001 .78 Parameter 3.8597 0.0014 0.0025 -0.0068 0.0195 -0.0013 R2 = .30 1997 Data SE 0.0919 0.0098 0.0017 0.0154 0.0032 0.0015 Adj R2= .27 o value .0001 .89 .15 .66 .0001 .39 001139 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 190.1 -0.99 0.74 -0.40 -0.04 -0.18 R2 = .02 Table 8 (continued) 1995 Data SE 22.1 1.77 0.39 3.04 0.67 0.31 Adj R2= .00 Cholesterol p value .0001 .58 .06 .90 .95 .56 Parameter 155.5 4.66 1.39 7.50 -0.40 -0.08 R2= .17 1997 Data SE 19.0 2.04 0.36 3.18 0.67 0.30 Adj R2= .14 d value .0001 .02 .0002 .02 .55 .80 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 135.26 0.50 0.58 -3.33 -6.67 -0.52 R2 = .05 LDL 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE_________ p value 20.27 .0001 114.69 17.06 .0001 1.61 .76 4.01 1.82 .03 0.36 .11 1.04 0.32 .002 2.78 .23 4.29 2.84 .13 0.61 .28 -1.09 0.60 .07 0.28 .07 -0.02 0.28 .95 o <NII Pi < Adj R2= .02 R2= .13 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 70.82 -1.17 -0.11 3.53 -0.63 -0.17 R2 = .28 HDL 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE_________ p value 5.73 .0001 66.71 4.52 .0001 0.46 .01 -0.22 0.49 .65 0.10 .27 0.12 0.09 .17 0.79 .0001 2.01 0.76 .009 0.17 .0004 -0.95 0.16 .0001 0.08 0.03 -0.16 0.07 .03 Adj R2= .26 R2= .29 Adj R2 = .26 001140 Intercept PFOS Age Alcohol BMI Cigarettes Parameter -103.29 -1.78 1.54 -0.43 6.71 2.70 R2= .24 Table 8 (continued) 1995 Data SE 45.97 3.68 0.82 6.32 1.39 0.64 Adj R2= .22 Triglycerides d value .03 .63 .06 .95 .0001 .0001 Parameter -187.99 -0.33 1.37 14.68 10.06 1.87 R2= .23 1997 Data SE 58.18 6.28 1.11 9.82 2.05 0.94 Adj R2= .20 d value .002 .96 .22 .14 .0001 .05 intercept PFOS Age Alcohol BMI Cigarettes Parameter 48.01 -0.06 0.0005 0.10 -0.08 0.05 R2= .08 1995 Data SE 1.36 0.11 0.02 0.19 0.04 0.02 Adj R2= .05 Hematocrit D value .0001 .57 .98 .58 .06 .007 Parameter 46.73 -0.25 -0.03 0.15 0.01 0.04 R2= .08 1997 Data SE 1.28 0.14 0.02 0.22 0.05 0.02 Adj R2= .04 d value .0001 .07 .20 0.50 0.75 .04 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 15.78 -0.02 0.004 -0.02 -0.03 0.02 R2= .06 1995 Data SE 0.45 0.04 0.008 0.06 0.01 0.006 Adj R2= .03 Hemoglobin d value .0001 .53 .57 .77 .06 .02 Parameter 15.43 -0.08 -0.006 0.01 0.008 0.02 R2= .06 1997 Data SE 0.43 0.05 0.01 0.07 0.02 0.007 Adj R2= .03 d value .0001 .07 .46 .85 .59 .03 001141 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 4.8766 -0.0026 -0.0022 -0.0323 0.0069 -0.0006 R2= .04 Table 8 (continued) 1995 Data SE 0.1559 0.0123 0.0028 0.0213 0.0047 0.0022 Adj R2= .01 RBC p value .0001 .83 .42 .13 .14 .77 Parameter 5.1416 -0.0208 -0.0068 -0.0111 0.0089 -0.0025 R2 = .06 1997 Data SE 0.1626 0.1756 0.0031 0.0274 0.0057 0.0026 Adj R2= .02 p value .0001 .24 .03 .69 .12 .35 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 32.292 -0.015 0.23 0.189 -0.093 0.0347 R2 = .20 1995 Data SE 0.725 0.058 0.013 0.099 0.022 0.010 Adj R2= .18 MCH p value .0001 .79 .08 .06 .0001 .001 Parameter 30.062 -0.047 0.028 0.095 -0.035 0.046 R2 = .13 1997 Data SE 0.706 0.076 0.013 0.119 0.025 0.011 Adj R2= .10 p value .0001 .54 .04 .43 .16 .0001 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 32.8525 -0.0015 0.0097 -0.1121 -0.0006 -0.0041 R2= .07 1995 Data SE 0.3341 0.0266 0.0059 0.0458 0.0100 0.0047 Adj R2 = .05 MCHC p value .0001 .96 .11 .02 .95 .39 Parameter 33.0109 -0.0092 0.0101 -0.0589 0.0072 0.0016 R2 = .07 1997 Data SE 0.2675 0.0289 0.0051 0.0451 0.0094 0.0043 Adj R2 = .04 p value .0001 .75 .05 .19 .45 .72 001142 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 98.29 -0.03 0.04 0.88 -0.28 0.12 R2 = .23 Table 8 (continued) MCV 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE__________p value 2.29 .0001 90.91 1.98 .0001 0.18 .88 -0.09 0.21 .67 0.04 .33 0.06 0.04 .15 0.31 :oi 0.49 0.33 .14 0.07 .0001 -0.12 0.07 .08 0.03 .0005 0.13 0.03 .0001 Adj R2= .21 R2= .15 Adj R2= .12 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 272.49 -3.37 -0.88 -4.80 -0.04 0.21 R2 = .06 1995 Data SE 23.37 1.86 0.41 3.20 0.70 0.33 Adj R2= .03 Platelets p value .0001 .07 .04 .14 .95 0.52 Parameter 260.92 -2.91 -0.25 2.37 -0.73 -0.30 R2= .03 1997 Data SE 25.27 2.73 0.48 4.26 0.89 0.41 Adj R2= .00 p value .0001 .29 .61 .58 .41 .46 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 3.9538 0.0893 0.0214 -0.0860 0.0492 0.1009 R2 = .36 WBC 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE__________p value 0.8120 .0001 5.1296 0.6927 .0001 0.0646 .17 -0.0048 0.0748 .95 0.0144 .14 0.0139 0.0132 .30 0.1111 .44 0.2471 0.1169 .04 0.0244 .05 0.0045 0.0244 .86 0.0115 .0001 0.0878 0.0111 .0001 Adj R2 = .34 R2= .33 Adj R2 = .31 001143 Table 9 Multivariable Regression of Total Bilirubin, Direct Bilirubin, Creatinine Cholesterol, LDL, HDL, Heamtocrit, Hemoglobin and Platelets - Examination of the Effect of PFOS Adjusting for Age, Alcohol, BMI and Cigarettes, by Location, 1995 and 1997 In Total Bilirubin - Both Locations Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter 0.2742 -0.0984 0.0086 0.0052 0.0738 -0.0273 -0.0158 R2= .32 1995 Data __________ 1997 Data SE_________ p value______ Parameter SE__________p value 0.1958 .16 -0.1395 0.1945 .47 0.0368 .008 -0.1620 0.0515 .002 0.0039 .03 0.0188 0.0070 .009 0.0035 .14 0.0015 0.0037 .70 0.0268 .007 0.1220 0.0327 .0003 0.0059 .0001 -0.0097 0.0068 .16 0.0028 .0001 -0.0054 0.0031 .09 Adj R2= .30 R2= .18 Adj R2 = .14 In Total Bilirubin - Antwerp Onlv Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter -0.0776 -0.0473 0.0035 0.0062 0.0398 -0.0115 -0.0176 R2 = .15 1995 Data SE 0.0475 0.0575 0.0072 0.0055 0.0316 0.0161 0.0056 Adj R2= .08 Dvalue .85 .41 .63 .27 .21 .48 .002 Parameter -0.3130 -0.2157 0.0336 -0.0036 0.0826 0.0114 -0.0102 R2= .14 1997 Data SE 0.5586 0.1779 0.0385 0.0081 0.0441 0.0222 0.0074 Adj R2= .05 p value 0.58 .23 .39 .66 .07 .61 .17 Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes In Total Bilirubin - Decatur Only 1995 Data__________ __________ 1997 Data Parameter SE_________ p value______ Parameter SE__________p value -0.4462 0.3154 .16 -0.9486 0.2932 .002 -0.0862 0.0509 .09 -0.1160 0.0578 .05 0.0081 0.0048 .10 0.0144 0.0072 .05 0.0128 0.0048 .01 0.0134 0.0046 .005 -0.0045 0.0828 .59 0.1708 0.1607 .29 -0.0179 0.0068 .01 -0.0045 0.0073 .54 -0.0117 0.0031 .0004 -0.0031 0.0032 .33 90N 11 R2= .29 Adj R2= .24 .17 Adj R2= .10 001144 Table 9 (continued) Intercept PFOS Age Alcohol BMI Cigarettes HDL - Both Locations 1995 Data (N = 178) ________ 1997 Data (N = 149) Parameter SE_________ p value______ Parameter SE_________ p value 70.82 5.73 .0001 66.71 4.52 .0001 -1.17 0.46 .01 -0.22 0.49 .65 -0.11 0.10 .27 0.12 0.09 .17 3.53 0.79 .0001 2.01 0.76 .009 -0.63 0.17 .0004 -0.95 0.16 .0001 -0.17 0.08 0.03 -0.16 0.07 .03 R2= .28 Adj R2= .26 R2= .28 Adj R2= .26 Intercept PFOS Age Alcohol BMI Cigarettes HDL - Antwerp Only 1995 Data (N = 86) Parameter SE p value 73.44 12.69 .0001 -1.55 0.73 .04 0.01 0.17 .93 3.31 0.98 .001 -0.87 0.50 .09 -0.19 0.17 .28 R2 = .16 Adj R2= .11 1997 Data (N = 631 Parameter SE D value 57.30 12.12 .0001 -2.26 1.17 .06 0.43 0.17 .01 2.78 0.99 .007 -0.82 0.48 .10 -0.45 0.16 .007 R2 = .28 Adj R2= .21 Intercept PFOS Age Alcohol BMI Cigarettes HDL - Decatur Onlv 1995 Data (N = 851 Parameter SE Dvalue 66.35 9.42 .0001 -0.75 0.60 .22 -0.14 0.14 .32 4.71 2.48 .06 -0.49 0.21 .02 -0.16 0.09 .09 R 2= .14 Adj R2 = .09 1997 Data (N = 83) Parameter SE d value 69.17 7.01 .0001 0.17 0.50 .73 0.01 0.11 .96 -3.07 3.87 .43 -0.90 0.18 .0001 -0.04 0.08 .61 R2 = .27 Adj R2 = .22 001145 Intercept PFOS Age Alcohol BMI Cigarettes Table 9 (continued) Platelets - Both Locations Parameter 272.49 -3.37 -0.88 -4.80 -0.04 0.21 R2= .06 1995 Data SE 23.37 1.86 0.41 3.20 0.70 0.33 Adj R2= .03 t>value ,0001 .07 .04 .14 .95 .52 Parameter 260.92 -2.91 -0.25 2.37 -0.73 -0.30 R2= .03 1997 Data SE 25.27 2.73 0.48 4.26 0.89 0.41 Adj R2= .00 d value .0001 .29 .61 .58 .41 .46 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 294.81 -2.57 -0.35 -3.65 -2.23 0.82 R2= .06 Platelets, Antwerp Only 1995 Data SE 48.19 2.77 0.65 3.74 1.91 0.65 Adj R2= .00 d value .0001 .36 .60 .33 .25 .21 Parameter 258.21 -1.99 0.41 -0.87 -1.16 -0.70 R2= .02 1997 Data SE 66.58 6.44 0.91 5.42 2.65 0.90 Adj R2= .00 d value .0003 .76 .65 .87 .66 .44 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 353.90 -6.08 -1.99 5.36 -0.64 -0.35 R2= .19 Platelets. Decatur Onlv 1995 Data SE 39.04 2.50 0.59 10.14 0.84 0.40 Adj R2= .14 d value .0001 .02 .001 .60 .45 .38 Parameter 225.66 -2.75 -0.16 8.45 0.14 -0.19 R2 = .01 1997 Data SE 42.96 3.09 0.69 23.70 1.10 0.47 Adj R2= .00 D value .0001 .38 .82 .72 .90 .69 001146 Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter 0.5744 -0.0223 0.0033 0.0059 -0.0260 0.0086 -0.0005 R2 = .29 Table 9 (continued) Creatinine - Both Locations 1995 Data SE 0.0779 0.0146 0.0016 0.0013 0.0106 0.0023 0.0011 Adj R2= .27 Dvalue .0001 .13 .04 .0001 .02 .0003 .65 Parameter 0.9478 -0.0302 0.0055 0.0027 -0.0242 -0.0022 -0.0019 R2= .16 1997 Data SE 0.0602 0.0159 0.0022 0.0012 0.0101 0.0021 0.0010 Adj R2= .13 Dvalue .0001 .06 .01 0.02 .02 .30 .05 Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter 0.7319 -0.0324 0.0008 0.0044 -0.0055 0.0037 0.0017 R2= .20 Creatinine - Antwerp Onlv 1995 Data SE 0.1382 0.0195 0.0024 0.0019 0.0107 0.0055 0.0019 Adj R2= .14 D value .0001 .10 .74 .02 .61 .50 .36 Parameter 0.8600 -0.0453 0.0074 0.0033 -0.0142 -0.0012 0.0041 R2= .14 1997 Data SE 0.1430 0.0455 0.0099 0.0021 0.0113 0.0057 0.0019 Adj R2= .05 Dvalue .0001 .32 .46 .11 .21 .83 .04 Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes .003 Parameter 0.8641 -0.0093 0.0036 0.0032 0.0082 0.0032 -0.0030 R2 = .23 Creatinine - Decatur Onlv 1995 Data SE 0.1300 0.0206 0.0019 0.0020 0.0336 0.0028 0.0013 D value .0001 .65 .07 .11 .81 .26 .02 Parameter 1.1363 -0.0430 0.0067 0.0006 -0.1133 -0.0038 -0.0034 1997 Data SE t>value 0.1021 .0001 0.0201 .04 0.0025 .009 0.0016 .72 0.0559 .05 0.0026 .14 0.0011 Adj R2= .18 R2= .27 Adj R2= .22 001147 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 190.1 -0.99 0.74 -0.40 -0.04 -0.18 R2= .02 Table 9 (continued) Cholesterol - Both Locations 1995 Data SE 22.1 1.77 0.39 3.04 0.67 0.31 Adj R2= .00 p value .0001 .58 .06 .90 .95 .56 Parameter 155.5 4.66 1.39 7.50 -0.40 -0.08 R2= .17 1997 Data SE 19.0 2.04 0.36 3.18 0.67 0.30 Adj R2= .14 p value .0001 .02 .0002 .02 .55 .80 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 62.17 -1.22 2.60 1.13 2.28 0.81 R2 = .25 Cholesterol - Antwerp Only 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE__________p value 43.72 .16 93.11 49.20 .06 2.51 .63 2.37 4.74 .62 0.59 .0001 2.54 0.67 .0004 3.39 .74 7.49 3.98 .06 1.73 .19 0.78 1.95 .69 0.59 .17 -0.23 0.66 .72 Adj R2= .20 R2= .33 Adj R2 = .27 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 304.27 -1.03 -1.05 -21.16 -1.03 -0.40 R2= .14 Cholesterol - Decatur Only 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE__________p value 33.14 .0001 207.01 31.55 .0001 2.19 .64 4.26 2.27 .06 0.51 .04 0.48 0.51 .35 8.88 .02 19.54 17.41 .27 .73 .17 -.78 0.81 .34 0.33 .23 -0.11 .35 .75 Adj R2= .09 R2= .07 Adj R2= .01 001148 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 135.26 0.50 0.58 -3.33 -6.67 -0.52 R2= .05 Table 9 (continued) 1995 Data SE 20.27 1.61 0.36 2.78 0.61 0.28 Adj R2= .02 LDL Dvalue .0001 .76 .11 .23 .28 .07 Parameter 114.69 4.01 1.04 4.29 -1.09 -0.02 R2= .13 1997 Data SE 17.06 1.82 0.32 2.84 0.60 0.28 Adj R2= .10 D value .0001 .03 .002 .13 .07 .95 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 12.14 0.24 2.15 -3.04 2.04 0.35 R2 = .20 LDL - Antwerp Only 1995 Data SE 41.44 2.38 0.56 3.21 1.64 0.56 Adj R2= .15 p value .77 .92 .0003 .35 .22 .53 Parameter 40.88 1.50 2.22 4.12 0.48 0.26 R2 = .29 1997 Data SE 42.42 4.08 0.58 3.43 1.68 0.57 Adj R2= .23 p value .34 .71 .0003 .23 .78 .65 Intercept PFOS Age Alcohol BMI Cigarettes LDL - Decatur Only Parameter 218.24 1.06 -0.79 -22.95 -1.38 -0.65 R2 = .20 1995 Data SE 30.58 1.96 0.46 8.04 0.67 0.30 Adj R2= .15 p value .0001 .59 .09 .006 .04 .03 Parameter 160.50 3.55 0.20 -12.68 -1.27 -0.13 R2= .07 1997 Data SE 28.17 2.03 0.45 28.72 0.72 0.31 Adj R2= .01 p value .0001 .08 .65 .66 .08 .67 001149 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 48.01 -0.06 0.0005 0.10 -0.08 0.05 R2 = .08 Table 9 (continued) Hematocrit - Both Locations 1995 Data SE p value Parameter 1.36 .0001 46.73 0.11 ,57 -0.25 0.02 .98 -0.03 0.19 .58 0.15 0.04 .06 0.01 0.02 .007 0.04 Adj R2= .05 R2= .08 1997 Data SE 1.28 0.14 0.02 0.22 0.05 0.02 Adj R2 = .04 p value .0001 .07 .20 0.50 0.75 .04 Intercept PFOS Age Alcohol BMI Cigarettes Hematocrit - Antwerp Only Parameter 44.00 0.06 0.06 -0.12 0.02 0.05 R2 = .08 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE_________ p value 2.30 .0001 45.63 3.63 .0001 0.13 .63 -0.29 0.35 .42 0.03 .06 0.04 0.50 .37 0.18 .51 0.06 0.30 .84 0.09 .80 -0.02 0.14 .89 0.03 .14 0.01 0.05 .79 Adj R2 = .03 R2 = .01 Adj R2 = .00 Intercept PFOS Age Alcohol BMI Cigarettes Hematocrit - Decatur Onlv Parameter 45.56 -0.05 0.01 -0.56 -0.02 0.08 R2= .10 1995 Data SE 2.62 0.17 0.04 0.68 0.06 0.03 Adj R2= .04 p value .0001 .79 .89 .41 .71 .004 Parameter 45.69 -0.28 -0.05 -1.52 0.08 0.06 R2= .19 