Environ. Sci. Technol. 2009, 43, 4589–4594
Observation of a Commercial Fluorinated Material, the Polyfluoroalkyl Phosphoric Acid Diesters, in Human Sera, Wastewater Treatment Plant Sludge, and Paper Fibers JESSICA C. D’EON,† PATRICK W. CROZIER,‡ VASILE I. FURDUI,† ERIC J. REINER,‡ E. LAURENCE LIBELO,§ AND S C O T T A . M A B U R Y * ,† Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada, Laboratory Services Branch, Ontario Ministry of the Environment, 125 Resources Road, Toronto, Ontario M9P 3V6, Canada, and Office of Pollution Prevention and Toxics, U.S. Environmental Protection Agency, Mail Code 7406C, 1200 Pennsylvania Avenue, Washington, DC 20460
Received January 12, 2009. Revised manuscript received March 11, 2009. Accepted March 23, 2009.
Sources of human exposure to perfluorinated carboxylic acids (PFCAs) are not well-characterized. Polyfluoroalkyl phosphoric acids (PAPs) are fluorinated surfactants used in human food contact paper products. PAPs can migrate into food and food simulants, and their bioavailability and biotransformation into PFCAs has been demonstrated using a rat model. To characterize human exposure to PAP materials, we analyzed pooled human sera samples collected in 2004 and 2005 (n ) 10) and 2008 (n ) 10) from the midwestern United States for the 4:2 through 10:2 PAP diesters (diPAPs). The 2004 and 2005 sera samples contained 4.5 µg/L total diPAPs, with the 6:2 diPAP dominating the congener profile at 1.9 ( 0.4 µg/L. DiPAP concentrations observed in the 2004 and 2005 human sera samples were similar to those of the C8 to C11 PFCAs (0.13 ( 0.01 to 4.2 ( 0.3 µg/L) monitored in the same samples. 6:2 diPAP was also consistently observed in the 2008 human sera samples at a mean concentration of 0.63 ( 0.13 µg/L. As diPAPs have been shown to degrade to PFCAs in vivo, our observation of diPAPs in human sera may be a direct connection between the legacy of human PFCA contamination and PAPs commercial applications. Wastewater treatment plant (WWTP) sludge and paper fibers were analyzed for diPAPs as a proxy for human use and potential exposure to diPAPs. DiPAPs were observed in WWTP sludge at concentrations ranging from 47 ( 22 to 200 ( 130 ng/g, a range similar to perfluorooctane sulfonic acid (PFOS) (100 ( 70 ng/g) and greater than the C8 to C11 PFCAs (1.6 ( 0.6 to 0.17 ( 0.10 ng/g) observed in the same samples. DiPAPs were observed in paper
fiber extracts at concentrations ranging from 34 ( 30 to 2200 ( 400 ng/g. The high diPAP concentrations in WWTP sludge suggest PAP materials may be prevalent in our daily lives.
Introduction Perfluorinated carboxylic acids (PFCAs) are consistently observed in North American human serum at microgram per liter concentrations (1-7). Potential sources of this contamination include industrial production of PFCAs or the transformation of other fluorinated materials into PFCAs (7). The majority of commercial fluorochemical production involves incorporation of fluorotelomer alcohols (FTOHs) or perfluorinated sulfonamides (PFSAms) into polymers for treating textiles and carpets or surfactants for use in food contact paper applications (8). Transformation of FTOHs into PFCAs can occur atmospherically (9) or biologically as observed in microbial cultures (10, 11), liver hepatocytes (12, 13), and whole rat models (14). Humans may be exposed to FTOHs through inhalation because FTOHs are present in urban North American air at picogram per cubic meter concentrations (15, 16) or ingestion of FTOH-based surfactants applied to food use paper packaging (17, 18). Polyfluoroalkyl phosphoric acids (PAPs) (Table 1) are one version of a family of fluorinated surfactants used to greaseproof food contact paper products (17-21). Migration from paper into food and food simulants has been observed for PAP materials (17, 18), suggesting exposure to PAPs can occur via ingestion of foods in contact with PAP-treated paper. Bioavailability and biotransformation of 8:2 PAP mono- and diesters (8:2 monoPAP and 8:2 diPAP) were investigated using a rat model (22). Biotransformation into PFOA was observed in the 8:2 monoPAP and 8:2 diPAP exposed animals. A small increase in the concentration of perfluoroheptanoic acid was also observed in the 8:2 monoPAP-dosed animals. The concentration of both PFCAs peaked after 24 h postdosing. In addition, 8:2 monoPAP and 8:2 diPAP were observed in the blood of the dosed animals, demonstrating that uptake of PAPs from food is possible. These studies together suggest human exposure to PAPs through food may be an important source of PFCAs to humans. PFSAm-based PAPs were used in human food contact products starting in 1974 (2, 23), until 2002 when 3M completed the voluntary phase-out of their perfluorooctyl sulfonyl fluoride (POSF) chemistries (24). Because 3M used the N-ethyl perfluorooctane sulfonamidoethanol (NEtFOSE,
