Document J3a6m63XG1jZap50M6Y52npRB

FINAL REPORT Epidemiology, 220-6W-08 `Medical Department 3M Company St. Paul, MN 55144 Date: August 20, 2010 `Title: Biomonitoring Assessment of the 3M Cottage Grove Building 15 Demolition and Disposal Project: PhasIe Study Start Date: August 26, 2009 Protocol Number N/A IRB Approval N/A `Principal Investigator: Co-investigators: Geary W. Olsen, D.V.M., PhD.' Barbara A. Gibson, M.D., M.P.H.! David J. Ehresman, B.S., MT (ASCP)! Diane C. Madsen, CMA-C' Study Director: Carol A. Ley, M.D, M.P.H. 1. Corporate Occupational Medicine, Medical Department, 220-6W-08, St. Paul, MN 55144 2. Toxicology Laboratory, Toxicology Assessment & Compliance Assurance, Medical Department, Mail Stop 236-1B-22, St. Paul, MN 55144 3M ComPpagaeny Summary In 2008-2009, 3M and contract workers were involved with the removal of equipment `and lines (referred to as Phase I) in Building 15 prior o its demolition and disposal (Phase IT). A total of 66 workers participated in Phase I of the 3M Cottage Grove Building 15 Demolition and Disposal project. Fifty-six workers (84.8%) had both baseline and end-of-project serum perfluorooctanoate (PFOA) and perfluorooctanesulfonate (PFOS) concentrations determined. `The other 10 workers (15.2%) only had a baseline assessment. Forty-five (80.3%) of the 56 workers had a higher end-of-project PFOA concentration than at baseline, indicating `occupational exposure to PFOA during the project. For these 45 workers with occupational exposure, the mean numberof days between their paired measurements was 96 days (range 23 251). The 45 subjects' mean baseline PFOA concentration was 22.7 ng/mL (95% C10 - 50.1) and their mean end-of-project PFOA concentration was 156.4 ng/mL (95 CI 103.4 - 209.4). The corresponding geometric means for PFOA were 6.0 ng/mL (95% CI 4.4 - 8.3) and 78.1 (95% CI 52.0- 117.1). The mean matched-pair PFOA concentration difference for the 45 exposed individuals was +133.7 ng/mL (p < 0.0001). The 45 subjects' mean baseline and end-of-project PFOS concentrations were 25.9 ng/mL (95% C1 15.3 - 36.5) and 30.9 ng/mL (95% CI 19.1 42.6), respectively. The mean matched-pair difference for PFOS was +5.0 ng/mL (p = 0.07). `The lackofany major changes for PFOS was due to the fact that, unlike PFOA, PFOS had not been manufactured in Building 15 for several decades. `There were no statistically significant adverse changes in this workforce's serum lipids (total cholesterol, non-HDL cholesterol, HDL cholesterol), renal function (BUN, creatinine), or liver clinical chemistries (total bilirubin, alkaline phosphatase, AST, ALT) related to the change (difference) in serum PFOA concentration from baseline to end-of-project. For example, using 3M Company Page3 `multiple regression models, a 100 ng/mL increase in PFOA resulted in a non-statistcally significant decreaseof 0.8 mg/dL. total cholesterol (95% CI (-6.6) - 49),a 2.2 mg/dL decrease in non-HDL cholesterol (95% CI (-8.0) - 3.6), and a 1.4 mg/dL increase in HDL (95% CI (6.2) 3.5). In the adjusted models for these three lipid-related variables, none of the 6 covariates. considered in the models were statistically significant at p < 0.10 except for sex in the HDL model. The other covariates considered for these models were age, BMI, alcohol use, lipid lowering medication use, and the number ofdays between baseline and end-of-project `measurement. For AST and ALT, a 100 ng/mL increase in PFOA resulted in an unadjusted 0.6 ULL increase in AST (95% CI (-1.7) - 0.5) and a statistically significant 1.8 U/L decrease in ALT (95% CI (3.3) 0.3). This statistically significant association likely has no clinical relevance because the ALT value decreased with increasing PFOA concentration. Inferences from the present analysis are limited by the number of workers who participated and the study's inability to acquire fasting blood samples due to the logistical necessity of obtaining the samples from a predominantly contract workforce throughout their available work day. The lack of fasting, however, would only bias the serum triglyceride, blood glucose, and the indirectly calculated LDL measurements. In conclusion, this longitudinal assessmentof45 workers with demonstrated occupational exposure to PFOA did not observe changes in blood lipids, renal function, or liver enzymes related to their change in PFOA concentrations measured between baseline and end-of-project measurements. This longitudinal assessment does not support some cross-sectional epidemiologic studies that have reported positive associations between PFOA and non-HDL cholesterol. Unlike the present study, these cross-sectional investigations were unable to assess temporality becauseoftheir study design 3M ComPpaagneys Introduction Beginning in 2008, exposure remediation projects have been launched regarding legacy perfluorochemicals at the 3M Company. These included the 3M Woodbury landfill remediation project, the 3M Cottage Grove Building 25re-roof project, the 3M Cottage Grove DI/D2 excavation project, the 3M Decatur Building 2/49 demolition and disposal project, and the 3M Cottage Grove Buildings 15 and 73 demolition and disposal project (hereafter referred to as only Building 15). For each of these projects baseline and end-of-project assessments were required of 3M employees and contractor workers who entered specified work zones where potential exposure