Small scale design of experiment investigation of equilibrium solubility in simulated fasted and fed intestinal fluid.

ABSTRACT

It is widely recognised that drug solubility within the gastrointestinal tract (GIT) differs from values determined in a simple aqueous buffer and to circumvent this problem measurement in biorelevant fluids is determined. Biorelevant fluids are complex mixtures of components (sodium taurocholate, lecithin, sodium phosphate, sodium chloride, pancreatin and sodium oleate) at various concentrations and pH levels to provide systems simulating fasted (FaSSIF) or fed (FeSSIF) intestinal media. Design of Experiment (DoE) studies have been applied to investigate FaSSIF and FeSSIF and indicate that a drug's equilibrium solubility varies over orders of magnitude, is influenced by the drug type and individual or combinations of media components, with some of these interactions being drug specific. Although providing great detail on the drug media interactions these studies are resource intensive requiring up to ninety individual experiments for FeSSIF. In this paper a low sample number or reduced DoE system has been investigated by restricting components with minimal solubility impact to a single value and only investigating variations in the concentrations of sodium taurocholate, lecithin, sodium oleate, pH and additionally in the case of fed media, monoglyceride. This reduces the experiments required to ten (FaSSIF) and nine (FeSSIF). Twelve poorly soluble drugs (Ibuprofen, Valsartan, Zafirlukast, Indomethacin, Fenofibrate, Felodipine, Probucol, Tadalafil, Carvedilol, Aprepitant, Bromocriptine and Itraconazole) were investigated and the results compared to published DoE studies and literature solubility values in human intestinal fluid (HIF), FaSSIF or FeSSIF. The solubility range determined by the reduced DoE is statistically equivalent to the larger scale published DoE results in over eighty five percent of the cases. The reduced DoE range also covers HIF, FaSSIF or FeSSIF literature solubility values. In addition the reduced DoE provides lowest measured solubility values that agree with the published DoE values in ninety percent of the cases. However, the reduced DoE only identified single and in some cases none of the major components influencing solubility in contrast to the larger published DoE studies which identified multiple individual components and component interactions. The identification of significant components within the reduced DoE was also dependent upon the drug and system under investigation. The study demonstrates that the lower experimental number reduces statistical power of the DoE to resolve the impact of media components on solubility. However, in a situation where only the solubility range is required the reduced DoE can provide the desired information, which will be of benefit during in vitro development studies. Further refinements are possible to extend the reduced DoE protocol to improve biorelevance and application into areas such as PBPK modelling.