Effects of Excluding Serum PCDD/F and Dioxin-Like PCB Data of Individuals with Incomplete Congener Profiles on Estimates of Total TEQ

Purpose We assessed whether excluding individuals without complete congener profiles biased various total TCDD TEQ estimates calculated using both the WHO1998 and WHO2006 TEFs. Methods PCDD/F and PCB serum concentration data collected as part of the 2001–2002 National Health and Nutrition Examination Survey were obtained from the National Center for Health Statistics, and total TCDD TEQ concentrations were calculated by summing the product of each congenerʹs concentration and its associated WHO1998 or WHO2006 TEF. Individuals 20 years and older who were not missing all 17 PCDD/F congeners or all 26 dioxin-like congeners were included in these analyses. Comparison of estimates reported previously with estimates calculated for these analyses were made by assessing both absolute differences and percent differences of each statistic by TEQ subgroup. Results The average percent difference in the mean, median, and 95th percentile TEQPCDD/F was 2.09, 1.15, and 0.89%, respectively, using the WHO1998 TEFs and 1.86, 1.39, and 0.56%, respectively, using the WHO2006 TEFs, while the average percent difference in the mean, median, and 95th percentile TEQ17–9 was 1.35, 0.79, and -0.85% (WHO1998 TEFs), respectively, and 1.60, 1.23, and -1.50% (WHO2006 TEFs), respectively. Examination of estimates for particular subgroups demonstrated that the 95th percentile TEQPCDD/F and TEQ17–9 reported previously for males 20–29 years of age were moderately underestimated and that the 95th percentile TEQPCDD/F and TEQ17–9 reported previously for males 30–44 years of age were slightly overestimated. Conclusion While the 95th percentile for certain subgroups were impacted more than others, summary statistics reported previously were not greatly influenced due to excluding individuals without complete congener profiles.