Document EqBBrbdXmVp9RneD8aKb1oknb

3M Company EPI-0013 Page I of 36 ( FINAL REPORT Epidemiology Medical Department 3M Company St. Paul. MN 55144 Date: February 25, 2002 Title: Identification of Fluorochemicals in Human Sera. I. American Red Cross Adult Blood Donors Study Start Date: September 29, 2000 Protocol Number EPI-0013 Principal Investigator:. G-earyW. Olsen, D.V.M., Ph.D) 3M Co-investigators: Jean M. Burris, M.P.H., R.NJ James K. Lundberg, Ph.D.2 Kristen I. Hansen, Ph.D.2 Jeffrey H. Mandel, M.DJ Larry R. Zobel, M.D.I ( . .q-. Study Sponsor:. Corporate Occupational Medicine, Medical Department, 3M Company, 220-3W-05, St. Paul, MN 55144 1. Medical Department, 3M Company, St. Paul, MN 55144 2. Environmental Laboratory,3M Company, St Paul, MN 55144 ABSTRACT f_''" 3M Company EPI-0013 Page 2 of 36 Through cooperation with six American Red Cross blood banks, 645 serum samples from adult donors (ages 20-69, equally represented of both sexes) were obtained for fluorochemical analyses. Blood bank locations were Los Angeles (CA), Portland (OR), Minneapolis-St. Paul _, Charlotte (NC), Hagerstown (MD) and Boston (MA). Samples were void of personal identifiers. Age, gender and location were the only known demographic factors. Sera samples were extracted and quantitatively analyzed for seven fluorochemicals using high-pressure liquid chromato_aphy/electrospray tandem mass spectrometry and evaluated versus an extracted curve from a human plasma matrix. The seven fluorochemicals were perfluorooctanesulfonate (PFOS, CsFtvSO3"); N-J_hyl perfluorooctanesulfonamidoacetate (PFOSAA, C_FI:SO2N(CH2CH3)CH2COO'); Ni methyl perfluorooctanesulfonarnidoacetate (M.570, C_FITSO.,N(CH3)CH.,COO"); perfluorooctanesulfonamidoacetate (M556, C,FtTSO2N(cH)CH2COO-); perfluorooctanesulfonylamide (PFOSA, C_FtTSO2NH2); perfluorooctanoate (PFOA, C7FI3COO'); and perfluorohexanesulfonate (PFHS, C_13SO3"). Overall, the geometric mean measured concentration of PFOS was 34.9 ppb (95% CI 33.3-36.5). The measured PFOS concentration ranged from less than the lower limit of quantitation (I.J..OQ) of 4.1 ppb to 1656.0 ppb. The geometric mean for PFOS was sigT1ificantly higher among males (37.8 ppb; 95% CI 35.5-40.3) than females (31.3 ppb; 95% C130.0-34.3). No significant difference was observed with age. Charlotte (NC) had the highest geometric mean serum PFOS concentration (51.5 ppb) and Boston (MA) the lowest (29.5 ppb). Bootstrap analyses were used to calculate a 95% tolerance limit for f' 3M Co_:,P_ge3 o_ _'36 ( PFOS of 88.5 ppb with an upper 95% confidence limit of 100.0 ppb. Additional geometric mean and tolerance limit data are reported for PFOA, PFI=IS,PFOSAA and M570. The geometric mean and 95% tolerance limits of these fluorochemicals were, cn average, an order of magnitude (or more) lower than PFOS. PFOS and PFOA were highly correlated (r = .63). PFOS had lower correlations with PFOSAA (r = .43), PFH_ (r = .38) and M_570(r = .20). The number of samples with measured PFOSA and M.556 concentrations below the LIOQ prohibited meaningful statistical analyses for these compounds. The findings from this analysis of serum PFOS concentrations are consistent with those previously reported. The human data, to date, suggests the approximate average serum concentration in non-occupational adult populations may be 30 to 40 ppb with 95% of a population's serum PFOS concentrations below 100 ppb. Since serum PFOS ( concentrations likely reflect cumulative human expo.s.-,ure, this information will be useful for risk characterization. ( <.... ,/ 3M Company EP[-0013 Page4 of36 _TRODUCTION InMay, 2000the3M Company (3M')announcedthatitwouldvoluntaricleyase manufacturipnegrfluorooctanesul(fPoOnSyFl,-CsFITSO2Fr'e)latemdateriaalfstetrhe compound,perfluorooctanesul(fPoFnOaSt,eC8F17SO3")w,as foundtobepervasivaend persisteinnthumanpopulationwsi,ldlifmea,rinemammals andpiscivorobuisrd(s3M Company 2000;Hansenetal2001;GiesyandKannan2001;Kannanetal2001a;2001b). POSF, producebdy anelectrochemifclaulorinatipornocessi,susedasthebasic buildinbglocktocreatueniquechemistritehsrougthhesulfonyflluorimdoeietyusing conventionhayldrocarborneactionFso.rexampleP,OSF canbereactewdithmethylor ethyalminestoproduceitheNr-ethyolrN-methylperfluorooctanesulfonaAmtitdhei.s staget,heseintermediactaensbeusedtomake amideso,xazolidinonseisl,anes, carboxylataensdalkoxylataesscommerciaplroductAsl.so,thesientermediactaensbe subsequentrleyactewdithethylencearbonatteoformeitheNr-ethyolrN-methyl perfluorooctanesulfonamidowchtihcahncoalnbe usedtomake adipatepsh,osphate esterfsa,ttaycidesterusr,ethanceo-polymerasndacrylataesscommercializperdoducts. Dependinguponthespeciffiucnctiondaelrivatizaotritohnedegreeofpolymerization, suchPOSF-basedproductmsay degradeormetabolizteo,anundetermindeedgreet,o PFOS,a stablaendpersistentd-produtchtathasthepotentitaolbioaccumulatWeh.ile nota majorcommerciaplroductP,FOS itsehlafsbeenusedinsomeproductisn,cluding fire fighting foams. The mechanisms and pathways leading to the presence of PFOS in human blood arc not well characterized but likely involve environmental exposure to PFOS or its precursor molecules and residual levels of PFOS or PFOS precursors in industrial and commercial 3M