Theme: 07. Per- and Polyfluorinated Alkyl Substances (PFAS)

Topic Area: 7l. PFAS Human Health and Ecological Risk Assessment and Toxicity

Formatted Title:

Integrating Evidence for Dose-Response Assessment in PFAS Regulatory Risk Assessment

Background/Objectives:

Background/Objectives. The United States Environmental Protection Agency (USEPA) has derived toxicity values for perfluorooctanoic acid (PFOA) and perfluorooctane sulfonic acid (PFOS) (as well as additional PFAS) based on epidemiological associations between PFOA and PFOS blood concentrations and decreased response to tetanus and diphtheria vaccines, increases in total cholesterol, and decreases in birth weight. The toxicity values based on epidemiological data are three to four orders of magnitude lower than previous values based on animal studies and were used to inform maximum contaminant levels (MCLs) as part of rulemaking under the National Primary Drinking Water Regulations (NPDWR).

Approach/Activities:

Approach/Activities. We conducted a critical evaluation of the strength of evidence for causality of these associations, considering both mechanistic data regarding possible modes of action as well as the potential for confounding in the epidemiological data. We also applied a toxicity pathway framework to explicitly identify data gaps from molecular initiating events to key events and to adverse outcomes. Finally, we considered whether different approaches for assessing individuals at an increased risk of disease versus population-wide interventions can be useful in the problem formulation stage of regulatory risk assessments.

Results/Lessons Learned:

Results/Lessons Learned. While the epidemiological associations were reported for blood concentrations as low as 20 to 30 ng/mL, rodent and human in vitro effects (gene expression, cell function) occur above 1000 ng/mL. Additionally, the use of biomarkers as a surrogate for health effects limits their use for causal inference in the absence of additional evidence. These issues are primarily related to confounding or reverse causation (i.e., pharmacokinetic bias) that links effects on both the biomarkers and blood PFOA and PFOS concentrations. In contrast to perchlorate toxicity, where the mode of action is well understood (inhibition of iodide uptake) for the relationship of the precursor (changes in thyroid hormones) to an adverse outcome (hypothyroidism), the mode of action for PFOA and PFOS and the clinical relevance of small decreases in vaccine response and birth weight reported in epidemiological studies are not understood. Thus, there are data gaps related to understanding the mode of action for health biomarkers that are themselves considered precursors to adverse effects (i.e., increased duration or frequency of infectious diseases, increased incidence of cardiovascular diseases, or increased incidence of developmental diseases) when there is mixed evidence or null associations between PFOA and PFOS blood concentrations and these adverse outcomes. Additional in vitro studies at human-relevant PFAS exposures are needed. The available evidence from animal and in vitro studies published to date does not support the biological plausibility of reported epidemiological associations for PFOA and PFOS at human-relevant concentrations. More human-relevant data are needed to inform the biological pathways from molecular initiating events to adverse outcomes to demonstrate causal relationships and scientifically justify the apparent conservatism in the USEPA toxicity values.

Harvey Clewell

Affiliation

Track

Refine Results

Integrating Evidence for Dose-Response Assessment in PFAS Regulatory Risk Assessment