Science of the Total Environment 408 (2010) 2155–2162 Contents lists available at ScienceDirect Science of the Total Environment j o u r n a l h o m e p a g e : w w w e l s e v i e r c o m / l o c a t e / s c i t o t e n v Accumulation of polychlorinated biphenyls and brominated flame retardants in breast milk from women living in Vietnamese e-waste recycling sites Nguyen Minh Tue a, Agus Sudaryanto b, Tu Binh Minh c, Tomohiko Isobe b, Shin Takahashi a,⁎, Pham Hung Viet c, Shinsuke Tanabe a a b c Center for Marine Environmental Studies (CMES), Ehime University, Bunkyo-cho 2-5, Matsuyama 790-8577, Japan Senior Research Fellow Center, Ehime University, Bunkyo-cho 2-5, Matsuyama 790-8577, Japan Centre for Environmental Technology and Sustainable Development, Hanoi University of Science, 334 Nguyen Trai, Hanoi, Vietnam a r t i c l e i n f o Article history: Received 28 October 2009 Received in revised form January 2010 Accepted January 2010 Available online February 2010 Keywords: Breast milk E-waste Exposure pathway HBCD PBDE PCB a b s t r a c t This study investigated the contamination status of PCBs, PBDEs and HBCDs in human and possible exposure pathways in three Vietnamese e-waste recycling sites: Trang Minh (suburb of Hai Phong city), Dong Mai and Bui Dau (Hung Yen province), and one reference site (capital city Hanoi) by analysing human breast milk samples and examining the relationships between contaminant levels and lifestyle factors Levels of PBDEs, but not PCBs and HBCDs, were significantly higher in Trang Minh and Bui Dau than in the reference site The recyclers from Bui Dau had the highest levels of PBDEs (20–250 ng g− lipid wt.), higher than in the reference group by two orders of magnitude and more abundant than PCBs (28–59 ng g− lipid wt.), and were also the only group with significant exposure to HBCDs (1.4–7.6 ng g− lipid wt.) A specific accumulation, unrelated to diet, of low-chlorinated PCBs and high-brominated PBDEs was observed in e-waste recyclers, suggesting extensive exposure to these compounds during e-waste recycling activities, possibly through inhalation and ingestion of dust The estimated infant intake dose of PBDEs from breast milk of some mothers occupationally involved in e-waste recycling were close to or higher than the reference doses issued by the U.S EPA © 2010 Elsevier B.V All rights reserved Introduction Waste electrical and electronic equipment, also known as e-waste, refers to end-of-life products encompassing information-communication devices, consumer electronics and household appliances Owing to the short life span of devices such as computers, television sets, stereo systems, printers and cell phones, e-waste is generated in large amounts, with an annual volume of 20–50 million tonnes world-wide and increases rapidly at a rate of 3–5% per year (UNEP, 2005) The hazard of e-waste lies in the high content of many toxic substances (BAN and SVCT, 2002) including heavy metals (lead, mercury, cadmium, etc.) and persistent organohalogen compounds such as polychlorinated biphenyls (PCBs) and brominated flame retardants (BFRs) PCBs are present in older electrical capacitors and transformers as coolants and dielectrics whereas BFRs are additives found in most polymeric parts (printed circuit boards, cable coatings, plastic casings, etc.) There are increasing evidences that BFRs, such as polybrominated diphenyl ethers (PBDEs) and hexabromocyclododecanes (HBCDs), exhibit a range of toxic effects similar to PCBs, including ⁎ Corresponding author Tel./fax: + 81 89 927 8196 E-mail address: shint@agr.ehime-u.ac.jp (S Takahashi) 0048-9697/$ – see front matter © 2010 Elsevier B.V All rights reserved doi:10.1016/j.scitotenv.2010.01.012 endocrine disruption as well as reproductive and neurodevelopmental toxicity (Legler and Brouwer, 2003; Birnbaum and Staskal, 2004) Uncontrolled e-waste recycling has become a topic of serious concern in