Document YKrppk55rVoVDJGN8d1Jve80

Corporate Occupational Medicine 3M Center, Building 220-3W-05 St. Paul, MN 55144-1000 651 737 4230 Telephone 651 733 9066 Fax ARZK-OH75 7. Serum Perfluorooctanoic Acid and Hepatic Enzymes, Lipoproteins, and Cholesterol: A Study of Occupationally Exposed Men. This is the published paper (Gilliland and Mandel; Am J Ind Med 1996;29:560-568) from the Gilliland doctoral dissertation (see study # 5) that evaluates the effect of PFOA on hepatic enzymes and lipoproteins. This study of 115 occupationally exposed workers examined the cross-sectional associations between serum PFOA (measured as serum total organic fluorine) and hepatic enzymes, lipoproteins, and cholesterol. There was no significant clinical hepatic toxicity at the PFOA (i.e., total organic fluorine) levels measured in this study. Based on multivariable models, PFOA may exacerbate the effect obesity has on liver transaminase tests and blunt the effect that alcohol has on HDL levels. [Note: In three subsequent time periods, the findings of this study could not be replicated. PFOA, assayed by mass spectrometry methods, did not appear in this production workforce to modulate hepatic responses to either obesity or alcohol consumption. See study #8 below.] 003477 \ AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 29:560-568 ( I 996) Serum Perfluorooctanoic Acid and Hepatic Enzym es, Lipoproteins, and Cholesterol: A Study of Occupationally Exposed Men Frank D. Gilliland, m o , P h o and Jack S. Mandel, PhD, m p h Perfluorooctanoic acid (PFOAI produces marked hepatic effects, including hepatomegaly, focal liepuiocyte necrosis, hypolipidemia. and alteration of hepatic lipid metabolism in a 'number o f animal species. In rodents. PFOA is a peroxisome proliferulor. an inducer of members o f the cytochrome P4S0 superfamUy and other enzymes involved in xenobiotic metabolism, an uncoupler of oxidative phosphorylation, and may be a cancer promoter. Although PFOA is the major orgunofluorine compound found in humans, tittle information is available concerning human responses to PFOA exposure. This study o f 115 occupationally exposed workers examined the cross-sectional associations between PFOA and hepatic enzymes, lipoproteins, and cholesterol. The findings indicate that there is no significant clinical hepatic toxicity at the PFOA levels observed in this study. PFOA may modulate the previously described hepatic responses to obesity and xenobiotics. /wrt Witry-Usx. Inc. KEY WORDS: perfluorooctanoic acid, human, hepatic enzymes, cholesterol, HDL INTRODUCTION Little is known about the toxic potential of PFOA in hu mans; however, studies have shown that the liver is an Perfluorooctanoic acid (PFOA) is a potent synthetic important site of toxicity in animals (Griffith and Long. surfactant that is used in a wide variety of industrial pro 1980; Kennedy. 1985; Kennedy et al.. 1986; Pastoor et al., cesses and products. Organic fluorine has been found in the 1987; Van Rafelghem et al.. 1987; Just et al.. 1989). serum of all human populations studied (Ubel et aL 1980; Animals treated with PFOA rapidly develop hepatome Taves. 1971; Taves et aL 1976: Guy. 1979. Belisle. 1981 j. galy with focal necrosis and show marked hepatic physio Guy and Taves reported that PFOA was the principal or logic responses that include hypolipidemia. peroxisome ganic fluorine compound in human serum (Taves. 