Detecting Chromatography Unit Degradation : Comparison of single- and multi-point techniques implemented in system control and monitoring software

Abstract: Chromatography units, used in the production of pharmaceuticals, degrade with use and need to be changed or repackaged. This study investigates the effectiveness of two statistical methods, principal component analysis and simple and one-point multiparameter technique, for determining degradation in the Fibro chromatography unit. The methods have been shown to be effective on resin chromatography columns but not before tested on the relatively new Fibro chromatography unit. The statistical methods are implemented in an unreleased version of the monitoring and control software Unicorn. This implementation aims to be a proof of concept for including more complex methods for monitoring runs directly in the software, easing the workflow of operators by removing the need to export measurements to a third-party program. The methods were tested on measurements of absorbance, conductivity, and pressure from two series of chromatograms performed on two Fibro chromatography units. One of the units was defective and broke down halfway through the series. Principle component analysis could clearly visualize a difference between early and late runs on the defective unit. The same could only be achieved for the non-defective unit by excluding measurements of pressure. Simple and one-point multiparameter technique visualized trends from early to late in the series which were much clearer for the defective unit. Both methods showed signs of predicting degradation in a Fibro chromatograpy unit but require validation on chromatogram series with more direct measurements of performance and a wider range of failure causes.