Anal. Chem. 2002, 74, 790-798
Determination of Polybrominated Diphenyl Ethers and Polychlorinated Biphenyls in Human Adipose Tissue by Large-Volume Injection-Narrow-Bore Capillary Gas Chromatography/Electron Impact Low-Resolution Mass Spectrometry Adrian Covaci,*,† Jacob de Boer,‡ John Jake Ryan,§ Stefan Voorspoels,† and Paul Schepens†
Toxicological Center, University of Antwerp, Universiteitsplein 1, 2610 Wilrijk, Belgium, Netherlands Institute for Fisheries Research, P.O. Box 68, 1970 AB IJmuiden, The Netherlands, and Health Canada, Bureau of Chemical Safety, Ottawa K1A 0L2, Ontario, Canada
A new analytical method has been developed for the quantification of polybrominated diphenyl ethers (PBDEs) in human adipose tissue samples. After Soxhlet extraction and a cleanup procedure using two successive solid-phase extraction cartridges containing acid silica and acid silica: neutral silica:deactivated basic alumina (from top to bottom), detection can been achieved by narrow-bore (0.10-mm i.d.) capillary gas chromatography/electron impact low-resolution mass spectrometry using a largevolume injection technique. Using narrow-bore capillaries, it is possible to analyze complex mixtures in a short time (up to 10 min), saving 50% or more of the analysis time of conventional columns while maintaining a similar resolution power. The method allows the determination of five major PBDE congeners (BDE 28, 47, 99, 100, and 153) at concentrations below 1 ng/g lipid weight. Detection limits in the selected ion mode varied between 0.05 and 0.30 ng/g lipid weight, depending on the degree of bromination. The sensitivity of this method can compete with low-resolution mass spectrometry with electron capture ionization, while a much better selectivity is obtained. Levels of PBDEs in 20 Belgian human adipose tissue samples ranged between 2.18 and 11.70 ng/g lipid weight and were similar to previously reported values from Europe. Polybrominated diphenyl ethers (PBDEs) have been used extensively over the past two decades as flame retardants in most types of polymers used in electronic circuit boards and computer and TV housing, furniture, building materials, textiles, carpets, and vehicles1. Because of their persistence and bioaccumulation potential, various PBDEs have been reported in various environmental matrixes2-4 and in humans.5-9 To assess possible oc* Corresponding author. Fax: +32 3 820 2722. E-mail:
[email protected]. † University of Antwerp. ‡ Netherlands Institute for Fisheries Research. § Health Canada. (1) Major Brominated Flame Retardants Volume Estimates; Bromine Science and Environmental Forum: Brussels, Belgium, 2000. (2) De Wit, C. Organohalogen Compd. 1999, 40, 329-332.
790 Analytical Chemistry, Vol. 74, No. 4, February 15, 2002
cupational and elevated human exposure to PBDEs and to monitor trends, it is essential to establish background values for the general population. PBDE monitoring in humans has started only recently, and only a limited data set is available from Sweden5,8,10, Spain,7 Finland,9 and Canada.11 Despite the fact that PBDE concentrations in humans are still significantly lower than those of polychlorinated biphenyls (PCBs) and p,p′-dichlorodiphenyldichloroethylene (p,p′-DDE) in human tissues,12 a dramatic increase in tetra to hexa BDE concentrations in Swedish human milk collected between 1972 and 2000 has been reported, corresponding to a doubling every 5 years until 1997, followed by a present steady state.13 More information is becoming available on the toxic aspects of PBDEs.14,15 Although PBDEs show a relatively low toxicity in several tests,16 recent data indicate that they may be more harmful than previously expected. Several PBDE congeners interfere (3) Sellstro ¨m, U.; Kierkegaard, A.; De Wit, C.; Jansson, B. Environ. Toxicol. Chem. 1993, 17, 1065-1072. (4) Alaee, M.; Luross, J.; Sergeant, D. B.; Muir, D. C. G.; Whittle, D. M.; Solomon, K. Organohalogen Compd. 1999, 40, 347-350. (5) Meironyte´, D.; Nore´n, K.; Bergman, A. J. Toxicol. Environ. Health, A 1999, 58, 101-113. (6) Sjo ¨din, A.; Hagmar, L.; Klasson-Wehler, E.; Kronholm-Diab, K.; Jakobsson, E.; Bergman, A. Environ. Health Perspect. 1999, 107, 643-648. (7) Meneses, M.; Wingfors, H.; Schumacher, M.; Domingo, J. L.; Lindstro¨m, G.; van Bavel, B. Chemosphere 1999, 39, 2271-2278. (8) Lindstro ¨m, G.; van Bavel, B.; Hardell, L.; Liljegren, G. Oncol. Rep. 1997, 4, 999-1000. (9) Strandman, T.; Koistinen, J.; Kiviranta, H.; Vuorinen, P. J.; Tuomisto, J.; Vartiainen, T. Organohalogen Compd. 