Document LoXGaakN5k4gJJLddDBnLyzjQ

Il Taft/ Taft Stettinius & Hollister LLP 425 Walnut Street, Suite 1 8 0 0 /Cincinnati, OH 4 5 2 0 2 -3 9 5 7 /Tel: 51 3.38 1.28 38/Fax: 513.381.0205 / www.taftlaw.com Cincinnati / Cleveland / Columbus / Dayton / Indianapolis / Northern Kentucky / Phoenix / Beijing R obert A. Bilott 513.357.9638 bilott@tafUaw.com March 6, 2009 ff* o<$- TSCA Confidential Business Information Center (7407M) EPA East - Room 6428, Attn: Section 8(e) & FYI U.S. Environmental Protection Agency 1200 Pennsylvania Avenue, NW Washington, DC 20460-0001 OCQ --J --fa c VjO c CO CO Re: Submission To TSCA 8(e)/FYI Database Re: PFQA/PFOS To TSCA 8(e)/FYI Database: W e are hereby providing the following information for inclusion in the TSCA 8(e)/ FYI databases with respect to PFOA/PFOS: 1. Costa, G., et al., "Thirty Years of Medical Surveillance in Perfluorooctanoic Acid Production Workers," 51 J.O.E.M. 1-9 (March 2009); and 2. Joensen, U.N., et al., "Do Perfluoroalkyl Compounds Impair Human Semen Quality?," Environ. Health Persp. (doi: 10.1289/ehp.0800517) (online March 2, 2009). -Very truly yours, RAB:mdm Enclosure tobert A. Bilott 11361469.1 CONTAINS NO CBI f JOEM Volume 51, Number 3, March 2009 iI p.2 Thirty Years o f Medical Surveillance in Perfluooctanoic Acid Production Workers Giovanni Costa, MD Samantha Sartori, BS Dario Consonni, MD Objective: To report health outcomes of 30 years (1978--2007) of medical surveillance of workers engaged in a perfluooctanoic add (PFOA) production plant. Methods: Fifty-three males workers (20 to 63 years) were submitted every year to medical examination and blood chemical chemistry tests, and serum PFOA dosage. Results: In the latest survey PFOA serum levels rangedfrom 0.20 to 47.04 xg/mL in currently exposed workers, and from. 0.53 to 18.66 Jig/mL in thoseformerly exposed. No clinical evidence ofany specific trouble or disease has been recorded over the 30 years, and all the biochemical parameters, including liver', kidney and hormonal junc tions, turned out to be within the reference ranges, but a significant association of total cholesterol and uric acid with and PFOA serum level was evidenced. Conclusions: A probable interference of PFOA on interme diate metabolism deservesfurther investigations. (J O ccup Environ Med. 2009;51:000-000) From the Department of Occupational and Environmental Health (Mr Costa, Mr Sartori), University of Milano, Italy; Unit of Epidemiology (Mr Consonni). 1RCCS Maggiore Hospital. Mangiagalli and Regina Elena Foundation, Milano, Italy; and Institute of Medical Statistics (Mr Sartori) and Biometrics "G.A. Maccacaro," University of Milano, Italy. Address correspondence to: Prof. Giovanni Costa, Department of Occupational and Environmental Health, University of Milano. Clinica del Lavoro "L. Devoto," Via S. Bamaba 8, 20122 Milano: E-mail: giovanni.costa @unimi.it. Copyright 2009 by American College of Occupational and Environmental Medicine DOI: 10.1097/JOM.0b013e3181965d80 erfluorooctanoic acid (PFOA, C7F 15COOH, CAS No. 335-67-1) is a fully fluorinated carboxylic acid, widely used as emulsifier in flu oropolymer polymerization. It is solu ble in water, where it forms a mixture with its dissociated perfluorooctanoate anion (C7F I5COO- ); it is absorbed by oral, dermal and inhalation routes, and mainly bound to serum albumin. It is mostly found in liver, serum and kid ney, and is excreted, not metabolized, mainly in urines, with large differences in elimination speeds between species and sexes. It is a mild irritant for skin and mucoses, it is not a sensitizer and has neither teratogenic nor fetotoxic effects, but it delays sexual maturation in animals.1-5 It is not a genotoxic, but in two chronic studies in rats it has been reported to increase the incidence o f tumors (adenomas) o f liver, pancre atic acinar cells, and testicular Ley dig cells,6,7 like other peroxisome proliferators, as it acts mainly as a peroxisome proliferator-activated re ceptor alpha (PPARa) agonist.8'9 In recent years, many studies have been carried out with the aim o f better understanding the toxico kinetics and mode o f actions, al though with not clear and definite c o n c lu s io n s .10-17 Its half-life of elimination in se rum markedly differs between spe cies and sexes, being 1 to 9 days in rats, 18 days in mice, 20 to 40 days in dogs, 20 to 32 days in monkeys, and 3.8 years in humans.18-20 Hence, its biopersistence in the human blood, its detection in the general population21-24 and its presence in several biota and animals all over the II p.3 Medical Surveillance of Workers Exposed to PFOA Costa et al world25-27 are a concern for possible chronic effects.28-29 Epidemiological studies on large cohorts o f occupationally exposed workers, mainly of US plants (3M and DuPont) in the period 1947 2002, could not find any association between PFOA exposure and signif icant adverse health effects in terms o f morbidity and mortality.30-34 For mortality in particular, neither G illi land and Mandel31 nor Alexander32 and Leonard et al33 found excess o f death for cancers o f any type in quite large 3M and DuPont cohorts. Also the slight excess o f prostate cancer signaled by Gilliland and Mandel,31 and o f bladder cancer found in the first study by Alexander,32 were not confirmed by subsequent more spe cific surveys both in 3M34 and Du Pont33 workers. However, some contrasting and somewhat inconsistent findings, mainly related to lipid metabolism, have been reported as concerns per turbation o f biochemical variables. The first study, made by Ubel et al30 in 1980 in about 300 workers of 3M Cottage Grove plant, having total serum organic fluorine (90% PFOA) up to 71 p.g/mL, did not report any significant deviations from the nor mal values as concerns hematology, liver enzymes, total cholesterol, glu cose, and uric acid. In a subsequent study on 115 workers o f the same plant,35 after adjusting for age, smoking, alcohol intake and BMI in a multivariate regression model, a total serum flu orine level above 10 ppm was asso ciated with slight increases in hepatic enzymes through an interaction with obesity. A slight interference with HDL cholesterol was also observed, when in association with alcohol consumption. In other two cross-sectional stud ies on 191 3M production workers,36 serum PFOA was not significantly associated with any o f the 11 mea sured hormones, except for a 10% increase in mean estradiol level among em ployees having the highest levels o f serum PFOA (> 3 0 ppm), although this association was con founded by body mass index. Extending the survey in the same production plant to 265 workers for a longer period,37 a negative associa tion was recorded between serum PFOA and plasma Cholecystokinin (although inside the normal range), and no effect on hepatic enzymes, bilirubin, triclycerides, HDL, and LDL cholesterol. However, in a subsequent survey concerning 518 workers o f both sexes from two US and European plants,38 the same authors found a modest positive association between PFOA and cholesterol and triglycer ides in both cross sectional and lon gitudinal (over a 6 year time period) analysis, after adjusting for potential confounders, such as age, BMI, smoking, alcohol, and location. More recently, Olsen and Zobel39 have re-examined the same parame ters in 506 3M workers from three US and European 3M plants, by multiple and logistic regressions and analysis o f covariance: serum PFOA concentrations were negatively asso ciated with HDL and bilirubin, and positively with triglycerides, after adjusting for possible confounders. Very recently two reports, one lon gitudinal and another cross-sectional, on DuPont production workers have been published. In the retrospective study o f 454 production workers, with repeated measurements con cerning liver function and lipids over a period o f 25 years,40 a positive association between serum PFOA and total cholesterol and AST was found, whereas a negative associa tion with bilirubin, after adjustment for age, BMI, gender, and decade o f hire, but without control for lipid lowering medications. In the other cross-sectional study on 1025 active workers,41 including hematology, liver and renal function, lipid and purine metabolism, and hormones, a modest but statistically significant positive relationship be tween PFOA and total cholesterol, LDL, VDLD, GGT and uric acid was recorded, after excluding people tak ing lipid lowering medications. The present study is aimed at providing a further epidem iological contribution to the understanding o f possible effects on humans. The data refer to the workers of a chemical plant (Miteni, Trissino, It aly), where PFOA has been produced by electrochemical fluorination since 1968. Since 1978 all the workers engaged in PFOA production have been submitted every year to the medical surveillance program by an expert Occupational Health Physi cian, according to the Italian legisla tion concerning health and safety at work. Materials and Methods Subjects As a whole, 53 workers, all males, engaged in the PFOA production de partment have been examined every year from 1978 to 2007, for a total o f 919 persons-years. Subjects' age ranged from 20 to 63 years and length o f work exposure varied from 0.5 to 32.5 years in the period. At the latest PFOA biomonitoring, in 2007, 37 were active workers in the PFOA department, whereas 16 were no longer exposed being retired or transferred to other departments in the meantime. All the other male workers, 12 executive clerks and 95 blue collars from the other departments o f the company, who have