Evaluation of Transmission and Reflection Modalities for Measuring Content Uniformity of Pharmaceutical Tablets with Near-Infrared Spectroscopy

This paper examines how one may assess spectral changes with instrument configuration (or composition), in combination with the spectral changes in the measurement that are caused by experimental effects, and subsequently select an appropriate measurement modality for tablet content uniformity determination with near-infrared (NIR) spectroscopy. Two NIR spectrometers furnished with three configurations in the sample measurement interface were evaluated. One spectrometer, Bruker MPA (multiple purpose analyzer), was equipped with two measurement modalities, diffuse transmission (DT) and diffuse reflection based on integrating sphere optics (DR/IS). The other spectrometer, Bruker StepOne, was equipped only with diffuse reflection mode based on a fiber-optic probe (DR/FO). The data were collected with each of the configurations for the tablets associated with two dosage strengths differing significantly in diameter and thickness. Spectral diagnosis was performed in terms of sensitivity and selectivity. The signal-to-noise ratio computed for the data collected with the DT and DR/IS spectrometers was approximately an order of magnitude greater than that computed for the DR/FO spectrometer. The net-analyte-signal-based selectivity analysis of NIR spectra associated with the sample tablet and the placebo tablet indicated that both transmission and reflection mode provided similar selectivity when the optimal spectral range was chosen. A partial least squares (PLS) calibration model was developed for each data set. The overall standard error of calibration for each DT and DR/IS measurement was approximately 0.3% in weight for each strength, significantly better than the value of 1.0% in weight produced by the DR/FO measurement. This result was consistent with the sensitivity analysis based on spectral noise characterization. The poor analytical performance of the DR/FO spectrometer was attributed to the small illumination spot size of the reflection probe and thus the sensitivity of the measurements to the tablet engraving. The PLS analysis and spectral diagnostics both showed that transmission and reflection modes based on the Bruker MPA provided similar measurement accuracy for each strength. However, the robustness study clearly revealed that the transmission mode would be more robust than the reflection mode when there is considerable variability in the chemical composition and physical properties of tablets.