A Margin-of-Exposure Approach to Assessment of Noncancer Risks of Dioxins Based on Human Exposure and Response Data

BACKGROUND: Risk assessment of human environmental exposure to polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/PCDFs) and other dioxin-like compounds is complicated by several factors, including limitations in measuring intakes because of the low concentrations of these compounds in foods and the environment and interspecies differences in pharmacokinetics and responses. OBJECTIVES: We examined the feasibility of relying directly on human studies of exposure and potential responses to PCDD/PCDFs and related compounds in terms of measured lipid-adjusted concentrations to assess margin of exposure (MOE) in a quantitative, benchmark dose (BMD)–based framework using representative exposure and selected response data sets. METHODS: We characterize estimated central tendency and upper-bound general U.S. population lipid-adjusted concentrations of PCDD/PCDFs from the 1970s and early 2000s based on available data sets. Estimates of benchmark concentrations for three example responses of interest (induction of cytochrome P4501A2 activity, dental anomalies, and neonatal thyroid hormone alterations) were derived based on selected human studies. RESULTS: The exposure data sets indicate that current serum lipid concentrations in young adults are approximately 6- to 7-fold lower than 1970s-era concentrations. Estimated MOEs for each end point based on current serum lipid concentrations range from < 10 for neonatal thyroid hormone concentrations to > 100 for dental anomalies—approximately 6-fold greater than would have existed during the 1970s. CONCLUSIONS: Human studies of dioxin exposure and outcomes can be used in a BMD framework for quantitative assessments of MOE. Incomplete exposure characterization can complicate the use of such studies in a BMD framework.