Near infrared spectroscopy for determination of moisture content in lyophilized formulation

Abstract

A non-invasive near infrared (NIR) spectroscopic method was developed for the quantitative moisture determination in a lyophilized injection formulation. The calibration samples were prepared by exposing lyophilized samples at different temperatures and relative humidity. The samples from different scales and different process parameters were considered for adding robustness to the model. The NIR spectra were collected using a Fourier- transform (FT) NIR with a diffuse reflectance probe and the same samples were further analyzed by the Karl Fisher (KF) method for moisture content. The pre-treated NIR spectra were used for quantitative method development for moisture content. Partial least squares (PLS) regression was used to develop calibrations in the 5600-4950 cm−1 region with calibration coefficient of determination (R2) of 0.96 and root mean square error of calibration (RMSEC) of 0.149. The model was cross-validated internally using the Kernel algorithm with r2 = 0.96 and RMSECV = 0.15. The accuracy of the NIR method against the KF method, precision, and reproducibility were good and the model was robust in predicting different external validation samples. This work allowed NIR as an alternative measurement for moisture analysis as well as facilitate 100% monitoring before packaging and save the cost of sample and time of KF analysis.