Document ZBo2d1aMoK87qv0JqJbk2pj1J

P-1 "McCrea, Deborah" <mccrea@taftlaw.com> 09/01/2009 11:30 AM To NCIC OPPT@ EPA cc "Bilott, Robert A." <bilott@ taftlaw.com> bcc Subject 09/01/2009 Letter To EPA Docket Center /? A a -e -M 0 7 Taft/ Deborah McCrea / Legal Assistant Taft Stettinius & Hollister LLP 425 Walnut Street, Suite 1800 Cincinnati, Ohio 45202-3957 Tel: 513.381.2838 Fax: 513.381.0205 www.taftlaw.com / mccrea@taftlaw.com o U3 Internal Revenue Service Circular 230 Disclosure: As provided for in Treasury regulations, advice (if any) relating to federal taxes that is contained in this communication (including attachments) is not intended or written to be used, and cannot be used, for the purpose of (1) avoiding penalties under the Internal Revenue Code or (2) promoting, marketing or recommending to another party any transaction or matter addressed herein. This message may contain information that is attorney-client privileged, attorney work product or otherwise confidential. If you are not an intended recipient, use and disclosure of this message are prohibited. If you received this transmission in error, please notify the sender by reply e-mail and delete the message and any attachments. 4957_001.pdf CONTAliyp NO CB P-2 Taft/ Taft Stettinius & Hollister LLP 425 W alnut Street, Suite 1800/Cincinnati, OH 45202-3957 /Tel: 513.381.2838 / Fax: 513.381.0205 / www.taftlaw.com Cincinnati /Cleveland /Colum bus /D ayton /Indianapolis /N orthern Kentucky /Phoenix /Beijing Ro b er t A . Bilo tt 513-357-9638 bilott@tafUaw.conn September 1, 2009 FEDERAL EXPRESS EPA Docket Center, MC 2822T U.S. Environmental Protection Agency EPA West, Room 3334 1301 Constitution Avenue, NW Washington, D.C. 20004 Re: Submission to IRIS and AR-226 Database For PFOA/PFOS: EPA-HQORD-2003-0016 To IRIS Database for PFOA/PFOS: In response to the Notice issued by USEPA on February 23, 2006, regarding USEPA's efforts to consider perfluorooctanoic acid ("PFOA") and perfluorooctane sulfonate ("PFOS") within the Integrated Risk Information System ("IRIS"), 71 Fed. Reg. 9333-9336 (Feb. 23, 2006), we are submitting the following additional information to USEPA for inclusion in that review, and for inclusion in the AR-226 database: 1. Stein, C.R., Savitz, D.A., and Dougan, M., "Serum Levels of Perfluorooctanoic Acid and Perfluorooctane Sulfonate and Pregnancy Outcome," Am. J. Epidem. (online version) (doi: 10.1093/aje/kwp212 (August 19, 2009). RAB:mdm Enclosure cc: Gloria Post (NJDEP)(w/ end.) (via U.S. Mail) Helen Goeden (MDH)(w/ end.) (via U.S. Mail) Lora Werner (ATSDR)(w/ end.) (via U.S. Mail) 11493355.1 CONTAINS NO CB P-3 American Journal of Epidemiology Advance Access published August 19,2009 American Journal of Epidemiology ThB Author 2009. Published by the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: joumals.permissionseosfordioumals.org. DOI: 10.1093/aje/kwp212 Original Contribution Serum Levels of Perfluorooctanoic Acid and Perfluorooctane Sulfonate and Pregnancy Outcome Cheryl R. Stein, David A. Savitz, and Marcelle Dougan Initially submitted March 27, 2009; accepted for publication June 19, 2009. The authors examined the association of serum perlluorooctanoic acid (PFOA) and perfluorooctane sulfonale (PFOS) with self-reported pregnancy outcome in Mid-Ohio Valley residents (2000-2006) highly exposed to PFOA. Data on 1,845 pregnancies within the 5 years preceding exposure measurement were analyzed for PFOA, and data on 5,262 pregnancies were analyzed for PFOS. Generalized estimating equations were used to calculate adjusted odds ratios and 95% confidence intervals. Neither PFOA nor PFOS showed any association with mis carriage or preterm birth. Preeclampsia was weakly associated with PFOA (adjusted odds ra tio 1.3, 95% confidence interval: 0.9,1.9) and PFOS (adjusted odds ratio = 1.3,95% confidence interval: 1.1,1.7) exposures above the median. PFOA was not associated with an increase in low birth weight, but PFOS showed an increased risk above the median (adjusted odds ratio = 1.5,95% confidence interval: 1.1,1.9) and a dose-response gradient. Birth defects were weakly associated with PFOA exposures above the 90th percentile (adjusted odds ratio = 1.7, 95% confidence interval: 0.8,3.6). This study identified modest associations of PFOA with preeciampsia and birth defects and of PFOS with preeclampsia and low birth weight, but associations were small, limited in precision, and based solely on self-reported health outcomes. congenital abnormalities; fluorocarbons; infant, low birth weight; pre-eclampsia; pregnancy; pregnancy outcome; premature birth Abbreviations: Cl, confidence interval; OR, odds ratio; PFC, perfluorinated compound; PFOA, perfluorooctanoic acid; PFOS, perfluorooctane sulfonate. Perfluorooctanoic acid (PFOA) is a synthetic chemical that has been used in the manufacture of fluoropotymers since the 1950s ( 1) and may also result from the breakdown o f a related group of chemicals called fluorinaled telomers (2). Fluoropolym ers are used in nonstick cookware and clothing made from waterproof, breathable fabric (2), PFOA and another perfluorinated compound (PFC) with comparahle industria! uses-- perfluorooctane sulfonate (PFOS)-- are persistent environmental pollutants that have been detected worldwide in both wildlife and humans, with higher exposure closer to urbanized and industrialized re gions (3). In the general US population, PFOA and PFOS were detected in all serum sam ples from the 1999-2000 National Health and Nutrition Examination Survey (4) and in 99.7% and 99.9%, respectively, of serum samples from the 2003-2004 National Health and Nutrition Examination Survey (5). In the 