to legacy perfluorochemicals (.g., PFOA (perfluorooctanoate) and PFOS (perfluorooctanesulfonate) was possible, but due to the uniquenessofthe work, the magnitude of potential exposure was unknown. The purpose of this report is to provide an analysisofthe PFOA and PFOS serum concentrations (ng/mL) measured at baseline and end-of-project time periods for those workers involved with equipment and lin removal from Building 15 at 3M Cottage Grove. This is referred to as Phase I ofthis project. The second phase of this project (hereafier referred to as Phase I involved the subsequent demolition and disposalof Building 15 and these results have: already been reported elsewhere (Olsen et al. 2010). Trend analysesof the changes in PFOA related to serum clinical chemistries were conducted becauseof the higher PFOA concentrations observed among the majorityofworkers during Phase I. Despite several epidemiology reports describing positive associations between serum lipids and PFOA, none conducted longitudinal assessments. Methods 3M ComPpaagneys 1. Informed consent `The purposeofthe project was explained in an informed consent. Subjects read and signed this informed consent at both baseline and end-of-project assessments. Subjects were informed they could not work on this specific project without such compliance. 2. Clinical chemistries Clinical chemistries included a lipid panel profile, blood glucose, BUN, creatinine, serum electrolytes, and liver enzyme tests. Fasting was nota requirement becauseofthe logistics of collecting blood samples during various timesofthe day workers would arrive to be tested. Clinical chemistries were analyzed by Quest Diagnostics. The specific tests measured were: Total Cholesterol (mg/dL) HDL Cholesterol (mg/dL) LDL Cholesterol (mg/dL ~ indirect calculation) Triglycerides (mg/dL) Non-HDL Cholesterol (mg/dL ~ calculated) CholesterolHDL Ratio (calculated) BUN (mg/dL) Creatinine (mg/dL) Total Bilirubin (mg/dL) Alkaline Phosphatase (U/L) AST (UL) ALT (UL) Total Protein (g/dL) Albumin (g/dL) Globulin (g/dL ~calculated) Albumin/Globulin Ratio (calculated) Sodium (mmol/L) Potassium (mmol/L) CChalrobroindeDi(omxmiodle/L()mmol/L) Calcium (mg/dL) 3M ComPpaagneys Individual clinical chemistry analyses at baseline and at end-of-project were medically reviewed by Dr. Gibson. Subjects were informed in writing that this examination program was not a full medical `check-up.' Values outofreference range were indicated with a notationforthe subjects to follow up with their primary care physicianifthey had abnormal test results. However, questions could be directed to Dr. Gibson regarding their clinical chemistry test results should the subject so desire. At the baseline exam subjects also responded to abriefmedical questionnaire. Obtained at the baseline assessment was a self-reported questionnaire that inquired about basic demographic data (age, height, and weight), abrief medical history, and current medication use (blood pressure, lipid lowering, and glucose lowering). Statistical analyses included basic descriptive measures of central tendency, a matchedpairanalysis examining the changes between baseline and end-of-project for PFOA, PFOS, and the clinical chemistries, and multiple regression analyses employing standard statistical methods. Analyses were performed using JMP-SAS (Cary, NC). 3. Analytical measurementsof PFOA and PFOS Serum samples were analyzed for PFOA and PFOS by state-of-the-art high performance liquid chromatography mass spectrometry methods by the 3M Medical Department'sToxicology Laboratory under the direction of Dave Ehresman. Medical Department personnel collected blood samples that were processed to provide serum samples for analysis. These samples were assigned unique identification numbers and randomized prior to the samples being delivered for analysis. The 3M Medical Department's Toxicology Laboratory was "blinded" to the identity of all samples received for analysis. 3M ComPpaagney Sample extractions were performed using solid phase extraction (SPE) technique. The extraction and sample clean-up was based on a 100 uL. sample size and utilized Waters (Milford, MA) Oasis hydrophilic-lipophilic balance (HLB) 3.0mL cartridges (Ehresman et al. 2007) `The method used two stable labeled intemal standards for quantitation. The internal standards used were a dual labeled PFOS where two '*0 molecules were included in the sulfonate group (internal standard, >99% purity, synthesized by Research Triangle Institute, Research Triangle Park, NC) and a dual labeled PFOA molecule, where the carboxyl and alpha carbons were labeled with C stable isotope (greater than 97%, provided by DuPont, Wilmington, DE). All quaniitations were based on matrix matched extracted standard curves. AS uL injection of the sample eluate was introduced into the High Pressure Liquid Chromatograph (HPLC) which was directly interfaced into the triple quadrupole mass spectrometer (Applied Biosystems/MDS-Sciex Instrument Corporation, Forest City, CA). Standard curves covered the range from 1.0- 150 ng/mL. Standard curves were evaluated using a quadratic regression model where the standards were weighted at 1x, and each curve had an *"R" value equal to or greater than 0.9998. Matrix spiked controls (QC samples) evaluated during this study all had acceptable results "with-in" their previously established ranges. Matrixmatched dilutions were used for samples requiring dilution to bring the samples into the linear range of the assay. Extracted serum and aqueous blanks remained below the lower limit of quantitation established at 1.0 ng/mL (lowest standard fitted on the standard curve used for this project). 