Bioequivalence and Bioavailability Forum

Dear Jaime!

❝ We found in one of the study subjects, drug levels above LLOQ at the pre-dose sample of the first study period.

❝ This came out on both analytes under study (ethinyl estradiol and gestodene), somehow confirming this was not an artifact.

❝

❝ Since levels were below 5% of the Cmax, our first approach was to include this subject in the analysis anyway.

• Noncompliance of the subject with an exclusion criterion (=intake of the respective combination contraceptive)? Ask the subject again.
  
• If the value was <5 % of C_max, but reasonably close to the LLOQ, the result may be pure chance.

❝ It was also noticed that the statistical analysis shows a significant sequence effect (p ~= 0,048)

Is ANVISA still asking for that? To be honest I have given up to fix the many broken links in the Guidance page. Do you have a working URL?

❝ Sequence effect does not show up taking this subject out of the analysis. I think this is just a coincidence since the pk profile for this subject is almost coincident for both ref and test drugs. (Cmax 325.46 vs 315.95 pg/mL and AUC 0-t 2859 vs 3017 [pg x hr]/mL, for ethinyl, for example).

❝

❝ I'm wondering what would be the "best" approach to report the final results - including or excluding this volunteer - since our regulators usually "don't like" volunteer exclusions.

What have you stated in the protocol?

❝ By the way, BE results are the same (bioequivalent) including or excluding this volunteer from the BE analysis.

Lucky you! I would submit whatever you have stated in the protocol as primary analysis and show the data after exclusion as a sensitivity analysis (supportive).