Rational design of multicomponent rivaroxaban solid dispersions using thermodynamic and statistical modeling

Abstract

The aim. The aim of the work is to study the effect of hydrophilic polymers (PVP K-12, PVP K-17, HPC) and excipients (sucrose, mannitol) on rivaroxaban supersaturation, as well as to apply thermodynamic modeling to determine the optimal composition and evaluate the robustness of ternary solid dispersion systems (SDS). Materials and methods. Rivaroxaban, polyvinylpyrrolidone (PVP K-12, K-17), hydroxypropyl cellulose (HPC), sucrose, and mannitol were used to obtain SDS. Fibers were obtained by centrifugal melt spinning at 160 - 200 °C. Equilibrium solubility was determined spectrophotometrically (250 nm) at 25 °C. A log-linear model and Jouyban-Acree polynomials were applied for optimization. The stability of the systems against fluctuations was evaluated using parametric bootstrapping and Monte Carlo simulation. Results. The intrinsic solubility of the drug is 0.017 g/L. The addition of excipients to PVP matrices promoted a nonlinear increase in solubility, with the PVP K-17 and sucrose composition exhibiting the highest effect. In HPC matrices, excipients decreased solubility due to gelation kinetics. According to the Akaike information criterion, the full Jouyban-Acree model was proven to be optimal for describing the systems. Modeling allowed shifting the theoretical optimum of the PVP K-17 + Sucrose composition to a ratio of 5:83:12 with a maximum calculated solubility of 1.470 g/L. Bootstrap analysis confirmed the thermodynamic stability of this system even under 5% technological noise, whereas HPC matrices demonstrated significant variability. Conclusions. The integration of the in silico approach within the Quality by Design framework effectively optimizes the composition of multicomponent rivaroxaban SDS. PVP matrices with sucrose increased the API solubility by more than 86 times. The phenomenological Jouyban-Acree model combined with statistical simulation minimizes the volume of routine testing and reliably predicts the stability of systems during their industrial scale-up