1999-2000 survey, geometric mean values were measured at 5.2 ng/mL for PFOA and 30.4 ng/mL for PFOS (4). These values dropped to 3.9 ng/m L and 20.7 ng/ mL, respectively, in the 2003-2004 survey (5). PFOA and PFOS have been detected in maternal and umbilical cord blood (6-8) and breast milk (9-13). The serum elimination half-life is approximately 4 years for PFOA and 5 years for PFOS (14). Toxicology studies highlight the potential for these PFCs to affect fetal growth and development. A series of studies suggest that in utero exposure to PFCs is associated with a range of nonspecific adverse infant outcomes in mouse, rat, and rabbit models (15), including reduced fetal weight and increased neonatal mortality. Several epidemiologic Correspondence to Dr, Cheryl R. Siein, Department of Community and Preventive Medicine, Mount Sinai School ol Medicine. One Gustave L. Levy Place, Box 1057, New York, NY 10029-6574 (e-mail: cheryl.s1ein@mssm.edu). 1 P-4 2 Stein et al. studies have evaluated the association between PFCs and reproductive health in the general population (16-19). These recently reviewed studies (20) suggest small, incon sistent decrements in birth weight and other infant anthro pometric measures in relation to maternal PFOA and PFOS exposure. Two studies examining highly exposed popula tions found no association between PFC exposure and birth outcome (21, 22). A chemical plant in the Mid-Ohio Valley near Parkersburg, West Virginia, has used PFOA in the manufacture of Auoropolymers since 1951. In 2001, a group of residents from the West Virginia and Ohio communities surrounding the plant filed a class action lawsuit alleging health damage from drinking water supplies drawing on PFOA-contaminated groundwater (23). Groundwater contamination from the Ohio River and air deposition is believed to be the primary exposure route for this population (24). The settlement of the class action lawsuit included a baseline survey, the C8 Health Project, conducted in 2005-2006 that gathered data from more than 69,000 people from 6 contaminated water districts surrounding the plant (23). (C8 is another name Tor PFOA, denoting its chain o f 8 carbons.) The C8 Health Project included demographic and health questionnaires and measurement o f 10 PFCs in serum. The current study used these data to examine the association between serum PFOA and PFOS measurements at C8 Health Project enroll ment and pregnancy outcomes in the 5 years prior to enrollment. M ATERIALS AND METHODS C 8 Health Project population The C8 Health Project enrolled subjects between August 2005 and July 2006. The Project's purpose was to collect health data from members of the class action lawsuit through written questionnaires and blood tests, including measurement of PFCs. This community was highly exposed to PFOA, but PFOS exposure reflects background levels. Individuals were eligible to participate in the C8 Health Project if they could provide documentation such as utility bills, properly deeds, employment pay stubs, or school rec ords proving they had consumed water for at least 1 year between 1950 and December 3,2004, while living, working, or attending school in a water district supplied by Little Hocking Water Association of Ohio; City of Belpre, Ohio; Tuppcr Plains-Chester District of Ohio; Village of Pomeroy! Ohio; I.ubeck Public Service District of West Virginia; Mason County Public Service District of West Virginia; or private water sources within areas of documented PFOA contamination. Participants were compensated $400 for completing the questionnaire and providing blood. The C8 Health Project collected data on 69,030 people. The total number of people eligible to join the class action lawsuit is unknown. Participation rates based on US Census counts of current residents of the eligible water districts are estimated at approximately 80% and, in some ZIP codes within eligible water districts, appear to be close to 100% (23). The overall participation rate is likely lower than 80% because former residents were also eligible to participate. In this population, the strongest predictor of PFOA serum level was current residence in a contaminated water district, with distance to the plant directly affecting PFOA levels (25). PFOS exposure is not determined by residence in a PFOAcontaminated water district. Study population This analysis o f serum PFOA and PFOS levels and recent pregnancy outcomes was restricted to pregnancies among female C8 Health Project participants with scrum PFC measurements. For the PFOA analyses, we included single ton miscarriages, stillbirths, or livebirths in the 5 years prior to C8 Health Project enrollment to women who lived in the same contaminated water district from the approximate start of the pregnancy through the time o f enrollment, The 5-year time frame and residency restrictions were used to ensure that the PFOA level measured at C8 Health Project enroll ment would reflect the level at the time o f pregnancy. For miscarriages, women must have lived in the same water district from 90 days before the pregnancy ended through enrollment. For stillbirths and livebirths, women must have lived in the same water district for 283 days before the pregnancy ended through enrollment. Each pregnancy was evaluated separately for eligibility. For the PFOS analyses, we included all singleton miscarriages, stillbirths, or live births in the 5 years prior to C8 Health Project enrollment. There was no need to restrict pregnancies to women living in the same water district from pregnancy through enrollment because PFOS exposure was not specific to the water district. Measures For laboratory analysis