Cotton:, /- products. PFOS has been detected at low parts per billion (ppb) concentrations in the general population (Hansen et al 2001; 3M Company 2000) although the scope of these investigations has been limited. Using high pressure liquid chromatography/electrospm)7 tandem mass spectrometry, Hansen et al (2001) detected an average PFOS concentrator: of 28.4 ppb (SD 13.6; range 6,7-81,5) in 65 commercial individual human sera samples. An analysis of pooled blood samples (n -- 3 to 6 pooled samples per location with 5 to I(C) donors per pooled sample) from 18 blood banks in the United States resulted in a mean measured PFOS serum concentration of 30 ppb with a range from 9 to 56 ppb (3M Company, 2000). Serum PFOS concentrations among production employees working in POSF-related processes were approximately 2 parts per million (ppm) depending on work activity (range O.l to 12 ppm) (Olsen et al 1999). The purpose of this study was to better characterize the distribution of seven ( fluorochemicals, including PFOS and some of its precu_0rs in an adult populatio..-n. by analyzing sera samples obtained from donors at six American Red Cross blood banks. A_.._. assessment of the serum fluorochemical distribution was performed in relation to three demographic attributes (age, gender and location) of the anonymous blood donors. METHODS Fluorochemicals The seven analytes detected and quantified in this study were: PFOS; N-ethyl perfluorooctanesulfonamidoacetate (PFOSAA, CsFIvSO2N(CH2CH3)CH2COO'); N- methyl perfluorooctanesulfonamidoacetate (M570, CsP)vSOEN(CH3)CH_,COO'): perfluorooctanesulfonamido acetate (M556. CsFIrSO_,N(CH)CHzCOO'); / 3M Company EP[-0013 Page6 of36 (< perfluorooctanesulfonylamide (PFOSA, C_FITSO2NH2); perfluorooctanoate (PFOA, CTFt3COO'); and perfluorohexanesulfonate (PFHS, C6F)3SO3"). PFOSAA is an oxidation product of N-ethyl peHluorooctanesulfonamidoethanol (N-EtFOSE) and is a residual in N-EtFOSE-related chemistry which was primarily used in paper and packa_ng protectant applications. M570 is an oxidation product of N- methyl perfluorooctanesulfonamidoethanol (N=MeFOSE) and is a residual of N- MeFOSE-related chemistry which was used primarily in surface treatment applications (e.g., carpets, textiles). Therefore, PFOSAA and MST0 can be considered markers of consumer=related exposure. Both PFOSAA and MS70 can metabolize to MS56 and FFOSA which, in turn can subsequently metabolize to PFOS. Unlike PFOSAA and M570, M556, PFOSA and PFOS are not specific to any one consumer application. Unlike the other analytes, PFOA and PFHS are not precursors, metabolites or residuals of ( PFOS. PFOA can be a residual by-product of the production of the POSF-related manufacturing electrochemical fluorination process and was produced by 3M tO be an emulsifier in a variety of industrial applications (e.g., ammonium salt) (Olsen et al 2000). PFOA can also be an oxidation product or metabolite of the widely used telomer-based fluorochemicals manufactured by other companies. PFHS, the sulfonate form of perfluorohexane sulfonyl fluoride (PI-ISF), is a residual by,product of POSF-related production. 3M produced the PHSF as a building block compound incorporated in fire fighting foams and specific post-market carpet treatment applications. /Y (..... _, f" Sample Collection 3M Corm:r,n:y EPl-0tl13 Page 7 of 36 Through cooperation with six American Red Cross blood banks, 645 serum samples from adult donors (ages 20 - 69, equally represented of both sexes) were obtained for analysis. [Note: This final report supercedes the June 25, 2001 3M interimreport whick indicated 652 samples. Seven samples were not included in the final reportbecause the subjects were determined to be ineligible by age _ 70). Additional separate3M sponsored studies were designed to examine the distribution of serum samples among elderly adults and children.] The six American Red Cross blood banks represented donors from the following areas: Los Angeles, CA; Portland, OR; Minneapolis-St. PauL MN; Charlotte, NC; Hagerstown, MD and Boston, MA. Samples were void of persona/. identifiers. The"only known demographic factors were age, gender and location. Each blood bank was requested to provide approximately lO samples per I0 year age intcrv_ (20-29, 30-39, 4049, 50-59 and 60-69) for each sex. Fluorochemical Analysis Northwest Bioanalytical (Salt Lake City, Utah) analyzed the serum for the targe-. fluorocbemicals using t_chniques similar to those described by Hansen et al (2001). Details of the specific analytical procedures are presented elsewhere (NWB 2002). Briefly, the analytical method consisted of a liquid:liquid extraction procedure followed by evaporation and reonsdtudon of the extract residue with 20 mM ammonium acetate in water:.20 mM ammonium acetate in methanol (30:70, v/v). The samples were analyzed by high pressure liquid chromatoo_'aphy/tandem mass spectrometry. Quantitation of the target analytes in serum samples was performed by comparing the (<..... ( chromatographic 3M Company EPI-0013 Fagc 8 of 36 peak areas for each compound to those generated in a series of extracted calibration standards prepared from control Chinese plasma. The samples were injected in a systematic order. Evaluation of quality control samples injected dunng each analytical run indicated that the reported quantitative results may have varied, on average, up to 26 percent using human plasma calibration curves for all analytes except PFOSA which may have varied on average up to 43 percent. Also presented in this report is a calculated total organic fluorine CTOF) index. TOF was the percent of each of the seven fluorochemicals' molecular weight that was attributed to organic fluorine [PFOS (64.7%); PFHS (61.9%); PFOA (69.0%); PFOSAA (55.3%); PFOSA (64.7%); M570 (56.6%) and M556 (58. 