A New Method of Constructing a Drug-Polymer Temperature-Composition Phase Diagram Using Hot-Melt Extrusion.

Current experimental methodologies used to determine the thermodynamic solubility of an API within a polymer typically involves establishing the dissolution/melting end point of the crystalline API within a physical mixture or through the use of the glass transition temperature measurement of a demixed amorphous solid dispersion. The measurable "equilibrium" points for solubility are normally well above the glass transition temperature of the system, meaning extrapolation is required to predict the drug solubility at pharmaceutically relevant temperatures. In this manuscript, we argue that the presence of highly viscous polymers in these systems results in experimental data that exhibits an under or overestimated value relative to the true thermodynamic solubility. In previous work, we demonstrated the effects of experimental conditions and their impact on measured and predicted thermodynamic solubility points. In light of current understanding, we have developed a new method to limit error associated with viscosity effects for application in small-scale hot-melt extrusion (HME). In this study, HME was used to generate an intermediate (multiphase) system containing crystalline drug, amorphous drug/polymer-rich regions as well as drug that was molecularly dispersed in polymer. An extended annealing method was used together with high-speed differential scanning calorimetry to accurately determine the upper and lower boundaries of the thermodynamic solubility of a model drug-polymer system (felodipine and Soluplus). Compared to our previously published data, the current results confirmed our hypothesis that the prediction of the liquid-solid curve using dynamic determination of dissolution/melting end point of the crystalline API physical mixture presents an underestimation relative to the thermodynamic solubility point. With this proposed method, we were able to experimentally measure the upper and lower boundaries of the liquid-solid curve for the model system. The relationship between inverse temperature and drug-polymer solubility parameter (χ) remained linear at lower drug loadings. Significantly higher solubility and miscibility between the felodipine-Soluplus system were derived from the new χ values.