Solubilization by cosolvents. Establishing useful constants for the log-linear model.

The purpose of this study was to develop constants for the log-linear cosolvent model, thereby allowing accurate prediction of solubilization in the most common pharmaceutical cosolvents: propylene glycol, ethanol, polyethylene glycol 400, and glycerin. The solubilization power (sigma) of each cosolvent was determined for a large number of organic compounds from the slope of their log-solubility vs. cosolvent volume fraction plots. The solubilization data at room temperature were either experimentally determined or obtained from the literature. The slopes of the nearly linear relationship between solubilization power and solute hydrophobicity (logK(ow)) were obtained by linear regression analysis for each considered cosolvent. Thus, knowing or calculating a compound's partition coefficient is all that is needed to predict solubilization.