Back to the Top
Dear all,
A dose-range study of our compound (administration by iv or po at 3 different
doses), had shown us that threre is nonlinearity in its kinetics(there is
incresed half-life, decreased clearance, disproportionality in AUC and so on
with increase in dose).
At lower doses the concentration-time profile clearly fits with two
compartmental model, at highest dose the data fits with one compartmental
model.
We do not know the actual Km and Vmax values for this compound and I can only
do very basic compartmental or noncompartmental analysis of the data with the
software on hand. My question is whether it would be right (or not) to
compare the PK parameters obtained at lower doses (by two-compartmental
fitting) with those obtained at the highest dose (by non-compartmental
analysis)? The reason that I had chosen non-compartmental analysis for the
highest dose is that presumably I should not be using compartmental modeling
in a nonlinear or mixed order situation. Can anyone provide me with any
suggestions?
Thanks
Back to the Top
You have stated the question in terms of characterizing the data that you
have...namely the plasma concentrations and sampling times. However, the
real question is how you characterize the data that you do not have (or have
not mentioned). Because you have very differently shaped profiles at a high
and low dose, you may be missing samples at late enough times, following in
the high dose, to see the complete profile. On the other hand, this may be
indicative of a type of nonlinearity different from the classical
Michaelis-Menten, capacity-limited clearance, i.e., creation of an
inhibitory metabolite, substrate depletion...
Others on this list may state it more bluntly, but the only way to account
for such deficiencies in design, knowledge of mechanisms, or other sources
of information is a model based approach that includes *ALL* the data.
There are numerous commercial and academic software packages that are
capable of analyzing this sort of data and they are all priced well below
the cost of having run this study, and most likely less expensive than the
cost of performing an analysis that leads to an incorrect conclusion. There
have been several discussions of different software packages on this list in
the past, and David Bourne has a nice list of resources published at the
PharmPK website (http://www.boomer.org/pkin/).
Regards,
Jeff Wald, PhD
Pharsight Corporation
Back to the Top
Hello Panteha
It would be better if you can analyse all the dose groups in one model based
on mechanism of structurally similar drugs. This would enable you to make
better predictions. The reliability of estimates is questionable. I am not
sure if you have your class of compounds and if there are reported
literature values, you can perhaps make a reasonable approximation.
Venkatesh Atul Bhattaram
Post-doctoral Fellow
University of Florida
Gainesville-32610
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)