Validation of proximity-based estimates of residential exposure to emissions of dioxins and dioxin-like compounds in the U.S.

Abstract

Background: Polychlorinated dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs), and polychlorinated biphenyls (PCBs) are persistent organic pollutants emitted from industrial sources. The relationship between proximity to emissions from these sources and personal exposure is poorly understood.Methods: We linked geocoded residential addresses and serum measurements of PCDDs, PCDFs, PCBs, and their derived TEQ (toxic equivalency quotient) for 5,725 participants in the 1999-2004 U.S. National Health and Nutritional Examination Survey (NHANES) to a U.S. Environmental Protection Agency database of 4,478 historical emission sources of dioxins and dioxin-like compounds (e.g., coal-fired power plants, solid waste incinerators). We calculated continuous and categorical (3, 5, 10km) exposure metrics reflecting distance to the nearest facility and estimated intensity of exposure with a distance- and toxicity-weighted average emissions index (AEI [ng TEQ/km2]). We used linear regression models to evaluate associations between proximity metrics and log-transformed sums of serum PCDDs, PCDFs, PCBs, and TEQ and used logistic regression to determine the extent these metrics predicted high (top quartile) serum levels.Results: Overall, 16%, 30%, and 53% of participants lived within 3, 5, or 10km of one or more facility, respectively. No consistent associations were observed between serum levels and simple proximity metrics reflecting distance to nearest facility. Mean serum concentrations of total PCDDs/PCDFs/PCBs and total TEQ were highest among participants in the top AEI5k quartile; an increasing trend was statistically significant only for PCBs (p<0.01). In adjusted models, participants in the highest AEI5k quartile had increased odds of having high levels of total PCBs (OR=1.8; 95%CI:1.3-2.5) and TEQ (OR=1.6; 95%CI:1.0-2.5) compared to those with no facilities within 5km.Conclusion: In this nationally representative population, we demonstrated that residential proximity to emissions was positively associated with serum total PCBs and TEQ. Findings from this effort support the use of distance- and emissions-based exposure proxies to estimate personal exposure.