Bioequivalence and Bioavailability Forum

Dear rams
Of course you can estimate missing concentrations (e.g. linear interpolation which would not influence the study result) but for which reason? I guess the assessor has a sensitivity analysis in mind in case the missing sample could have represent the Cmax, right?
Kind regards
Dr_Dan

Hi Rams,

❝ Can we estimate Missing concentration for particular subject at particular time point.

Since you posted in the Regulatives / Guidelines-category: AFAIK nothing is mentioned in any of them.

• I would suggest to calculate the AUC by the lin-up/log-down trapezoidal rule. Missing values shouldn’t bias the estimate – contrary to the linear trapezoidal which will give a positive bias in the distribution / elimination phases.
  
• If the missing value is close to the expected C_max, bad luck. Hopefully you have a tight sampling schedule around.
  
• If your last sample is missing, who cares? GLs ask for a comparison of AUC_0–t. You might end up with some apples-and-oranges statistics (e.g., comparing AUC_0-48 of the test with AUC_0–72 of the refe­rence) but the same would be true if you face a BLQ only after one formulation. Shit happens.
  
• If you have to deal with a comparison of C_τ in steady state (EMA) I always use an estimate. Of course, the procedure has to be unambiguously stated in the protocol.