Document zd4gkGOeq4pnrKnobq0mwdqg6

Corporate Occupational Medicine 3M Center, Building 220-3W-05 St. Paul, MN 55144-1000 651 737 4230 Telephone 651 733 9066 Fax PRX%rOH% 8. Plasma Cholecystokinin and Hepatic Enzymes, Cholesterol and Lipoproteins in Ammonium Perfluorooctanoate Production Workers. Toxicology research suggested, to a limited extent, that the incidence o f pancreas acinar cell adenomas in rats fed perfluorooctanoic acid may be the result o f a mild but sustained increase in cholecystokinin (CCK) as a consequence o f hepatic cholestasis. To assess this hypothesis, plasma CCK levels were measured by radioimmunoassay for those Cottage Grove employees participating in the 1997 fluorochemical surveillance examinations (n = 84). The mean serum PFOA level, as measured by high performance liquid chromatography mass spectrometry methods, was 6.8 ppm (median 1.3 ppm, range 0.1 - 81.3 ppm). CCK values approximated the assay's reference range for a 12 hour fast and were negatively, not positively, associated with employees' serum PFOA levels. In addition to the CCK analyses, medical surveillance examinations were reviewed for the 1993, 1995 and 1997 time periods to assess the initial hypothesis generated from the 1990 fluorochemical medical surveillance data from this Cottage Grove workforce that PFOA (measured as total organic fluorine) may modulate hepatic responses to obesity and alcohol consumption (see studies # 5 and #7). In these three subsequent time periods, PFOA, assayed by mass spectrometry, did not appear to modulate hepatic responses to either obesity or alcohol consumption. Regardless of the surveillance year, there was no indication of significant clinical hepatic toxicity at the PFOA levels G034S5 observed in this workforce. The study protocol for the assessment of cholecystokinin levels, the internal 3M Final Report and a manuscript that has been accepted for publication (Drug and Chemical Toxicology) are submitted. [Note: the PFOA category levels for the univariate analyses in the publication manuscript (0 - < 1 ppm, 1 - < 10 ppm, > 10 ppm) are different than those reported in the 3M final report (0 - < 1 ppm, 1 < 1 0 ppm, 10 - < 30 ppm, > 30 ppm). This change in the publication manuscript was done per a request made during the peer review process. This difference in categories did not affect the results or conclusions.] G034S6 * '* [ Ip press 2000/ Drug and Chemical Toxicology] PLASMA CHOLECYSTOKININ AND HEPATIC ENZYMES, CHOLESTEROL AND LIPOPROTEINS IN AMMONIUM PERFLUOROOCTANOATE PRODUCTION WORKERS Geary W. Olsen*. Jean M. Burris. Michele M. Buriew, Jeffrey H. Mandel Medical Department. 3M Company. 220-3W-05, St. Paul. MN 55144-1000 ABSTRACT Ammonium perfluorooctanoate is a potent synthetic surfactant used in industrial applications. It rapidly dissociates in biologic media to perfluorooctanoate [CFj(CF2)6C 02-). which is the anion o f perfluorooctanoic acid [PFOA, CFj(CF2)sCOOH]. PFOA is a peroxisome proliferator known to increase the incidence o f hepatic, pancreas and Leydig cell adenomas in rats. The pancreas acinar cell adenomas may be the consequence of a mild but sustained increase o f cholecystokinin as a result o f hepatic cholestasis. Although no significant clinical hepatic toxicity was observed, PFOA was reported to have modulated hepatic responses to obesity and alcohol consumption among production workers. To further assess these hypotheses, we examined medical surveillance data of male workers involved in ammonium perfluorooctanoate production in 1993 (n = 111), 1995 (n - 80) and 1997 (n = 74). Serum PFOA was measured by high-performance liquid chromatography mass spectrometry methods. Plasma cholecystokinin was measured (only in 1997) by the use o f direct radioimmunoassay. Serum biochemical tests included hepatic enzymes, cholesterol and lipoproteins. Serum PFOA levels, by year, were: 1993 (mean 5.0 ppm, SD 12.2, median 1.1 ppm, range 0.0 - 80.0 ppm); 1995 (mean 6.8 ppm, SD 16.0, median 1.2 ppm, range 0.0 - 114.1 ppm); and 1997 (mean 6.4 ppm, SD 14.3, median 1.3 ppm, range 0.1 - 81.3 ppm). CCK values (mean 28.5 pg/ml, SD 17.1, median 22.7 pg/ml, range 8.8-86.7pg/ml) approximated the assay's reference range (up to 80pg/ml)fo r a 12 hour fa st and were negatively, not positively, associated with employees' serum PFOA levels. Our findings continue to suggest there is no significant clinical hepatic toxicity associated with PFOA levels as measured in this workforce. Unlike a previously reported observation, PFOA did not appear to modulate hepatic responses to either obesity or alcohol consumption. Limitations of these findings include: 1) the crosssectional design as only 17 subjects were common for the three surveillance years; 2) the voluntary participation that ranged between 