Bioequivalence and Bioavailability Forum

Hi BE-proff,

the BE-assessment is always done with the PP-population. Imagine a single dose study where 2 × 50mg (Test) is compared to 1 × 100mg (Reference) and you discover that one subject managed it to spit one of the T-tablets into the toilet. Expected T/R 50%. Doesn’t make sense. Exclude the subject.
Same in steady state. If a subject didn’t take the drug(s) the T/R-ratio will be biased. The bias on the profile day will not be so dramatic as in the single dose example but it still exists. Whatever clever things you will do (e.g., PK-modeling,…) it is only exploratory. No regulator would buy that.
Subject wasn’t fully compliant ⇒ exclusion. Full stop.
However, the case is not lost. The impact of dropouts & excluded subjects on power is overrated. There are good chances that the study will still pass. Prepare for an inspection.