Bioequivalence and Bioavailability Forum

Dear Prakash!

❝ i would like to keep few subjects extra/standbys to account for

❝ withdrawal/dropped out in bioeuquivalence study. could you let me know the

❝ maximum allowable number or percentage of subjects?

Hmm, I would guess some ethics committees to be concerned if your study would be powered to >90% (see also this thread).

❝ is there any specific regulatoyr guideline to outline the details on

❝ inclusion of withdrawal/dropped out in sample size calculation?

Canada 1992 (Section 3.3.d), USA 2001 (Section V.B), Brazil 2006 (Annex, Section 1.j; in Portuguese), WHO 2006 (Section 6.3.2), and many others - you know this post for sure...

Generally you should estimate the expected drop-out rate based on

• site's & study's conditions (facility, inhouse or outpatient,...)
  
• expected adverse reactions (drug/formulation/dosage/regimen-specific)

Some examples:
For a drug causing gastrointestinal disturbances the likelihood of drop-out rates is higher after multiple doses than after a single one. With a given multiple dose regimen an ambulatory setting in the saturation phase will be better than hospitalisation for the entire study because the 'stimulating' effect of watching other subjects vomiting will be lower.
Even weather conditions may be influential - drop-outs rates will be higher during monsoon...
Don't forget discomforting addtional measurements. If you had drop-out rates of 5-10% for a drug in one study and are planning some annoying PD measurements (e.g., bleeding time) in the next one, drop-out rates will be higher.

In any case you should perform some kind of 'pre-study-sensitivity-analysis', which is also recommended by the ICH E9 (Section 3.5, 3rd paragraph).