50 and 70 percent; and 3) the few subjects with serum levels > 10 ppm. f Corresponding Author C004S7 INTRODUCTION Ammonium perfluorooctanoate [APFO; C FjfC FjleC O V N W ] is a potent synthetic surfactant used in industrial applications which rapidly dissociates in biologic media to perfluorooctanoate [CFjfCFijCOi'], which is the anion of perfluorooctanoic acid [PFOA, CF^CFi^CO O H ]. In laboratory animals, PFOA and its salts are: 1) absorbed by ingestion, inhalation or dermal contact;1'3 2) distributed primarily in the liver and blood;4 3) not biotransformed, conjugated, incorporated into lipids or form coenzyme A conjugates;5'8 and 4) eliminated in the female rat at a greater rate of renal excretion than the male rat although no gender differences in excretion of PFOA have been seen in other laboratory animal species.1-4-9 In rats, administration of APFO results in peroxisome proliferation, uncoupling of mitochondrial oxidative phosphorylation and altered lipid metabolism.10-11-9 In a 90-day gavage study of rhesus monkeys, all animals in the 100 mg/kg/day group and 3 of the 4 animals in the 30 mg/kg/day group died before the end of study.1-12 Histopathologic examination revealed marked diffuse lipid depletion in the adrenals, slight to moderate hypocellularity of bone marrow, moderate atrophy of lymphoid follicles in the spleen and moderate atrophy of the lymphoid follicles of the lymph nodes in the two highest treatment groups. There were no histopathologic changes in the 0, 3 and 10 mg/kg/day dose groups. In lifetime feeding bioassays of rats,13-14 APFO in the diet at 300 ppm (daily dose of 15 mg/kg/day) increased the incidence of liver. Leydig cell and pancreas acinar cell adenomas. The liver and testicular tumors most likely occur via nongenotoxic mechanisms: oxidative stress and apoptosis in the development of the liver tumors; and enhanced hepatic aromatase activity which results in a hormone-mediated mechanism (increased estradiol) for the formation of Leydig cell tum ors.15' 17 The pancreas acinar cell adenomas were hypothesized to be a result of a mild but sustained increase in cholecystokinin (CCK) levels secondary to hepatic cholestasis.18 CCK has been shown in animal models to produce pancreatic hypertrophy, hyperplasia and neoplasia.19' 23 Hepatic toxicity and hypolipidemia have not been observed in ammonium perfluorooctanoate production workers.24,25 Gilliland and Mandel did report that PFOA 003468 may negatively modulate the effect alcohol has on high-density lipoprotein (HDL) levels and exacerbate the effect that obesity has on hepatic enzyme tests.25 This workforce was not found to be at an increased mortality risk for liver cancer or liver disease.26 However, there were four pancreatic cancer deaths compared to two expected deaths (Standardized Mortality Ratio 2.0, 95% Confidence Interval 0.5-5.0). One of these four pancreatic cancer deaths had worked in the building where ammonium perfluorooctanoate was produced. The purpose of this analysis was to examine several additional years of medical surveillance data at this ammonium perfluorooctanoate production plant in order to determine: 1) whether CCK levels are positively associated with serum PFOA levels among production employees; 2) whether PFOA results in clinical hepatic toxicity; and 3) whether PFOA may modulate hepatic responses to obesity and alcohol. METHODS PFOA Production Ammonium perfluorooctanoate production at this 3M plant began in 1947. Ammonium perfluorooctanoate, a white powder, is produced via a five-stage process: electrochemical fluorination; isolating and converting the chemical to a salt slurry; converting the slurry to a salt cake; drying the cake; and packaging. The greatest likelihood for exposure has occurred in the drying area. Subject Selection and Data Collection Voluntary medical surveillance examinations were offered to the fluorochemical production workers in 1993, 1995 and 1997. The total number of male subjects, by year, who participated in these three cross-sectional investigations were: 1993 (n = 111); 1995 (n = 80); and 1997 (n = 74). (There were too few female employees to include in the data analysis.) Eligible voluntary participation rates among these production workers ranged 003459 from approximately 50 (1997) to 70 (1993) percent. There were 68 subjects in common for 1993 and 1995; 21 subjects in common between 1995 and 1997 (lower number due to employee turnover and re-assignments); and 17 subjects in common for all three years. Surveillance activities included a self-administered questionnaire, measurement of height, weight and pulmonary function, standard biochemical and urinalysis tests, serum PFOA determination and several male reproductive hormone assays. The hormone data were collected in 1993 