1971; proliferation, induction of xenobiotic metabolic enzymes, Taves et al,, 1976: Guy. 1979). PFOA is found in serum increased hepatic tumor incidence, uncoupling of mitochon because PFOA has a lone biological half-life, allowing ac drial oxidative phosphorylation, and alterations in lipid me cumulation of small doses over time (Ubel et al.. 1980). tabolism (Griffith and Long, 1980: Kennedy. 1985; Kennedy et al.. 1986: Pastoor et al.. 1987: Van Rafelghem et al.. 1987: Just et al.. 1989: Takagi et al.. 1991. Permadi et al.. 1992: Haughom et al.. 1992: Sohlenius et al.. 1992: Oivision o< Environmental and Occuoabonai Health. School ol Puttie Health, University ot Minnesota. Minneaottts iF.D.G.. J.S.M .). Keller etal.. 1992: Handler. 1992). Rats treated with PFOA Oeoanment of internal Medicme. Occupational and Environmental Medicine and other peroxisome proliferators (PPs). such as clofibrate. Section. S I Paul Ramsey Medical Center. St. Paul. Minnesota iF.D.G.). 'ddres* repnnt requests to Frank 0. Gilkland. University ol New Meitco School tte n e . Ettdemioiojy and Cancer Control Program. 900 Cammo de Saiud show a MYX reduction of serum cholesterol and changes in the hepatic production and processing of lipoproteins. JbuQuerque. NM 67121. 'H aughom et al. (1992) showed that the hypolipidemic re Accepted tor putticaDon April 25, 1995 sponse results from downregulution of HMG-CoA reduc- t 1996 Wiley-Uss. me. 003478 PFOA, Cholesterol, Lipoproteins, and Hepatic enzymes 561 T A B L E I- Distribution of Exposed Workers by Tout Serum Fluorine Category in 3M Chemolite Plant. Cottage Grove. MN Age* BMI (kg/m*)' Alcohol use* <! 02/flay 1-3 oz/day Nonresponse Tobacco use* Smoker Nonsmoker Nonresponse Total <1 3 9 .9 (1 0 .2 ) 27.6 (5.3) 1 7 (7 3 .9 ) 2 (8.7) 0 (0 ) A (17.4) 3 (13.0) 1 9 (8 2 .7 ) 1 (4.3) 23 (100) 1-3 39.6 (8.5) 26.6 (2.6) 51 (78.5) 13 (20.0) 0 (0 ) 1 (1.5) 16 (24.6) 49 (75.4) 0 (0 ) 65(100) Total serum fluorine (ppm) >3-10 >10-15 36.0 (7.5) 26.3 (3.3) 39.3(11.1) 29.4 (3.7) 9 (56.3) 4 (25.0) 0 (0 ) 3(18.7) 5 (83.3) 1 (16.7) 0 (0 ) 0 (0 ) 6 (37.5) 9(56.2) 1 (6.3) 16(100) 2 (33.3) 4 (66.7) 0 (0 ) 6(100) >15-26 41 6 (1 0 .5 ) 26.0 (1.4) 5 (100) 0 (0 ) 0 (0 ) 0 (0 ) 1 (20.0) 4 (80.0) 0 (0 ) 5 (100) Total 39.2 26.9 87 (75.6) 2 0 (1 7 4) 0 (0 ) 8 (7.0) 85 (73.9) 28 (24.4) 2 (1.7) 115 BMI. body nass index. Values art mtan (SO). Values art n (percent). tase. in addition. PFOA has been associated with hepatocyte TABLE I I. Distribution of Age. Alcohol, and Tobacco Use in Participants necrosis and increased hepatic enzymes, suggesting that ir by Body Mass Index in Study of Workers Exposed to PFOA reversible cell damage occurs (Kennedy, 1985; Just et aL 1989). Hepatomegaly and alterations in lipid metabolism BMI mg/kg* appear to be rapidly reversible; however, other hepatic changes are not rapidly reversed (Perkins. 1992; Sohlenius <25 25-30 >30 et al.. 1992). Based on findings from the studies of rodents and in vitro experiments, some investigators have suggested that PFOA is likely to present a health risk to humans (Just et al., 1989; Takagi et al., 1991). If the observations in rodent species are relevant to humans exposed to PFOA, it is rea sonable to hypothesize that changes in human hepatic en zymes and lipid metabolism are similar to those observed in rodents. Limited data are available to assess the hepatic responses to PFOA in humans. Ubel and coworkers (1980) and Griffith and Long (1980) reported that PFOA-exposed workers showed no clinical evidence of adverse hepatic effects. Furthermore, a retrospective cohort mortality study of exposed workers found no excess mortality from liver cancer or liver disease (Gilliland and Mandel. 