never been ex posed to PFOA but submitted to the same periodical medical surveil lance, were considered as "controls." Table 1 summarizes the main characteristics o f the examined workers. Methods All the workers (exposed and not exposed) have been submitted every year to physical examination, includ ing recording o f blood pressure, height and weight, and to blood chemical chemistry tests, including hematology (Ht, Hb, WBC, RBC, PLTS), liver (albumin, a 1-a 2-3 -y 'I P-4 JOEM Volume 51, Number 3, March 2009 globulins, bilirubin, AST, ALT, ALP GGT) and renal (urea nitrogen, cre atinine) functions, glucose and lipid metabolism (total and HDL Choles terol, triglycerides), and uric acid. A ll the variables have been measured over the years by autoanalysers (SM A ) in the same laboratory o f the nearby regional public general hos pital, undergoing a periodical quality control program. In addition, in the workers o f the PFOA department, also Apo-A and Apo-B lipoproteins, protein C reac tive, immunoglobulins IgA-IgGIgM, Testosterone, Estradiol, FSH, TSH, FT3, FT4, and Prostatic Serum Antigen (PSA) were tested in blood in 2002 and 2006. The biological monitoring o f se rum PFOA started in April 2000 and was repeated in May 2001, Decem ber 2001, June 20Q2, September 2003, September 2004, May 2006, and May 2007. It included not only the workers o f the PFOA department currently exposed in those years, but also those formerly exposed who had been transferred to other departments or had retired in the meantime, as well as some non-exposed workers (clerks), taken as controls. PFOA serum levels were deter mined by HPLC-Electrospray-Tandem M ass Spectrometry42 in two internationally accredited laborato ries: the first one carried out the analysis in 2000, 2001, and 2002, the second one in the following years. A double blind comparison between the two laboratories carried out in 2003 gave a correlation index equal to 0.90. In 2000 and 2001, the method had an upper scale detection limit fixed at 45.5 pg/m L, thereafter no detection limit was present. Statistical Analysis A N O V A , l test and multiple linear regression models were used to eval uate the relationships between PFOA exposure and hematochemical ef fects while adjusting for relevant co variates. Natural log transformation o f the dependent variables was per formed where appropriate. General- TABLE 1 Demographic Characteristics of the Workers Examined PFOA Currently PFOA Previously W orkers Exposed Workers No Sex Age (mean and SD)* Years of work (mean and SD)* Years elapsed since PFOA exposure (mean and SD)* Person-years of observation BMI (mean and SD) Smoking (%) Alcohol consumption (%) 37 All males 41.6(8.31) 14.3(9.1) 487.2 26.2 (2.9) 38.2 73.5 16 All males 52.0 (8.7) 16.3 (9.9) 9.5 (8.3) 431.8 25.7 (3.1) 47.6 71.4 At the latest medical check. Control Group 107 All males 41.9(9.2) 14.0(9.3) 1499.3 25.1 (2.8) 36.5 62.6 ized Estimating Equations (GEE) models with an exchangeable corre lation matrix were fitted when ana lyzing repeated measurements and relationships between PFOA serum level and the different (continuous or dichotom ous) health outcom es, while accounting for within-subject correlations.43 In the regression models the follow ing covariates were included: age, years o f expo sure, year o f PFOA sampling, BMI, smoking, and alcohol consumption. Stata 1044 software was used to perform all the statistical analyses. TABLE 2 PFOA Serum Level (^g/mL), Recorded in the Year 2007, in Currently and Formerly Exposed Workers Currently Formerly Exposed Exposed No. Min 25th percentile Median Mean arithmetic Mean geometric 75th percentile Max Standard deviation 39 0 .2 0 2.25 5.71 12.93 4.02 23.55 47.04 14.43 11 0.53 2.61 4.43 6.81 3.76 9.24 18.66 6.06 Results From the clinical point of view, all the workers were in good health and no excess o f morbidity for any disease has been recorded over the years. In the latest survey, carried out in May 2007, the PFOA serum levels ranged from 0.20 to 47.04 pg/mL (ppm) in the currently exposed work ers, and from 0.53 to 18.66 pg/mL in the formerly exposed workers (Table 2): in the seven non-exposed subjects (clerks/staff) examined in 2006, the PFOA serum levels ranged from 0.05 to 0.181 pg/mL. The highest level recorded was 91.9 pg/mL in 2002. Table 3 shows the PFOA serum levels recorded over the years in the production workers. As above mentioned, in 2000 and 2001 the method had an upper scale detec tion limit fixed at 45.5 pg/mL; conse quently, in order to compare the data over the years the maximum level in 2000 and 2001 was set by default as the highest level recorded in 2002, when the method had no upper scale detection limit. A significant decrease o f both peak and mean levels was recorded after 2002, following a renovation o f the plant carried out in that year, including a partial automation o f the process, and the adoption o f more strict working procedures and use o f suitable protective devices. In the 27 currently production workers who had their serum PFOA measured in all the four last occa sions, median value decreased from 13.6 pg/mL in 2002 to 7.1 pg/mL in 2007 ( --47.8%), mean value de creased from 22.2 pg/mL to 14.0 pg/m L ( --37.0%), and peak level decreased from 86.3 pg/mL to 47.0 Pg/mL ( --45.5%), respectively. AH the biochemical parameters, turned out to be on average within 1I P-5 Medical Surveillance of Workers Exposed to PFOA Costa et al the laboratory reference ranges, while som e subjects had some pa rameters above the upper reference values, as shown in Table 4, which refers to the results o f the latest check carried out in 2007. As com pared with the control group some differences, both in terms o f mean TABLE 3 Serum PFOA Levels (p.g/mL) Recorded in Production Workers Over the Year 2000 2001 2002 2003 2004 2006 2007 No. Min 25th percentile Median Mean arithmetic Mean geometric 75th percentile Max Standard deviation 25 1.54 5.53 11.92 18.8 11.7 32.0 86.3* 2 0 .0 42 0.73 4.35 11.07 19.7 1 0 .2 19.72 91.9* 23.6 46 0.34 4.57 10.15 19.3 9.3 20.80 91.9 23.0 41 0.38 4.11 6.25 13.7 6.9 14.20 74.7 16.6 34 0.54 2.84 6.82 11.4 6.5 18.97 46.3 1 2 .0 49 0.54 2.36 5.27 1 0 .8 5.8 16.31 41.9 1 1 .8 50 0 .2 0 2.18 3.89 1 1 .6 5.4 18.66 47.0 13.3 'Upper scale detection limit at 45.5 gg/mL (discussed in text). values and number o f persons above the upper references limits, were noted for some parameters such as uric acid, cholesterol, triglycerides, and liver enzymes. A s the crude correlations between PFOA and the biochemical variables, recorded in the last 7 years, gave some indication on a possible asso ciation between PFOA serum levels and some lipids and uric acid (data not shown), three more in-depth analyses in this respect were per formed. The aim was clarifying this possible association by controlling the most important confounding fac tors, and after preliminary exclusion TABLE 4 Haematology: Results of the Last Check in the Currently Exposed Workers and Control Group Exposed Control Group Glucose (mg/dL) Urea nitrogen (mg/dL) Creatinine (mg/dL) Uric acid (mg/dL) Cholesterol total (mg/dL) Cholesterol HDL (mg/dL) Triglycerides (mg/dL) Bilirubin total (mg/dL) Bilirubin conjugated (mg/dL) AST (U/L) ALT (U/L) 7 GT (U/L) ALP (U/L) Proteins total (g/dL) Albumin (%) od globulins (%) a 2 globulins (%) globulins (%) 7 globulins (%) WBC (X109/L) RBC (X10,2/L) Hb (g/dL) Ht (%) Platelets (X109/L) Protein C reactive (mg/L) Apo A (g/L) Apo B (g/L) igG (g/L) IgA (g/L) IgM (g/L) PSA (ng/mL) Testosterone (ng/mL) Estradiol (pg/mL) TSH (UI/mL) FT3 (pg/mL) FT4 (ng/dL) q CO I d Reference Values Min-Max 70-110 7-22 0.80-1.30 3.5-7.2 191-240 >39 30-180 < 1 .0 <.30 <50 <50 8-61 40-129 6.4-8.2 52-67 3-7 5-9 5-12 1 0 -2 2 4.0-10.9 4.5-5.9 13.5-17.5 41-53 140-440 0.0-5.0 1.10-2.05 0.55-1.40 7.0-16.0 0.70-4.0 0.40-2.30 <4.0 10-50 0.27-4.20 2.0-4.4 0.80-1.70 Mean (SD) 92.7 (8.7) 15.2(3.1) 0.92(0.12) 6 . 2 (1 .1 ) 239.7(50.9) 55.9 (14.7) 178.9 (118.6) 0.58 (0.22) 0.15 (0.05) 32.4 (8 .8 ) 47.8 (22.3) 53.8 (46.7) 71.8(22.8) 7.71 (0.50) 61.6(3.0) 4.1 (0.9) 8.9 (1.1) 5.9 (0.9) 14.7 (2.6) 7.02(2.13) 5.06 (0.37) 15.6 (0.88)1 45.9 (2.2) 241.3(55.0) 2.59 (4.21) 1.43(0.24) 1.25(0.31) 10.6 (2.4) 3.1 (1.2) 1.2(0.76) 0.80 (0.94) 5.88 (0.57) 30.8(1.3) 1.99 (5.48) 3.62 (0.67) 1.31 (13.5) % Outside Reference Range 2 .6 0 0 17.9 35.9 0 7.7 0 2 .6 2 .6 17.9 28.2 0 1 0 0 1 2 .8 0 0 2 .6 0 2 .6 0 0 5.1 5.1 5.1 2 .6 1 2 .8 5.1 2 .6 5.1 2 .6 2 .6 2 .6 2 .6 Mean (SD) 89.3 (16.5) 15.2(3.0) 0.93(0.11) 5.7 (1.2) 213.5(39.7) 56.7 (12.3) 146.3(97.9) 0.67 (0.27) 0.18(0.11) 29.7 (10.7) 40.6(21.8) 44.5(41.4) 67.7 (19.6) 7.81 (0.56) 62.3 (3.9) 5.5 (9.4) 8 . 