Abstract:
In this dissertation it is shown how the systematic use of computional materials science molecular simulation can on one hand lead to an optimization of the experiments done and on the other hand to a reduction of the amount of excipients and active pharmaceutical ingredient (API) needed for the development of a formulation. This is achieved by the calculation of the partial solubility parameters and the subsequent calculation of the free energy change of mixing of hypothetical mixtures of the included substances. Based on molecular dynamic simulation (MD), the calculation shows the best miscibility of the plasticiser triethyl citrate (TEC) and the applied cationic polymethacrylate Eudragit RS (ERS), whereas the formation of a solution of the API theophylline (THEO) and ERS is not possible. The combination of these three substances promises the most stable incorporation of high amounts of crystalline API into a polymermatrix that is stable against phase separation. These findings based on energetic calculations are verified by experimental results.
Subsequently using the matrix polymer ERS and the plasticiser TEC in hot melt extrusion, sustained release matrix pellets containing 50 to 70 % (w/w) THEO are produced to analyse possible influences on the dissolution properties of the dosage form. It is shown that the release of the API is, resulting from the production method, anisotropic. The remaining factors influenceing the dissolution properties are shown to be: pellet dimensions, glass transition temperature Tg of the matrix and the amount of the incorporated API. Depending on the factors, the cumulative release of THEO after 8 hours of dissolution is between 30 and 100 %. Using the Weibull distribution function, the complete dissolution data is parametrised without loss of any data.
The extent of the effect of pellet dimensions and the Tg of the matrix on the parametrised data is analysed and quantified. Based on each quantification, a combined drug release prediction model is exemplarily calculated for 60 % of incorporated THEO. The prediction model allows, distinctive changes of the resulting dissolution profiles in an expansive range. The prediction model is verified by using pellets that were not used for the development of the model. Using MD simulations, an additional prognostic tool for the Tg of plasticiser containing polymermatrices is developed. Hence, an analysis of the influence of changes within the matrix composition can be performed without the need of experiments.
The combination of the experimental and computional methods within this work enables a comprehensive and reliable characterisation of the prepared dosage form. By applying the two computional simulation methods (mixture analysis and calculation of Tg) it is possible, not only to choose the best excipient, but also to estimate the release properties of virtual compositions. This is possible at a very early point of drug dosage form development where no sufficient amount of API is available. The semiempirical API release prediction model enables the description of the sustained release matrix system with a reduced number of experiments, while still being related to the underlying physical diffusion process.