Environ. Sci. Technol. 2010, 44, 9298–9303
Dechlorane Plus in Human Hair from an E-Waste Recycling Area in South China: Comparison with Dust J I N G Z H E N G , † J I N G W A N G , ‡,§ X I A O - J U N L U O , * ,‡ M I T I A N , ‡,§ LUO-YIYI HE,† JIAN-GANG YUAN,† B I - X I A N M A I , ‡ A N D Z H O N G - Y I Y A N G * ,† State Key Laboratory for Biocontrol, School of Life Sciences, Sun Yat-Sen University, Guangzhou 510275, China, State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China, and Graduate University of Chinese Academy of Science, Beijing 100049, China
Received September 10, 2010. Revised manuscript received November 8, 2010. Accepted November 9, 2010.
Dechlorane Plus (DP) and a dechlorination product, 1,6,7,8,9, 14,15,16,17,17,18-octadeca-7,15-diene (anti-Cl11-DP), were measured in human hair and indoor dust collected from an e-waste recycling area and two control areas (rural and urban) in South China. DP was detected in hair and dust samples at concentrations ranging from 0.02-58.32 ng/g and 2.78-4197 ng/ g, respectively. anti-Cl11-DP, mainly detected in human hair and dust samples from the e-waste recycling area, ranged from nd (nondetected) to 0.23 ng/g in hair and from nd to 20.22 ng/g in dust. Average values of anti-DP fractional abundance (fanti ratio) in hair of e-waste dismantling workers (0.55 ( 0.11) and dust from e-waste recycling workshops (0.54 ( 0.15) were significantly lower than those in other groups (0.62-0.76 means for hair and 0.66-0.76 means for dust). Significantly positive correlation between DP concentrations in dust and hair and similarity in fanti ratios between hair and dust suggest that ingestion of dust comprise one of the major routes for DP exposure. Significantly positive relationships were also observed between anti-Cl11-DP and anti-DP for both hair and dust samples with similar regression line slopes. The ratios of anti-Cl11-DP to antiDP between hair and dust show no significant difference. These results suggest that anti-Cl11-DP in the human body is likely accumulated from the environmental matrix and not formed from biotransformation of the parent DP.
Introduction Dechlorane Plus (DP, C18H12Cl12) is an additive chlorinated flame retardant. It has been used in electrical wire and cable coatings, computer and televisions connectors, and plastic roofing materials for over 40 years (1). Technical DP mixture comprises primarily of two isomers, the syn- and antiisomers. Its annual production is estimated to be as high as 10 million pounds (1). It is thus considered to be a high production volume (HPV) chemical. Although there are no * Corresponding author phone: +86-20-85290146; fax: +86-2085290706; e-mail:
[email protected] (X.-J.L.); phone: +86-2084112008; fax: +86-20-84112008; e-mail:
[email protected] (Z.-Y.Y.). † Sun Yat-Sen University. ‡ Guangzhou Institute of Geochemistry. § Graduate University of Chinese Academy of Science. 9298
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 24, 2010
regulations for the use and production of DP, it is subject to the HPV challenge of the United States Environmental Protection Agency (U.S. EPA). It is also listed on Canada’s Domestic Substances List (2). Many recent reports have demonstrated the presence of DP isomers in numerous environmental compartments worldwide, including sediment, air, dust, tree barks, fish, bird, and human beings (3-12). Their concentrations in sediment samples from some locations in the Great Lakes region exceed those of the widely used brominated flame retardants (BFRs) known as polybrominated diphenyl ethers (PBDEs) (13). Several studies have also shown that both isomers of DP can be biomagnified in some aquatic food webs (9, 14). Recently, Ren et al. (11) detected a dechlorination product of DP in human serum, which raises the question as to whether or not biotransformation of DP occurred in biota. Currently, little information is available on this issue. Human blood, human adipose tissue, and human milk have been used to monitor human exposure to halogenated flame retardants in most studies. There are some limitations, however, associated with sampling of these biological samples in humans. For example, human milk is available only in lactating women, and blood and adipose tissue samples must be collected by a medical professional personnel. Hair, which originates from the hair follicle, is a metabolic end product. It is another human sample that