TABLE 1. Structures, Acronyms, and Description of the Congeners Monitored for the Analytes of Interest
* Corresponding author phone: (416) 978-1780; fax: (416) 9783596; e-mail:
[email protected]. † University of Toronto. ‡ Ontario Ministry of the Environment. § U.S. Environmental Protection Agency. 10.1021/es900100d CCC: $40.75
Published on Web 04/29/2009
2009 American Chemical Society
VOL. 43, NO. 12, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4589
F(CF2)8SO2N(CH2CH3)CH2CH2OH) moiety in their PAPs formulations alone, this moiety can be used as a tracer for PFSAm-based PAPs exposure. Olsen et al. (2) found the acetate adduct of NEtFOSE increased 4-fold in human blood from 1974 to 1989 (25). The authors attributed this increase to the commercialization of PFSAm-based PAPs in food packaging. A separate study of fluorinated chemicals in human blood also observed the acetate adduct of NEtFOSE in samples collected in 2001 and 2002 (4). As it is suspected to be quickly eliminated, the observation of the acetate adduct of NEtFOSE may indicate recent exposure to PAPs materials (4). Correlation between these observations in human sera to human exposure via food packaging are supported by an analysis by Tittlemier et al. (26) of Canadian food samples for PFSAms, where a precursor in the manufacture of NEtFOSE, N-ethyl perfluorooctane sulfonamide [NEtFOSA, F(CF2)8SO2N(CH2CH3)], was detected in a variety of food samples, with the highest concentrations observed in fast food items. After the phase-out of the POSF-chemistries by 3M (24), FTOH-based products assumed the vacated market share (27, 28). This shift in fluorochemical production has been reflected in temporal analyses of human sera (3, 6). Calafat et al. (6) compared human sera samples collected in the United States in 1999-2000 with those collected in 2003-2004 and found a 32% decrease in PFOS concentrations and a 25% decrease in PFOA concentrations. Olsen et al. (3) analyzed samples collected in the United States in 2000-2002 and 2006 and observed a 60% decrease in PFOS concentrations and a 25% decrease PFOA concentrations over this time period. As PFOA (t1/2 ) 3.5 years) has a shorter serum elimination half-life in humans than PFOS (t1/2 ) 4.8 years) (29), the smaller percentage decline observed for PFOA in both studies suggests there is an additional and current source of PFOA to the human population aside from the historic burden. Human exposure to FTOH-based materials is supported by the isomer profile of PFOA in 16 human sera samples collected from the midwestern United States in 2004 and 2005 (7). The concentration of PFCAs in human sera is controlled by direct exposure to PFCAs and exposure to PFCA precursor compounds. Extrapolation of human PFCA exposure via the concentrations of PFCAs and precursor compounds in relevant media (air, water, food, etc.) has been attempted (30-32); however, accumulating all of the data required is time and resource intensive, given the large number of samples and analytes. To specifically interrogate whether the North American population was exposed to PAPs materials, we determined the concentration of the 4:2 through 10:2 diPAPs in twenty pooled human sera samples from the midwestern United States collected either in 2004 and 2005 (n ) 10) or 2008 (n ) 10). Wastewater treatment plant (WWTP) sludge and paper fiber samples were also analyzed for the suite of diPAP congeners as these media may reflect human exposure. DiPAPs are a prevalent commercial product (8, 19-21) and a known PFCA precursor (22). The observation of diPAPs in human sera would be direct evidence of human exposure to PFCA precursors via PAPs-based commercial materials.
