(652a) The Role of Modeling, Simulation and Optimization Approaches in the Design of Pharmaceutical Production Processes

Authors: 
Sugiyama, H., The University of Tokyo

Achieving cost-effectiveness while maintaining good manufacturing practice (GMP) is an emerging topic in pharmaceutical manufacturing. In this presentation, the role of modeling, simulation and optimization (SMO) approaches is explored by showcasing two recent research examples on active pharmaceutical ingredients (APIs) and excipients, respectively.

The first example is on optimizing the purification process of thermally unstable APIs. The target here was an industrial process for sterilizing API as the raw material of a Parenterals (or injectable) product. The process consists of dissolution of crude API with water by heat exchanger, sterile filtration, crystallization and drying, and as a product sterile API crystal is obtained. In the heat exchanger, the crude API needs to be dissolved completely while degradation reaction of the API occurs in the liquid phase. This dissolution process of thermally unstable API was modeled with the overall yield as the objective function, and the flow rate and the temperature in the heat exchanger as design variables. The process model also considered a couple of characteristics such as particle size distribution, non-linear temperature profile in the heat exchanger, and sterility requirements by GMP. The simulation result indicated an optimal set of the design variables as well as certain ranges where the yield decreases dramatically due to the degradation or non-dissolution.

The second example is on the multiobjective optimization of the use process of pharmaceutical excipients, i.e., reception, storage, testing, and application of the material. As an example, benzyl alcohol, an oxidation-sensitive material, was investigated, and a model was created, which defines the amount of oxidized impurities, and the costs associated with testing as the quality and economic objectives, respectively. The process was modelled and simulated, considering container volume and purchase frequency as the design variables, and also the use of Raman spectroscopy as an alternative to conventional manual testing. The multiobjective evaluation indicated options on the Pareto frontier of the design variables as well as the effectiveness of the use of Raman spectroscopy.

In these two examples, the SMO approaches showed its supportive role for shifting the design concept of pharmaceutical production processes from “by-testing” to “by-design”. The goal of the design can cover not only a single aspect but also to multiple ones, such as quality, economy, supply stability or risks regarding environment, health and safety.