Mathematical Modeling and Optimization to Inform Impurity Control in an Industrial Active Pharmaceutical Ingredient Manufacturing Process

Mathematical modeling of pharmaceutical manufacturing processes can provide insights and understanding regarding the key factors impacting product quality. In this study, we describe the development of a dynamic model for a stage in an active pharmaceutical ingredient (API) manufacturing process, its calibration and validation versus industrial experimental data, and its use to address three objectives: (1) assessment of process operating parameter criticality on key performance indicators (KPIs); (2) confirming whether the considered process operating space safely respected limits of critical quality attribute (CQA) impurities; and (3) finding process setpoints that can potentially improve the KPIs. Objective 1 used global sensitivity analysis (GSA) to find that only operating parameters associated with the reactor were significant. Objectives 2 and 3 used nonlinear optimization, confirming that impurity limits are respected at any point in the considered process operating space and suggesting a shifted process setpoint that could allow enhanced yield (∼4% absolute increase) and reduced impurity content (∼0.5 mol % absolute reduction).