and 1995 and the findings have been reported elsewhere.27 Serum hepatic and lipoprotein-related biochemical tests included: alkaline phosphatase (IU/L); gamma glutamyl transferase (GGT, IU/L); aspartate aminotransferase (AST, IU/L); alanine aminotransferase (ALT, IU/L); total bilirubin (mg/dl); direct bilirubin (mg/dl); total cholesterol (mg/dl); low-density lipoproteins (LDL, mg/dl); high-density lipoproteins (HDL, mg/dl); and triglycerides (mg/dl). In 1997, employees' plasma CCK33 (pg/ml) levels were determined. CCK exists in various forms and lengths, although sulfated CCK-33 (i.e., a 33 amino-acid arrangement) appears to be the predominant form. For purposes of brevity, we will refer to CCK-33 as CCK. Employees were required to have fasted for 12 hours prior to their venipuncture. Because CCK analyses were not a standard analysis of the company's fluorochemical medical surveillance program, a study protocol was reviewed and approved by the company's human subjects committee and a signed, informed consent was obtained from each participant. Serum chemistries and hematology were evaluated at Allina Laboratories (Minneapolis, Minnesota). Plasma CCK was measured by direct radioimmunoassay by Inter Science Institute (Inglewood, California). Serum PFOA (i.e., perfluorooctanoate) was determined f by thermospray (1993 and 1995) and electrospray (1997) high-performance liquid chromatography/mass spectrometry methods.28,29 Data Analysis Simple and stratified analyses, analysis of variance (ANOVA), and multivariable regression techniques were used to evaluate linear and nonlinear associations between PFOA and the biochemical parameters with adjustment for potential confounding variables.'0 Various serum category levels were used for the stratified analysis with no 003490 significant differences observed based on cutpoints as high as > 30 ppm. However, the number of employees at > 30 ppm was 5 or fewer (based on year). For purposes of this report, employees were stratified into three PFOA categories (0 - <1 ppm; 1 - <10 ppm; and >10 ppm) in order to provide a greater number of employees in the highest (> 10 ppm) category. For multivariable regression analyses, PFOA, age, body mass index (BMI), alcohol use, and cigarette use were examined as both categorical and continuous variables. Alcohol use was analyzed as less than 1 drink per day, > 1 drink per day and non-response to this questionnaire item (almost all subjects reported between < 1 - 3 drinks/day). Linear and nonlinear transformations of PFOA were used to test for associations. In particular, the multivariable models employed by Gilliland and M andel25 were used to determine whether PFOA has a modulating effect on obesity or alcohol consumption in regards to hepatic serum chemistries (ALT, AST and GGT) and HDL, respectively. RESULTS Serum PFOA levels, by year, were; 1993 (mean 5.0 ppm, SD 12.2, median 1.1 ppm, range 0.0 - 80.0 ppm); 1995 (mean 6.8 ppm, SD 16.0, median 1.2 ppm, range 0.0 - 114.1 ppm); and 1997 (mean 6.4 ppm, SD 14.3, median 1.3 ppm, range 0.1 - 81.3 ppm). Provided in Table I are the mean, standard deviation, median and range of the employees' age, BMI and hepatic, cholesterol and lipoprotein serum chemistry data stratified by PFOA level and the year of the medical surveillance examination. Depending upon the surveillance year, one to two orders of magnitude of difference were observed between the means (and medians) of the lowest and highest serum PFOA categories. There was f no evidence for abnormal liver function tests, hypolipidemia or cholestasis associated with increasing employees' serum PFOA levels. Controlling for potential confounders, multivariable regression analyses did not suggest otherwise. Other measures including renal function, blood glucose and hematology were not associated with serum PFOA levels (data not shown). The mean CCK value was 50 percent lower among employees with serum PFOA values > 10 ppm than for those employees with serum PFOA levels < 1 ppm (Table I), Figure 003491 1 is a scatter plot of the natural log of CCK and PFOA. All but two CCK values were within the assay's reference range (up to 80 pg/ml). These two CCK values (80.5 pg/ml and 86.7 pg/ml) were of employees with 0.6 ppm and 5.6 ppm serum PFOA levels, respectively. Adjusting for potential confounding variables, we continued to observe a negative association between the natural log of CCK and serum PFOA levels (Table II) 'y although minimum variation was explained (R- = .08). Provided in Table HI are the multivariable regression models (as originally reported by Gilliland and Mandel with the 1990 medical surveillance data- ) which examined the potential modulating effect of PFOA on the association between alcohol and HDL. The coefficients of determination (R-, adjusted R-) were not large for any model. Based on these models, Table IV shows the change in HDL levels associated with a 10 ppm increase in serum PFOA levels among light and moderate drinkers (> 1 drink/day) compared to light drinkers (< 1 drink/day). [Note: total serum organic fluorine measurements, rather than serum PFOA, were used in 1990.] Unlike the findings from the 1990 data, the effect of alcohol use on increasing HDL levels was not blunted by a 10 ppm increase in PFOA in any of the subsequent medical surveillance examination years. ; Presented in Table V are multivariable regression models, including those originally reported using the 1990 surveillance data,25 regarding the potential effect of PFOA on hepatic enzyme responses to obesity as measured by BMI. Again, the coefficients of determination were not large for any model. In the 1990 surveillance data, ALT increased among obese (BMI = 35 kg/m2) but not non-obese (BMI = 25 kg/m2) workers f who had a 10 ppm change in serum PFOA (Table VI). However, this interaction was not observed in the 1993, 1995 or 1997 medical surveillance examinations. Likewise, we did not observe associations with AST or GGT (data not shown) as was reported in the 1990 medical surveillance examinations.23 003492 DISCUSSION We observed a negative association between serum PFOA and plasma CCK among 74 workers engaged in the production of ammonium perfluorooctanoate. This observation is opposite that proposed by Obourn et al who questioned whether chronic exposure to peroxisome proliferators, like PFOA, can cause pancreatic adenomas in the rat as the consequence of a mild but sustained increase in CCK levels secondary to hepatic cholestasis.17 We do not believe the negative association observed in our study represents an entirely different biological relationship than what was originally postulated because: 1) all CCK values observed in this study were within the assay's reference except for two values (which were not associated with high serum PFOA values); and 2) there was no suggestion of cholestasis which was considered the underlying reason for the elevated CCK levels in the rat. There are several explanations for the lack of a positive association between PFOA and CCK in our study. First and foremost, the serum PFOA measurements in these production workers may have been too low to cause an increase in CCK if such a mechanism exists in humans. Second, the mechanistic reason for the elevated CCK levels in the Obourn et al study17 was not clearly established. Obourn et al examined the effects of Wyeth 14,643, a more potent peroxisome proliferator than PFOA; thus their findings may not be directly related to PFOA. The clinical pathology data indicative of cholestasis included alterations in bile flow and bile acid output. Absolute bile flow and flow relative-to-body weight were marginally increased and acinar cell proliferation, altliough numerically increased at 3 months, returned to control levels at 6 months. Obourn et al also conducted in vitro experiments of both W yeth-14,643 and PFOA which argued against other biological pathways known to elevate plasma CCK levels including CCK-A receptor agonism, trypsin inhibition and increased dietary fat content. Third, the primary set of biochemical and cellular events identified in rodents susceptible to the hepatic tumor effects of peroxisome proliferators has not been identified in either liver biopsies from humans exposed to peroxisome proliferators or in in vitro studies with human hepatocytes; however, the peroxisome proliferator-activated receptor (PPAR-a) is 003493 expressed at very low levels in the human liver.17,31 It should also be noted that unlike the pancreas of the rat. the human pancreas has no detectable CCK-A receptors and little to no mRNA for the receptor.31'34 Finally, whether CCK initiates or promotes pancreatic cancer is a controversial issue." Data from more than seventy laboratory animal studies have variably suggested that CCK has positive trophic effects, inhibitory effects, or no involvement in pancreatic tumor growth.36 CCK has promoted growth of human pancreatic cancers in cell cultures.37 On the other hand, fasting plasma concentrations of CCK in unresected pancreatic cancer patients did not differ from healthy controls.38 Also acinar cell malignancies in rats are rare in the human.39 Activation of the c-K-ras gene is frequent in both human and hamster pancreatic cancer but is not found in azaserineinduced pancreatic cancer acinar cell adenoma models in the rat.40,41 Neither Gilliland and Mandel2'^ nor ourselves observed significant clinical hepatic toxicity associated with the serum PFOA levels measured in this workforce. In the three surveillance years we examined (1993, 1995 and 1997), 88 percent, 