To assess the probability of bioequivalence (BE) between orally disintegrating tablets (ODTs) taken without water and conventional tablets (CTs) taken with water, an in vitro biorelevant methodology was developed using the BE Checker, which reproduces fluid shifts in the gastrointestinal tract and drug permeation. In addition to the fluid shift from the stomach to the small intestine, the process of ODT disintegration in a small amount of fluid in the oral cavity and the difference in gastric emptying caused by differences in water intake were incorporated into the evaluation protocol. Assuming a longer time to maximum plasma concentration after oral administration of ODTs taken without water than for CTs taken with water due to a delay in gastric emptying, the fluid shift in the donor chamber of the BE Checker without water was set longer than that taken with water. In the case of naftopidil ODTs and CTs, the values of the f₂ function, representing the similarity of the permeation profiles, were 50 or higher when the fluid shift in ODTs taken without water was set at 1.5 or 2 times longer than that of the CTs taken with water. The values of the f2 function in permeation profiles of pitavastatin and memantine ODTs were both 62 when the optimized experimental settings for naftopidil formulations were applied. This methodology can be useful in formulation studies for estimating the BE probability between ODTs and CTs.

Objective: The objective of this study was to develop a Physiologically Based Biopharmaceutics Modeling (PBBM) framework that can predict PK profiles in humans based on data generated from the BE Checker. Methods: Metoprolol and dipyridamole were selected as model drugs. A mathematical model was developed to describe drug dissolution, membrane permeation, and dynamic changes in pH and fluid volume within the BE Checker system. Using data generated under various experimental conditions, dissolution rate constants were estimated. For dipyridamole, the precipitation rate constant was also estimated, assuming simultaneous dissolution and precipitation processes. The estimated parameters were subsequently incorporated into the human PBBM to simulate PK profiles. Finally, the predictive accuracy of PK parameters such as Cmax and AUC was assessed. Results: For metoprolol, the PK profiles using the paddle revolution rates of 100 and 200 rpm closely matched the observed human data, particularly for Cmax and AUC, a key indicator of BE. In the case of dipyridamole, accurate predictions of the mean human PK profile were achieved when using BE Checker data obtained under high paddle speed (200 rpm) and longer pre-FaSSIF infusion times (20–30 min). Conversely, simulations based on lower paddle speed (50 rpm) and shorter pre-FaSSIF infusion time (10 min) underestimated plasma concentrations in humans. Conclusions: These findings suggest that the combination of BE Checker data acquired under high agitation conditions and the in silico mathematical model developed in this study enables accurate prediction of average human PK profiles.