Bioequivalence and Bioavailability Forum

Hi HS,

very good post and very well illustrated C_last isn't easy to work with.
Here is a perspective from the world of pharmacodynamics for inhaled products.
In some cases therapeutic equivalence is evaluated by giving patients a puff of the medicines and then measuring and comparing the lung function (often an endpoint called FEV1) over time. In certain cases clients of mine have been suggested to include FEV1 at the last sampling point in the comparison as one of the primaries.
This is every bit as bad as the story with C_last. I can certainly see why a regulator might want 'something like that' but I think we are getting away from proper science to semi-thought-through-solutions that sound intuitively correct but by closer scrutiny aren't really practical, feasible and valid for the reasons you already gave. For inhaled products FEV1_max may have an intraCV of around 20% on a good day whereas the intraCV for FEV1_last can exceed 80% when the same product (same brand, same batch) is dosed. That's pretty showstopping.
Solution: Treat guidelines as variable constants and confront regulators in an sc. advice application - is there any acceptable design that does not require evaluation of C_last (or FEV1_last)? In one case I know of they caved in and accepted an alternative. Just unfortunate that this adds 3-4 months to the project plan.