1997 Data SE 1.95 0.14 0.03 1.08 0.05 0.02 Adj R2= .13 p value .0001 .05 .09 .16 .11 .006 001150 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 15.78 -0.02 0.004 -0.02 -0.03 0.02 R2 = .06 Table 9 (continued) Hemoglobin - Both Locations 1995 Data SE 0.45 0.04 0.008 0.06 0.01 0.006 Adj R2= .03 p value .0001 .53 .57 .77 .06 .02 Parameter 15.43 -0.08 -0.006 0.01 0.008 0.02 R2 = .06 1997 Data SE 0.43 0.05 0.01 0.07 0.02 0.007 Adj R2= .03 d value .0001 .07 .46 .85 .59 .03 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 14.72 0.008 0.02 -0.04 -0.007 0.005 R2= .08 Hemoglobin - Antwerp Only 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE_________ p value 0.74 .0001 15.32 1.15 .0001 0.04 .86 -0.10 0.11 .40 0.01 .02 0.02 0.02 .33 0.06 .51 0.02 0.09 .81 0.03 .81 -0.01 0.05 .81 0.01 .61 -0.006 0.02 .70 Adj R2= .02 R2= .03 Adj R2 = .00 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 15.87 -0.03 -0.004 -0.22 -0.02 0.02 R2 = .10 Hemoglobin - Decatur Only 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE_________ p value 0.90 .0001 15.58 0.71 .0001 0.06 .63 -0.10 0.05 .07 0.01 .79 -0.02 0.01 .12 0.23 .35 -0.47 0.39 .23 0.02 .32 0.02 0.02 .25 0.009 .01 0.02 0.008 .004 Adj R2= .04 R2= .17 Adj R2 = .12 001151 Table 9 (continued) Intercept PFOS Age Alcohol BMI Cigarettes Parameter 0.2196 -0.0009 0.0006 0.0074 -0.0012 -0.0007 R2 = .06 Direct Bilirubin - Both Locations 1995 Data SE 0.0259 0.0021 0.0005 0.0036 0.0008 0.0004 Adj R2= .03 Dvalue .0001 .68 .17 .04 .13 .06 Parameter 0.1929 -0.0061 -0.0002 0.0106 -0.0017 -0.0009 R2 = .14 1997 Data SE 0.0264 0.0029 0.0005 0.0045 0.0009 0.0004 Adj R2= .11 d value .0001 .03 .69 .02 .07 .05 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 0.1794 -0.0034 0.00004 0.0102 0.0017 -0.0016 R2 = .17 Direct Bilirubin - Antwerp Onlv 1995 Data SE 0.0461 0.0026 0.0006 0.0036 0.0018 0.0006 Adj R2= .12 Dvalue .0002 .21 .95 .006 .36 0.009 Parameter 0.1963 -0.0077 -0.0017 0.0112 0.0006 -0.0017 R2= .14 1997 Data SE 0.0813 0.0079 0.0011 0.0066 0.0032 0.0011 Adj R2= .06 Dvalue .02 .33 .13 .10 .84 .12 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 0.1972 0.0025 0.0012 -0.0114 -0.0016 -0.0001 R2 = .07 Direct Bilirubin - Decatur Onlv 1995 Data SE 0.0478 0.0032 0.0007 0.0128 0.0011 0.0005 Adj R2= .02 d value .0001 .44 0.11 .37 .13 .79 Parameter 0.1205 -0.0038 0.0014 -0.0042 -0.0018 -0.0004 R2= .15 1997 Data SE 0.0347 0.0025 0.0006 0.0191 0.0009 0.0004 Adj R2= .10 Dvalue .0008 .13 .02 .83 .04 .27 001152 Table 10 Summaries of Simple Linear and Quadratic Regression Models for Total and Unconjugated Bilirubin, 1995 and 1997 1995 Data 1997 Data Both Locations Total Bilirubin Linear Ouadratic U n co n iu sa ted Linear Ouadratic Total Bilirubin Linear Ouadratic U n co n iu ea ted Linear Ouadratic R2 .002 Intercept 0.78247 PFOS(p value) -0.02927 (.047) PFOS2 (p value) .052 0.86532 -0.00918 (.03) 0.00846 (.02) .024 0.56837 -0.0285 (.04) .059 0.65239 -0.09941 (.002) 0.00858 (.01) .016 0.8175 -0.02583 (.12) .046 0.78963 -0.10456 (.01) 0.01200 (.03) .012 0.57745 -0.01940 (.18) .046 0.64247 -0.09138 (.01) 0.01097 (.03) Antwerp R 2 .018 Intercept 0.92001 PFOS (p value) -0.02927 (.21) PFOS2 (p value) .028 0.96548 -0.07872 (.18) 0.00697 (.31) .017 0.69447 -0.02604 (.23) .028 0.73919 -0.07468 (.17) 0.00686 (.32) .014 0.85170 -0.03647 (.35) .052 0.96350 -0.22847 (.08) 0.04620 (.20) .012 0.69161 -0.02872 (.39) .050 0.78708 -0.19268 (.09) 0.03945 (.13) Decatur R2 .004 Intercept 0.59712 PFOS (p value) -0.014363 (.53) PFOS2 (d value) .010 0.66974 -0.05867 (.09) 0.00518 (.12) .008 0.39803 -0.01166 (.39) .029 0.48279 -0.06966 (.03) 0.00605 (.05) 0.009 0.59609 -0.00894 (.53) .071 0.68287 -0.09408 (.02) 0.001143 (.02) .002 0.47268 -0.00489 (.69) .077 0.55440 -0.08508 (.01) 0.01076 (.01) 001153 Table 11 Mean Values of PFOS, Demographic, Serum Chemistry and Hematologic Parameters for Antwerp and Decatur Combined Locations, 1995 and 1997 Examinations for Employees Who Participated and Did Not Participate in Both Years Variable PFOS (ppm) 1995 Data Both Years Only 1995 IN = 61) IN = 117) 2.40 2.08 ______ 1997 Data Both Years Only 1997 (N = 61) (N = 88) 2.34*** 1.34 Age 39* 42 41* 38 BMI 27.1 26.3 27.3 26.9 Cigarettes 5.4 6.7 5.2 6.8 Alcohol 0.8 0.6 0.4* 0.4 Aik Phosphatase 87 86 79 80 GGT 44 44 37* 27 AST 29 27 26 26 ALT 46 46 33 32 Total bilirubin 0.71 0.72 0.65 0.69 Direct bilirubin 0.21 0.22 0.13 0.13 BUN 16.2 15.7 14.6 14.3 Creatinine 1.0 1.0 1.0 1.0 Glucose 86 87 93 89 Cholesterol 212 213 225*** 201 LDL 142 134 145*** 128 HDL 50 47 46 45 001154 Table 11 (continued) Variable 1995 Data Antwerp Decatur 1997 Data Antwerp Decatur Triglycerides 139 157 167 149 Hematocrit 45 47 46 46 Hemoglobin 15.0 15.5 15.3 15.4 RBC 4.9 5.0 5.0* 5.1 MCH 30.8 31.2 30.8* 30.2 MCHC MVC 33.2 33.1 93 94 33.6 92** 33.6 90 WBC 6.6 7.0 6.5 6.4 Platelets 226 227 228 224 * p < .05; * * p < .0 1 ; ***p<.001 001155 Table 12 Mean Values of PFOS, Demographic, Serum Chemistry and Hematologic Parameters for Antwerp and Decatur, 1995 Examinations for Employees Who Participated and Did not Participate in Both 1995 and 1997 Variable PFOS (ppm) Antwerp_______ Both Years Only 1995 (N = 27) (N = 61) 2.30 1.76 _______ Decatur Both Years Only 1995 (N = 34) (N = 56) 2.48 2.42 Age 34 37 42* 46 BMI 24.0 23.9 29.6 29.0 Cigarettes 5.4 4.3 5.4 9.4 Alcohol 1.7* 1.0 0.1 0.2 Aik Phosphatase 75 75 96 98 GGT 46 39 43 50 AST 28 25 30 29 ALT 43 45 48 47 Total bilirubin 0.91 0.84 0.54 0.59 Direct bilirubin 0.23 0.22 0.19 0.21 BUN 17.7 16.7 15.0 15.0 Creatinine 0.9 0.9 1.1 1.1 Glucose 78* 83 92 92 Cholesterol 209 216 232** 209 LDL 133 140 150** 128 HDL 56 53 46 42 001156 Table 12 (continued) Variable 1995 Data Antwerp Decatur 1997 Data Antwerp Decatur Triglycerides Hematocrit Hemoglobin 100 46 15.3 122 47 15.5 170 44** 14 8** 197 46 15.4 RBC 4.9 4.9 4.9* 5.0 MCH 31.2 31.5 30.5 30.8 MCHC 33.0 32.8 33.4 33.4 MVC 95 96 91 92 WBC 6.4 6.3 6.8* 7.8 Platelets 222 225 23.0 228 * p < .05; * * p < .0 1 ; ***p<.001 001157 Table 13 Mean Values of PFOS, Demographic, Serum Chemistry and Hematologic Parameters for Antwerp and Decatur, 1997 Examinations for Employees Who Participated and Did not Participate in Both Years Variable PFOS (ppm) Antwerp_______ Both Years pnly 1997 (N = 27) (N = 38) 2 33*** 0.88 _______ Decatur Both Years Only 1997 (N = 34) (N = 50) 2.35 1.69 Age 30 45 44 BMI 24.4* 22.8 29.5 29.9 Cigarettes 5.9 5.2 4.7 7.9 Alcohol 1.0 0.7 0.1 0.2 Aik Phosphatase 72 69 85 88 GGT 33* 21 41 32 AST 27 26 26 26 ALT 32 30 34 34 Total bilirubin 0.77 0.81 0.56 0.59 Direct bilirubin 0.14 0.16 0.12 0.11 BUN 15.2 14.6 14.1 14.0 Creatinine 0.9 0.9 1.0 1.0 Glucose 80 81 103 95 Cholesterol 224*** 192 226* 208 LDL 148 123 143 132 HDL 51 49 43 42 001158 Table 13 (continued) Variable 1995 Data Antwero Decatur 1997 Data Antwerp Decatur Triglycerides 125 101 201 185 Hematocrit 46 46 45 46 Hemoglobin 15.5 15.4 RBC 5.0 5.1 15.2 49** 15.4 5.1 MCH 30.8 30.4 30.9 30.0 MCHC MVC 33.4 33.3 92 91 33.7 33.8 91* 89 WBC 6.8 6.3 6.4 6.5 Platelets 241 234 218 217 * p < .05; ** p < .01; ***p<.001 001159 Table 14 Multivariable Regression of Serum Chemistries and Hematological Parameters - Examination of the Effect of PFOS Adjusting for Age, Alcohol, BMI and Cigarettes, Antwerp and Decatur Employees (N = 61) Who Participated in Both the 1995 and 1997 Examinations Intercept PFOS Age Alcohol BMI Cigarettes Parameter 55.23 -1.71 0.42 -5.34 0.70 0.82 R2= .24 1995 Data SE 22.22 1.77 0.43 2.47 0.67 0.34 Adj R2= .17 Alkaline Phospatase d value .02 .34 .33 .04 .30 .02 Parameter 45.74 0.46 0.33 -2.93 0.62 0.76 R2= .15 1997 Data SE 22.07 1.71 0.37 2.36 0.61 0.32 Adj R2= .07 o value .04 .79 .37 .22 .31 .02 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 2.5359 -0.0028 -0.0078 0.0795 0.0469 0.0023 R2= .14 1995 Data SE 0.5561 0.0443 0.0107 0.0619 0.0167 0.0084 Adj R2= .06 In GGT d value .0001 .95 .47 .20 .007 .79 Parameter 2.5660 -0.0357 -0.0095 0.0673 0.0463 -0.0025 II CM P5 1997 Data SE 0.5965 0.0461 0.0010 0.0638 0.0165 0.0087 Adj R2= .07 d value .0001 .44 .35 .30 .007 .77 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 15.27 0.40 0.17 1.01 0.24 -0.26 R2= .08 1995 Data SE 11.19 0.89 0.22 1.25 0.34 0.17 Adj R2= .00 SGOT d value .18 .65 .42 .42 .47 .13 Parameter 21.68 -0.22 0.001 0.47 0.17 0.07 R2= .03 1997 Data SE 6.80 0.53 0.11 .73 0.19 0.10 Adj R2= .00 d value .002 .67 .99 52 .38 .49 001160 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 26.32 -0.03 -0.20 0.13 1.05 -0.25 R2 = .10 Table 14 (continued) 1995 Data SE 17.83 1.42 0.34 1.98 0.54 0.27 Adj R2= .02 SGPT Dvalue ,15 .98 .57 .95 .06 .35 Parameter 15.34 0.27 -0.34 0.22 1.15 -0.04 R2= .18 1997 Data SE 13.86 1.07 0.23 1.48 0.38 0.20 Adj R2= .11 v value .27 .80 .14 .88 .004 .83 Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter 0.7086 -0.0874 0.0081 -0.0030 0.0344 -0.0319 -0.0130 R2 = .23 1995 Data SE 0.3971 0.0806 0.0068 0.0073 0.0436 0.0115 0.0060 Adj R2= .14 In Total Bilirubin Dvalue .08 .28 .24 .69 .43 .01 .03 Parameter -0.3739 0.0202 -0.0014 0.0072 0.1045 -0.0182 -0.0105 R2= .19 1997 Data SE 0.3996 0.0308 0.0026 0.0068 0.0426 0.0112 0.0058 Adj R2= .10 Dvalue .35 .51 .58 .29 .02 .11 .08 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 0.2579 -0.0027 -0.0019 0.0049 0.0010 -0.0005 R2 = .13 1995 Data SE 0.0469 0.0037 0.0009 0.0052 0.0014 0.0007 Adj R2= .05 Direct Bilirubin d value .0001 .47 .04 .35 .49 .50 Parameter 0.1711 -0.0026 0.0004 0.0089 -0.0021 -0.0008 R2= .14 1997 Data SE d value 0.0462 .0005 0.0036 .47 0.0008 .60 0.0049 .08 0.0013 .10 0.0007 .22 Adj R2= .07 001161 Table 14 (continued) Intercept PFOS Age Alcohol BMI Cigarettes Parameter 18.33 -0.09 -0.01 0.59 -0.04 -0.15 R2 = .16 1995 Data SE 3.62 0.29 0.07 0.40 0.11 0.05 Adj R2= .09 BUN p value .0001 .75 .87 .15 .71 .007 Parameter 14.06 0.29 0.07 0.18 -0.12 -0.03 R2= .08 1997 Data SE 3.09 0.24 0.05 0.33 0.09 0.04 Adj R2= .00 o value .0001 .22 .16 .59 .18 .55 Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter 0.5185 -0.0515 0.0068 0.0073 -0.0100 0.0099 -0.0003 R2 = .49 1995 Data SE 0.1308 0.0264 0.0022 0.0024 0.0143 0.0038 0.0019 Adj R2= .43 Creatinine p value .0002 .06 .004 .004 .49 .01 .89 Parameter 0.8785 0.0168 -0.0005 0.0036 -0.0266 -0.0030 -0.0028 R2= .18 1997 Data SE 0.1333 0.0103 0.0009 0.0023 0.0142 0.0037 0.0020 Adj R2= .09 Dvalue .0001 .11 .57 .12 .07 .43 .15 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 3.7479 -0.0017 0.0008 -0.0050 0.0236 0.0023 R2= .46 1995 Data SE 0.1311 0.0103 0.0025 0.0145 0.0039 0.0020 Adj R2= .41 In Glucose p value .0001 .87 .74 .73 .0001 .26 Parameter 3.4111 0.0036 0.0063 0.0005 0.0306 -0.0051 R2= .48 1997 Data SE 0.2013 0.0156 0.0034 0.0215 0.0056 0.0029 Adj R2 = .43 d value .0001 .82 .07 .98 .0001 .09 001162 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 184.37 -1.15 1.37 -2.72 -0.24 -0.82 R2 = .13 Table 14 (continued) 1995 Data SE 34.59 2.76 0.67 3.85 1.04 0.52 Adj R2= .05 Cholesterol p value .0001 ,68 .04 .48 .81 .12 Parameter 179.50 3.75 0.73 6.57 0.09 -0.07 R2= .09 1997 Data SE 38.87 3.01 0.65 4.16 1.07 0.57 Adj R2 = .01 p value .0001 .22 .27 .12 .94 .90 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 115.89 0.36 1.26 -3.74 -0.61 -0.72 R2 = .15 LDL 1995 Data__________ __________ 1997 Data SE_________ p value______ Parameter SE__________p value 30.35 .0004 150.88 34.90 .0001 2.38 .88 2.68 2.70 .33 0.58 .88 0.44 0.58 .45 3.35 .27 3.81 3.73 .31 0.91 .51 -1.20 0.96 .22 0.45 .12 -0.10 0.51 .85 Adj R2= .07 R2= .07 Adj R2 = .00 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 75.25 -0.68 -0.0005 2.31 -0.86 -0.30 R2 = .25 1995 Data SE 11.01 0.86 0.21 1.21 0.33 0.17 Adj R2 = .18 HDL p value .0001 .44 .99 .06 .01 .08 Parameter 68.22 -0.33 0.11 2.10 -0.97 -0.12 R2 = .34 1997 Data SE 8.91 0.69 0.15 0.95 0.25 0.13 Adj R2 = .28 p value .0001 .63 .47 .03 .0002 .37 001163 Intercept PFOS Age Alcohol BMI Cigarettes Parameter -89.52 -3.78 0.72 -1.93 7.50 1.40 R2 = .26 Table 14 (continued) 1995 Data SE 66.79 5.32 1.29 7.43 2.00 1.01 Adj R2= .19 Triglycerides D value .19 ,48 .58 .80 .0004 .17 Parameter -198.05 7.03 0.87 3.28 11.26 0.73 R2= .38 1997 Data SE 82.02 6.35 1.37 8.77 2.26 1.19 Adj R2= .32 Dvalue .02 .27 .53 .71 .0001 .54 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 50.87 -0.03 -0.05 0.09 -0.13 0.05 R2 = .20 1995 Data SE 2.15 0.17 0.04 0.24 0.06 0.04 Adj R2= .13 Hematocrit v>value .0001 .87 .19 .70 .04 .21 Parameter 46.61 -0.20 -0.04 0.31 0.02 0.06 r 2= .io 1997 Data SE 2.56 0.20 0.04 0.27 0.07 0.04 Adj R2= .02 d value .0001 .33 .38 .26 .77 .14 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 17.09 -0.02 -0.01 -0.002 -0.06 0.003 R2= .18 1995 Data SE 0.72 0.06 0.01 0.08 0.02 0.01 Adj R2= .10 Hemoglobin o value .0001 .70 .36 .98 .01 .79 Parameter 15.55 -0.07 -0.01 0.05 0.01 0.02 R2 = .07 1997 Data SE 0.88 0.07 0.01 0.09 0.02 0.01 Adj R2= .00 d value .0001 .33 .44 .61 .70 .12 001164 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 5.21 0.002 -0.003 -0.006 -0.007 -0.005 R2 = .06 Table 14 (continued) 1995 Data SE 0.22 0.018 0.004 0.025 0.007 0.004 Adj R2= .03 RBC d value .0001 ,89 .55 .81 .28 .18 Parameter 5.02 -0.003 -0.003 0.025 0.004 -0.005 