Cartas de controle multivariadas para o monitoramento simultâneo do teor de isoniazida e rifampicina em uma formulação farmacêutica empregando a espectroscopia no infravermelho próximo

Abstract

Statistical process control is a strategy that has been increasingly used for online monitoring of industrial processes. It is the main tool used for control charts, which can be univariate or multivariate, according to the number of variables involved in monitoring. Basically, control charts are statistical graphs that have a tolerance range bounded by upper and lower lines, calculated based on the standard deviation of the samples in control conditions. This paper presents the application of multivariate control charts based on two philosophies: the NAS (Net Analyte Signal) vector and the principal component analysis (Principal Component Analysis - PCA), using Near Infrared Spectroscopy for the simultaneous monitoring of level two drugs; isoniazid and rifampicin in pharmaceutical formulations produced by NUPLAM (Center for Research on Food and Drug Administration) at UFRN. The charts were constructed using samples from an experimental design and the NUPLAM production line. Control Charts via PCA scores for the simultaneous monitoring of both APIs showed accuracy rates above 92% in the classification of samples for calibration and validation, and 100% accuracy in prediction. The charts via NAS to isoniazid had 100% accuracy in classifying samples of calibration and prediction. The NAS charts to rifampicin achieved hit rates above 90.0% in the calibration and prediction. The results obtained demonstrate the applicability of multivariate control charts for the control of pharmaceutical processes, thereby reducing cost and time of analysis, generating less waste and thus optimizing the quality control process.