3M ComPpaagneys 4. Communication After each blood collection, individual letters were sent to the participants describing their results. In the baseline assessments, two letters were sent. One referred to the clinical chemistries and the other leter provided baseline PFOA and PFOS concentrations. At end-ofproject, two letters were again sent to each participant. The first letter provided the individual's clinical chemistry results. The second letter compared baseline to end-of-project serum PFOA `and PFOS concentrations. Results A totalof66 individuals were tested at baseline for serum PFOA and PFOS `concentrations and the aforementioned clinical chemistries. The majority were non-3M employees (86%). A subsetof 56 were also tested at end-of-project ofwhich 51 (91%) were non3M employees. Figure 1A is a scatterplotofthe 56 individuals who had baseline and end-of-project PFOA concentrations (ng/mL). As shown on Figure 1, the identity line represents those values where the x axis (baseline) is equal to they axis (end-of-project). Eleven (19.6%) individuals had serum PFOA concentrations at baseline that were greater than their end-of-project value, This is represented in Figure 1A by the 11 points to the right of the identity line. Forty-five (80.4%) individuals had their end-of-project serum PFOA concentrations greater than their baseline as representedby the points to the lef of the identity line in Figure 1A. Figure 1B is restricted 10 these 45 individuals whose end-of-project values were greater than their baseline. Ina similar manner, Figure 24 is a scatterplotof the baseline and end-of-project PFOS concentrations (ng/mL) for the same 56 individuals. Unlike Figure 1A with PFOA, PFOS is 3M ComPpaagneyd distributed around the identity line indicating there were minimal PFOS differences between baseline and end-of-project concentrations. Twenty-nine (52%) individuals had serum PFOS baseline concentrations greater thantheirend-of-project values. Figure 2B represents the PFOS baseline and end-of-project concentrations for those 45 individuals who had higher end-ofproject serum PFOA concentrations than their baseline values. Again, the majorityofpoints center around the identity line for PFOS in Figure 2B indicating minimal change in PFOS concentrations. Based on the findings in Figures 1A through 2B, three groups were identified for further description and statistical analyses: the 10 individuals who had baseline measurements only; the 56 individuals who had baseline and end-of-project measurements; and the 45 individuals (subset of the 56) who had2higher end-of-project PFOA level than baseline. For eachofthese groups, their distributions of serum PFOA and PFOS concentrations are found in Tables 2 and 3, respectively. Among the 45 individuals whose PFOA concentration increased over bascline, their arithmetic mean PFOA concentration was 22.7 ng/mL at baseline. The PFOA mean increased to 156.4 ng/mL at end-of-project (see Table 2). The PFOA geometric mean increased 1078.1 ng/mL at end-of-project from a baseline of 6.0 ng/mL. However, there was less difference in these 45 subjects' mean PFOS concentrations: 25.9 ng/mL at baseline and 30.9 `ng/mL at end-of-project (Table 3). The difference in PFOS geometric means was even less: 13.1 ng/mL at baseline compared to 21.2 ng/mL at end-of-project (Table 3). `The age, BMI, and numberofdays betweenbaselineand end-of-project for the 3 groups are found in Table 4. Among the 45 individuals whose PFOA concentration increased, the mean numberofdays between baseline and end-of-project measurements was 96 days. Provided in Table 5 are additional demographic data for these three groups as well as the prevalence of self- 3M CoPmapgean1y0 reported medical history, medication status, and alcohol usage. Most individuals were males and non-3M employees. Because the data were paired samples, a matched-paired statistical analysis was performed where the mean difference is the average of the sumofthe individual differences for the 56 or 45 subjects. See Equation 1 Equation 1. Mean Difference = 3s(teontsd-ofproject baseline)6or 45 Provided in Tables 6 and 7 are the mean baseline and end-of-project clinical chemistry Values for the 56 and 45 subjects, respectively, along with their mean matched-pair difference between baseline and end-of-project. For the 45 subjects whose PFOA values increased over baseline, the mean matched-pai difference was 133.7 ng/mL (p <0.0001)compared to 5.0 ng/mL for PFOS (p = 0.07) (Table 7). No statistically significant difference in the mean `matched-pwaaisr observed for total cholesterol, non-HDL cholesterol, HDL, or LDL. There was a 71 ng/mL reduction in the mean matched-pair triglyceride value (p < 0.0003). Other statistically significant (p < 0.05) mean matched-pair differences were observed for BUN, creatinine, total bilirubin, potassium, and carbon dioxide. The magnitude of these differences, however, was minimal andofno clinical significance. No statistically significant differences were observed for any liver enzyme mean matched-pair. Tables 8 and 9 