Bioequivalence and Bioavailability Forum

Hi BJ!

❝ According to the reference, ISCV in AUC and Cmax values was considerable, with the CV ranging from 46.79% to 61.89% and 35.81% to 68.18%, respectively

❝ (Clinical Therapeutics/Volume 31, Number 5, 2009)

I don’t have the reference, but in the abstract (arithmetic?) means ±SD of the doses are given. If you calculate CV=100·SD/x these values are ~CV_total (see this post). Since generally CV_inter > CV_intra I would not worry. Furthermore, n=10 with three treatments have few degrees of freedom. Even if the CI is given in the paper (please check; you can calculate CV_intra from it) the CV would be rather uncertain…

I would trust your own data the most. If you plan another study allow for a safety margin since the PE of C_max is not that nice. The upper confidence limit of the CV of C_max is 26.81%. With an expected T/R-ratio of 0.86 you would need 166 (!) subjects in a 2×2 cross-over to obtain 80% power… Since you had significant differences of AUC_t and C_max (CIs did not contain 100%) consider reformulating.