of PFCs, automated solid-phase extraction combined with reverse-phase high-performance liquid chromatography was used (12). Samples below the limit of detection (0.5 ng/mL) were assigned a value of 0.25 ng/mL. We examined PFCs as natural log-transformed con tinuous variables, and here we present the odds ratios reflecting a shift from the 25th to the 75th percentile, which corresponds to an increase from 10.3 ng/mL to 49.8 ng/mL for PFOA and from 9.0 ng/mL to 17.7 ng/mL for PFOS. We also categorized PFCs to enable identification of a high-exposure effect: <50th percentile (the referent), 50th-<75th percentile, 75th-90th percentile, and >90th percentile. The outpoints for the interquartile shift in PFOA do not match the cutpoints for the categorical measure; the categorical cutpoints were based on all eligible pregnancies, and the interquartile shift reflects the pregnancies included in the analysis. We examined the relation between PFC level and self reported pregnancy outcomes from the C8 Health Project questionnaire. Examined pregnancy outcomes included miscarriage (pregnancy loss from 6 weeks' to <20 weeks' gestation), preterm birth (livebirth at <37 weeks' gestation) and low birth weight (livebirth <5.5 pounds, approximately equivalent to 2,500 g). The number of reported stillbirths (fetal death from 20 weeks' to 42 weeks' gestation) was too small to analyze separately, but stillbirths were included in P-5 Serum PFCs and Pregnancy Outcome 3 the denominator in our analysis of miscarriages. We also evaluated preeclampsia and reported birth defects. Because of the cross-sectional nature of the C8 Health Project, some of the covariates in the questionnaire were not specific to the time of pregnancy. Covariates defined explic itly for a given pregnancy were maternal age (15-19, 20-34, 35-55 years) and parity (0, >1 previous livebirths). Mater nal education 1 2 , 12, >12 years) and smoking status (never, current, former smoker) were defined at the time of enrollment. Additionally, we restricted the analysis to white women (97.3% of the population) with nonmissing covariate data and no prepregnancy diabetes, although miss ing outcome information was permitted because each out come was analyzed separately. Institutional review board approval was granted from the Mount Sinai Program for the Protection of Human Subjects. Statistical analysis Analysis was performed by using SAS version 9.1 soft ware (SAS Institute. Inc., Cary, North Carolina). Covariates associated with both outcome and exposure--age, parity, ed ucational level at interview, smoking status at interview-- were included in all adjusted models together with either PFOS or PFOA. We ran logistic regression models using generalized estimating equations and an exchangeable corre lation structure to obtain crude and adjusted odds ratios and 95% confidence intervals accounting for the lack of indepen dence across multiple pregnancies to the same woman. RESULTS Women in the C8 Health Project reported 5,663 singleton pregnancies ending in a miscarriage, stillbirth, or livebirth during the 5 years prior to enrollment. Of these pregnancies, 5,624 were to women with PFC measurements and 5,262 (93.6% of pregnancies with PFC measurements) were to white women with complete covariate data and no prepreg nancy diabetes, and they were included in the analyses of PFOS (Table 1). Exclusion of pregnancies to women not liv ing in the same water district from the approximate start of pregnancy through enrollment left 1,972 pregnancies. Much of this attrition occurred because we were unable to determine water district lor at least part of the time from pregnancy through enrollment because of extensive missing information on self-reported water district. Restriction to pregnancies to white women with complete covariate data and no prepreg nancy diabetes yielded 1,845 pregnancies (93.5% of those otherwise eligible) (Table 1) for the analyses of PFOA. The pregnancies excluded for lack of residential stability in a con taminated water district (i.e., pregnancies in the analysis of PFOS but not PFOA) were to younger (mean age, 26.5 years vs. 27.3 years), more nulliparous (48.3% vs. 45.4%) women, more of whom had fewer than 12 years of education (10.8% vs. 8.3%) and were current or former smokers (56.2% vs. 50.4%). The descriptive statistics presented in the text are for the smaller population relevant to PFOA; Table 1 provides descriptive statistics for both populations. Mean serum PFOA was 48.8 ng/mL (standard deviation, 77.8), and the median value was 21.2 ng/mL (interquartile range, 10.3-49.8) (Table 1). Mean and median serum PFOS concentrations were 15.0 (standard deviation, 8.2) and 13.6 ng/mL (interquartile range, 9.4--18.7), respectively. The cor relation between PFOA and PFOS concentration was min imal (Spearman p = 0.2, P < 0.001). Most women were pregnant between ages 20 and 34 years (82.2%) and had at least 12 years of education (91.7%). Smoking was rela tively common, with 30.8% smoking at enrollment in the C8 Health Project and 19.6% self-identifying as former smokers. Of all pregnancies, 249 (13.5%) were reported to have ended in miscarriage. Of livebirths, 329 (20.9%) were reported as preterm, 80 (5.0%) as low birth weight, and 156 (9.8%) as preeclampsia, and 74 women (4.7%) noted a birth defect. Neither PFOA nor PFOS was related to miscarriage (Table 2). Preeclampsia was weakly associated with PFOA (Table 3) (adjusted odds ratio (OR) = 1.3,95% confidence interval (Cl); 0.9, 1.9) and PFOS (adjusted OR = 1.3, 95% Cl: 1.1, 1.7) for exposures above the median, but without a dose-response gradient. The adjusted odds ratio for pre