Near infrared spectroscopy for blend uniformity monitoring: An innovative qualitative application based on the coefficient of determination

Abstract

Blending process is a critical unit operation in the pharmaceutical industry during the solid dosage form production. Near infrared (NIR) spectroscopy is a powerful analytical tool to assess the blend homogeneity in real-time. In this paper, a new methodology for blending process monitoring and for end point confirmation is proposed. Quantitative procedure validation and maintenance of NIR procedures are time-consuming activities that can prevent the adoption of PAT tools in the pharmaceutical industry. Clearly, there is a need in the industry for simpler and more intuitive qualitative blend monitoring analytical procedure that are easy to build, validate and maintain. The method introduced herein consists of tracking the trend of the Coefficient of Determination (CD) between a mean reference spectrum from a homogeneous batch and the NIR spectra that are recorded during the blending operation. Four formulations of commercial products were selected from different scales–including low dosage solid form-to show the usefulness of the method. In addition, this analytical procedure is tested with data from two different types of spectrometers (diode array instruments). Method calibration was performed with five batches (representing expected process variability) for each product: one for the computation of the homogeneous batch target spectrum and four to compute the limit of the CD values related to anticipated and acceptable homogeneity. Method validation was performed with homogeneous batches and with challenge spectra for assessing the specificity of the method. Real-world examples ( e.g. technical, validation batches and clinical batches) were presented in order to demonstrate that this method is able to detect inhomogeneous batches. The new qualitative method presented in this paper is useful for determination of the blending endpoint, in assessing the blend uniformity in real-time and in increasing process understanding during early development and troubleshooting.