Experimental Section Chemicals. See the Supporting Information for a complete list of the chemicals used. DiPAPs (y ) x only) were synthesized as described in D’eon and Mabury (22). Human Sera Samples. Human sera samples were obtained from Golden West Biologicals (Temecula, CA). The samples were collected across the midwestern United States from donors varying in age (18-70 years old), gender, and blood type. Each sample pool consisted of at least 10 individual donors, with no overlap in donors between pooled 4590
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 12, 2009
samples. Ten pooled samples were obtained in June 2005 from samples collected in 2004 and 2005, and an additional ten pooled samples were obtained in June 2008 from samples collected in 2008. Similar pooled human sera samples have previously been used to obtain representative populationbased estimates of human PFCA contamination (7). The use of pooled samples has the advantages of reduced analytical costs and lower biosafety risks associated with using samples from a commercial supplier that screens for hepatitis and HIV. It must be noted that the analysis of pooled human sera samples does not provide information on the contaminant concentrations present in single individuals, but rather an estimate of the overall contamination present in the population. One human sera sample pooled from 1000-1500 males with type AB blood was purchased from Sigma Aldrich (Oakville, ON) for use as a recovery matrix. All human sera samples were stored at -20 °C prior to extraction. Wastewater Treatment Plant Sludge and Paper Fiber Samples. Extracts (n ) 1) from six WWTP sludge samples (samples 1-6 in the Supporting Information) and four paper fiber samples (samples 1-4 in the Supporting Information) analyzed in this study were initially collected and extracted for purposes of government monitoring programs for PFCAs and PFSAs. Sludge samples had been collected from WWTPs across Ontario in 2002 and paper fiber from paper mills in Ontario in 2002 and 2003. A subsample of dried sludge from WWTP sample 3 was acquired for use as a recovery matrix. An additional paper fiber sample (sample 5 in the Supporting Information) was collected from a paper mill in Ontario in 2008. The National Institute of Standards and Technology (NIST) WWTP sludge standard reference material (SRM) 2781 (sample 7 in the Supporting Information) was also included in the analysis. Extractions. The 2004 and 2005 human sera samples and the WWTP sludge and paper fiber samples were extracted using a modified version of the ion-pairing method developed by Hansen et al. (1). Briefly, 4 mL of a 0.25 M sodium carbonate buffer and 1 mL of a 0.5 M tetrabutyl ammonium hydrogen sulfate (TBAS) solution, adjusted to pH 10, was added to 1 mL of sera or 2 g of sludge or paper fiber. The samples were then extracted with two 5 mL aliquots of methyl tert-butyl ether (MTBE). These MTBE aliquots were combined, brought to dryness under nitrogen, reconstituted in 0.5 mL of methanol, and filtered using a 0.2 µm nylon filter. Three subsamples were extracted for each human sera sample, and one extraction blank (HPLC grade water) was extracted with each sample. One extract was obtained for the WWTP sludge samples 1-6 and for the paper fiber samples 1-4 as well as one extraction blank (Ottawa sand). Paper fiber sample 5 and WWTP sludge sample 7 were extracted in triplicate, with one extraction blank each (HPLC water). Contamination is always a concern when analyzing for perfluorinated acids. Unfortunately, 8:2 diPAP was consistently observed in the extraction blanks of the 2004 and 2005 human sera samples. During this analysis any lab specific contamination was eliminated by performing the extractions in a newly renovated lab, where PAPs had never been specifically analyzed or used. Despite all attempts to limit potential sources of contamination, low-level 8:2 diPAP contamination could not be eliminated from the analysis of the 2004 and 2005 human sera samples. However, upon further investigation 8:2 diPAP contamination could be controlled by removing the sodium carbonate buffer from the ion-pairing extraction method. A new extraction method was designed, which involved adding 0.5 mL of TBAS solution to 2 mL of sera, and continuing with the liquid-liquid extraction as described above. Extraction efficiencies were similar between the two methods (Supporting Information). This improved extraction method was used to analyze the
2008 human sera samples. Unfortunately, there was not enough sample remaining for the 2004 and 2005 human sera samples to be re-extracted using this improved method. Instrumental Analysis. All samples were analyzed by LCMS/MS using an API 4000 mass spectrometer (Applied Biosystems/MDS Sciex) coupled to an Agilent 1100 LC system. Two LC-MS/MS methods were used here, and the fundamental difference between these two methods was the LC column used. The 2004 and 2005 human sera samples, WWTP sludge samples 1-6, and paper fiber samples 1 and 2 were analyzed using a Gemini C18 LC column (50 mm × 4.6 mm, 3 µm; Phenomenex, Torrance, CA). The 2008 human sera samples, WWTP sludge sample 7, and paper fiber samples 3-5 were analyzed using an Ascentis Express C18 LC column (50 mm × 4.6 mm, 2.7 µm; Supelco, Sigma Aldrich, Oakville, ON). Instrumental details including multiple reaction monitoring (MRM) transitions and LC gradients are provided in the Supporting Information. Analysis of four human sera samples using both LC columns gave similar values (Supporting Information). As no standards were available or synthesized for the mixed diPAPs (y ) x + 2), we inferred analytical parameters and MRM transitions from adjacent diPAPs with synthesized standards (Supporting Information). Quality Control. The C8 to C11 PFCAs and PFOS were quantified using the following mass-labeled internal standards: 13C4-PFOA (>99%), 13C5-PFNA (>99%), 13C2-PFDA (>99%), 13C2-PFUnA (>99%), and 13C4-PFOS (>99%). DiPAPs (x ) y only) were quantified by standard addition, a quantification technique demonstrated to be appropriate for the analysis of fluorinated acids in the absence of masslabeled internal standards (33). Mixed diPAPs (y ) x + 2) were quantified using the standard additions of the adjacent diPAPs as matrix-matched standards (Supporting Information). Values are presented using the arithmetic mean and standard error. Full data sets are provided in the Supporting Information. The average per sample relative standard error for the twenty human sera samples (n ) 3) was
