Analytical Performance and Greenness Evaluation of Five Multi-Level Design Models Utilized for Impurity Profiling of Favipiravir, a Promising COVID-19 Antiviral Drug

Abstract

In 2018, the discovery of carcinogenic nitrosamine process related impurities (PRIs) in a group of widely used drugs led to the recall and complete withdrawal of several medications that were consumed for a long time, unaware of the presence of these genotoxic PRIs. Since then, PRIs that arise during the manufacturing process of the active pharmaceutical ingredients (APIs), together with their degradation impurities, have gained the attention of analytical chemistry researchers. In 2020, favipiravir (FVR) was found to have an effective antiviral activity against the SARS-COVID-19 virus. Therefore, it was included in the COVID-19 treatment protocols and was consequently globally manufactured at large-scales during the pandemic. There is information indigence about FVR impurity profiling, and until now, no method has been reported for the simultaneous determination of FVR together with its PRIs. In this study, five advanced multi-level design models were developed and validated for the simultaneous determination of FVR and two PRIs, namely; (6-chloro-3-hydroxypyrazine-2-carboxamide) and (3,6-dichloro-pyrazine-2-carbonitrile). The five developed models were classical least square (CLS), principal component regression (PCR), partial least squares (PLS), genetic algorithm-partial least squares (GA-PLS), and artificial neural networks (ANN). Five concentration levels of each compound, chosen according to the linearity range of the target analytes, were used to construct a five-level, three-factor chemometric design, giving rise to twenty-five mixtures. The models resolved the strong spectral overlap in the UV-spectra of the FVR and its PRIs. The PCR and PLS models exhibited the best performances, while PLS proved the highest sensitivity relative to the other models.