Date: Fri, 04 Feb 2000 16:47:41 -0500
From: Meiyu Shen <M_Shen@fccc.edu>
Subject: Coding problem
I would like to use a 2-compartment indirect response model as a simulation tool. In this model, the link between the PK and Pd model is via AUC(t) rather that C(t) [plasma drug concentration]. I am unsure of how to write the code that would allow me to use AUC in this type of model. I want to generate different C-t profiles, which calculate the associated AUC, and then input the AUC values into the PD model. I would like to all these calculations within NONMEM.
I'd appreciate any help.
Meiyu Shen
Fox Chase Cancer Center
Tel: 215-728-3513
Email: M_Shen@fccc.edu
Coding problem
Date: Sat, 05 Feb 2000 18:42:42 +1300
From: Nick Holford <n.holford@auckland.ac.nz>
Subject: Re: Coding problem
At first sight there seems to be some misthink going on here. AUC as I understand the term is the integral of conc over some time interval so that AUC(t) does not make sense unless the integration interval is specified AND the value of t for AUC(t) refers to some specific time related to the integration interval e.g. the mid-point.
It is commonly believed in oncology that AUC has some special property for predicting drug effects but in fact it is just an empirical statistic that has some correlation with intensity of drug effect.
If you insist on using AUC then you *can* use AUC(t) in an indirect effect model e.g.
dadt(1) = -CL*A(1) ; solution is amount(t) in cpt 1
dadt(2) = A(1)/V ; solution is AUC(t)
dadt(3) = Kin*(1-A(2)/(A(2)+IC50)) - Kout*A(3) ; soln is effect(t)
Note that this model can only predict inceasing AUC(t) so that effect(t) will change monotonically and the effect of the drug cannot "wear off". Each successive dose will lead to a further increase of drug effect.
If you arrange to "reset" the solution in compartment 2 after every dose (or dose cycle) (using EVID=4) then it is possible to relate the AUC(t) after each dose to the effect observed after that dose. However, overall I suggest it is makes more sense to use concentration(t) rather than AUC(t) to affect the indirect effect model because this will naturally take care of the onset and wearing off of effect after each dose.
If you are smart enough to understand about indirect effect models which can describe the time course of response of biological mechanisms then I suggest you try hard to throw away the AUC concept and think more about biology when devising models for the time course of drug effect. Remember that the basic property of AUC is in some sense to discard time by integrating it away so it is not likely to be a helpful quantity to understand time related phenomena.
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, Private Bag 92019, Auckland, New Zealand
email:n.holford@auckland.ac.nz tel:+64(9)373-7599x6730 fax:373-7556
http://www.phm.auckland.ac.nz/Staff/NHolford/nholford.htm
Date: Sat, 05 Feb 2000 11:28:35 -0500 (EST)
From: "Jogarao Gobburu 301-594-5661 FAX 301-480-8329" <GOBBURUJ@cder.fda.gov>
Subject: Re: Coding problem
Hello,
While I am not sure why AUC needs to be used instead of concentration, the concentration profile can be integrated by adding a dummy compartment, as:
DADT() = A(?)/S? ; where A(?) is the amount in plasma, S? is the
; volume of distribution
This should give you the AUC(0-TIME).
Hope this helps.
Regards,
Joga
Joga Gobburu
Pharmacometrics,
CDER, FDA