Helmut
★★★
avatar
Homepage
Vienna, Austria,
2013-02-12 20:12
(4061 d 17:09 ago)

Posting: # 10005
Views: 7,188
 

 Puzzle: Estimation of half life [Surveys]

Dear all!

Another one from a consultant’s diary. For a similar story see this thread.
Another (but also large) CRO, another continent. PK report from 2013. 4-period replicate; the double peak is specific for the formulation.

In four cases the last concentration was increasing. The CRO followed EMA’s GLs and could not reanalyse the samples (PK reason alone not sufficient). Obviously they tried to ‘save’ the profiles by including more data points…

Below the most extreme case. Two samples (at 10 and 12 h) were BLQ; the last 5.47 ng/mL (~2.7× LLOQ).
Left what the CRO did (first data point in the estimation of λz was at tmax); right what I suggest:

[image][image]

My half life of 2.074 h is consistent with what I see in the other periods of the same subject (2.116, 1.999, 2.185 h). Clearly the CRO’s procedure is crap (but backed by SOPs the size of a telephone directory).

Since nothing about excluding data points was stated in the protocol I think the sponsor has two options:
  • Follow my approach and keep the profile for the AUC comparison.
    Justification: Half life similar to the ones in the other three periods, two values between 8 and 16 hours were BQL – which agrees with what I predict from my λz.
  • Drop the profile from the AUC comparison, but keep Cmax (higher variable anyway and scaling intended in the protocol).
How is your experience? Regulatory acceptance?

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
ElMaestro
★★★

Denmark,
2013-02-12 20:42
(4061 d 16:40 ago)

@ Helmut
Posting: # 10006
Views: 5,564
 

 Puzzle: Estimation of half life

Hi Helmut,

this happens occasionally with inhalation products and leaves everybody puzzled. If several subjects have such a profile one initial idea is to audit.
If nothing is found, and if SOPs are silent in relation to this situation, then I would say you can calculate AUCt but not kel or AUCinf for such subjects.
Hence I believe the subjects are not lost to evaluation of the primary endpoints Cmax and AUCt. You will of course get questions about the typical max. 20% extrapolation requirement which is impossible to answer in a meaningful manner. If you do business with some authority that requires AUCinf as primary then you have "an issue" and I have no solution which is universally useful.

Pass or fail!
ElMaestro
d_labes
★★★

Berlin, Germany,
2013-02-12 20:46
(4061 d 16:36 ago)

@ Helmut
Posting: # 10007
Views: 5,506
 

 Estimation of half life - LLPout rule

Dear Helmut!

This story reminds me on the "leave last point out" rule I (we) have as a standard. I.e. leave last point(s) out from the terminal decreasing part for estimating the half life if they are increasing.
Remember this post?

After answering the deficiency letter the described way I never heard any complaints. Thus assuming that the "leave last point out" rule was accepted by the regulatory body EMA (in this special case?).

Regards,

Detlew
jag009
★★★

NJ,
2013-02-12 21:50
(4061 d 15:32 ago)

@ Helmut
Posting: # 10008
Views: 5,537
 

 Puzzle: Estimation of half life

Hi Helmut,

The CRO that I outsourced before would consider the t1/2 to be "N.C" or Not calculated if the last pt (or 2 pts) showed an increase in concentration. I ran like 200 studies with them and about 100–150 studies were for FDA ANDA submission. The agency never came back with questions. Granted that this issue occurred in 2–5 subjects (with n=24+).

John
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2013-02-13 13:41
(4060 d 23:40 ago)

@ jag009
Posting: # 10012
Views: 5,478
 

 Excluding increasing concentration(s)

Dear ElMaestro, Detlew, and John,

THX for sharing your experiences & thoughts!

@ElMaestro: I would not audit the study. We can expect such values due to the variability close to the LLOQ (see this post). In the good ol’ days of common sense one would have reanalysed the sample in duplicate and likely the original value would have turned out to be an artifact. Only for EMA & TGA according to the current GLs such a procedure is taboo.

@Detlew: Like you I have this rule in place for ages. As I said there I never got a question. Hasn’t changed ever since. Good to know that your response was accepted.

@John: I was suspecting something along those lines from NDAs/ANDAs accessible under FOI. Many studies reported a smaller sample size for AUC than for AUCt.

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
22,957 posts in 4,819 threads, 1,639 registered users;
79 visitors (0 registered, 79 guests [including 3 identified bots]).
Forum time: 13:22 CET (Europe/Vienna)

Nothing shows a lack of mathematical education more
than an overly precise calculation.    Carl Friedrich Gauß

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