Environ. Sci. Technol. 2004, 38, 2345-2350
Preliminary Assessment of U.K. Human Dietary and Inhalation Exposure to Polybrominated Diphenyl Ethers S T U A R T H A R R A D , * ,† RAMANEE WIJESEKERA,‡ STUART HUNTER,† CHRIS HALLIWELL,† AND ROBERT BAKER† Division of Environmental Health and Risk Management, Public Health Building, School of Geography, Earth, and Environmental Sciences, University of Birmingham, Birmingham, B15 2TT, United Kingdom, and Department of Chemistry, University of Columbo, Columbo, Sri Lanka
This study reports concentrations of BDEs 47, 99, 100, 153, and 154 in outdoor air [median ∑PBDE (sum of BDEs 47, 99, 100, 153, and 154) ) 18 pg m-3] in air from a range of office and home indoor microenvironments (median ∑PBDE ) 762 pg m-3) and vegan and omnivorous duplicate diet samples (median ∑PBDE ) 154 and 181 pg g-1 dry weight for vegan and omnivorous diets, respectively). Median daily human exposure to ∑PBDE via inhalation is 6.9 ng/person and 90.5 ng/person via diet but the relative significance of these pathways may vary considerably between individuals. Median concentrations in indoor air were higher in workplace (∑PBDE ) 1082 pg m-3) than in domestic (∑PBDE ) 128 pg m-3) microenvironments, and substantial differences in concentrations in air from different rooms in the same office building were found. When data from the only mechanically ventilated room was excluded, a significant positive correlation (p < 0.001) was observed between PBDE concentrations and both the number of electrical appliances and polyurethane foamcontaining chairs. Concentrations of ∑PBDE and BDEs 47 and 99 were significantly higher (p < 0.1) in omnivorous diet samples than in vegan diet samples, implying that while plant-based foods contribute appreciably, higher exposure occurs via ingestion of animal-based comestibles.
Introduction Polybrominated diphenyl ethers (PBDEs) are a group of brominated compounds widely used as flame retardants. In recent years, production and use of PBDEs has been in the guise of three formulations: penta (consisting primarily of BDEs 47 and 99-37% each, alongside smaller amounts of other tetra-, penta-, and hexa-BDEs), octa [a mixture of hexa (10-12%), hepta (44-46%), octa (33-35%), and nona (1011%)], and deca (98% decabromodiphenyl ethersBDE 209s and 2% various nona-BDEs) (1, 2). Worldwide, PBDE production is dominated by the deca commercial formula* Corresponding author e-mail:
[email protected]; telephone: +44 121 414 7298; fax: +44 121 414 3078. † University of Birmingham. ‡ University of Columbo. 10.1021/es0301121 CCC: $27.50 Published on Web 03/18/2004
2004 American Chemical Society
tion, with global demand in 2001 an estimated 56 100 t (3). This is similar to the 1999 estimate of 54 800 t (4). By comparison, 2001 global demand for the penta product was 7500 t (3), down slightly from 8500 t in 1999 (4). Production and use of commercial PBDE formulations in North America exceed considerably that in Europe; for example, in 2001, 7100 t of penta product was used in North America as compared to just 150 t in Europe (3). The uses for these commercial formulations are myriad: the penta product is employed mainly to flame retard polyurethane foams in carpet underlay, furniture, and bedding; the octa formulation is used to flame retard thermoplastics such as high-impact polystyrene; and the deca product is used principally in plastic housings for electrical goods such as TVs and computers as well as textiles (1). Despite concerns surrounding these contaminants owing to their presence in human milk (5-7), adipose tissue (8-10), and blood (11, 12), coupled with evidence relating to their potential adverse effects on human health (2, 13)swhich have led to EU restrictions on and from July 2003 the banning within the EU of the penta product (3)scomparatively few data are available relating to both (a) the magnitude of human exposure and (b) the relative significance of inhalation and diet as human exposure pathways. Table 1 summarizes the available data relating to measurements of human dietary exposure in Canada (7), The Netherlands (14), Spain (15), and Sweden (13, 16). To our knowledge, there are currently no human dietary exposure data available for the U.K. Furthermore, despite reports of elevated airborne concentrations of PBDEs in occupational environments such as electronics recycling plants (17, 18) and the presence of a wide range of PBDEtreated goods in both domestic and common occupational environments such as offices, there has been no systematic experimental evaluation of the relative significance of exposure via inhalation of outdoor and indoor air. This latter environment is likely to be of particular importance given the number of PBDE sources indoors and the low proportion of time spent outdoors by the population of the UK and other temperate industrialized regions. Investigation of the potential contribution of inhalation