1999, 40, 355-358. (10) Meironyte´, D.; Bergman, A.; Nore´n, K. Arch. Environ. Contam. Toxicol. 2001, 40, 564-570. (11) Ryan, J. J.; Patry, B. Organohalogen Compd. 2000, 47, 57-60. (12) Sjo ¨din, A.; Hagmar, L.; Klasson-Wehler, E.; Bjork, L.; Bergman, A. Environ. Health Perspect. 2000, 108, 1035-1048. (13) Nore´n, K.; Meironyte´, D. Chemosphere 2000, 40, 1111-1123. (14) De Boer, J.; De Boer, K.; Boon, J. P. The Handbook of Environmental Chemistry; Paasivirta, J., Ed.; Springer-Verlag, New York, 1999; Vol. 3, Part K, pp 61-95. (15) Pijnenburg, A. M. C. M.; Everts, J. W.; De Boer, J.; Boon, J. P. Rev. Environ. Contam. Toxicol. 1995, 141, 1-26. (16) WHO. Brominated Diphenyl Ethers; Environmental Health Criteria 162; International Program on Chemical Safety: World Health Organization, Geneva, 1994. 10.1021/ac010784e CCC: $22.00
© 2002 American Chemical Society Published on Web 01/18/2002
weakly with the aryl hydrocarbon (Ah) receptor.17 BDE 47 is transformed in rats and mice to hydroxylated metabolites18 that compete with thyroxin for the binding site on transthyretin,19 and some hydroxylated PBDE metabolites also bind to the thyroid receptor.20 Together, these observations suggest that PBDEs might act as endocrine disruptors. Because concentrations of PBDEs in humans are on the order of nanogram per gram of lipid weight (∼50-200 times lower than PCBs), most of the analytical work has been carried out by highly sensitive systems. Brominated substances are often analyzed under electron capture negative ionization (ECNI) conditions21 using low-resolution mass spectrometry (LRMS) by monitoring the negative ions formed by electron capture reactions. The predominant ions formed under such conditions are the bromine isotopes m/z 79 and 81. This technique is more sensitive and less costly than other alternatives, such as electron impact highresolution mass spectrometry11 (EI-HRMS); however, the latter technique has a higher selectivity than the ECNI-LRMS, since the accurate mass of the molecular ion or fragment ion for each level of bromination is recorded. The higher specificity of the electron ionization mode is needed to reduce the risk of misinterpretations of interfering substances. Furthermore, it allows the use of 13C-labeled standards (as internal standards), which makes the quantification procedure more accurate. The use of electron impact low-resolution mass spectrometry (EI-LRMS) can be useful because of easy maintenance of the instrumentation and lower costs. Until now, EI-LRMS in combination with a classical injection technique (e.g., 1-2 µL injected in hot splitless) was used for the determination of PBDEs in samples with relatively high concentrations.22 However, its use in human monitoring was limited because of a lower detectability (1-2 orders of magnitude) for congeners with more than four Br atoms. For lower brominated congeners, the sensitivity of EI increases. In the present study, the use of large-volume (up to 20 µL) injection (LVI) combined with narrow-bore capillary column gas chromatography and EI-LRMS for the determination of individual PBDE congeners in human adipose tissue at the low parts-perbillion level was investigated. EXPERIMENTAL SECTION Materials. Human adipose samples were obtained in 2000 by post-mortem abdominal surgery at the University Hospital of Antwerp, Belgium. Samples were collected in hexane-washed polyethylene containers and stored at -20 °C until analysis. Hexane, acetone, dichloromethane, and isooctane in residue analysis grade were obtained from Merck (Darmstadt; Germany). Silica gel 60 (63-230 mesh), alumina 60 (70-230 mesh), and anhydrous sodium sulfate p.a. (Merck) were heated at 150 °C for 24 h. The preparation of silica impregnated with concentrated sulfuric acid p.a. (Merck) was previously described.23 The acid (17) Meerts, I.; Luijks, E.; Marsh, G.; Jakobsson, E.; Bergman, A.; Brouwer, A. Organohalogen Compd. 1998, 37, 147-150. (18) O ¨ rn, U.; Klasson-Wehler, E. Xenobiotica 1998, 28, 199-211. (19) Meerts, I.; Marsh, G.; van Leeuwen-Bol, I.; Luijks, E.; Jakobsson, E.; Bergman, A.; Brouwer, A. Organohalogen Compd. 1998, 37, 309-312. (20) Marsh, G.; Bergman, A.; Bladh, L. G.; Gillner, M.; Jakobsson, E. Organohalogen Compd. 1998, 37, 305-308. (21) Sellstro ¨m, U.; Jansson, B.; Kierkegaard, A.; De Wit, C. Chemosphere 1993, 26, 1703-1718. (22) Lindstro¨m, G.; Wingfors, H.; Dam, M.; van Bavel, B. Arch. Environ. Contam. Toxicol. 1999, 36, 355-363.