recent years It is estimated that up to 80% of e-waste from industrialised countries is exported to Asian developing countries for recycling, exploiting the inexpensive cost of labour and weak enforcement of environmental laws (BAN and SVCT, 2002) Large scale waste processing operations employing primitive practices such as uncontrolled dismantling, acid stripping and open burning in Chinese e-waste recycling sites (EWRSs) have resulted in severe environmental contamination (Wong et al., 2007) The human exposure levels to PCBs in Taizhou region (Zhao et al., 2007) and to PBDEs in Guiyu town (Bi et al., 2007) are two of the highest ever reported In other Asian developing countries such as Cambodia, India, Philippines and Viet Nam, although open waste disposal sites have been identified as potential sources of PCBs and dioxin-like compounds (Minh et al., 2003), information on EWRS are lacking Available data, albeit limited, indicate that the levels of PCBs and BFRs in the general populations from these countries are lower than those from developed countries (Schecter et al., 2004; Sudaryanto et al., 2005, 2008a; Malarvannan et al., 2009); nevertheless EWRSs are suspected as potential hotspots of these contaminants In this context, the present study was carried out to assess the human exposure to persistent organohalogen compounds related to e-waste recycling in 2156 N.M Tue et al / Science of the Total Environment 408 (2010) 2155–2162 Viet Nam Breast milk was selected as bioindicator on account of the simple and non-invasive sampling, relatively high content of lipid and lipophilic contaminants, and relevance to infants' health PCBs, PBDEs and HBCDs were determined in human breast milk collected from three EWRSs and a reference site in view of contamination levels, congener patterns, influence of lifestyle factors and potential health risk for breastfeeding infants Materials and methods 2.1 Study locations End-of-life electrical and electronic equipment in Viet Nam are often picked up by scrap collectors, through purchase from their former owners or by scavenging domestic waste disposal sites The wastes are then transported to trade villages specialised in recycling The present study investigated three e-waste recycling sites located in the Red River delta region, in the northern part of the country; Dong Mai (DM, Hung Yen province, battery recycling), Trang Minh (TM, suburb of Hai Phong City, e-waste dismantling) and Bui Dau (BD, Hung Yen province, e-waste dismantling) All study locations were small villages with approximately 80 households or less E-waste recycling as a trade started in these villages very recently at the beginning of the decade, except in the case of DM where the recycling of lead from batteries has been occurring for almost four decades Most businesses are family-based and e-waste is recycled in the backyard of the house The waste processing often consists of only dismantling, retrieval of easily accessible metals and grinding plastic casings intro reusable pellets Circuit boards are not processed and instead sold to foreign buyers Alongside these recent recycling activities, the villages still maintain traditional crop production A minority of the families also raise livestock or aqua-culture fish in limited quantities in-house, but the main supplies of livestock produces and fish for all the three communities are from the neighbouring communes 2.2 Sample collection The sampling survey was conducted in August 2007 In each of the three recycling sites, the number of nursing mothers was limited to around 20 or less Four breast milk samples from DM, eleven from TM and nine from BD were collected from the mothers who agreed to donate As reference, another nine samples were collected from the capital Hanoi (HN), a typical urban area Informed consents were obtained from all donors Questionnaire survey was also conducted to acquire personal