PK Analysis in Toxicokinetic Studies

NONMEM Topic 4

Keywords: Toxicokinetic, Animal Study

Topic started by: ganesda@pr.cyanamid.com (Derick Ganes) - 28 Jan 1994

What is the most appropriate way to naively pool plasma
concentrations for noncompartmental pharmacokinetic analysis
of data obtained in routine toxicokinetic studies in
rodents?  Typically, many individual plasma concentrations
are below the limit of quantification (BQL) at this stage of
drug development.

A typical study design is as follows: 6 rats per sex per
dose are divided into two sets of 3.  A maximum of 3 blood
samples is obtained from each set on an alternating basis.
Blood sampling times are 0.5, 1, 2, 4, 6 and 8 hr after
dosing.

Is it appropriate to substitute missing BQL values with
zero, obtain the mean, then determine the pharmacokinetic
parameters?  Is there a superior method for summarizing the
data?  Should NONMEM be used?  I look forward to your
comments.

Sincerely, Derek A. Ganes, Ph.D.

Reply by nholford@ccu1.auckland.ac.nz (Nick Holford)

> What is the most appropriate way to naively pool plasma
> concentrations for noncompartmental pharmacokinetic analysis
> of data obtained in routine toxicokinetic studies in
> rodents?
There is no *appropriate way* to naively pool etc.! But less naive
methods may be useful.

> Is it appropriate to substitute missing BQL values with
> zero, obtain the mean, then determine the pharmacokinetic
> parameters? 
NO. (some questions are very easy to answer)
Reason - Not so quick to write. BQL means less than some non-zero value.
If you think your errors inmeasurement may be proportional to conc then
putting a zero in is a long way from non-zero.

> Is there a superior method for summarizing the
> data?  Should NONMEM be used?  I look forward to your
> comments.
NONMEM could be used. If I understand your design each rat may have more
than one conc measured. Thats all you need to get started on a
non-naive analysis.

Nick Holford

End of Topic - 8 Apr 94