A new PAT/QbD approach for the determination of blend homogeneity: combination of on-line NIRS analysis with PC Scores Distance Analysis (PC-SDA).

A novel and straightforward multivariate analytical tool for the qualitative determination of powder blend uniformity using on-line Near-Infrared Spectroscopy (NIRS) is presented. The approach combines current chemometric methods, e.g. spectral pre-processing and Principal Component Analysis (PCA), with (1) a new approach of data analysis to determine the end-point of the blending process, (2) building a design space (DS) for blend homogeneity and (3) developing a solid statistical rationale to stop blending according to Quality-by-Design (QbD) principles of FDA's Process Analytical Technology (PAT) initiative. The new approach comprises calculation of Euclidean distances between PCA scores in a multidimensional space and determination of Moving Block Standard Deviations (MBSDs) of successive Principal Component (PC) scores distances to estimate a time-window during blending where spectral variability decreases to a preset minimum. Hotelling's T(2) statistics is then used to monitor and report blend homogeneity. This technique is called "Principal Component Scores Distance Analysis" (PC-SDA). A Central Composite Design resulting in 10 batches mixed in a bin-blender (same composition, different blender fill level, different number of revolutions) was executed. NIR Chemical Imaging (NIR-CI) in combination with Symmetry Parameter Image Analysis (SPIA) was used to verify the NIRS analyzer response and assess homogeneity of all NIR-active components.