R2= .04 1997 Data SE 0.30 0.024 0.005 0.033 0.008 0.004 Adj R2= .04 d value .0001 .88 .53 .44 .64 .28 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 32.97 -0.05 -0.01 0.04 -0.07 0.04 R2 = .18 1995 Data SE 1.25 0.10 0.02 0.14 0.04 0.02 Adj R2= .10 MCH d value .0001 .60 .67 .79 .07 .06 Parameter 31.13 -0.11 -0.004 -0.06 -0.007 0.07 R2 = .19 1997 Data SE 1.43 0.11 0.024 0.15 0.039 0.02 Adj R2= .12 o value .0001 .31 .85 .71 .86 .001 Intercept PFOS Age Alcohol BMI Cigarettes oo <NII U. Parameter 33.55 -0.03 0.01 -0.08 -0.02 -0.02 1995 Data SE 0.53 0.04 0.01 0.06 0.02 0.01 Adj R2= .10 MCHC d value .0001 .55 .20 .18 .17 .01 Parameter 33.40 -0.005 0.005 -0.10 0.001 0.003 R2= .09 1997 Data SE 0.50 0.039 0.008 0.05 0.014 0.007 Adj R2= .01 d value .0001 .89 .55 .07 .92 .69 001165 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 98.33 -0.09 -0.07 0.30 -0.15 0.19 R2= .27 Table 14 (continued) 1995 Data SE 3.59 0.29 0.07 0.40 0.11 0.06 Adj R2= .20 MCV d value .0001 ,76 .34 .45 .18 .003 Parameter 92.85 -0.28 -0.02 0.19 -0.03 0.20 R2= .22 1997 Data SE 3.82 0.30 0.06 0.41 0.11 0.06 Adj R2 = .15 d value .0001 .34 .76 .65 .78 .0008 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 295.55 -3.96 -0.84 -10.54 -0.54 -0.78 R2 = .19 1995 Data SE 37.09 2.96 0.71 4.13 1.12 0.63 Adj R2= .11 Platelets Dvalue .0001 .19 .24 .01 .63 .22 Parameter 392.12 -3.84 -2.31 -6.59 -1.85 -0.82 R2= .21 1997 Data SE 48.22 3.73 0.81 5.16 1.33 0.70 Adj R2 = .14 D value .0001 .31 .006 .21 .17 .25 Intercept PFOS Age Alcohol BMI Cigarettes Parameter 3.10 0.03 0.002 0.07 0.10 0.12 R2 = .43 1995 Data SE 1.19 0.09 0.023 0.13 0.04 0.02 Adj R2= .38 WBC v value .01 .77 .93 .62 .007 .0001 Parameter 4.46 -0.04 -0.007 0.21 0.06 0.11 R2= .40 1997 Data SE 1.36 0.11 0.02 0.15 0.04 0.02 Adj R2= .34 d value .002 .70 .75 .15 .09 .0001 001166 PFOS (ppm) 0-< 1 1-<3 3-<6 >= 6 Table 15 Employee Distribution as to Plant Location and Whether Hormones Were Measured, 1995 Total 45 91 35 7 178 Both Locations Hormones Measured Yes No 10 (22%) 35 (78%) 46 (51%) 45 (49%) 27 (77%) 8 (23%) 5(71%) 2 (29%) 88 90 _______________ By Location___________ Antwerp Decatur Hormones Measured Hormones Measured Yes No Yes No 9 (27%) 25 (73%) 1 (9%) 10(91%) 21 (66%) 11 (34%) 25 (43%) 34 (57%) 18 (95%) 1 (5%) 9 (56%) 7 (44%) 2 (67%) 1 (33%) 3 (75%) 1 (25%) 50 38 38 52 Table 16 Mean Values for PFOS, Demographic, Serum Chemistries and Hematology, by Plant Location and Whether Hormones Were Measured, 1995 Both Locations Hormones Measured Yes No Variable___________ N = 88 N = 90 PFOS 2.87*** 1.52 Age 38.4*** 42.7 Alcohol 1.0** 0.4 BMI 26.5 26.7 Cigarettes 4.3 BUN 16.3 15.5 Creatinine 0.97* 1.02 Glucose 86 88 AlkalinePhosphatase 86 86 Antwerp Hormones Measured Yes No N = 50 N = 38 2.69*** 0.92 35.0* 38.4 1.6** 0.8 23.6 24.4 6.7** 1.9 17.1 16.8 0.87*** 0.94 80 84 77 73 Decatur Hormones Measured Yes No N = 38 N = 52 3.10** 1.96 42.9 45.9 0.2 0.2 30.3 28.5 10.3 6.0 15.1 14.5 1.10 1.08 94 90 99 96 89TT00 00 * Both Locations Hormones Measured Yes No Variable___________ N = 88 N = 90 GGT 47 42 AST 29 27 ALT 47 45 Total Bilirubin 0.7 0.8 Direct Bilirubin 0.2 0.2 Cholesterol 217 215 LDL 136 Oo HDL 49 Triglycerides 158 cr Hematocrit 47 137 48 144 46 Hemoglobin 15.4 15.3 Table 16 (continued) Antwerp Hormones Measured Yes No N = 50 N = 38 45 37 27 24 44 45 0.8 0.9 0.2 0.2 210 220 133 144 52 55 124 104 47 47 15.4 15.4 Decatur Hormones Measured Yes No N = 38 N = 52 50 46 31 29 51 45 0.5 0.6 0.2 0.2 227 211 141 132 43 43 205 174 46 45 15.3 15.1 Variable Both Locations Hormones Measured Yes No N = 88 N = 90 RBC 5.0 4.9 MCH 31.1 31.0 MCHC 33.0* 33.3 MCV WBC 94 __ _** 7.3 93 6.5 Platelets 230 *p < .05; **p < .01 ;*** p < .001 223 001170 Table 16 (continued) Antwerp Hormones Measured Yes No N = 50 N = 38 4.9 4.9 31.5 31.3 32.9 32.8 96 95 6.8*** 5.7 233** 212 * O OO Decatur Hormones Measured Yes No N = 38 N = 52 5.0 4.9 30.5 30.8 33.2* 33.5 92 92 7.0 226 231 Table 17 Mean, Median (Med), Standard Deviation (SD) of Mean and Range of PFOS, Demographic, Serum Chemistries and Hematological Values for N = 88 Employees, Antwerp and Decatur Combined, who Had Hormone Measurements, 1995 PFOS (ppm)___ Mean Med SD Range V1 o 1 -<3 3-<6 >6 PFOS 0.681 0.75 0.21 0.37-0.90 1.961 1.96 0.62 1.00-2.90 4.151 3.97 0.85 3.00-5.80 8.671 8.50 2.85 6.06-12.83 F value = 121.3, p ==.0001 0-< 1 1 -<3 3-<6 >6 Age 32.7 32.5 6.3 21.0-43.0 38.7 39.0 8.4 25.0-58.0 39.0 39.0 7.6 26.0-54.0 43.0 42.0 7.8 37.0-56.0 F value = 2.4, p = .08 0-< 1 1 -<3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 Alcohol 1.4 1.1 1.2 0.0-3.6 0.6 0.2 0.8 0.0-3.6 1.4 0.5 2.0 0.0-6.0 0.9 0.0 1.3 0.0-2.9 F value = 2.1, P = -11 BMI 24.5 24.5 2.4 20.4-28.0 27.9 26.2 7.2 19.6-60.7 24.7 24.2 4.2 17.9-32.5 27.5 29.4 4.9 20.6-33.0 F value: = 2.1, p = . 10 001171 PFOS (ppm) 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0- < 1 1 - <3 3- <6 >6 0-< 1 1 -< 3 3-<6 >6 0- < 1 1- <3 3- < 6 >6 Table 17 (continued) Mean Med SD Range Cigarettes 6.0 o.o 10.5 7.7 0.0 11.2 11.6 12.0 11.8 0.6 0.0 1.3 F =1.8, p = .16 0.0-25.0 0.0-35.0 0.0-40.0 0.0-3.0 BUN 17.0 16.0 3.1 14.0-22.0 15.8 15.0 3.8 8.0-26.0 17.0 17.0 3.8 10.0-23.0 14.4 14.0 4.0 10.0-21.0 F value = 1.1, p = .35 Creatinine 1.0 1.0 0.1 0.8-1.1 1.0 0.9 0.2 0.7-1.6 0.9 0.9 0.1 0.7-1.2 1.1 1.2 0.4 0.6-1.6 F value = 2.1, P = .05 Glucose 82 82 7 70-93 89 84 29 62-260 81 82 12 66-112 87 83 16 71-105 F value = 0.8, p = .52 Alkaline Phosphatase 82 82 14 66-103 88 89 25 49-146 85 82 21 32-124 88 74 29 63-136 F value; = 0.3, P = .84 PFOS (DDm) 0-< l 1 -< 3 3-<6 >6 0-< l 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 Table 17 (continued) Mean Med SD Ranee GGT 38 28 27 23-111 53 41 49 2-293 39 39 15 21-80 48 49 19 32-79 F value = 0.9, p = .44 AST 31 25 23 17-96 30 27 14 14-90 25 24 6 13-37 31 30 4 26-37 F value =1.2, p = .31 ALT 52 48 21 36-108 48 43 24 25-183 42 41 7 30-59 54 57 20 29-82 F value =1.1, p = .36 Total Bilirubin 0.86 0.70 0.45 0.40-2.00 0.67 0.65 0.28 0.20-1.30 0.65 0.60 0.31 0.20-1.40 0.68 0.70 0.15 0.50-0.90 F value; = 1.2, p = .30 Direct Bilirubin 0.22 0.20 0.04 0.20-0.30 0.20 0.20 0.05 0.10-0.30 0.22 0.20 0.04 0.20-0.30 0.22 0.20 0.04 0.20-0.30 F value: = 0.6, p = .63 001173 PFOS (ppm) 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< l 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 Table 17 (continued) Mean Med SD Range ... Cholesterol 213 207 33 180-290 220 227 42 144-315 217 214 29 171-270 200 208 34 160-240 F value = 0.5, p = .69 LDL 129 128 22 106-177 136 146 39 65-228 139 135 29 84-190 129 130 29 95-172 F value = 0.3, p = .85 HDL 52 57 12 31-63 50 46 15 28-94 46 48 12 23-74 45 46 10 34-61 F value = 0.6, p = .64 Triglycerides 157 121 167 41-622 163 129 118 41-651 156 138 97 34-413 128 151 52 64-187 F value = 0.1, p = .94 Hematocrit 47 48 2 44-49 46 46 3 39-52 47 47 2 43-52 47 48 1 45-49 F value = 2.3, p = .08 001174 PFOS (ppm) 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 Table 17 (continued) Mean Med SD Ranee Hemoglobin 15.6 15.8 0.5 14.6-16.1 15.2 15.3 0.9 13.0-17.1 15.6 15.6 0.7 13.8-17.4 15.5 15.4 0.7 14.8-16.2 F value = 1.8, p = .16 RBC 4.9 5.0 0.3 4.3-5.2 4.9 4.9 0.3 4.3-5.7 5.0 5.0 0.2 4.6-5.4 5.0 5.2 0.7 4.0-5.7 F value = 0.4, p = .74 MCH 32.0 31.7 1.1 30.8-33.9 30.8 30.8 1.4 27.3-33.2 31.2 31.8 1.5 26.0-33.3 31.1 30.2 3.4 28.2-36.9 F value = 1.9, p = .14 MCHC 32.9 33.1 0.8 31.9-34.5 33.0 33.1 0.7 31.7-34.5 32.9 32.8 0.7 31.3-34.2 33.1 33.3 0.8 32.2-34.0 F value = 0.2, p = .93 MCV 97 95 5 92-106 93 93 4 83-101 95 95 5 81-104 94 91 12 85-115 F value>=1.9, p = .13 001175 PFOS (ppm) Table 17 (continued) Mean Med SD Ranee 0-< 1 1 -< 3 3-<6 >6 WBC 6.7 6.5 1.8 4.4-9.4 7.3 6.9 2.2 3.6-15.5 7.7 7.5 2.3 4.1-13.3 7.1 6.9 0.6 6.4-7.8 F value = 0.7, P = .59 0-< 1 1 -<3 3-<6 >6 Platelets 243 230 43 189-309 234 227 44 153-365 229 230 50 132-344 177 182 29 143-205 F value = 2.7, p = .05 1. Significantly different (p < .05) than the remaining three PFOS exposure categories. 2. Significantly different (p < .05) than the 0 - < 1 ppm PFOScategory. 3. Significantly different (p < .05) than the 1 - < 3 ppm PFOScategory. 4. Significantly different (p < .05) than the 3 - < 6 ppm PFOScategory. 5. Significantly different (p < .05) than the > 6ppm PFOS category. Sample sizes: PFOS Level Both Locations 0 - < 1 ppm 10 1 - < 3 ppm 46 3 - < 6 ppm 27 > 6 ppm _5 88 Antwerp 9 21 18 2 50 Decatur 1 25 9 _3 38 001176 Table 18 Multivariable Regression of Serum Chemistries and Hematological Parameters in Relation to PFOS Adjusting for Age, Alcohol, BMI and Cigarettes, Antwerp and Decatur Data Combined, For Those Employees Who Had Hormone Measurements in 1995 Alkaline Phosphatase Variable Intercept PFOS Age Alcohol BMI Cigarettes Parameter 64.57 0.002 0.13 -4.17 0.60 0.60 R2= .22 SE 15.23 1.11 0.30 1.75 0.39 0.20 Adj R2= .17 D value .001 .99 .66 .02 .13 .004 Variable Intercept PFOS Age Alcohol BMI Cigarettes ln GGT Parameter 1.9570 0.0077 0.0157 0.1043 0.0341 0.0096 R2 = .20 SE 0.4203 0.0301 0.0084 0.0484 0.0107 0.0056 Adj R2= .15 D value .0001 .80 .07 .03 .002 .09 Variable Intercept PFOS Age Alcohol BMI Cigarettes AST Parameter 16.13 -0.13 0.006 1.02 0.51 -0.20 R2= .08 SE 9.90 .72 0.20 1.14 0.25 0.13 Adj R2 = .02 d value .11 .86 .98 .37 .04 .13 001177 Variable Intercept PFOS Age Alcohol BMI Cigarettes Table 18 (continued) ALT Parameter 25.19 -0.12 -0.04 0.20 0.95 -0.20 R2= .10 SE 14.22 1.04 0.28 1.64 0.36 0.19 Adj R2 = .04 D value .08 .90 .89 .90 .01 .29 Variable Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes In Total Bilirubin Parameter 0.6133 -0.0885 0.0073 -0.0025 0.0235 -0.0270 -0.0162 R2= .39 SE 0.2726 0.0518 0.0048 0.0054 0.0309 0.0068 0.0036 Adj R2= .34 D value .03 .09 .13 .64 .45 .0002 .0001 Variable Intercept PFOS Age Alcohol BMI Cigarettes Direct Bilirubin Parameter 0.225 0.001 0.0007 0.004 -0.002 -0.0003 R2= .09 SE 0.0335 0.002 0.0007 0.004 0.0009 0.0004 Adj R2= .03 p value .0001 .64 .31 .26 0.05 .44 001178 Table 18 (continued) Variable Intercept PFOS Age Alcohol BMI Cigarettes BUN Parameter 18.73 . -0.10 -0.05 0.37 -0.003 -0.05 R2= .08 SE 2.72 0.20 0.05 0.31 0.07 0.04 Adj R2= .02 D value .0001 .60 .35 .24 .97 .14 Variable Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Variable Intercept PFOS Age Alcohol BMI Cigarettes Creatinine Parameter 0.5031 -0.0488 0.0051 0.0086 -0.0119 0.0090 -0.0017 R2 = .37 SE 0.1178 0.0223 0.0020 0.0023 0.0133 0.0029 0.0016 Adj R2= .32 p value .0001 .03 .01 .0004 .38 .003 .28 In Glucose Parameter 3.8241 -0.0029 0.0051 -0.0052 0.0166 -0.0018 R2= .42 SE 0.1090 0.0079 0.0022 0.0125 0.0028 0.0014 Adj R2= .38 Dvalue .0001 .71 .02 .68 .0001 .22 001179 Table 18 (continued) Variable Intercept PFOS Age Alcohol BMI Cigarettes Cholesterol Parameter 198.04 . -2.83 1.26 -1.32 -0.73 -0.07 R2= .09 SE 26.88 1.97 0.54 3.09 0.68 0.35 Adj R2= .03 D value .0001 .15 .02 .67 .29 .85 Variable Intercept PFOS Age Alcohol BMI Cigarettes LDL Parameter 152.37 -0.45 0.70 -5.58 -1.19 -0.59 R2 = .14 SE 23.90 1.74 0.48 2.75 0.61 0.32 Adj R2 = .08 D value .0001 .80 .14 .05 .05 .07 Variable Intercept PFOS Age Alcohol BMI Cigarettes HDL Parameter 77.68 -0.66 -0.31 2.68 -0.66 -0.07 R2= .32 SE 8.51 0.62 0.17 0.98 0.22 0.11 Adj R2= .28 Dvalue .0001 .29 .07 .008 .003 .56 001180 Variable Intercept PFOS Age Alcohol BMI Cigarettes Table 18 (continued) Triglycerides Parameter -178.52 -7.89 , 4.22 9.90 6.07 3.17 R2= .32 SE 71.97 5.26 1.44 8.28 1.83 0.95 Adj R2= .28 Dvalue .02 .14 .005 .24 .001 .001 Variable Intercept PFOS Age Alcohol BMI Cigarettes Hematocrit Parameter 48.03 -0.05 0.02 -0.003 -0.10 0.04 R2= .12 SE 1.66 0.12 0.03 0.19 0.04 0.02 Adj R2= .06 Dvalue .0001 .70 .48 .99 .02 .05 Variable Intercept PFOS Age Alcohol BMI Cigarettes Hemoglobin Parameter 15.76 -0.02 -0.04 0.01 -0.03 0.009 R2 = .09 SE 0.58 0.04 0.07 0.01 0.01 0.008 d value .0001 .67 .28 .58 .03 .25 Adj R2= .03 001181 Variable Intercept PFOS Age Alcohol BMI Cigarettes Table 18 (continued) MCH Parameter 33.52 0.04 -0.006 0.11 -0.10 0.03 R2= .24 SE Dvalue 1.05 .0001 0.08 .59 0.02 .79 0.12 .35 0.03 .0003 0.01 .03 Adj R2= .20 Variable Intercept PFOS Age Alcohol BMI Cigarettes MCHC Parameter 32.87 -0.001 0.008 -0.08 -0.0006 -0.01 R2= .07 SE 0.50 .037 0.01 0.06 0.01 0.007 Adj R2= .02 p value .0001 .98 .42 .19 .96 0.08 Variable Intercept PFOS Age Alcohol BMI Cigarettes MCV Parameter 101.97 0.14 -0.04 0.56 -0.31 0.12 R2= .28 SE 3.36 0.25 0.07 0.39 0.09 0.04 Adj R2= .23 D value .0001 .56 .53 .15 .0005 0.007 001182 Table 18 (continued) Variable Intercept PFOS Age Alcohol BMI Cigarettes RBC Parameter 4.673 . -0.009 0.005 -0.026 0.006 -0.002 R2= .08 SE 0.214 0.016 0.004 0.025 0.005 0.003 Adj R2= .02 D value .0001 .58 .28 .29 .24 .48 Variable Intercept PFOS Age Alcohol BMI Cigarettes Variable Intercept PFOS Age Alcohol BMI Cigarettes Platelets Parameter 320.83 -6.73 -0.94 -6.95 -1.26 0.43 R2= .18 SE 32.21 2.36 0.64 3.70 0.82 0.43 Adj R2= .13 o value .0001 .006 .15 .06 .13 .32 WBC Parameter 3.24 -0.03 0.07 0.03 0.03 0.09 R2= .35 SE 1.32 0.10 0.03 0.15 0.03 0.02 Adj R2= .31 D value .02 .74 .01 .82 .42 .0001 001183 Table 19 Mean, Median (Med), Standard Deviation (SD) of Mean and Range of PFOS, Hormonal Measurements for N = 88 Employees, Antwerp and Decatur Combined, 1995 PFOS (Dpm) Mean Med SD Ranee 0-< 1 1 -< 3 3-<6 >6 Cortisol 19.3 17.5 7.0-29.0 17.7 18.0 7.2 1.0-42.0 21.4 23.0 7.2 7.0-31.0 17.0 19.0 6.2 9.0-23.0 F = 1.7, p = .18 0-< 1 1 -< 3 3-<6 >6 DHEA-S 3883'5 358 168 2342'4 210 95 3163 318 106 1942 190 19 F = 8.3, p = .0001 88-605 69-460 90-530 176-215 0-< 1 1 -< 3 3-<6 >6 Estradiol 67.1 67.0 12.8 60.3 59.0 15.2 60.5 61.0 10.1 64.8 65.0 18.7 F = 0.8, p = .49 50.0-87.0 35.0-101.0 42.0-81.0 47.0-92.0 0-< 1 1 -< 3 3-<6 >6 FSH 3.8 3.5 1.5 2.0-6.0 5.6 4.0 4.2 1.0-26.0 5.6 4.0 3.9 2.0-18.0 6.6 6.0 3.4 3.0-12.0 F value = 0.8, p = 0.48 0-< 1 1 -< 3 3-<6 >6 17-Hvdroxvprogesterone 170 164 41 121-245 131 123 52 53-294 150 153 52 65-245 119 98 4590-197 F value = 2.4, CO II a, 001184 PFOS (dditi) 0-< 1 1 -< 3 3-<6 >6 0-< 1 1-<3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 0-< 1 1 -< 3 3-<6 >6 Table 19 (continued) Mean Med SD Range LH 3.8 4.0 0.9 2.0-5.0 4.6 4.0 3.0 1.0-21.0 4.6 5.0 1.9 2.0-9.0 4.8 5.0 1.3 3.0-6.0 F value = 0.3, p = .81 Prolactin 13.5 13.0 7.0 6.0-29.0 11.8 11.0 5.0 3.0-30.0 13.4 10.0 7.9 5.0-39.0 13.6 10.0 6.3 9.0-24.0 F value = 0.5, p = .67 SHBG 0.9 0.9 0.3 1.0 0.9 0.4 1.0 0.9 0.3 1.2 1.3 0.6 F = 1.1, p = .35 0.5-1.3 0.4-1.9 0.4-1.7 0.6-2.1 Free testosterone 20.53 20.2 5.2 10.2-28.2 16.22 16.1 3.3 8.9-27.1 17.7 18.2 3.2 12.2-25.2 17.5 17.7 2.0 15.3-20.5 F value: = 4.5, p = .006 Bound testosterone 7393 757 175 5281094 5802,4 589 n o 278-762 6763 659 171 410-1039 711 752 160 462-883 F value:= 5.2, p = .003 001185 Table 19 (continued) PFOS (ppm) Mean Med SD Ranee 0-< 1 1 -< 3 3-<6 >6 TSH 1.9 1.4 1.3 0.6-4.5 1.0 1.5 1.4 0.5-8.1 1.5 1.4 0.8 0.5-3.4 2.0 1.6 1.2 0.7-3.8 F value = 0.6, P = .62 Sample sizes: PFOS Level Both Locations 0 - < 1 ppm 10 1 - < 3 ppm 46 3 - < 6 ppm 27 > 6 ppm _5 88 Antwerp 9 21 18 2 50 Decatur 1 25 9 3. 