Environ. Sci. Technol. 2005, 39, 5177-5182
Polybrominated Diphenyl Ethers and Polychlorinated Biphenyls in Human Adipose Tissue from New York
America, although these efforts do not address continuing emissions from existing sources, such as polyurethane foams.
B O R I S J O H N S O N - R E S T R E P O , †,| K U R U N T H A C H A L A M K A N N A N , * ,† DAVID P. RAPAPORT,‡ AND BRUCE D. RODAN§
Polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs) are widespread environmental contaminants. Chemical stability and lipophilicity of PCBs and PBDEs, in addition to resistance to metabolism, cause these compounds to persist in the environment. They have been detected in human tissues as well as in breast milk (1-6). Whereas body burdens of PCBs have been declining in most industrialized nations (7, 8) following the ban on production, concentrations of PBDEs have increased over the past 20 years (1, 5, 9, 10, 11). PBDEs have been used as flame retardants in various consumer products and electronics. Although the manufacture of commercial penta- and octaBDE mixtures has recently ceased in the United States (69 FR 70404, December 2004), amounts contained in existing consumer products are likely to contribute to environmental loadings for the foreseeable future. Most of the studies on PBDEs in humans have focused on northern and western Europe and Canada. Few data are available on the levels of PBDEs in the U.S. population. PBDEs have been determined in blood serum samples (8, 12, 13), breast tissue samples (14, 15), and breast milk (5) collected from the U.S. population. Concentrations of PBDEs in the U.S. population were higher than those reported for other countries, consistent with the elevated usage of PBDEs in North America (2, 5, 11). Human milk and breast tissue samples collected from female donors in San Francisco, California, had 6- to 10-fold higher PBDE levels than did samples from European populations (1, 11, 14). PBDE levels in human tissues have increased exponentially, by a factor of ∼100, during the last 30 years, with a doubling time of ∼5 years (11). Studies of PBDEs in human tissues from the U.S. population have been limited thus far to samples collected from California, Texas, Illinois, and Indiana (12-15). These measurements were generally performed on blood sera or breast milk, raising concerns that residue levels of lipophilic contaminants in these matrixes may fluctuate with surges in blood lipids (16, 17). In contrast, this study measures contaminant residues in adipose tissue, providing steadystate concentrations that integrate levels of lipophilic chemicals such as PBDEs accumulated over time. We report the occurrence of elevated concentrations of PBDEs in human adipose tissue collected from New York City. Concentrations of PCBs were also analyzed to allow comparison of residue levels of these two classes of halogenated compounds in human adipose tissues from this city.
Wadsworth Center, New York State Department of Health and Department of Environmental Health Sciences, State University of New York at Albany, Empire State Plaza, P.O. Box 509, Albany, New York 12201-0509, Plastic Surgery, Lipoworks, 580 Park Avenue, New York, New York, U.S. Environmental Protection Agency, Office of Research and Development, Washington, DC, 20460
Human adipose tissue samples (n ) 52) collected in New York City during 2003-2004 were analyzed for the presence of polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs). Concentrations of PBDEs in adipose tissues ranged from 17 to 9630 ng/g, lipid wt (median: 77; mean: 399 ng/g, lipid wt; sum all di- through hexaBDE congeners). Average PBDE concentrations in human adipose tissues from New York City were 10- to 100times greater than those reported for European countries. A concentration of 9630 ng/g, lipid wt, found in a sample of adipose tissue, is one of the highest concentrations reported to date. PBDE 47 (2,2′,4,4′-tetraBDE) was the major congener detected in human tissues, followed by PBDE congeners #99 (2,2′,4,4′,5-pentaBDE), 100 (2,2′,4,4′,6pentaBDE), and 153 (2,2′,4,4′,5,5′-hexaBDE). A few individuals contained PBDE 153 as the predominant congener in total PBDE concentrations, suggesting alternative exposure sources, possibly occupational. Principal component analysis of PBDE congener composition in human adipose tissues revealed the presence of five clusters, each