provides a method for noninvasive sampling with additional merits, such as a stable matrix and ease of collection as well as short- and long-term exposure tracings (15). Several studies have shown that significant relationships exist between concentrations of halogenated compounds in hair and internal tissues (16, 17). Altshul et al. (16) observed significant correlations between hair and the serum of 10 people for p,p′-DDE (r ) 0.8) (16). A positive relationship between BFRs levels in hair and internal tissues for sum PBDEs and BDE 47 (0.37 < r < 0.78) was found in mammals (17). Hence, using hair as a bioindicator of endogenous tissue exposed to halogenated compounds is promising. Dust is an important carrier for many contaminations, especially for those with low volatility. House dust has been identified as the primary exposure pathway of PBDEs (18, 19). An EPA review has concluded that 82% of PBDE exposure is from incidental ingestion and dermal contact with house dust (18). Studies have also shown that dust PBDE levels are correlated with serum (19) and breast milk (20). DP and BDE209 have similar physicochemical properties, with a high molecular weight and high log Kow (9.3). Thus, it may be possible to evaluate DP exposure in humans through dust. DP has been detected in house dust from Ottawa Canada (8). However, no study has been conducted to investigate relationships between DP concentrations in both indoor dust and humans. In the present study, human hair samples from occupationally exposed workers, nonoccupationally exposed residents in an e-waste recycling area, and residents from rural and urban control areas were collected. Dust samples in corresponding indoor environments, such as e-waste recycling workshops, residences in e-waste recycling area, and residences in two control areas, were also collected. The concentrations of DP and a dechlorination product of DP, 1,6,7,8,9,14,15,16,17,17,18-octadeca-7,15-diene (anti-Cl11DP), were determined in hair and dust samples. The objectives of this study are (a) to investigate DP and its dechlorination product concentrations in hair and dust samples in different areas and (b) to assess the role played 10.1021/es103105x
2010 American Chemical Society
Published on Web 11/24/2010
by dust in human DP exposure by examining the relationship between DP and its dechlorination derivatives in hair and dust.
Materials and Methods Chemicals. anti-DP, syn-DP, and 1,6,7,8,9,14,15,16,17,17,18octadeca-7,15-diene (anti-Cl11-DP, lot no. a-Cl11DP 0708) standards were purchased from Wellington Laboratories (Ontario, Canada). BDE118, BDE128, BDE77, and BDE181 were obtained from Ultra Scientific (North Kingdom, RI). 13 C-PCB 208 was obtained from Cambridge Isotope Laboratories (Andover, MA). Sample Collection. Hair samples were collected from volunteer participants in an e-waste recycling area located in Longtang Town, Qingyuan County, using stainless steel scissors during their routine haircut sessions. A detailed description of the sample area has been given in a previous report (21). All the participants gave permission for their participation; consent from parents of children was also obtained after being clearly informed of the objectives of this study. A short questionnaire was completed for each participant covering age, gender, occupation (related to e-waste recycling or not), and place of residence. Dyed or bleached hair samples were excluded. In all, 30 and 82 hair samples were obtained from e-waste dismantling workers and nonoccupationally residents in the e-waste area, respectively. In addition, 32 hair samples were collected from residents living a rural area (Yuangtan Town), and 29 samples were obtained from residents in Guangzhou City, the largest urban center in South China. The same method was used for sample collection, which was done in the same period. Yuangtan Town, located 15 Km northeast of Longtang Town, is an agricultural area with a ceramic industry and no e-waste relative activities. Guangzhou, located 60 Km south of Longtang, is the capital of Guangdong Province; it is densely populated and has a rapidly developing economy. The locations of sampling areas are shown in Figure S1 in the Supporting Information. Dust samples were collected using the method described in a previous study (22). For this study, indoor dust samples from 13 e-waste recycling workshops were collected to represent the working environment of