Bioequivalence and Bioavailability Forum

jag009
★★★

NJ,
2013-04-25 19:45
(4397 d 21:56 ago)

Posting: # 10494
Views: 5,042

Question about subject with very few timepoints [Study Assessment]

Hi all,

Question... If I have a subject with only 2 out of 22 quantifiable plasma concentration(total 22 timepoints) while the rest are BLQs, is there any justification to drop him out of the analysis or run two sets of analysis (1 with all subjects, 1 without this subject)? His two quantifiable timepoints are adjacent to each other (lets say 1 and 1.5). He did take the medication.

All others subjects have quantifiable concentrations at most if not all timepoints. Concentration values show high degree of variability among subjects.

Thanks
John

Edit: Category changed. [Helmut]

Helmut
★★★

Vienna, Austria,
2013-04-25 20:16
(4397 d 21:25 ago)

@ jag009
Posting: # 10495
Views: 4,200

Reliable profile

Post reply

Hi John,

some counter-questions first:

PPI?
Test or Reference?
Sample size?
2×2 or replicate?

Since you posted in the Reg/GL-category: I don’t know of any GL specifying what a “reliable profile” is. There was a workshop in Canada a while ago, where TPD claimed two points to be sufficient to calculate AUC and one for C_max (not kiddin’). For EMA you could take the subject out if the AUC is <5% of the rest of the subjects (if prespecified in the protocol). You can perform both evaluations (regulators love sensitivity analyses), but very likely you will fail with the full data set. In the past FDA recommended re-testing of the subject with both test and reference (sample size ≥6 or 20% of subjects, whichever is larger). If the subject show a ‘normal’ profile in the retest you may exclude him/her from the analysis. I’m rather skeptic whether FDA still accepts such a procedure nowadays. :confused:

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

jag009
★★★

NJ,
2013-04-25 22:19
(4397 d 19:21 ago)

@ Helmut
Posting: # 10496
Views: 4,172

Reliable profile

Post reply

Hi Helmut!

❝ • PPI? :-D

Nope.

❝ • Test or Reference?

Reference

❝ • Sample size?

❝ • 2×2 or replicate?

Rep - of the 2 reference arm

His other reference and test arms have more concentration values but still many BLQs. His "meaningful" profiles are lower in magnitude than other subjects and yes there are subjects with comparable "small" values. The questionable part is the reference arm with only 2 datapoints. AUC is defintely smaller than other subject.

From what I understand and being taught (by old schoolers) is never drop a subject... But this one is strange because there are only two points and they are next to each other (I guess, Cmax and post Cmax)...

Thanks
John

ElMaestro
★★★

Denmark,
2013-04-26 00:40
(4397 d 17:01 ago)

@ jag009
Posting: # 10497
Views: 4,151

Reliable profile

Post reply

Hi jag009,

❝ ❝ • PPI? :-D

❝

❝ Nope.

❝

❝ ❝ • Test or Reference?

❝

❝ Reference

I had really hoped you would answer "Nope" to this one as well. Would have left 3000 regular readers of the bebac forum rather baffled.

Unless the protocol or some internal SOP explicitly deals with this situation I don't think the subject can be dropped. I am, however, 100% sure the FDA will be evaluating the dataset without that subject when a PP analysis is presented. If the subject suffers from low values in both periods then I guess (s)he is just either a terrible absorber or has warp speed elimination, and there may be little you can actually do. But in certain fields the FDA allows population enrichment (genotyping, responders only etc) to make sure drug exposures or effects are measured relevantly and meaningfully, and that of course has to do with sensitivity.

In a sense you could try to argue that removal of a poor absorber / ultrametaboliser is not worse than population enrichment by genotyping or responders only. For example ICH E9 (Published in the Federal Register, 16 September 1998) reads:
The primary variable (‘target’ variable, primary endpoint) should be the variable capable of providing the most clinically relevant and convincing evidence directly related to the primary objective of the trial.
...so you could argue that no relevant or convincing info about the comparative product performance is coming from the subject in question.
It is perhaps a long shot :-D

—
Pass or fail!
ElMaestro

Dr_Dan
★★

Germany,
2013-04-26 10:58
(4397 d 06:43 ago)

@ ElMaestro
Posting: # 10498
Views: 4,158

CVintra result of outlier

Post reply

Hi jag009,
as far as I understood the subject completed two reference treatments, one with detectable drug concentrations and one without = no profile. You could argue that the reference product is of such a poor quality that inclusion of the period without profile would on the one side affect the comparison with the test product in a negative way and on the other side would give a wrong impression on the intra-individual variability of the reference product. According to EMA guideline:
“The applicant should justify that the calculated intra-subject variability is a reliable estimate and that it is not the result of outliers.”
Maybe this helps.
Kind regards
Dan

—
Kind regards and have a nice day
Dr_Dan

Question about subject with very few timepoints [Study As­sess­ment]

Reliable profile

Reliable profile

Reliable profile

CVintra result of outlier

Question about subject with very few timepoints [Study Assessment]