Use of multivariate statistical analysis to evaluate experimental results for certification of two pharmaceutical reference materials

This paper demonstrates the use of the multivariate analysis for the quick and easy evaluation of the experimental results from the homogeneity test of two new certified reference materials (CRM) of active pharmaceutical ingredients (API): metronidazole and captopril. The principal component analysis (PCA) and the hierarchical cluster analysis (HCA) indicated that some results from the homogeneity test were statistically different when the concentrations of all API impurities were considered simultaneously. Through the use of these statistical tools, it was possible to reduce the standard uncertainty due to between-bottle (in)homogeneity (u _bb) and consequently the combined standard uncertainty of the certified reference materials (u _CRM) with 95% confidence level.