Correlating pharmacokinetics and teratogenic endpoints.

The use of pharmacokinetics can improve the extrapolation of animal teratology data for human risk evaluation. Before one can extrapolate between species, however, the pharmacokinetic model must be predictable within the species for which it was developed. This article summarizes an approach being used for correlating pharmacokinetics and teratology endpoints in the same animal and predicting the teratogenic outcome for other animals of the same species. With the aid of micro-sampling procedures, and sensitive and rapid analytical techniques, blood, urine and feces samples are obtained from individual animals following dosing and the data are simulated using a hybrid computer to develop a pharmacokinetic model. The model is validated in other animals by measuring the parent compound and metabolites in various "compartments" predicted by the model. Then the pharmacokinetic model is tested by predicting the teratogenic outcome in single ani-analyses indicated the most predictive pharmacokinetic parameters to be two maternal blood concentration values. Prediction of the teratogenic outcome based on these parameters was accurate for 74% of the litters in the 95% confidence interval. This approach is discussed as it relates to its utility for other exposure routes and for extrapolation to other species.