1%)] multiplied by the ppb measured for e_/ch fiuorochemical and then summed across all seven fluorochemicals. Data Analysis Measures of central tendency applicable to log normally distributed data (median, geometric mean) were used for descriptive analyses. In those instances where a sample was measured below the lower limit of quantitation (LLOQ), the midpoint between zero and the LLOQ was used for calculation of the geometric mean. An assessment of this midpoint assumption and how it affected the calculation of the geometric mean was performed using the l0 ta and 90ta percentile values between zero and the LLOQ for those values <LLOQ. In order to minimize parametric assumptions in the estimation of extreme percentiles of the population, the bootstrap method of Efron (1993) was used to generate confidence intervals around the empirical percentiles for serum concentrations. In this EPI-.,):_ Page 9 of_6 - method, a large number of replicated estimates of the percentile are generated from full- size samples of the original observations drawn with replacement. The distribution of _ deviations of replicates from the original-sample estimate mimics the underlying sampling distribution for the estimate. Bias-corrected, accelerated percentiles were use_ to minimize residual bias. The bias correction factor is derived by comparing empirical percentiles to bootstrap percentiles and acceleration is accomplished by partial jackknifing. Twenty-four samples were split and analyzed to provide an estimate of the reliabiloiftyheanalysecsonductedT.heanalyticlaalboratowrays blintdotheidentity. ofthesseplistamplesT.heseanalysewserel_rformecdoncurrentwliythallother analyseosfthestudytominimize xperimentearlrorF.ivesplistampleswereanalyzed fromCharlottLeo,sAngelesH,agerstowanndPortlanadndfoursplistamplefsrom ( Boston. Inadvertently, no reliability analyses were performed on the Minnesota sample_. RESULTS The results for the reliability analysis for PFOS, PFOA and PFHS are displayed :m Figure 1. Only six split samples for PFOSAA and seven split samples for M570 had values that were above the LLOQ. None of the PFOSA and M.556 split samples were above Me LLOQ. Therefore, only the reliability results for PFOS, PFOA and PFHS are displayed in Figure 1. ,There was a strong correlation between the split samples (r = .9' for each of these three fluorochemicals. It should be noted that 13 of the split sample analyses for PFHS had the identical LLOQ (2.1 ppb). This is represented in the graph ,_ the single point near the abscissa (0,0) on the identity (In y = In x) line. 3M Company EPI-O013 Page I0 of 36 . Provided in Table 1 is the distribution of the donor subjects by 10 year age intervals, gender and location. Altogether there were 332 male donors and 313 female donors. As could be expected with the age stratification design used for sample collection, the study subjects' mean ages were comparable by gender: 44.6 years for males and 43.9 years for females. The measured concentrations of PFOSA and M556 were predominantly below the LLOQ. For PFOSA, there was one subject with a value above (2.1 ppb) the LLOQ, 196 subjects 1I.OQ (1.0 ppb) and 448 subjects II.OQ (1.4 ppb). The different LLOQ values were determined on different analytical runs. For M556, twelve subjec_ were above the LLOQ. Their M556 values ranged from 3.6 to 12.9 ppb. There were 145 subjects with M_556values <LLOQ (2.5 ppb) and 488 subjects with M556 values J.,LOQ (3.2 ppb). Because of the few subjects whose serum concentrations exceeded the LLOQ, !l' statistical analyses are not presented for PFOSA and M556. Although PFOSA and M556 are not presented in the subsequent analyses, they were included in the calculation of the TOF index. For those measured concentrations of PFOSA and M556 < LLOQ, the midpoint between zero and the LLOQ was used in the calculation of the TOF index. The frequency distributions of the five remaining fluoroehemicals, PFOS, PFOA, PFI-IS, PFOSAA and M570, are displayed in Figure 2. Although the graphs are suggestive of log normal distributions, only the PFOS distribution met such criteria based on the Shapiro-Wilk test. This lack of log normality is due to the geater percentage of subjects with values LLOQ for PFOA, PFI-IS,PFOSAA and M570. The range, interquartile range, number of samples < LLOQ. raw cumulative 90_ percentile, median, geometric mean and 95% confidence interval of the geometric mean 3M Cor_'-_n y I:PI-}_)13 Page 11 ._,'36 (_.. for PFOS, PFOA, PFHS, PFOSAA and M570 are provided in Table 2 for all subjects, males only and females only. Overall, the geometric