eclampsia for PFOS exposure above the 90th percentile (adjusted OR = 1.6, 95% Cl: 1.2, 2.3) was elevated. Pre term birth (Table 4) was unrelated to PFOA. The only sug gested association between PFOS and preterm birth was for exposures above the 90th percentile (adjusted OR = 1.4, 95% Q : 1.1, 1.7). Low birth weight followed no discemable pattern in re lation to PFOA (Table 5), with some tendency for risk to decrease as exposure increased, but was compatible with no effect in the analysis of a continuous measure of PFOA. When the categorical exposure measure was used, women in the 75th to 90th percentile of exposure were markedly less likely to deliver a low birth weight baby (adjusted OR -- 0.3, 95% Cl: 0.1, 0.9) compared with women below the 50th percentile of PFOA exposure, but no dose-response gradient was found, PFOS showed evidence of a positive association with all 3 exposure metrics. The adjusted odds ratio for exposure above the median was 1.5 (95% Cl: 1.1, 1.9). There was also a dose-response gradient when the categorical measure was used, with adjusted odds ratios ranging from 1.3 (95% Cl: 0.9, 1.8) for the 50th to 75th percentile of exposure to 1.8 (95% Cl: 1.2, 2.8) for expo sures above the 90th percentile, compared with exposures below the 50th percentile. Birth defects (Table 6) were weakly associated with PFOA, but not PFOS. For PFOA, the elevation in risk was found only above the 90th percentile (adjusted OR = 1.7, 95% Cl: 0.8, 3.6). This estimate was based on 12 cases in the uppermost exposure category. In the PFOS analysis, the association with birth defects remained imprecise and was weaker than for PFOA (adjusted OR = 1.3,95% Cl: 0.8,2.1). Overall, results were similar, but less precise when PFOS was analyzed by using the population restricted to women living in a single contaminated water district from the time of pregnancy through enrollment (data not shown). The primary differences were for low birth weight, with an increased risk above the median but no dose-response gradient, and for birth defects, with the association with PFOS exposure stronger in the smaller population. We then restricted our analyses to the 1,197 pregnancies within the 3 years prior to measurement, presumably P-6 6 Stein et al. Table 2. Crude and Adjusted3 Associations Between Serum Perfluorinated Compounds and Singleton Miscarriage1' Within 5 Years of C8 Health Project Enrollment, Mid-Ohio Valley, 2000-2006 Serum PFC Per increase from the 25th to toe 75to percentile' <50th percentile >50th percentile <50th percentile 50th-<75th percentile 75th-90th percentile >901h percentile PFOA Exposure (n - 1,845) No. of Cn m Crude OR Adjusted on 95% Cl 249 0.9 0.9 0.7, 1.2 122 1.0 1.0 127 1.0 1.0 0.7. 1.4 122 1.0 1.0 74 1.3 1.2 0.8, 1.7 35 0.8 0.8 0.5, 1.3 18 0.7 0.9 0.5, 1.6 PFOS Exposure (n = 5,252) No. of Cases Crude OR Adjusted OR 5% Cl 657 1.1 1.0 0.9,1.1 470 1.0 1.0 187 1.2 1.0 0.9, 1.3 295 1.0 1.0 175 1.1 1.1 0.9. 1.3 114 1.2 1.1 0.8, 1.4 73 1.2 0.9 0.7, 1.3 Abbreviations: Cl, confidence interval; OR, odds ratio; PFC, perfluorinated compound; PFOA, periluoroodanoie acid; PFOS, perfluorooctane sulfonate. 3 Adjusted for maternal age, parity, educational level at Interview, smoking status at interview, and PFOS in the analysis of PFOA and PFOA in the analysis of PFOS. b Modeled by using generaBzed estimating equations with an exchangeable correlation structure to account for multiple pregnancies per woman. c Change from toe 25th to the 75th percentile corresponds to a shift from 10.3 ng/m l to 49.8 rtg/mL for PFOA and from 9.0 ng/ml to 17.7 ng/mL for PFOS. value, but with imprecise confidence intervals. Nonetheless, these authors found that prenatal PFOA levels were inversely associated with mean adjusted birth weight (-10.6 g, 95% Cl: -20.8, -0.5) per ng/mL increase (17). A smaller effect was observed for PFOS. Two highly exposed populations have been studied to date (21. 22). An occupational cohort from a facility manufacturing PFOS-based chemicals and specialty films included 421 Iivebirths (21). No association was found be tween maternal PFOS exposure, as determined by geometric mean exposure for various job categories, and self-reported birth weight. The second study, in the same geographic area as the present study, used ZIP code-based ecologic mea sures of PFOA exposure, US Census-derived indicators of socioeconomic status, and gestational age and birth weight data from vital records for 1,555 singleton births (22). Among women in PFOA-exposed ZIP code areas, mean adjusted birth weight was reduced by 8.8 g (95% Cl: -86.1, 68.5), while the adjusted odds ratio for birth weight of less than 2,500 g was 0.43 (95% Cl: 0.17, 1.12), which Table 3. Crude and Adjusted3Associations Between Serum Perfluorinated Compounds and PreedampsiabWithin 5 Years of C8 Health Project Enrollment, Mid-Ohio Valley, 2000-2006 Serum PFC Per increase from the 25th to the 75th percentile' <50th percentile >50th percentile <50th percentile 50th-<75th percentile 75tb-90lh percentHe >90th percentile PFOA Exposure (n =.>1,589} No. of Cases Crude OR Adjusted OR 95% Cl O) d 156 1.2 1.1 64 1.0 1.0 92 1.5 1.3 0.9, 1.9 64 1.0 1.0 52 1.8 1.5 1.0, 2.3 27 1.3 1.2 0.7, 2.1 13 0.9 0.9 0.5. 1.8 PFOS Exposure (n = 4,556) No. Of Cesse 407 Crude OR 1.2 Adjusted on 95% Cl u 0.9, 1.3 280 1.0 1.0 127 1.5 1.3 1.1, 1.7 163 1.0 1.0 117 1.4 1.3 1.0, 1.7 65 1.3 1.1 0.8, 1.6 62 2.0 1.6 1 5 ,2 .3 Abbreviations: Cl, confidence interval; OR, odds ratio; PFC, perfluorinated compound; PFOA, periluorooctanoic acid: PFOS, perfluorooctane sulfonate. Adjusted lor maternal age, parity, educational level at interview, smoking status at Interview, and PFOS in the analysis ol PFOA and PFOA in toe analysis of PFOS. Modeled by using generalized estimating equations with an exchangeable correlation structure to account for m ultiple pregnancies per woman.