(T1530-11-60) Evaluation of Near Infrared (NIR) for the Quantitation of Low Dose Drugs

Purpose: The demand for advanced manufacturing techniques such as continuous manufacturing is increasing in the pharmaceutical industry. Near infrared (NIR)-based multivariate calibration models (e.g., chemometric models) are emerging as a popular application for the detection of active pharmaceutical ingredient (API) content and have the potential to support real-time release testing. With increasing utilization of NIR model surrogates in place of traditional offline low dose drug content measurements (e.g., HPLC, UPLC, UV, or titrimetry), the new models could pose high-risk when the suitability of the quantitative NIR model of the low dose drug is not properly justified. This project evaluates the capability of NIR spectroscopy as a process analytical technology tool to accurately monitor API content in 3% w/w acetaminophen (APAP) powder blends.

Methods: APAP powder blends with drug loadings spanning 70–130% of the target 3% w/w drug loading were prepared for development and validation of a quantitative model. NIR spectra were collected via a Bruker MATRIX-F emission spectrometer with contactless measurements. Both HPLC and UPLC methods were developed to acquire reference drug concentration data. Savizky-Golay derivative, standard normal variate, and orthogonal signal correction techniques were performed on the NIR spectra. After data preprocessing, Unscrambler 11 software was used for spectral analysis, as well as NIR model development and validation.

Results: A partial least squares (PLS) regression model was successfully developed based on offline NIR spectral measurements of powder blends with a target APAP content of 3.0% w/w. Internal validation of the PLS model yielded a Pearson R² of 0.9737, root mean square error of prediction of 0.1037% APAP (w/w), and bias of -0.0257% APAP (w/w). External validation of the PLS model yielded a Pearson R² of 0.9800, root mean square error of prediction of 0.09478% APAP (w/w), and bias of -0.0408% APAP (w/w). The PLS model was validated for specificity, accuracy, precision, linearity, limits of detection and quantitation, as well as robustness, all of which demonstrated sufficient discrimination between different powder blends around the target of 3.0% w/w APAP.

Conclusion: A quantitative NIR chemometric model for measuring 3.0% w/w APAP blends was successfully developed and validated (Figure 1). Results from this work will provide general evaluation profiles that are compliant with pharmaceutical regulatory requirements to assess the suitability of NIR-related quantitative chemometric models to measure API content.

Acknowledgements: Disclaimer - This abstract reflects the views of the authors and should not be construed to represent FDA’s views or policies.


Figure 1. Workflow diagram describing the development and validation of quantitative NIR models for low dose API powder blends.