Estimating points of departure (PODs) for human toxicity effects from experimental animal data

Chemical management and assessment frameworks, as life cycle impact assessment, aim at evaluating toxicological impacts on human health from chemical exposures. Such assessments typically rely on chemical-specific points of departure (PODs) from regulatory toxicity data sources, but such data are not available for the majority of chemicals to which people are exposed. Thus, experimental animal data may complement regulatory data to derive PODs that most closely mimic one that would be selected in regulatory assessments. This study aims to propose a method for consistently deriving PODs for substances for which regulatory data are missing. As starting point, we extracted and curated experimental animal toxicity data from the US EPA CompTox Dashboard. We considered only oral repeat-dose studies and three non-cancer effect level types: lowest-observed-adverse-effect level (LOAEL), no-observed-adverse-effect level (NOAEL) and benchmark dose lower bound (BMDL). Curation steps included harmonization of units in mg/kg/d and extrapolation of LOAELs and BMDLs to NOAELs. We then estimated PODs for 1625 chemicals based on the 5th percentile of all data available for the same substance across animal species, assuming a lognormal distribution. These PODs correlate well with the 447 available regulatory PODs (adjusted test R2 =0.69; RSE=0.65 log10 units). This method significantly broadens the coverage of chemicals by estimating PODs consistent with regulatory values for a larger dataset. Next steps include using this curated dataset to train a machine-learning-based prediction model to estimate PODs for the even larger population of substances without experimental animal data. This abstract does not necessarily reflect US EPA policy.