1993). To assess whether the changes in cholesterol, lipoproteins, and Total Tobacco use Smoker Nonsmoker Nonresponse Alcohol use <1 oz/day 1-3 oz/day Nonresponse Age <40 years 240 years Total serum flourine Mean ppm (SD) 41 (100% ) 11(26.8% ) 29 (70.7%) 1 (2.5%) 31 (75.6% ) 6 (14.6%) 4 (9.8%) 31 (75.6% ) 10 (24.4% ) 2.8 (3.7) 57 (100%) 17(100% ) 15 (26.3% ) 41 (7 1.9% ) 1 (1.8%) 2(11.8% ) 15 (88.2% ) 0(0% ) 43 (75.4% ) . 13 (76.4% ) 11 (19.3% ) 3 (17.7%) 3 (5.3%) 1 (5.9% ) 2 8 (49.1%) 29 (50.9%) 6 (35.3%)* 11 (8 4.7% ) 4.0 (5.5) 2.1 (3.5) hepatic enzymes observed in rodents treated with PFOA *p .005. occur in humans, we studied I IS occupationally exposed BMI. body mass mdex employees at a plant that produces PFOA. Production work ers with the highest PFOA exposures had serum PFOA levels similar to those in rodents that developed hepatome MATERIALS AND METHODS galy when treated orally with low doses of PFOA (Ubel ei al.. 1980). We examined the cross-sectional association be- Subject Selection ten serum PFOA. a validated surrogate measure of total rum fluorine, and cholesterol, lipoproteins, and hepatic Participants were recruited from current employees at a enzymes in this group of occupationally exposed men. PFOA production plant that has operated since 1947. The # 003479 S62 Gilliland and Mandel TABLE III. Serum Cholesterol. Low Density lipoprotein and High Density Lipoprotein Py Total Serum Fluoride in SluOy o! Workers Exposed to PFOA Total fluoride N Mean SO Median Range Test* TABLE V. Serum Cholesterol by BbOy Mass Inoex Age Smoking ana Drinking Status 3M Chemoiite Plant. Cottage Grove Minnesota Cholesterol |mg/dl) N (% ) Mean SO Median Range Test* Cholesterol (mg/dl) <1 ppm 23 1 -3 >3-10 65 16 >10-15 6 >15-26 5 Total 115 LDL (mg/dl) <1 ppm 23 21-3 65 >3-10 16 >10-15 6 >15-26 5 Total 115 HDL (mg/dl) <1 ppm 23 21-3 ' 65 >3-10 16 >10-15 6 >15-26 5 Total 115 201 211 206 226 214 210 132 136 134 124 143 135 45.9 46.1 41.8 46.5 45.6 45.4 34.7 40.0 37.7 40.0 27.0 38.1 32.4 34.5 34.5 44.0 20.B 33.6 11.7 10.0 10.2 6.8 10.2 10.2 203 212 198 216 204 210 137 131 133.5 139 144 134 47 44 40 44 49 43 132-268 130-349 150-277 163-298 184-244 130-349 F *.06 6 p < .62 BMI <25 25-30 >30 Age 30 >30-40 41 (35 7) 195 40.1 186 130-277 57 (49.6) 219 36.2 220 146-349 17(14.8) 214 29.3 216 163-268 21 (18.3) 196 37.8 201 130-254 48(41.7) 2.9 43.8 204 132-349 70-196 70-264 83-217 36-156 117-171 36-264 F * 0.31 p * .87 >40-50 >50-60 Alcohol <1 oi/d 1-3 oi/d Missing Tobacco 27 (23.5) 216 30.2 216 163-263 19(16.5) 219 29.7 224 164-268 87 (81.3) 209 38.6 204 135-349 20(18 .7) 216 33.5 218 130-277 8 207 45.5 213 132-261 Smoker 28 (24.8) 233 41.6 238 167-349 10-67 F > 0.66 Nonsmoker 85 (75.2) 203 32.9 203 130-268 30-79 p * .66 Missing 2 198 89.1 198 135-261 29-68 Total 115 40-59 29-54 19-79 Anova. BMI. body mass index. F = 5 10 p * .008 F . 1.60 p .19 F .63 p * .43 F * 15.63 p .0001 Anova. LOL low density lipoprotein; HOL lugn density lipoprotein. period were considered highly exposed. This group in cluded maintenance and engineering supervisors, as well as TABLE IV. Pearson Correlation Coefficients Between Total Serum Fluoride. Age. Body Mass Index. Daily Alcohol Use, Daily Tobacco Consumption, and Lipoproteins production workers. Forty-eight (96%) of 50 exposed work ers agreed to participate in the study, in addition, a sample of workers employed in jobs with no apparent PFOA expo sure was asked to participate. Those without direct contact Total Agt BMI Alcohol Tobacco lluoride (ppm) (y e a n ) (kg/m1) (oi/day) (clgs/day) with PFOA for at least 5 years were considered to have low exposure. A