4 replies to this topic

[Angus McLean]

Usually I find that for a Phase 1 study e.g. a bioequivalence type study that the blood sampling times actually obtained in the clinical study are nearly always EXACTLY the same as the nominal times in the clinical protocol. The clinics are very efficient. I have been able to use the nominal times for the pharmacokinetic data analysis. I specify limits for the time of blood samples in the clinical protocol e.g. for a modified release (slow) formulation plus or minus 15 minutes deviation from nominal and for samples up to 24 hours and thereafter plus or minus 30 minutes is allowed. If the time of the blood sample is outside of these limits for nominal then I would use the actual time of that one sample instead of the nominal time.

 

If you consider the pipette used to aliquot a blood sample is accurate to plus or minus 2% and think of the other error sources then I feel time differences inside the stipulated time windows in AUC calculation as stipulated are not significant.

 

The other option is to always use actual times, but this creates a problem since the charts and tabulations in a report use the nominal times. So there is a difference and it serves to confuse.

 

Any thoughts on this topic?

 

Angus

[Helmut Schütz]

I have been able to use the nominal times for the pharmacokinetic data analysis. I specify limits for the time of blood samples in the clinical protocol e.g. for a modified release (slow) formulation plus or minus 15 minutes deviation from nominal and for samples up to 24 hours and thereafter plus or minus 30 minutes is allowed. If the time of the blood sample is outside of these limits for nominal then I would use the actual time of that one sample instead of the nominal time.

That’s not a good idea. Too lazy to check FDA’s requirements, but EMA states in their BE GL: “Actual time of sampling should be used in the estimation of the pharmacokinetic parameters.”
The purpose of “Time allowance windows” is to relax the requirement to document the reason of a deviation in the CRF – not to use the nominal time instead.

For a recent discussion see also this thread.

If you consider the pipette used to aliquot a blood sample is accurate to plus or minus 2% and think of the other error sources then I feel time differences inside the stipulated time windows in AUC calculation as stipulated are not significant.

Nope, that’s another pot of tea. If we f.i. estimate λ_z by log-linear regression one of the main assumptions is that the independent variable (time) is known free of error. All the error resides in the dependent one (concentration).

The other option is to always use actual times, but this creates a problem since the charts and tabulations in a report use the nominal times. So there is a difference and it serves to confuse.

Why? For every datum have two entries; the nominal and actual times. For e.g. mean plots run descriptive statistics sorted by the nominal time. For subject’s plots use the actual times.
In Phoenix it is easy to generate a table containing the actual times / measured concentrations sorted by the treatment/subject/nominal times. Looks like:

scheduled sampling time
                 0     0.3333       0.6667       1.0000        1.2500
form.     subj.  t  C   t      C     t      C     t       C     t      C
reference     1  PD BQL 0.3333   BQL 0.6667 8.494 1.0000 26.518 1.2500 27.692
              2  PD BQL 0.3667 0.345 0.6667 3.424 1.0000  9.084 1.2500 14.300

So you can present the data in an nonambiguous manner.

[Angus McLean]

Thanks Helmut: There is no actual FDA guidance on this topic. I do see the European Guidance.

 

I like your idea of non-ambiguity in the presentation, since there is a lot of confusion on this point in companies.

 

Angus

[Simon Davis]

Ditto Helmut's response, I actually think performing the NCA using a mix of nominal times within a window and and actual times outside the window is in fact much more work, more error prone and less accurate.

 

  Hope everyone can have a nice break and then make it's their New Year's resolution to put together a neat NCA workflow template using Helmut's proposals!

 

 Simon.

[Angus McLean]

Simon: Thanks for you comments: I am going to pursue Helmut's suggestion to see how it goes.

 

Best wishes over the festive season and the New Year.

 

 

Angus