Bioequivalence and Bioavailability Forum

Dear Ram!

❝ we have a study with William's design now with only 8 subjects (as every guidelines say BE study should be done with minimum of 12 subjects, IS IT ETHICAL to do study with 8 subjects) with 3 test formulations (T1, T2, T3) and one reference.

IMHO simply not. The search feature of forum will give you some ideas why small sample sizes for pilot studies are not giving us reliable estimates of both the PE and CV_intra. You may go with the minimum sample size of 12 if you repeat the 4 sequences 3 times; but I would recommend this only if you already know that you are dealing with a drug of low variability (let's say 10% or less).

❝ I don't know why sponsor want to do study with so few subjects ?

An unfortunate combination of costiveness and lack of knowledge?
If you are working in a CRO, you will have to swallow the bitter pill and start some education...

❝ pre-pivotal, to know which of his test fomulation is having good bioavailability in comparision to reference.

The smaller the sample size the less reliable are the estimates…

❝ Is it possible to do statistical analysis with pk samples of just 4 subjects as in this situation (2T1&2R, 2T2&2R, 2T3&2R).

If you really adopt this design, you will have 8 subjects (not 4). You have 4 sequences, each of administered to 2 subjects. Another reason not to perform such a small sized study are potential drop outs. Since you have 4 periods, it is much more likely that you will have to deal with a heavily inbalanced evaluation…