randomly selected low-exposure group of workers was frequency matched in 5-year age groups to the CHOLESTEROL .07 .25 .19 .09 .35 high-exposure workers. Sixty-five employees from jobs p .008 p * .05 p > .0001 thought to involve no PFOA exposure volunteered for the LDL .02 .13 .06 -.0 0 8 .28 study. The total number of the presumed unexposed em p .00 ployees invited to participate was not recorded: however, HDL -.01 .03 - .1 3 .18 - .0 9 few individuals in this group declined to participate. We p * .06 estimate that more than 80% of those invited agreed to participate in the study. LOL. low-density lipoprotein: HOL. high-density lipoorotem: BMI. body mass index. Total serum fluorine was used as a surrogate variable for PFOA exposure. We assayed total serum fluorine rather than measuring PFOA directly because the assay was less plant produces a number of specialty chemicals in addition expensive and technically easier to perform on the large to PFOA. Details about the plant have been described pre- number of samples collected in this study. Furthermore, the ously (Gilliland and Mande!. 1993). All workers em use of total serum fluorine has been validated as a surrogate ployed in PFOA production during the period 1985-1989- marker for PFOA in past biological monitoring in the plant were invited to participate in the study. Workers with jobs and other plants using PFOA (Ubel et al.. 1980). Approxi involving direct contact with PFOA during the 1985-1989 mately 90% of total serum fluorine in workers was reported CQ3480 PFOA, Cholesterol, Lipoproteins, and Hepatic enzymes S63 TABLE VI. Serum Low Density lioooroiem by Body Mass index Age. Smoking, and Drinking Status: 3M Ctiemolite Plant, Cottage Grove. Minnesota lD L(m g /d l) N (% ) Mean SO Median Range Test* BMI <25 25-30 41 (35.7) 130 22.8 133 70-217 F .65 57 (49.6) 138 34.2 135 36-264 p .52 >30 17(14.8) 136 33.0 137 71-196 Age $30 >3<M0 21 (18.3) 130 29.6 131 75-177 F . .37 48(4 1 .7 ) 136 36.2 135 70-264 p .77 >40-50 27 (23.5) 133 34.5 135 36-193 >50-60 19(16 .5) 140 32.3 137 20-196 Alcohol <1 oz/d 8 7(81 .3) 135 34.5 1-33 oz/d 20(1 8 .7 ) 135 31.4 133 36-264 F .01 137 78-217 p > .93 Missing 8 134 35.6 141 70-174 Tobacco Smoker 28 (24.8) 152 35.6 Nonsmoker 85 (75.2) 130 31.3 Missing 2 115 55.9 146 99-264 F 9.42 133 78-217 P .003 115 70-174 Total 115 Anova. BMI, body mass miti: LD L low density lipoprotein. TABLE VII. Serum Hign Oensity Lipoprotein by Booy Mass inoex Age Smoking, and Drinking Status HOL tmg/dl) N (% l Mean SD Median Range BMI <25 25-30 >30 Age 41 (35.7) 46.0 107 43 19-68 57 (49.6) 45.6 10.5 44 22-79 17(14.8) 43.6 7.7 43 32-55 30 >30-40 >40-50 >50-60 Alcohol 21 (18.3) 43.5 14.3 40 19-79 48 (41.7) 46.7 9.9 46 22-65 27 (23.5) 46.0 8.3 45 29-61 19 (16.5) 46.6 7.9 43 32-67 <1 oz/d 87 (81.3) 44.3 9.2 43 19-65 1-3 oz/d 20 (1 8 .7 ) 49.3 13.5 45 29-79 Missing 8 48.3 9.3 53 32-55 Tobacco Smoker 28 (24.8) 44.3 8.9 43 29-68 Nonsmoker 85 (75.2) 45.6 10.6 43 19-79 M is sin g 2 54.5 21.2 55 53-56 Total 115 Tesi* F = .38 p 69 F 72 p * .55 F 3.88 p .05 F . .35 P * .56 Anova. BMI, body mass index: HOL hign density lipoprotein. to be in the form of PFOA (Venkateswarlu, 1982). Because the vast majority of total serum fluorine in plant employees is in the form of PFOA, total serum fluorine closely reflects serum PFOA in production workers, and its use is unlikely to introduce substantial error into the study. We expected the group of workers who were selected for the unexposed group based on job history to have total serum fluorine levels similar to the general