Application of pharmaceutical quality approaches and the development of process analytical technologies to improve pharmaceutical product and process understanding

The pharmaceutical industry is undergoing a paradigm change with the advancement towards the Fourth Industrial Revolution. The manufacturing framework is moving from batch to continuous, which has been accompanied by the necessity of the implementation of process analytical technologies (PAT) for continuous process monitoring and control. There has also been a need to develop advanced quality approaches such as Quality by Design (QbD) and Quality by Control (QbC). Within these quality approaches, modelling has been used to expand knowledge related to raw material attributes and transformations that happen during the process, and to monitor and control process parameters. Besides the development of models for process and product transformations understanding and monitoring, predictive models have been an important advancement to reduce experimental cost during formulation development and process design. There are challenges that need to be addressed in order to complete the implementation of continuous process and advanced quality approaches in that pharmaceutical sector. It is necessary to extend the full mechanistic knowledge of processes, to develop data analysis and computer simulation, to expand the number of equipment and models to scale up solutions, and to upskill personnel to work with the technologies required. The general aim of this thesis was to increase understanding of pharmaceutical materials and processes by the development of PAT and models for quantitative characterization and prediction of drug substance and drug product critical properties. To achieve this aim, several specific objectives were established focused on different aspects of oral dosage form manufacture. Initially, a PAT was developed to rapidly characterize microcrystalline cellulose (MCC) percentage crystallinity index. MCC crystallinity was determined from Raman spectral data. Then, MCC moisture sorption and moisture monolayer were determined from isotherms obtained using dynamic vapor sorption (DVS). A correlation between both attributes was investigated. However, it was not possible to confirm the relationship between the monolayer moisture sorption and the percentage crystallinity. A total of 30 commercial batches of MCC were used to design and validate the MCC crystallinity model. Furthermore, a web application, McCrystal, was designed to disseminate the model. Available at https://sspc.ie/mccrystal/. Secondly, a PAT was developed to determine the endpoint and monitor form changes during slurry co-crystallization. This PAT tool involved the application of in situ Raman spectroscopy combined with principal component analysis (PCA). An additional PAT tool was developed using Mid Infrared spectroscopy (MIRS) combined with multivariate curve resolution (MCR) in order to quantitatively evaluate phase purity of co-crystals produced by slow evaporation and slurry co-crystallization. The models developed were applied to different combinations of active pharmaceutical ingredients (APIs) and coformers, and were able to determine the co-crystallization endpoint for all systems where the solvents exhibited a weak Raman signal. Moreover, the phase purity for all co-crystals systems investigated was determined. The third challenge addressed by this study relates to particle-particle interactions in powder blends to predict the general behaviour of the blend when it is directly compressed into pharmaceutical tablets. Percolation threshold model was used to determine a critical drug load, above which the powder blend behaves similarly to the drug substance. As a result, the formulation may show undesirable properties above the threshold, which can lead to non-compliance or process difficulties. The application of PCA to determine the critical drug loading was investigated. The predicted values of critical drug loading were confirmed by changes in powder flowability, tablet weight variation, and Raman spectroscopic analysis of drug distribution. PCA was not able to precisely predict the threshold, however, it aided in clarifying differences between the blends containing different ibuprofen loadings and different MCC grades. Expanding on the impact of critical drug loading on tablet critical quality attributes, the final objective was to investigate the impact of critical drug loads, determined from compaction data, on disintegration and dissolution performance. The application of novel process analytical technologies, in-process video microscopy (PVM) and focused-beam reflectance measurement (FBRM), to study tablet disintegration was also proposed and investigated in-depth. FBRM and PVM provided important information of tablet disintegration behaviour that cannot be captured by the traditional disintegration testing. The critical drug load predicted from compaction data was also relevant for disintegration and dissolution behaviour. Above the threshold drug loading, disintegration and drug release rates were decreased. A PCA model developed from Raman spectra acquired from tablet surfaces further supported the finding that blends manufactured with different grades of MCC differed more significantly above the critical drug load. The work presented in this thesis is a contribution to solid dosage formulation design, co-crystallization understanding, tableting process scale up and scale down, and data handling. The models and PAT developed can be used to predict formulation performance, determine process parameters based on raw material critical attributes, determine process endpoint, and thoroughly assess final product critical quality attributes