are the cross-sectional analyses for the clinical chemistry measurements performed at baseline and end-of-project, respectively, forthe 56 subjects who participated at both times. At baseline, there were no statistically significant PFOA coefficients with either lipid, renal, or hepatic clinical chemistry measurements. The same was true with the end-of- 3M CoPmapgean1y1 project cross-sectional analyses. There were statistically significant unadjusted PFOA (negative) coefficients for chloride at baseline and for calcium at end-of-project. Of much greater inferential importance than the cross-sectional analyses are the longitudinal assessmentsof change (i.e. difference) in the clinical chemistries between end-ofproject versus baseline versus the concurrent change in PFOA (and PFOS) concentrations. Presented in Table 10A are the unadjusted and adjusted models for the 56 subjects for these longitudinal assessments. The adjusted models are the full hierarchical model that included 6 independent variables, besides the change (difference) in PFOA concentration. These covariates, included age, sex, calculated BMI from height and weight, alcohol use (2 4 drinksperweek, lipid lowering medication use (yes/no), and the numberofdays between baseline and end-ofproject measurement. See Equation 2. Equation 2. AY = a+ BAPFOA + 7X, WAh=erdiefference between end-of-project minus baseline value XY,==tthheedeip=e1ndtehnrtu 6clicnoivcaarliactheesmidsetsrcyrivbaerdiaabbloeve aB==irnetgerrecespstion coefficient of APFOA = regression coefficient for each covariate Xi Through backward stepwise regression procedures with removal of any variable setat p> 0.10, except for PFOA which was forced ("locked") into all models, those models that did not revert back to the unadjusted model shown in Table 10A, are found in Table 10B. Ina similar `manner, Tables 114 and 11B present the same data restricted to the 45 subjects who only had PFOA values that increased between baseline and end-of-project. 3M CoPmagpean1y2 Among the 45 subjects who had an increase in PFOA concentration during Phase I, there: were no statistically significant changes in total cholesterol, non-HDL cholesterol, HDL, LDL, or the cholestero/HDL ratio associated with the higher PFOA concentrations (Table 114). For example, a 100 ng/mL increase in PFOA was associated with an unadjusted non-staristically significant decrease of 0.8 mg/dL total cholesterol (95% CI (-6.6) - 4.9), a 2.2 mg/dL decrease in non-HDL cholesterol (95% CI (-8.0) - 3.6), and a 1.4 mg/dL increase in HDL (95% CI (6.2) 3.5) (Table 114). In these models, noneofthe covariates were statistically significant at p < 0.10 except for sex in the HDL model (Tables 114 and 11). Ilustrated in Figures 3A - 3C are the non-significant lincar trends and 95% confidence intervals for the unadjusted association between the changes in total cholesterol, non-HDL cholesterol, and HDL with the unadjusted change in PFOA concentration for the 45 subjects. A decrease in serum triglycerides was statistically significantly associated with an increase in PFOA for the unadjusted and hierarchical models (Table 114) as well as the backward regression model (Table1B). However, it is not possible to interpret this association since fasting was not a requirement. There were no statistically significant changes in liver enzymes associated with increasing PFOA concentrations among the 45 subjects (Table 114). Scatter plots for the unadjusted associations between changes in PFOA with alkaline phosphatase, AST, and ALT, are shown in Figures 4A~ 4C. Adding the change in PFOS, as well as PFOA, into the models displayed in Tables 11A did not result in statistically significant associations for PFOS (data not shown). 3M CoPmapgean1y3 Discussion The collectionofbaseline and end-of-project samples for analysis of PFOA and PFOS provided an excellent method to assess each worker's exposure experience involved with the 3M Building 15 Demolition and Disposal Project. There was an increase in serum PFOA concentrations at end-of-project among the majorityofworkers involved in Phase I of the Building 15 Demolition and Disposal project, indicating exposure occurred during the removal of equipment and breaking of lines (pipes). Specific exposure tasks were not identified for each employee; thus it is not possible to ascertain which tasks yielded the highest potential for exposure. Increased exposure, determined by paired biomonitoring analyses, has not been observed at other 3M remediation sites that involved workers at the 3M Woodbury landfill remediation site, the 3M Cottage Grove Building 25re-roofproject, the Cottage Grove DI/D2 remediation site or Phase ILofthis Building 15 Demolition and Disposal project. "The higher end-of-project serum concentrations from Phase I provided an opportunity to conducta longitudinal assessmentofthese 45 individuals to determine whether their serum lipids, renal function, and liver enzymes were affected. The data do not suggest that serum. PFOA concentrations, that increased an average of 133 ng/mLover a 3+ month time period, modified total cholesterol, non-HDL cholesterol, or HDL measurement. "The baseline averageof22.6 ng/mL for the 45 workers is approximately 4 times the general population average (Nelsoent al. 2010), but slightly less than the median PFOA concentration of 28 ng/mL reported for the affected mid-Ohio River valley community by Frisbee et al. (2009). Iti 1 to 2 ordersof magnitude lower than averages reported for occupationally exposed workers. 