81 percent and 85 percent of those employees who volunteered had serum PFOA levels less than 10 ppm, respectively. In a laboratory study, no significant hepatic toxicity was reported for the lowest dose (3 mg/kg/day) group of rhesus monkeys administered APFO by gavage for 90 days.1,12 One of four primates in the next highest dose (10 mg/kg/day group) developed anorexia and black stools during the course of the study. There were no other abnormalities reported for this group. Only one animal survived in the 30 mg/kg/day group and none survived in the highest (100 mg/kg/day) dose group. For the 3 and 10 mg/kg/day dose groups, their mean serum total organic fluorine levels were 54 and 67 ppm, respectively.1,12 Sixty-nine percent of the molecular weight of PFOA is organic fluorine; therefore these total organic fluorine levels in the lower dose groups may correspond to serum PFOA levels of 80 to 100 ppm. Total organic fluorine levels were analyzed in the liver for two animals in each of the two lowest two dose groups and the means were 5 and 10 ppm, respectively. " Liver total organic fluorine levels were also analyzed for the 30 mg/kg/day (n = 4, mean = 98 ppm) and the 100 mg/kg/day (n = 2, mean = 213 ppm) dose groups; however, only the sole surviving animal in the 30 mg/kg/day group was analyzed for serum total organic fluorine (145 ppm which 003494 approximates 210 ppm of PFOA). The 30 mg/kg/day group did have higher hepatic transaminase values after 30 days of compound administration than the control group. Serum chemistries were not performed for the 100 mg/kg/day group after the onset of compound administration. Relative (% body weight) liver weights were higher among the 30 mg/kg/day (3.84%) and 100 mg/kg/day (3.31%) treatment groups than the control (2.36%), 3 mg/kg/day (2.36%) and 10 mg/kg/day (2.32%) dose groups. No abnormal liver function results were observed among employees with the highest serum PFOA levels; nevertheless these workers have been restricted from potential high exposure workplace areas. We were unable to replicate, in three subsequent medical surveillance examinations, an earlier investigation's finding that PFOA may modulate hepatic responses to obesity and alcohol.2'^ Total serum organic fluorine was used as a surrogate variable for PFOA in the 1990 medical surveillance exams. The use of total serum organic fluorine constitutes additional potential exposure to perfluorocarbons; however, data suggest that PFOA would represent the greatest fraction of total serum organic fluorine levels in this employee population.* Another explanation for the disparate findings, in particular as related to BMI, is that there may have been measurement error regarding BM I in the original study as well as in our investigation. We have previously noted the lack of an expected positive association between BMI and estradiol in the 1990 data.27 Yet in the present study we did not observe the anticipated strong positive correlation between BMI and ALT except in 1997. Correlation coefficients (p value in parentheses) were; 1993, r = .16, (p = .09); 1995, r = .13, (p = .27); and 1997, r = .43 (p = .0001). Self-reported alcohol data collected in the occupational setting can be questioned for its reliability as well as validity. The latter was not feasible to address; however, to partially examine the issue of reliability, we examined the analyses of the 68 employees who participated both in 1993 and 1995. The data showed good correlation between these two years for the potential confounding factors of BMI (r = .94, p = .0001), alcohol consumption (r = .67, p = .0001) and cigarette smoking (r = .84, p = .0001). 003495 Several additional issues are worthy of consideration. The medical surveillance program is voluntary. Overall participation rates declined from approximately 70 percent in 1993 to 50 percent in 1997. Serum PFOA levels could differ between participants and nonparticipants. The high turnover of employees between 1995 and 1997 detracted from the opportunity for a longitudinal assessment. In this regard, we did examine the change in several parameters among the 68 subjects in common for 1993 and 1995. For example, the average difference in serum PFOA was + 0.1 ppm (Wilcoxin signed-rank test = -540.5, p = .0001), the mean change in ALT was +0.5 IU/L (Wilcoxin signed-rank test = 50.0, p = 0.6) and the mean change in cholesterol was -1.6 mg/dL (Wilcoxin / signed-rank test = -60.0, p = 0.4). The change in serum PFOA levels did not predict, via regression analyses, the change in ALT or cholesterol. Besides the few employees in common across all three years of medical surveillance data (n = 17), another difference that occurred in 1997 was the method of analysis of PFOA changed from thermospray (1993 and 1995) to