as an exposure pathway is especially pertinent as (i) while concentrations of PBDEs in human tissue in North America far exceed those in Europe (5, 19), dietary exposure estimates are similar (see Table 1) and (ii) the exponential temporal increase in PBDE concentrations in Swedish human milk is at considerable variance with that in Swedish pike and guillemotsin which concentrations have leveled off or may be decreasing (20). Both observations suggest that current human exposure may not solely originate from the diet. This study reports concentrations of a number of PBDE congeners in outdoor air, in air from a range of indoor microenvironments, and both typical omnivorous and vegan duplicate diet samples. While recognizing that other PBDE congeners such as BDE 28, 66, 85, etc. may be present in measurable quantities in both food and air, we focused on BDEs 47, 99, 100, 153, and 154. These congeners were selected for two principal reasons, specifically: (i) they have been identified as the most abundant in organisms (21), and (ii) they are the principal congeners monitored in previous dietary exposure assessments (see Table 1). Although decabromodiphenyl ether has been detected in human tissues (19), it was not included in this study owing to the difficulties in achieving its reliable determination (22). Our principal objectives were as follows: (i) To make a preliminary evaluation of both the overall magnitude of U.K. human exposure to PBDEs via diet and VOL. 38, NO. 8, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2345
inhalation and the relative significance of exposure via these exposure pathways. (ii) To evaluate the relative contributions of plant-based and animal-based comestibles to dietary exposure. (iii) To investigate the possible sources of PBDEs in indoor air. In particular, the potential influence of items treated with PBDEs, viz., electrical appliances and polyurethane foam (PUF)-containing chairs.
Experimental Section Indoor Air Sampling. Seventeen sampling sites located in and around the University of Birmingham were investigated between June 12 and June 28, 2001 (n ) 9), and between June 16 and June 30, 2002 (n ) 8). Air samples were collected using a Graseby-Anderson high-volume sampler fitted with a total suspended particulate inlet and modified to hold both a Teflon-coated glass fiber filter (GFF, 0.6 µm pore size) and a precleaned polyurethane foam (PUF) plug (0.016 g cm-3, 827 cm3). Sample flow rates were typically 0.6-0.8 m3 min-1, yielding sample volumes of typically 300 m3, collected with the windows closed, over a period not exceeding 24 h. Becausesowing to the noise generated by the sampler motorssampling was conducted in the absence of the room occupants, all electrical appliances were switched off during sampling. To minimize underestimation of concentrations that could occur when the volume of air sampled exceeded the volume of the room (23), sampling was conducted for a period equivalent to sampling 1 room volume and allowed to reequilibriate for an identical period, and the process was repeated until ca. 300 m3 of air had been sampled. Outdoor Air Sampling. A total of six samples of outdoor air were taken: two in July and August 2002 and four in April and May 2003. All samples were collected at our Elms Road Observatory Site (EROS) location on the University of Birmingham campus, for which PCB concentrations have been previously reported (24). Sampling was conducted using the same equipment used for indoor air sampling, but in view of the anticipated lower concentrations likely to be present in outdoor air, sampling was conducted continuously for 48 h per sample, yielding an air volume of typically 2000 m3. Note that apart from three samples where the GFF and PUF were analyzed separately, for the remaining samples, the GFF and PUF were combined for analysis. Duplicate Diet Samples. Accurately weighed ca. 50-g samples of freeze-dried duplicate diet samples collected in March-April 1999 and February-April 2000 were investigated. The samples comprised both omnivorous (n ) 10) and vegan (n ) 5) diets. Each sample represented the dietary intake collected for a given individual (diet samples relating to a total of 10 individuals were analyzed in this study) over a period of 7 consecutive days. A detailed account of the collection, preparation, and storage of these samples is available elsewhere (25), but a summary is given here. It is important to note thatsalthough actual consumption was accurately replicatedsin each trial, subjects were free to choose their meals from a menu, albeit more limited for the vegan diet. Subjects’ meals were taken under supervision in the metabolic unit of the Department of Nutrition and Dietetics of King’s College, London. Where food was consumed outside of the metabolic unit, then subjects kept detailed notes and provided duplicates of what was consumed. Individual duplicate diet samples of all food and drink were collected daily in solvent precleaned glass jars and aluminum boxes for the full duration of each trial. Each day’s diet was then frozen until the end of that 7-day trial. At the end of each trial, all diet samples for each subject were weighed, thawed, and homogenized. Appropriate quantities of each homogenate were immediately freeze-dried and stored in a freezer until analysis. 2346