occupationally exposed workers. Ten residences in the e-waste recycling area, 10 residences in the rural control area, and 27 residences in the urban control area were selected for dust sampling to reflect the indoor living environment of hair sample donors. Dust samples in the e-waste recycling workshops were collected from the surface of floors, tables, and windowsills in the work place, using solvent-cleaned brushes. For residences, dust samples were obtained from the surface of furniture, tables, windowsills, and the floors of living rooms and bedrooms. Hair and dust samples were wrapped in aluminum foil and sealed in polyethylene zip bags before being transported to the laboratory where the samples were kept at -20 °C until chemical analysis. DP Extraction and Analysis. Hair sample preparation and extraction methods were similar to previous studies (17, 23). with some modifications. Briefly, individual hair samples were incubated in Milli-Q water in a shaking incubator (1 h, 40 °C) twice to remove external contamination (e.g., fine soil particles, dust). After rinsing with Milli-Q water, samples were freeze-dried, cut into small pieces (2-3 mm), and then thoroughly mixed. Approximately 2 g of hair samples was spiked with BDE77 and 181 and incubated overnight (12 h) at 40 °C with 40 mL of hydrochloric acid (4 M) and 40 mL of hexane/dichloromethane (4:1, v/v). Extraction of target analytes from the incubation medium was done by a liquid-liquid extraction (LLE) with 3 × 40 mL of hexane/ dichloromethane (4:1, v/v). The extracts were further cleaned by a multilayer silica/alumina column and finally condensed
to 50 µL under a gentle stream of N2. Known amounts of internal standards (13C-PCB 208, BDE 118 and 128) were added before instrument analysis. Dust samples were sieved through a stainless steel sieve (500 µm) to remove large debris before analysis. The samples (0.5-2 g) were spiked with internal standards and Soxhlet extracted with a mixture of acetone and hexane (1:1, v/v) for 48 h. Clean-up procedures are the same in processing hair samples. The target compounds were analyzed by a Shimadzu 2010 gas chromatograph coupled with a mass spectrometer (GCMS) with electron capture negative ionization (ECNI) in the selected ion monitoring mode. A DB-5HT (15 m × 0.25 mm i.d., 0.1 µm film thickness) capillary column was used. The GC oven temperature program had the following parameters: 110 °C (held for 2 min), raised to 280 at 20 °C/min (held for 3 min), and finally at 310 at 30 °C/min (held for 15 min). Methane was used as a chemical ionization moderating gas at an ion source pressure of 2.4 × 103 Pa. Helium was the carrier gas at a flow rate 1 mL/min. Injection of 1 µL sample was conducted with an automatic sampler in the splitless mode. The injection port temperature was set at 280 °C. Ion source and interface temperatures were set to 200 and 290 °C, respectively. Ion fragments m/z 653.8 and 651.8 were monitored for DPs; m/z 618.0 and 620 were monitored for a-Cl11-DP; m/z 79 and 81 were monitored for BDE77 and 81; and m/z 476 and 478 were used for 13C-PCB208. QA/QC. A procedural blank and a spiked blank were run with each batch of samples. Procedural blanks contained no target analytes. The recoveries of surrogate standards were in the range of 87-124% and 72-110% for BDE77 and BDE181, respectively. Recoveries of syn-DP and anti-DP in spiked blanks for hair extraction were between 91-98% and 93-96%, respectively. For dust extraction, the recoveries of syn-DP and anti-DP ranged from 77-89% and 87-112%, respectively. Repeatability of analysis was assessed by analyzing three duplicate hair samples. The relative standard deviation was within 10% for both DP isomers. The limit of quantification (LOQ) was estimated based on a signal-tonoise ratio of 10. Using an average sample mass of 2 g, the LOQs for syn-DP, anti-DP, and anti-Cl11-DP were 3.1, 2.8, and 2 pg/g, respectively. The final results were not recovery corrected. Statistical Analysis. Data analysis was performed using SPSS for Windows Release 13.0 (SPSS Inc.). The statistical significance of hair DP concentrations between the e-waste area and the control area was conducted by t test. Linear regression analysis was conducted to investigate correlations between DP in hair and dust and between anti-Cl11-DP and anti-DP in both hair and dust samples. Statistical significance was set at p < 0.05 throughout the manuscript.