3M CoPmapgean1y4 As reviewed by Steenland et al. (2010), positive associations have been reported between PFOA concentrations and non-HDL cholesterol levels in different populations including the general population (Nelson et al. 2010), a community affected with contaminated drinking `water (Emmett et al. 2006; Steenland et al. 2009), and in some (Costa etal. 2009; Olsen etal. 2003; Sake et al. 2007a; 20075), but not all occupational investigations (Olsen et al. 1999; 2000; 2007). Mostofthese studies, however, have been cross-sectional investigations that can not fundamentally address the issueof causality due to their inability to assess temporal relationships. To add further to the confusionofthese published studies, the strongest statistical associations reported have been in those populations with the lowest PFOA concentrations (Nelson et al. 2010), whereas the weakest associations have been reported in occupational populations whose serum PFOA concentrations are 2 to 3 ordersof magnitude higher than the `general population (Figure 5). Steenland et al. (2010) hypothesized that such findings could be the consequenceof a biological pathway with non-HDL cholesterol that may be saturated at relatively low (i.., general population level) PFOA concentrations. Data presented graphically by Nelson et al. (2010) based on the National Health and Nutrition Examination Survey (NHANES) general population, and by Steenland et al. (2009) from a community exposed through contaminated drinking water, do suggest that this saturation might exist between serum PFOA concentrations ranging between 20 and 50 ng/mL. Although no association was seen with non-HDL or HDL cholesterol in the present study, the average baseline concentration was near this threshold level. `The present study did not indicate perturbations in renal function or liver enzymes with increasing PFOA concentrations. Therefore, the present study's longitudinal assessment does not support the findings by Lin etal. (2009) who conducted a cross-sectional analysis of 3M CoPmapgean1y5 NHANES data and reported a 1.86 unit increase in ALT with a 1 unit increase in log PFOA concentration. If ths statistical association by Lin et al. was extrapolated to the present study, then the ALT concentrations among the 45 workers with increasing PFOA concentrations would have increased several ordersofmagnitude. This did not occur. Nor have there been clinically relevant increases in liver enzyme values observed in a series of cross-sectional or longitudinal occupational studies of ammonium perfluorooctanoate production workers (Costa tal. 2009; Olsen et al. 1999; 2000; 2003; 2007; Sakr et al. 2007a; 2007b; Steenland et al. 2010). Ina longitudinal assessmentof 454 DuPont workers whose medical records were abstracted for many years, Sakr etal. reported a 1,000 ng/mL increase in serum PFOA concentration was statistically significantly associated with a 0.008 mg/dL decline in total bilirubin (95% CI -0.014 t0 -0.002), 20.35 U/L increase in AST (95% CI -10t0 40.60), and a non-significant 0.54 U/L increase in ALT (95% C1 -0.46 to +1.54). These statistical associations would not have clinical relevance if extrapolated to the much lower concentrationsof PFOA reported in the general population. "The present study is limited by the number of workers who participated and the inability to obtain fasting blood samples. This inability to request fasted blood sampleswas a logistical necessity because the predominantly contract workforce would participate throughout the normal work day. The lack of fasting, however, would only bias the serum triglyceride and glucose measurements. The lack of fasting could also affect the LDL cholesterol results due to its indirect calculation by the Friedewald formule. Consequently, these specific clinical chemistries are uninterpretable. Total cholesterol, non-HDL cholesterol, HDL cholesterol, renal function, and the liver enzymes would not have been biased by the lackoffasting. Also, it could be argued that the average assessment periodof approximately 3 months between baseline and end-ofproject may have been too short of follow-up time to observe an altered clinical chemistry 3M CoPmapgaeny1.6 profile. However, increasing serum PFOA concentrations did not result in higher blood lipid values in male cynomolgus monkeys dosed daily for 6 months that resulted in steady serum concentrations ranging between 81,000 ng/mL and 156,000 ng/mL depending upon administered dose (range 3 to 20 mg/kg/day via oral capsule) (Butenhofetf al. 2002; 2004). There was minimal change in PFOS concentrations in this study. The lackofany major changes was due to the fact that, unlike PFOA and its salts, PFOS had not been manufactured in Building 15 for several decades. Conclusion Based on biomonitoring data, serum PFOA concentrations increased over baseline for 45 (80.3%)of 56 workers during Phase I ofthe Building 15 Demolition and Disposal project. Phase Tinvolved removalofequipment and lines. There was an arithmetic mean matched pair increase of 133.7 ng/mL PFOA (range 4.6 - 697 ng/mL) above the baseline average of22.7 ng/mL. The `geometric mean matched pair increase was 78.1 ng/mL PFOA. There was minimal change in PFOS owing to a lackofexposure in Building 15. There were no statistically significant changes in these 45 individuals' serum lipids (total cholesterol, non-HDL cholesterol, HDL cholesterol), renal function (BUN, creatinine), or liver clinical chemistries (total bilirubin, alkaline phosphatase, AST, ALT) associated with the increased PFOA concentrations that occurred over a mean of 96 days (range 23 - 251) between baseline and end-of-project `measurements. 