A Combined Experimental and Predictive Modeling Study of Diltiazem Release From Natural Polymer Microcapsules in Different Media

Abstract

The release behavior of diltiazem-loaded microcapsules prepared from gelatin B and gum arabic was evaluated experimentally and mechanistically in distilled water and simulated saliva. Release amounts were quantified by liquid chromatography over 5–720 min and used directly as input for kinetic and mechanistic modeling. Kinetic analysis was performed using different models, and model performances were compared using the Akaike and Bayesian information criteria (AIC, BIC). The Weibull model provided the best fit in both media (R² = 0.989 for distilled water, R² = 0.994 for simulated saliva). A two-compartment mechanistic model developed in SimBiology/MATLAB yielded release rate constants of 0.0111 min⁻¹ (water) and 0.0026 min⁻¹ (saliva). Agreement between the Weibull mechanistic model and experimental results was R² = 0.989 for distilled water (RMSE = 0.524 mg) and R² = 0.994 for simulated saliva (RMSE = 0.348 mg). Model validation was performed using Visual Predictive Check (VPC) with N = 200 virtual individuals, confirming that all observed data points fell within the 5%–95% percentile prediction interval. The validated model was further extended to a sublingual pH variability scenario (pH 6.2–7.4), demonstrating that saliva pH influences cumulative release through its effect on the unionized fraction of diltiazem.