8 (1.4) 5.8 (0.5) 14.7(2.3) 6.96 (1.88) 5.11 (0.39) 15.7(0.86) 46.0 (2.4) 237.1 (54.4) % Outside Reference Range 6.5 0.9 0 4.7 26.2 6 .6 27.1 10.5 7.6 4.8 26.2 17.9 0 0 0 0.9 14 0 0 5.6 1 .8 0.9 0 0 T 'l P-6 JOEM Volume 51, Number 3, March 2009 o f people under treatment o f primary hyperlipidemias and with a history of chronic hepatitis. In the first analysis, 34 currently exposed workers were matched with 34 workers from the other depart ments with the same age ( l year) and job seniority ( l year) in the company, never exposed to PFOA, having the same working hours (shift work or daywork), and being resi dent in the same area with similar housing and living conditions. Table 5 shows that the two groups signifi cantly differ only for total choles terol and uric acid mean levels, being slightly higher in workers exposed to PFOA. In the second analysis, the same 34 currently exposed workers were compared with all the other 107 male workers o f the company. The multi ple regression analysis, adjusted by age, job seniority, body mass index, smoking and alcohol consumption, confirmed a significant effect of PFOA exposure as concerns an in crease in mean total cholesterol and uric acid (Table 6). The third analysis considered the 56 subjects (currently, formerly and never exposed) who had serum PFOA assessed concurrently with the biochemical parameters in the last 7 years. Multivariate GEE mod els (including age, job seniority, body mass index, alcohol consump tion, and years o f observation as potential confounders) show that to tal cholesterol and uric acid levels were weakly but significantly corre lated to serum PFOA (Table 7) as well as some liver enzymes and alpha2 globulins; however, total biliru bin appears to be inversely related to PFOA. As expected, also body mass in dex was found to be significantly correlated to total (Coef. = 0.028, P < 0.01) and HDL (Coef. = - .0 2 9 , P < 0.001) cholesterol, triglycerides (Coef. = 0.065, P < 0.001) and uric acid (Coef. = 0.022. P < 0.001), as well as with some liver enzymes (ALT, ALP), suggesting a possible interaction, like in previous surveys.35 TABLE 5 Comparison Between 34 Exposed and Non-Exposed Workers, Matched By Age, Work Seniority, Day/Shiftwork, and Living Conditions No. BMI Glucose (mg/dL) Urea nitrogen (mg/dL) Creatinine (mg/dL) Uric acid (mg/dL) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL) Total bilirubin (mg/dL) AST (U/L) ALT (U/L) GGT (U/L) ALP (U/L) Total proteins (g/dL) Albumins (%) ct1 globulins (%) a2 globulins (%) globulins (%) 7 globulins (%) WBC (X109/L) RBC (X1012/L) Haemoglobin (g/dL) Haematocrit (%) Platelets (x 109/L) Non-Exposed 34 25.94 89.41 14.82 0.93 5.73 206.4 56.68 150.03 0.58 31.62 42.74 48.52 69.41 7.72 62.87 3.82 8.76 5.68 14.44 6.64 5.08 15.69 45.86 233.03 Exposed 34 26.16 93.56 19.47 0.94 6.29 237.0 57.82 155.35 0.55 31.44 35.29 43.88 67.65 7.67 61.51 3.94 8.90 5.76 14.47 7.53 5.00 15.37 45.36 238.97 t Test 0.34 1.30 1.14 0.19 2 .1 0 3.06 0 .2 1 0.77 0.18 0.41 1.35 0.05 0.24 0 .2 2 0.99 0.59 0 .2 0 0.49 0 .1 0 1 .8 6 0.60 0.40 0.74 0.84 P NS NS NS NS 0.039 0.003 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS TABLE 6 Multiple Regression Analysis Comparing the 34 Exposed Workers (E) With All the Other 107 Workers (ALL) of the Plant _________ Parameter___________ Coef. (E vs ALL) 95% C.l. P Glucose (mg/dL) Urea nitrogen (mg/dL) Creatinine (mg/dL) Uric acid (mg/dL) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL) Total bilirubin (mg/dL) AST (U/L) ALT (U/L) GGT (U/L) ALP (U/L) Total proteins (g/dL) Albumin (%) a 1 globulins (%) a2 globulins (%) P globulins (%) 7 globulins (%) WBC (x109/L) RBC (X1012/L) Haemoglobin (g/dL) Haematocrit (%) Platelets (X109/L) 3.53 3.92 0.004 0.50 21.7 2.42 -.15 -0.09 1.35 -5.18 0.32 -0.78 - 0 .2 0 -0.73 -1.82 0.27 -0.003 -0.53 0.58 -0.08 0.27 -0.51 1.31 -2.21/9.28 -0.95/8.79 -0.04/0.05 0.06/0.94 6.83/36.6 -2.30/7.13 -34.6/34.3 -0.19/0.01 -2.72/5.41 -13.7/3.32 -17.5/18.1 -8.51/6.95 -0.57/0.17 -3.44/1.97 -8.18/4.54 -0.75/1.28 -0.37/0.36 -2.29/1.24 -0.19/1.35 -0.23/0.07 -0.60/0.07 -1.42/0.40 -18.8/21.4 NS NS NS 0.027 0.005 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS !I p.7 6 Medical Surveillance of Workers Exposed to PFOA Costa et al TABLE 7 Results of the Multivariate Analysis (GEE Model) in the 56 Subjects (Currently, Formerly, and Never Exposed) Who Had Serum PFOA Assessed Concurrently With the Biochemical Parameters in the Last Six Yr (Controlled for Age, Job Seniority, Body Mass Index, Alcohol Consumption, and Yr of Observation) Coef. Pfoa 95% C.l. P Glucose Urea nitrogen Creatinine Uric acid Total cholesterol HDL cholesterol Triglycerides Total bilirubin Conj. bilirubin AST ALT GGT ALP Apo A Apo B IgG IgM Estradiol Prot C reactive PSA Testosterone Total proteins Albumins a 1 globulins a 2 globulins fi globulins 7 globulins WBC RBC Haemoglobin Haematocrit Platelets 0.005 0.017 - 0 .0 1 1 0.026 0.028 -0.018 0.055 -0.080 -0.034 0.038 0.116 0.177 0.057 - 0 .0 0 1 0.023 -0.017 0.048 - 0 .0 1 2 - 0 .0 2 0 -0.032 -0.005 - 0 .0 0 1 -0.009 0.026 0.026 0 .0 1 1 0.013 0.029 - 0 .0 0 2 0.008 0 .1 2 2 1.987 - 0 .0 1 1 / 0 . 0 2 0 -0.015/0.049 -0.025/0.004 0.001/0.053 0.002/0.055 -0.047/0.012 -0.036/0.147 -0.137/-0.024 -0.099/0.031 -0.003/0.080 0.054/0.177 0.076/0.278 0.007/0.107 -0.038/0.035 -0.041/0.087 -0 .1 1 5 /--0.080 -0.093/0.190 -0.080/0.057 -0.268/0.228 -0.286/0.222 - 0 .0 2 1 / 0 .0 1 1 -0.010/0.008 -0.017/0.001 -0.001/0.053 0.007/0.045 -0.008/0.030 -0.005/0.031 -0.011/0.071 -0.009/0.006 -0.113/0.129 -0.132/0.376 -7.36/11.34 NS NS NS <0.05 <0.05 NS NS < 0 .0 1 NS NS < 0 .0 1 < 0 .0 1 <0.05 NS NS NS NS NS NS NS NS NS NS NS < 0 .0 1 NS NS NS NS NS NS NS 60 years after exposure Fig. 1. PFOA serum level (pg/mL) in the 16 formerly exposed workers, who retired or were transferred to other departments. Discussion In our cohort neither clinical evi dence o f specific disturbances nor health disorders have been recorded over 30 years of medical observa tions o f workers exposed to PFOA, having serum levels ranging from 0.20 to 91.9 pg/mL. A significant decrease in PFOA blood levels (--37% in mean level and --45.5% in peak level) was re corded in the 4 last years after plant renovation and improvement o f working conditions. However, in evaluating this trend it is necessary to take into account the long biolog ical half-life o f the substance, so that the present blood levels largely re flect the exposure conditions of the previous years. In fact, contrary to the rather rapid rates o f elimination reported in lab oratory animals, PFOA appears to be slow ly eliminated in humans.3-30 Very recently Olsen et al20 have determined the serum elimination half-life o f PFOA in 26 retired fluorochemicals production workers over a 5 year span; the arithmetic mean half-life was 3.8 years (95% Cl = 3.1 to 4.4) and the geometric mean half-life o f 3.5 years (95% Cl = 3.0 to 4.1), with a range from 1.5 to 9.1 year. A similar trend can be found also in our data (Fig. 1). In the 16 for merly exposed workers, who retired or were transferred to other depart ments, the decrement o f PFOA blood concentration over the last 7 years appears to be more or less steeper depending on both extent, and dura tion o f the past exposure, and years elapsed since leaving the job. A l though the data are not complete and homogeneous, in terms o f PFOA level, number and years o f measure ments, time elapsed since exposure, the observed trend in the decrement suggests an estimated half-life o f 5.1 (sd 1.7) years, varying from 2.6 to 9.7 years (geometric mean 4.8). It is noteworthy that even after 20 years from exposure some people still have 1 to 3 jxg/mL of PFOA in their blood. i'l p.8 JOEM Volume 51, Number 3, March 2009 7 As regards biological effects, no significant perturbation o f liver and renal functions, immunology, and hormonal secretion (including estra diol, testosterone, thyroid, and pros tate hormones) were recorded in our workers, except for a significant in terference with lipids (cholesterol) and purines (uric acid) metabolism, beside the influence of BMI. A s mentioned in Introduction, some previous cross sectional studies in 3M workers did not find any consistent association between se rum PFOA and liver, renal and lipid functions, and hormonal secretion. As concerns liver function and lipids metabolism in particular, Gilli land and Mandel35 did not record any significant difference between ex posed and non-exposed workers in any biological parameter o f liver function, and no significant correla tion between total serum fluorine (above 10 |xg/mL) and cholesterol, LDL or HDL, after adjusting for age, BM I, alcohol, and smoking. The same negative results were obtained in three successive cross-sectional investigations in workers with serum PFOA up to 114.1 p.g/niL.37 More recently, in two other 3M groups, including both European and US occupationally exposed workers, Olsen et al38 found no significant effect on liver function but a posi tive, although modest, association between serum PFOA (range: 0.01 to 12.7 |xg/mL) and cholesterol as well as triglycerides, both in a cross sectional and in a 6-year longitudinal analysis. M oreover, in a further study in workers with serum PFOA ranging from 0.007 to 92.03 p.g/mL, Olsen and Zobel39 recorded a weakly negative association with HDL and total bilirubin, and a positive one with triglycerides, after excluding people taking cholesterol lowering medications; however, these findings were explained by demographic dif ferences across the three locations. However, Sakr et al in a cross sectional study on DuPont work ers,41 having serum PFOA from 0.005 to 9.55 pg/mL, reported a modest, but significant positive cor relation between serum PFOA and total cholesterol, LDL, and VLDL, whereas in another longitudinal study40 they found a positive associ ation o f serum PFOA (range: 0 to 22.66 p.g/mL) with total cholesterol and AST, and a negative association with total bilirubin, after adjusting for potential confounders (age, BMI, gender, and decade o f hiring). However, it is worth mentioning that serum cholesterol appears to move in the opposite direction than observed in experimental studies in rodents, in which PFOA has a signif icant hypolipidemic effect in serum, being a strong PPARa agonist.45 No study reported any hypolipidemic ef fect in exposed workers, whereas three studies including the present one recorded a significant increase in cholesterol levels, besides other non consistent increases o f triglycerides and liver enzymes. A lso in cynomolgus monkeys, after 