Assessment of methods to determine PCB levels in blood serum: interlaboratory study

Twenty-five laboratories participated in a study to determine the levels of polychlorinated biphenyls (PCBs) in 3 bovine serum pools, referred to as trial environmental pools (TEPs). TEPs 2 and 3 contained, respectively, low (9.97 ppb) and high (49.64 ppb) levels of PCBs (as in vivo-bound Aroclor 1254) and the same level of 9 commonly occurring chlorinated hydrocarbons. TEP 1 contained only naturally occurring levels of these analytes. Laboratories analyzed each sample in duplicate by the method used in their laboratory for measuring PCBs in blood serum. The coefficients of variation (CV) for the 12 laboratories reporting quantitative data and the required number of analyses for TEP 2 and TEP 3 were 37.0 and 30.7%, respectively. The mean recoveries for these 12 laboratories were 239.3 and 165.4% for TEP 2 and TEP 3, respectively. Three laboratories reported data with mean values for TEP 2 and TEP 3 within +/- 3 standard deviations of the CDC characterized mean. Their coefficients of variation were 12.4 and 18.8% for TEP 2 and TEP 3, respectively. The mean recoveries for these 3 laboratories were 150.7 and 98.4% for TEP 2 and TEP 3, respectively. Our most significant observations were the laboratories' failure to separate PCBs from DDTs and the excessive background of the reagent blanks. The widely discrepant results indicate a definite need to standardize methodology for this analysis.