38 001186 Table 20 Multivariable Regression Analysis of Hormones in Relation to PFOS Adjusting for Age, Alcohol, BMI and Cigarettes, Antwerp and Decatur Data Combined, 1995 Variable Intercept PFOS Age Alcohol BMI Cigarettes Cortisol Parameter 27.13 0.27 -0.10 1.56 -0.25 0.01 R2= .22 SE 4.84 0.35 0.10 0.56 0.12 0.06 p value .0001 .45 .31 .006 .04 .87 > & & N> II Variable Intercept PFOS Age Alcohol BMI Cigarettes DHEAS Parameter 499.17 -2.95 -5.30 17.88 -1.82 2.25 R2= .28 SE 76.98 5.63 1.54 8.84 1.95 1.01 Adj R2= .24 p value .0001 .60 .0009 .05 .35 .03 Variable Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Estradiol Parameter 53.56 -3.82 0.44 -0.13 2.42 0.57 0.07 R2= .18 SE 9.76 1.85 0.17 0.19 1.10 0.24 0.13 Adj R2= .12 p value .0001 .04 .01 .51 .03 .02 .58 001187 Table 20 (continued) Estradiol (without employee C) Variable Intercept PFOS PFOS2 Age Alcohol BMI Cigarettes Parameter 53.43 -3.56 , 0.40 -0.13 2.43 0.56 0.69 R2= .13 SE 9.84 2.39 0.28 0.19 1.11 0.25 0.13 Adj R2= .06 D value .0001 .14 .15 .51 .03 .03 .60 Variable Intercept PFOS Age Alcohol BMI Cigarettes FSH Parameter 3.64 0.02 0.14 0.04 -0.13 0.005 R2= .10 SE 2.79 0.20 0.06 0.32 0.07 0.04 Adj R2= .05 Dvalue .20 .91 .02 .89 .07 .90 17-Hvdroxvorogesterone Variable Intercept PFOS Age Alcohol BMI Cigarettes Parameter 306.10 -0.04 -2.13 -0.76 -3.47 0.79 R2 = .33 SE 32.69 2.36 0.67 3.68 0.82 0.42 Adj R2= .28 p value .0001 .99 .002 .84 .0001 .07 001188 Variable Intercept PFOS Age Alcohol BMI Cigarettes Variable Intercept PFOS Age Alcohol BMI Cigarettes Variable Intercept PFOS Age Alcohol BMI Cigarettes Table 20 (continued) LH Parameter 5.65 0.05 -0.02 -0.24 -0.03 0.03 R2= .04 SE 1.86 0.13 0.04 0.21 0.05 0.02 Adj R2 = .02 d value .003 .69 .69 .26 .56 .21 Prolactin Parameter 16.67 0.34 -0.15 1.75 0.004 -0.14 R2 = .31 SE 3.99 0.29 0.08 0.46 0.10 0.05 Adj R2= .27 d value .0001 .25 .07 .0003 .97 .01 SHBG Parameter 0.03 0.005 0.02 -0.005 0.005 0.004 R2= .28 SE 0.24 0.02 0.005 0.03 0.006 0.003 Adj R2= .23 Dvalue .89 .77 .0001 .86 .40 .19 001189 Variable Intercept PFOS Age Alcohol BMI Cigarettes Variable Intercept PFOS Age Alcohol BMI Cigarettes Table 20 (continued) Free Testosterone Parameter 27.37 0.02 -0.14 -0.11 -0.19 0.01 R2= .19 SE 2.57 0.19 0.05 0.30 0.06 0.03 Adj R2= .14 D value .0001 .90 .01 .69 .005 .68 Bound Testosterone Parameter 993.88 6.91 -2.43 7.41 -11.43 1.22 R2= .28 SE 100.93 7.28 2.06 11.37 2.52 1.31 Adj R2= .23 p value .0001 .35 .24 .52 .0001 .35 Variable Intercept PFOS Age Alcohol BMI Cigarettes TSH Parameter -0.873 0.004 0.003 0.078 0.096 -0.011 R2= .23 SE 0.826 0.060 0.016 0.095 0.021 0.011 Adj R2= .18 v value .29 .95 .85 .41 .0001 .32 001190 FIGURE I S c a t t e r P lo t o f E s t r a d io l and PFOS, B oth L o c a tio n s C om bined, 1 9 9 5 Estradiol 8y PFOS H0 000 100 000 90 000 - 80.000 70.000 - 60.000 50.000 - 40.000 - 30.000 ------- 1----------1--------- 1----------1--------- 1----------J------- OOO 2.00 0 4.000 6.000 8.000 10.000 12.000 14.000 PFOS Estradiol Polynomial Ft degree-2 Linear Fit Polynomial Rt degree-2 Estradiol - 66.4703 - 3.60448 PFOS + 0.42197 PFOSA2 Summary of Frt R S q u are RSquare Adj R oot M ean Square Error M ean of Response O bservations (or Sum W gts) 0 .0 7 6 1 6 0.054423 13.33314 61.36364 68 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 1245.701 622.850 3 .5 0 3 6 85 15110.663 177.773 Prob>F 87 16356.364 0 .0 3 4 5 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 66.470276 -3.604477 0.4219692 Std Error 3.632698 1.7997 0.167705 t Ratio 18.30 -2.00 2.52 Prob>[t| <.0001 0 .0 4 8 4 0 .0 1 3 7 001191 Estradiol - 59.7222 + 0 57198 PFOS Summary o( Fit RSquara RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.00735 -0.00419 13.74017 61.36364 88 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 1 120.226 120.226 0.6368 86 16236.138 188.792 Prob>F 87 16356.364 0.4271 Term Intercept PFOS Parameter Estimates Estimate 59.722236 0.5719843 Std Error 2.525093 0.716766 t Ratio 23.65 0.80 Prob>[t| <.0001 0.4271 001192 FIGURE 2 S c a t t e r P lo t o f E s t r a d io l and PFOS, B oth L o c a tio n s C om in b ed , 1 9 9 5 Estradiol By P F O S W ith o u t E m ployee A Estradiol PFOS Polynomial Fit degree-2 Linear Fit Polynomial Fit degree-2 Estradiol - 65.8629 - 3.10569 PFOS + 0.34755 PFOSA2 Summary of Fit R S q u are RSquare Adj Root M ean Square Error M ean of Response Observations (or Sum Wgts) 0.020552 -0.00277 13.40325 61.01149 87 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 316.639 158.320 0.8813 84 15090.349 179.647 Prob>F 86 15406.989 0.4180 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 65.862871 -3.105693 0.3475472 Std Error 4.074112 2.339491 0.278213 t Ratio 16.17 -1.33 1.25 Prob>|t| <.0001 0.1879 0.2151 001193 linear Fit Estradiol - 62 0271 - 0 36861 PFOS Summary of Fit RSquare RSquare Adj Root M ean Square Error M ean of Response Observations (or Sum Wgts) 0 002356 -0.00938 13.44737 61.01149 87 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 36.295 36.295 0 .2 0 0 7 85 15370.694 180.832 Prob>F 86 15406.989 0 .6 5 5 3 Term Intercept PFOS Parameter Estimates Estimate 62.027069 -0.368607 Std Error 2 .6 8 6 4 8 8 0 .8 2 2 7 6 8 t Ratio 23.09 -0.45 Pfob>|t| <.0001 0.6553 001194 APPENDIX A Total Bilirubin and PFOS Scatter Plots DRAFT 001195 Scattec Plot of Total Bilirubin and PFOS, Both Locations Combined/ 1995 Total Blirubin Linear Frt Polynomial Fit degree-2 Linear Fit Total Bilirubin - 0.78247 - 0.02927 PFOS Summary of Frt RSquare R S q u areA d j Root M ean Square Error Mean of Response Observations (or Sum Wgts) 0.022279 0.016692 0.363488 0.718644 177 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 1 0.526856 0.526856 3 .9 8 7 6 175 23.121618 0.132124 Prob>F 176 23.648475 0.0474 Term Intercept PFOS Parameter Estimates Estimate 0.7824732 -0.029269 Std Error 0.04205 0.014657 t Ratio 18.61 -2.00 Prob>|t| <.0001 0.0474 Polynomial Fit degree-2 Total Bilirubin - 0.86532 - 0.09918 PFOS + 0.00846 PFOSA2 Summary of Fit 001196 RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0 051674 0.040773 0 359009 0.718644 177 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 1.222006 0.611003 4.7406 174 22.426468 0.128888 Prob>F 176 23.648475 0.0099 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 0.865318 -0.099182 0.008459 Std Error 0.054748 0.033404 0.003642 t Ratio 15.81 -2.97 2 .3 2 Prob>|t| <.0001 0 .0 0 3 4 0 .0 2 1 4 001197 S c a t t e c P l o t o f T o ta l B i l i r u b i n and PFOS, A n tw erp , L 995 Total 8ilirubm By PFOS Total Birubin Linear Fit Polynomial Fit degree-2 Linear Fit Total Bifirubin - 0.92001 -0 .0 2 9 2 7 PFOS Summary of Fit RSquare RSquare Adj Root M ean Square Error Mean of Response Observations (or Sum Wgts) 0.017786 0.006365 0.407279 0.863636 88 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.258313 0.258313 1.5573 86 14.265323 0.165876 Prob>F 87 14.523636 0.2155 Term Intercept PFOS Parameter Estimates Estimate 0.9200069 -0.029266 Std Error 0.062654 0.023452 t Ratio 14.68 -1.25 Prob>|t| <.0001 0.2155 Polynomial Fit degree-2 Total Bilirubin - 0.96548 - 0.07872 PFOS + 0.00697 PFOSA2 Summary of Fit 001198 R Square RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0 027505 0 004623 0 407635 0.863636 88 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 0.399470 0.199735 1.2020 85 14.124166 0.166167 Prob>F 87 14.523636 0.3056 Term Intercept PFOS PFOSA2 Parameter Estimates Estimate 0.9654789 -0.078719 0.0069749 Std Error 0.07979 0.058565 0.007568 t Ratio 12.10 -1.34 0.92 Prob>lt| <.0001 0.1825 0.3593 001199 Scatter Plot of Total Bilirubin and PFOS, Deeatuc, 1995 Total Birubin Linear Fit Polynomial Fit degree-2 Linear Fit Total Bilirubin - 0.59712 - 0.00898 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.004472 -0.00697 0.249626 0.575281 89 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 1 0.0243540 0.024354 0.3908 87 5.4212639 0.062313 Prob>F 88 5.4456180 0.5335 Term Intercept PFOS Parameter Estimates Estimate 0.5971232 -0.008979 Std Error 0.043827 0.014363 t Ratio 13.62 -0.63 Prob>|t| <.0001 0.5335 Polynomial Fit degree-2 Total Bilirubin - 0.66974 - 0.05867 PFOS + 0.00518 PFOSA2 Summary of Fit RSquare 0.032545 001200 RSquare Adj Root Mean Square Error Mean of Response Observations (or Sum Wgts) 0 010046 0 247508 0 575281 89 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 0.1772262 0.088613 1.4465 86 5.2683918 0.061260 Prob>F 88 5.4456180 0.2411 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 0.6697418 0.058672 0 .0 0 5 1 8 3 3 Std Error 0 .0 6 3 2 5 8 0.034531 0.003281 t Ratio 10.59 -1.70 1.58 Prob>|t| <.0001 0.0929 0.1178 001201 Scattar PLot of Total Bilirubin and PFOS, Both Locations Combined 1997 Linear Fit Polynomial Fit degree-2 Linear Fit Total Bilirubin - 0.7185 - 0.02583 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.016221 0.009483 0.322046 0.672973 148 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.249674 0.249674 2.4073 146 15.142218 0.103714 Prob>F 147 15.391892 0.1229 Term Intercept PFOS Parameter Estimates Estimate 0.7184974 -0.025825 Std Error 0.039518 0.016645 t Ratio 18.18 -1.55 Prob>|t| <.0001 0.1229 Polynomial F t degree-2 Total Bilirubin - 0.78963 - 0.10456 PFOS + 0.012 PFOSA2 Summary of Fit RSquare RSquare Ad] 0.045789 0.032628 001202 Root Mean Square Error Mean of Response Observations (or Sum Wgts) 0 018261 0 672973 148 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 0.704787 0.352393 3.4790 145 14.687105 0.101290 Prob>F 147 15.391892 0.0334 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 0.7896264 -0.104564 0.0120049 Std Error 0 .0 5 1 4 9 0 .0 4 0 6 2 5 0 .0 0 5 6 6 3 t Ratio 15.34 -2.57 2.12 Prob>|t| <.0001 0.0111 0 .0 3 5 7 001203 Scattec Plot of Total BiLirubin and PFOS, Antwerp, 1997 Linear Fit Polynomial Fit degree-2 Linear Fit Total Biirubin - 0.8517 - 0.03647 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.014291 -0.00161 0.385334 0.796875 64 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.1334670 0.133467 0.8989 62 9.2059080 0.148482 Prob>F 63 9.3393750 0.3468 Term Intercept PFOS Parameter Estimates Estimate 0.8516999 -0.036474 Std Error 0.075259 0.038471 t Ratio 11.32 -0.95 Prob>|t| <.0001 0.3468 Polynomial F t degree-2 Total Bilirubin - 0.9635 - 0.22847 PFOS + 0.0462 PFOSA2 Summary of F t RSquare RSquare Adj 0.051848 0.020761 001204 Root Moan Square Error Mean of Response Observations (or Sum Wgts) 0.318261 0 672973 148 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 0.704787 0.352393 3.4790 145 14.687105 0.101290 Prob>F 147 15.391892 0.0334 Term intercept PFOS P F O S *2 Parameter Estimates Estimate 0.7896264 0.104564 0.0120049 Std Error 0 .0 5 1 4 9 0 .0 4 0 6 2 5 0.005663 t Ratio 15.34 -2.57 2 .1 2 Prob>(t| <.0001 0.0111 0.0357 001205 Scatter Plot of Total Bilirubin and PFQS, Antwerp/ 1997 ------ Linear Fit ------ Polynomial Fit degree-2 Linear Fit Total Bifiubin - 0.8517 - 0.03647 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.014291 -0.00161 0.385334 0.796875 64 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.1334670 0.133467 0 .8 9 8 9 62 9.2059080 0.148482 Prob>F 63 9.3393750 0.3468 Term Intercept PFOS Parameter Estimates Estimate 0.8516999 -0.036474 Std Error 0 .0 7 5 2 5 9 0.038471 t Ratio 11.32 0.95 Prob>|t| <.0001 0 .3 4 6 8 Polynomial Fit degree-2 Total Bilirubin - 0.9635 - 0.22847 PFOS + 0.0462 PFOSA2 Summary of Fit RSquare RSquare Adj 0.051848 0.020761 001206 Root Moan Square Error Moan of Rospansa Observations (or Sum Wgts) 0 381007 0 796875 64 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 0.4842255 0.242113 1.6678 61 8.8551495 0.145166 Prob>F 63 9.3393750 0.1971 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 0.963501 -0.228466 0.0461951 Std Error 0.103492 0 .1 2 9 2 3 8 0.029718 t Ratio 9.31 -1 .7 7 1.55 Prob>|t| <.0001 0.0821 0 .1 2 5 3 001207 Scatter PLot of Total Bilirubin and PFQS, Decatur, 1997 Linear Fit Polynomial Fit degree-2 Linear Fit Total Bilirubin - 0.59609 - 0.00894 PFOS Summary of Fit RSquare RSquare Adj Root M ean Square Error M ean of Response Observations (or Sum Wgts) 0.004924 -0.00721 0.229 0.578571 84 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.0212775 0.021278 0 .4 0 5 7 82 4.3001510 0.052441 Prob>F 83 4.3214286 0.5259 Term Intercept PFOS Parameter Estimates Estimate 0.5960926 -0.008937 Std Error 0.037161 0.01403 t Ratio 16.04 -0.64 Prob>|t| <.0001 0.5259 Polynomial Fit degree-2 Total Bilirubin - 0.68287 - 0.09408 PFOS + 0.01143 PFOSA2 Summary of Fit R S quare RSquare Adj 0.071405 0.048477 001208 Root Mean Square Error Mean of Response Observations (or Sum Wgts) 0.222579 0.578571 84 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 0.3085734 0.154287 3.1143 81 4.0128552 0.049541 Prob>F 83 4.3214286 0.0498 Term Intercept p ro s P F O S A2 Parameter Estimates Estimate 0.6828676 -0.094082 0.011426 Std Error 0 .0 5 1 0 2 0 .0 3 7 8 9 6 0 .0 0 4 7 4 5 t Ratio 13.38 -2.48 2.41 Prob>|t| <.0001 0.0151 0.0183 001209 FIGURE L S catter Plot of E strad io l and PFOS, Both Locations Combined, 1995 Gsir.jdiol 8y PFOS PFOS Polynomial R t degree-2 Linear Fit Polynomial R t degre e d Estradiol - 66.4703 - 3.60448 PFO S + 0.42197 PFOSA2 Summary of Fit R S q u are RSquare Adj R oot M ean Square Error M ean of Response O bservations (or Sum Wflte) 0 .0 7 6 1 6 0.054423 13.33314 6 1 .3 6 3 6 4 88 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 1245.701 622.850 3 .5 0 3 6 85 15110.663 177.773 Prob>F 8 7 16356.364 0 .0 3 4 5 Term Intercept PFOS PFOSA2 Parameter Estimates Estimate 66.470276 -3.604477 0 .4 2 1 9 6 9 2 Std Error 3 .6 3 2 6 9 8 1.7997 0 .1 6 7 7 0 5 t Ratio 18.30 -2.00 2.52 Prob>|t| <.0001 0 .0 4 8 4 0 .0 1 3 7 001210 Estradiol - 597222 * 0 57198 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.00735 -0.00419 13.74017 61.36364 86 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 120.226 120.226 0 .6 3 6 8 86 16236.138 188.792 Prob>F 87 16356.364 0.4271 Term Intercept PFOS Parameter Estimates Estimate 59.722236 0.5719843 Std Error 2 .5 2 5 0 9 3 0 .7 1 6 7 6 6 t Ratio 23.65 0.80 Prob>|t| <.0001 0.4271 001211 FEGURE 2 Scatter Plot o f E strad io l and PFOS, Both Locations Cominbed, 1995 Estradiol By PFOS Without Employee A PFOS ------ Polynomial Ft degree-2 ------ Linear Fit Polynomial Fit degree-2 E strad o ! - 65.8629 -3 .1 0 5 6 8 PFO S + 0.34755 PFOSA2 Summary of Fit R S q u are RSquare Ad] R oot M ean Square Error M ean of Response Observations (or Sum Wgts) 0 .0 2 0 5 5 2 -0 .0 0 2 7 7 13.40325 6 1 .0 1 1 4 9 87 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 316.639 158.320 0.8813 84 15090.349 179.647 Prob>F 86 15406.989 0.4180 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 65.862871 3.105693 0 .3 4 7 5 4 7 2 Std Error 4 .0 7 4 1 1 2 2.339491 0.278213 t Ratio 16.17 -1.33 1.25 Prob>|t| <.0001 0 .1 8 7 9 0.2151 001212 linear F'l Estradiol - 62 0 2 / 1 - 0 0 6 3 6 1 PFQS Summary oI Fit RSquare RSquare Adj Root Mean Square Error Maan of Response Observations (or Sum Wgts) 0 002356 -0.00933 13.44737 61.01149 87 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 1 3 6 .2 9 5 36.295 0 .2 0 0 7 85 15370.694 180.832 Prob>F 86 15406.989 0 .6 5 5 3 Teem Intercept PFOS Parameter Estimates Estimate 62.027069 *0.368607 Std Error 2 .6 8 6 4 8 8 0 .8 2 2 7 6 8 t Ratio 23.09 -0.45 Prob>|t| <.0001 0 .6 5 5 3 001213 APPENDIX Unconjugated Bilirubin and PFOS Scatter Plots DRAFT 001214 S catter Plot of Unconjugated B iliru b in and PFOS/ Both Locations Combined/ Unconjugated bilirubin By PFOS 1995 Unconjugated bifrubin Linear Fit Polynomial Fit degree-2 Linear Fit Unconjugated bilirubin - 0.56837 - 0.0285 PFOS Summary ot Fit RSquare RSquare Adj Root M ean Square Error M ean of Response Observations (or Sum Wgts) 0.024418 0.018843 0.337753 0.506215 177 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.499660 0.499660 4 .3 8 0 0 175 19.963503 0.114077 Prob>F 176 20.463164 0.0378 Term Intercept PFOS Parameter Estimates Estimate 0.5683746 -0.028503 Std Error 0.039072 0.013619 t Ratio 14.55 -2.09 Prob>|t| <.0001 0.0378 Polynomial Fit degree-2 Unconjugated bilirubin - 0.65239 - 0.09941 PFOS 0.00858 PFOSA2 Summary of Fit RSquare RSquare Adj 0.059356 0.048544 001215 Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0332602 0.S06215 177 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 1.214616 0.607308 5 .4 8 9 8 174 19.248548 0.110624 Prob>F 176 20.463164 0 .0 0 4 9 Term Intercept PFOS PFOSA2 Parameter Estimates Estimate 0.6523913 *0.099406 0.0085787 Std Error 0.050721 0 .0 3 0 9 4 7 0 .0 0 3 3 7 4 t Ratio 12.86 -3.21 2.54 Prob>|t| <.0001 0 .0 0 1 6 0 .0 1 1 9 001216 Scatter Plot of Unconjuyated BiLicubin and PFOS# Antwerp# 1995 ------ Linear Fit ------ Polynomial Fit degree-2 lin ear Fit Unconjugated bilirubin - 0.69447 --0.02604 PFOS Summary of Fit RSquare RSquare Adj Root M ean Square Error M ean of Response Observations (or Sum Wgts) 0.016651 0.005217 0.374674 0.644318 88 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 1 0.204426 0.204426 1.4562 86 12.072733 0.140381 Prob>F 87 12.277159 0.2308 Term Intercept PFOS Parameter Estimates Estimate 0.6944654 -0.026035 Std Error 0.057638 0.021575 t Ratio 12.05 -1.21 Prob|t| <.0001 0.2308 Polynomial Fit degree-2 Unconjugated bilirubin - 0.73919 - 0.07468 PFOS + 0.00686 PFOSA2 Sum m ary of Fit RSquare RSquare Adj 0.027775 0.004899 001217 Root Moan Square Error M ean of Response Observations (or Sum Wgts) 0 374734 0.644318 88 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 0.340994 0.170497 1.2141 85 11.936165 0.140425 Prob>F 87 12.277159 0.3021 Term Intercept PFOS PFOSA2 Parameter Estimates Estimate 0.7391923 -0.074878 0.0068606 Std Error 0.07335 0.053838 0.006957 t Ratio 10.08 1.39 0.99 Prob>|t| <.0001 0.1690 0.3268 001218 Scatter PLot of Unconjugated Bilirubin and PPOS/ Decatur, 1995 Uncon|ugatod bilirubin By PFOS Unconjugated bkubin Linear Rt Polynomial R t degree-2 Linear Fit Unconjugated biRrubin - 0.39803 - 0.01166 PFOS Summaiy of Rt RSquare RSquare Adj Root M ean Square Error Mean of Response Observations (or Sum Wgts) 0.008475 -0.00292 0.235059 0.369663 89 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.0410886 0.041089 0.7436 87 4.8070013 0.055253 Prob>F 88 4.8480899 0.3909 Term Intercept PFOS Parameter Estimates Estimate 0.3980339 -0.011663 Std Error 0.04127 0.013525 t Ratio 9.64 0.86 Prob>|t| <.0001 0.3909 Polynomial Fit degree-2 Unconjugated bilirubin - 0.48279 - 0.06966 PFOS + 0.00605 PFOSA2 Summary of Fit RSquare 0.051431 001219 RSquare Adj Root Mean Square Error M ean oI Response Observations (or Sum Wgts) 0 02937 t 0.231244 0.369663 89 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 0.2493418 0.124671 2.3314 86 4.5987481 0.053474 Prob>F 88 4.8480899 0 .1 0 3 3 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 0.4827917 -0.069663 0.0060498 Std Error 0 .0 5 9 1 0 2 0 .0 3 2 2 6 2 0 .0 0 3 0 6 6 t Ratio 8.17 -2.16 1.97 Prob>|t| <.0001 0 .0 3 3 6 0 .0 5 1 7 001220 Scattec Plot of Unconjuyated BLlicubin and PFOS, Both Locations Combined, 1997 Unconjugated biirubm lin e a r Fit Polynomial Fit degree-2 Linear Fit Unconjugated bilirubin - 0.57745 - 0.0194 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.012338 0.005573 0.277986 0.543243 148 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 1 0.140939 0.140939 1.8238 146 11.282304 0.077276 Prob>F 147 11.423243 0.1789 Term Intercept PFOS Parameter Estimates Estimate 0.577447 -0.019403 Std Error 0.034111 0.014368 t Ratio 16.93 -1.35 Prob>|t| <.0001 0 .1 7 8 9 Polynomial Fit degree-2 Unconjugated bilirubin - 0.64247-0.09138 PFOS + 0.01097 PFOSA2 Summary of Fit RSquare 0.045634 001221 RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.03247 0.274201 0.543243 148 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 0.521284 0.260642 3.4666 145 10.901959 0.075186 Prob>F 147 11.423243 0.0338 Term Intercept PFOS PFOSA2 Parameter Estimates Estimate 0.6424713 0.091384 0.0109746 Std Error 0.044361 0.035001 0 .0 0 4 8 7 9 t Ratio 14.48 -2.61 2 .2 5 Prob>|t| <.0001 0 .0 1 0 0 0 .0 2 6 0 001222 Scatter Plot of Unconjugated Bilicubin and PFOS/ Antwerp, 1997 Unconiugated bilirubin By PFOS Linear Fit Polynomial Fit degree-2 Linear Fit Unconjugated bilirubin - 0.69161 - 0.02872 PFOS Summary of Fit R S q u are RSquare Adj Root M ean Square Error M ean of Response Observations (or Sum Wgts) 0.01217 -0.00376 0.329151 0.648438 64 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.0827520 0.082752 0.7638 62 6.7170917 0.108340 Prob>F 63 6.7998437 0.3855 Term Intercept PFOS Parameter Estimates Estimate 0.6916073 -0.02872 Std Error 0.064286 0.032862 t Ratio 10.76 -0.87 Prob>|t| <.0001 0.3855 Polynomial F t degree-2 Unconjugated bilirubin - 0.78708 - 0.19268 PFOS + 0.03945 PFOSA2 Summary of Fit R S q u are RSquare Adj 0.049788 0.018634 001223 Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0 325458 0 648438 64 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 0.3385540 0.169277 1.5981 61 6.4612897 0.105923 Prob>F 63 6.7998437 0.2106 Term Intercept PFOS P F O S A2 Parameter Estimates Estimate 0.7870832 -0.192677 0.0394497 Std Error 0.088403 0.110395 0.025386 t Ratio 8.90 -1.75 1.55 Prob>|t| <.0001 0 .0 8 6 0 0 .1 2 5 4 001224 Scatter Plot of Unconjugated Bilirubin and PFOS/ Decatur/ 1997 Unconjugated bilirubin By PFOS Unconjugated birubin .00 2 .0 0 4.00 6 .0 0 8.00 10.00 PFOS Linear Fit Polynomial Fit degree-2 Linear Fit Unconjugated bilirubin - 0.47268 - 0.00489 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.001886 -0.01029 0.202702 0.46309S 84 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.0063673 0.006367 0 .1 5 5 0 82 3.3692280 0.041088 Prob>F 83 3.3755952 0.6949 Term Intercept PFOS Parameter Estimates Estimate 0.47268 -0.004889 Std Error 0.032893 0.012419 t Ratio 14.37 -0.39 Prob>|t| <.0001 0.6949 Polynomial Fit degree-2 Unconjugated bilirubin - 0.5544 - 0.08508 PFOS + 0.01076 PFOSA2 Summary of Fit RSquare 0.077372 001225 RSquare Adj Root Moan Square Error M ean of Response Observations (or Sum Wgts) 0 054591 0.196086 0.463095 84 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 0.2611750 0.130588 3.3963 81 3.1144202 0 .0 3 8 4 5 0 Prob>F 83 3.3755952 0.0383 Term intercept PFOS P F O S *2 Parameter Estimates Estimate 0 .5 5 4 4 0 1 5 -0.085075 0.0107606 Std Error 0.044647 0.033385 0.00418 t Ratio 12.33 -2 .5 5 2 .5 7 Prob>[t| <.0001 0 .0 1 2 7 0 .0 1 1 9 001226 APPENDIX C H D L and PFOS Scatter Plots DRAFT 001227 Scatter Plot of HDL Cholesterol and PFOS, Both Locations Combined 1995 Linear Fit Polynomial R t degree-2 Linear Fit HDL-Cholesterol - 5 1 .0 1 9 7 -1 .1 7 5 0 9 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.030337 0.0247 12.55611 48.45977 174 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 1 848.392 848.392 5 .3 8 1 3 172 27116.827 157.656 Prob>F 173 27965.218 0 .0 2 1 5 Term Intercept PFOS Parameter Estimates Estimate 51.01971 -1.17509 Std Error 1.457347 0.506557 t Ratio 35.01 -2.32 Prob>|t| <.0001 0.0215 Polynomial Rt degree-2 HDL-Cholesterol - 53.5286 - 3.30537 PFOS + 0.25758 PFOSA2 Summary of Rt RSquare RSquare Adj 0.053266 0.042193 001228 Root Mean Square Error M ean of Response Observations (or Sum Wgts) 12 443 48 45977 174 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 1489.605 744.802 4.8105 171 26475.614 154.828 Prob>F 173 27965.218 0.0093 Term Intercept PFOS PFOSA2 Parameter Estimates Estimate 53.5286 -3.305373 0.257576 Std Error 1.898656 1.160937 0 .1 2 6 5 7 t Ratio 28.19 2.85 2.04 Prob>|t| <.0001 0.0050 0 .0 4 3 4 001229 Scatter Plot of HDL and PFOS, Antwerp, 1995 HOL Cholesterol By P FO S HDL-Cholestefol Linear Fit Polynomial Fit degree-2 Linear Fit HDL-Cholesterol - 55.2981 - 0.93355 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.019736 0.008337 12.32088 53.5 88 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 262.842 2 6 2 .8 4 2 1.7315 86 13055.158 151.804 Prob>F 87 13318.000 0.1917 Term Intercept PFOS Parameter Estimates Estimate 55.298149 -0.933553 Std Error 1.895379 0.709469 t Ratio 2 9 .1 8 -1.32 Prob>|t| <.0001 0.1917 Polynomial F t degree-2 HDL-Cholesterol - 56.91 - 2.68656 PFOS + 0.24725 PFOSA2 Summary of Fit RSquare 0.033054 001230 R S q u areA d j Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0 010302 12.30867 53 5 88 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 440.212 220.106 1.4528 85 12877.788 151.503 Prob>F 87 13318.000 0.2397 Term Intercept PFOS P F O S *2 Parameter Estimates