3M CoPmapgean1y7 References ButenhofJt, Costa G, Elcombe C, Farrar D, Hansen K, Iwai H, Jung R, Kennedy G, Lieder P, Olsen G, Thomford P. 2002. Toxicity ofammonium perfluorooctanoate (APFO) in male cynomolgus monkeys after oral dosing for six months. Toxicol Sci 69:244-257. ButenhoffJL, Kennedy GL, Hinderliter PM, Lieder PH, Jung R, Hansen KJ, Gorman GS, Noker PE, Thomford PJ. 2004. Pharmacokineticsof perfluorooctancate in cynomolgus monkeys. Toxicol Sci 82:394-406. Costa G, Sartori S, Consonni D. 2009. Thirty years of medical surveillance in perfluorooctancic acid production workers. J Occup Environ Med 51:364-372. Ehresman DJ, Froehlich JW, Olsen GW, Chang SC, ButenhoJfLf. 2007. Comparison of human whole blood, plasma, and serum matrices for the determination of perfluorooctanesulfonate (PFOS), perfluorooctanoate (PFOA), and other fluorochemicals Environ Res 103:176-184. Emmett EA, Shofer FS, Zhang H, Freeman D, Desai C, Shaw LM. 2006. Community exposure to perfluorooctanoate: relationships between serum levels and certain health parameters. I Occup Environ Med 48:771-779. 3M CoPmapgean1y8 Frisbee SJ, Brooks AP, Maher A, Flensborg P, Amold S, Fletcher T, Steenland K, Shankar A, Knox SS, Pollard C, Halverson JA, Viera VM, Jin C, Leyden KM, Ducatman A. 2009. The C8 Health Project: Design, methods, and participants. Environ Health Perspect 117:1873-1882. Frisbee SJ, Shankar A, Knox SS, Steenland K, Fletcher T, Savitz DA. 2010. The C8 Health Project: associations between perfluorooctanoic acid and perfluorooctanesulfonic acid and serum lipids in children. Arch Pediatr Adolesc Med (in press). Lin CY, Lin LY, Chiang CK, Wang WJ, Su YN, Hung KY, Chen PC. 2009. Investigation of the associations between low-dose serum perfluorinated chemicals and liver enzymes in U.S. adults. Am J Gastroenterol doi:10.1038/aig.2009.707. Nelson JW, Hatch EE, Webster TF. 2010. Exposure to polyfluoroalkyl chemicals and cholesterol, body weight, and insulin resistance in the general U.S. population. Environ Health Perspect 118:197-202. Olsen GW, Burris JM, Mandel JH, Zobel LR. 1999. Serum perfluorooctane sulfonate and hepatic and lipid clinical chemistry tests in fluorochemical production employees. J Occup Environ med 41:799-806. Olsen GW, Burris JM, Burlew MM, Mandel JH. 2000. Plasma cholecystokinin and hepatic enzymes, cholesterol, and lipoproteins in ammonium perfluorooctanoate production workers. Drug Chem Toxicol 23:603-620. 3M CoPmapgean1y9 Olsen GW, Burris JM, Burlew MM, Mandel JH. 2003. Epidemiologic assessmentofworker Serum perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) concentrations and medical surveillance examinations. J Occup Environ Med 45:260-270. Olsen GW, Zobel LR. 2007. Assessmentoflipid, hepatic, and thyroid parameters with serum perfluorooctanoate (PFOA) concentrations in fluorochemical production workers. Int Arch Occup Environ Health 81:231-246. Sakr CJ, Kreckmann KH, Green JW, Gillies PJ, Reynolds JL, Leonard RC. 2007a. Crosssectional studyoflipids and liver enzymes related to a serum biomarkerofexposure (ammonium perfluorooctanoate or APFO) as part of a general health survey in a cohort ofoccupationally exposed workers. J Occup Environ Med 48:1088-1096. Sakr CJ, Leonard RC, Kreckmann KH, Slade MD, Cullen MR. 2007b. Longitudinal study of Serum lipids and liver enzymes in workers with occupational exposure to ammonium perfluorooctanoate. J Occup Environ Med 49:1086-1096. Steenland K, Tinker S, Frisbee S, Ducatman A, Vaccarino A. 2009. Association of perfluorooctanoic acid (PFOA) and perfluorooctanesulfonate (PFOS) with serum lipids among adults living near a chemical plant. Am J Epidemiol 170:1268-1278. 3M CoPmapgean2y0 Steenland K, Tinker S, Frisbee S, Ducatman A, Vaccarino V. 2009. Association of perfluorooctanoic acid (PFOA) and perfluorooctanesulfonate (PFOS) with uric acid among adults with elevated community exposure to PFOA. Environ Health Perspect 118:229-233. Steenland K, Fletcher T, Savitz DA. 2010. Epidemiologic evidence on the health effects of perfluorooctanoie acid (PFOA). Environ Health Perspect 118:1100-1108. 3M Company Page2] `Table 1. Distribution of Workers by Company and Participation for Phase Iofthe Building 15 Demolition and Disposal Project Company Non-3M employee 3M emplovee Total Baseline 57 9 66 End-of-Project si 5 56 Reason for No End-of-Project Sample Unknown Unknown 3M CPoamgpean2y2 Table 2. PFOA Baseline and End-of-Project Concentrations (ng/mL) for 10, 56, and 45 Subjects Involved with Phase Iofthe Building 15 Demolition and Disposal Project Participated Only at Baseline (N = 10) Baseline Arithmetic Geometric: Min Ql _ MedinQ3 Max Mean 95%CI Mean 95%Cl 29 76 333 1963 345 1029 54-1983 334 96-1160 Participated at Baseline and End-of-Project (N = 56) Arithmetic Geometric Min Ql Median Q3 Max Mean 95%ClL Mean 95%Cl Baseline 13 40 61 224 1325 1000 354-1646 125 7.6-205 End-of-Project 4.6 37.6 1138 2795 1,026 194.1 