population. However, we found that this group of workers was not unexposed, having levels 20-50 times higher than levels reported for the general population. We concluded that job histoty was not an accurate metric for exposure. Because job history performed poorly for exposure assessment, we used measured total serum fluorine to classify individuals in the analyses. Data Collection Participants completed a medical history questionnaire, were measured for height and weight, and donated a blood sample by venipuncture for assays of total serum Fluorine. rum glutamyl oxaloacetic transaminase (SCOT), serum glutamyl pyruvic transaminase (SGPT). gamma glutamyl transferase (GGT). cholesterol, low-density lipoproteins (LDL). and high-density lipoproteins (HDL). The blood sample for total serum fluorine (TSF) was collected in a fluorine-free 15-m! Vacutainer. Divided aliquots of serum collected for total fluorine assay were frozen at --70C. After all total fluorine samples had been received, batches of 15 samples were assayed on successive work days. Total serum fluorine, reported as a mean value, was determined using sodium biphenyl extraction and atomic absorption spectroscopy (Venkateswarlu. 1982). Each sample was as sayed twice. Each batch included high- and low-quality con trol samples. Analysis Stratified analysis. Anova. Pearson correlation coeffi cients. and linear multivariate regression were used to eval uate associations between PFOA and the biochemical end points. For stratified analyses. Anova procedures were used to assess differences in mean values. Total serum fluorine was divided a priori into five categories-- <1 ppm. 1-3 ppm. >3-10 ppm. >10-15 ppm. and >15 ppm-- based on the distribution of previous monitoring data. Age. body mass index (BMI). alcohol use. and tobacco use were in cluded in regression models as potential confounders. Num ber of cigarettes smoked per day was used as a continuous 0034S1 564 Gilliland and Mandel A B L E V III. Linear Multivariate Regression Moflei ot Factors Predicting the Hign Density Lipoprotein in Study ot Workers Exposed to PFOA Variable P SE(P) p value T A B L E X . Pearson Correlation Coefficients Between Total Serum Fluonne. Age. Body Mass index. Daily Alcohol Use. Daily Tooaccc Consumption, and Hepatic Parameters in Study ol Workers Exposed to PFOA intercept Total fluonne Alcohol1 Low (<1 oz/day) Nonresponstve (NR) Low X total fluonne11 NR X total fluonne' 65.00 -1.61 - 9 .9 2 - 6 .7 7 1.62 2.05 10.07 .77 3.51 5.73 .80 1.63 R* .17. Reference category is annkers who consumed 1-3 at elhanolfday. interaction terms between total fluonne and alcohol category. Adiusiid lor age. body mass index, smoking, and testosterone. .0001 .04 Total fluorine (ppm) Age (years) BMI (kg/m1) Alcohol (ot/day) Tobacco (cigs/day) .006 SGOT .24 SGPT .04 .21 GGT .01 .01 - .0 4 -.1 0 .01 12 .09 .20 p 02 .27 p .004 12 03 15 - 11 - 11 03 SGOT. serum glutamic oxaloacetic transaminase: SGPT. serum glutamic pymvic tran saminase: GGT. gamma glutamyl transterase: BMI. body mass index. TABLE IX . Serum Glutamic Oxaloacetic Transaminase. Serum Glutamic F*yruvic Transminase. and Gamma Glutamyl Transferase by Total Serum Fluorine in Study of Workers Exposed to PFOA TABLE XI. Serum Glutamic Oxaloacetic Transaminase by Body Mass Index. Age. Smoking, and Dnnking Status in Study ol Workers Exposed to PFOA Total fluorine N Mean SO Median Range Test* SGOT (lU/dl) SGOT (lU/dl) <1 ppm 1-3 >9-10 >10-15 >15-26 SGPT (lU/dl) <1 21-3 >3-10 >10-15 >15-26 GGT (lU/dl) <1 ppm 21-3 >3-10 >10-15 >15-26 Total 23 22.5 65 24.1 16 25.8 6 25.7 5 22.2 23 47.7 65 51.3 16 53.0 6 73.2 5 44.6 23 37.2 65 32.4 16 35.4 6 38.3 5 22.2 115 33.7 4.1 8.6 14.5 11.3 5.1 10.7 30.2 14.0 53.2 8.6 29.4 26.7 35.4 16.7 11.5 27.6 22 23 22.5 22.5 22 46 45 50.5 52.5 42 27 25 26 36.5 20 26 13-29 10-74 17-77 17-47 14-27 F 0.41 p * .80 30-69 4-263 29-40 38-177 34-54 F 1.19 p * .32 6-117 5-174 10-158 19-60 11-37 5-174 F . 