Bioequivalence and Bioavailability Forum

Hi Lucas,

❝ I don't even know if that concept is known by you guys, but a triangulation point is a null point (concentration <LLOQ) between two non-null points.

Never heard this term. I have some sympathies with ANVISA’s approach.

❝ […] EMA and FDA regulations, if anywhere else this is a standard procedure.

Though they don’t like it, the FDA accepts exclusion of data points based on PK reasons. Makes sense to me. Even if you confirmed a concentration to be <LLOQ such a value might be physiologically impossible – especially if the value is embedded by two high concentrations. You should unambigously state in an SOP (or better in the protocol) how to deal with this situations. You may opt for the lin-up/log-down trapezoidal method to calculate the AUC. If the value is not close to t_max another option is to exclude the subject from the comparison of AUCs and keep him/her for the comparison of C_max.

If EMA is concerned BLQs cannot be removed.

❝ When we exclude points like those with the intention of eliminating a bias, we may be in fact creating another (big) one and we'll never know for sure.

In science we never can be sure. That’s the field of religion(s).

❝ How would we know if the point that was excluded was wrong and not the ones surrounding it?

I don’t like the common practice of repeating only the doubtful value. We always re-analyze the two neighbouring samples.

❝ That exclusion can cause a massive difference in AUC, changing even the study's conclusion in some cases.

Sure. Some ideas above. As long as you follow predefined procedures and the study has not be unblinded yet, I don’t see a problem.