Back to the Top
The following message was posted to: PharmPK
Dear all,
Recently, there were a series of discussions on how to assess the
linearity of a compound. Most comments referred to some kind of
regression analysis of Cmax or AUC values.
I am interested to hear some comments on using a modeling approach in
addressing the same question. Dose-linearity studies normally include
data from more than 2-3 dose levels and relatively extensive sampling.
This should make these studies suitable for modeling. The modeling
process can in turn include testing of whether non-linear processes are
present in the system.
Any thoughts?
Toufigh Gordi
Back to the Top
The following message was posted to: PharmPK
Dear Toufigh,
I believe we do not need to make things more complicated than
necessary. I
cannot think of any way of postulating PK models (time vs. conc) without
looking at the non-compartmental analysis (nca) and/or graphs first.
Most
dose-disproportionalities (that we really care about) are detected by
the
nca in my experience. So, why model if that is only the question we are
interested in: Does this drug exhibit dose-disproportional kinetics?
From a purely academic perspective, linear and most non-linear PK
models
(say, first-order vs. michaelis-menten elimination) are not nested.
So, the
usual model testing criteria (like log-likelihood ratio test) per se
cannot
be applied to derive significance levels. You need to use empirical
computationally intense methods (such as permutation tests) to derive
those.
Is it worth it? In my opinion, no. There are other empirical PK
models to
describe non-linearity (for eg, linear increase in bioavailability with
dose) that might be amenable to the usual criteria. One can test
constant F
vs. dose-dependent F (but need to conduct nca first though).
Regards,
Joga
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)