mean levels of PFOS was 34.9 pD_ (95% CI 33.3-36.5). The range of PFOS values was < LLOQ (4.3 ppb) to 1656.0 ppb. Male subjects had si_ificantly (p < .05) higher geometric means for PFOS than female subjects [male geometric mean = 37.8 ppb (95% CI 35.5-40.3) vs female geometric me = 32.1 ppb (9:5%C130.0-34.3)]. Males also had significantly higher serum levels of PFOA and PFI-IScompared to females although the mean levels for both sexes were approximately one order of magnitude lower than that of PFOS. It should be noted that the overall geometric mean for the calculated TOF index was 31.7 ppb (95% C130.4 - 33.0) (data not shown). The calculated TOF index ranged from 5.'7ppb to 1083.2 ppb. Provided in Figure 3 is a graphical distribution (natural log scale) of the five fluorochemicals by l0 year age intervals stratified by gender. The box covers the ( interquartile range of the natural log distribution. The circle within the box is the mean. The whiskers extend to the last observation within I.S times the interquartile range. Th_ dots with lines throu_ them represent observations outside the 1.5 times interquartile range. As shown in Fibre 3, age was not an important predictor of adult serum fluorochemieal concentrations. In those instances where there were many outliers (e._.. M570 concentrations in males aged 40-49 and 60-69), this was the result of a large percentage of values <LLOQ that were within the 1.5 x interquartile range. As discussed previously in the Methods, the geometric mean data were calculac_._l under the assumption that for individual serum fluorochemical values <IJ OQ the midpoint between zero and the LLOQ was assi_ed. For PFOS, only one subject had a_ value <LLOQ: thus this assumption did not affect its calculation of the geometric mean. 3M Company EPg-OGt 3 Page IZof 36 ( However, considerably more subjects had LLOQs for PFOA, PFHS, PFOSAA and M570 (see Table 2). If these values were assumed to be 10% or 90% of this range between zero and the LLOQ the respective range of the geometric means (95% confidence interval in parenthesis) became: PFOA 4.0 ppb (3.7-4.1) to 4.8 ppb (4.6-5.0); PFHS 0.9 ppb (08-1.0) to 2.5 ppb (2.4-2.6); PFOSAA 0.8 ppb (0.7-0.9) to 2.8 ppb (2.7-3.0) and M570 0.5 ppb (0.5-0.6) to 1.9 ppb (1.8-2.0). These geometric mean values were not substantially different than those calculated using the midpoint between zero and the <LLOQ as presented in Table 2." Consequently, the midpoint between zero and the LLOQ was used for the analyses. Presented in Table 3 are the range, interquartile range and medians for the six locations combined across age and gender for the five fluorochemicals. ;_.graphical presentation of these data (natural log scale) is presented in Figure 4. Interpretation of ( the graphs is comparable to those discussed above for Figure 3. Provided in Table 4 are the results from a bootstrap analysis which calculated mean serum fluorochemical values for each of the six locations adjusted for 10 year age intervals, gender and their interaction terms. The highest mean value for PFOS was Charlotte (39.0 ppb) with the lowest being Boston (29.0 ppb). Los Angeles, Minneapolis-St. Paul and Hagerstown had comparable mean PFOS levels of approximately 35.0 ppb with Portland slightly lower (32.8 ppb). The range of means for the other fluorochemical analytes was narrow and thus difficult to distinglaish any substantial differences by location. Because PFOS is the primary contributor to the calculated TOF index, the bootstrap analysis findings for TOF mirrored those of PFOS. 33,! Cotr4:_,, _. BPl..O0!L3 Page 13o/_ ('" Scatter plots (log scale) between the five fluorochemicals are displayed in Figur_ 5. PFOS and PFOA were highly correlated (r = .63). PFOS had a lower correlation witm PFOSAA (r = .42) and lower yet with M570 (r = .20). The correlation between PFOSA.._ and M570 was weak (r = .12). The remaining scatter plots display the correlation between PFOS and PFI-IS(r = 0.38) and PFOA and PFHS (r = 0.32). Both PFOSAA and M570, adjusted for age, gender and their interactionl were significant predictors of PFOS in a multivariable model albeit PFOSAA was the stronger of the two independent variables (Table 5). Nevertheless, approximately eighty percent of the variation of PFOS was left unexplained. Age and gender were not significant predictors in models that examined the significant association between PFOS and PFOA (Table 6) or PFI-IS (Table 7). None of themodels in Tables 5 through 7 had lack of fit F ratios that were statistically significant (p < .05). Presented in Table 8 are the results from bootstrap analyses conducted to provide .. mean concentrations of several tolerance limits. The tolerance limits represent the limit of each fluorochemical wi_in which the stated proportion of the population is expected to be found. Presented are the mean values of the five serum fluorochemicals and TOF for the 90'_, 95" and 99'_ percent tolerance limits along with the upper limit (bound) from the 95% confidence interval. For example, the mean of the 95% tolerance limit for PFOS was 88.5 ppb with an upper 95% confidence limit of 100.0 ppb. At the lowest tolerance limit analyzed (90%), the mean for