Release testing is used in the development of new medicines, for predicting in vivo behaviour of pharmaceutical forms and active substances, and in quality control.
The purpose of our work was the evaluation of the flow-through cell method, Apparatus 4 (Ph Eur and USP), to monitor the release of poorly soluble active substance from a pharmaceutical form with immediate release, using a statistical approach in experiment planning, for which the MODDE program was used.
Various factors were changed, as were the interactions between them: two tablets with varying content of excipients (T1 and T2); four different dissolution media
(water, potassium phosphate buffer pH 6.8, acetate buffer pH 4.5, and simulated gastric fluid pH 1.2); the flow through the cells was set to 4, 8 and 16 mL/min; two different cells were used: a smaller one with the diameter of 12 mm and a bigger one with the diameter of 22.6 mm. For the analysis, an already developed method of analysis was used (High-performance liquid chromatography). We also specified the solubility of the model active substance in the presence of all components of the tablets. The design of the experiment included a selection of 28 best experiments for solving our problem. In addition, four more experiments were conducted to compare the release from T1 and T2. In all tests, the adequacy of the flow through the cells was monitored. The results were sorted and entered in the MODDE program. The program produced a model, which we also interpreted. This was followed by the use of the model.
The model was compatible with the responses (percent of model substance released in a certain time interval). In almost every time interval, the model had an excellent predictive power, sufficient validity and very good repeatability. The interpretation showed, that the media for release testing, most notably water, affect the responses the most. The flow through the cells and the interaction between the flow and the individual media also have a great influence. The model was used to calculate the responses for all combinations of factors. Using these data, the ratios between the percent of released active substance from T1 and T2 were calculated. The ratios calculated from the predicted data were compared with the ratios calculated from the experimental data. The best match was obtained for the medium of water and the medium of potassium phosphate buffer pH 6.8. We also found that the most comparable results with in vivo data were obtained by experiments of release testing in water and potassium phosphate buffer pH 6.8 The examination of all the ratios (predicted responses) showed that the in vivo data were most closely approximated by using a smaller cell with the diameter of 12 mm and adjusting the higher flow rate for release testing.
|
