Chemoinformetrical evaluation of granule and tablet properties of pharmaceutical preparations by near-infrared spectroscopy Makoto Otsuka * Research Institute of Pharmaceutical Sciences, Musashino University, Shinmachi 1-1-20, Nishi-Tokyo, 202-8585, Japan Received 1 December 2004; received in revised form 1 February 2005; accepted 4 April 2005 Available online 6 October 2005 Abstract The purpose of this research was to apply near infrared spectrometry (NIR) with chemometrics to predict the change of pharmaceutical properties of antipyrine granules during granulation by regulation of the added water amount. Methods: The various kinds of granules (mean particle size, 70 – 750 Am) were obtained from the powder mixture (1 g of antipyrine, 6 g of hydroxypropylcellulose, 140 g of lactose and 60 g of potato starch) by regulation of the added water amount (11 –19 wt/wt.%) in a high-speed mixer. The granules were characterized as the mean particle size, angle of repose, compressibility, and tablet porosity and tablet hardness as parameters of pharmaceutical properties. To predict the pharmaceutical properties, NIR spectra of the granules were measured, and analyzed by principal component regression analysis (PCR). The mean particle size of the granules increased from 81 to 650 Am in diameter with an increase in the amount of water, and it is possible to make larger spherical granules with narrow particle size distribution using a high-speed mixer. The pharmaceutical properties, such as angle of repose, compressibility and porosity of the tablets decreased with an increase of added water, but the tablet hardness increased. The calibration models to evaluate particle size, angle of repose and tablet porosity and hardness were established by using PCR based on NIR spectra of granules, respectively. The correlation efficient constants of calibration curves for prediction of mean particle size, angle of repose, tablet porosity and tablet hardness were to be 0.9109, 0.8912, 0.7437 and 0.8064, respectively. It is possible that the pharmaceutical properties of the granule, such as mean particle size, angle of repose, tablet porosity and tablet hardness could be predicted by NIR-chemometric method. D 2005 Elsevier B.V. All rights reserved. Keywords: Agitating granulation; Physical properties; Near infrared spectroscopy; Chemometrics; Principle component regression analysis 1. Introduction Various kinds of formulations, techniques and equipment have been developed to obtain high quality granular materials [1–11], since high quality granules have the number of potential advantages to the pharmaceutical industry in the production of beads or granules as both finished and intermediate products. However, physical properties of the granules, such as granule size, hardness and porosity, etc. were affected by their granulation formulations and methods, and the properties influence reproducibility of the preparations by affecting granular flowability, tablet compressibility, disinte- gration time, and dissolution rate. Therefore, in order to ensure the manufacturing of safe and efficacious pharmaceutical products, production process validation is required in order to meet regulatory requirements. Consequently, the near infrared (NIR) spectroscopy is fast becoming an important technique for analysis use in validation of pharmaceuticals in the industry, since the NIR spectra were measured directly on the surface of non-destructed samples without any pretreatment [12]. Addi- tionally, chemometrics provides an ideal method of extracting quantitative information from samples through NIR spectros- copy [13] spectra of multi-component samples. A number of chemoinfometric and statistical techniques are employed in NIR quantitative and qualitative analysis because these approaches have been proven to be successful in extracting the required information from unprocessed NIR spectra. Calibration methods such as multiple linear regression (MLR), principal component analysis/principal component regression (PCA/PCR) and partial least squares regression (PLS) are commonly used [13]. Chemoinfometrical NIR spectroscopic methods were reported to determine tablet 0169-7439/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.chemolab.2005.04.015 * Tel./fax: +81 424 68 8658. E-mail address: motsuka@musashino-u.ac.jp. Chemometrics and Intelligent Laboratory Systems 82 (2006) 109 – 114 www.elsevier.com/locate/chemolab