6 months o f oral dosing o f PFOA from 3 to 30/20 mg/ kg/die (with serum PFOA ranging from 20 to 467 |xg/mL) no decrease of cholesterol, but a significant increase o f triglycerides was observed.18 This might be due to different mechanisms of action and/or receptor agonisms, ie, PPARa/y and/or CAR29 in rodents rather than in primates, which deserves further investigation. For example it is well known that there are different expressions and properties o f PPAR isoforms (a, (3, and *y) in animals and humans9,13-15,46-48 as well as there might be different bindings with serum proteins (ie, albumin in animal, CETP) and in renal readsoiption (Oatpl and Oat3),49,50 or interference with hepatic metabolism o f fatty acids, ie, via CYP450. As to the latter, it is worth men tioning that three studies (Olsen and Zabel,39 Sakr et al,41 and the present one), also recorded a significant neg ative association between serum PFOA and total bilirubin, that was not documented in rodents and pri mates. It is worth noting in this regard that perfluorolauric acid proved to cause a marked bilirubin destruction in vitro when added to a pure bacterial enzyme (CYP102 or BM 3) specialized in fatty acid me tabolism. By analogy, it is possible that a similar effect might take place in humans on interaction o f PFOA with a fatty acid metabolizing CYP e n z y m e .51 Moreover, we recorded a positive significant association between se rum PFOA and uric acid, that was also reported by Sakr et al in DuPont workers.41 This may be suggestive o f some interference in liver purines metabolism, which deserves specific investigations. In our study, we tried to control as many confounding variables as pos sible, in particular age, BMI, smok ing, alcohol consumption, and shift work. The first four factors were taken into consideration in the GEE analysis, whereas the possible inter ference o f shift and night work on metabolic parameters, as evidenced in literature,52,53 was controlled by comparing exposed workers and controls matched for (shift) work seniority (Table 5). So, beside some inconsistent find ings among the studies, it is probable that PFOA exerts some interference on intermediate metabolism also in humans. Owing to the small number o f exposed subjects it is not possible to draw any firm conclusion, though the long time span o f the careful medical surveillance may allow to exclude a severe impact o f PFOA un human health. However, the possi ble interference on the intermediate metabolism should not be underes timated and deserves further inves tigations on mechanisms for action, and as a possible risk factor for metabolic disorders and/or CVD diseases. References 1. Kennedy GL, Bmenhoff JL Jr, Olsen GW, et al. The toxicology of perfluorooctanoate. Crit Rev Toxicol. 2004;34: 351-384. 2. Lau C, Anitole K. Hodes C, Lai D, Pfahles-Hutchens A, Seed J. Perfluoroalkyl acids: a review of monitoring and !I P-9 8 Medical Surveillance of Workers Exposed to PFOA Costa et al toxicological findings. Toxicol Sci. 2007; 99:366-394. 3. Butenhoff JL, Kennedy GL, Frame SR, O'Connor JC, York RG. The reproductive toxicology of ammonium perfluorooctanoate (APFO) in the rat. Toxicol ogy. 2004;196:95-116. 4. Lau C, Thibodeaux JR, Hanson RG, et al. Effects of perfluorooctanoic acid expo sure during pregnancy in the mouse. Toxicol Sci. 2006;90:510-518. 5. Abbott BD, Wolf CJ, Schmid JE, et al. Perfluorooctanoic acid (PFOA)-induced developmental toxicity in the mouse is dependent on expression of peroxisome proliferator activated receptor-alpha (PPARa). Toxicol Sci. 2007;98:571-581. 6. Sibinski LJ, Allen JL, Erickson EE. Two Year Oral (diet) Toxicity/Carcinogenicity Study o f Fluorochemical FC-143 in Rats. St. Paul, MN: Riker Laboratories, Inc., Expt. No. 0281CR0012. U.S. Environ mental Protection Agency Administrative Record 226-0437-0440-, 1987. 7. Cook JC, Murray SM, Frame SR, et al. Induction of Leydig cell adenomas by ammonium perfluorooctanate: a possible endocrine related mechanism. Toxicol Appl Pharmacol. 1992;192:209-217. 8. Bicgel LB, Hurtt ME, Frame SR, O'Connor JC, Cook JC. Mechanisms of extrahepatic tumor induction by peroxi some proliferators in male CD rats. Toxi col Sci. 2001;60:44-55. 9. Maloney EK, Waxman DJ. Transactiva tion of PPARa and PPAR-y by structur ally diverse environmental chemicals. Toxicol Appl Pharmacol. 1999:161:209 218. 10. Butenhoff JL, Gaylor DW, Moore JA, et al. Characterization of risk for general population exposure to pcrfluorooctanoate. Regul Toxicol Pharmacol. 2004;39: 363-380. 11. Andersen ME, Clewell HJ, Tan YM. Butenhoff JL, Olsen GW. Pharmacokinetic modeling of saturable, renal resorption of perfluoroalkylacids in monkeys--probing the determinants of long plasma half lives. Toxicology. 2006;227:156-164. 12. Klaunig JE, Babich MA, Baetcke KP, et al. PPARalpha agonist-induced ro dent tumors: modes of action and hu man relevance. Crit Rev Toxicol. 2003; 33:655-780. 13. Palmer CN, Hsu M-H. Griffin KJ, Raucy JL. Johnson EF. Peroxisome proliferator activated receptor-a expression in human liver. Mol Pharmacol. 1998.53:14-22. 14. Cheung C, Akiyama TE, Ward JM, et al. Diminished hepatocellular proliferation in mice humanized for the nuclear recep tor peroxisome proliferator-activated re ceptor alpha. Cancer Res. 2004;64: 3849 -3854. 15. Morimura K, Cheung C, Ward JM, Redd JK, Gonzalez FJ. Differentia] susceptibil ity of mice humanized for peroxisome proliferator-activated receptor alpha to Wy-14,643-induced liver tumorigenesis. Carcinogenesis. 2006;27:1074 -1080. 16. Rosen MB, Lee JS, Ren H, et al. Toxicogenomic dissection of the perfluo rooctanoic acid transcript profile in mouse liver: evidence for the involve ment of nuclear receptors PPARalpha and CAR. Toxicol Sci. 2008;103:46-56. 17. Kudo N, Kawashima Y. Toxicity and toxicokinetics of perfluorooctanoic acid in humans and animals. J Toxicol Sci. 2003;28:49-57. 18. Butenhoff J, Costa G, Elcombe C, et a). Toxicity of ammonium perfluorooctanoate in male cynomolgus monkeys after oral dosing for 6 months. Toxicol Sci. 2002;69:244-257. 19. Harada K, Inoue K, Morikawa A, Yoshinaga T, Saito N, Koizumiet A. Renal clearance of perfluorooctane sulfonate and perfluorooctanoate in humans and their species-specific excretion. Environ Res. 2005;99:253-261. 20. Olsen GW, Burris JM, Ehresman DJ, el al. Half-life of serum elimination of perfluorooctanesulfonate, perfluorohexanesulfonate, and perfluorooctanoate in retired fluorochemical production work ers. Environ Health Perspect. 2007;! 15: 1298-1305. 21. Calafat AM, Kuklenyik Z, Reidy JA, Caudill SP, Tully JS, Needham LL. Se rum concentrations of 11 polyfluoroalkyl compounds in the U.S. population: data from the National Health and Nutrition Examination Survey (NHANES) 1999 2000. Environ Sci Technol. 2007;41: 2237-2242. 22. Emmett EA, Shofer FS, Zhang H, Free man D, Desai C, Shaw LM. Community exposure to perfluorooctanoate: relation ships between serum concentrations and exposure sources. J Occup Environ Med. 2006;48:759-770. 23. Kannan K, Corsolini S, Falandysz J, et al. Perfluorooctanesulfonate and related fluorochemicals in human blood from several countries. Environ Sci Technol. 2004;38:4489-4495. 24. Olsen GW, Huang H-Y, Helzlsouer. KJ, Hansen KJ, Butenhoff JL. Mandel JH. Historical comparison of perfluo rooctanesulfonate, perfluorooctanoate, and other fluorochemicals in human blood. Environ Health Perspect. 2005: 113:539-545. 25. Giesy JP, Kannan K. Global distribution of perfluorooctane sulfonate and related perfluorinated compounds in wildlife. Environ Sci Technol. 2001 ;35:1339-- 1342. 26. Nakata H, Kannan K, Nasu T, Cho H-S, Sinclair E, Takemura A. Perfluorinated contaminants in sediment and aquatic organisms collected from shallow water and tidal flat areas of the Ariake Sea, Japan: environmental fate of perfluo rooctane sulfonate in aquatic ecosystems. Environ Sci Technol. 2006;40:49164921. 27. Tao L, Kannan K, Kajiwara N, el al. Perfluorooctanesulfonate and related fluorochemicals in albatrosses, elephant seals, penguins, and polar skuas from the Southern Ocean. Environ Sci Technol. 2006;40:7642-7648. 28. US Environmental Protection Agency. Perfluorooctanoic Acid. 2008. Available at: http://www.epa.gov/oppt/pfoa. Ac cessed October 3, 2008. 29. UK Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment. Second Draft Working Pa per on the Tolerable Daily Intake fo r Perfluorooctanoic Acid. 2006. Available at: http://cot.gov.uk/pdfs/tox-200618.pdf. Accessed October 3, 2008. 30. Ubel FA, Sorenson SD, Roach DE. Health status of plant workers exposed to fluorochemicals-- a preliminary report. Am Ind Hyg Assoc J. 1980;41:584-589. 31. Gilliland FD, Mandel JS. Mortality among employees of a perfluorooctanoic acid production plant. J Occup Med. 1993;35:950-954. 32. Alexander BH. Mortality Study o f Work ers Employed at the 3M Cottage Grove Facility. St. Paul, MN: 3M Company; W ashington DC: U.S. EPA Public Docket AR-226-1030a018; 2001. 33. Leonard RC, Kreckmann KH, Sakr CJ, Symons JM. Retrospective cohort mor tality study of workers in a polymer production plant including a reference' population of regional workers. Ann Epi demiol. 2008;18:15-22. 34. Alexander BH, Olsen GW. Bladder can cer in perfluorooctanesulfonyl fluoride manufacturing workers. Ann Epidemiol. 2007;17:471-478. 35. Gilliland FD, Mandel JS. Serum perfluo rooctanoic acid and hepatic enzymes, li poproteins and cholesterol: a study of occupationally exposed men. Am J Ind Med. 1996,29:560-568. 36. Olsen GW, Gilliland FD, Burlew MM, Burris JM, Mandel JS, Mandel JHL. An epidemiologic investigation of reproduc tive hormones in men with occupational exposure to pefluorooctanoic acid. J Oc cup Environ Med. 1998;40:614 - 621. 