* * Change from the 25th to the 75th percentile corresponds to a shift from 10.3 ng/mL to 49.8 ng/mL for PFOA and tram 9.0 ng/m l to 17.7 ng/mL for PFOS. p .7 Serum PFCs and Pregnancy Outcome 7 Table 4. Crude and Adjusted* Associations Between Serum Perfluorinated Compounds and Singleton Birth at Less Than 37 Weeks' Completed Gestation0 Within 5 Years of C8 Health Project Enrollment, Mid-Ohio Valley, 2000-2006 Serum PFC Per Increase from the 25lh to the 75th percentile* <50lh percentile >50th percentile <501h percentile 50th-<75th percentile 75th-90th percentile >90th percentile PFOA Exposure (n 1.571) No. of Cases Crude OR Adjusted OR 95% Cl 329 0.9 0.8 0.8, 1.1 PFOS Exposure (n = 4,512) No. ol Cases Crude OR Adjusted OR 95% Cl 1.015 1.1 1.1 1.0,1.3 166 1.0 1.0 745 1.0 1.0 163 0.9 1.0 0.7, 1.3 270 1.1 1.1 1.0, 1.3 166 1.0 1.0 481 1.0 1.0 89 1.1 1.1 0.8, 1.5 264 1.1 1.1 0.9, 1.3 42 0.7 0.7 0.5, 1.1 153 1.0 1.1 0.9, 1.3 32 0.9 0.9 0.6, 1.5 117 1.3 1.4 1.1, 1.7 Abbreviations: Cl, confidence Interval; OB, odds ratio; PFC, perfluorinated compound; PFOA, perfluorooctanoic acid; PFOS, perfluorooctane sulfonate. " Adjusted for maternal age, parity, educational level at Interview, smoking status at Interview, and PFOS in the analysis of PFOA and PFOA in the analysis of PFOS. 6 Modeled by using generalized estimating equations with an exchangeable correlation structure to account for multiple pregnancies per woman. c Change from the 25th to the 75th percentile corresponds to a shift from 10.3 ng/mL to 49.8 ng/mL tor PFOA and from 9.0 ng/mL to 17.7 ng/mL for PFOS. conversely suggests a lower risk in the exposed community. Cumulatively, these studies provide little basis for suggest ing a positive relation between PFOA and PFOS exposure and low birth weight. To our knowledge, no associations have been observed between gestational age examined either continuously (16, 22) or dichotomized as birth at less than 37 completed weeks' gestation (17,22). Studies of other indications of fetal growth, such as head and abdominal circumference, birth length, and ponderal index, have yielded equivocal results (16, 18, 19) that are not directly comparable to the measures used in the present study. Ours is the first known epidemiologic study to assess the association between PFC exposure and miscar riage, preeclampsia, and birth defects, with no opportunity to compare findings with those of previous studies. A central concern regarding interpretation of our results is the quality of the information on self-reported pregnancy outcome. Numerous studies show high concordance Table 5. Crude and Adjusted* Associations Between Serum Perfluorinated Compounds and Singleton Birth Weight ot Less Than 5.5 Pounds6-' Within 5 Years of C8 Health Project Enrollment. Mid-Ohio Valley, 2000-2006 Serum PFC Per increase from the 25th to the 75th percentile11 <50th percentile >50th percentile <50th percentile 50th-<75th percentile 75th-9Qth percentile >90th percentile PFOA Exposure (n = 1,589) No. of Ca m s Crude OR Actuated OR 95% Cl 80 0.7 0.7 0.5, 1.0 47 1.0 1.0 33 0.7 0.7 0.5, 1.2 47 1.0 1.0 21 0.9 1.0 0.6, 1.7 5 0.3 0.3 0 .1 ,0 9 7 0.7 0.8 0.3, 1.9 PFOS Exposure (n 4,561) NO. Of Cases Crude OR Adjusted OR 95% Cl 243 1.2 1.3 1.1, 1.6 168 1.0 1.0 75 1.4 1.5 1.1, 1.9 102 1.0 1.0 66 1.2 1.3 0.9. 1,8 43 1.4 1.6 1.1, 2.3 32 1.6 1.8 1.2, 2.8 Abbreviations: Cl, confidence interval; OR, odds ratio: PFC, perfluorinated compound; PFOA, pertluorooctenolc acid; PFOS, perlluorooclane sulfonate. * Adjusted tor maternal age, parity, educational level at Interview, smoking status at Interview, and PFOS in the analysis of PFOA and PFOA in the analysis ot PFOS. B Modeled by using generalized estimating equations with an exchangeable correlation structure to account for multiple pregnancies per woman. * One pound = 0.454 kg. d Change from the 25th to the 75th percentile corresponds to a shift from 10.3 ng/mL lo 49.8 ng/mL for PFOA and from 9.0 ng/mL to 17.7 ng/mL tor PFOS. p .8 4 Slein et al. Table 1. Univariate Characteristics of Singleton Pregnancies Within 5 Years of C8 Health Project Enrollment, Mid-Ohio Valley, 2000-2006 Unique women Pregnancies per woman 1 2 3 4 5 Pregnancy outcome Miscarriage Stillbirth Uvebirth Serum PFOA (ng/mL) at interview Mean (SD) Median (IQR) Serum PFOA (ng/mL) at interview 0 .2 5-< 2 i .3 ng/mL (<50th percentile) 21 3-<50.0 ng/mL {50th-<75th percentile) 50 0 -< 120.6 ng/mL (75th-90th percentile) 120.6-694.4 ng/mL (>90th percentile) Serum PFOS (ng/mL) at interview Mean (SO) Median (IQR) Serum PFOS (ng/mL) at interview 0.25-C12.7 ng/mL (<50th percentile) 12 .7-< i7 .7 ng/mL (50th-<75th percentile) 17.7-<23.2 ng/mL (75th-90tti percentile) 23.2-83.4 ng/mL (>90th percentile) Water district Beipre. Ohio Little Hocking, Ohio Pomeroy, Ohio Tupper Plains-Chester, Ohio Lubeck, West Virginia Mason County, West Virginia PFOA Population (n = 1.S4S) No. % All pregnancies 1.505 1,215 247 38 3 2 80.7 16.4 2.5 0.2 0.1 249 6 1,590 13.5 0.3 86.2 48.8 (77.8) 21.2(10.3-49.8) 925 50.1 460 24.9 281 15.2 179 9.7 15.0 (82) 13.6(9.4-18.7) n/a n/a n/a n/a 163 8.8 350 19.0 73 4.0 471 25.5 301 16.3 487 26.4 PFOS Population (n = 5,262} No. % 3.996 2.942 870 159 22 3 73.6 21.8 4.0 0.5 0.1 857 36 4,569 12.5 0.7 86.8 32.1 (59.7) 13.6 (7.4-292) n/a n/a n/a n/a 14.1 (7.7) 12.8 (9.0-17.7) 2,582 1,345 802 533 49.1 25.6 15.2 10.1 n/a n/a n/a n/a n/a n/a Teble continues generating values with more concordance between the mea surement and the