Fermentation Optimization: Using Comparative Statistics To Enhance Large-Scale Process Productivity

When manufacturing pharmaceutical compounds, process optimization often begins at scale-up. But it should not end there. To maximize yields, process improvement must continue throughout a compound’s lifecycle. By their very nature, new processes are subject to variability that cannot be fully understood until they are run at commercial scale over time. Rigorous in-process monitoring and ongoing statistical analysis of performance at scale are critical to establishing effective process control and robust production yields.

In this article, we focus on fermentation as part of API synthesis at Pfizer’s Kalamazoo, Michigan, U.S.A., facility, although we apply the same basic approach to most processes used at the site. Enzymes generated by large-scale (>100,000 liters) bioproduction methods are often subject to a great degree of yield variability due to the nature of the organisms used to generate a protein of interest. In this example, we demonstrate how we identified key variables that enabled us to improve control and output for a current fermentation process.