37. Olsen GW, Burris JM, Burlew MM, JOEM Volume 51, Number 3, March 2009 9 Mande] JH. Plasma cholecystokinin and hepatic enzymes, cholesterol and lipopro teins in ammonium perfluorooctansulfonate production workers. Drug Chem Toxicol. 2000;23:603-620. 38. Olsen GW, Burris JM, Burlew MM, Mandel JH. Epidemiologic assessment of workers serum perfluorooctansulfonate (PFOS) and pefluorooctanoate (PFOA) concentrations and medical surveillance examinations. J Occup Environ Med. 2003;45:260-270. 39. Olsen GW, Zobel LR. Assessment of lipid, hepatic, and thyroid parameters with serum perfluooctanoate (PFOA) concentrations in fluochemical produc tion workers. Ini Arch Occup Environ Health. 2007;81:231-246. 40. Sakr CJ, Leonard RC, Kreckmann KH, Slade MD, Cullen MR. Longitudinal study of serum lipids and liver enzymes in worker with occupational exposure to ammonium perfluooctanoate. J Occup Environ Med. 2007;49:872-879. 41. Sakr CJ, Kreckmann KH, Green JW, Gilles PJ, Reynolds JL, Leonard RC. Cross-sectional study of lipids and liver enzymes related to a serum biomarker of exposure (ammonium perfluooctanoate or APFO) as part of general health survey in a cohort of occupationally exposed workers. J Occup Environ Med. 2007;49: 1086-1096. 42. Hansen KJ, Clemen LA, Ellefson ME. Johnson JHO. Compound-specific, quanti tative characterization of organic fluorochemicals in biological matrices. Environ Sei Technol. 2001;35:766-770. 43. Twisk JWR. Applied Longitudinal Data Analysis fo r Epidemiology. Cambridge: Cambridge University Press; 2003. 44. StataCorp. Stata Statistical Software: Re lease 10. College Station, TX: StataCorp LP; 2007. 45. Xie Y, Yang Q, Nelson BD, DePierre JW. The relationship between liver per oxisome proliferation and adipose tissue atrophy induced by peroxisome proliferator exposure and withdrawal in mice. Biochem Pharmacol. 2003;66:749-756. 46. Michalik L, Desvergne B, Dreyer C, DeGavillet M, Laurini RN, Wahl W. PPAR expression and function during vertebrate development. Int J Dev Biol. 2002;46:105-114. 47. Rosen ED, Spiegelman BM. PPARy: a nuclear regulator of metabolism, differ entiation, and growth. J Biol Chem. 2001; 276:37731-37734. 48. Toth B, Hornung D, Scholz C, Djalali S, Friese K, Jeschke U. Peroxisome proliferatoractivated receptors: new players in the field of reproduction. Am J Reprod Immunol. 2007;58:289-310. 49. Katakura M, Kudo N, Tsuda T, Hibino Y, Mitsumoto A, Kawashima Y. Rat organic anion transporter 3 and organic anion transporting polypeptide 1 mediate perfluorooctanoic acid transport. J Health Sci. 2007;53:77-83. 50. Kudo N, Katakura M, Sato Y, Ka washima Y. Sex hormone-regulated renal transport of perfluorooctanoic acid. Chem Biol Interact. 2002;139:301-316. 51. Pons N, Pipino S, De Matteis F. Interaction of polyhalogenated compounds of appro priate configuration with mammalian or bacterial CYP enzymes. Increased bilirubin and uropofobylinogen oxidation. Biochem Pharmacol. 2003;66:405-414. 52. Karlsson B, Knutsson A, Lindahl B, Alfredsson L. Metabolic disturbances in male workers with rotating threeshift work. Results of the WOLF study. hit Arch Occup Environ Health. 2003; 76:424-430. 53. Biggi N, Consonni D, Galluzzo V, Sogliani M, Costa G. Metabolic syndrome in permanent nightworkers. Chronobiol lnt. 2008;25:1-12. ehponline.org ENVIRONMENTAL HEALTH PERSPECTIVES Do Perfluoroalkyl Compounds Impair Human Semen Quality? Ulla Nordstrom Joensen, Rossana Bossi, Henrik Letters, Allan Astrup Jensen, Niels E. Skakkebk, and Niels Jorgensen doi: 10.1289/ehp.0800517 (available at http://dx.doi.org/) Online 2 March 2009 NIEHS ^ N a t i No antioanall Institute of t r Environmental Health Sciences N a tion al Institutes o f Health U .S. D epartm ent o f Health and Hum an Services Page 1 of.2Q Do Perfluoroalkyl Compounds Impair Human Semen Quality? ii p. 12 Ulla Nordstrom Joensen1, Rossana Bossi2, Henrik Leffers1,Alian Astrup Jensen3, Niels E. Skakkebaek1, Niels Jprgensen1. 1University Department of Growth and Reproduction, Rigshospitalet, Copenhagen, Denmark 2National Environmental Research Institute, University of Aarhus, Roskilde, Denmark 3FORCE Technology, Brpndby, Denmark Address of institution where work was done: University Department of Growth and Reproduction Rigshospitalet, section 5064 Blegdamsvej 9 2100 Copenhagen Denmark Corresponding author: Ulla Nordstrom Joensen, MD University Department of Growth and Reproduction, Rigshospitalet Blegdamsvej 9 2100 Copenhagen Denmark Telephone: +45 3545 5064 Fax: +45 3545 6054 E-Mail: ulla.nordstroem.ioensen@rh.regionh.dk 1 Acknowledgements/grant support: The authors greatly appreciate the support from the European Union (contract no. QLK4CT- 2002-00603), the Danish Agency for Science, Technology and Innovation (grant nos. 9700833 and 271070678), the Danish Ministry of Health (ISMF) (Grant no. 7-302-02-9/3), the Danish Environmental Protection Agency, and the University of Copenhagen. The authors declare they have no competing financial or non-financial interests. Running title: PFAAs and Human Semen Quality. EH P article descriptor: Reproductive Health Key words: Endocrine disruptors, male reproductive health, perfluoroalkyl compounds, PFAA, PFC, semen quality, sperm morphology, testosterone. Abbreviations: BMI: body mass index Cl: confidence interval ESI: electrospray ionization FAI: free androgen index FSH: follicle stimulating hormone HPLC: high performance liquid chromatography LC-MS-MS: liquid chromatography-tandem mass spectrometry LH: luteinizing hormone LOD: limit of detection PFAA: perfluoroalkyl acids PFC: polyfluorinated compounds PFDA: perfluorodecanoic acid PFDoA perfluorododecanoic acid PFHpA: perfluoroheptanoic acid PFHxS: perfluorohexane sulfonic acid PFNA: perfluorononanoic acid PFOA: perfluorooctanoic acid PFOS: perfluorooctane sulfonic acid PFOSA: perfluorooctane sulfonamide PFTrA: perfluorotridecanoic acid PFUnA: perfluoroundecanoic acid SHBG: sex hormone binding globulin TDS: testicular dysgenesis syndrome 2 Page 3 of. 26. Outline of section headers Abstract Introduction Materials and methods Study population PFAA analysis Reproductive hormone analysis Semen analysis Statistical analysis Results Levels ofPFAAs in serum PFAAs and semen variables PFAAs and reproductive hormones Smoking and BMI Discussion and conclusions References Tables Figure legends Figures 1 p. 14 3 ' Abstract p. 15 Page 4 of 26 B ackground: Perfluoroalkyl acids (PFAAs) are found globally in wildlife and humans, and are suspected to act as endocrine disruptors. There are no reports of PFAA levels in adult men from Denmark, and no reports of a possible association between semen quality and PFAA exposure. Objectives: To investigate possible associations between PFAAs and testicular function. The hypothesis was that higher PFAA levels would be associated with lower semen quality and lower testosterone levels. Methods: We included 105 Danish men (median age 19 years) from the general population, and analyzed serum samples for levels of 10 different PFAAs and reproductive hormones, and assessed semen quality. Results: Considerable levels of PFOS, PFOA and PFHxS were found in all young men (median 24.5, 4.9 and 6.6 ng/mL, respectively). Men with high combined levels of PFOS and PFOA had a median of 6.2 million normal spermatozoa in their ejaculate in contrast 15.5 million among men with low PFOS-PFOA (p=0.030). In addition we found non-significant trends with regard to lower sperm concentration, lower total sperm counts and altered pituitary- gonadal hormones among men with high PFOS-PFOA levels. Conclusion: High PFAA levels were associated with fewer normal sperms. Thus, high levels of PFAAs may contribute to the otherwise unexplained low semen quality often seen in young men. However, our findings need to be corroborated in larger studies. 4 Page 5 of 26 1I p. 16 Introduction Perfluoroalkyl acids (PFAAs) are degradation products of many man-made polyfluorinated compounds (PFCs) used in consumer and industrial products, for example for impregnation of carpets, textiles and paper (Jensen et al. 2008; Jensen and Leffers 2008; Kissa 2001). Studies of environment, wildlife and humans suggest widespread presence and exposure, as well as persistence in the environment and bioaccumulation (Giesy and Kannan 2001; Kannan et al. 2004). For perfluorooctanoic acid (PFOA), perfluorooctane sulfonic acid (PFOS) and perfluorohexane sulfonic acid (PFHxS), three of the most abundant PFAAs, half-lives for humans have been estimated as 3.8, 5.4 and 8.5 years, respectively (Olsen et al. 2007). Some studies suggest that men may have higher serum concentrations of PFAAs than women, and younger individuals may have higher levels than older (Calafat et al. 2006). Thus, young men may have particularly high levels of exposure and may therefore be a group at risk for potential adverse effects of PFAAs. PFAAs can cross the placental barrier and therefore have the potential to affect the fetus. In humans, levels of PFOS and PFOA in umbilical cord blood have been inversely related to birth weight (Apelberg et al. 2007). In addition, PFOS, PFOA and PFHxS have been detected in human seminal plasma samples (Guruge et al. 2005). However, data on effects in humans are sparse, and most come from studies of occupationally exposed individuals. These studies have not given conclusive evidence of adverse effects. A recent study, however, measured PFAA levels in early pregnancy found that higher levels of PFOS and PFOA was associated with