Bioequivalence and Bioavailability Forum

Hi Mittyri,

I agree with what ElMaestro and Angus wrote. Essentially it is very simple: Assuming that you sample long enough – what you “see” in the terminal’ phase is always driven (but not identical to) the slowest pro­cess. For IR it is likely elimination (sometimes also redistribution from a deep compartment) and for CR it may be absorption.
Think about a simple one-compartment model, po administration. In a mathematical sense the two rate constants are not identifiable. You only assume that k_a > k_el. You can only be sure once you have ad­mi­nistered the drug iv. Have a look at this presentation, slide 19. I simulated a true elimination of eight hours (iv, yellow line). Hey, it’s easy to create new formulations on the desk! I started with an IR (ab­sorption t_½ one hour) and slowed it down gradually. Once we have k_a = k_el we enter flip-flop PK. Not so evident in the linear plot but the lin/log-plot is interesting. As long as k_a > k_el we get an un­biased estimate of the elimination (the lines are only shifted to the right; same slope). In the flip-flop any beyond absorption drives the process.
Another nice observation: If we chance the absorption and keep elimination constant, the curves inter­sect the iv-curve at their respective t_max/C_max. Of course, AUC_∞ is identical.
Now see the next slide. Until we reach flip-flop PK, AUC_∞ is practically unbiased. Some bias creeps in afterwards. Truncated AUC will not work any more – that’s why the MR-GL tells us for CR forms AUC₇₂ is not acceptable as the only metric for extent of BA (like for IR and DR).
But – slide 22: If you have k_a ≪ k_el, the (still ongoing) absorption might be interrupted if the for­mu­la­tion leaves the GIT. Shit happens sometimes with matrix forms and osmotic pumps…

❝ Which type of halflife should be used for washout calculation?

The predominant (i.e., relevant) one. I don’t know your formulation.