(426b) Building the Predictive Process Engineering Toolbox to Define Mixing Parameters for Early-Stage Drug Product Manufacturing of Large Molecules | AIChE

(426b) Building the Predictive Process Engineering Toolbox to Define Mixing Parameters for Early-Stage Drug Product Manufacturing of Large Molecules

Authors 

Kuo, R. - Presenter, Merck & Co., Inc.
Kalal, Z., Merck & Co.
Deshmukh, S., Bristol-Myers Squibb
Flamm, M., Merck & Co., Inc.
Mittal, S., Merck & Co., Inc.
Procopio, A., Merck & Co.
Mixing is an integral operation used throughout the manufacturing process train of pooling, formulation, and primary container closure filling for large molecule sterile drug products. As such, evaluating appropriate mixing parameters and associated ranges is necessary to understand their impact on product uniformity and physical stability. For this work, a broad toolbox of at-scale, computational, and scale-down approaches were used to demonstrate successful mixing operating ranges. Saline and water were used as a model system to evaluate mixing at-scale (65 L) with agitation rates ranging from 37 to 100 RPM. In parallel, computational fluid dynamics (CFD) was used to predict at-scale homogeneity with further refinement and validation from the at-scale mixing results. It was observed that an agitation rate of 80 RPM yielded a uniform mixture within 30 minutes. At these mixing conditions, the computational model predicted maximum shear rates in the impeller region to be approximately 1,000 s-1. A scale-down approach was developed using a rheometer to subject active large molecules to uniform shear rates based on the predicted range. The stressed molecules were then analyzed for particle aggregates as an indication of physical stability at shear rates similar to those shown in the CFD models. This scale-down approach is particularly useful during early development to evaluate the shear sensitivity of large molecules under anticipated mixing conditions when material quantities are limited. Setting the operating parameters for the scale-down model required the predictions from the CFD models, which were optimized with data collected from at-scale mixing. When used together, the toolbox of at-scale, computational, and scale-down approaches implemented within this work are broadly applicable to evaluate and define successful mixing parameters for the manufacture of large molecule sterile drug products.