Estimate 56.910034 2.686556 0.2472466 Std Error 2 .4 0 9 2 7 6 1 .7 6 8 3 9 2 0 .2 2 8 5 0 8 t Ratio 23.62 -1.52 1.08 Prob>|t| <.0001 0 .1 3 2 4 0 .2 8 2 3 001231 Scatter Plot of HDL and PPOS, Decatur/ 1995 Linear Fit Polynomial Fit degree-2 Linear Fit HDL-Cholesterol - 45.0317 - 0.7097 PFOS Summary of Fit RSquare RSquare Adj Root M ean Square Error M ean of Response Observations (or Sum Wgts) 0.014998 0.003272 10.89577 43.30233 86 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 151.843 151.843 1.2790 64 9972.296 118.718 Prob>F 85 10124.140 0.2613 Term Intercept PFOS Parameter Estimates Estimate 45.031672 -0.709695 Std Error 1.928382 0.627527 t Ratio 23.35 -1.13 Prob>|t| <.0001 0.2613 Polynomial Fit degree-2 HDL-Cholesterol - 47.1802 - 2.18908 PFOS + 0.15422 PFOSA2 Summary of Fit RSquare 0.028286 001232 RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.004872 10.88702 43.30233 86 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 286.376 143.188 1.2081 83 9837.764 118.527 Prob>F 85 10124.140 0.3040 Term Intercept PFOS P F O S *2 Parameter Estimates Estimate 47.180201 -2.189082 0.1542222 Std Error 2 .7 8 9 2 0 8 1.523604 0.144758 t Ratio 16.92 -1.44 1.07 Prob>|t| <.0001 0 .1 5 4 5 0 .2 8 9 8 001233 Scatter Plot of HDL and PFOS/ Both Locations Combined/ 1997 Linear Fit Polynomiai Fit degree-42 linear Fit HDL-Cholesterol - 46.3275 - 0.45133 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.004651 -0.00212 10.57408 45.53691 149 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 76.804 76.804 0.6869 147 16436.243 111.811 Prob>F 148 16513.047 0.4086 Term Intercept PFOS Parameter Estimates Estimate 46.327525 -0.451333 Std Error 1.288559 0.544563 t Ratio 35.95 -0.83 Prob>|t| <.0001 0.4086 Polynomial Fit degree-2 HDL-Cholesterol - 46.8473 - 1.03203 PFOS + 0.08884 PFOSA2 Summary of Fit RSquare RSquare Adj 0.006173 -0.00744 001234 Root Mean Square Error M ean of Response Observations (or Sum Wgts) 10 60212 45.53691 149 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 101.936 50.968 0 .4 5 3 4 146 16411.111 112.405 Prob>F 148 16513.047 0 .6 3 6 3 Term Intercept PFOS P F O S *2 Parameter Estimates Estimate 46.847332 -1.032032 0.0888404 Std Error 1.696367 1.343985 0 .1 8 7 8 8 2 t Ratio 27.62 -0.77 0 .4 7 Prob>[t| <.0001 0 .4 4 3 8 0 .6 3 7 0 001235 Scatter PLot of HDL and PFOS, Antwecp, 1997 Linear Rt Polynomial Fit degree-2 Linear Fit HDL-Cholesterol - 50.2864 - 0.41131 PFOS Summary of Fit RSquare RSquare Adj Root Mean Square Error M ean of Response Observations (or Sum Wgts) 0.00232 -0.01352 10.86492 49.67692 65 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 1 17.2928 17.293 0 .1 4 6 5 63 7436.9226 118.046 Prob>F 64 7454.2154 0 .7 0 3 2 Term Intercept PFOS Parameter Estimates Estimate 50.286404 -0.41131 Std Error 2.086111 1.074642 t Ratio 24.11 -0.38 Prob>|t| <.0001 0 .7 0 3 2 Polynomial Fit degree-2 HDL-Cholesterol - 52.4254 - 4.16781 PFOS + 0.90954 PFOSA2 Summary of Fit RSquare RSquare Adj 0.020968 -0.01061 001236 Root Mean Square Error M ean of Response Observations (or Sum Wgts) 10.84934 49.67692 65 Source Model Error C Total Analysis of Variance OF Sum of Squares Mean Square F Ratio 2 156.3026 78.151 0.6639 62 7297.9128 117.708 Prob>F 64 7454.2154 0.5184 Term Intercept PFOS P F O S *2 Parameter Estimates Estimate 52.425357 -4.167812 0.9095417 Std Error 2.865909 3.619456 0.836957 t Ratio 18.29 -1.15 1.09 Prob>|t| <.0001 0.2539 0 .2 8 1 4 001237 Scatter Plot of HDL, and PFOS, Decatur, 1995 ---- Linear Fit -- Polynomial Fit degree-2 Linear Fit HDL-Cholesterol - 42.3281 + 0.00268 PFOS Summary of Fit RSquare R S q u areA d j Root M ean Square Error M ean of Response Observations (or Sum Wgts) 2 .7 1 2 e -7 -0.01219 9.293759 42.33333 84 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 1 0.0019 0.0019 0.0000 82 7082.6647 86.3740 Prob>F 83 7082.6667 0 .9 9 6 2 Term Intercept PFOS Parameter Estimates Estimate 42.328069 0.002685 Std Error 1.508131 0.569385 t Ratio 28.07 0.00 Prob>|t| <.0001 0.9962 Polynomial Fit degree-2 HDL-Cholesterol - 42.5348 - 0.20021 PFOS + 0.02723 PFOSA2 Summary of Fit RSquare 0.000231 001238 RSquare Adj Roof Mean Square Error M ean of Raspons Observations (or Sum Wgts) -0 02446 9 349375 42 33333 84 Source Model Error C Total Analysis of Variance DF Sum of Squares Mean Square F Ratio 2 1.6333 0.8166 0.0093 81 7081.0334 8 7 .4 2 0 2 Prob>F 83 7082.6667 0 .9 9 0 7 Term Intercept PFOS P F O S *2 Parameter Estimates Estimate 4 2 .5 3 4 8 4 7 020021 0 .0 2 7 2 2 7 3 Std Error 2 .1 4 3 1 8 9 1 .5 9 1 8 9 2 0 .1 9 9 3 1 3 t Ratio 19.85 -0 .1 3 0 .1 4 Prob>|t| <.0001 0 .9 0 0 2 0 .8 9 1 7 001239 JOEM Volume 41, Number 9, September 1999 799 Serum Perfluorooctane Sulfonate and Hepatic and Lipid Clinical Chemistry Tests in Fluorochemical Production Employees JJLGeeaeafrafrnryreyMyRW..H.ZB.OouMbrlsraeiesnl,n,d,MeRDlDN,V,,MMMMD,PP,PHHMhPDH The 3M Company manufactures jluorochemicals, which have as a precursor perfluorooctane sulfonyl fluoride (CsF]7S 0 2F). These com pounds may be expected to transform metabolically, to an undetermined degree, to perfluorooctane sulfonate (PFOS, CfFI7S 0 3~) as an endstage metabolite. Subchronic studies in rats and primates indicate a potential for cumulative toxicity with PFOS with the primary effect related to metabolic wasting with hypolipidemia as a consistentfinding. Biennial medical surveillance has been offered to the company's fluorochemical production 'workers located in Decatur, Alabama, and Antwerp, Belgium. In 1995, the mean serum PFOS level, as measured by high-performance liquid chromatography mass spectrometry, for 178 male employees was 2.19 parts per million (ppm; range, 0.00 to 12.83 ppm), and in 1997, for 149 male employees, it was 1.75 ppm (0.10 to 9.93 ppm). Our analyses suggest that among these production employ ees, there were no substantial changes in serum hepatic enzymes, cholesterol, or lipoproteins associated with PFOS levels less than 6 ppm. It was not possible to derive inferences from thefew employees who had serum PFOS levels S 6 ppm. These results may be due to the lower levels of serum PFOS measured among these production employees, compared to those suspected to cause effects in laboratory animals. From the Medical Department. 3M Company, St. Paul. Minn. Address correspondence to: Geary Olsen. DVM. PhD, Medical Department. 3M Company. 220-3W-05. St. Paul. MN 55144. Copyright by American College of Occupational and Environmental Medicine he 3M C om pany m anufactures prod ucts that contain fluorochem icals, e i ther as intentional com ponents or re sidual im purities, that have as a precursor m olecule perfluorooctane su lfo n y l flu o rid e (C 8F | 7S 0 2F ). W ork place exposure m ay occur by inhala tion, ingestion, and derm al routes. T h ese fluorochem icals m ay transform m etabolically, to an undeterm ined de gree, to perfluorooctane sulfonate (P F O S ; C 8F i7S 0 3~ ) as an en d -sta g e m eta b o lite.1 P otassiu m p erflu o ro o cta n e su lfo n a te (C 8F 17S 0 3~ K + ) is itself a surfactant used as a w etting and foam ing agent in industrial and com m ercial processes. PFO S is k n ow n to concentrate pri m arily in the liver and, to a 10-fold lesser extent, in the p lasm a o f rats.2 There appears to be significant enterohepatic circulation o f PFO S with b o th urin ary an d f e c a l e x c r e tio n .2-3 C holestyram ine decreased the reten tion o f rad iolab eled P F O S in the liver and p lasm a and increased its elim ination via fe c e s.3 Subchronic stu d ies in rats and prim ates su ggest there is cu m u lative toxicity w ith P F O S.4' 7 L ow ered serum total ch o lesterol lev els appear to be an early consistent finding, w ith cum ulative to x icity resulting in m etab olic w ast ing and ultim ately death. A lthough the m ech an ism o f to x icity in labora tory anim als is not fu lly understood, it m ay be d ue to an e ffe c t on fatty acid transport and m etabolism , m em brane function, p eroxisom e prolifer ation, and/or m itochondrial bioener g e tic s .8' " 001240 800 PFOS, Hepatic Enzymes, and Cholesterol Olsen et al V oluntary biennial m edical sur veillance o f fluorochem ical produc tion em p lo y ees has been routinely perform ed sin ce the late 1 970s at 3 M 's D ecatur, A labam a, and A n t werp, B elgium , locations. Total se rum organic fluorine le v els w ere an alyzed until the m id -1 9 9 0 s w h en serum PFO S determ ination, quantifi able by h igh-perform an ce liq u id chrom atography m ass spectrom etry, becam e available. The purpose o f this study w as to p rovide a crosssectional analysis o f the m edical sur v eilla n ce data in relation to the e m p lo y ees' serum PFOS levels. Methods Fluorochemical Production Fluorochem ical production began in D ecatur in 1961 and A n tw erp in 1976. In gen eral, p erflu orin ated chem icals are produced via an e le c trochem ical process: a solu tion o f organic substrate is electro ly zed in anhydrous hydrogen fluoride at a lo w v o lt a g e .12 T h e p r o d u c ts o f th is electrolysis cell reaction are h ighly fluorinated com p ou n d s, w ith the end-product defined by the starting m aterial. Potential for w ork place e x posure to those flu oroch em icals that m etabolize to PFO S can occu r in the electroch em ical cell, the reactor, m ixing, drum m ing, and packaging areas, as w ell as in the plant's quality assurance and research and d evelop m ent laboratories. Subject Selection V oluntary fluoroch em ical m edical surveillance exam inations are o f fered biennially to approxim ately 200 A n tw er p an d 300 D e c a tu r p ro d u c tio n e m p lo y e e s . In 1995, a to ta l o f 178 m ale em p loyees (A ntw erp: n = 8 8 ; D eca tu r: n = 90) p a r tic i pated in the m ed ical su rv e illa n ce e x am ination s. In 1997, 149 m a le e m p lo y e e s (A n tw erp : n = 65; D ecatu r: n = 84) participated. (T here w ere too few fem ale e m p lo y e e s to in clu d e in the data an alysis.) S ix ty -o n e e m p lo y e e s participated in both yea rs. T h is reduction w as due to a large turnover o f e m p lo y ee s at both plant location s during 1996 and 1997. T he su rveil lance consisted o f a m edical q u es tionnaire: m easurem ent o f height, w eight, and blood pressure; standard clinical chem istry and hem atology tests: and determ ination o f serum PFOS levels. PFOS Analysis In 1995, the a n a ly sis for serum PFOS w as conducted by 3 M 's E nvi ronm ental Laboratory in St. Paul. M innesota. T he m ethod used tetrabutylam m onium to ion-pair w ith PFOS in s e r u m .13 T h e io n -p a ir s w e r e then extracted w ith ethyl acetate. The ab straction product w as then analyzed using high-perform ance liquid chrom atography-therm ospray m ass sp ec trom etry. In 1 997, the serum sam p les w ere analyzed by liquid chrom atography/m ass spectrom etry, using se lected ion m onitoring in the n ega tiv e-io n m o d e .1415 Laboratory Analyses For both tim e periods and plant locations, U nited Laboratory Services (St. Paul, M N ) perform ed the standard hem atological and serum chem istry tests. These included the follow ing: alkaline phosphatase (IU /L), gam m a glutam yl transferase (IU /L ), aspartate am inotransferase (IU /L ), alanine am i notransferase (IU /L ), total and direct bilirubin (m g/dL ), cholesterol (m g/ dL), low -d en sity protein (m g/dL ), high-density cholesterol (H D L; m g/ dL) and triglycerides (m g/dL). Clinical chem istry, h em atology, and serum PFOS determ inations were performed on overnight fasted blood sam ples. Data Analysis D escriptive, sim p le, and stratified analyses, analyses o f variance, and or dinary m ultivariable regression were used to evaluate associations between PFOS and each hem atological and clinical chem istry test. A ge. body m ass index (BM I: k g /n r ). current alcohol consum ption (drinks per day), and cig arette use (cigarettes sm oked per day) were potential con foun din g factors that w ere considered in the analyses. For stratified analyses, em p loyees were divided into four PFOS catego ries (0 to < 1 ppm . 1 to < 3 ppm. 3 to < 6 ppm . and > 6 ppm ) to determine w heth er an effect cou ld be detected at the highest serum PFOS levels. Other categorical le v els w ere used that pro vided sim ilar results. For multivariable regression analyses, serum PFOS and the potential confounders o f age. BM I, alcohol use, and cigarette use were exam ined as continuous and categori cal explanatory variables in the m od els. M ultivariable regression m odels w ere fitted, with PFO S analyzed as a continuous variable, using linear as w ell as non-linear transformations in order to m axim ize the possibility o f finding associations betw een PFOS and the dependent variable o f interest. Natural log transform ations o f the d e pendent variables w ere performed, w hen n ecessary, to norm alize vari ab les and to enhance m odel fit. Tradi tional stepw ise selection procedures w ere also utilized (selection in and out o f m odel w as set at P = 0.1), as w ell as taking into account other covariates that m ay be on the biologic pathway o f e f f e c t .16 W e d id n ot ex a m in e ch a n g es in m easured PFO S betw een the tw o tim e periods because the estim ated h a lf-life o f P F O S in the serum is b e lieved to range between 1000 and 15 0 0 d ays (J. H . M andel, M D , unpub lished data based on 4 years o f serum m easurem ents o f three retirees, 1998). Study results w ere analyzed using the S A S S y s te m .17 Results T he distribution o f em p loyees, by serum PFO S exposure categoriza t io n , is p r e s e n te d in T a b le 1. W hereas 20% o f the Decatur em p lo y e e s had exp osu res at ^ 3 ppm for b oth y ea rs, th is proportion in A n t w erp d eclin e d from 25% in 1995 to 13% in 1997. For both years. 