Use of reference pools to compare the qualitative and quantitative determination of polychlorinated biphenyls by packed and capillary gas chromatography with electron capture detection. Part 1. Serum

Serum for reference pools of in vivo polychlorinated biphenyls (PCBs) was obtained from four goats that had received one dose (100 mg kg–1) of a selected technical Aroclor (AR)(1016, 1242, 1254 or 1260) and were allowed to recover for 30 d. These pools were used to assess the differences in an analytical method that determines and quantifies PCBs using packed-column gas chromatography (PCGC)(quantified on the basis of mean mass percent. data for grouped PCB peaks) and capillary-column gas chromatography (CCGC)(quantified on the basis of percent. composition data for specific congeners). With CCGC, results were statistically significantly different (p⩽ 0.0002) from results with PCGC for ARs 1016, 1242 and 1254 but not for AR 1260 (p= 0.23). When comparing these gas chromatographic methods using bovine serum spiked in vitro with the same ARs at 17–25 p.p.b., it was found that the methods were not statistically significantly different for any of the ARs (p= 0.30–0.92). Levels of serum PCB determined by the two methods for 12 persons, divided into two groups according to exposure, were compared using the paired t-test. Group 1 consisted of three persons with dietary and/or environmental exposure; one with dietary and/or environmental exposure in addition to occupational exposure dating back 20 years. Group 2 consisted of eight persons with recent occupational exposure. The PCBs that eluted prior to 1,1-dichloro-2,2-di(p-chlorophenyl)ethylene (DDE) were not statistically significantly different for either group (n= 4, p= 0.45; n= 8, p= 0.07). The PCBs that eluted after DDE were found to be statistically significantly different for those persons with recent occupational exposure, Group 2 (n= 8, p= 0.004) but not for others, Group 1 (n= 4, p= 0.11). Total PCBs for the two methods were not statistically significantly different for Group 1 (n= 4, p= 0.82) but were for Group 2 (n= 8, p= 0.018). In this limited evaluation of the two quantifying techniques, using serum from 12 persons with different potential exposure histories, no consistent statistically significant differences were observed. Larger data sets are needed before definite conclusions can be drawn about the comparability of these two quantifying techniques.