Dissolution behavior of various drugs in different FaSSIF versions

Biorelevant media have proven to be useful in predicting the performance of poorly soluble drugs in the gastrointestinal tract. Several versions of fasted state simulated intestinal fluids have been published and compared with respect to their physical chemical properties and solubilization of drugs. However, to date there have been no reports in the literature comparing dissolution of poorly soluble drugs in these media. In this study eleven BCS Class II compounds (five nonionized compounds, three weak bases and three weak acids) were investigated with respect to their thermodynamic solubility and dissolution behavior in three biorelevant media simulating conditions in the small intestine (FaSSIF V1, FaSSIF V2 and FaSSIF V3). It was shown that the maximum percentage release of drugs from their commercial formulations can differ from the results for the thermodynamic solubility of the pure drug; these differences can be largely attributed to API presentation, composition of the formulation and manufacturing effects. The results were additionally compared with data for solubility in HIF taken from the literature in order to determine which version of FaSSIF most closely corresponds to the physiological conditions. The different versions of FaSSIF are able to achieve solubility results similar to those in HIF, with closest results generally achieved in FaSSIF V1. The magnitude of solubility/dissolution differences among the three FaSSIF versions is dependent on the drug's characteristics. In the case of weakly basic compounds, the differences among the FaSSIF versions are minor. For weakly acidic compounds the behavior in the different versions is primarily pH dependent and influenced by the lipid composition of the FaSSIF only to a minor extent. The differences in solubility and dissolution of the nonionized compounds among the three versions of FaSSIF becomes apparent above a log P value of 2.5, with larger differences among the versions at high log P values.