Background: In Physiologically Based Pharmacokinetic (PBPK) models, the most important input parameter
is tissue-to-plasma partition coefficient (Kp). Over the years, several empirical methods have been developed
to predict Kp in animals.
Objectives: The objective of this study was to propose two allometric methods to predict Kp from experimentally
determined in vivo volume of distribution at steady state (Vss).
Method: In one method, Vss was allometrically predicted (using a fixed exponent 1.0 or 0.8) in a given tissue of the
rat and then Kp was predicted for that tissue. In another method, an allometric plot (Kp versus Vss) was developed to
predict Kp in a tissue of the rat. In total, Kp values were predicted for 46 drugs and 338 tissues. The predicted Kp
values by the proposed two methods were compared with the experimentally determined Kp values as well as empirically
predicted Kp values by other investigators.
Results: Comparison of the predicted Kp values by the two proposed methods with experimentally determined Kp
values indicated that 67-72% of the predicted Kp values were within two-fold prediction error. The predictive power
or accuracy of method 1 (when taken all tissues and all classes of drugs into account) was 19%, 35%, 14%, 35%,
14%, and 13% better than the methods proposed by Arundel, Berezhkovskiy, Jansson et al., Poulin et al., Poulin-
Theil, and Rodgers et al. respectively.
Conclusion: The proposed two allometric methods are slightly more accurate than other empirical methods in their
predictive performance for the prediction of tissue Kp values for acidic, weak bases and neutral drugs.