significantly longer waiting time to pregnancy (Fei et al. 2009). Animal studies provide some evidence for adverse reproductive effects on animals exposed as adults or in utero. Exposure of adult male rats to PFOA reduced their testosterone levels and increased their estradiol levels, which may partly explain earlier 5 p. 17 Page 6 of 26 findings of induction of Leydig cell hyperplasia and/or adenomas in the testes of exposed animals (Biegel et al. 1995; Cook et al. 1992). Our objective was to investigate the associations between PFOS, PFOA, PFHxS and other PFAAs and testicular function. Our primary hypothesis was that high concentrations of PFAAs would be associated with low testosterone levels and secondly that high PFAA levels are negatively associated to semen quality variables. Materials and methods Study population. Since 1996, semen quality of young men in Denmark has been surveyed in a cross-sectional study (Andersen et al. 2000; Jprgensen et al. 2002). All young Danish men must report for military draft, and annually new cohorts of approximately 300 men from the Copenhagen area in Denmark have been included. They each provided one semen sample and had a venous blood sample drawn. Of the 546 men examined in 2003, we selected 105 for the investigation of associations between PFAAs and testicular function. The 105 men included the 53 men (group 1) with the highest testosterone levels (median 31.8 nmol/1, range 30.1 - 34.8), and the 52 men (group 2) with the lowest testosterone levels (median 14.0 nmol/1, range 10.5 - 15.5). We chose the men examined in 2003, as this was the latest year from which we had completed analyses of reproductive hormones. The median age of men in group 1 was 18,9 years (range 18,2 - 24,6), median age in group 2 was 19,0 years (range 18,2 - 25,1), and median age for all 105 young men was 19.0 years (range 18.2 - 25.2) years. Information on ejaculation abstinence period and hour of blood sampling was recorded. All samples of semen and blood were collected between 8.30 a.m. and 1.15 p.m. Serum was stored at -20C until chemical analysis. The Danish National Committee on Biomedical Research Ethics, Copenhagen Region, approved the research, and all young men gave written informed consent. 6 Page 7 of 26 tJ p. 18 PFAA analysis. Thawed serum samples were analyzed in January 2008 for ten different perfluorinated chemicals, with carbon chain length from C6 to Cl 3: PFHxS (perfluorohexane sulfonic acid), PFHpA (perfluoroheptanoic acid), PFOA (perfluorooctanoic acid), PFOS (perfluorooctane sulfonic acid), PFOSA (perfluorooctane sulfonamide), PFNA (perfluorononanoic acid), PFDA (perfluorodecanoic acid), PFUnA (perfluoroundecanoic acid), PFDoA (perfluorododecanoic acid), and PFTrA (perfluorotridecanoic acid). One mL of serum was spiked with the surrogate standards ,3C8PFOA, 13C2-PFDA and 13C4-PFOS and extracted according to the ion pairing method described previously (Hansen et al. 2001). Matrix-matched standards were prepared by spiking rabbit serum (Sigma Aldrich, Schnelldorf, Germany) with the analytes and the surrogate standards. Blank samples consisted of rabbit serum spiked with only surrogate standards. Standards and blanks were extracted together with each batch of samples. Instrumental analysis was performed by liquid chromatography-tandem mass spectrometry (LC-MS-MS) with electrospray ionization (ESI). The extracts (20 pL injection volume) were chromatographed on a C l8 Betasil column (2.1 x 50 mm, Thermo Hypesil-Keystone, Bellafonte, PA) using an Agilent 1100 Series HPLC (Agilent Technologies, Palo Alto, CA). The high performance liquid chromatography (HPLC) was interfaced to a triple quadrupole API 2000 (Sciex, Concorde, Ontario, Canada) equipped with a Turboion Spray source operating in negative ion mode. Chromatographic conditions and transition MS-MS ions have been described in details previously (Bossi et al. 2005). The limits of detection (LODs) ranged from 0.1 to 0.5 ng/mL. Reproductive hormone analysis. Thawed serum samples were analyzed for the levels of testosterone, estradiol, SHBG (sex hormone binding globulin), LH (luteinizing hormone), 7 p. 19 . Page 8 of 26 FSH (follicle stimulating hormone) and inhibin B as described previously (Paasch et al. 2008). Free androgen index (FAI) was calculated as [testosterone x 100 / SHBG). Ratios between hormones were calculated by simple division. Semen analysis. Semen volume was assessed by weight, and the sperm concentration by use of a Brker-Trk haemocytometer. Total sperm count was calculated as semen volume x sperm concentration. The percentage of motile spermatozoa (WHO class A+B+C) was assessed on fresh samples. Sperm morphology slides were fixed, Papanicolaou stained and all assessed according to strict criteria (Menkveld et al. 1990) by one trained technician over a period of one week. Further details of the semen analysis can be seen in previously published work (Jprgensen et al. 2002). Statistical analysis. Medians and 5-95,h percentiles were used to describe the levels of PFAAs in group 1 and 2. Mann-Whitney t/-test was used to compare the groups 1 and 2 with respect to PFAA levels, BMI (body mass index) and smoking status. Samples with values below LOD were set to 0 ng/mL. We used Pearson's correlation coefficients and related p-values to describe correlations between levels of different PFCs. Univariate regression analysis was done for comparison of hormone levels between group 1 and 2, and to describe associations between PFAAs and hormones or semen variables. Sperm concentration, semen volume and total sperm count were adjusted for the effect of ejaculation abstinence period. Sperm motility was adjusted for time between ejaculation and assessment of motility. Sex hormone concentrations were adjusted for hour of blood sampling. Semen variables and hormone levels and ratios, except sperm morphology and total testosterone, were In transformed to obtain normality of the residuals. Smoking and 8 BMI were tested for confounding effects but were found to be non-significant, and therefore not included in the final analyses. We proceeded to analyze for associations between PFAAs and testicular function (hormone levels and semen quality) for the whole group of 105 men. We singled out PFOS and PFOA, and results were calculated as estimated changes in endpoint (reproductive hormones and semen variables) with a change in serum concentration of 1 ng/mL of PFOS, PFOA concentrations, and the summed concentration of PFOS and PFOA. We divided the men into three groups from the combined concentrations of PFOS and PFOA. Each sample was given a quartile score of 1 to 4 for PFOS and PFOA levels separately. Score 1 was given to samples with levels within the lowest quartile, and score 4 within the highest quartile. We then summed the quartile scores for PFOS and PFOA, giving each sample a possible score from 2 to 8. Samples were then divided into 3 quartile groups for PFOS and PFOA combined: "Low PFAA" group (N=29) with summed quartile score from 2 to 3, "intermediate PFAA" (N=48) group with score 4 to 6, and "high PFAA" (N=28) group with score 7 to 8. Analysis for association between quartile group and hormone levels or semen variables was done using univariate regression analysis, adjusted for the abovementioned confounders. Statistical analysis was performed using SPSS statistical software version 16.0 (SPSS Inc., Chicago, IL, USA). Results Levels o f PFAAs in serum. The serum levels and number of samples above LOD for all PFAAs are shown for the low- and high testosterone groups and the entire group of 105 men (Table 1). Except for PFOSA, which was detected in only 56 men, there was no significant difference in levels of any PFAA between group 1 and 2. The median PFOS, 9 p. 21 . Page 10 of 26 PFOA and PFHxS concentrations for the whole group of men were 24.5,4.9 and 6.6 ng/mL, respectively, and only these were included in the final regression analyses. The remaining PFAAs were found in much lower concentrations, and therefore these results are not discussed further. PFOS levels were positively correlated with PFOA (r = 0.594, p < 0.0005) and PFHxS (r = 0.304, p = 0.002) levels. PFOA and PFHxS levels were positively correlated, but not statistically significantly (r = 0.136, p = 0.2). PFAAs and semen variables. In the whole group of 105 men, there was a tendency toward reduced levels of all semen variables in the "high" PFAA quartile group compared to the "low" group (groups constructed to include PFOS and PFOA levels, see statistical analysis), Table 2. The difference in percentage of morphologically normal spermatozoa as well as in the total number of normal spermatozoa (total sperm count x % morphologically normal sperms) was statistically significant (p = 0.037 and 0.030, respectively). In the high PFOS-PFOA group the median number of normal spermatozoa in the ejaculate was 6.2 million vs. 15.5 million in the low group (Figures 1 and 2). When analyzing associations of semen variables to PFOS and PFOA separately, as well as the simple summed concentration of PFOS and PFOA, estimated changes in semen variables with a change in serum PFAA concentration of 1 ng/mL indicated negative but non-significant associations between the PFAAs and semen variables (Table 3). PFAAs and reproductive hormones. There was no significant association between testosterone levels and PFAA levels, and no significant difference in PFAA levels between the high- and low-testosterone groups. For the whole group of 105 men, adjusted medians for the hormones point to a negative association to PFAA levels - however, none of these tendencies were statistically significant (Table 2). 