0.39 p * .81 Anova. SGOT. serum glutamic oxaloacetic transaminase: SGPT, serum glutamic pynmc tran saminase: GGT. gamma glutamyl transferase. N (% ) Mean SO Median Range BMI <25 41 (35.7) 24 12.4 22 13-77 25-30 57 (49.6) 23 5.8 23 10-42 >30 17(14.8) 27 8.1 26 17-47 Age S30 >30-40 21 (18.3) 25 12.7 23 17-77 48 (41.7) 24 9.1 23 10-74 >40-50 27 (23.5) 22 5.4 23 13-40 >50-60 19 (16.5) 26 7.8 23 14-47 Alcohol <1 oz/d 87 (81.3) 26 13.5 22 16-77 1-3 oz/d 2 0(18 .7) 24 8.0 23 10-74 Missing 8 23 4.3 21 19-31 Tobacco Smoker 28 (24.8) 24 8.4 23 13-77 Nonsmoker 85 (75.2) 24 11.0 22 10-42 Missing 2 20 3.5 20 17-47 Total 115 Anova. SGOT. serum glutamic oxaloacetic transaminase: BMI. body mass index. Test* F . .92 p < .40 F . .78 p .51 F * .61 p .44 F . .02 p .89 variable if model fit was improved compared with the model using categorical variables. BMI was categorized into three categories. <25 kg/m: . 25-30 kg/rrr. and >30 F/m*. Alcohol use was divided into three categories: <1 k per day. 1-3 drinks per day. and no response to the uestionnaire item, and was entered into the models as a set of indicator variables. Significant nonlinear dose-response relationships were evaluated by comparing model fit using residual analysis and by comparing parameter estimates us ing indicator variables and continuous variables. Interac tions between total serum fluorine and the covariaies were evaluated based on biologic plausibility. Interaction terms "were included in the final model if the parameter estimate had a p value S0.05. The two nonrespondents to the smok- G034S2 PFOA, Cholesterol, Lipoproteins, and Hepatic enzymes 567 partteipan*s only. SCOT ;ind SCPT increased with incrcus- patotoxins. Because PFOA has a long biological half-life in im: PFOA. The hypothesis that PFOA may modulate (he huinans. is absorbed easily, and is hcpatotoxic in rodents. hepatic elTeels of obesity is consistent with these changes in PFOA production workers have been under medical surveil enzyme profile. This hypothesis has biologic plausibility lance for more than 20 years. No adverse clinical outcomes because obesity has been associated with elevation of tran related to PFOA exposure have been observed in these em saminases through fatty infiltration (Ludwig et al.. 1980: ployees. Hodgson et al.. 1989). PFOA may directly or indirectly In summary. PFOA was not associated with the marked potentate this effect in susceptible individuals. PFOA alters hepatic changes in humans that have been observed in ro hepatic lipid metabolism and may block the metabolism of dents. This finding is consistent with the results of a retro accumulated fatty acids, resulting in an exacerbation of the spective mortality study that found no increased mortality pathologic process (Haughom et al.. 1992). from liver disease (Gilliland and Mandel. 