Relationship between dissolution efficiency of Oxazepam/carrier blends and drug and carrier molecular descriptors using multivariate regression analysis

Quantitative structure-property relationships were developed for predicting the enhancement of dissolution rate of the model lipophilic drug Oxazepam (Oxa) from blends (BLs) with 12 structurally different carriers at three different drug/carrier weight ratios (1/5, 1/10, and 1/20). To this end, 36 BLs were prepared by the solvent-evaporation method and characterized by spectroscopic (FT-IR), thermoanalytical (DSC) and X-ray diffraction studies. The dissolution rate of the examined systems was quantified by log DE/DEOxa, where DE and DEOxa are the dissolution efficiencies of the BL and pure drug, respectively. Twenty molecular descriptors, including parameters for size, lipophilicity, cohesive energy density (CED), and hydrogen bonding capacity were calculated and together with the experimental melting point (MP), were used in multivariate analysis. Twelve pertinent variables were detected after looking at the results of principal component analysis (PCA) and cluster analysis, and reliable six-descriptor models generated by Partial Least Squares-Projection to Latent Structures (PLS) method. Satisfactory coefficient of determination values were obtained (i.e., R-2 equal to 0.794 and Q(2) equal to 0.705). The equations generated can predict with reasonable accuracy the dissolution rate increase of the model lipophilic drug/carrier BLs. (C) 2008 Elsevier B.V. All rights reserved.