exposure at the time of the pregnancy. Results for PFOA and PFOS were similar, but less precise (data not shown). D IS C U S S IO N We found little or no support for a possible association between PFOA or PFOS exposure and miscarriage or pre term birth, nor between PFOA and low birth weight. Both PFOA and PFOS were weakly related to precclampsia and PFOA was weakly associated with birth defects, but without dose-response gradients. The association between PFOS and low birth weight was stronger and indicated increasing risk with increasing exposure, warranting examination in other suitable populations. The limited reproductive and developmental toxicology literature suggests possible adverse effects of PFOA and PFOS on fetal growth and viability and on postnatal growth (26, 27). The evidence regarding teratogenicity is unclear, P-9 Serum PFCs and Pregnancy Outcome 5 Table 1. Continued Age at pregnancy, years 15-19 20-34 35-55 Parity 0 previous livebirths >1 previous livebirths Education at interview <12 years 12 years >12 years Smoking status at interview Never Former Current Unique women Births per woman 1 2 3 4 Preeclampsia Yes No Preterm birth Preterm birth <37 weeks Term birth Low birth weight Birth weight <5.5 pounds* Birth weight >5.5 pounds Birth defect Yes No PFOA Population (n = 1,845) No. % 123 1,517 205 6.7 82.2 11.1 838 1,007 45.4 54.6 154 614 1,077 8.3 33.3 58.4 915 49.6 362 19.6 568 30.8 Pregnancies ending in livebirths 1,395 1,207 181 7 0 86.5 13.0 0.5 156 1,433 9.8 90.2 329 1,242 20.9 79.1 80 1,509 74 1.516 5.0 95.0 4.7 95.3 PFOS Population (n - 5,262) No. % 426 4,367 469 8.1 83.0 8.9 2,528 2,734 48.0 52.0 493 1.706 3,063 9.4 32.4 58.2 2,415 1,136 1,711 45.9 21.6 32.5 3,748 2,986 706 53 3 407 4,159 79.7 18.8 1.4 0.1 8.9 91.1 1,015 3,497 22.5 77.5 243 4,318 188 4,381 5.3 94.7 4.1 95.9 Abbreviations: IQR, interquartile range: n/a, not applicable; PFOA, pertluorooctanoic acid; PFOS, perfluorooctane sultonate; SD. standard deviation. * One pound - 0.454 kg. with adverse effects observed for PFOS possibly attributable to maternal toxicity at higher doses (27). Few epidemiologic studies have examined the association between PFC exposure and birth outcome in humans. Of these studies, most have looked at a continuous measure of birth weight and found mixed results. In a pregnancy cohort in Japan (n - 428). adjusted analyses showed an association between higher PFOS in maternal serum and reduced infant birth weight (19). For each log-10 change in PFOS. infants weighed 148.8 g less (95% Cl: -297.0, -0.5). PFOA had less of an effect on birth weight (-75.1 g, 95% (3: -191.8, 41.6) (19). A reduction in mean adjusted birth weight (-104 g, 95% Cl: -2 1 3 ,5 ) per log-unit increase in cord serum PFOA was also noted in a hospital-based study (n = 293) in Baltimore, Maryland (16). In this study, PFOS, compared with PFOA, had a smaller effect on birth weight (-69 g, 95% Cl: -149, 10), A subset of the Danish National Birth Cohort (n = 1,400) examined birth weight of less than 2,500 g and reported no association with either PFC (17). The adjusted odds ratios were above the null 8 Stein et al. Teble 6. Crude and Adlusted* Associations Between Serum Perftuorinaiad Compounds and Singleton Birth Defects Within 5 Years of C8 Health Project Enrollment, Mid-Ohio Valley, 2000-2006 Serum PFC Per increase from the 25!h to the 75th percentile' < 50th percentile >50th percentile <50th percentile 50lh-<75th percentile 75th-90th percentile >90th percentile PFOA Exposure (n 1,590) No, of Cases Cnide OR /tojuried OR 95% Cl 74 1.2 1.1 0.8, 1.6 35 1.0 1.0 39 1.1 1.0 0.6, 1.7 35 1.0 1.0 18 1.1 1.0 0.5, 1.8 9 0.8 0.7 0.3, 1.6 12 1.8 1.7 0.8, 3.6 PFOS Exposure (n 4,659) No. of Cases Crude OR Adjusted OR 95% Cl 188 1.0 1.1 0.9, 1.3 135 1.0 1.0 53 1.0 1.0 0.8. 1.4 92 1.0 1.0 43 0.9 0.9 0.6, 1.3 29 1.0 1.1 0.7, 1.6 24 1.3 1.3 0.8, 2,1 u n , ouus rauo; r r u , peniuormated compound; PFOA, perfluorooctanoic acto; PFOS, perfluorooctane sulfonate. Adjusted for matama! age, parity, educational level at interview, smoking status at Interview, and PFOS in the analysis of PFOA and PFOA In the analysis of PFOS. * Modeied by using generalized estimating equations with an exchangeable correlation structure to account for multiple pregnancies per woman. ` C,,hf 9 fr m m2sth to the 75h percentile corresponds to a shift from 10.3 ng/mLto 49.8 ng/mL for PFOA and from 9.0 ng/mL to 17.7 ng/mL lor PFOS. between maternal report of birth weight and birth weight recorded in a medical record (28-31). Recall of gestational age is also fairly accurate (29, 32). The validity of self reported preeclampsia is less clear. Maternal recall of preeclampsia had low to moderate sensitivity and high spec ificity compared with medical records (33) and a positive predictive value o f only 59.2% (34). Whether maternal re call would be comparable in our population, with relatively low educational attainment but recent births, is unknown. A high proportion ot births in this population were re ported to have ended in miscarriage (13.5%), with frequen cies of 10% more common based on self-report. Of livebirths, 20.9% were reported to be preterm, higher than would be expected in a white population (8%-15%). Pre eclampsia was likewise reported more commonly (9.8%) than through medical record review (3%-7%). Reported bmh defects in 4.7% of livebirths are near the range of that expected (3%). Given the limited quality of self-reported inlormation on such outcomes, it seems likely that these high frequencies are an artifact of reporting quality rather than a reflection of a high background risk in the study area. Even with the frequency o f self-reported