1993) and with PFOA may also modulate the effect of alcohol on he the results from an earlier morbidity study that found no patic metabolism. PFOA is associated with changes in the adverse hepatic effects (Ubel et al.. 1980). PFOA may mod effect of alcohol consumption on HDL levels, essentially ulate the effect of alcohol use and obesity on hepatic lipid blocking the rise in HDL associated with alcohol consump and xenobiotic metabolism. Continued epidemiologic sur tion. GOT was inversely associated with PFOA in drinkers. veillance is appropriate in workers exposed to PFOA. Perfluorooctanoic acid may decrease serum GGT by alter ing cell membrane permeability, by reducing the alcohol- ACKNOWLEDGMENTS mediated induction of GGT, or by changing alcohol oxida tion pathways and reducing the production of such toxic This study was supported in part by NIOSH grant intermediates as acetaldehyde (Bates. 1981: Schuckit and T150H07098-16 and the 3M Medical Department. Griffiths, 1982; Onego et al.. 1985: Schuckit and Irwin, 1988). These findings support the hypothesis that PFOA REFERENCES modulates the effects of endogenous and exogenous deter minants of hepatic metabolism. Abdellatif A. Preat V. Vamecq J. Nilsson R. Roberfroid M fl990). Per- Interpretation of these findings is limited by a number of factors. Only active workers in PFOA production were included in this study. It is unlikely that workers who had significant exposure during the previous 5 years would have been lost to this study because of transfer out of the PFOA production division. Transfer as a result of subclinica! changes in such biochemical parameters as SGOT is un likely. Because of the low turnover rate in plant employees (3% per year) and the inclusion of most current employees with appropriate job histories, selection bias is not a likely explanation for the findings in this study. Given the high participation (>80%). nonresponse bias is likely to be small. Information on smoking and alcohol consumption was col oxysome proliferation and modulation o f rat liver carcinogenesis by 2.3dichlorophenoxyacetic acid. 2.4.5 trichlororphenoxyacetic acid, perfluo rooctanoic acid and nafenopin. Carcinogenesis 11:1199-1902. Bates H (1981): GGTP and alcoholism: A sober look. Lab Management 19:1-3. Belisle J (1981): Organic fluoride in human serum: Natural versus indus trial sources. Science 212:1509-1310. Gilliland F. Mandel J (1993): Mortality among employees in a PFOA production plant. J Occup Med. 35:950-954. Griffith F. Long J (1980): Animal toxicity studies with ammonium perfluoroocunoate. Am Ind Hyg Assoc 41:576-583. Guy W (1979): ionic and organic fluorine in blood. In Johansen E Taves D. Olsen T (eds): Continuing Evaluation o f the Use of Fluoride. Boulder. CO: Westview Press. lected and used in the analyses: however, measurement er Haughom B. Spydevold O. (1992): The mechanism underlying the hyporor for these variables could allow residual confounding. lipemic effect of perfluorooctanoic acid (PFOA). perfluorooctane sul- Because smoking and alcohol consumption are not strong phonic acid (PFOSA) and clofibric acid. Biochim Biophys A cu 1128:65determinants for the endpoints in this study, the magnitude 72. of any residual confounding is likely to be small. The du ration of exposure may be an important determinant of PFOA level and effect: however, information on the dura tion of employment in exposed jobs was not available be cause plant records did not contain sufficient information to reconstruct exposures. Furthermore, the use of job history resulted in marked miscla.vsificution of exposure status, in dicating that the use of job duration would be of limited Hodgson M. Van Theil D. Lauschus K. Karpf M (1989): Liver injury tests in harzardous waste workers: The role o f obesity. J Occup Med 31:238242. Just W. Gorgas K. Hanl F. Heinemann P. Salazar M. Schimassek H (1989): Biochemical effects and zonal heterogeneity of peroxisome proliferation induced by