information on age, height and weight measurements, number of deliveries and duration of nursing periods, occupation (recyclers or non-recyclers), period of involvement in recycling activities and dietary habit (Table 1) All donors were nonsmokers and appeared healthy Breast milk was expressed by the donor or with the help of a midwife into a solvent-precleaned glass container with Teflon-lined screw caps The samples were kept in gel ice immediately after collection and then sent within h to our laboratory in Viet Nam and frozen at −20 °C Later the frozen samples were air-transported with gel ice to the Environmental Specimen Bank (es-BANK, Ehime University, Japan) and stored at − 25 °C in until analysis 2.3 Chemical analyses Approximately 40 g of sample was freeze-dried and then Soxhletextracted with 300 ml diethyl ether and 100 ml hexane for h About 10% of the extract was used for determination of the lipid content using the gravimetric method The remaining extract was spiked with 13 C12–PCBs, 13C12–PBDEs (5 ng each) and 13C12–HBCDs (10 ng each) as standard surrogates, subjected to gel permeation chromatography (packed Bio-Bead S-X 3, Bio-Rad Laboratories, USA) for lipid removal using a hexane/dichloromethane (DCM) mixture (1:1 v/v) as eluant The cleaned-up extract was then concentrated and passed through g of activated silica gel (Wakogel® DX, Wako, USA) packed in a glass column for separation of HBCDs from PCBs and PBDEs The first fraction containing PCBs and PBDEs, eluted with a mixture of hexane/ DCM (95:5 v/v, 80 ml), was spiked with ng 13C12–BDE-139 as internal standard and solvent-exchanged into iso-octane The HBCD fraction, eluted with another mixture of hexane/DCM (75:25 v/v, 100 ml), was concentrated until dryness and then redissolved in a methanol solution containing deuterium-labelled HBCDs-d18 (10 ng each) as internal standards PCBs (62 congeners) and PBDEs (14 congeners commonly found in abundance in the technical PBDE mixtures (La Guardia et al., 2006), namely BDE-3, -15, -28, -47, -99, -100, -153, -154, -183, -196, -197, -206, -207, and -209) were quantified using a gas chromatograph (Agilent 7890 series) equipped with a mass-selective detector (Agilent 5975 series) and a DB-1 fused silica column (J&W Scientific) having 0.25 mm i.d × 30 m × 0.25 µm film thickness for PCBs and mono- to hepta-BDEs or 0.25 mm i.d × 15 m ì 0.1 àm lm thickness for octa- to deca-BDEs The GC oven temperature program for PCBs was: from 120 °C to 180 °C at 20 °C min− 1, °C min− to 260 °C, °C min− to 300 °C and held for PCBs were monitored in selective ion monitoring mode (EI-SIM) at masses of [M]+ and [M+ 2]+ for mono- to tetra-CBs, [M+ 2]+ and [M+ 4]+ for penta- to nona-CBs and [M+ 4]+ and [M+ 6]+ for deca-CB The instrumental set-up for PBDE analysis has been described previously (Sudaryanto et al., 2008b) HBCDs (α, β and γ isomers) were quantified by LC-MS/MS according to the methods described elsewhere (Isobe et al., 2007) Procedural blanks were analysed simultaneously with samples to check for interferences and contamination Detection limits were calculated as three times the procedural blank, and were in the range of 0.01–0.05 ng g− lipid wt for all target compounds Recoveries of 13 C-labelled surrogates were in the range of 57%–107% for PCBs, 63%– 110% for PBDEs and 84%–113% for HBCDs The analysis of human breast milk described above followed our common procedure employed for animal tissues For quality assurance and control, our laboratory participated in an intercalibration exercise on BFRs using the blubber of a marine animal as the reference material and our data were in good agreement with those reported by other laboratories Table General characteristics of the donors of breast milk Parameters Hanoi (n = 9) Dong Mai (n = 4) Trang Minh (n = 11) Bui Dau (n = 9) Age (year) Weight (kg) Height (cm) Body mass index (kg m− 2) Number of children Total nursing time (month) Occupation 23–35 (28.2) 46–62 (54.3) 155–167 (160.9) 16.5–25.4 (21.1) 1–2 (1.67) 4.5–25 (12) 89% office worker 11% housewife 25–31 (28.3) 42–52 (46.0) 153–155 (154.5) 17.9–21.6 (19.3) 1–4 (2.3) 4–57 (26) 25% recycler 75% farmer 18–26 (23.6) 41–50 (45.5) 150–160 (154.6) 16.4–21.3 (19.1) 1–2 (1.55) 1.5–24 (12) 55% recycler 45% housewife 24–35 (29.3) 40–46 (44.0) 150–155 (154.0) 16.7–21.6 (19.1) 1–4 (2.2) 9.5–79 (34) 55% recycler 45% others Values between parentheses are arithmetic means N.M Tue et al / Science of the Total Environment 408 (2010) 2155–2162 2157 HBCDs were calculated using RfD values of (Oostdam et al., 1999), 0.1, 0.1, 0.2 (EPA, 2008) and 0.2 µg kg− day− (European Chemicals Bureau, 2007), respectively (Isobe et al., 2007) Concentrations were expressed on a lipid weight basis unless otherwise specified 2.4 Statistical analyses Results The Wilcoxon rank sum test was used for assessing whether the contaminant levels between groups were significantly different In this analysis, non-detectable levels were set to zero Principal component analysis (PCA) was employed to categorise PCB and PBDE congeners according to their variation patterns This analysis included only the congeners detected in at least 70% of the samples The influence of socio-demographic parameters on the total concentration of each category was then analysed qualitatively using multiple linear regressions The parameters used as independent variables included: age, body mass index (BMI), total duration of nursing, period of involvement in recycling activities and consumption rate of food from animal origin (total rate of meat, fish and dairy products, servings per week) Other location-specific factors were assumed to be negligible and not considered Parameters with a pvalue of less than 0.05 were considered as having significant relationship with contamination level In these later two statistical analyses, non-detectable concentrations were set to half of the detection limit and then all concentrations were log-transformed to bring the data distribution closer to normality All calculations were performed using the statistical software package R version 2.9.0 PCBs, PBDEs and HBCDs were detected in all the samples analysed (Table 2) In terms of total concentrations, PCBs were more abundant than PBDEs and HBCDs by one order of magnitude or higher in majority of the samples However, PBDEs were more abundant than PCBs in three donors who were involved in the recycling of e-waste in BD Total PCB levels were statistically comparable among locations but significant differences were observed with individual PCB congeners The residents of the three recycling sites had statistically lower (p < 0.05) levels of high-chlorinated congeners such as CB-138, -153 and -180 but somewhat higher levels of tri- and tetra-chlorinated congeners (tri- and tetra-CBs) than those of the urban reference site CB-28 was detected at significantly higher (p < 0.05) levels in TM and BD (recyclers only) than in HN, and the highest levels of CB-28 and CB-74 were found in battery recyclers from DM In each recycling site, the total levels of tri- and tetra-CBs were higher in the recycler group than in the non-recycler group, with the median being 30 vs 10, 6.7 vs 4.9 and 5.0 vs 2.8 ng g− lipid wt in DM, TM and BD, respectively However, these differences were not statically significant (p > 0.05), probably due to the limited number of samples The preferential accumulation of these low-chlorinated biphenyls in the recycling sites resulted in distinctive PCB congener profiles As seen in Fig 1, the profile in the reference site followed the order of CB-138 > CB153 > CB-118 > CB-180 > CB-99 > CB-74 > CB-28 whereas in TM and BD, CB-28 and CB-74 were as abundant as CB-99 and more than CB180 In DM, the battery recycling site, CB-28 and CB-74 were the dominant congeners Based on the PCA results PCB congeners were divided