Purpose: Previous investigations have shown that the Rodgers and Rowland method [Rodgers 2007] for prediction of tissue:plasma partition coefficients (Kps) provides good prediction for compounds with low to moderate lipophilicity, but it often fails when applied to highly lipophilic compounds. The reasons for the unreasonably high Kp predictions for lipophilic compounds were investigated.
Methods: The effects on errors in predictions of experimental measurements of logP, pKa, Fup and Rbp on the accuracy of Kp prediction were evaluated. The main focus was on prediction of Kps, and the resultant volume of distribution, using the Rodgers & Rowland method for highly lipophilic compounds. The study revealed that this method tends to overpredict Kps especially for lipophilic compounds which also have fairly high measured fraction unbound in plasma (Fup). This could be due to the inability of current experimental techniques to capture the possible binding of drug to plasma lipids in Fup measurements. We have derived an equation which corrects the experimental Fup for binding to plasma lipids, assuming that the experimental Fup is an accurate estimate of drug binding to plasma proteins, and that octanol/water partition coefficient (logP) can be used as a surrogate for the description of drug partitioning to plasma lipids
American Association of Pharmaceutical Scientists (AAPS), November 17-19, 2008, Atlanta, GA
By Viera Lukacova, N. J. Parrott, T. Lavè, Grace Fraczkiewicz, Michael B. Bolger, Walter S. Woltosz