silica was always used under a laminar flow hood to ensure that the analyst does not inhale particles of silica. The following individual PCB standards were obtained in isooctane from Dr. Ehrenstorfer Laboratories (Augsburg, Germany): IUPAC no. 28, 31, 46, 52, 74, 95, 99, 101, 105, 110, 118, 128, 138, 143, 146, 149, 153, 156, 163, 167, 170, 171, 172, 177, 178, 180, 183, 187, 189, 190, 194, 195, 196, 199, 203, 206, and 209. Native PBDE standards in isooctane (no. 28, 47, 66, 71, 75, 77, 85, 99, 100, 138, 153, and 154) were obtained from Promochem (Wessel, Germany), and 13C-labeled PBDEs (13C-BDEs 47, 99, and 153) were obtained in nonane from Wellington Laboratories (Guelph, Ontario, Canada). Extraction of Adipose Tissue. A 1-g portion of each sample was accurately weighed and mixed with 6 g of anhydrous Na2SO4 until a fine floating powder was obtained. Aliquots of 100 µL from a mixture of 13C-BDE 47, 99, and 153, 13.06 pg/L in isooctane, and 200 µL from a mixture of PCB 46 and 143, 100 pg/L in isooctane, were added. The powder was then extracted by automated hot Soxhlet for 2 h with 75 mL of hexane:acetone: dichlormethane (3:1:1, v/v). The extract was evaporated to dryness, and the extracted lipids were determined gravimetrically. Cleanup. Two successive solid-phase extraction (SPE) cartridges containing 6 g of acid silica and 2 g of acid silica:1 g of neutral silica:2 g of deactivated basic alumina (from top to bottom), respectively, were used for cleanup. After rinsing each cartridge with 20 mL of hexane, the sample extract was applied to the upper acid silica SPE cartridge. From this moment, eluate collection started under the second cartridge. The eluate from the first cartridge was directed into the second cartridge. The first cartridge was eluted using 25 mL hexane and then removed. PBDEs, together with PCBs and DDTs, were eluted from the second cartridge using 15 mL of hexane and 20 mL of hexane: dichloromethane (1:1, v/v). The total eluate (60 mL) was concentrated to near dryness, and 100 µL of the syringe standard (polybromobiphenyl (PBB) 80, 18 pg/L in isooctane) was added. The extract was concentrated under a N2 stream to ∼60 µL and transferred to an injection vial. GC/MS System. All measurements were performed using a HP 6890/5973 GC/MS. For PBDEs and PCBs, the system was equipped with a 10 m × 0.10 mm × 0.10 µm AT-5 (5% phenyl polydimethyl siloxane) capillary column (Alltech, Lokeren, Belgium). For PCBs only, a 50 m × 0.22 mm × 0.25 µm HT-8 (1,7dicarbaclosododecarborane 8% phenyl methyl siloxane) capillary column (SGE, Zulte, Belgium) was used. Sample introduction was performed by a HP 6890 automated liquid sampler. The temperature of the ion source was set at 230 °C and the interface, at 300 °C. Helium was used as the carrier gas at constant flow (0.4 and 0.7 mL/min for the 10-m and 50-m columns, respectively). The low-resolution quadrupole mass spectrometer (EI) was operated at 70 eV in the selected ion monitoring (SIM) mode. The ion source and quadrupole temperatures were 230 and 250 °C, respectively. The electron multiplier voltage was set at 2300 V. PBDE Determination. Twenty µL (4 × 5 µL) of the extracts was injected into a Gerstel (CIS 4) programmable temperature vaporizer (PTV) in the solvent vent mode (vent flow, 100 mL/ min for 1.1 min; injector at 70 °C for 1.1 min and then heated at 700 °C/min to 270 °C) with the split outlet opened after 2.1 min (Figure 1). The temperature program of the AT-5 column was (23) Covaci, A.; Schepens, P. Chemosphere 2001, 43, 439-447.
Analytical Chemistry, Vol. 74, No. 4, February 15, 2002
791
Figure 1. Time relationships of oven temperature, PTV temperature, and split status in large-volume injection mode.
