Model-Based Process Challenge of an Industrial Ion-Exchange Chromatography Step



The separation of one weakly and one strongly binding impurity from a therapeutic protein by ion-exchange chromatography was studied with a mechanistic model. The effect of disturbances in eight potentially critical process parameters on two critical quality attributes was evaluated. It was found that the outgoing concentration of the weakly binding impurity was only affected by the ingoing concentration of this impurity, while the target product purity was affected by the amounts of buffer and acid used to prepare the elution buffers and the protein concentration and purity of the feed. Full factor analysis of these parameters showed that there were interaction effects between the parameters due to the pH dependence of the chromatographic separation. Sampling-based robustness analysis estimated the probability of batch failure to be above 1 %. The model improved the analysis as it was possible to study the effects of more process parameters and their interactions than would have been possible in an experimental study.