electrospray (1997) high-performance liquid chromatography mass spectrometry. Also, the laboratory reference range substantially changed for ALT in 1997 (as can be seen in Table 1 by examining the lower mean values for ALT in 1997). Finally, the issue remains that the lack of a clinical hepatotoxic effect reported by Gilliland and Mandel25 and ourselves does not rule out the possibility that PFOA may result in a subclinical hepatic effect in this production population that we have yet to observe. Results from ongoing laboratory animal studies, including a 6 month APFO gelatin capsule feeding study of cynomolgus primates, may provide further insight into the direction of medical surveillance activities for this workforce. REFERENCES 1. Griffith FD, Long JE. 1980. Animal toxicity studies with ammonium perfluorooctanoate. Am Ind Hyg Assoc J 41:576-583, 1980. 2. Kennedy G. Dermal toxicity of ammonium perfluorooctanoate. Toxicol Appl Pharmacol 81:348-355, 1985. 3. Kennedy G, Hall G, Brittelli J, Chen H. Inhalation toxicity of ammonium perfluorooctanoate. Fd Chem Toxicol 24:1325-1329, 1986. 4. Vanden Heuvel J, Kuslikis B, Van Refelghem M, Perterson R. Tissue distribution, metabolism and elimination of perfluorooctanoic acid. J Biochem Toxicol 6:83-92, 1991. C03496 5. PastoorTP, Lee KP, Perri MA, Gillies PJ. Biochemical and morphological studies of ammonium perfluorooctanoate-induced hepatomegaly and peroxisome proliferation. Exp Mol Pathol 47:98-109, 1987. 6. Ophaug R, Singer L. Metabolic handling of perfluorooctanoic acid in rats. Proc Soc Exp Biol Med 163:19-23, 1980. 7. Ylinen M, Koho A. Hanhijarvi H, Peura P. Disposition of perfluorooctanoic acid in the rat after single and subchronic administration. Bull Environ Contam Toxicol 44:46-53, 1990. 8. Kuslikis BI, Vanden Heuvel JP, Peterson RE. Lack of evidence for perfluoroodecanoyl- or perfluorooctanoyl-coenzyme A formation in male and female rats. J Biochem Toxicol 7:25-29, 1992. 9. Hanhijarvi H, Phaug R, Singer L. The sex-related difference in perfluorooctanoate excretion in the rat. Proc Soc Exp Biol M ed 171:51-55, 1982. 10. Haughom B, Spydevold O. The mechanism underlying the hypolipmie effect of perfluorooctanoic acid (PFOA), perfluorooctane sulphonic acid (PFOSA) and clofibric acid. Biochim Biophys Acta 1128:65-72, 1992. 11. Keller B, Marsman D, Popp J, Thurman R. Several nongenotoxic carcinogens uncouple mitochondrial oxidative phosphorylation. Biochim Biophys Acta 1102:237-244,1992. 12. Goldenthal El, Jessup DC, Geil RG, Mehring JS. Ninety-day Subacute Rhesus Monkey Toxicity Study. Study No. 137-090. International Research Development Corp, Mattawan (MI), November, 1978. 13. Sibinski LJ. Two-year oral (diet) toxicity/carcinogenicity study of fluorochemical FC-143 in rats. Riker Laboratories, St. Paul (MN), 1987. 14. Rao MS, Reddy JK. Hepatocarcinogenesis of peroxisome proliferators. Ann N Y AcadSci 804:573-587,1997. 15. Cook JC, Murray SM, Frame SR, Hurtt ME. 1992. Induction of Leydig cell adenomas by ammonium perfluorooctanoate: a possible endocrine-related mechanism. Toxicol Appl Pharmacol 113:209-217,1992. 16. Cook JC, Hurtt ME, Frame SR, Biegel LB. 1994. Mechanisms of extrahepatic tumor induction by peroxisome proliferators in CrkCD BR (CD) rats. Toxicologist 14:301, 1994. 17. Biegel LB, Liu RCM, Hurtt ME, Cook JC. Effects of ammonium perfluorooctanoate on Leydig cell function: in vitro, in vivo, and ex vivo studies. Toxicol Appl Pharmacol 134:18-25, 1995. 18. Obourn JD, Frame SR, Bell RH, Longnecker DS, Elliott GS, Cook JC. Mechanisms for the pancreatic oncogenic effects of the peroxisome proliferator W yeth-14,643. Toxicol Appl Pharmacol 145:425-436, 1997. 19. Pandol SJ. Pancreatic physiology and secretory testing. (In) Sleisenger M, Fordtran JS (eds) Gastrointestinal and Liver Diseases, Volume 1, W.B. Saunders Co., Philadelphia, 1998, pages 771-782. 20. Longnecker DS. Experimental models of exocrine pancreatic tumors. (In) Go VLW (ed) The Exocrine Pancreas: Biology, Pathology and Diseases, Raven, New York, 1986, pages 443-458. 21. Longnecker DS. Experimental pancreatic cancer: role of species, sex and diet. Bull Cancer 77:27-37, 1990. 003497 22. Pour PM, Runge RG, B in D, Gingell R, Lawson T, Nagel D, Wallcave L, Salmasi S. Current knowledge of pancreatic carcinogenesis in the hamster and its relevance to the human disease. Cancer 47:1573-1587, 1981. 23. Pour PM, Lawson T, Helgeson S, Donnelly T, Stepan K. Effect of cholecystokinin on pancreatic carcinogenesis in the hamster model. Carcinogenesis 9:597-601, 1988. 24. bel F, Sorenson S, Roach D. Health status of plant workers exposed to fluorochemicals: A preliminary report. Am Ind Hyg Assoc 4 1:584-589, 1980. 25. Gilliland FD, Mandel JS. Serum perfluorooctanoic acid and hepatic enzymes, lipoproteins and cholesterol: a study of occupationally exposed men. Am J Ind M ed 129:560-568, 1996. 