Bioequivalence and Bioavailability Forum

My Capt’n!

❝ It sounds like you do not have a procedure that describes the choice of points in the elimination phase? Would be a good idea to have one for future purposes, especially when or if you do a pivotal study.

Remember some old posts, especially that one? My resume:

• All automated procedures are restricted to methods available in the software. AFAIK all com­mercial programs use solely max. R²_adj. Following our discussions here only bear offers five (!) different automated methods: R²_adj (ARS), Akaike’s information criterion (AIC), Two-Times-T_max method (TTT), TTT-ARS, TTT-AIC.
  
• We (Martin & Jack with some moral support from myself) have done simulations comparing the automated methods (our manuscript is at the bottom of the to-do-pile; volunteering co-authors with some knowledge in R are mostly welcome!). As expected especially flip-flop-PK (k_a≤k_el) is nasty. Though we have found some differences between methods in terms of bias for λ_z, the differences diminish when it comes to the calculation of AUC_∞.
  
• There is a rank order where automated methods might fail: 1 compartment < 2+ com­part­ments < MR formulations (especially with “flat” peaks) < pulsed formulations. I work a lot on the latter and don’t use the automated method even as an initial guess…
  
• All (!) papers and textbooks emphasize the importance of inspection of automated fits and corrective measures if deemed necessary (=eye-balling).

In other words: no rule rulez. But: it must be clear, that the selection of time points is not driven by the “desire” to obtain AUC_t ≥80% of AUC_∞. Preferably estimation of λ_z is done blinded for treat­ment – before NCA!
I would not be so radical like Maria Durisova stating at David's list:

AUC_t to AUC_inf ratio is meaningless. The reason for this is that time "t" is selected by a phar­ma­cokinetician, consequently AUC_t to AUC_inf depends on the pharmacokinetician choice.

—
I could be wrong, but…