PFOS was 70.7 ppb with art upper 95% confidence limit of 74.3 ppb. At the higlaest tolerance limit analyzed (99%), the mean was 157.3 ppi_ with an upper 95% confidence limit of 207.0 ppb. For other fluorochemicals analyzed, the mean of the 95% tolerance Iimit for PFOA was 12.1 ppb with an upper 95% 3M Company EPbOO[3 Page 14 of 36 / confidence limit of 13.6 ppb. For PFHS, the mean of the 95% tolerance limit was 9.5 ppb with an upper 95% confidence limit, of 10.8 ppb. The mean of the 95% tolerance limit for PFOSAA was 7.6 ppb with an upper 95% confidence limit of 8.5 ppb. For M570, the mean was 5.0 ppb for a 95% tolerance limit with an upper 95% confidence limit of 5.4 ppb. Finally, for the calculated index of TOF, the mean was 75.1 ppb for the 95% tolerance limit with an upper 95% confidence limit of 80.9 ppb. DISCUSSION The findings from this analysis of serum PFOS concentrations in 645 adult donors am consistent with previous human data. Previous measurements of human serum samples obtaindd in the United States showed mean PFOS concentrations of 30 ppb in 18 pooled blood banks, 44 ppb from a pooled commercial sample of 500 donors, 33 pph _- from a different pooled commercial sample of 200 donors and 28.ppb in 65 commercial individual human sera samples (3M Company 2000; Hansen et al 2001). These findings were also comparable to a limited number of European samples which found mean serum PFOS concentrations at 17 ppb in 5 pooled samples from a Belgium blood bank, 53 ppb in 6 pooled samples from the Netherlands and 37 ppb from 6 pooled blood samples from Germany and 39 individual Swedes whose serum PFOS ranged between <rx OQ32 ppb to 85 ppb (3M Company, 2000). The mean calculated TOF index of 31.7 ppb in the present study was also consistent with the low ppb total organic fluorine measurements of general population samples that have been reported since the 1960's (Tares 1968; Tares et al 1976; Singer and Ophaug 1979; Belisle 1981). ..= 3,%C1orno=._ny EPr_.OOI3 Puge15c_;_.36 : Unique to the present study was its large individual sample size which facilitate_l the characterization of the serum fluorochemical distribution. This included the calculation of tolerance limits and their upper bounds. The highest serum PFOS measurement (confirmed by re-assay of the sample) was 1656.0 ppb. Because donor samples were anonymous, it is not possible to determine anything about this individuai besides gender (male), age (67 years) and location (Portland). This PFOS sample approximated the average serum PFOS levels observed for POSF-related production workers (Olsen et al 1999). The next highest donor level for PFOS was considerably lower at 329 ppb (also a male, age 62 from the Portland area) and the subsequent next eight highest serum PFOS values (range 139 ppb to 226 ppb) were measured in four females and foul"males representing Charlotte (n = 4), I-Iagerstown (n = 2), Los Angele_ (n "--1) and Minneapolis-St. Paul (n = 1). Also distinctive to this effort was the study's capability to examine potential associations between PFOS and age (no association) an_. ..=, gender (males slightly higher than females). Our findings showed a strong correlation between PFOS and PFOA. Whereas PFOS has been routinely measured in human populations, wildlife, marine mammals amd piscivorous birds (Giesy and Kannan 2001; Kannan et al 20Ola; 2001b; Hansen et al 2001), serum PFOA concentrations, to date, have been consistently quantified (i.e., measured above the LLOQs) primarily in humans. This association is of significant interest because PFOAcannot convert directly from PFOS (or vice versa). Whether tha__association is due to the presence of PFOA as a by-product in POSF-related production :or to other non-related environmental exposures or consumer products from other manufacturers (e.g.. higher chain telomers) remains to be answered. Another ... (".... \ e< ( _M Compaay EPI-0013 Page16of 36 unanswered question is whether perfluoroctanesulfonamides can metabolize in humans to PFOA. Any of these explanations coupled with the suspected long serum half-lives in humans for PFOS (8.7 years (SD = 6.1) and PFOA (4.4 years (SD = 3.5) as reported by Burris et al (2002), could explain the strong correlation between PFOS and PFOA. PFOS was also correlated with .two fluorochemicals, PFOSAA and M570, known to be analytes from exposure to consumer products involving paper/packaging and carpeVtextile protectants, respectively. Overall, the data, to date, indicate that PFOS bioaccumulation in animals may be primarily through environmental sources whereas both environmental and product exposures likely contribute to serum PFOS concentrations in humans (Giesy and Kannan 2001; Kannan et al 2001a; 2001b). As with any interpretation of data obtained from a study population, questions arise regarding the representativeness and ability to generalize the data collected. Clearly, American Red Cross blood donors are a self-selected group from the United States population (Leibrecht et al 1976; Oswalt 1977; Burnett 1982; Allen and Butler 1993; Andaleeb and Basu 1995: Chiavette et al 2000). Motivations to donate have included altruism, desire for personal credit, low self-esteem, social