26. Gilliland FD, Mandel JS. Mortality among employees of a perfluorooctanoic acid production plant. J Occup Med 35:950-954,1993. 27. Olsen GW, Gilliland FD, Burlew MM, Burris JM, Mandel JS, Mandel JH. An epidemiologic investigation of reproductive hormones in men with occupational exposure to perfluorooctanoic acid. J Occ Env Med 40:614-622, 1998. 28. Johnson JD, Wolter JT, Colaizy GE, Rethwill PA, Nelson RM. Quantification of perfluorooctanoate and perfluorooctanesulfonate in human serum using ion-pair extraction and high performance liquid chromatography-thermospray mass spectrometry with automated sample preparation. 3M Company, St. Paul (MN), 1996. 29. Anderson DJ, Mulvana DE. Analytical report for the determination of perfluorooctanaoate and perfluorooctanesulfonate in human serum by LC/MS. Advanced Bioanalytical Services, Inc., Ithaca (NY), 1997. 30. SAS Institute, Inc. SAS Users Guide: Statistics. Version 6. SAS Institute, Inc., C ary(NC), 1990. 31. Cattley RC, DeLuca J, Elcombe C, Fenner-Crisp P, Lake BG, Marsman DS, Pastoor TA, Popp JA, Robinson DE, Schwetz B, Tugwood J, Wahil W. Do peroxisome proliferating compounds pose a hepatocarcinogenic hazard to humans? Reg Toxicol Pharmacol 27:47-60, 1998. 32. Wank SA, Pisegna JR. deWeerth A . Cholecystokinin receptor family. Ann NY Acad S c i l 13:49-66, 1994. 33. Gavin CE, Martin NP, Schlosser MJ. Absence of specific CCK-A binding sites on human pancreatic membranes. Toxicologist 30:334,1996. 3 4 / Gavin CE, Malnoske JA, White J, Schlosser MJ. Species differences in expression of pancreatic cholecystokinin-A receptors. Toxicologist 36:232 1997 35. Axelson J, Ihse I, Hakanson R. Pancreatic cancer: the role of cholecystokinin? Scand J Gastroenterol 1992;27:993-998. Toxicol Appl Pharmacol 134:18-25, 1995. 36. Herrington MK, Adrian TE. On the role of cholecystokinin in pancreatic cancer. Int J Panereatol 17:121-138, 1995. 37. Palmer-Smith J. Krame ST, Solomon TE. CCK stimulates growth of six human pancreatic cancer cell lines in serum-free medium. Reg Peptides 22:341-349, 1991. 38. Rehfeld JF. van Solinge WW. The tumor biology of gastrin and cholecystokinin. Adv Cancer Res 63:295-347,1994. 003498 39. Anderson KE, Potter JD, Mack TM. Pancreatic cancer. (In) Schottenfeld D, Fraumeni IF (eds) Cancer Epidemiology and Prevention (2nd edition). Oxford University Press, New York, 1996, pages 725-771. 40. van Kranen HJ, Vermeulen E, Schoren L, Bas J, Woutersen RA, van Iersel P, van Kreijl CF, Scherer E. Activation of c-K -raj is frequent in pancreatic carcinomas of Syrian hamsters, but is absent in pancreatic tumors of rats. Carcinogenesis 12:1471482, 1991. 41. Caldas C, Kern SE. K-ras mutation and pancreatic adenocarcinoma. Int J Pancreatol 18:1-6, 1995. i i 003499 Table 1. Mean, Median, Standard Deviation (S.D.) of Mean and Range of Demographic, Hepatic, Cholesterol and Lipoprotein Values by Serum PFOA Categories and Medical Surveillance Year (1993, 1995, 1997) PFOA* Category (ppm) Mean 1993 Median S.D. Range 0-<l 1 -<10 > 10 0.48 3.38 30.88 0.48 2.50 19.50 0.27 2.17 25.12 F Value = 68.3, p = .0001 0.00 - 0.99 1.03-8.92 11.90-80.00 1995 Mean Median S.D. PFOA (ppm) 0.31 3.03 30.06 0.20 2.40 25.50 0.32 1.84 26.58 F Value = 39.1, p = .0001 Range 0.00 - 0.90 1.10-8.20 10.3-114.1 1997 Mean Median S.D. Range 0.47 3.13 32.13 0.55 2.30 22.52 0.26 2.12 24.84 F Value = 48.7, p = .0001 0.05 - 0.92 1.05-7.66 10.50-81.35 0 -< l 1- <10 > 10 43 44 9.2 39 38 7.8 39 38 6.6 F Value = 3.7, p = .02 27-61 27-60 25-49 Age 42 41 8.3 41 40 8.6 43 45 8.4 F Value = 0.2, p = .85 29-60 24-58 27-55 40 39 9.1 41 41 8.7 42 46 10.9 F Value = 02, p = 0.81 25-61 26-58 28-57 0-<l 1 - <10 > 10 28.0 27.6 26.9 26.3 28.4 28.5 4.3 2.5 2.4 F Value = .14 20.9-42.0 21.6-32.5 22.4-32.0 BMI (kg/m2) 27.6 26.8 28.6 27.9 28.4 28.8 4.2 3.4 3.5 F Value = 0.7, p = .51 21.9-45.2 22.1 -38.3 21.2-34.8 28.7 27.5 29.5 29.5 27.6 28.5 3.9 5.3 3.5 F Value = 0.8, p = .47 21.5-35.0 21.9-46.8 22.0 - 33.0 II a. C03500 0 -< l 1- <10 > 10 Not done in 1993 CCK (pg/ml) Not done in 1995 33.4 31.3 15.7 28.0 21.1 19.2 17.4 16.4 5.6 F value = 3.8, p = 0.03 13.4-80.5 8.8 - 86.7 11.4-29.9 003501 PFOA* Category (ppm) Mean 1993 Median S.D. 0-<l 1 - <10 > 10 88 82 26 82 78 23 83 75 24 F Value = 0.7, p = .52 1995 Range 3 7 - 161 47-151 58-132 Mean Median S.D. Alkaline Phosphatase (IU/L) 78 76 18 80 76 25 89 76 31 F Value = 1.4, p = .25 Range 40-114 48 - 165 55 - 146 0-<l 1 - <10 > 10 33 30 19 50 30 70 37 33 17 F Value = 1.5, p = .24 11-84 6-472 19-77 GGT (RJ/L) 42 34 27 51 36 41 40 38 13 F Value = 0.9, p = .41 16- 149 19-190 21-61 0-<l 1 - <10 > 10 23 22 7 26 24 11 24 24 5 F Value = 1.1, p = .33 11-60 12-83 16-35 AST (IU/L) 21 20 6 24 20 13 21 20 4 F Value = 