A NOVEL MULTIVARIATE APPROACH FOR ESTIMATING THE DEGREE OF SIMILARITYIN BIOAVAILABILITY BETWEEN 2 PHARMACEUTICAL PRODUCTS

Statistical assessment of bioavailability and bioequivalence of drug p roducts is generally carried out with a univariate analysis by indepen dently comparing each relevant parameter [such as, area under the drug concentration curve (AUG) and peak drug concentration (C-max)] of the test and reference products. The assumption is made that AUC and C-ma x are independent of each other. In reality, AUG, C-max, and other pha rmacokinetic parameters are dependent on each other. Therefore, a mult ivariate approach is theoretically superior. In this report, a novel m ultivariate approach to determine the bioavailability and bioequivalen ce of pharmaceutical products is described. The method determines the bivariate confidence ellipse (CE) between C-max and AUC and the trivar iate confidence ellipsoid (CED) among C-max AUC, and either T-max or M RT for test and reference products. The similarity between the test an d reference products can be assessed by the fraction of overlapping ar eas of CE and volumes of CED. The method and its comparison with conve ntional univariate analysis are demonstrated with data obtained from a bioequivalence study of ketoconazole tablets.