pregnancy out comes qualitatively higher than expected, there was no clear indication that these outcomes were associated with PFC exposure. Despite the limitations of relying on maternal report to evaluate pregnancy complications and birth outcomes, we noted several associations consistent with expectation. Ad vanced maternal age was clearly associated with an in creased risk of miscarriage. Smokers had a small increase m preterm births and a larger increase in low birth weight Risk ol preeclampsia was reduced for both smokers and parous women, as expected. Our focus on, at most, 5 years between pregnancy and recall should facilitate better recall. Hie potential for differential recall in relation to exposure status is unlikely, except insofar as residents of the more highly exposed water districts were aware of their exposure and reported adverse outcomes either more completely or excessively. The generally null results make differential re call unlikely, although nondifferential misclassification could have attenuated our results. Potential bias from non participation is also a concern; however, none of the known participation biases (smaller proportion of men, elderly, young children, nonlocal residents) would have impacted our study findings for PFOA because these potential partic ipants would not have been eligible for our study. Although residents of more highly exposed areas, or people experi encing more health problems, may be more likely to partic ipate, it is unlikely that there would have been a joint effect of exposure and outcome on participation, as needed to pro duce biased measures of association (35). The laboratory-determined exposure measurement does not suffer from the same quality issues as self-reported preg nancy outcomes, Our restriction to pregnancies occurring in the 5 years prior to exposure measurement, and for PFOA residential stability within affected water districts, should ensure that the measured value reflects the level at the time of pregnancy. Maternal serum PFOA and PFOS values ap pear to drop during pregnancy, but they remain highly cor related (6, 17). To the extent that pregnancy modifies exposure levels, the measurements will be inaccurate How ever, the relative ranking of women based on the measure ments would still have value. Although other populations can be studied to corroborate or refute our findings, the large study size and high PFOA exposure call for more complete evaluation of current and past residents o f the Mid-Ohio Valley. We are developing exposure reconstruction models that will enable us to in clude greater numbers of pregnancies to women enrolled in the C8 Health Project, enhance precision, and refine Serum PFCs and Pregnancy Outcome 9 exposure estimates even for the recent pregnancies included in this paper. We are also planning to link self-reported in formation on livebirth outcomes to birth records from Ohio and West Virginia, with vital records superior to maternal recall for birth weight and preterm birth. With the link to vital records, we will also gain additional relevant informa tion on preeclampsia and birth defects, despite the known limitations in vital records pertaining to these outcomes (36, 37). The ubiquitous exposure, suggestive evidence from toxicology, and sporadic indications in epidemiologic stud ies ol' possible adverse reproductive effects warrant contin ued and more refined evaluation of the association between PFC exposure and pregnancy outcome. ACKNOWLEDGMENTS Author affiliation: Department of Community and Pre ventive Medicine, Mount Sinai School of Medicine, New York, New York (Cheryl R. Stein, David A. Savitz, Marcelle Dougan). The entirety of the funding for the C8 Health Project, including cash payments to participants, was achieved pur suant to the settlement agreement in the case of Leach v E.J. DuPont de Nemours & Co., Civil Action No. 01-C-608. Specific funds for the project were administered as de scribed in the " Order Approving Final Settlement and Notice Plan and for Entry of Final Judgment" (publically available): a "Settlement Fund" was created with the agreed-upon $70 million and was administered by a named, court-approved health project administrator. Conflict of interest: none declared. REFERENCES 1. US Environmental Protection Agency. Petfluorooctunoic A d d (PFOA). Washington, DC: US Environmental Protection Agency; 2009. (http://www.epa.gov/oppt/pfoa/index.html). (Accessed January 21. 2009). 2. US Environmental Protection Agency. Per/luoroodanoic A d d (PFOA) and Fluorinated Tdomers. Washington, DC: US Environmental Protection Agency; 2009. (http://www.epa.gov/ oppi/pfoa/pubs/pfoainfo.html). (Accessed January 21, 2009). 3. Moude M . Martin JW, Letcher RJ, cl al. Biological monitoring o f polyfluoroalky) substances: a review. Environ S d Technol 2006:40(11):3463-3473. 4 Culufat A M . Kuklcnyik Z, Reidy JA, et al. Serum concentra tions o f 11 polylluoroalkyl compounds in the U.S. population; data from the National Health and Nutrition Examination Survey (NHANES). Environ S d Technol. 2007;4 1(7):22372242. 5. Calafut A M , Wong LY, Kuklenyik Z, el al. Polyfluoroalky) chemicals in the U.S. population: data from the National Health and Nutrition Examination Survey (NHANES) 2003 2004 and comparisons w ith NHANES 1999-2000. Environ Health Persped. 2007,115(11): 1596-1602. 6. Monroy R. Morrison K , Teo K, et al. Serum levels o f perfluoroalkyl compounds in human maternal and umbilical cord blood samples. Environ Res. 2008; 108( 1);56-62, 7. lnoue K, Okada F, Uo R, et al. Perfluorooctane sulfonate (PFOS) and related perfluorinated compounds in human ma ternal and cord blood samples: assessment of PFOS exposure in a susceptible population during pregnancy. Environ Health Persped. 