Identifying sources of batch to batch variation in processability

The source of variation in the full scale processability of 131 batches of the active pharmaceutical ingredient 5-aminosalicylic acid was investigated. The variation in processability, seen in the amount of granulation liquid needed for extrusion, was found to be related to a difference in the combined effects of particle size and packing behaviour. Interactions of particle size, specific surface areas and packing behaviour caused the variation in the individual variables to be mistakenly considered unimportant. Instead, multivariate analysis had to be introduced in order to realise their effect. The combination of the packing related compressed density and the 90% percentile diameter of the volume distribution provided especially good separation of the batches according to processability. Low-pressure compression in combination with particle size data, measured by laser diffraction, was found to be a quick, powerful and relevant tool for powder characterisation. The combination allows quick screening of many batches, thereby aiding rational selection of representative samples for further investigations. The results strongly support use of multivariate analysis for investigation of sources of batch to batch variation in processability.

Graphical abstractBatch to batch variation in processability of 131 full scale batches of an active pharmaceutical ingredient was studied. Multivariate analysis was used to identify the source of the variation. The variation was attributed to differences in the combined effect of particle size and packing behaviour. A combination of low-pressure compression and laser diffraction measurements allowed swift screening of many batches.Figure optionsDownload full-size imageDownload as PowerPoint slide