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 8, 2004
Analytical Protocols. Samples were treated with known quantities of internal standards (PCB 173 for diet samples, changed to 13C12-BDEs 47, 99, and 153 for the air samples that were analyzed later), prior to Soxhlet extraction for ca. 16-24 h with dichloromethane:hexane (50:50 v/v). Concentrated crude extracts were washed with concentrated H2SO4 prior to sequential further purification via elution through a H2SO4-impregnated silica multicolumn with hexane, lipid removal via solvent exchange between dimethyl sulfoxide and hexane (diet samples only), and florisil chromatography. After concentration and exchange of solvent to nonane, GC/ MS analysis was carried out on a Fisons MD-800 instrument fitted with a HP-5 trace analysis column (50 m × 0.25 mm i.d., 0.25 µm film thickness). One microliter of sample extract was injected in the splitless mode at an injector temperature of 285 °C. The oven temperature program was 140 °C for 2 min, 5 °C/min to 200 °C; 2°C/min to 280 °C, 5 °C/min to 295 °C; 295 °C for 18 min. Six ions (for BDE 47: 485.8, 487.8; BDEs 99 and 100: 403.8, 405.8; and BDEs 153 and 154: 481.7, 483.7) were monitored in EI selected ion monitoring mode (ionization voltage, 70 eV; ion source temperature ) 250 °C). To ensure accurate and precise measurement, peaks were only accepted if the following criteria were met: (i) Signal-to-noise ratios for the least abundant ion exceeded 3:1. (ii) Peaks eluted within 5 s of standards run in the same batch as the samples. (iii) Isotope ratios for peaks were within 15% of those obtained for standards run in the same batch as the samples. Blanks consisting of a pre-extracted Soxhlet thimble for diet samples (n ) 3) and field blanks consisting of a GFF and a PUF plug (treated in identical fashion to those used for sampling, except that no air was aspirated through them) for air samples (n ) 3) were analyzed and found to contain concentrations of target PBDEs that were no greater than 5% of the concentrations found in the corresponding samples. Our data are thus not corrected for blank concentrations. Recoveries of internal standards ranged between 54 and 104% for all samples, with an average of 85%. Analytical precision was assessed by replicate (n ) 3) analysis of a homogenized sediment sample and found to be 4.3%, 3.0%, 5.6%, 22%, and 17% for BDEs 47, 99, 100, 153, and 154, respectively. Method detection limits for individual BDEs were typically 0.1 pg m-3, 1 pg m-3, and 5-20 pg g-1 dry weight for outdoor air, indoor air, and diet samples, respectively. At the time of writing, no certified reference materials are available for PBDEs; however, the accuracy of our methods is indicated by our satisfactory performance in the 2002 BSEF/QUASIMEME interlaboratory comparison on brominated flame retardants (26). Table 2 compares our reported values for freeze-dried sediment and wet mussel tissue with the assigned values and errors reported by BSEF/QUASIMEME.
Results and Discussion PBDE Concentrations in Outdoor Air. Table 3 shows the concentrations (sum of vapor and particle-bound phases) of BDEs 47, 99, 100, 153, and 154 detected in outdoor air samples and in the two categories (i.e., workplace and domestic) of microenvironments studied. Concentrations in the outdoor air samples analyzed in this study are of comparable magnitude to those reported for Chicago in the late 1990s (27), although there is a shift toward higher concentrations of the hexabrominated BDEs 153 and 154 in the Birmingham samples that may reflect greater North American use of the penta product. Concentrations in these Birmingham samples are slightly higher than those reported at two semi-rural U.K. locations in 2001 (28), confirming previous observations that atmospheric concentrations of PBDEs are greater in conurbations (27, 29).
TABLE 1. Measurements of Human Dietary Exposure to ∑PBDE location (ref)
daily exposure (ng d-1)
Canada (7) The Netherlands (14) Spain (15) Sweden (13) Sweden (16) a
comments based on foodstuffs collected 1998; ∑PBDE ) ∑BDE 28, 47, 99, 100, 153, and 154 did not include fruit and vegetables; ∑PBDE ) ∑BDE 28, 47, 99, 100, 153 and 154 based on foodstuffs collected 2000; ∑PBDE ) ∑tetra through octa-BDEs based on foodstuffs collected 1999; ∑PBDE ) ∑BDE 47, 99, 100, 153, and 154 based on foodstuffs collected 1998-1999; did not include fruit and vegetables; females only; ∑PBDE ) ∑BDE 47, 99, 100, 153, and 154
44 213a; 13b 97.3a; 81.9b 51 40.8
Upper bound estimate