2004;112(11); 1204 -1207. 8. Midasch 0 , Drexler H, Hart N, et al. Transplacental exposure of neonates to perfluorooctanesulfonate and perfluorooctanoate: a pilot study. Int Arch Occup Environ Health. 2007; 80(7):643-648. 9. Tbo L, Kannan K, Wong CM, et al. Perfluorinated compounds in human milk from Massachusetts, U.S.A. Environ S d Technol. 2008;42(8):3096-3101. 10. Tao L, Ma J, Kunisue T, et al. Perfluorinated compounds in human breast milk from several Asian countries, and in infant formula and dairy milk from the United States. Environ S d Technol. 2008;42(22):8597-8602. 11. Volkel W, Genzel-Boroviczeny O, Demmelmair H, et al. Perfluorooctane sulphonate (PFOS) and perfluDrooctanoic acid (PFOA) in human breast milk: results of a pilot study. Int J Hyg Environ Health. 2008;211(3-4):440-446. 12. Kuklenyik Z, Reich JA, Tully JS, et al. Automated solid-phase extraction and measurement of perfluorinated organic acids and amides in human senim and milk. Environ S d Technol. 2004;38( 13):3698-3704. 13. So MK, Yamashita N, Taniyasu S, et al. Health risks in infants associated with exposure to perfluorinated compounds in hu man breast milk from Zhoushan, China. Environ S d Technol. 2006;40(9):2924-2929. 14. Olsen GW, Burris JM. Ehresman DJ, et al. Half-life of serum elimination of perfluorooctanesulfonate, perfluorohexanesul- fonate. and perfluorooctanoate in retired fluorochemical pro duction workers. Environ Health Persped. 2007-.115(9): 1298-1305. 15. Lau C, Anitole K, Hodes C, et al. Perfiuoroalkyl acids: a re view of monitoring and toxicological findings. Toxicol Sd. 2007;99(2):366-394. 16. Apelberg BJ, Witter FR, Hcrbstman JB, et al. Cord serum concentrations of perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA) in relation to weight and size at birth. Environ Health Persped. 2007;115(11):1670-1676. 17. Fei C, McLaughlin JK, Throne RE, et al. Perfluorinated chem icals and fetal growth: a study within the Danish National Birth Cohort. Environ Health Persped. 2007;! 15(11): 1677-1682. 18. Fei C, McLaughlin JK, Tarone RE, et al. Fetal growth indi cators and perfluorinated chemicals: a study in the Danish National Birth Cohort. Am J Epidemiol. 2008;168(l):66-72. 19. Washino N, Saijo Y, Sasaki S, et al. Correlations between prenatal exposure to perfluorinated chemicals and reduced fetal growth. Environ Health Persped. 2009; 117(4):660-667. 20. Olsen GW, Butenhoff JL, Zobel LR. Perfiuoroalkyl chemicals and human fetal development: an epidemiologic review with clinical and toxicological perspectives, Reptod Toxicol. 2009; 27(3-4):212-230. 21. Grice MM, Alexander BH, Hoffbeck R, et al. Self-reported medical conditions in pcrfluorooctanesulfonyl fluoride manufacturing workers. J Occup Environ Med. 2007;49(7): 722-729. 22. Nolan LA, Nolan JM, Shofer FS. et al. The relationship be tween birth weight, gestational age and perfluorooctanoic acid (PFOA)-contarainated public drinking water. Reprod Toxicol. 2009;27(3-4):231-238. 23. Frisbce SJ, Brooks AP, Maher A, et al. The C8 Health Proj ect: design, methods, and participants. Environ Health Persped. Advance Access: July 13,2009. (DOI: 10.1289/ehp. 0800379). 10 Stein et al. 24. Emmett EA, Shofer FS, Zhang H, et al. Community exposure to perfluoroocianoale: relationships between serum concen trations and exposure sources. J Occup Environ Med. 2006; 48(8):759-770. 25. Steenland K, JinC, MacNeil J, etal. Predictors of PFOA levels in a community surrounding a chemical plant. Environ Health Perspeci. 2009;] 17(7): 10 8 3 -1088. 26. Butenhoff JL, Kennedy GL Jr, Frame SR, el al. The repro ductive toxicology of ammonium perfluorooctanoate (APFO) in the rat. Toxicology. 2004; 196(1-2):95-116. 27. I.uu C. BuienholT JL, Rogers JM. The developmental toxicity of perfluoroalkyl acids and their derivatives, Toxicol Appl Pharmacol. 2004;198(2):231-241. 28. Adegboye AR. Heiimann B. Accuracy and correlates of ma ternal recall of birthweight and gestational age. BJOG. 2008; 115(7):886-893. 29. Rice F. Lewis A, Harold G, et al. Agreement between maternal report and antenatal records for a range of pre and peri-natal factors: the influence of maternal and child characteristics. Early Hum Dev. 2007;83(8):497-504. 30. Tate AR, Dezateux C, Cole TJ, et al. Factors affecting a mother's recall of her baby's birth weight. Int J Epidemiol. 2005;34{3):688-695. 31. Gofin R, Neumark YD, Adler B. Birthweight recall by mothers of Israeli children. Public Health. 2000;! 14(3):161-163. 32. Seidman DS, Slater PE, Ever-Hadani P, et al. Accuracy of mothers' recall of birthweight and gestational age. BrJO bstet Gynaecol. 1987;94(8):731-735. 33. Sou SC, Chen WJ, Hsieh WS, et al. Severe obstetric compli cations and birth characteristics in preterm or term delivery were accurately recalled by mothers. J Clin Epidemiol. 2006; 59(4):429-435. 34. Klemmensen AK, Olsen SF, 0sterdal ML, et al. Validity of preeclampsia-related diagnoses recorded in a national hospital registry and in a postpartum interview of the women. Am J Epidemiol. 2007; 166(2): 117-124. 35. Greenland S. Basic methods for sensitivity analysis of biases. Int J Epidemiol. 1996;25(6): 1107-1116. 36. Lydon-Rochelle MT, Holt VL, Cardenas V, et al. The re porting of pre-existing maternal medical conditions and complications of pregnancy on birth certificates and in hos pital discharge data. Am J Obstet Gynecol. 2005:193(1): 125-134. 37. Watkins ML, Edmonds L, McCleam A, et al. The surveillance of birth defects: the usefulness of the revised US standard birth certificate. Am J Public Health. 1996;86(5):731-734.