10 Page 11 of 26 Estimated associations between reproductive hormones and PFOS and PFOA separately and the summed concentration of PFOS and PFOA, showed no significant associations (Table 4). Smoking and BMI. 35% of the 105 men were smokers, and there were more smokers in the high testosterone group compared to the low testosterone group (49% and 21% smokers, respectively, p = 0.003). Smoking status was significantly associated to higher testosterone, lower estradiol and higher SHBG when entered as a confounder in univariate analyses for these three variables only (p = 0.004 to 0.009). Smoking was not associated with any semen quality variables (p = 0.1 to 0.4), or levels of any PFAA (p = 0.1 to 1.0). Including smoking as a confounder did not considerably change the presented results or significance levels. Therefore, smoking was not included in the final analyses. BMI was not associated to levels of any PFAA (p = 0.1 to 1.0), nor was there any confounding effect on any semen variables (e.g. p = 0.5 for morphologically normal sperms, and p = 0.9 for total number of morphologically normal sperms in analyses for difference between high and low PFAA groups). Discussion and conclusions This study examined PFAA levels in young adults. High serum concentrations of PFAAs were significantly associated with reduced numbers of normal spermatozoa. In addition, sperm concentration, total sperm count and sperm motility showed some tendency towards lower levels in men with high PFAA levels, although not at a statistically significant level. A tendency toward lower inhibin B/FSH ratio with high PFAA levels was in agreement with these findings, as these hormones reflect the spermatogenetic activity. 11 p. 23 Page 12 of 26 Testosterone, FAI, testosterone/LH ratio and testosterone/estradiol ratio could suggest a poorer Leydig cell function in the "high" compared to the "low" PFAA quartile group. However, the associations between reproductive hormones and PFAAs were not completely consistent, and all were non-significant. This study therefore cannot demonstrate an adverse effect of PFAAs on Leydig cell function. We singled out PFOS and PFOA, as there are no data that specifically support PFHxS as an endocrine disruptor, and divided the men into three groups from the combined concentrations of PFOS and PFOA to account for a potential different effect at same concentration levels. As we could find no significant association between testosterone levels and PFAA levels, and no significant difference in PFAA levels between high- and low-testosterone groups, we could analyze associations between PFOS-PFOA levels and reproductive hormones or semen variables for the whole group. Controlling for confounding effect of smoking or BMI did not change estimates or significance levels. To our knowledge, there have been no consistent reports of associations between PFAA levels and smoking or BMI. Our material included only 105 men, and in addition, we had selected the men based on their serum testosterone values. The selection of two groups with high and low testosterone was done to test our primary hypothesis and therefore affects the homogeneity of the group when correlations are analyzed for the group as a whole. This could potentially influence the subsequent analysis of semen quality by bias or confounding, and may affect the general applicability of the results. A larger follow-up study should preferably include randomly selected men from the general population. Our study is, to our knowledge, the first report of a correlation between semen quality and PFAAs. Very few studies of other endocrine disruptors (e.g. phthalates and pesticides) have previously demonstrated such an association (Duty et al. 2003; Hauser et 12 Page 13 of 26 1.1 p. 24 al. 2003; Hauser et al. 2007; Meeker et al. 2008; Swan et al. 2003). If the results from our preliminary study of an association between high levels of PFAAs and decreased number of normal sperms are confirmed, then high levels of PFAAs may be regarded as another endocrine disrupting factor contributing to the low semen quality seen among many young men. However, the importance of mixture effects of low-dose exposure to multiple compounds is becoming evident from animal studies but remains to be studied in humans (Hass et al. 2007; Kortenkamp et al. 2007). The mode of action by PFAAs is not clear and only a few animal studies have explored mechanistic issues. These show decreased testosterone levels and reduced expression of steroidogenesis genes associated with Leydig cell hyperplasia in adult animals (Biegel et al. 1995; Shi et al. 2007), suggesting a direct testicular effect. Recent studies have indicated that the fetal gonad is particularly sensitive to exogenous factors (Skakkebaek et al. 2001). However, our results could indicate that exposures later in life may contribute to impairment of semen quality, in line with other recent studies (Hauser et al. 2007). We speculate that morphology is perhaps more susceptible to this than sperm concentration or total sperm count. Sperm morphology has proven an important indicator of semen quality and fertility in a clinical setting, even in men with normal sperm concentration (Guzick et al. 2001). Interestingly, a recent study showed higher levels of maternal PFOS and PFOA levels in early pregnancy was associated with significantly longer waiting time to pregnancy (Fei et al. 2009). We speculate that men and women living together may have similar exposure to PFAAs and that decreased semen quality caused by high PFAA levels may contribute to the longer waiting time to pregnancy found in that study. The use and emission polyfluorinated compounds continue to increase, and they are not readily cleared from the environment (Jensen et al. 2008; Prevedouros et al. 2006). 13 p. 25 Page 14 of 26 Therefore, humans and wildlife worldwide will be exposed for years to come. We found positive correlations between levels of different PFAAs, as has been found previously (Apelberg et al. 2007; Calafat et al. 2007; Fei et al. 2007), suggesting common sources of exposure. The PFAAs levels we have detected are comparable to those found in other countries like Sweden and the Faroe Islands (Karrman et al. 2007; Weihe et al. 2008), but lower than earlier results from Denmark from 1996-2002 (Fei et al. 2007). Thus, the effects we have indicated may also be true for other than the Danish population. In conclusion, our results indicate that higher PFAA levels were associated with lower numbers of normal sperms. In addition, we found non-significant negative associations between PFAA levels and other semen variables and reproductive hormones. Thus, high levels of PFAAs may contribute to the otherwise unexplained low semen quality seen in many young men. However, results from this first and preliminary study should be corroborated in larger studies. 14 Page 15 of 26 p. 26 References Andersen AG, Jensen TK, Carlsen E, Jorgensen N, Andersson AM, Krarup T et al. 2000. High frequency of sub-optimal semen quality in an unselected population of young men. Hum Reprod 15:366-372. Apelberg BJ, Witter FR, Herbstman JB, Calafat AM, Halden RU, Needham LL et al. 2007. Cord serum concentrations of perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA) in relation to weight and size at birth. Environ Health Perspect 115:1670-1676. Biegel LB, Liu RC, 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. Bossi R, Riget FF, Dietz R. 2005. Temporal and spatial trends of perfluorinated compounds in ringed seal (Phoca hispida) from Greenland. Environ Sci Technol 39:7416-7422. Calafat AM, Kuklenyik Z, Caudill SP, Reidy JA, Needham LL. 2006. Perfluorochemicals in pooled serum samples from United States residents in 2001 and 2002. Environ Sci Technol 40:2128-2134. Calafat AM, Wong LY, Kuklenyik Z, Reidy JA, Needham LL. 2007. Polyfluoroalkyl chemicals in the U.S. population: data from the National Health and Nutrition Examination Survey (NHANES) 2003-2004 and comparisons with NHANES 1999 2000. Environ Health Perspect 115:1596-1602. 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. Duty SM, Silva MJ, Barr DB, Brock JW, Ryan L, Chen Z et al. 2003. Phthalate exposure and human semen parameters. Epidemiology 14:269-277. 15 p. 27 Page 16 of 26 Fei C, McLaughlin JK, Lipworth L, Olsen J. 2009. In press. Maternal levels of perfluorinated chemicals and subfecundity. Hum Reprod. Fei C, McLaughlin JK, Tarone RE, Olsen J. 2007. Perfluorinated chemicals and fetal growth: a study within the Danish National Birth Cohort. Environ Health Perspect 115:1677-1682. Giesy JP, Kannan K. 2001. Global distribution of perfluorooctane sulfonate in wildlife. Environ Sei Technol 35:1339-1342. Guruge KS, Taniyasu S, Yamashita N, Wijeratna S, Mohotti KM, Seneviratne HR et al. 2005. Perfluorinated organic compounds in human blood serum and seminal plasma: a study of urban and rural tea worker populations in Sri Lanka. J Environ Monit 7:371-377. Guzick DS, Overstreet JW, Factor-Litvak P, Brazil CK, Nakajima ST, Coutifaris C et al. 2001. Sperm morphology, motility, and concentration in fertile and infertile men. N Engl J Med 345:1388-1393. Hansen KJ, Clemen LA, Ellefson ME, Johnson HO. 2001. Compound-specific, quantitative characterization of organic fluorochemicals in biological matrices. Environ Sei Technol 35:766-770. Hass U, Scholze M, Christiansen S, Dalgaard M, Vinggaard AM, Axelstad M et al. 2007. Combined exposure to anti-androgens exacerbates disruption of sexual differentiation in the rat. Environ Health Perspect 115 Suppl 1:122-128. Hauser R, Chen Z, Pothier L, Ryan L, Altshul L. 2003. The relationship between human semen parameters and environmental exposure to polychlorinated biphenyls and p,p'-DDE. Environ Health Perspect 111:1505-1511. Hauser R, Meeker JD, Singh NP, Silva MJ, Ryan L, Duty S et al. 2007. DNA damage in human sperm is related to urinary levels of phthalate monoester and oxidative metabolites. Hum Reprod 22:688-695. 