Re: Absorption
Dear Adithya,
Your problem has attracted some good advice from the list, but if you are
still interested in trying out the transit model, perhaps we could help,
if you were able to furnish some more details about what exactly went
wrong with your implementation of it in NONMEM V - posted code snippets
and precise error messages would be of particular use.
It ought to be possible to make it work under NMV, although the richer the
data, the better, and NMV is known to be less robust at working with
differential equations than NMVI. Problems can often be resolved with some
minor tweaking, given sufficient information in the data to support the
model - how many observations per individual are you working with, and how
well covered is the absorption phase? You might also try hard-coding the
transit compartments, i.e. not estimating the number as a model parameter,
but fixing it to different values until you find an optimum.
You might also want to take a look at some previous work on the subject,
in which the method was established for single dosing:
Savic R, Jonker DM, Kerbusch T, Karlsson MO. PAGE 13 (2004) Abstr 513
[www.page-meeting.org/?abstract=513]
Have you tried normalizing the concentrations by dose and plotting them on
the same axes? That should give you information with respect to whether
you have any dose-related nonlinearity - if the curves for each dose
levels are roughly superimposed, kinetics are linear, and if not, the
relative positions of the curves should provide some useful clues about
what is going on.
Best regards
Justin
Justin Wilkins
Novartis Pharma AG
PH346, MODELING & SIMULATION
CHBS, WSJ-027.1.084
Novartis Pharma AG
Lichtstrasse 35
CH-4056 Basel
Switzerland
Phone: +41 61 324 6549
Fax: +41 61 324 3039
Mobile: +41 76 561 0949
Email : [EMAIL PROTECTED]
"Adithya Muralidharan" <[EMAIL PROTECTED]>
Sent by: [EMAIL PROTECTED]
03.09.2007 19:10
To
[email protected]
cc
Subject
[NMusers] Absorption
Dear nmusers,
I am working on a rising single dose study, with 5 cohorts (10, 25,
50, 100 and 200 mg dose groups). This is a oral drug. There are a
total of 45 subjects in the study, 9 subjects in each cohort.
The data fits well to a two compartment model, but at higher doses 100
and 200mg dose groups the concentrations leading to Cmax is
overpredicted significantly that the model seems to be completely
ineffecient in explaining the absorption phase at higher doses.
I have tried using a
Mixture model assuming that there is a sub population with a totally
different absorption, this seemed to work but visual predictive check
failed.
Used a cobination zero and first order absorption model which couldn't
fix the issue at higher doses
Transit compartment model (A flexible approach to modeling variable
absorption in the context of repeated dosing: illustrated with
rifampicin, J. Wilkins, Page 2007). This approach did not work with
NONMEM V.
i assumed that the drug at higher doses has saturable absorption but
even that does not help much.
Can someone please comment on this.
Thanks
Adithya