into three categories (Fig 2): (a) high-chlorinated PCBs (HighCB), including hepta to deca congeners and two hexa congeners (CB-138 and CB-153), aligned with the first principal component (PC1) which accounted for 57% of the total variance; (b) low- 2.5 Risk assessment The risk for infants exposed to contaminants in mother milk was assessed using hazard quotients (HQs) An HQ is defined as the ratio of the estimated daily intake dose (DI) of the compound via breastfeeding to the corresponding maximum acceptable oral dose for human, or reference dose (RfD) A value of HQ higher than indicates potential risk The DIs were calculated based on the assumption that an infant in average weighs kg and consumes 700 g mother milk daily (Oostdam et al., 1999) HQs of total PCBs, BDE-47, BDE-99, BDE-153 and total Table Concentrations (ng g− lipid wt.) of major PCB congeners, PBDE congeners and HBCD isomers in human breast milk collected from Hanoi and Vietnamese e-waste recycling sites Compound CB-28 CB-74 CB-99 CB-118 CB-138 CB-153 CB-180 ΣPCBs BDE-15 BDE-28 BDE-47 BDE-99 BDE-100 BDE-153 BDE-154 BDE-183 BDE-196 BDE-197 BDE-206 BDE-207 BDE-209 ΣPBDEs α-HBCD β-HBCD γ-HBCD ΣHBCDs Bui Dau non-recyclers Bui Dau recyclers Median Hanoi Range Median Dong Mai Range Median Range Median Range Median Range 1.1 2.8 2.9 4.8 9.7 8.2 3.7 46 0.040 0.029 0.13 0.057 0.040 0.098 n.d 0.029 0.010 0.045 n.d 0.025 n.d 0.57 0.33 n.d n.d 0.33 0.42–7.6 1.3–6.0 1.2–7.3 2.0–13 3.8–17 3.8–16 1.9–5.5 20–100 0.014–0.073 0.017–0.067 0.070–0.25 n.d.–0.12 n.d.–0.10 0.062–0.14 n.d.–0.037 n.d.–0.038 n.d.–0.015 0.016–0.11 n.d.–0.027 n.d.–0.057 n.d 0.24–0.8 0.066–1.4 n.d n.d.–0.13 0.070–1.4 5.6 5.9 2.5 3.5 4.0 3.5 1.0 50 0.011 0.010 0.097 n.d n.d 0.10 n.d 0.050 0.010 0.046 0.024 0.030 0.17 0.73 0.43 n.d n.d 0.42 1.0–34 1.2–15 0.88–4.6 1.2–8.3 2.1–11 1.7–9.0 0.51–3.5 11–69 0.010–0.032 n.d.–0.044 0.041–0.20 n.d.–0.028 n.d 0.061–0.25 n.d 0.029–0.11 n.d.–0.018 0.026–0.14 n.d.–0.10 0.011–0.11 0.069–0.50 0.26–1.1 0.11–0.97 n.d n.d 0.11–0.97 2.0 2.7 2.1 3.3 5.8 5.7 1.8 33 0.030 0.074 0.40 0.11 0.082 0.40 0.021 0.13 0.036 0.26 0.05 0.10 0.42 2.3a 0.38 n.d n.d 0.38 0.72–10 0.42–7.6 0.60–4.3 1.0–6.9 1.9–11 1.8–11 0.71–4.3 11–73 0.010–0.34 0.024–1.0 0.11–1.8 0.032–0.51 0.022–0.21 0.021–1.5 n.d.–0.069 0.022–1.2 n.d.–0.14 0.032–0.78 0.012–0.23 0.041–0.41 0.12–7.3 0.55–13 0.12–3.3 n.d.–0.051 n.d.–0.27 0.11–3.3 0.85 1.6 1.7 2.8 4.8 3.6 1.4 24a 0.057 0.21 0.81 0.38 0.13 0.65 n.d 0.14 0.037 0.47 0.014 0.15 0.11 3.2a 0.36 n.d n.d 0.36 0.42–14 0.56–3.7 0.60–2.0 1.0–3.3 1.7–6.0 1.5–5.2 0.50–2.1 8.4–28 0.022–0.075 0.13–0.27 0.63–1.0 0.22–0.56 0.083–1.6 0.27–1.0 n.d.–0.035 0.053–0.22 0.027–0.056 0.17–0.61 n.d.–0.03 0.10–0.18 n.d.–0.16 2.0–4.0 0.29–1.2 n.d n.d 0.29–1.2 1.9 2.7 2.8 5.5 4.6 1.6 34 0.35 0.96 4.8 3.2 0.80 4.4 0.39 1.2 0.59 7.5 0.26 3.5 4.1 84a 1.9 n.d 0.045 2.0a 1.5–14 1.6–8.6 1.9–3.6 3.6–5.6 4.6–6.7 3.6–5.6 1.3–1.7 28–59 0.16–1.5 0.82–8.2 3.5–32 2.2–15 0.54–3.1 2.1–23 0.27–2.1 0.49–2.7 0.25–5.1 3.1–29 0.16–4.5 1.3–51 0.87–96 20–250 1.4–7.5 n.d n.d.–0.099 1.4–7.6 n.d.: not detected a Significantly different from reference site (p < 0.05) Trang Minh 2158 N.M Tue et al / Science of the Total Environment 408 (2010) 2155–2162 chlorinated PCBs (LowCB), consisting of mainly tri and tetra congeners, aligned with PC2 (17% of variance); and (c) others (MediumCB), mostly penta and a few hexa congeners including CB99, -101, -118, -128, etc., contributing to both PCs PBDE levels in the battery recycling site DM were comparable with those in the reference site HN whereas the levels in the two e-waste dismantling sites, TM and BD, were statistically higher (Table 2) In TM, the recycler and non-recycler groups had similar levels of total PBDEs, with an overall median higher than in the reference group by a factor of In BD, although the levels in the non-recycler group were only comparable to those in TM, the recycler group had the