16 Page 17 of 26 II p. 28 Jensen A.A., Poulsen P.B., Bossi R. 2008. Survey and environmental/health assessment of fluorinated substances in impregnated consumer products and impregnating agents. Survey of chemical substances in consumer products no. 99. Danish Agency for Environmental Protection. Http://www.mst.dk/Udgivelser/PubIications/2008/10/978-87-7052-845-0.htm [accessed 10 February 2009]. Jensen AA, Leffers H. 2008. Emerging endocrine disrupters: perfluoroalkylated substances. IntJA ndrol 31:161-169. Jprgensen N, Carlsen E, Nermoen I, Punab M, Suominen J, Andersen AG et al. 2002. East- West gradient in semen quality in the Nordic-Baltic area: a study of men from the general population in Denmark, Norway, Estonia and Finland. Hum Reprod 17:2199-2208. Kannan K, Corsolini S, Falandysz J, Fillmann G, Kumar KS, Loganathan BG et al. 2004. Perfluorooctanesulfonate and related fluorochemicals in human blood from several countries. Environ Sci Technol 38:4489-4495. Karrman A, Ericson I, van Bavel B, Damerud PO, Aune M, Glynn A et al. 2007. Exposure of perfluorinated chemicals through lactation: levels of matched human milk and serum and a temporal trend, 1996-2004, in Sweden. Environ Health Perspect 115:226-230. Kissa E. 2001. Fluorinated surfactants and repellents, 2nd edition. Surfactant science series, Vol. 97. New York: Marcel Dekker. Kortenkamp A, Faust M, Scholze M, Backhaus T. 2007. Low-level exposure to multiple chemicals: reason for human health concerns? Environ Health Perspect 115 Suppl 1:106-114. 17 p. 29 . Page 18 of 26 Meeker JD, Barr DB, Hauser R. 2008. Human semen quality and sperm DNA damage in relation to urinary metabolites of pyrethroid insecticides. Hum Reprod 23:1932 1940. Menkveld R, Stander FS, Kotze TJ, Kruger TF, van Zyl JA. 1990. The evaluation of morphological characteristics of human spermatozoa according to stricter criteria. Hum Reprod 5:586-592. Olsen GW, Burris JM, Ehresman DJ, Froehlich JW, Seacat AM, Butenhoff JL et al. 2007. Half-life of serum elimination of perfluorooctanesulfonate,perfluorohexanesulfonate, and perfluorooctanoate in retired fluorochemical production workers. Environ Health Perspect 115:1298 1305. Paasch U, Salzbrunn A, Glander HJ, Plambeck K, Salzbrunn H, Grnewald S et al. 2008. Semen quality in sub-fertile range for a significant proportion of young men from the general German population: a co-ordinated, controlled study of 791 men from Hamburg and Leipzig. Int J Androl 31:93-102. Prevedouros K, Cousins IT, Buck RC, Korzeniowski SH. 2006. Sources, fate and transport of perfluorocarboxylates. Environ Sei Technol 40:32-44. Shi Z, Zhang H, Liu Y, Xu M, Dai J. 2007. Alterations in gene expression and testosterone synthesis in the testes of male rats exposed to perfluorododecanoic acid. Toxicol Sei 98:206-215. Skakkebk NE, Rajpert-De Meyts E, Main KM. 2001. Testicular dysgenesis syndrome: an increasingly common developmental disorder with environmental aspects. Hum Reprod 16:972-978. Swan SH, Kruse RL, Liu F, Barr DB, Drobnis EZ, Redmon JB et al. 2003. Semen quality in relation to biomarkers of pesticide exposure. Environ Health Perspect 111:1478 1484. 18 Page 19-of 26 1 .1 p. 30 Weihe P, Kato K, Calafat AM, Nielsen F, Wanigatunga AA, Needham LL et al. 2008. Serum concentrations of polyfluoroalkyl compounds in Faroese whale meat consumers. Environ Sci Technol 42:6291-6295. 19 Table 1 i! p. 31 Page 20 of 26 PFAA median concentrations (5-95* percentiles) for high and low testosterone groups and all study subjects (ng/mL), and p-values for difference between the two groups. Samples High testosterone > LOD N=53 PFHxS PFHpA PFOA PFOS PFOSA PFNA PFDA PFUnA PFDoA PFTrA 105 98 105 105 56 105 104 101 102 7 6.6 (4.0-13.0) 0.2 (0.01 - 0.9) 4.4 (2.6 - 7.0) 25.5 (14.2-39.6) 0.1 (0.00-3.7) 0.8 (0.4-1.8) 0.9 (0.2-1.1) 0.1 (0.04-0.3) 0.08 (0.04-0.8) 0.00 (0.000 - 0.4) Low testosterone N=52 6.6 (3.5-12.1) 0.3 (0.00-1.3) 5.0 (2.7 - 7.5) 23.9 (12.8-45.2) 0.00 (0.00 - 3.5) 0.8 (0.4 - 2.0) 0.8 (0.4-1.2) 0.2 (0.00 - 0.4) 0.08 (0.02-0.8) 0.00 (0.00 - 0.06) Whole group N=105 6.6 (4.0-12.1) 0.2 (0.00-1.1) 4.9 (2.7 - 7. 2) 24.5 (14.2-42.1) 0.06 (0.00 - 3.5) 0.8 (0.4-1.8) 0.9 (0.3-1.1) 0.1 (0.02-0.4) 0.08 (0.04 - 0.8) 0.00 (0.00 - 0.2) Pvalue 0.8 0.8 0.1 0.4 0.008 0.6 0.9 1.0 0.8 0.2 20 Page 21 .of 26 II p. 32 Table 2 Adjusted means (95% Cl) for PFAA quartile groups, and p-value for difference between the low and high PFAA quartile groups. Sex hormones3 Low PFAA Intermediate PFAA High PFAA p-value N Testosterone (nmol/L) Estradiol (pmol/L) SHBG (nmol/L) LH (IU/L) FSH (IU/L) Inhibin-B (pg/mL) FAI Testosterone/LH FA1/LH Estradiol/Testosterone Inhibin/FSH n = :29 25.2 (21.7-28.7) 77.6 (70.3 - 85.8) 27.8 (24.0 - 32.2) 3.4 (2.8-4.1) 2.7 (2.1 -3.5) 181 (142 - 232) 84.1 (74.1 -95.5) 6.9 (5.6 - 8.4) 24.7 (19.9-30.7) 3.3 (3.0-3.7) 66.8 (42.1 - 106.0) N = 48 22.3 (19 .6 -2 5 .0 ) 72.4 (67.0 - 84.7) 25.4 (22.6-28.5) 2.9 (2.5 - 3.4) 2.7 (2.2 - 3-3) 175 (144-212) 80.2 (7 2 .6 -8 8 .7 ) 7.0 (5.9 - 8.2) 27.4 (23.1 -3 2 .5 ) 3.6 (3.2 - 3.9) 66.0 (45.9 - 94.8) n = :28 22.3 (18.8-25.8) 76.6 (69.3 - 84.7) 26.1 (22.5 - 30.3) 3.7 (3.0 - 4.4) 3.0 (2.3 - 3.8) 152 (119-195) 77.8 (68.5 - 88.3) 5.5 (4.5 - 6.8) 21.2 (17.1 -26.4) 3.8 (3.4 - 4.2) 51.3 (32.3-81.4) 0.2 0.9 0.5 0.6 0.6 0.3 0.4 0.1 0.3 0.1 0.4 Semen quality15 Volume (mL) 4.0 (3.2 - 5.0) Concentration (mio./mL) 59 (36 - 96) Total count (mio.) 228 (134-389) Motile sperms (%) 73 (69 - 77) Morphologically normal (%) 8.8 (7.2- 10.4) Total morphologically normal (mio.) 15.5 (7.3 - 33.0) 3.4 (2 .9 -4 .1 ) 51 (35 - 74) 172 (114-261) 70 (66 - 73) 7.7 (6.4 - 9.0) 10.0 (5 .6 -1 7 .9 ) 3.5 (2.9 - 4.4) 40 (25 - 64) 143 (86 - 237) 71 (66 - 75) 6.3 (4.6 - 8.0) 6.23 (3.0- 12.8) 0.3 0.2 0.1 0.4 0.037 0.030' a Hormone levels are adjusted for time of blood sampling. b Volume, concentration and total count are adjusted for duration of abstinence. Motility is adjusted for time between ejaculation and semen analysis. Morphology is not adjusted for confounders. 21 1i p. 33 Page 22 of 26 Table 3 Estimated change in semen variables3with a change in PFAA of 1 ng/mL (95% Cl). All 105 men included. PFOS PFOA PFAA sumb In Volume In Concentration In Total count In Motility Morphology 0.000 (-0.012 - 0.011) -0.020 (-0.044 - 0.005) -0.018 (-0.045 - 0.010) -0.006 (-0.019 - 0.007) -0.085 (-0.200 - 0.026) -0.002 -0.080 -0.074 -0.027 -0.540 (-0.070 - 0.066) (-0.230 - 0.066) (-0.230 - 0.086) (-0.110 - 0.053) (-1.200 - 0.110) 0.000 -0.018 -0.016 -0.006 -0.082 (-0.010-0.010) (-0.040 - 0.004) (-0.041 -0.008) (-0.018-0.007) (-0.181 -0.018) 3Volume, concentration and total count are adjusted for duration of abstinence. Motility is adjusted for time between ejaculation and semen analysis. Morphology is not adjusted for confounders. b"PFAA sum" is PFOS and PFOA mass concentrations summed (ng/mL). 22 Table 4 Estimated change in reproductive hormones" with a change in PFAA concentration of 1 ng/mL (95% Cl). All 105 men included. PFOS PFOA PFAA sumb Testosterone -0.087 (-0.32 - 0.15) In Estradiol - 0.001 (-0.008 - 0.005) ln SHBG 0.002 (-0.007 -0.012) ln LH 0.000 (-0.014 -0.012) InFSH 0.004 (-0.13- 0.22) In Inhibin-B -0.004 (-0.21 - 0.12) lnFAI -0.006 (-0.015 - 0.002) In Testosterone/LH -0.003 (-0.017 - 0.011) ln FAI/LH -0.006 (-0.020 - 0.009) In Estradiol/ Testosterone 0.003 (-0.005 -0.010) In Inhibin/FSH -0.009 (-0.039 - 0.022) -0.98 (-2.33 - 0.37) -0.012 (-0.051 - 0.027) -0.009 (-0.067 - 0.048) -0.010 (-0.084 -0.064) -0.037 (-0.14- 0.064) 0.012 (-0.084 -0.11) -0.038 (-0.087 -0.011) -0.037 (-0.12- 0.045) -0.028 (-0.114 - 0.058) 0.035 (-0.010 -0.081) 0.049 (-0.13- 0.23) -0.093 (-0.303 -0.116) - 0.001 (-0.007 - 0.005) 0.002 (-0.007 -0.011) 0.000 (-0.012 - 0.010) 0.003 (-0.013--0.018) -0.003 (-0.018 - 0.012) -0.006 (-0.014 - 0.001) -0.003 (-0.016 - 0.009) -0.005 (-0.018 - 0.008) 0.003 (-0.004 - 0.010) -0.006 (-0.034 -- 0.022) dHormone levels are adjusted for time of blood sampling. b"PFAA sum" is PFOS and PFOA mass concentrations summed (ng/mL). Figure legends Figure 1. Morphologically normal spermatozoa (%) and PFAA quartile groups (adjusted means and 95% Cl). All 105 men included. * p = 0.037 for difference compared to Low PFAA quartile group. Figure 2. Total morphologically normal spermatozoa (mio.)a and PFAA quartile groups (adjusted means and 95% Cl). All 105 men included. a Total morphologically normal spermatozoa is adjusted for duration of abstinence. * p = 0.030 for difference compared to Low PFAA quartile group. Page 25 of 26 Figure 1. Morphologically normal .spermatozoa (%). Low PFAA Intermediate PFAA High PFAA p. 36 Morphologically normal spermatozoa (% ) and PFAA quartile groups (adjusted means and 9 5% C l). All 105 men included. * p = 0.0 3 7 for difference compared to Low PFAA quartile group. 201x285mm (150 x 150 DPI) figure 2. Total morphologically normal spermatozoa (into.). Low PFAA intermediate PFAA High PFAA Total morphologically normal spermatozoa (m io.)a and PFAA quartile groups (adjusted means and 95% C l). All 105 men included. a Total morphologically normal spermatozoa is adjusted for duration of abstinence. * p = 0.0 3 0 for difference compared to Low PFAA quartile group. 201x285mm (150 x 150 DPI)