Statistical investigation of simulated intestinal fluid composition on the equilibrium solubility of biopharmaceutics classification system class II drugs.

A drug's solubility and dissolution behaviour within the gastrointestinal tract is a key property for successful administration by the oral route and one of the key factors in the biopharmaceutics classification system. This property can be determined by investigating drug solubility in human intestinal fluid (HIF) but this is difficult to obtain and highly variable, which has led to the development of multiple simulated intestinal fluid (SIF) recipes. Using a statistical design of experiment (DoE) technique this paper has investigated the effects and interactions on equilibrium drug solubility of seven typical SIF components (sodium taurocholate, lecithin, sodium phosphate, sodium chloride, pH, pancreatin and sodium oleate) within concentration ranges relevant to human intestinal fluid values. A range of poorly soluble drugs with acidic (naproxen, indomethacin, phenytoin, and piroxicam), basic (aprepitant, carvedilol, zafirlukast, tadalafil) or neutral (fenofibrate, griseofulvin, felodipine and probucol) properties have been investigated. The equilibrium solubility results determined are comparable with literature studies of the drugs in either HIF or SIF indicating that the DoE is operating in the correct space. With the exception of pancreatin, all of the factors individually had a statistically significant influence on equilibrium solubility with variations in magnitude of effect between the acidic and basic or neutral compounds and drug specific interactions were evident. Interestingly for the neutral compounds pH was the factor with the second largest solubility effect. Around one third of all the possible factor combinations showed a significant influence on equilibrium solubility with variations in interaction significance and magnitude of effect between the acidic and basic or neutral compounds. The least number of significant media component interactions were noted for the acidic compounds with three and the greatest for the neutral compounds at seven, with again drug specific effects evident. This indicates that a drug's equilibrium solubility in SIF is influenced depending upon drug type by between eight to fourteen individual or combinations of media components with some of these drug specific. This illustrates the complex nature of these fluids and provides for individual drugs a visualisation of the possible solubility envelope within the gastrointestinal tract, which may be of importance for modelling in vivo behaviour. In addition the results indicate that the design of experiment approach can be employed to provide greater detail of drug solubility behaviour, possible drug specific interactions and influence of variations in gastrointestinal media components due to disease. The approach is also feasible and amenable to adaptation for high throughput screening of drug candidates.