highest PBDE levels of all donors, approximately two orders of magnitude higher than those in the residents of HN In the reference site, BDE-47 and BDE-153 were the dominant PBDE congeners whereas BDE-209 was at nondetectable levels (Fig 1) BDE-209 was detected in most of the samples collected from the recycling sites with varying proportions up to 50% Octa to nona congeners were also observed at higher percentages BDE197 and -207 were prominent in samples from recyclers living in BD, with levels comparable to BDE-47 PCA of PBDE congeners indicated that these compounds could be considered as a single category because 86% of the variance could be represented by PC1 (Fig 2), suggesting a high degree of correlation among congeners with the exception of BDE-209, main contributor to PC2 (only 7.4% of the variance) HBCDs levels were not statistically different among residents of HN, DM and TM (both recyclers and non-recyclers) and BD (non-recyclers) Recyclers from BD had significantly higher levels than the other groups, Fig PCA loading plot of individual congeners of PCBs (left) and PBDEs (right) Numbers represent IUPAC numbers of PCB/PBDE congeners N.M Tue et al / Science of the Total Environment 408 (2010) 2155–2162 2159 Table Coefficients (β) and p-values of socio-demographic parameters in linear models of contaminant concentrations Parameters Age BMI Nursing time Recyclinga Dietb Model a b LowCB MediumCB HighCB PBDEs HBCDs β p β p β p β p β p − 0.028 0.062 0.008 0.145 0.008 0.19 0.36 0.11 0.009 0.85 − 0.015 0.046 − 0.002 0.076 0.070 0.23 0.09 0.49 0.022 0.013 − 0.007 0.023 − 0.0074 0.061 0.080 0.51 0.22 0.007 0.34 0.002 0.005 − 0.073 0.013 0.180 − 0.037 0.88 0.25 0.12 0.029 0.64 − 0.020 0.051 0.009 0.101 0.003 0.45 0.25 0.13 0.13 0.96 R2 p R2 p R2 p R2 p R2 p 0.40 0.013 0.37 0.023 0.48 0.002 0.38 0.018 0.24 0.17 Period of involvement in recycling activities (years) Consumption rate of food from animal origin (total rate for meat, fish and dairy products, servings per week) former were not major sources of PCBs Total PCB levels in Vietnamese human breast milk were in comparable ranges with those reported in other Asian developing countries and lower than in developed nations (Table 4) The levels observed in this study were similar to the serum PCB levels in the residents of Guiyu (median 52 ng g− lipid wt.), the largest recycling site of electronic waste in China, reported by Bi et al (2007) These authors also did not find any significant difference between the exposed and the reference populations Thus e-waste related PCB contamination is believed to involve old electric materials, especially transformers, rather than electronic waste The case of Luqiao, the largest Chinese disassembly site of electrical waste, is an example where very high PCB levels in human breast milk associated with e-waste recycling were reported (median 359 ng g− lipid wt., Zhao et al., 2007) The absence of substantial PCB contamination suggests that in the Vietnamese recycling sites, the occurrence of waste materials containing PCBs may be uncommon and the recycling of these materials is of limited scale Significant exposure to PBDEs was observed in the two e-waste dismantling sites, TM and BD, especially in BD recyclers Compared with the levels reported by other studies, the PBDE levels in the with a 6-fold difference compared with the reference group (Table 2) α-HBCD was the dominant isomer in all the samples, accounting for more than 90% of the total HBCD levels γ-HBCD was detected in eight samples with a proportion of less than 10% and β-HBCD was detected in only one sample Results from the fitting of chemical concentrations to multiple linear regression models of socio-demographic parameters (Table 3) showed that LowCB and PBDEs were associated only with involvement period in recycling activities, HighCB correlated positively with consumption of food from animal origin but negatively