Results and Discussion Levels of DP in Hair. Concentrations of ΣDP (sum of the syn- and anti-DP isomers) in all hair samples ranged from 0.02-58.3 ng/g dry weight. Levels of hair DP were in the following order: e-waste dismantling workers > nonoccupationally exposed residents living in e-waste recycling area > residents living in the rural and urban control areas (Table 1). No significant difference was found between hair DP in the rural and urban control area (p > 0.05). However, the concentrations of ΣDP in nonoccupationally exposed residents in the e-waste area were significantly higher than those in two control areas (p < 0.01). Meanwhile, occupationally exposed workers showed significantly higher DP concentrations than nonoccupationally exposed residents (p < 0.01). This distribution pattern is consistent with the expected elevation in exposure of workers in the e-waste recycling area due to the dismantling of e-waste products. VOL. 44, NO. 24, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9299
4.62 (nd-21.6) 14.3 (1.83-62.1) 18.9 (2.78-70.4) nd 0.70 (0.43-0.88) 0.22 (0.004-1.1) 0.65 (0.01-3.91) 0.87 (0.02-4.97) nd 0.74 (0.45-0.85) nd: nondetected; *: mean is not given since only two samples have detected anti-Cl11-DP. a
syn-DP anti-DP ΣDP anti-C11-DP fanti
6.86 (0.66-19.7) 8.52 (0.8-43.6) 15.4 (1.46-58.3) 0.06 (0.01-0.23) 0.55 (0.25-0.75)
723 (158-3038) 793(121-1902) 1515 (343-4197) 8.78 (1.78-20.2) 0.54 (0.28-0.77)
2.48 (0.06-11.4) 3.6 (0.13-25.4) 6.08 (0.19-35.7) 0.03 (0.004-0.17) 0.62 (0.34-0.9)
101 (19.2-475.04) 276 (26.1-1323) 377 (45.2-1798) 2.71 (nd-7.54) 0.66 (0.35-0.77)
0.19 (0.02-1.1) 0.84 (0.07-7.28) 1.03 (0.09-8.38) nd 0.76 (0.54-0.94)
15.6 (6.62-26.9) 49.3 (25.5-91.4) 64.9 (32.6-118) (nd-2.6)* 0.76 (0.68-0.8)
dust (n ) 27)
urban control area
hair (n ) 29) dust (n ) 10) hair (n ) 32) dust (n ) 10) hair (n ) 82) dust (n ) 13) hair (n ) 30)
resident area in e-waste recycling area
rural control area 9
e-waste recycling workshop
TABLE 1. Concentrations of syn-, anti-, ΣDP, anti-C11-DP (ng/g dry weight), and fanti in Hair and Dust Samplesa 9300
FIGURE 1. Boxplots of DP concentrations in the hair samples from the residents grouped by age and the e-waste dismantling workers in the e-waste recycling area.