95% o f the em p lo y ees' serum PFOS levels w ere below 6 ppm. There were no P F O S m ea su rem en ts > 6 ppm in A n tw erp in 1997. 001241 001242 JOEM Volume 41, Number 9, September 1999 801 TABLE 1 Distribution of Employees, by Year, Location, and Perfluorooctane Sulfonate (PFOS) Exposure Level (in parts per million [ppm]) 1995 Data 1997 Data All Employees Antwerp Decatur All Employees Antwerp Decatur PFOS n % n %n % n % n %n % 0 to <1 ppm 1 to <3 ppm 3 to < 6 ppm 2 6 ppm 45 91 35 7 25 34 39 h 12 60 51 32 36 59 66 63 20 19 22 16 18 21 4 3 34 4 5 40 31 43 25 14 9 30 48 29 38 38 14 12 05 35 45 14 6 Total 178 100 88 100 90 100 149 100 65 100 84 100 TABLE 2 Mean Values of PFOS, Demographic, Serum Chemistry, and Hematologic Parameters for Antwerp and Decatur, 1995 and 1997 Examinations" 1995 Data 1997 Data Variable PFOS, ppm Age, years BMI, kg/m2 Cigarettes, no. per day Alcohol, drinks per day Aik Phos, IU/L GGT, IU/L AST, IU/L ALT, IU/L T. bilirubin, mg/dL D. bilirubin, mg/dL Cholesterol, mg/dL LDL, mg/dL HDL, mg/dL Triglycerides, mg/dL Antwerp 1.93 37*** 23.9*** 4.7* 1.3*** 75*** 41 26* 44 0 .8 6 *** 0.22 214 138 54*** 115*** Decatur 2.44 45 29.2 7.9 0.2 97 48 29 47 0.58 0.21 218 136 43 187 Antwerp 1.48 33*** 23.5'** 5.5 1.1' " 70"* 26* 27 31 0.80'" 0.15"* 206 134 50*** 111**' Decatur 1.96 44 30.0 6.6 0.1 87 36 26 34 0.58 0.12 215 137 42 192 * P < 0.05; ** P < 0.01 ; *** P < 0.001. a BMI, body mass index; Aik Phos, alkaline phosphatase; GGT, gamma glutamyl trans ferase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; T., total; D., direct; LDL, low-density lipoprotein; HDL, high-density lipoprotein. T he m ean values for PFO S, d em o graphic, and liver and lipid test re sults, by location , are presented in T able 2. T he A ntw erp m ale em p loyee population w as sign ifican tly you n ger than that at D ecatur, had low er BM Is, and had higher selfreported daily consum ption o f a lco h o l. In ad d itio n , th eir c lin ic a l p r o file s w ere different for several tests. For both tim e periods, the A ntw erp em p loyees had low er m ean alkaline phosphatase and triglyceride values and higher total bilirubin and H D L v a lu e s. T able 3 lists the m ean, m edian, standard deviation, and range o f the covariates and hepatic enzym es, cho lesterol, and lipoproteins by four lev e l s o f P F O S c a te g o r iz a tio n (0 to < 1, 1 to < 3 , 3 to < 6 , and s 6 ppm ) for the com bined populations for each surveillance year. (H em atology and other clinical ch em istry data were unrem arkable and are not show n.) Several observations are noteworthy. First, the m ean for the ^ 6 ppm PFOS category w as one order o f m agnitude higher than the low est PFO S category (0 to < 1 ppm ) for both years. A lso , the m eans o f the four PFOS categories w ere sign ifi cantly different from each other. S ec ond, there w as on ly on e variable. total bilirubin, that had significant ( P < 0 .0 5 ) F tests for d ifferen ces in m eans in both years o f analysis. Stratification by plant location and su rv e illa n ce y ea r d id n ot result in sign ifican t find in gs for total bilirubin or direct bilirubin (T able 4). Third, m ean serum cholesterol levels re m ained constant (1 9 9 5 ) or increased (1 9 9 7 ) w ith higher PFO S category levels, although the H D L m ean val ues trended low er am ong those em p loyees w ith the highest PFO S cate gorization s. Stratification by plant location and su rveillan ce year did not result in sig n ifica n tly different m ean ch olesterol or H D L values by PFO S categories (T able 5). Finally, it sh ould be n oted that em p lo y ees in the h ighest PFO S category were old er and had higher B M Is than did em p lo y ees in the lo w est PFO S cate gory. Furtherm ore, in 1997, the em p loyees w ith s 6 ppm serum PFOS levels w ere on ly from D ecatur (Ta b le 1); thus the a n a ly ses m ay be confounded becau se D ecatur em p loyees w ere generally older and h eavier than A ntw erp em p loyees. For exam p le, in 1997 the m ean H D L level for the ^ 6 ppm group w as 40 m g/d L (T ab le 3). H ow ever, this m ean value w as so lely from Decatur em ployees. Linear and nonlinear relationships betw een PFO S and the dependent variables o f interest, taking into ac count the potential confounding af fects o f age, BM I, alcohol, and cig arettes, resu lted in m any an alyses. Total bilirubin sh ow ed a significant 001243 8 0 2 PFOS, Hepatic Enzymes, and Cholesterol Olsen et al TABLE 3 Mean, Median, Standard Deviation (SD) of Mean and Range of PFOS, by Demographic and Serum Chemistries for Antwerp and Decatur Employees Combined, for 1995 (rt = 178) and 1997 (n = 147) 1995 Data 1997 Data PFOS* (ppm) PFOS, ppm 0 to <1 1 to <3 3 to < 6 >6 Age, years Oto <1 1 to <3 3 to <6 >6 Alcohol, drinks per day 0 to <1 1 to < 3 3 to < 6 6 BMI, kg/m2 0 to <1 1 to <3 3 to <6 >6 Cigarettes, no. per day 0 to <1 1 to <3 3 to <6 >6 Alkaline phosphatase, IU/L Oto <1 1 to <3 3 to <6 6 GGT, U/L 0 to <1 1 to <3 3 to < 6 >6 AST, U/L 0 to <1 1 to < 3 3 to <6 a6 ALT, IU/L 0 to <1 1 to <3 3 to < 6 s6 T. bilirubin, mg/dL 0 to <1 1 to <3 3 to <6 5=6 Mean Median SD 0.491 0.50 0.27 1.82' 1.77 0.58 4.12' 3.97 0.81 8.17' 7.73 2.52 F value =321.9, F < 0.0001 37 36 422 41 40 40 45 43 F value = 3.7, P = 0.02 8 9 7 7 0.8 0.6 0.9 0.5 0.1 0.7 1.2 0.3 1.9 0.7 0.0 1.1 F value = 4.0, P = 0.009 25.5 24.8 27.7 26.3 24.93 25.0 27.7 29.4 F value = 3.7, P = 0.02 4.2 5.8 3.8 4.2 2.6 0.0 6.4 6.8 0.0 11.3 10.6 8.0 12.4 0.4 0.0 1.1 F value = 4.8, P = 0.003 80 78 22 89 89 27 86 85 21 88 85 24 F value = 1.3, P = 0.28 43 31 28 47 36 39 40 39 15 43 33 18 F value = 0.5, P = 0.71 27 25 13 29 26 12 25 24 5 33 33 6 F value = 1.8, P = 0.14 48 43 20 46 42 21 42 41 7 51 49 17 F value = 1.0, P = 0.38 0.88 0.70 0.50 0.662 0.60 0.30 0.642 0.60 0.28 0.76 0.70 0.23 F value = 4.4, P = 0.005 Range Mean Median SD Range 0.00 to 0.90 1.00 to 2.91 3.00 to 5.80 6.06 to 12.83 21 to 58 25 to 60 26 to 55 37 to 56 0 to 3.6 0 to 3.6 0 to 6.0 0 to 2.9 17.9 to 38.7 19.6 to 60.7 17.9 to 32.5 20.6 to 33.0 0.0 to 25.0 0.0 to 40.0 0.0 to 40.0 0.0 to 3.0 31 to 158 49 to 191 32 to 124 63 to 136 16 to 155 2 to 293 21 to 80 28 to 79 15 to 96 14 to 90 13 to 37 26 to 43 27 to 118 18 to 183 30 to 59 29 to 82 0.40 to 2.90 0.20 to 1.50 0.20 to 1.40 0.50 to 1.20 0.52' 0.52 0.27 1.78' 1.64 0.56 3.87' 3.59 0.70 7.201 6.68 1.59 F value = 367.6, P < 0.0001 36 34 11 422 41 9 41 42 5 42 45 9 F value = 5.1, P = 0.002 0.5 0.1 0.8 0.5 0.1 0.8 1.0 0.1 1.8 0.2 0.1 0.2 F value = 1.8, P = 0.15 26.0 24.9 4.9 27.7 26.4 5.7 27.3 27.9 4.4 30.8 29.7 4.0 F value = 2 .1 ,P = 0.10 4.7 0.0 9.4 8.2 0.0 11.3 4.1 0.0 8.3 6.0 0.0 13.4 F value = 1 .5 ,P = 0.23 77 73 17 83 79 23 76 74 22 88 84 18 F value = 1.2, P = 0.32 28 22 20 36 25 33 28 27 14 33 37 12 F value = 1 .1 ,P = 0.34 27 25 7 26 25 7 25 23 7 29 28 3 F value = 0.5, P = 0.67 31 30 11 33 29 16 34 31 18 41 45 10 F value = 0.9, P = 0.46 0.77 0.60 0.40 0.612 0.60 0.21 0.63 0.50 0.31 0.58 0.50 0.24 F value := 2.9, P = 0.04 0.10 to 0.97 1.02 to 2.89 3.09 to 5.30 6.05 to 9.3 21 to 62 24 to 63 32 to 54 29 to 52 0 to 4.3 0 to 5.0 0 to 7.1 0.1 to 0.8 20.1 to 41.7 18.1 to 48.5 19.1 to 36.0 26.1 to 36.2 0 to 40 Oto 40 Oto 30 0 to 30 49 to 132 41 to 163 29 to 120 65 to 114 10 to 142 10 to 179 13 to 71 17 to 48 13 to 53 15 to 56 14 to 43 26 to 34 13 to 60 10 to 89 14 to 82 25 to 49 0.30 to 2.30 0.30 to 1.30 0.40 to 1.30 0.40 to 1.00 001244 JOEM Volititi* 41, Num b*r 0, September 1099 803 TABLE 3 Continued. 1995 Data 1997 Data PFOS* (ppm) D. bilirubin, mg/dL 0 to <1 1 to <3 3 to <6 a6 Cholesterol, mg/dL 0 to <1 1 to <3 3 to < 6 >6 LDL, mg/dL 0 to <1 1 to <3 3 to <6 >6 HDL, mg/dL Oto <1 1 to <3 3 to <6 26 Triglycerides, mg/dL Oto <1 1 to <3 3 to <6 26 Mean Median SD 0.22 0.20 0.05 0.21 0.20 0.06 0.21 0.20 0.04 0.20 0.20 0.02 F value = 0.6, P = 0.58 219 215 47 216 213 43 214 214 35 213 221 36 F value = 0.1, P = 0.96 140 137 43 134 134 40 137 135 34 142 136 32 F value = 0.2, P = 0.87 53 53 13 48 47 13 45 46 11 45 46 9 F value = 2.9, P = 0.04 129 96 98 161 133 107 158 142 88 132 151 45 F value = 1.1, P = 0.35 Range Mean Median SD 0.02 to 0.40 0.10 to 0.40 0.20 to 0.30 0.10 to 0.30 100 to 340 118 to 315 128 to 278 160 to 251 29 to 261 44 to 234 65 to 190 95 to 178 31 to 94 26 to 94 23 to 74 34 to 61 41 to 622 41 to 651 34 to 413 64 to 187 0.15 0.10 0.07 0.122 0.10 0.04 0.12 0.10 0.04 0.10 0.10 0.00 F value = 3.5, P = 0.02 198 197 40 216 219 42 2292 224 29 229 238 26 F value = 4.3, P = 0.006 124 128 34 141 134 38 1482 142 24 145 156 26 F value = 3.7, P = 0.01 46 48 11 44 45 10 48 47 10 40 38 4 F value = 1.1, P = 0.34 148 107 162 156 122 108 166 158 92 220 191 83 F value = 0.5, P = 0.67 * See Table 1 for sample size by year. 1 Mean is significantly different (P < 0.05,Bonferroni (Dunn) test) from the mean of the other PFOS categories. 2 Mean is significantly different (P < 0.05,Bonferroni (Dunn) test) from the mean of 0 to <1 ppm PFOS category. 3 Mean is significantly different (P < 0.05,Bonferroni (Dunn) test) from the mean of 1 to < 3 ppm PFOS category. Range 0.10 to 0.40 0.10 to 0.20 0.10 to 0.20 0.10 to 0.10 110 to 280 116 to 365 192 to 321 186 to 250 50 to 205 61 to 290 111 to 196 103 to 164 19 to 74 28 to 69 32 to 69 37 to 45 38 to 1209 44 to 534 45 to 394 149 to 352 quadratic association w ith P F O S for D ecatur em p loyees. That is, total b il irubin levels in itially d eclin ed but subsequently increased w ith in creas ing PFO S lev els. T his association w as not ob served in the A n tw erp em p lo y ee p opulation, w hich had h igh er total bilirubin le v els. In 1997, serum cholesterol levels w ere p o si tively associated (linearly) w ith se rum PFO S, after adjustm ent for p o tential con foun din g factors am ong D ecatur em p loyees (data not sh ow n ). T his association w as not observed w ith serum PFO S and ch o lestero l in 1995 am ong D ecatur em p loyees or in eith er tim e period w ith the A n t werp em p loyees. Stratification o f the data by plant (T able 5) and m ultiva- riable an alyses (data not sh ow n ) show ed no consistent associations betw een PFO S and H D L . For both plants com b ined, H D L lev els were significantly n egatively associated (linearly) w ith P F O S in 1995, after adjustm ent for possib le confounding factors, but this w as not ob serv ed in 1997. Discussion W e conducted tw o cross-sectional analyses o f surveillance data to e x am ine the associations betw een se rum PFOS levels and several hem a tological and clinical chem istry tests in m ale fluoroch em ical production em ployees. For both years, 95% o f the m easured serum PFOS levels w ere b elow 6 ppm . B ecau se the A nt werp and D ecatur em ployees w ere dissim ilar by age, BM Is, and selfreported alcohol use, w e conducted com bined as w ell as separate analy ses by plant location. T hese three dem ographic factors explain, at least partially, w hy the A ntw erp em p loy ees had low er m ean serum triglycer ides and h igher H D L le v els than the D eca tu r e m p lo y e e s .18-22 C linical hepatic en zym e tests were not associated w ith the em p lo y ees' serum PFO S levels. This w as an im portant q uestion to address b e cause PFO S has been reported (1) to be a p e r o x iso m e p roliferator in rats810; (2 ) to result in increased p lasm a liver transam inase tests in a 001245 804 PFOS, Hepatic Enzymes, and Cholesterol * Olsen et al TABLE 4 Mean, Median (Med), Standard Deviation (SD) of Mean and Range of PFOS, and Total and Direct Bilirubin Levels, by Year and Plant Location Antwerp Decatur PFOS* (ppm) 1995 Data T. bilirubin, mg/dL 0 to <1 1 to <3 3 to <6 6 D. bilirubin, mg/dL 0 to <1 1 to <3 3 to <6 >6 1997 Data T. bilirubin, mg/dL 0 to <1 1 to <3 3 to <6 a6 D. bilirubin, mg/dL 0 to <1 1 to <3 3 to <6 2=6 Mean Med SD Range Mean Med SD Range 0.96 0.80 0.55 0.83 0.80 0.26 0.75 0.70 0.30 0.93 0.90 0.25 F value = 1.2, P = 0.31 0.23 0.20 0.06 0.22 0.20 0.04 0.21 0.20 0.03 0.20 0.20 0.00 F value = 0.7, P = 0.55 0.90 0.80 0.46 0.68 0.70 0.23 0.79 0.70 0.40 -- --_ F value = 2.3, P = 0.11 0.16 0.20 0.08 0.13 0.10 0.05 0.14 0.10 0.05 -- ---- F value = 2.2, P = 0.12 0.40 to 2.90 0.40 to 1.30 0.30 to 1.40 0.70 to 1.20 0.20 to 0.40 0.20 to 0.30 0.20 to 0.30 0.20 to 0.20 0.40 to 2.30 0.30 to 1.30 0.30 to 1.30 -- 0.10 to 0.40 0.10 to 0.20 0.10 to 0.20 -- 0.65 0.60 0.16 0.57 0.50 0.28 0.51 0.50 0.18 0.63 0.65 0.10 F value = 0.7, P = 0.54 0.20 0.20 0.00 0.20 0.20 0.06 0.20 0.20 0.04 0.20 0.20 0.08 F value = 0.4, P = 0.74 0.63 0.60 0.30 0.56 0.50 0.18 0.51 0.50 0.16 0.58 0.50 0.24 F value = 1.0, P = 0.41 0.13 0.10 0.06 0.11 0.10 0.03 0.11 0.10 0.03 0.10 0.10 0.00 F value = 1.3, P = 0.28 0.40 to 0.90 0.20 to 1.50 0.20 to 1.00 0.50 to 0.70 0.20 to 0.20 0.10 to 0.40 0.20 to 0.30 0.10 to 0.30 0.30 to 1.40 0.30 to 1.00 0.30 to 0.90 0.40 to 1.00 0.10 to 0.30 0.10 to 0.20 0.10 to 0.20 0.10 to 0.10 * See Table 1 for sample sizes by year. subchronic rat study4; and (3 ) to increase serum aspartate am inotrans ferase and low er alkaline phospha tase lev els in a subchronic rhesus m onkey study.5 There appears to be significant enterohepatic circulation o f PFO S w ith both urinary and feca l excretion in th e rat.2-3 A lth o u g h w e o b s e r v e d a quadratic relation b etw een P F O S and total bilirubin in our m ultivariable analyses w ith the D ecatur em p loyee population, interpretation o f these data is d ifficu lt, g iv en the narrow range o f PFOS serum levels a ssoci ated w ith the total bilirubin values. Interestingly, the A ntw erp em p lo y e e s ' total bilirubin lev els w ere sig n if icantly higher than the D ecatur em p lo y ees' levels. W e suspect there may be a greater prevalence o f G il b ert's sy n d r o m e 20 23 a m o n g th e A n t w erp e m p lo y e e s. In I9 9 5 . 