pressure and testseeking behavior. Motivations not to donate have included fear, medical excuses, apathy and inconvenience. Donors tend to have a greater trust in institutions, more interest in their personal health and higher risk-taking behavior. Demo_aphically, white males and older individuals have historically constituted a larger percentage of the donor pool. Donors also tend to be more educated, married and have more children than nondonors. In the present study, donors were not asked whether a sample of their blood donation could be used for fluorochemical assay, nor was there any linkage between the sample 3M ConT_mv _ 17 c_."36 -- collected and personal identifiers. No information was obtained about past exposure histories to POSF-related chemistries and materials (or the other fluorochemicals that were analyzed). Therefore, we believe the selection process of donors used in this stuc.,:, resulted in a reasonable rgpresentation of the overall blood donor population that was providing blood at the time when those donors were sampled. We are unaware of any database that can be considered generalizable to the diverse United States general adui:population without measures of random and systematic bias incorporated in the data analysis. Given the consistency of the data analyzed, to date, we hypothesize that the average serum PFOS concentrations in non-occupational adult populations likely range_ between 30 to 40 ppb with 95% of a population's serum PFOS concentrations below 10_ ppb. Understanding these serum PFOS levels in human populations will be useful for ( risk characterization since serum PFOS likely reflects cumulative human exposure (3M .... Company 2000). Currently available data (unpublished reports to U.S. EPA:Docket No. FYI-0500-01378) suggest that the serum concentrations observed in humans are substantially less than those required to cause adverse effects in laboratory animals (3M, Company 2000). (...... 3M Company EPI-O013 PagsI$ of36 ACKNOWLEDGm', r=NTS We wishto acknowledgemany cone,'ibutorsto this 3M final report. AmericanRedCross . blood donorcollectionwascoordinatedunderthe guidanceof Dr. John_filler (North Central Blood Services) with regional assistance from the following: Drs. John Armitage (Carolinas Region) Ross Herron and George Garratty (Southern California Region), Zahra Medhdizadehkashi (Pacific Northwest Regional), John Nobiletti (Greater Alleghenies Region) and Mary O'Neill (New England Region). Laboratory analysis of the seven fluorochemicals was provided by a dedicated team at Northwest Bioanalytical which included Ann Hoffman, Connie Sakashita, Patrick Bennett, Dr. Rodger Foltz, Suzanne Newman, Toni Peacock and Emily Yardimici. Biostatistical assistance was provided by Dr. Tim Church of the University of Minnesota. - /r l ( REFER.ENCES 3M Con's:my EPI-:_O 13 Page 19 o_E36 3M Company (2000). SIDS Initial Assessment Report Perfluorooctane Sulfonic Acid and its Salts. St. Paul:3M Company, September 20, 2000. Allen J, Butler DD (1993). Assessing the effects of donor knowledge and perceived risk on intentions to donate blood. J Health Care Market I3:26-33. Andaleeb SS, Basu AK (1995). Explaining blood donation: the trust factor. J Health Care Market 15:42-48. Belisle J (1981). Organic fluorine in human serum: natural versus industrial sources. Science 212:1509-1510. Bumett JJ (1982). Examining the profiles of the donor and nondonor through a multipi= discriminant approach. Transfusion 22:136-142. Burris JM, Lundberg JK, Olsen GW, Simpson C, Mandel JH (2002). Interim Report: Determination of serum half-lives of several fluorochemicals. St. Paul:3M Company, January 11, 2002. .... Chiavetta J, Ennis M, Gula CA, Baker AD, Chambers TL (2000). Test-seeking as (' motivation in volunteer blood donors. Transfusion Med Rev 14:205-215. Efron B, Tibshiarani RJ. An Introduction to the Bootstrap. In: Cox DR, Hinkley DV. Reid N, Rubin DB, Silvernan BW, eds. Monographs on Statistics and Applied Probability. Vol 57 New York:Chapman H Hall. Giesy JP, Kannan K (2001). Global distribution of perfluorooetane sulfonate in wildlife:. Environ Sci Technol 35(7): 1339-1342. Hansen KJ, Clemen LA, Ellefson ME, Johnson HO (2001a). Compound-specific quantitative characterization of organic fluorochemicals in biolo_cal matrices. Envirorr Sci Technol 35:766-770. Kannan K, Koistenen J', Beckmen K, Evans T, Gorzelany J'F, Hansen K.I, Jones PD, I-Ie/lie E, Nyman M, Giesy J1a (2001 b). Accumulation of perfluorooctane sulfonate in marine mammals. Environ Sci Technol 35(8): 1593-1598. Kannan K, Franson JC, Bowerman WW, Hansen KJ, J'ones PD, Giesy JP (2001). Perfluorooctane Sulfonate in fish-eating water birds including bald ea_es and albatrosses. Environ Sci Technol 35(15):3065-3070. 