Tablet production is a complex process that can be influenced by many interrelated factors. Even minor variations in the raw materials and/or the manufacturing processes can affect the quality of the final product. Thus, it is essential to set the material specifications and process parameters correctly. The key research objective was to understand the influence of physicochemical properties of the raw materials and the process parameters of roller compaction and compression on dissolution as relevant biopharmaceutical parameters of immediate-release tablets. This study aimed initially to optimize the industrial tablet-manufacturing process using multivariate analysis. Partial least squares method was retrospectively applied to the data obtained from two years production, to study the influence of 90 factors on tablets that contained two active pharmaceutical ingredients. Eight critical material attributes and four critical process parameters were identified. With the optimal settings of the tablet compression as the final process step, the combined variability of all of the materials and previous process steps was compensated. Remarkably, our findings were supported with process validation on an industrial scale. In the last part of the study, artificial neural networks were used on the same data set to identify non-linear relationships, which are common for pharmaceutical processes. We identified four new critical material attributes and three new critical process parameters, not being identified by multivariate analysis. This technique assures better process understanding and creates opportunities for continuous improvements. Demonstrated approach shows many benefits and can be transferred to other pharmaceutical processes. Improved quality of the drug will decrease number of batches of insufficient quality, lower costs for drug production and provide better patient care.
|