characterized by varying composition. No significant difference was found in the concentrations of PBDEs between gender. Concentrations of PBDEs were, on average, similar to those for PCBs in human adipose tissues, and substantially higher when PBDE outliers were retained. PBDE and PCB concentrations were not correlated. PBDE concentrations did not increase with increasing age of the subjects, whereas concentrations of PCBs increased with increasing age in males but not in females in this study. These results suggest differences between PBDEs and PCBs in their sources or time course of exposure and disposition. The presence of comparable or greater concentrations of PBDEs, relative to PCBs, highlights the importance of recent voluntary and regulatory efforts to cease production of commercial penta- and octa-BDE in North * Corresponding author phone: (518)474-0015; fax: (518)473-2895; e-mail:
[email protected]. † Wadsworth Center and State University of New York at Albany. ‡ Plastic Surgery. § U.S. Environmental Protection Agency. | Present address: Environmental and Computational Chemistry Group, Universidad of Cartagena, Cartagena, Colombia. 10.1021/es050399x CCC: $30.25 Published on Web 05/21/2005
2005 American Chemical Society
Introduction
Materials and Methods Sampling. Institutional Review Board (IRB) approval was obtained from the New York State Department of Health for the analysis of adipose tissue samples from humans. Adipose fat tissue samples (n ) 52) were collected during October 2003-October 2004 from an office based surgery center in New York City. Adipose fat tissue samples had been collected from patients who underwent liposuction procedures. The samples were devoid of personal identifiers. The only known demographic factors were age, gender, ethnicity, and date of collection. Samples were stored in precleaned glass bottles at -20 °C until analysis. Chemical Analysis. PCB and PBDE congeners were analyzed following the method described elsewhere with VOL. 39, NO. 14, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5177
TABLE 1. Demographic Characteristics and PBDE and PCB Concentrations in Human Adipose Tissue Samples from New York City ∑PBDE (ng/g, lipid wt)
Donor Age (yr) 31.7 ( 8a: 31.5 (18-51)b 40 Gender female male Ethnicity African American Asian Caucasian Hispanic a
Mean ( SD.
b
mean ( SD
median (range)
mean ( SD
median (range)
52 22 (42%) 22 (42% 8 (16%)
399 ( 1410a 331 ( 838 579 ( 1980 90 ( 64
77.3 (17-9630)b 81 (26-4060) 75 (20-9630) 75 (17-250)
144 ( 138a 128 ( 82 189 ( 182 62 ( 43
110 (18.9-816)b 110 (47-356) 135 (36.2-816) 40 (18.9-149)
40 (77%) 12 (23%)
253 ( 639 885 ( 2640
85.7 (20-4060) 68.7 (17-9630)
157 ( 152 98 ( 50
127 (18.9-816) 95 (40-230)
9 (17%) 1 (2%) 35 (67%) 7 (14%)
100 ( 75 61.5 518 (1700 236 ( 322
73 (17-296) 61.5 80 (26-9630) 53 (20-961)
144 ( 85 816 132 ( 104 61 ( 19
134 (25.6-356) 816 117 (26-567) 72 (18.9-356)
Mean (range).
some modifications (18, 19). Approximately 5 g of the adipose fat tissue samples was homogenized with anhydrous sodium sulfate and extracted in a Soxhlet apparatus for 16 h using dichloromethane and hexane (3:1; 400 mL). The extract was rotary-evaporated at 40 °C to 11 mL, and an aliquot of 1 mL of the extract was used for the determination of fat content by gravimetry. The remaining extract was spiked with 13Clabeled PCB congeners 3, 15, 31, 52, 118, 153, 180, 194, 206, 209, and 13C-labeled PBDE congeners 3, 15, 28, 47, 99, 100, 118, and 153 as internal standards. PCB congener 30 (2,4,6triCB) was spiked as a surrogate standard. Sample extracts were then purified by passage through a series of layers of silica gel (Davisil, 100-200 mesh, Aldrich, Milwaukee, WI) in the following order: 1 g of silica gel, 2 g of 40% acidic-silica gel, 1 g of silica gel, 2 g of 40% acidic-silica gel, and 1 g of silica gel at the top. The silica gel layer was cleaned by passage of 50 mL of hexane prior to the transfer of the sample extracts. Samples were then eluted with 150 mL of 15% dichloromethane in hexane and rotary-evaporated to 10 mL. The extract was then treated with concentrated sulfuric acid (5 mL) and passed through a glass column (10 mm i.d.) packed with 1 g of silica gel-impregnated active carbon (Wako Pure Chemical Industries, Tokyo, Japan). PCBs and PBDEs were eluted with 15% dichloromethane in hexane (150 mL) and then concentrated to 1 mL. Identification