Chemometric Models for Quantification of Carbamazepine Anhydrous and Dihydrate Forms in the Formulation

Abstract

Carbamazepine (CBZ) exists in anhydrous and dihydrate forms. These forms differ in their solubility, dissolution rate, and subsequently in their oral bioavailability. The objective of this study is to develop multivariate chemometric models for estimation of the low level of carbamazepine dihydrate (CBZ-DH) in the CBZ formulations containing excipients of the commercial formulation. The selected excipients were mixed in proportions to make sample matrices ranging from 0% to 50% CBZ-DH. Fourier transform infrared (FTIR), near infrared (NIR), and hyperspectral imaging data were mathematically pretreated before the development of partial least square and principal component analysis regression models. The developed partial least squares regression and principal component analysis models demonstrated predictability of CBZ and CBZ-DH by multiple scattering correction and standard normal variate processing methods. Among the spectroscopic techniques used the model performance parameters such as root-mean-square error, standard error, and bias were found to be low for NIR compared to FTIR. The treated data have shown better model fitting than without treatment, which was demonstrated by correlation coefficient of 0.9778, 0.9824, and 0.9852 for FTIR, NIR, and hyperspectral imaging, respectively. Furthermore, the predicted values were found to be very close to the selected low level of independent samples having 5% CBZ-DH in tablet formulation.

Keywords: chemometrics; differential scanning calorimetry; fourier-transform infrared spectroscopy; near-infrared spectroscopy; polymorphism; principle component analysis; thermogravimetric analysis.