Predicting passive intestinal absorption using a single parameter

A model is proposed for the prediction of either high or low fraction absorbed for an orally administered, passively transported drug on the basis of a new absorption parameter, ∏. The model includes only two inputs: the octanol-water partition coefficient (K_ow) and the dimensionless oversaturation number (O_Lumen). The latter is the ratio of the concentration of drug delivered to the gastro-intestinal (GI) fluid to the solubility of the drug in that environment. Thus, O_Lumen is equal to the dose-normalized solubility for suspensions and unity for solutions. The value of ∏ increases with an increase in K_ow and a decrease in O_Lumen for suspensions, and is equal to K_ow for solutions. The effectiveness of the model is validated using experimental human gastrointestinal absorption data for 98 compounds. About 88% of these drugs are correctly predicted to be either well absorbed or poorly absorbed based solely upon whether their ∏ value is greater than or less than unity. Thus, the use of a single absorption parameter, ∏ provides a simple means to estimate whether or not an orally administered drug undergoing passive transport will be absorbed efficiently. The advantage of this parameter is that it is based upon simple, easily measured (or calculated) physical chemical data. It is especially noteworthy that experimental measurement of in vitro membrane transport is not required. The model based on the new absorption parameter is shown to have wider applicability than current available models for predicting the fraction absorbed.