perlluorocarhosylic acids in rat liver. Hepatology 9:570-581. Keller B. Vtarsman D. Popp J. Thurman R (1992): Several nongenotosic carcinogens uncouple mitochondrial oxidative phosphors latntn. Biochim Biophys A cu 111)2:237-244. value in determining duration of exposure. Kennedy G il9K5>: Dermal toxicity of ammonium pcrlluorinn'iuniuic. Many of the participants were employed in the produc Toxicol Appl Pharmacol 81:348-355. tion of compounds other than PFOA: however, none of ~ Kennedy G. Hall G. Briiiclli J Chen H <I9R6i: Inhalation insicnv of these processes involve substantial exposure to known he- ammonium pcrlluoruciunoalc. Final Chem Toxic 24:1325-1324 C034S3 568 Gilliland and Mandel usreoP Y avauable Ludwig J. V iggianoT. McGill U. On U tl)Kll): Nonalcoholic stcaiohcpic11in. Mayo Clin Pmc 55:43-* --d.*. Nilsson R. Bcijc B. Preat V. Erixon K. Kamel C i 1991): On the mechanism iif Ihc hcp41m.un.1nogcnK.1ly 1 peroxisome prolilcrators. Chem Biol InIcruel 711:2.15-30. Soiilcniux K. Andersson K. IV Picnv J il992l: Tile cllccts <11 (vrtluornocianoti acid on licpalic penixisiMiie proliferation and retail'd parameterslinw mi sex-related differences in inice. Rnwhciu Toxicol. 285 7 7 9 -7 S V Statistical Analysis System 11992). "SAS User s Guide. Version (1 " C.irs . NC: SAS Institute. Inc. Onego H. Blake J. IsrjcI Y (1985): Rclaiionship between gumma glutamyl irunspcptiiia.sc and mean urinary alcohol levels in alcoholics while drinking and in withdrawal. Alcohol Clin Exp Res 9:10-13. Pasioor T. Lee K. Perri M. Gillies P (1987): Biochemical and morpholog ical studies of ammonium pertluoroucianoate-induced hepatomegaly and peroxisome proliferation. Exp Mol Pathol 47:98-109. Takagi A. Sai K. Umemura T. Ha-cgawa R Kurukawa Y 1 19*111: Short term exposure In the peroxisome pniliferalors. pcrl)uisKx:ianoic acid and pcrfluomdccanoic acid, causes significant increases of 8-hydrnxydcoxyguanosine in liver DNA of rats. Cancer Leu 57:55-(it) Tavcs D (1971): Comparison of "organic" tluoride in human and nonhu man serums. 1 Dent Res 50:783. Perkins R (1992): Investigation of ammonium pcrfluorooctanate effect on hormone levels and peroxysome proliferation in the rat. Toxicologist 12: 38. Permadi H. Lundgren B. Andersson K. DePiemc. J. (1992): Effects of pcrfluoro fatly acids on xenobiotic-metaholizing enzymes, enzymes which detoxify reactive forms o f oxygen and lipid peroxidation in mouse liver. Biochem Pharmacol 44:1183-1191. Reddy i. AzamofT 0 . Hignite C (1980): Hypolipidemic hepatic peroxi some proliferators form a novel class of chcnical carcinogens. Nature 283:397-398. Schuckit M. Griffiths J (1982): Gamma glutamyltransferase values in non alcoholic drinking men. Am J Psychiatry 139:227-228. Schuckit M. Irwin M (1988): Diagnosis of alcoholism. Med Clin North Am 72:1133-1153. Tavcs D. Guy W. B a y W (1976): Organic fluorucartmns in human plasma Prevalence and characterization. In Filler R led): "Biochemistry Involvinc Carhon-Fluorinc Bonds" Washington DC: American Chemical Society, pp 117- 134. libel F. Sorenson S. Roach D (198(1): Health status of plum workers ex posed to fluorochemicals: A preliminary report. Am Ind Hyg Assoc 41: 584 -589. Van Rafelghem M. Mattie D. Bruner R. Andersen M (1987): Patholgoical and hepatic ultraslnictuial effects o f a single dose o f pcrtluoro-n-dccanoic acid in the rat. hamster, mouse and guinea pig. Fundam Appl Toxicol 9:522-54. Venkateswarlu P (1982): Sodium biphenyl method for determination of covalently bound fluorine in organic compounds and biological materials. Anal Chem 54:1132-1137. 0Q3484