0.8. p = .45 13-36 13-75 15-29 0-<l 1 - <10 > 10 45 42 14 48 45 29 46 47 8 F Value = 0.2, p = .82 22-88 22-221 35 - 62 ALT (IU/L) 44 40 13 53 40 34 47 48 12 F Value = 1.2, p = .30 27-80 27-175 28-71 1997 Mean Median S.D Range 79 78 19 87 84 23 80 73 23 F Value = 1.3,p = .28 27 - 122 47 - 164 61 - 142 34 25 27 36 32 25 30 29 10 F Value = 0.2, p = .78 15-230 14-162 15-50 26 23 25 25 25 25 7 7 4 F Value = 0.2, p = .83 13-41 14-48 22-34 31 30 10 33 28 15 35 31 13 F Value = 0.3, p = .73 13-59 14-80 18-57 1993 1995 PFOA* Mean Median S.D. Range Mean Median S.D. Category (ppm) __________________________________________________________________ LDL (rng/dl) Range 0-<l 1 - <10 > 10 138 142 143 137 140 143 F Value = 0.2, p = .84 40 38 42 27 - 227 72 - 223 60-188 131 130 133 137 130 118 32 40 39 F Value = 0.1,p = .96 31 - 191 28-210 62-211 0-<l 1 -<10 > 10 171 145 124 205 129 408 221 223 159 F Value = 0.3,p = .77 37 - 636 47 - 2845 41-564 Triglycerides (mg/dl) 170 152 93 175 123 144 254 183 154 F Value = 2.7, p = .07 57-371 59 - 743 77 - 563 1997 Mean Median S.D. Range 114 119 134 134 134 127 28 44 38 F Value = 2.3, p = .11 4 0 - 158 26-253 79 - 206 219 186 176 161 251 188 116 85 192 F Value = 21, p = .13 53-445 44-360 63-718 * Study population size by PFOA category (ppm) and year Sample Size PFOA Category 0< 1 1 -< 10 > 10 Total 1993 52 46 13 111 1995 39 26 15 80 1997 29 34 11 74 003502 Table II. Multiple Regression M odel o f Factors Predicting the Natural Log o f Plasm a Cholecystokinin in W orkers with Serum PFO A Levels Variable Intercept PFOA Age Alcohol BMI Cigarettes B 3.02 -0.008 0.0001 -0.005 0.009 -0.009 SEfB) 0.48 0.004 0.007 0.086 0.015 0.007 p value .0001 .07 .98 .95 .53 .17 R2 = .08 Adj R2 = .02 1 003503 Table III. Multivariable Regression Models of Factors Predicting High Density Lipoprotein in Workers with Serum PFOA Levels 1990 1993 1995 1997 Variable B SE(J3) p value B SE(B) p value B SE(B) p value B SE(5) p value Intercept PFOA Light Drink Interaction* 65.00 -1.61 -9.92 1.62 10.07 0.77 3.51 0.80 .0001 .04 .006 .04 55.00 -0.14 -4.83 0.02 16.70 0.33 3.76 0.35 .001 .67 .20 .96 52.10 -0.10 -5.11 0.02 11.92 0.08 2.61 0.13 .0001 .18 .05 .87 57.82 -0.19 -5.75 0.36 10.15 0.13 3.13 0.18 .0001 .16 .07 .05 > & ii UJ R2= .17 Adj R2= Not reported r 2= .io Adj R2= .02 R2= .30 Adj R2= .19 R2= .11 *Adjusted for age, BMI, cigarette use, and non-respondents to alcohol question (all four years) and testosterone level (1990, 1993 and 1995). The 1990 regression model used total organic fluorine as the dependent variable (see reference #25). *Interaction = PFOA x Light Drink G03504 Table IV. Change in HDL* from Light Alcohol Drinker (< 1 drink/day) to: 1) a Light Drinker with a 10 ppm Change in Serum PFOA; 2) a Moderate Alcohol Drinker (> 1 drink/day); and 3) a Moderate Alcohol with a 10 ppm Change in Serum PFOA. Light Drinker with a 10 pm increase in PFOA Moderate Drinker Moderate Drinker with 10 ppm increase in PFOA Year Chanee in HDL fme/dli 1990# 0.1 9.9 -6.2 1993 -1.3 4.8 3.4 1995 -0.8 5.1 4.1 1997 1.7 5.8 3.9 `Determined from multivariable models (see table 3) adjusted for age, body mass index cigarette use, and non-respondents to alcohol question (all four years) and testosterone (1990, 1993 and 1995 only). # Data in 1990 analyzed for total serum organic fluorine (see reference #25). 003505 Table V. Multivariable Regression Models* of Factors Predicting ALT in Workers with Serum PFOA Levels 1990 1993 1995 1997 B SE(B) p value B SE(Z?) p value B SE() p value B SE(fi) p value Intercept PFOA BMI Interaction* 58.13 -15.80 0.30 0.62 24.60 4.58 0.82 0.17 .02 .0008 .72 .0004 27.32 0.89 1.07 -0.03 19.13 2.88 0.67 0.10 .16 .76 .11 .79 40.93 0.81 1.08 -0.03 22.90 2.62 0.74 0.09 .08 .75 .15 .76 3.93 2.77 1.54 -0.09 11.34 1.27 0.35 0.42 .73 .03 .0001 .04 R2= .21 Adj R1= Not reported R2= .06 Adj RJ = .01 R2= .ll Adj RJ = .01 R2= .32 Adj RJ = .26 'Adjusted for age, alcohol and cigarette use. The 1990 regression model used total organic fluorine as the dependent variable (see reference #25). "interaction = PFOA x BM1 003506 Table VI. Change in Alanine Aminotransferase (ALT)* Associated with a 10 ppm Change in Serum PFOA for Three Body Mass Indices _________ BMI (kg/m2l 25 30 35 Year 1990* 1993 1995 1997 Chanee in ALT CIU/L! -3.0 28.0 59.0 2.2 0.9 -0.5 1.5 0.1 -1.2 5.3 0.8 -3.7 'Determined from multivariable regression model (see table 5) adjusted for age, alcohol and cigarette use. * Data in 1990 analyzed total serum organic fluorine (see reference #26). > 003507 Figure 1. Scatter Plot of Nautural Log CCK (pg/ml) by Serum PFOA Level (ppm) [Linear Regression Model: Ln CCK = 3.24 - 0.006PFOA; p value of PFOA coefficeint = 0.19; R2= .02] / 003508 Ln CCK (pg/ml) o KJ LO u> 4--- --- - < MO UO> 3 2>rao<L- oai *D w3. oo> o''I ooo o(O oo ti Co (*