15 (17 % ) A ntw erp em p loyees had total biliru bin valu es > 1 . 2 m g/dL , com pared with three (3%) D ecatur em p lo y ees. Com parable percentages also existed in 1997. W e are uncertain w hether p ost-collection factors, including he m olysis, light, and heat, could have contributed to the d ifferen ces in total bilirubin le v els b etw een the A ntw erp an d D e c a tu r e m p lo y e e s .18 U n fo r tu nately, total bilirubin w as not ana lyzed in tw o su bchronic rhesus m o n key studies, w hich resulted in death o f all an im als in the 4 .5 m g/k g/d ay P F O S d o s e g ro u p o r h ig h e r .5'6 O n go in g to x ic o lo g y stud ies in rats and prim ates should provide additional perspectives regarding any b io lo g i cal effect b etw een PFO S and biliru bin levels. Our data do not suggest a reduction in total serum cholesterol with PFO S at the serum le v els m easured am ong these production em ployees. PFOS is a p e r o x iso m e p roliferator in the rat. and hypolipidem ia has been consistently o b serv ed in su b ch ro n ic rat and prim ate to x ic o lo g y stu d ie s.5 - 8 10 R h esu s m o n k eys that had been adm inistered PFO S at 4.5 m g/kg/day had m ean serum ch olesterol values that decreased from 183 m g/dL to 99 m g/dL w ithin 30 days.5 R hesus m onkeys adm inistered 1.5 m g/kg/day had m ean cholesterol le v els that decreased from 195 m g/dL to 111 m g/dL w ithin 9 0 days. A noobservable-effect level w as seen for the 0.5 m g/kg dose group at 9 0 days. Serum PFOS m easurements were not d eterm ined in these rhesus m onk eys. H ow ever, data from a recent cyn om olgus m onkey dose-range-finding study su ggested that hypolipidem ia m ay be initially associated with serum PFO S levels in the range o f 100 to 200 ppm .7 H aughom and Spydevold suggested that the hypolipidem ic effect o f PFO S m ay be due to reduced liver activity o f hydroxym ethyl glutaric acid-C oA re ductase and acyl-C oA cholesterol acyltransferase. with enhanced fatty acid oxid ation in the liver.0 N abb efeld et al observed that PFO S has a high affinity for the fatty acid carrier pro teins album in and i.-fatty a cid -b in d in g 001246 JOEM Volum 41, Number 9, September 1999 805 TABLE 5 Mean, Median (Med), Standard Deviation (SD) of Mean and Range of PFOS, and Total and HDL Cholesterol, by Year and Plant Location Antwerp Decatur PFOS* (ppm) 995 Data Cholesterol, mg/dL 0 to <1 1 to <3 3 to <6 =r6 HDL, mg/dL 0 to <1 1 to < 3 3 to < 6 >6 997 Data Cholesterol, mg/dL 0 to <1 1 to <3 3 to < 6 >6 HDL, mg/dL 0 to <1 1 to <3 3 to < 6 s6 Mean Med SD 220 219 206 211 217 215 223 221 F value = 0.6, P = 0.61 50 49 30 16 56 57 53 51 50 49 53 49 F value = 1.1, P = 0.35 13 13 11 7 193 190 41 213 205 48 228 223 38 -- F value = 2.9, P = 0.07 51 50 12 48 46 10 51 50 10 F value = 0.8, P = 0.47 Range Mean Med SD 100 to 340 118 to 315 178 to 266 208 to 240 31 to 94 33 to 79 31 to 74 48 to 61 110 to 277 116 to 365 192 to 321 19 to 74 34 to 68 39 to 69 -- 215 208 221 218 209 213 206 206 F value = 0.5, P = 0.69 39 39 42 47 43 41 45 44 39 39 39 39 F value = 1.4, P = 0.25 9 12 9 5 204 208 218 226 230 230 229 238 F value = 2.0, P = 0.13 38 39 23 26 42 41 42 41 45 45 40 38 F value = 0.5, P = 0.67 9 10 10 4 Range 154 to 276 132 to 300 128 to 278 160 to 251 31 to 59 26 to 94 23 to 51 34 to 46 145 to 277 152 to 290 197 to 280 186 to 250 26 to 59 28 to 69 32 to 62 37 to 45 * See Table 1 for sample sizes by year. protein in vivo. T his could potentially alter fatty acid transport, biochem istry, and m etabolism , w hich conceivab ly could then lead to a decrease in ch o lesterol esterification and m etabolic w a s tin g .11 Several m eth od ological issu es should be considered in evaluating the results from our study. First, the crosssectional design does not allow for a direct analysis o f the tem porality o f an association. Second, the voluntary par ticipation rates in the fluorochem ical m edical surveillance were not ideal am ong the eligib le fluorochem ical pro duction em p loyees. Third, the serum levels o f PFO S that were m easured m ay be b elo w the no-effect lev el in laboratory anim als. Fourth. PFO S c o n centrates prim arily in the liver o f lab oratory anim als. Serum m easurem ents o f PFOS m ay not adequately reflect body burden. Fifth, the tw o crosssectional analyses cannot be view ed as independent populations. S ixty-on e em p lo y ees w ere studied in both years. T his w as due, in part, to a large turn over o f em p loyees at both plants b e tw een exam inations. Sixth, there could be m easurem ent error in important confounding variables. A n alysis o f the data o f the 61 subjects w h o partici pated in both years sh ow ed that there w as excellent correlation for the con founding factors o f BM I (r = .92, P = 0.0001), self-reported aspects o f alco hol consum ption (r = .88, P = 0.0001), and cigarette sm oking (r = .79, P = 0.0001). A s expected, these 61 em p loyees' serum PFOS levels for the 2 years were highly correlated (r = .92, P = 0.0001). Their results did not differ from those o f the entire study population. Finally, the quality o f m edical surveillance data, prior to its use for studying an a priori hypothesis, can often be evaluated by whether know n positive associations are ob served. In this regard, w e did observe various expected associations, such as cigarette sm oking and elevated white blood cell counts and large BM Is. associated w ith elevated liver transam in a se le v e ls .24-25 In sum m ary, our find in gs su ggest that am ong these A ntw erp and D eca tur m ale flu oroch em ical production em p loyees, there w ere no substantial ch an ges in serum hepatic en zym es, cholesterol, or lipoproteins associ ated w ith PFO S lev els less than 6 ppm . It is not p o ssib le to derive inferences from the few em ployees w ith serum lev els ^ 6 ppm . Lim ita tions o f this study in clu d e its crosssectional d esign , the lo w voluntary participation rates am ong the em p loyees, and the low er levels o f se rum PFOS m easured am ong these em p loyees, com pared w ith those sus p ected to ca u se effe c ts in laboratory anim als. Acknowledgment The authors gratefully acknowledge the assistance of Michele Burlew. 001247 806 PFOS, Hepatic Enzymes, and Cholesterol Olsen et al References 1. Gibson SJ, Johnson JD, Ober RE. A b- sorption and B iotransform ation o f N- Etli \ / F O SE a n d issu e D istribution a n d Elim ination o f Carbon-1-I A fter A dm inis tra tio n o f N -E th y l F O S E -IJC In F eed. St. Paul. MN: Riker Laboratories Inc; 1983. 2. Johnson JD, Gibson SJ. Ober RE. E xten t and R oute o f Excretion a n d Tissue D is tribution o f Total C arb o n -14 in R ats A fte r a S in g le i.v. D ose o f F C -9 5 -'4C. P ro jec t N o. 8 9 0 0 3 1 0 2 0 0 . St. Paul. MN: Riker Laboratories Inc; 1979. 3. Johnson JD. Gibson SJ, Ober RE. Cho lestyramine-enhanced fecal elimination of carbon-14 in rats after administration of ammonium [WC| perfluorooctanoate or po tassium [l4C] perfluorooctanesulfonate. F u n d A p p i Toxicol. 1984;4:972-976. 4. Goldenthal El, Jessup DC, Geil RG, Jef ferson ND, Arceo RJ. N in e ty -D a y S u b acute R at Toxicity Study. Study No. ST OSS. Mattawan, MI: International Research Development Corp; November 1978. 5. Goldenthal El, Jessup DC, Geil RG, Jef ferson ND. N in e ty -D a y S u b a c u te R h esu s M onkey Toxicity Study. Study N o 1 3 7 - 085. Mattawan, MI: International Re search Development Corp; December 1978. 6. Goldenthal El, Jessup DC, Geil RG, Jef ferson ND. N in ety -D a y S u b a c u te R h e s u s M onkey Toxicity Study. Study No. 1 3 7 - 087. Mattawan, Ml: International Re search Development Corp; January 1979. 7. Thomford PJ. F our-W eek C a p su le Toxicity Study With P erfluorooctane Sulfonic A cid Potassium Salt (PFO S) in Cynom ologus M onkeys. C o vance S tu d y 6 3 2 9 - 2 2 2 . Mad ison, WI: Covance Co; August 1998. 8. Ikeda T, Fukuda K. Mori I, Enomoto M, Komai T, Suga T. Induction of cytochrome P-450 and peroxisome proliferation in rat liver by perfluorinated octanesulfonic acid. In: Fallimi HD, Sies H, eds. P eroxisom es in B io lo g y a n d M e d ic in e . New York: Springer Verlag; 1987:304-308. 9. Haughom B. Spydcvold O. The mecha nism underlying the hypolipmie effect of porfluorooetanoic acid (PFOAl. perfluoroctunc sulphonic acid iPFOSA) and clofibric acid. B io ch em B io p h ys A cta. 1992; 1128:65-72. 10. Sohlenius AK, Eriksson AM. Hogstrom C. Kimland M. DePierre JW. Perfluo rooctane sulfonic acid is a potent inducer of peroxisomal fatty acid beta-oxidation and other activities known to be affected by peroxisome proliferators in mouse liver. P harm acol Toxicol. 1993;72:90-93. 11. Nabbefeld J, Butenhoff, J, Bass N, Seacat A. Displacement of a fluorescently la beled fatty acid analogue from fatty acid carrier proteins by Wyeth-14.643, ammo nium perfluorooctanoate, potassium per fluorooctane sulfonate and other known peroxisome proliferators (abstract). T o x icologist. 1998;42:395. 12. Bryce H. Industrial and utilitarian aspects of fluorine chemistry. In: Simons J, ed. F lu o rin e C h em istry. New York: Aca demic Press; 1964:297-492. 13. Johnson JD, Wolter JT, Colaizy GE, Rethwill PA, Nelson RM. Q uantification o f Perflurooctanoate and Perfluorooctanesul fo n a te in H um an Serum Using Ion-P air Extraction and High Performance Liquid C hrom atography-T herm ospray M ass Spectrom etry W ith A utom ated Sam ple Preparation. St. Paul, MN: 3M Environ mental Laboratory; 1996. 14. Anderson DJ, Muivana DE. A n a ly tic a l R eport fo r the D eterm ination o f P erfluo rooctanoate and Perfluorooctanesulfon a te in H u m a n S e ru m b y L C /M S . Ithaca, NY: Advanced Bioanalytical Services Inc; August 1997. 15. Anderson DJ, Muivana DE. A n a ly tic a l R eport fo r the D eterm ination o f P erfluo rooctanoate and P erfluorooctanesulfon a te in H u m a n S e ru m b y L C /M S . Ithaca, NY: Advanced Bioanalytical Services Inc: September 1997. 16. Greenland S. Modeling and variable se lection in epidemiologic analysis. A m J P ublic H ealth. 1989:79:340-349. 17. SAS Institute. Inc. SA S U sers G uide: S ta tistics. V ersion 6. Cary. NC: SAS Institute. Inc; 1990. 18. Davern TJ. Scharschmidt BF. Biochemi cal liver tests. In: Feldman M, Sleisenger MH, Scharschmidt BF, eds. G a stro in te s tin a l a n d L iv e r D ise a se , 6th ed. Philadel phia, PA: W.B. Saunders Co.; 1993: 1112- 1122. 19. Lewis JA. Illu s tr a te d G u id e to D ia g n o s tic Tests. Springhouse, PA: Springhouse Corp; 1994:100-107, 194-207. 20. Friedman LS. Liver, biliary tract and pancreas. In: Tierney LM, McPhee SJ, Papadakis WA. eds. M e d ic a l D ia g n o sis a n d T reatm ent, 37th ed. Stamford, CT: Appleton and Lange; 1998:628-630. 21. Fu PC. Lipids, lipoproteins, and apolipoproteins. In: Howanitz JH, Howanitz PJ, eds. L a b o r a to r y M e d ic in e : T e st S e le c tio n a n d In te r p r e ta tio n . New York: Churchill Livingstone Inc; 1991:173198. 22. Wolf PL. Liver function. In: Howanitz JH, Howanitz PJ, eds. L a b o ra to ry M e d i cine: Test Selection a n d Interpretation. New York: Churchill Livingstone Inc; 1991:67-84. 23. Lidofsky S, Scharschmidt BF. Jaundice. In: Feldman M, Sleisenger MH, Scharschmidt BF, eds. G a stro in te stin a l a n d L iv e r D isea se, 6th ed. Philadelphia: W.B. Saunders Co; 1998:220-224. 24. Olsen GW, Kusch GD, Stafford BA, Gudmundsen SL, Currier MF. The posi tive known association design: a quality assurance method for occupational health surveillance data. J O ccu p M ed. I991;33: 998-1000. 25. Bums CJ, Boswell JM, Olsen GW. Liver enzyme activity and body mass index. J O c cu p E n viro n M ed. 1997;38:1248 -- 1252. 001248