Adapting existing toxicokinetic models to relate perfluoroalkyl and polyfluoroalkyl intake to biomarkers in humans.

Exposures to per- and polyfluoroalkyl substances (PFAS) are associated with various adverse health outcomes, and a wide range of PFAS compounds have been detected in human serum, the environment, and food. Toxicokinetic models, however, have been developed for only a subset of the compounds of interest. To facilitate reverse dosimetry and risk assessment for the less studied PFAS compounds in food, we developed and evaluated an approach to adapt existing toxicokinetic models for nonhuman primates to predict human serum levels. The approach was validated with perfluorooctanoic acid and perfluorooctanesulfonic acid data and applied to perfluorohexanesulfonate. Results indicate that the approach yields similar dosimetry estimates to those of other models, particularly those used for regulatory purposes, suggesting the methodology can be leveraged to inform decision-making in data-sparse spaces. Applying and adapting the framework will improve our ability to connect dietary PFAS exposures to endpoints of concern for a wide range of PFAS compounds.