Evaluation of the effect of granule size of raw tableting materials on critical quality attributes of tablets during the continuous tablet manufacturing process using near-infrared spectroscopy.

The effects of granule size of raw materials on tablet hardness (TH) and weight (TW) in the continuous tablet manufacturing process (CTMP) were investigated using near-infrared spectroscopy (NIRS). Granule materials of different sizes were prepared by extrusion granulation from a standard granule formula powder containing lactose/starch and 4.5% acetaminophen. Large-, small-, and medium-sized granules were sequentially filled in a hopper, and tablets were produced continuously using a single-shot tableting machine. After arranging approximately 500 tablets in order, the tablets were subjected to NIRS. A total of 450 NIRS datasets were divided into three groups of 150 each (calibration, validation 1, and validation 2 datasets). The best fitted calibration models for predicting TH and TW were obtained, with sufficient accuracy, based on NIRS using the partial least squares regression, and comprised both physical and chemical information. The regression and loading vectors of the calibration models suggested that the models used to predict TH and TW involve physical information based on geometrical factors of the tablet and chemical information related to binder-related intermolecular interactions. The changes in the predicted value profiles of TH and TW using NIRS reflected the changes in the measured values depending on the raw granule size during CTMP.