1358-2524 04.1 650-1364 Participated at Baseline and End-of-Project and Had Higher PFOA Concentrations at End-ofProject (N = 45) Arithmetic Geometric: Min Ql Median Q3 Max Mean 95%CI Mean 95%CL Baseline 13 32 53 71 611 227 0-501 60 44-83 End-of-Project 4.6 37.6 939 2065 697 1564 1034-2094 781 520-1171 3M CoPmapgean2y3 Table 3. PFOS Baseline and End-of-Project Concentrations (ng/mL) for 10, 56, and 45 Subjects Involved with Phase I ofthe Building 15 Demolition and Disposal Project Participated at Baseline Only Baseline Arithmetic Min 146 Ol 247 MedianQ3 365 3108 Max 8725 Mean 181.8 95%Cl 0-386 MeGaenome9t5r%icCl 64.1 225-1828 Participated at Baseline and End-of-Project (N - 56) Baseline Arithmetic Min Ql Median Q3 Max Mean 95%CI 44 133 184 366 258 456 292-620 End-ofProject 5.9 160 242 446 229 468 317-620 Geometric Mean 95% Cl 254 195-331 281 217-364 Participated at Bascline and End-of-Project and Had Higher PFOA Concentrations at End-ofProject (N = 45) Bascline Arithmetic Geometric Min 44 Ql 122 Median Q3 160 258 Max 187 Mean 95%Cl 259 153-365 Mean 95%Cl 18.1 146-225 End-ofProject $9 126 197 304 218 309 101-426 212 168-268 2, LIER 3d 82 A oaE oe ZV| 3%2 2 d=a Be: Ay sw 9 El &08 az8 2 2 oun 2g ozee < g ga 2 2 gles a 77 S3 J3YH a230 3< 2 2 2s 28scy| E2 < EEE g2 2#23 2432 37= z } I goigeze ZEoS lsBAf erns z2 2? g dl i es gaz :3 3 :3 ,Af H* gg 2 a os ssez ogzgo eee a ena g sar o a a=n 2BE33 Bem mam Sea gi 32s3s33s3s sgszszs3sssz g REg : Ego lan ee 8 neces aoe : s zilagiigiliy E g ge 8 2 23 i: 73 3 3c 23 353 2 252:822 31 5 22:5828S382%Ff3555382 2 2 & 5& 3 3 I IEEEE ! f 5333 132 i i2 y ; . z 2z $2Te 3s 22 a8=: 28zTz27F8E8E zi 83 ~ I. Zo g g2 g fF2 ETfgfgF s i ce3 i cee n iI8 lis e ra cilT s t EsI g% 3 =| zs Fe3osFdaie=z2= s-8- <5 s = z2% 2= 2 = 3 ga = 3 H2323IEdoiE g1 c5iE 2 EoH EsEiR 8 Ri2 EIR .iE E7s3 lER 2s St2 SE 83 2 ogo J =. g% 3 2 $8 2 = 8 8 2 5 2 38 EZ gE5nr5 F So os - 5_ ga3 g4g 27d7 Y88TT =8 E7 s 27 s 2 o9z s3f2 JH EEeEaEEfEE ER se l= ad-Ss 28312 15 g5f it E3gE8z82E8%3g8eds EH _5 2 g2 5 $aB3E8 ii $83i2 -z8:2z533:z221,:%gF3E%5 Ee EE EEBEBECSEE 2 s= 2 =& 2dg$:3s3e5s%%8 325a=35%3=83338883 8%ox g &E 5 333| 233a3 97 .12432208a 82 I2 fdzlzslisiziiiig = 2gg22% cco 2& E=38a = T Je omeSa $ Es 2o33. dtd Tliliglgaed 48 a28 8 225s 284 54 o f ea . T- e on owZ222 ow aTE 2 L2 S oo dea=2%as1184%g dT 030i ede 4 8&8 3IFTI AISI & Seas oa=ss Eds g 3 2<g =t gE f23fi E 3 : 282 2 igsgiiis iffy & 2 22 RR E522 EES "2 z z < yz 48. 2022 222 828 z2% Jd 3222e 8E83 33 i2 gE 8E, HP8 JPIEgsa= Lei w, f 5Te Lf I E g 3L R B5r3 ogI g gceI s iTl az iaf za iles cse s EH 28 85 a 2 g : = ip fEsdivatzniooe HE i2E 3E823 _2o0.%4 $2 fg: sR 52 HBfH 2E 3 82 2 3 2 & EFE g8 4 e 388 32gg 59 2S2 2s :3 qd 33 3g 5 ER ggfE 28282 88c8 s 32 2 x 55283 SF3 og &8T2F8Tz e F5 z=o 2eg t3i 4L 5328o fl Isls xegT dge zd dde % 3 = 5 f 4 5 EE 9 i2f 2H 3 3: 2=% 21 gz 8oo8-8= 2= 22 o8 o 82223 k Lf3 i3 Ei ; ff 2 i 82 32s, f: EE2 $8: EE & 2 i 3%zE3 0 228878 3 357 %2 88< z$3gi fIgd s2282 3 v2 x 2 2 83 55835 88 E8 B g8 Jds = 32 234d 32 8 2: A 9 i.i iiia $5. 53553355 EEC A 22 cc 00 2% oT "woe on = 3 =ZsF oo Hggaeasa0 T 225d 2L 8S T 2T e 83g 1 3 3g 2-- fs x 2 $3 2% 2 8 FY des JHe2 48 32 5 Sk 8 2g 3nolgowlEzds 2 2 3e : 5 d3 2 2 5 8 82 3 %39 s a= 2 325% 3% E353 g <3 3 5- |. .EszEzF: :E55:3%8;:2;S & u2z 25 8 3 223 2%28 &3822 :38d% gir z I $383E E5E 3%E8E E8:E5E%8E E 8E %E3 :z22 SgH5yEssE 3 s$323 s5I3s353:83 823 8535s3 883 %43s 23 i dF2 gz E58g E52 835288X3 T %Ts38Os88@ss= 8= Z Hq ss 9 sc 35 5 3 3 3 3 zflr5 s egzofgEefzgiEoi8 oie t 33 FE29 83S 3 3253 8 ss35 8 288 8= 2 < IEEEEEEEEEEER & 8888866&&&& I, ass zeszz3:z583 S$yF 35 5 SES2SE888E8828:8285:82325588388 [EI [H2HEI2EER 7 E H88II]sESE 2 E32E S E3 38S E E28S E E32I REs 5E R8z cE R%i8E Rs%e EEsEE Ee 5ReR8s3 33 0%Z:izsgpzsogis. zsd IBSF PEES $ s:3 zf3f985z9ScsS s3 sE9s3 i Te I25z iTzT 8 54 ys 23 32E3H 3{ <2Ft8s: eRRaE3: gB2sE%,zl5$:if2 iEyg FE 3 33223$355388833888858 233 Tle 2 2 2 . : IE EEE EREE EI gos & 8 3 3 3S 3 3fle sc ss =z8z4 giz gs:8E8:88 B33 fF 9 9 2 % 9 7 3.0. o g 3 B$32<88fd58 582 8 2 E a =8 8 8E 28 %=8 E o yg +4 I REIEIIIEEIEEE 5i 59$38325332s 8d %3s %5gyd Y32 J( z sE sEzEEsE EsEsEsE EsNzE2 gos SS 8 3s 33 3S 9 % Ji3 8 ss82ss883382528882833 82 g3s 9 5 59 3% 9 3 wu - EE EEEE EE RE 558 s E8888% = 5883 $< 3 5 3 <5 9 g 3 . 2 IT. 5P3ss:t:27 5:2%.3%1 & J2 232 3 388 PM Zz ER g i3 Z3 z3 I % z zPf = & 25g Sz |)3< f35 Ek] 3Iz5 SE 3ii25 5 2 8 8 943 88 3 s 8] e3 3 = 3 83S 73 S= S3 | ,rsressezgzfigss 4g85 2 2352 2S 82 8 SS823 898 2S 2 83 33 fl 2 np 22 2% 35 9.8 = FEE EEEEREEERERE 8535333538338:3 3 3I.ii.gg.zg ca ezzBsgiagsssg agd 2F09E 8S22 3825 328 23 338 38 382 g8 38 13c5 s 2 22 85S 28538 88888%292 88 gg)8 =2g5x882 3 52285 2988g% tIg5IESE3E3E33ERE3E 53E%E3E32R3E3]2 EElJ e8E5v R8w E3wEEz ERs EEs EEe REEeE EzasE D zl dy 2 5 9 2 9 5% 9% q3 eHctgii.zg:gziaizizzziecoia 588 gz 3 f 9 F&F <2 9:3 J 21 S2 2S 8 2%3 2, : Zz83 2i8i:& 2:zE3E3zS:E:C:23R 31S z2E g23  z 12 2vw3v228282281 88e 8od 9FT : 7 5 S 2 3 238ggE SgEsE E3 E3 S 3 3 E 3 3 S 3 3 E 3 g9 3IE 22 E 2 5EEEEE8 E8 EY %8EEF q83sEE E E9I.I9 E2 E% E E9;3 a ssf ii EE fEifigd 22d 2 335338243 93 2 y Jz28zs3285gE38g 323g 3g3og 9g:t - i 2E8s: s8 ga s58aa s8 eg gs z2 jEEEigsziiig 2Slee F188 8c3 83: gl Cgs His g8 e 2 8 e 2 s 8 d3 e g8 i2 222 9 99 2 93 32 Z% fderasgsezaeg; i5838538 s 3Es2 s888 &58 Pg g. fs o %2 5.18ol1 E$s AH E Ei iif :3ii s5: 4% 43 :2 : 2 - zz3 585538 Ig8s8 L28s3s8 s3883 88:8 