Impact of gastric mixing and water intake on postprandial pharmacokinetics of orally disintegrating tablets: A PBBM-based study

OBJECTIVE: Food intake substantially influences the oral absorption of many drugs, and the ability to quantitatively predict food effects in humans prior to clinical studies is critical for optimizing dosing regimens and formulation design. Although various in silico approaches have been proposed to predict food effects on oral solid dosage forms, their predictive performance remains limited. Orally disintegrating tablets (ODTs), which can be administered without water, are known to exhibit characteristic postprandial absorption profiles that are inadequately captured by conventional models. This study aimed to develop a novel physiologically based biopharmaceutics model (PBBM) capable of quantitatively predicting the impact of postprandial water intake on drug absorption, using tolvaptan and rivaroxaban as model compounds Methods: To better represent the postprandial gastric environment, a novel model (Assumption 2) was developed in which the stomach was functionally divided into two compartments: the fundus and the antrum. For administration without water, the drug was assumed to remain in the fundus and transfer to the antrum after a defined lag time. In contrast, for administration with water, drug distribution between the fundus and antrum was assumed, accounting for the formation of a flow channel (Magenstraße). In addition, a mathematical description of drug-chyme mixing in the antrum was incorporated to reflect dissolution under high-viscosity conditions. RESULTS AND DISCUSSION: The conventional single-compartment gastric model (Assumption 1) substantially overestimated postprandial absorption rates. In contrast, the proposed compartment-divided model (Assumption 2) markedly improved the prediction of Tmax for both tolvaptan and rivaroxaban ODTs and successfully reproduced the characteristic postprandial absorption delay observed under water-free administration. These findings demonstrate that abandoning the conventional assumption of instantaneous homogeneous gastric mixing and explicitly accounting for administration-dependent (i.e., presence or absence of water intake) gastric distribution and drug entrapment within solid food substantially improves the prediction of postprandial pharmacokinetics. The remaining tendency toward slightly earlier predicted absorption, particularly for rivaroxaban, suggests that diffusion delays in ultra-high-viscosity environments generated by solid meals are not fully captured by parameters derived from liquid meal data. Overall, the proposed model structure is mechanistically valid, and further improvements in prediction accuracy are expected through the incorporation of parameters describing particle behavior in highly viscous gastric contents, informed by complementary in vitro and in vivo studies.