and Quantification. Extracts were injected into a gas chromatograph (Hewlett-Packed 6890) coupled with a mass selective detector (MS) (Hewlett-Packed, series 5973) for the determination of PCBs and PBDEs. A capillary column coated with RTX-5MS (30 m × 0.25 mm i.d. × 0.25 µm film thickness; Restek Corp, Bellefonte, PA) was used for the separation of individual isomers. The column oven temperature was programmed from 100 °C (1 min) to 160 °C (3 min) at a rate of 4 °C/min, and then to 250 °C at 3 °C/min, with a final hold time of 5 min for PCBs. For PBDEs, the column temperature was programmed from 100 °C (1 min) to 160 °C (3 min) at a rate of 10 °C/min, and then to 260 °C at 2 °C/min, with a final hold time of 5 min. The MS was operated in an electron impact (70 eV), selected ion monitoring mode. PCB congeners were monitored at the two most intense ions of the molecular ion cluster. An equivalent mixture of Kanechlor (KC 300, 400, 500, and 600) with known PCB composition was used in the identification of PCB congeners (20). Quantification of PCB congeners was based on external calibration standards containing known concentrations of di- through deca-CB congeners. Concentrations of individually resolved peaks of PCB isomers were summed to obtain total PCB concentrations. Total PCBs represent the sum of all tri- through deca-chlorobiphenyl congeners. PBDE congeners were monitored at the molecular ion clusters [M]+ and [M + 2]+. Eight major PBDE congeners 5178
9
∑PCB (ng/g, lipid wt)
N (%)
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 14, 2005
28, 30, 47, 85, 99, 100, 153, and 154 and three unidentified di-, tri-, and penta-BDE congeners, which were also found in our tissue samples, were quantified in this study. Total PBDE concentrations represent the sum of all di- through hexa-BDE congeners. PBDE congeners were quantified using an external calibration standard. PCB and PBDE congeners are identified by their IUPAC numbers. Quality Assurance/Quality Control. The extraction, cleanup, and fractionation steps were evaluated by measurement of the absolute recoveries of the compounds spiked and passed through the entire analytical procedure. Mean ((standard deviation) recoveries of 13C-labeled PCB congeners 30, 118, 153, and 194 spiked into the samples were 77 ( 14%, 77 ( 17%, 75 ( 12%, and 75 ( 14%, respectively. Overall recoveries of PCB congeners, including PCB 30, ranged from 60 to 118%. Mean ((standard deviation) recoveries of 13C-labeled PBDE congeners 28 and 47 were 92 ( 14% and 91 ( 14%, respectively. Overall recoveries of PBDEs ranged from 61 to 123%. The reported concentrations were not corrected for recoveries. The matrix-spike studies suggest that the data can be considered accurate to within one standard deviation (i.e., (15%) of the fortified sample recovery. Procedural blanks were analyzed for every set of 10 samples, as a check for interferences. Calculated concentrations were reported as less than the limit of detection, either if the observed isotope ratio was not within (20% of the theoretical ratio, or if the peak area was not greater than the specified threshold (3 times the noise). PBDE congeners 47 and 99 were found in blanks at 20 pg/g, on a wet wt basis. The detection limits of individual PCB and PBDE congeners varied from 40 to 400 pg/g, wet wt.
Results and Discussion Age of the subjects, on average, was 31 years (range: 18-51 years; Table 1). The subjects were 77% females and 23% males. The overall ethnic distribution was 67% Caucasian, 17% African-American, 14% Hispanic, and 2% Asian. Lipid content of the adipose tissue samples ranged from 37% to 89% (mean: 58%). Eleven PBDE congeners (BDE-28, 30, 47, 85, 99, 100, 153, 154, and unidentified di-, tri-, and penta-BDEs) were identified in human adipose tissue samples. Three of the congeners were unidentified (U-diBDE, U-triBDE, and U-pentaBDE) because of the lack of pure reference standards, although the fragment ions and their ratios were characteristic of brominated diphenyl ethers. We also examined the occurrence of deca-BDE (qualitatively), but it was not detected in the majority of samples. Concentrations of ∑PBDE in adipose tissues ranged from 17.4 to 9630, ng/g lipid wt (median: 77.3; mean: 399 ng/g, lipid wt; Table 1; Figure 1). These con-
by the predominant (>90% of global production) use of penta-BDE mixture in North America (11).