3M Compaay EPI-0013 Page 20 of 36 _' ...... Leibrecht BC, Hogan JM, Luz GA, Tobias KI (1976). Donor and nondonor motivations. Transfusions 16:182-189. Northwest Bioanalytical (NWB, 2002). Quantitative determination of PFOS and related compounds in human serum by LC/lVIS/MSJanuary 9, 2002. Olsen GW, Bun-is JM, Mandel JH, Zobel LR (1999). Serum perfluorooctane suIfonate and hepatic and lipid clinical chemistry tests in fluorochemical production employees. JOEM 41:799-806. Olsen GW, Bun'is J'M,Burlew MM, Mandel JH (2000). Plasma cholecystokinin and hepatic enzymes, cholesterol and lipoproteins in ammonium pedluorooctanoate production workers. Drug Chem Toxicol 23:603-620. Oswalt RM (1977). A review of blood donor motivation and recruitment. Transfusion, 17:123-135. Singer L, Fhaug RH (1979). Concentrations of ionic, total, and bound fluoride in plasma. Clin Chem 25:523-525. Tares D (1968). Evidence that there are two forms of fluoride in humanserum. Nature 217:1050-1051. f ' Taves D, Guy W, Brey W (1976). Organic fluorocarbons in human plasma: Prevalence ,, and characterization. In: Filler R, e_. Biochemistry Involving Carbon-Huorine Bonds. Washin_on DC:American Chemical Society, pp 117-134. Table 1 Distribution of American Red Cross Blood Donor Subjects by Age, Gender and Location 3M Company EPI-O013 Page 21 or 36 Males Localion 20-29 30-39 40-49 50-59 Los Angeles Boston 14 II 12 "11 15 13 !I 11 Minncapt_lis- I0 I0 I0 10 St. Paul Charlotte Portland 10 10 10 10 10 II 10 15 Hagerstown 10 l0 l0 10 66 63 66 69 60-69 Subtotal 10 63 12 57 I0 50 7 47 10 56 19 59 68 332 Females 20-29 30-39 40-49 50-59 60-69 Subtotal 16 12 12 12 10 62 i0 iI 10 iI 10 52 10 10 10 10 10 50 10 11 10 10 10 11 10 iI 8 49 9 51 10 10 l0 10 9 49 66 65 62 64 56 313 Overall Total 125 109 I(X) 96 107 108 645 ,, " Table 2 3MCompany EPI-O013 Page22 of36 Measures of Central Tendency of Serum Fluorochemicals for American Red Cross Blood Donors (H = 645) by Gender ,e PFOS PFOA PF! IS PFOSAA M570 All (N r. 645) Range < LOQ (4.3)- 1656.0 < LOQ (1.9)- 52.3 < LOQ (!.4)- 66.3 < LOQ (1.6)-60.1 < LOQ (I.0)- 16.4 QI -Q3 24.7-48.5 3.4-6.6 < LOQ (2'.1)- 3.4 < LOQ (2.8)- 3.4 < LOQ (1.8)- 2.2 : < LOQ (Number) < 4.3 (!) < 1.9 (2) < 1.4 (72) < 1.6 (101) < 1.0 (63) Cumulative 90% 70.7 < 2.1 (48) 9.4 < 2. I (235) 6.3 < 2.8 (27 I) 5.2 < i.8 (326) 3.8 Median Geomelric Mean 95% C.I. Geometric Mean 35.8 34.9 33.3 - 36.5 4.7 4.6 4.3 - 4.8 1.5 1.9 1.8 - 2.0 < LOQ (2.8) 2.0 1.9 - 2.1 < LOQ (1.8) 1.3 1.3 - 1.4 Males (N = 332) Range Qi -Q3 < LOQ (Number) Cumulative 90% Median Geometric Mean 95% C.I. Geometric Mean <LOQ (4.3) - 1656.0 28.3-49.7 < LOQ (1.9) - 29.0 3.6-7.0 < 2.1 (19) 72.6 37.4 37.8 35.5 - 40.3 10.1 4.9 4.9 4.6 - 5.3 < LOQ (l.4) - 66.3 < LOQ (2.1)- 3.8 < 1.4 (30) <2.1 (104) 7.9 2. I 2.2 2.0 - 2.4 < LOQ (l.6) - 60.1 < LOQ (2.8)- 3.3 < 1.6 (58) <2.8 (146) 4.7 < LOQ (2.8) 1.9 1.8 - 2.1 < LOQ (l.0) - 16.4 < LOQ (1.8) - 2.2 < 1.0 (36) < 1.8 (163) 3.5 < LOQ (1.8) !.3 1.2 - 1.4 Females (N = 313) Range QI -Q3 < LOQ (Number) Cumulative 90% Median Geometric Mean 95% C.I. Geomelric Mean 3M Company EPI-0013 P.',gt:23 of 36 6.0- 226,0 22.0-45.8 69.7 31.3 32,1 30.0 - 34.3 < LOQ (2.1)- 52.3 3.1 -6.2 < 1.9 (2) <2.1 (29) 8.4 4.4 4.2 3.9 - 4.5 < LOQ (1.4)- 15.3 < LOQ (2.1)-2.8 < 1.4 (42) <2.1 (131) 5.0 t < LOQ (2.1) 1.6 1.5 - 1.8 < LOQ (1.6)- 27.6 < LOQ (2.8)- 3.6 < i.6 (43) <2,8 (125) 6.1 < LOQ (2.8) 2.1 2.0 - 2.3 < LOQ (I.0)- 10.6 < LOQ (!,8)- 2,2 < 1.0 (27) < 1.8 (163) 4.0 < LOQ 1.8) 1.3 1.2 - 1.4 3M Company EPI-flO! 3 Page 24 of'36 Table 3 Measuresor Cenlral Tendency of Fluorochemicalsby the Six American Red CrossBlood Bank Locations PFOS PFOA PFHS PFOSAA M570 Los Ant_eles Range Q I - Q3 Cumulative 90% 6.6-205.0 29.5 - 53.7 70. I <LOQ (2.1) - 34.1!, <LOQ(2.1)- 12.4 <LOQ(2.8)-27.6 2.9 - 6.6 <LOQ (2.1) - 3.0 <LOQ (2.8) - 4.2 9.2 5.8 6.1 <LOQ(I.8)- 16.4 <LOQ (1.8) - 2.2 4.8 Medi an 42.2 Geometric Mean 40.4 95% C.I. 37.0 - 44.0 : Geometric Mean 4.6 4.1 3.6 - 4.7 <LOQ (2.1) 1.9 1.7 - 2.1 3.1 2.6 2.4 - 3.0 <LOQ ( 1.8) 1.4 1.2 - 1.6 Boston Range Q I - Q3 , Cumulative 90% <LOQ (4.3) - 87.2 20.8 - 39.1 48.7 Median Geometric Mean 95% C.i. Geometric Mean 29.5 28.0 25.4 - 3 i.0 1.5 - 13.9 4.1 - 7.3 9.7 5.5 5.4 5.0 - 5.8 <LOQ (1.4) - 12.6 <LOQ (2.1) - 3.0 5.4 <LOQ (1.6) - 6.7 <LOQ (1.8) - 13.0 < LOQ (1.6) - 2.9 4.1 <LOQ (1.8) - 2.4 3.5 2.1 1.9 1.6 - 2.2 <LOQ (2.8) 1.6 1.4 - 1.8 1.8 1.4 1.2 - 1.6 MDls. - St. Paul Range Q l - Q3 Cumulative 90% Median Geomelric Mean 95% C.I. Geometric Mean Charlotte Range Q 1 - Q3 Cumulative 90% Median Geometric Mean 959'0 C.I. Geometric Mean 3M (7ompany EPI-O013 Page25 of 36 ! 7.7 - 207.0 23.9 - 43.3 71.7 31.7 33.1 29.8 - 36.7 <LOQ (I.9) - 20.0 3.2 - 5.7 9.9 4.4 4.5 4.0 - 5.0 <LOQ (1.4) - 15.2 <LOQ (1.6) - 12.9 <LOQ (1.0) - 10.6 <LOQ (1.4) - 2.8 5.:/ <LOQ (1.6) - 2.6 4.9 <LOQ (l.0) - 2.3 4.0 1.4 1.4 1.3 1.5 1.6 1.2 1.3 - 1.8 1.4 - 1.8 1.0 - 1.4 19.3 - 166.0 36.3 - 70.9 105.3 48.9 51.5 46.8-56.8 <LOQ (2. I) - 29.0 <LOQ (1.4) - 22.4 <LOQ (1.6) - 60. l <LOQ (l.0) - 10.8 4.5 - 8.6 " 13.3 1.5 - 4.8 10.9 <LOQ (2.8) - 4.2 8.6 <LOQ (1.8) - 2.8 4.7 6.3 2.8 1.8 1.2 6.3 2.8 2.4 1.5 5.6-7.1 2.4-3.4 1.9-2.9 1.3-1.8 3M Ctanpany EPI-(X)13 I_dg2. 