NPavan
☆    

India,
2009-06-04 14:29
(6224 d 00:45 ago)

Posting: # 3819
Views: 5,107
 

 Missing Samples [General Sta­tis­tics]

Dear all,

We conducted a study with 23 sampling time points (upto 72 hrs). In that for reference product 23rd sampling time point, the concentration is showing missing.
Now for comparing test and reference products, AUC for test product is AUC0_72 and reference product is AUC0_36.

Now my question is, How to compare the test and reference products with respect to AUCt.

Thanks in advance.

Regards,
Pavan
Nirali
★    

India,
2009-06-17 09:14
(6211 d 06:00 ago)

@ NPavan
Posting: # 3862
Views: 4,247
 

 Missing Samples

Dear Npavan,

pl. specify the regulatory of the study.

As in my view, we shall follow below steps:

1) calculate AUC %extrapolation for the specific subject (who has missing sample at ambulatory time point) for reference product to check absorbtion process is complete or not.

2) if AUC %extrapolation is <20%, it assures that absorbtion process is complete. Thus, we can consider that subject for PK and statistical analysis and compare profile of test and referece product.

3) if AUC %extrapolation is >20%, it says absorbtion process is incomplete. Therefore, subject required to exclude from the PK and statistical analysis of the study since reference profile is not complete.


Dear Friends,
pls correct this if anything is lacking here.


Thanks & Regards,
NIRALI
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2009-06-17 13:56
(6211 d 01:18 ago)

@ Nirali
Posting: # 3865
Views: 4,127
 

 Bias?!

Dear Nirali!

❝ 2) if AUC %extrapolation is <20%, it assures that absorbtion process is

❝ complete. Thus, we can consider that subject for PK and statistical

❝ analysis and compare profile of test and referece product.


See the first plot in this post. The absorption phase (red line) is almost complete after 2 hours. Imagine a situation where both formulations give identical profiles (FT=FR; blue line) and one sample of T is missing at 6 hours. Both AUCt will be >80% of AUCinf (T: AUC5 41.5, 83.1% of AUCinf; R: AUC6 44.3, 88.7%). Following your suggestion (absorption complete, comparing AUCt) would lead to a biased result of F=93.7% (see also this post).
  • Comparison of AUCinf is unbiased - if a reliable estimate of t½ for both formulations is possible (n>3, first time point in estimation >2-4times tmax).
  • Comparison of AUCts, where tlast of formulations is different is biased.
  • Truncation of both AUCs at the smaller tlast is unbiased.
  • Imputation of an estimated Clast is unbiased (see also this post).
The regulatory acceptance of the last two options is unclear. IMHO if stated in the protocol or SAP no big deal (at least I am using the last method for many years without any problems).

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
UA Flag
Activity
 Admin contact
23,654 posts in 4,992 threads, 1,570 registered users;
178 visitors (0 registered, 178 guests [including 26 identified bots]).
Forum time: 15:15 CEST (Europe/Vienna)

Actually, science starts to become interesting
only where it ends.    Justus von Liebig

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5