Modeling the Simultaneous Effects of Co-solvent and Polymer on the Solubility of Drugs

Abstract

Having knowledge about the solubility of drugs is important in pharmaceutical sciences and one of the common methods to increase the solubility of low soluble drugs is cosolvensy. In this work the experimental solubilities of acetaminophen, ibuprofen, diazepam, lamotrgine, clonazepam, pioglitazone-HCl and glibenclamide are reported in binary and ternary, aqueous and non-aqueous mixtures of CMC, PVP, PEGs 600, 400 and 200 with NMP, PG and ethanol along with the density of the solute free and saturated solutions. Cosolvents addition to their effect on the solubility, influence the stability of the drugs, so for investigating the effects of the cosolvents on drugs stability ibuprofen was chosen for thermal stability analysis at four temperatures by HPLC system. In modeling part, three models have been used 1) JouybanتAcree model 2) Combined version of the Jouyban-Acree model and the Hansen solubility parameters 3) ANN model. For the models 1 and 2, first the binary solvents data were fitted to the models and the sub-binary constants were produced, then by using these constants and fitting the ternary solvents data to the models the ternary constants were calculated, and the fitting OMRD for models 1 and 2 were 64.5 % and 48.0 %, respectively. For ANN method the mass fractions of the cosolvents, the logarithm of the mono-solvents solubility data and the special composition of the Hansen solubility parameters were included as the input data and the logarithm of the solubility in the mixture was as output data, the fitting OMRD for this part was 32.3 %. For the densities, by fitting the solute free solutions density several trained version of the Jouyban-Acree model have been proposed. The fitting OMRD value for all data points was 0.6%. Then the saturated solutions densities have been predicted using the trained versions, the prediction OMRD value was 1.7% for all density data points. The predicted and experimental saturated solutions densities have been used to convert the reported molar solubilities into mole fractions, and the differences between these two groups have been computed by MRD values, and the OMRD for all density data points was 2.4%. For the reported stability data points, the Jouyban-Acree model was used and the prediction OMRD was 22.4%.Also, as adjuant works two projects were done along with the thesis; 1) Investigations on the retention factors of some benzodiazepine analytes and modeling the data by the Jouyban-Acree model. 2) Proposing four trained versions of the Jouyban-Acree model for calculating the dielectric constants of the aqueous and non-aqueous binary solvents and using the calculated dielectric constants in predicting the drugs solubility.