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 24, 2010
The large sample size of hair from nonoccupationally exposed residents in the e-waste recycling area makes it possible to analyze the differences in DP concentrations between age groups (Figure 1). Significant differences were not observed between hair DP levels in children (aged below 6), adolescents (7-18), and adults (19-60). This result is in line with the observation for serum DP in another e-waste recycling area (11). However, as a whole group, seniors (aged more than 60) have much higher hair DP levels than others (Figure 1). It was reported that seniors have a relatively low hair growth rate compared with other age groups (24), which could lead to a higher accumulation of contaminants in the hair of seniors than in other age groups for identical length of hair. The proportions of anti-DP (fanti: concentration of antiDP divided by total DP) were calculated for hair samples coming from people in different areas (Table 1). The fanti of hair samples from residents in the two control areas (mean values of 0.76 and 0.74 for the rural and urban areas, respectively) were similar to the proportions of anti-DP in the commercial DP product (fanti ) 0.65-0.80) in other reports (3, 13) and results from commercial products (fanti ) 0.70 ( 0.00) manufactured in China (9). In the e-waste area, however, fanti values (a mean value of 0.55 for occupationally exposed workers and 0.62 for nonoccupationally exposed residents) were significantly lower than those in the two control areas and those in commercial products. Ren et al. have also reported that the fanti in serum from dismantling workers in an e-waste area (0.58 ( 0.11) were lower than those in the control area (0.64 ( 0.05) (11). Interestingly, the fanti values in nonoccupationally exposed residents in e-waste areas were between occupationally exposed workers and the two control areas. A combination exposure to DP released from e-waste recycling activities, along with DP from background likely contributed to this effect. Assuming that DP composition in the background environment of the e-waste area is the same as in the rural control area and that no or insignificant stereoselective metabolism occurs in humans. The contribution of e-waste recycling derived DP can be calculated by a two-end member mixture model. The calculated result is 0.65, which means that approximately two-thirds of DP in the residents living in the e-waste area come from e-waste recycling activities. Given that this is the first known measurement of DP in human hair samples, comparison with other studies for this medium is therefore impossible. Nevertheless, a previous study has evaluated DP levels in serum from electronics dismantling workers in another e-waste recycling area in South China (Guiyu) (11). In the study, the concentration of
FIGURE 2. Relationship between DP concentrations in hair and dust. (a): syn-DP and (b) anti-DP. Error bars represent (1 standard error. ΣDP was three times higher in e-waste recycling areas than in control areas. The median fanti values in serum from workers in Guiyu were 0.53 (range: 0.40-0.77) (11). This result is similar with hair samples for the occupationally exposed workers found in the current work. Levels of DP in Dust. Concentrations of DP in the dust samples collected from e-waste recycling workshops and residences in the e-waste recycling area varied from 342.79-4197 ng/g and 45.22-1798 ng/g, respectively. These values are significantly higher than those in the two control areas (32.62-118.26 ng/g for the rural area and 2.78- 70.38 ng/g for the urban area) (Table 1). This spatial distribution is similar to that identified in hair DP. The dust samples from e-waste workshops showed a mean fanti value of 0.54 (range: 0.28- 0.77), which is significantly lower (p < 0.01) than dust from residences (mean of 0.66) and from the two control areas (mean values of 0.76 and 0.70 for rural and urban areas, respectively). The fanti values in dust from the two control areas and the residences are close to the reported values in commercial DP products, implying that no remarkable stereoselective degradation of DP occurred in dust. No significant difference in the stereoisomer profile of DP between the air and the commercial mixture was also observed in other studies (3, 5). The depletion of anti-DP in dusts from e-waste recycling workshops is likely due to the photolytic or thermal degradation of DP during the use, repair, and recycling of electronic products before release into the ambient environment. To date, only a single study has reported DP in indoor dust samples (8). DP was found in dust collected in Ottawa, Canada at concentrations ranging from 2.3s182 ng/g with an outlier of 5683 ng/g. These values are in the same range as the two control areas found in this study. The median of DP (14 ng/g for 2002-2003 and 22 ng/g for 2007) in Ottawa, Canada is much lower than those in the e-waste area (153.3 for dust samples in residences and 1037 for e-waste recycling workshops) in the present study. Correlation between DP in Hair and Dust. To demonstrate correlation between DP in both dust and hair, linear regression analysis was performed for concentrations and isomer compositions of DP between hair samples and corresponding dust samples. Strong positive correlations were found for both concentrations of syn-DP and anti-DP between human hair and dust samples (Figure 2). This consistency, on one hand, suggests that hair analysis could be a valid screening tool for assessing human DP exposure. On the other hand, it implies that dust is one of the major pathways for DP exposure. Previous studies have shown that dust ingestion is an important exposure pathway for flame retardants such as PBDEs (19, 20, 25, 26). It is assumed that the residents and occupationally exposed workers in the