Table 1. Acquisition Parameters, Recoveries, and Detection Limits for Target BDE Congeners. target compd tritetra-
penta-
hexa-
PBB 80 BDE 28 13C-BDE 47* BDE 47 BDE 66 BDE 71 BDE 75 BDE 77 13C-BDE 99* BDE 85 BDE 99 BDE 100 BDE 119 13C-BDE 153* BDE 138 BDE 153 BDE 154
Tr (min)
ions
7.02 6.07 7.05 7.05 7.17 6.94 6.82 7.37 7.94 8.38 7.93 7.72 7.80 9.06 9.79 9.06 8.64
470, 472 406, 408 496, 498 484, 486 484, 486 484, 486 484, 486 484, 486 576, 578 564, 566 564, 566 564, 566 564, 566 496, 498 484, 486 484, 486 484, 486
recovery (%)
81 ( 12
84 ( 17
103 ( 21
LOD (ng/g lipid) 0.05 0.1 0.1 0.1 0.1 0.1 0.20 0.15 0.15 0.15 0.30 0.25 0.25
programmed from 70 °C (2.2 min) to 230 °C at a rate of 40 °C/ min, then to 280 °C (5 min) at a rate of 25 °C/min. The run time was 13.2 min. Dwell times were set at 10 ms. The two most abundant ions (M+ and [M + 2]+ for the tri-, tetra-, and pentacongeners and [M - 2 Br]+ and [M + 2 - 2 Br]+ for the hexacongeners) were monitored for each level of bromination for native and labeled PBDEs (see Table 1). PCB Determination. For the 10-m AT-5 column, 1 L of the extract was injected in a cold splitless (injector temperature at 100 °C (0.1 min), then heated at 700 °C/min to 270 °C). The splitless time was 1 min. The temperature program of the AT-5 column was programmed from 90 °C (1 min) to 200 °C (0.5 min) at a rate of 50 °C/min, then to 250 °C (0.2 min) at a rate of 25 °C/min, and finally, to 280 °C (2 min) at a rate of 75 °C min-1. Run time was 8.3 min, and dwell times were set at 10 ms. For the 50-m HT-8 column, 1 L of the extract was injected in a hot pulsed splitless (pressure pulse 30 psi, pulse time 1.20 min) at an injector temperature of 270 °C. The splitless time was 1.25 min. The temperature was programmed from 90 °C (1 min) to 170 °C (2 min) at a rate of 15 °C/min, and finally, to 290 °C (14 min) at a 792
Analytical Chemistry, Vol. 74, No. 4, February 15, 2002
rate of 4 °C/min. The run time was 52.3 min, and dwell times were set at 50 ms. The two most abundant ions (m/z ) 256, 258 for tri-CBs; 290, 292 for tetra-CBs; 324, 326 for penta-CBs; 360, 362 for hexa-CBs; 394, 396 for hepta-CBs; 428, 430 for octa-CBs; 464, 466 for nona-CBs; and 498, 500 for deca-CBs) were monitored for each level of chlorination. Quality Control. Retention times, ion chromatograms, and intensity ratios of the monitored ions were used as identification criteria. A deviation of the ion intensity ratios within 20% of the mean values of the calibration standards was considered acceptable. PBDE. Recoveries of 13C-labeled internal standards, calculated on the basis of PBB 80, were between 81 and 103%, with a standard deviation of 0.995) was achieved for each compound between 2 × LOD and 10 ng/g lipid weight. Two samples of biota (eel and porpoise liver) analyzed in the first worldwide interlaboratory test on PBDE24 were used for method evaluation (Table 2). Furthermore, external quality control was achieved through comparison of PBDE concentrations in human adipose tissue measured with three different methods (Table 2). PCBs. Mean recoveries of the internal standards were 73 ( 7 and 79 ( 10% for PCBs 46 and 143, respectively. LODs ranged between 0.2 and 0.5 ng/g lipid weight. Coeluting PCB congeners were identified on the basis of their retention times and their reported elution order25 for each capillary column. The method performance was assessed through rigorous internal quality control, which included daily checks of calibration curves and the regular analysis of procedural blanks and certified material CRM 350 (PCBs in mackerel oil). The method was assessed through participation to an interlaboratory test organized by the Institute for Reference Measurements and Materials (IRMM, Geel, Belgium). Seven PCB congeners (no. 28, 52, 101, 118, 138, 153, and (24) De Boer, J. Organohalogen Compd. 2000, 45, 118-121. (25) Frame, G. Fresenius’ J. Anal.Chem. 1997, 357, 701-714.
Table 2. Interlaboratory Comparison for PBDEs in Different Samplesa BDE sample
description
A
HATb
B
HAT
C
HAT
D
HAT
E
eel
F
porpoise liver
detection
28
47
99
100
153
sum
difference (%)
EI-LRMSc ECNI-LRMSd EI-LRMS ECNI-LRMS EI-LRMS EI-HRMSe EI-LRMS EI-HRMS EI-LRMS BSEFg EI-LRMS BSEF
