Bioequivalence and Bioavailability Forum

¡Hola Armando!

IMHO you were courageous performing this study, since no Two-Stage method for such a design is published (currently only 2×2 cross-over and parallel). Did you perform simulations in order to demonstrate that the patient’s risk is maintained?

❝ […] so that the 18 subjects completed the four periods.

In the randamization you needed a multiple of 4. Just to clarify: There were two drop-outs?

❝ Once finished the first stage, the statistical analysis resulted on different CV intra for each form of administration with and without chamber, (29% and 35% respectively). So, finally 36 and 60 volunteers respectively should be added in in the second stage.

More information, please. Did you completely randomize subjects (4 sequences) or did you “stack” two 2×2 cross-overs? For details see beman’s post.

❝ Would it be necessary to complete the second stage with 60 subjects for both forms of administration? or could we reduce to 36 subjects in the case of chamber?

If you used beman’s stacked 2×2 cross-overs, the second stage could be of different sizes. If you used full randomization, no idea. IMHO, that would require simulations – which should have been performed before submitting the protocol.