FIGURE 3. Relationship between fanti in hair and dust. Error bars represent (1 standard error. e-waste recycling area share the similar dietary habits since they live in same region. Dietary exposures thus could not explain the hair DP difference between e-waste recycling related workers and nonoccupationally exposed residents. From these results, deducing that dust ingestion plays an important role in determining human DP exposure in the e-waste recycling area is, therefore, reasonable. The fanti values in hair samples were also correlated with those in dust samples (Figure 3). The fanti values in hair samples (0.55 ( 0.11 for occupationally exposed workers; 0.62 ( 0.09 for nonoccupationally exposed residents in e-waste recycling areas; 0.76 ( 0.07 for residents in the rural control area; and 0.74 ( 0.07 for residents in the urban control area) were similar to those in their corresponding dust samples (0.54 ( 0.05, 0.66 ( 0.12, 0.76 ( 0.03, and 0.70 ( 0.11). The correlation of fanti values between hair and dust further supports the conclusion that dust is a major route for the human DP exposure. It is also inferred that no stereoselective enrichment of DP in human hair compared to dust. This result is dramatically different from the observation in aquatic organisms, such as fish (9) and water birds (27) in the same sampling region. In those studies, an enrichment of syn-DP has been found in aquatic organisms compared to superficial sediments. The different exposure pathway, assimilation efficiency, and metabolic capabilities for DP between human beings and aquatic organisms, and the difference between measuring DP in human hair versus in aquatic muscle, might be responsible for this difference. However, further studies are needed to clarify this issue. Dechlorination Product of DP in Hair and Dust. In previous studies, both Sverko et al. (4) and Ren et al. (11) have VOL. 44, NO. 24, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9301
FIGURE 4. Correlation of concentrations of anti-DP plotted against anti-Cl11-DP in dust (a) and hair (b) samples from the e-waste recycling area. detected a dechlorination DP species with retention time between syn-DP and anti-DP in the sediments from the lower Great Lakes and in serum samples from electronic dismantling workers, respectively. This compound has been tentatively identified as a -1 Cl dechlorination product of DP (4). This compound was also found to be present in the dust and hair samples in the present study. Comparing GC retention times and full scan mass spectra between samples and standards (detailed information given in Figures S2 and S3 in the Supporting Information), we finally confirmed that the suspected dechlorination product of DP is 1,6,7,8,9,14,15,16,17,17,18-octadeca-7,15diene, briefly termed as anti-Cl11-DP. anti-Cl11-DP was detected in 93% hair samples from the electronics dismantling workers and 83% hair samples from nonoccupational residents, whereas no anti-Cl11-DP was detected in hair samples from the two control areas. As for dust samples, anti-Cl11-DP was detected in 100% and 70% of samples from e-waste recycling workshops and residences in the e-waste recycling areas, respectively. Only two dust samples from the control areas have detectable anti-Cl11DP. The distributions of anti-Cl11-DP in human hair and dust samples were consistent with that of DP. Relatively high concentrations of anti-Cl11-DP were found in hair from e-waste dismantling workers and dust from e-waste recycling workshops (Table 1). No significant differences were found in anti-Cl11-DP concentrations between different age groups. Meanwhile, 100% detection frequency and high concentration of anti-Cl11-DP in dust from the e-waste recycling workshop indicate that anti-DP dechlorination could occur during e-waste recycling. This can partially explain the low values of fanti in human hair and dust samples from the e-waste recycling area. To determine whether Cl11-DP was the degradation product of anti-DP from injector breakdown (4), we check the system by regular injection of a commercial DP solution. Trace amounts of Cl11-DP were present in the commercial solution, indicating that in situ linear dechlorination of DP does occur and/or the commercial mixture used in our study may contain trace amounts of low chlorination DP species. However, the proportion of anti-Cl11 (anti-Cl11/anti-DP) in the checking solutions were significantly lower (