FIGURE 1. Frequency distributions of concentrations of PBDEs and PCBs in human adipose tissue samples from New York City (two outliers were excluded for PBDEs). centrations are 10- to 100-fold greater than the concentrations reported for human adipose tissue collected from several European countries (6). Mean ∑PBDE concentrations in human adipose tissue collected in the mid- to late-1990s from Belgium, Spain, Finland, and Sweden ranged from 4 to 13 ng/g, lipid wt (2, 6). Concentrations of PBDEs in adipose tissue from Japan collected in 2000 ranged from 0.5 to 2.8 ng/g, (median 1.3) lipid wt (10). An earlier study determined PBDE concentrations in human breast tissue, collected from California as ranging from 17 to 462 ng/g, lipid wt, with a mean value of 86 ng/g, lipid wt (14). Concentrations of PBDEs in human milk from the United States were 10 to 100-times greater than the concentrations reported previously for breast milk from Europe (5). When we compared our data with data reported for adipose tissue samples from European countries, we found average concentrations of PBDEs in our samples to be 10- to 100-fold greater than those reported for the European population. A concentration of 9630 ng/g, lipid wt, found in a 32-year old male subject, was the highest concentration ever reported in the literature. The next highest concentration, 4060 ng/g, lipid wt, was found in adipose tissue of a 23-year old female donor. The high concentrations of PBDEs found in these two individuals qualified as outliers, since they were >4-times the standard deviation of the mean concentration. When these two outlier samples were excluded, the mean concentration of PBDEs in adipose tissue samples from New York City was 141 ng/g, lipid wt. However, data for these two individuals were confirmed by repeated analysis. High concentrations of PBDEs in the U.S. population, as compared to European populations, can be explained
PBDE concentrations in human adipose tissue were not correlated with age, either in a combined analysis or when examined by gender (Figure 2; outliers excluded). A similar lack of correlation was found in human breast tissue collected from San Francisco, California (14). When we grouped the PBDE concentrations into three categories, namely, 40 yr, individuals 40 yr (third group; Table 1). This finding is tempered, however, by the low number (n ) 8) in the >40 yr age group and our inability to adjust for confounding factors pertinent to the participants in this study. Several earlier studies have reported the absence of agedependent accumulation of PBDEs in humans (6, 8, 21-24). The lack of age-dependency for PBDE accumulation differs from the pattern that has been observed for persistent organic pollutants such as PCBs (23). In this study, too, a significant age-related increase in PCB concentrations was found for men. In women, the age-dependent increase in PCB concentrations was not statistically significant (Figure 2), a finding that has been noted previously and related to lactational transfer. The absence of an age-related increase in PBDE concentrations compared to PCBs may be explained by the relatively recent introduction of PBDEs (less than 30 years ago), different exposure pathways, and the greater metabolism and elimination of PBDEs than of PCBs. No correlation was found between PBDE and PCB concentrations in human adipose tissue samples (Figure 3). This may suggest differences in the sources of human exposures to PCBs and PBDEs or their differing pharmacokinetics in human body. The production and use of PCBs have been regulated for three decades, whereas PBDEs are still used in several types of consumer goods and are emitted to the environment. Major sources of PCBs are expected to be dietary, particularly from fish and meat. Exposure pathways for PBDEs are not as well established, with diet and inhalation both cited as potential sources (5, 25-27), including uptake from house dust (28, 29). PBDE 47 was the prevalent congener, accounting for, on average, 33% of the total PBDE concentrations in human tissues, followed by PBDE 153 (23%), 99 (18%), and 100 (17%)
FIGURE 2. Relationship between age and PBDE (closed circles) or PCB (triangles) concentrations in human adipose tissue from New York City. In the top panel, two PBDE outliers were excluded. The bottom two panels show PCB concentrations (all closed circles) versus age, stratified by gender. VOL. 39, NO. 14, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5179
FIGURE 3. Relationship between total PCB and total PBDE concentrations in human adipose tissue from New York City (two PBDE outliers excluded).
TABLE 2. PBDE Congener Concentrations (ng/g, Lipid Wt) in Human Adipose Tissue Samples from New York City PBDE congeners
mean
median
range
fat content (%) U2BDE a 30 (2,4,6-tri-BDE) 28 (2,4,4-tri-BDE) U3BDEb 47 (2,2′,4,4′-tetra-BDE) 100 (2,2′,4,4′,6-penta-BDE) 99 (2,2′,4,4′,5-penta-BDE) 85 (2,2′,3,4,4′-penta-BDE) U5BDEc 154 (2,2′,4,4′,5,6-hexa-BDE) 153 (2,2′,4,4′,5,5′-hexa-BDE) ∑PBDE
58.3 6.4 0.6 3.3 14.6 132 67.7 74.4 6.9 7.0 8.3 91.8 398
58.4 4.6 0.7 1.9 12.1 29.3 12.0 10.3