6 of 36 Portland Range Q 1 - Q3 Cumulative 90% Median Geometric Mean 95% C.I. Geometdc Mean 6.0 - 1656.0 17.2 - 37.7 49.4 ' 26.0 27.0 23.5 - 31.1 <LOQ (2. i) - 16.7 <LOQ (2.1) - 11.8 2.8 - 4.7 <LOQ (2. I) - 2.5 6.8 5.5 <LOQ 2.8) - 36.9 <LOQ (2.8) - 3.8 7.4 <LOQ (1.8) - 6.9 <LOQ ( i .g) - 2.9 2.9 3.8 <LOQ (2. I) <LOQ (2.8) <LOQ (1.8) 3.6 1.6 2.5 1.3 3.2 - 4.0 1.4 - 1.8 2.2 - 2.9 1.2 - 1.5 Haeerstown Range QI - Q3 Cumulative 90% 7.6 - 226.0 24.4 - 48. I 69.8 2 <LOQ (2.1) - 52.3 <LOQ (2.1) - 66.3 <LOQ (2.8) - 21.2 <LOQ (1.8) - 7.9 3.2 - 5.9 <LOQ (2.1) - 3.8 <LOQ (2.8) - 1.4 <LOQ (l.g) - 1.9 7.6 7.3 3.4 3. l Median 35.7 Geometric Mean 35.3 ) 95% C.I. Geometric Mean 31.8 - 39.2 4.7 4.2 3.8 - 4.8 <LOQ (2.1) 2. l 1.7 - 2.4 <LOQ (2.8) 1.7 1.5 - 1.9 <LOQ (1.8) 1.2 l.l - 1.4 Los Angeles Boston Minneapolis St. Paul Charlotte Portland Hagerstown PI:OS Mean 95% CI 35.0 33.4 - 36.5 29.0 26.0 - 30.3 34.8 31,9 - 36.3 39.0 36.2 - 40.7 32.8 30.5 - 34.2 34.9 32.8 - 36.5 Table 4 Mean and 95% Confidence Intervals Calculated from Bootstrap Analyses l:or the Six Locations, Adjusted for Age, Gender and Their Interaction Term PFOA Mean 95% CI 4.6 4.4 - 4.8 5.3 4.6 - 5.6 4.5 4,1 - 4.7 PF! IS Mean 95% CI 1.9 1.8- 2.0 !.9 : 1.7 - 2.3 1.8 1.6- 2.0 PFOSAA , Mean 95% Cl 2.0 1.9 - 2.2 !.6 1A - 1.8 1.8 1.7 - 2.1 M570 Mean 95% CI t .3 1.3 - ! .4 1.3 1.2 - !.5 1.3 1.1 - 1.4 5.0 4.6 - 5. I 4.3 4.0 - 4.4 4.5 4,2 - 4.6 2.2 2.0 - 2.4 i.8 1.7 - 2.0 1.9 1.8- 2.1 2. I 1.9 - 2.4 2. I 2.0 - 2.3 1.9 ! .8 - 2.1 ! .4 1.3 - 1.5 1.3 1.2 - 1.4 1.3 1.2 - 1.4 3M Company P-I'I-00I 3 P-.,g2"/of 36 .. TOF Mean 95% CI 30.2 29.0 - 31.3 26.4 23.9 - 29.2 29.7 27.4 - 32. I 33.4 3 t.3 - 35.7 28.5 26.7 - 30.4 30.1 28.4 - 31.9 3MCompany EPI-0013 PaSe28 of 36 Table 5 ( Multivariable Regression Model of PFOS" by PFOSAA', M570", Age, Gender and Age x Gender Interaction Intercept PFOSAA" M570" Age Gender Age x Gender Coefficient 3.3 0.3 0.1 - 0.001 - 0.2 0.001 SE 0.07 0.03 0.03 0.001 0.07 0.001 t ratio 46.3 11.7 4.1 -0..8 -2.4 0.9 p value < .0001 < .0001 < .0001 .40 .02 .35 N= 645 *Natural log / ( Adjusted r2 = 0.22 .... Gender: females =1; males = 0 t ratio = coefficient/SE (standard error) .... ( Intercept PFOA* Age Gender Age x Gender Table 6 MuldvariabRlegression Model of PFOS by PFOA" Age, Gender and Age x Gender Interaction 3M Compan,, EPI-O0! Page29 of.36 Coefficient 2.7 0.6 - 0.001 0.02 -0.001 SE 0.0"/ 0.03 0.001 0.06 0.001 t ratio 36.4 20.3 -1.0 0.4 -1.0 p value < .0001 < .0001 .29 .72 .31 N= 645 "Natural log Adjusted r2 = 0.40 Gender: females - 1; males ; 0 t ratio - oefficiengSE (standard error) =. Table 7 Multivariable Regression Model of PFOS"by PFI-IS Age, Gender and Age x Gender Interaction 3M Company EPI-O013 Page 30 of 36 " Intercept PFHS" Age Gender Age x Gender Coefficient 3.4 0.3 -0.0003 -0.06 0.0002 SE 0.07 0.03 0.002 0.07 0.002 t ratio 45.5 9.9 -0.2 -0.8 0.2 p value < .0001 < .0001 .84 .43 .87 N =645 *Natural log Adjusted r2 = 0.15 /_ Gender: females = I; males = 0 . t ratio = coefficient/SE / ( .... " ( .. \ Table 8. Tolerance Limits and Their Associated Mean and Upper 95th Percent Confidence Limits for Serum Fluorochemicals and Calculated Total Organic Fluorine Index Tolerance Limit Mean Upper 95thPercent Confidence Limit PFOS 90% 70.7 74.3 95% 88.5 100.0 99% 157.3 207.0 PFOA 90% 9.4 10.1 95% I2.1 13.6 99% 19.8 25.8 PFHS 90% 6.3 7.0 95% 9.5 10.8 99% 17.0 22.4 .o-. PFOSAA M.570 TOF 90% 95% 99% 90% 95% 99% 90% 95% 99% 5.3 7.6 19.4 3.7 5.0 8.1 59.9 75.1 137.3 5.9 8.5 27.6 4.0 5.4 10.3 63. I 80.9 187.5 3M Company EPI-O013 Page 31 of 36 ._ Figure 1. Analysis of"Split Samples for Reliability Assessment for PFOS, PFOA and PFHS 2 S 4 $ / i I 0 ,, t 2 I ga I*_OA_m Jl ! / ; _ /% ! ( 3MC_y F.PI.OOi3 Page 32 of 36 t.i PFttS_ Rt ( Figure 2. Adult Study Population Distribuion of Mc_ur_.d Fluorochemical Concentrations I,- li " hul tot_S ;ppl_ "li o _ i .... .Rnn.. _._ _, i j'! 0 . _ EP[430013 Page._3_.f36 Figure 3. Box and Whisker Plots of Serum FluorochernicM Concentrations by Age and Gender 3M Company EPI-OOL3 lhge 34 of 36 --,4--- '3. ! o_" (/) S.S 2.2 ' t. z2 _= 0.0 s,_= oo 4[ &s .-e.-- : O.S age at last birthday' arw 7 , ,, , '0. ee age at last birthday _'u "4 e ; ; i , ! o. , , -- i age at last birthday' age at last birthday I _ - I _"2.5 | g u _- _= i_ ,, _ _ ,2 I" e-, rj 0 2.S ;;" s.s 6.t "2.; O.I ": 1 0"t i _.4.4 _ ..-..... ,,O.S ! ' age at last birthday 20-29 3_39 40-4,9 50-59 age at last birthday 80-89 Figure 4. Box and _,_"niskePr lots of Serum Huorochemical Concentrations by Age ant. - -..... - Sex'd --41-- --r-- ,. f _ -'_ -4.4 _, _ , _- .... A _ -2.2 --= I 3_ F 6 ' 0.0 C O -.e-- rj _ i, " 3M Comply EPI-0QI_ _g 3Sof 36 '3. .2. ,_. 4 ' '_' 2.2 ' --e-- city 2::1. _= -_- Set: P -- _ 0..5 _ -- 3El --e-- "4 --4)-- 2 _- = m o ,e5_ 0 _ ,,( j _- . -4 _ -2 " "0 N O' _ _ clty . O" -=_ _IS"q . ", " N = . I ,.0 d I mi i _- '-*- = " ,m I I ...... (: Figure5. ScattePrlot(slogscaleo)fFluorochemicAaslsoc/ations I:3 jI ' i , ' 3M Company EP[-oo_3 Pag3e6of36 - s. : _,": , -, " ; ". ,.rr:.'_. -. _-".... s : _ _....-_._..o.._.,._., I aI ,"'",,-__''.-.._"'.k._-"'_':" " s F | t -, o , , s 4 s ., ka PF_A _ , , _ ., , La PFH8 (Rm) t , II ?' T, II- S " %e ,_ .-", .'...,.'-"-,: .' I',"-....... ,.... "" i i "" .._,'.-'.',:"" -" - _.:".y -. 2 | " " IJ , . . ..'..'.:. _ ._..'."''":_&'__.2..,'_ _:.-; _.".{' " : .:".,:.._,_*. -._r:;'".,.,:*.. ... "" _.! t -I O _ ;e 3 4 S t , ta PFOE_ (1m6) -I 0 t 2 a 4 S I- M$_0 (M)) l I i ' :_ : :i , ,., ," .". _. . - . -, .. ,,_:-- ! ... _"v" "" I i -.... - " m. ou ! usye tlmt) i (I _ "" " "" _ .." " ".. _ :. . _-- : _. -,_ ..'._." ,...;.. _: -- .. i _ ".:""".."_,:,"..,:':___y!"'"_'"..'...._":"': iI , _,_,,- .._.... i1 t ! _ 4 t..JPP't4U_ mOt