with nursing time whereas MediumCB had positive associations with both food consumption and recycling activities Age and BMI did not show any significant influence on contaminant concentrations Discussion 4.1 Contamination levels The similarity in PCB levels in the three Vietnamese e-waste recycling sites as also seen in the reference site indicates that the Table Comparison of PCBs and BFRs in human breast milk from Viet Nam with other countries Country Survey year PCBs 206 7.1 India 2003–2005 2004 2007 2002–2003 34 0.6a Indonesia Philippines Vietnam (HN) Vietnam (DM) Vietnam (TM) Vietnam (BD non-recylers) Vietnam (BD recyclers) 2001–2003 2004 2007 2007 2007 2007 2007 27 60 46 51 33 28 34 1.5 3.8 0.57 0.73 2.3 3.2 84 65 1.54 Belgium Norway Russia Spain 2004 2006 2006 2000–2002 2002 2003–2004 123 165 175 111 2.01 3.19 0.96 6.1 1.5 0.13 0.62 27 Sweden United Kingdom United States 2006–2007 2002–2003 2001–2003 2002–2003 111 180 126 2.93a 6.3b 50.4 0.35 Asian developing countries China Other countries Japan PBDEs HBCDs 0.86 0.62 0.33 0.42 0.38 0.36 2.0 1.4–4.0c 0.5 Concentrations are given as median if available (or arithmetic mean otherwise) and expressed in nanogram per gram lipid wt a Mono- to hepta-BDEs only b Mono- to hexa-BDEs only c Pooled values References Zhao et al (2007) Sudaryanto et al (2008b) Shi et al (2009) Subramanian et al (2007), Sudaryanto et al (2005) Sudaryanto et al (2008a) Malarvannan et al (2009) This study This study This study This study This study Inoue et al (2006) Kakimoto et al (2008) Colles et al (2008) Polder et al (2008b) Polder et al (2008a) Bordajandi et al (2008), Gómara et al (2007), Eljarrat et al (2009) Lignell et al (2003) Kalantzi et al (2004) Ryan et al (2006), She et al (2007) 2160 N.M Tue et al / Science of the Total Environment 408 (2010) 2155–2162 reference group in this study (HN) were among the lowest in the world (Table 4), consistent with the findings of Schecter et al (2004) Levels in TM and BD non-recyclers were higher in Indonesia and Japan and close to those in European countries whereas levels in BD recyclers were in the same range with American levels which are the highest reported among non-occupationally exposed populations The differences in waste materials, workload and processing methods may contribute to the variation of PBDE contamination levels in the Vietnamese recycling sites For instance, PBDEs may be minor contaminants in waste batteries as suggested by the low levels accumulated in the battery recyclers in DM On the other hand, the elevated levels in recyclers from BD may be the result of processing large amounts of waste materials with high PBDE contents using methods which facilitate the release of these contaminants from the waste matrices, such as burning and other thermal processes It is also remarkable that in BD recyclers, the PBDEs levels exceeded those of PCBs, despite the long legacy of the latter compounds This unusual feature (comparison in Table 4) has so far been observed only in case of uncontrolled e-waste recycling (Bi et al., 2007) It clearly indicates an extensive exposure to PBDEs with different sources from those of legacy persistent organic pollutants (POPs); nevertheless the highest PBDE levels in breast milk in this study were still lower than those in serum of Chinese e-waste dismantling workers from Guiyu (median 600 ng g− lipid wt., Bi et al., 2007) by an order of magnitude This large difference in lipid weight-normalised concentrations, albeit in two separate human matrices, again infers that Vietnamese recycling sites are still less contaminated in terms of PBDEs than their Chinese counterparts, probably on account of a smaller scale of e-waste processing Although information on HBCDs in human matrices is still limited, available data indicate that the accumulation levels of HBCDs in human breast milk from Asia, USA and several European countries are very low (