3 3E< ogi|EEe EiE=iE2zEoioEi5EiFEoi3E oieE ois gE ncoE aee 1 Ho: ' E1E S12 382 $ 88823837 fs 2 F8y2g2ssS Eg$sE% L% g3 E5 d5s 22 3g o2_o 90582888888 2 8 zz 8 5g sg gagesg Z 2 98 s 8 383538%559%298&8<8< g2 9 = &2 Ji8m2a3283f5s3z8a2g3s3e3 s 28: JiET5iE5sE3fRgsE3dR sgsEz8sE s:R z38zREE sEzR3E:EE3 g gg 2<28 T4F 5Z4|85a z&e=gez c ageEzigazcEBoce 9 EE FS J f3 5SS8s 3s 8s 288e 8c 8zeS 8 3 33 Is dE egEfE igs EE E iE aiEgE EsEgEt] 3E 2 33 233 ss 3% 3 23 += 2cE82 _zs I3 32 3 3 25 4 gi)i2c2 B2E 3fp, 28 2: 2 RB 22g 2EF EF 5 2: (iE ziiiliiin $ z Sz 5 2 28832289 J FEggr E = 2E E8E Z= Ea: E9Eo Jg f353s2 8s88 88 88 8 88 g 8 8%38 33 18s ss83z32R83 FE se EEEEEEg Bs =25529 HE EEEE REE: 5s 88:388 EQ? 3 8 8 & & EERE EERE FHlyeg E2 E= 5s vw $ 2 89 2= 82 33 IEEEEEEEEEE Egg 3go 8 & 8 ss S 4 32o8,8 2 Fig E E2a82 222 8 8 8 8z=8 E s 8: E gE:;5 5 83% 5% S9993 3 SAREERRRRRE Be 2 2 a ow PRE s EES 232888833 I 2 2 g 3 Sg 334 2 2 : 52f3:3F %;-:83 8 Bg EE & 2E8E32E&EE8R38EJE Table 10B. Adjusted PFOA Regression Coefficients for All (N = 56) Subjects with Baseline and End-of-Project Measurements. Includes Only Covariates withp < 0.10 Coefficients, Phase I of the Building 15 Demolition and Disposal Project Adjusted Dependent PFOA 95% Confidence Interval Varisble Coefficient Lower Upper HDL 000321 001265 0.01908 Triglycerides 045527 ~~ -066000 -0.25054 Cho/HDL 0.00187 000023 000350 Glucose 002069 -007138 0.02999 Albumin/GIob0.000001 0.00026 0.00027 Sodium 000223 000122 000568 Potassium 0.00011 -000066 0.00087 Chloride ~~ 0.00215 -000588 000157 Carbon Dioxide 0.00476 __ 0.00037 0.00914 pvalue 0.69 <0.0001 0.03 042 099 020 0.78 025 0.03 Adjusted for in Model sex lipid lowering med status sex lipid lowering med status sex age, sex sex age BML sex 22 Eg 2 2 338 FEe s$23 dg2g%3=8 223358388 4338z 3 53z 32580253555 8838:c3 g8z8sgzgcdE a3 gJ g398 338s8 83298:5 38338 85 3858 85:8 z gz [3EF 33I/E8E5E8E2E8E 55E 835E R5EE 3Es z I 3S 2 9 93 9 FS 5 5 <2 232 23oBo5 of2 ggaep0 ig2 g2 sg8 e8e22 8 ag 5V%E Sg E $ 8 8 3 83 88 3 88 383 8 383 EH 2T5gE2 Efr3iz&z8s8233 8I8f3g 888588S8 2 5S2E Js 2 8s 28 288 x2 88% Sz BEfp 34 fyyfst sE2 3i3 f38as8s2S3 8s8 2:e85gg 8 28 2 23 Z85z25: f9 d3rd 2 2@as8E8ga s8 e 2c8a i 58g 8S 2IE Ts 9 E: gSg 23 SH 83 93 9% 3 33 ||s3t.Eds tglezcie:oza iE:omioza o:m 35 [gd ss88&833533:33 i2f3 iz 3 E23 3 3 S3<5p=2 2 p=|dt3 " Ba ly FF ,fFd 2E2z F28RE3 E2E2 E 3 3E8 iETE EEF iE SE n& $ z EF I=Ss 35 32ds3 52 2E88 3= 32 3 oy Tflge% Z2 Z8 Zss 5 238 2 4 23 4 2g4525 S s 32 23 88 2228%3 2 H2 2d E5 Jlga28222 52555 525=55a 2 %8885 g <| ? 9 9 FF % 3 9 Ss992z2 2 8 388 &Z5s=83g33z2g| gaz s:gssssscs fyS $$s533838%233s3858233 Bgizezczczcig gys s 3 Ss 832 3 2 3,8 2 52s s 08 02228 2 @88 72 E58888 ; H4 y 2 - 2 8 eg oom 23 3s%ez g 8 35g 2gg88g% =z gg 238338 :588; 3 E o g 5d : 2 fog:2 $2 ,8 3 3 E ig28E 2 f E2r E8s E2t E3E2f3 Es8Et8E3R. `ETnadb-loef-1P1rB.ojeAcdtjuMsetaesduPreFmOeAntRseTghreastsHiaondCIonecfrfeiacsieedntPsFfoOrAAlClon(cNen=tr4a5t)ioSnusb.jeMcotdsewlisthInBcalsuedleinOenlaynd Those Covariates Disposal Project Where Cocflicient(s) p < 0.0001, Phase I of the Building 15 Demolition and Dependent APdFjOusAted 95% Confidence Interval Variable Coefficient _ Lower Upper HDL 0.00899 0.01093 0028913 Chol/HDL 000204 -000015 0.00423 Creatinine -0.00024 ~~ -0.00047 0.00001 Total Bilirubin 0.00002 0.00004 0.00004 AlkPhos 0.00465 -001536 0.02466 Sodium 000089 -0.00498 0.00321 Potassium 0.00064 -0.00028 0.00156 Carbon Dioxide 0.00385 -0.00046 0.00816 pvalue Adjusted for in Model 037 sex 0.07 sex 0.04 age 092 alcohol, days between tests 0.64 sex, lipid lowering med status 0.66 age, BMI, sex, alcohol 0.17 sex 0.08 sex. alcohol FPihgausree 11AB.ase3lMineCoatntdagEendG-roofv-PeroBjueicltdiPngFO15ADCeomnocleinttiroatniaonnds D(insgp/omsLa)l, Project. N= 56 Subjects. wo mm 3 Se = Sw val 2 Soe Sa. { 4 : : I) w on HON) 3M CoPmapag4ne1y SM Copmpeanty Figure 1B. 3M Cottage Grove Building 15 Demolition and Disposal Project. PhasIe for NN = 45 Subjects with Increased End-of-Project PFOA Concentrations (ng/mL) over Baseline. . Ra Lr wl La : :: Cowl Zu J |: o @ wo Fonsi Figure 2A. 3M Cottage Grove Building 15 Demolition and Disposal Project. Phase I. Baseline and End-of-Project PFOS Concentrations (ng/mL) for N= 56 Subjects -- I | 33 | | : Lat | 2J WE | A | : es = p|s mpady awy= Figure 2B. 3M Cottage Grove Building 15 Demolition and Disposal Project. Phase I for PFOS Concentrations (ng/mL) for N = 45 Subjects with Increased End-of-Project PFOA Concentrations (ng/mL) over Baseline. i. m Z:, poee : 7W PFOSBaseline (nginl) - Figure 3a Total Cholesterol Difference by Change in PFOA. i Ho. : FI rBeRE. eEe, Figure 3b. Non-HDL Cholesterol Difference by Change in PFOA . iil .i rAaK! Figure 3c. HDL Cholesterol Difference by Change in PFOA To i:: AK ma wo so ew 70 smim Figure 4a. Difference in Alkaline Phosphatase by Change in PFOA :a ! i %o an Bo ah wo oh 7 GrangeinPFOA rt) Figure 4b. Difference in AST by Change in PFOA = i" HL e5 o] -He LC 0 0 0 mo on chianPFnOAgGroem) Figure dc. Difference in ALT by Change in PFOA = - i ps -HR dre a wwPFnOwRore sao) mo Too) 3M CoPmapgean4y6 SM Cormpayny FChiogluerset5e. roMl eian nEpPiFdeOmAioCloongcyenRtersaetairocnh(Sntgu/dimeLs). and Slope Adapted of Linear Relationship`with from Steenland et al. (2010) - - . . 3. S g ory ii--igsz3 001 casero 3. .: i3is= g3el iE | = | _ He 3 s5 [[|Hs|| oozres | cMEid Rsiageneals| wh o2f 83P8RR3ls $2 0.001 L