Bioequivalence and Bioavailability Forum

Hi John!

❝ This just popped into my head (flashback... Nightmare, whatever you called it)...

For me – flashback. We performed a pilot study of a drug with a nice average [sic] half-life of ~50 hours, sampling up to 216 hours (since truncated AUC wasn’t common at that time), washout 2 weeks. No predose concentrations… The pivotal was a 6×3 in 36 subjects, but:

• only ½ of the dose of the pilot was administered and
  
• the LLOQ of analytical method was improved by a factor of 9 [sic].

Result: We measured pre-dose values (maximum 7.8% of C_max) in 54% of samples of periods 2 & 3 (none in period 1). Since we had a blinded plausibility review in place we were able to come up with a statistical amendment while analytics were still running:

• Primary analysis based on period 1’s data as a parallel design. The sponsor had too much money – so the cross-over was extremely overpowered. The total CV was quite low and we expected still ~75% power. Since no pretesting (sequence or unequal carry-over) was done (Grizzle’s flawed method), the patient’s risk was not compromised.
  
• Exploratory 1: Ignoring the carry-over and performing the comparison as originally planned. Nowadays we would go a similar track: 86% of the subjects with pre-dose values <5% would stay in the analysis.
  
• Exploratory 2: Estimating λ_z not only from data of the respective treatment, but including the pre-dose value of the next period. Subtracting estimated concentrations of the next period from measured ones.

Results:

• Confirmatory: AUC passed, upper limit of C_max failed.
  
• Exploratory 1 & 2: AUC and C_max passed easily.

Lessons learned: Never base the washout on mean-values (large studies not even on x±SD) and think twice if you improve the analytical method too much.

P.S.: The product was approved based on this study in the EU in 2002 and by the FDA in 2003.

❝ Should the lab adjust the concentration profile to remove the residual levels caused by the pre-dose concentration???? ie. using the t1/2 and determine the amount of pre-dose concentration remaining throughout the timepoints and subtract it from the respective concentrations.

See above. I cannot imaging that any regulatory agency would accept this (any more).

❝ I remember one lab doing this and they said it's part of their SOP...

Referring to an SOP is alway the worst (pseudo-)justification I can think of.

“Hey, let’s jump from the cliff!” Lemming’s SOP