Mechanistic prediction of food-induced variability in drug release and pharmacokinetics using a dynamic gastrointestinal system and convolution-based modeling.

Oral sustained-release (SR) formulations are widely used in clinical practice, yet reliable prediction of their in vivo performance remains challenging due to the dynamic and heterogeneous gastrointestinal (GI) environment. Here, a bioinspired dynamic human stomach-intestine system (DHSI-IV) was used to simulate physiologically relevant GI environments under different prandial states. Ibuprofen SR capsules were evaluated under fasted (water), milk, carbonated beverage, and high-fat meal conditions. The DHSI-IV reproduced pronounced prandial-dependent GI dynamics, with gastric half-emptying times ranging from 16.7 min (water) to 90.2 min (high-fat meal). These differences led to distinct dissolution behaviors, with cumulative release of 42.1 ± 1.5% (water), 33.15 ± 2.05% (carbonated beverage), 41.95 ± 3.35% (milk), and 44.35 ± 1.45% (high-fat meal). Coupling in vitro dissolution data with a simplified convolution-based pharmacokinetic (PK) back-calculation approach enabled prediction of systemic exposure. Predicted fasted-state PK parameters agreed with clinical data (Tmax 4.0 ± 0.5 h vs. ∼ 4.5 h; Cmax 12.5 ± 2.7 μg/mL vs. 13.88 ± 2.6 μg/mL). Under fed (high-fat meal) conditions, the model captured delayed absorption (Tmax 5.5 ± 0.5 h vs. ∼ 6.0 h) and increased exposure (Cmax 20.2 ± 2.3 μg/mL; AUC0-24h 153.7 ± 22.6 μg·h/mL). Overall, DHSI-IV provides a predictive platform for evaluating SR formulations under physiologically relevant GI conditions, supporting food-effect assessment and formulation optimization.