(547b) PAT Development Strategy in the Feed Frame As an Enabler for Faster Continuous Manufacturing Implementation

Continuous Manufacturing has been slowly but steadily adopted by many pharmaceutical companies in the development and manufacturing of their products. The control strategy is among the topics that are most discussed, although its definition is very dependent of the equipment design and formulation. Despite different approaches, the use of Process Analytical Technology (PAT) equipment (e.g., Near Infrared - NIR) generally plays an important role on the strategy, as it provides a fast and accurate measurement of the product in different stages of the process. Among them, a PAT equipment installed in the feed frame of a tablet press is very popular and subject of many studies as it is commonly used to monitor and control the blend uniformity / assay. Therefore, as a critical component in the overall control strategy of a continuous tableting process, the high impact chemometric model (typically quantitative) must be thoroughly developed and validated. In this sense, it is fundamental during the development stage to capture as much variability as possible to enhance model robustness that will prevent model downtime when implemented - a chemometric model that needs frequent redevelopment / revalidation leads to low technology adherence and has an important business impact as the control strategy highly relies on it. On the other hand, the (low) availability of Active Pharmaceutical Ingredient (API) during development, the different concentrations of powder required to build a calibration curve and the variability desired to achieve robustness (API – e.g., particle size – and process – e.g., sample presentation) presents a challenge to the scientist responsible for model building. Taking this into consideration, the development of chemometric models in a scaled down laboratory equipment that mimics the feed frame of a tablet press (i.e. the feed frame simulator), presents many advantages: a) reduction of API requirements and other excipients – it is possible to have a recirculation close system with a limited amount of powder (100-200g, depending on the powder density), under a constant paddle rotation; b) possibility to replicate several process conditions occurring in the normal operation of the commercial tablet press – mainly related to sample presentation to the probe, such as different volume inside the feed frame, paddle speed and distance of paddle to probe; c) reduction of commercial line occupancy; d) possibility to easily evaluate the suitability and selection of the PAT equipment to be chosen to monitor the process. On the other hand, some disadvantages of this development strategy might also occur: a) although the simulator intends to mimic the environment in the feed frame of the tablet press, there is a difference in the powder dynamics between a recirculation closed system (simulator) and the continuous flow of a tablet press which might reduce representativity of sample presentation; b) a closed system as in the simulator might present limitations of powder reuse, especially in formulations with a tendency to segregate, which leads to a more API consumption or time consuming experiments (constant re-homogeneity of the blend in use is required); c) if different PAT equipment is used in the simulator (for model calibration) and in the tablet press (for model validation), and especially for the typical diode array NIR equipment usually implemented in the tablet press equipment, direct model validation might not be possible due to intra-equipment variability. Therefore, if observed, all these factors might trigger the need to perform experiments in the commercial setup (tablet press feed frame interfaced with the GMP qualified PAT equipment that will be used in the process). Nevertheless, the extend of those experiments might be significant lower if a preliminary calibration in the simulator is performed.

In this study, it is presented the development strategy of a quantitative chemometric model, using the NIR SentroPAT FO coupled with a SentroProbe DRLS from Sentronic as a PAT equipment, interfaced in the feed frame simulator from Expo Process Analytics and in a Model-P / Model S tablet presses from GEA. The results intend to demonstrate in real cases the advantages and limitations mentioned above and the points to consider when this strategy for chemometric model development is used. Despite some disadvantages that are harder to overcome, the strategy is very beneficial for key points that are critical during product development.