SOT FDA Colloquium: Route-to-Route Extrapolation in the 21st Century

The Society of Toxicology in conjunction with the US FDA Center for Food Safety and Applied Nutrition (CFSAN) have partnered to provide this colloquia series. The series presents scientific information that is high-quality, cutting-edge, future-oriented toxicological science to provide a well-grounded foundation to inform the work of US FDA employees.

On February 19, 2020, SOT and the US FDA hosted a webinar that presented considerations for conducting route-to-route extrapolations and discussed the possibility of developing consistent methods for utilizing such extrapolations for risk assessment.

The toxicity and pharmacokinetic profiles of substances may vary with the exposure route. If appropriately-conducted studies are available for a relevant route of exposure, generally the point of departure (POD) is calculated based on the data from the studies. However, if adequate data based on the relevant route of exposure are not available, a route-to-route extrapolation methodology may be employed to predict toxicity and estimate POD for risk assessment. This methodology evaluates data from studies based on other non-relevant routes of exposure, provided that the observed toxicity or biomarkers of toxicity are systemic and not related to portal of entry. Route-to-route extrapolation-based approaches utilize equivalent internal dose rather than external dose for predicting effects. Physiologically-based pharmacokinetic (PBPK) models enable route-to-route extrapolations of pharmacokinetics and systemic toxicity by normalizing internal dosimetrics for different routes of exposure. Given that route-to-route extrapolations do not incorporate differences in modes of action between exposure routes, there could be a certain degree of uncertainty associated with the model. However, extrapolation uncertainty reduces with the incorporation of additional data or assumptions (such as 100% absorption, in vitro, or QSAR-based predictions), when sufficient information is available to support the changes to the model. Although route-to-route extrapolation-based approaches are being explored for predicting effects and estimating POD, there are some inconsistencies in methods used by different organizations, which may introduce variability.