Bioequivalence and Bioavailability Forum

Dear Thirupathi!

❝ We have to plan for withdrawal of ambulatory sample with the window period of plus 30 min to schedule time with NO Protocol deviation (e.g. for 08:00 clock amb. sample, if the sample withdrawn at 08:29/30 we will consider 08:00 clock sample).

First I agree with Ohlbe. Experienced CROs define ‘time allowance windows’ (based on the PK of the drug: narrower in the earlier parts of the profile and wider in the later parts) where deviations don’t have to be commented (!) in the CRF and/or EDCS. However, all deviations are recorded. Some ideas behind:

• Large deviations are unlikely in the early parts where mainly a venflon is used.
  
• Deviations are more likely in the later part; clotting of tubes may occur, or repetitive venipuncture may be applied.
  
• The window for recording a protocol deviation should be set realistically! It is by far more important to record the exact time point, than to write down a detailed explanation. If the personnel introduces further delays in the clinical performance just recording reasons for delays, such a practice is counterproductive.

❝ Is this samples are impact on the value of AUCt/other?

Sure. No good idea. Let’s see an example (i.v. administration, one-compartment open model): C = 100e^{–ln(2)/8 × t}, theoretical AUCs: AUC_∞ = 100/[ln(2)/8] = 1154.2, AUC_t = AUC_∞ – AUC_∞ × e^{–ln(2)/8 × t} (48 h: 1136.1, 48.5 h: 1136.9).
dataset           1.                  2.                  3.
time            s i m u l a t e d   c o n c e n t r a t i o n
 0              100                 100                 100
 2               84.0896             84.0896             84.0896
 4               70.7107             70.7107             70.7107
 8               50                  50                  50
16               25                  25                  25
24               12.5                12.5                12.5
48                1.5625                                  1.4963
48.5                                  1.4963
t½    (%bias)     8.0000 (±0.0)       8.0000 (±0.0)       7.8676 (-1.7)
lin-trapezoidal rule
AUCt  (%bias)  1199.1 (+5.5)       1201.8 (+5.7)       1198.3 (+5.5)
AUC∞  (%bias)  1217.1 (+5.5)       1219.0 (+5.6)       1215.3 (+5.3)
log-trapezoidal rule
AUCt  (%bias)  1136.1 (±0.0)       1136.9 (±0.0)       1134.3 (-0.2)
AUC∞  (%bias)  1154.2 (±0.0)       1154.2 (±0.0)       1151.3 (-0.2)

The log-trapezoidal (or lin-up/log-down for extravascular data) rule is the better choice.

1. The original dataset with no deviations. t_1/2 from the last three values is estimated with 8 hours (unbiased); correct AUC_t and AUC_∞.
   
2. A delay of 30 minutes at 48 hours and the actual time point is used: Unbiased estimate of t_1/2; correct AUCs.
   
3. Delayed sampling treated as scheduled (48.5 → 48): Biased estimate of t_1/2 with 7.87 hours; underestimated AUCs.