Bioequivalence and Bioavailability Forum

Hi Mikalai,

» What would be a sample size for a BE study when the CV of the drug is 30% but bioequivalence should be in range 90_110%?

Unless the drug belongs to the last generation anticoagulants (dabigatran, rivaroxaban) that's a contradiction. NTIDs show low variability. Otherwise, there would have been serious problems in phase III: toxicity, lacking efficacy – no approval – end of story.
BTW, the acceptance limits for NTIDs in the EEA are 90.00–111.11%.

» Our assumption is around 400 subject if the design is standard, and 200 if the design is replicative

library(PowerTOST)
sampleN.TOST(CV = 0.3, theta1 = 0.9, design = "2x2x2")
sampleN.TOST(CV = 0.3, theta1 = 0.9, design = "2x2x4")