From: "Hu, Chuanpu"Subject: [NMusers] Capturing AUC between observation time intervals Date: Wed, 3 Apr 2002 14:41:04 -0500 Dear NMusers, A less trivial problem according to Bill Bachman as of some weeks ago. I would like to capture the AUC (either central or peripheral compartment) at time intervals between observations. One use of this could be to predict a weekly-observed PD endpoint using the weekly AUC. Could anyone let me know how something like this could be done? Thanks, Chuanpu ---------------------------------------------------------------- Chuanpu Hu, Ph.D. Modeling and Simulation, U.S. CPDM GlaxoSmithKline Tel: 919-483-8205 Fax: 919-483-6380 ---------------------------------------------------------------- ******* From: Leonid Gibiansky Subject: Re: [NMusers] Capturing AUC between observation time intervals Date: Wed, 03 Apr 2002 16:42:40 -0500 I think you have a choice: 1. Create a data set with a lot of intermediate time points, and then come up with the cumulative AUC=concentration*time 2. Compute AUC analytically based on the dose history and PK parameters (the quickest way if the dose history is not too complicated) 3. Use differential equations with one of the variable being AUC (via equation dAUC/dt=C) 4. Try to imitate (3) via additional compartment with fixed rate constants: dA4/dt=kC4*A1-k4C*A4, k4C is about 0, kC4 is also very small but significantly higher than k4C. Then (A4/kC4)/VC is an approximation of AUC, if I am not mistaken (C is the central compartment). Leonid ******* From: "Sam Liao" Subject: RE: [NMusers] Capturing AUC between observation time intervals Date: Wed, 3 Apr 2002 16:52:37 -0500 Hi Chuanpu: For your question, please see the following archived nmusers msg. http://www.cognigencorp.com/nonmem/nm/98feb192002.html Best regards, Sam Liao, Ph.D. PharMax Research 270 Kerry Lane, Blue Bell, PA 19422 phone: 215-6541151 efax: 1-720-2946783 ******* From: "Hu, Chuanpu" Subject: RE: [NMusers] Capturing AUC between observation time intervals Date: Wed, 3 Apr 2002 17:02:15 -0500 Leonid/Sam, I guess my original message was not clear. Say that I have time points T1=0, T2, T3, ... I would like to do keep AUC(T1, T2), AUC(T2,T3), etc., during NONME fit. That is, I want to simutaneously fit PK (which drives the AUCs) and the PD parameters to fit the PD endpoint. Via eauation dAUC/dt=C I can compute the cumulative AUC, but somehow a variable needs to be created to memorize the cumulative AUC at the previous time. I do not see how the suggested approaches could be used. Am I missing something? Chuanpu ******* From: "Hu, Chuanpu" Subject: RE: [NMusers] Capturing AUC between observation time intervals Date: Wed, 3 Apr 2002 17:29:54 -0500 Leonid has a nice rephrase of the question: "how to store some value computed at one time point to access and update it at another time point". Exactly what I would like to do. Thanks for any ideas. Chuanpu ******* From: "Sam Liao" Subject: RE: [NMusers] Capturing AUC between observation time intervals Date: Wed, 3 Apr 2002 19:05:41 -0500 Chuanpu: My mehtod will work if you fit PK and PD data sequentially. I think I understand what you are trying to do now. But you may need a supper computer to run this model even if you find the way to calculate the AUC in your PK/PD model fitting. The AUC has to be recalculated for every iteration for the whole treatment period. It may take a long time to run this model fitting. For example, some time ago, I tried to calculate AUC based on the PK fit parameters and daily dosing information for a 6-months PK/PD study in over 500 patients, then used the AUC to model the PD end points. I added 10 data points in each day to run NONMEM in simulation mode for just one patient in order to estimate how much time will it take for over 500 patients. I found that for the 60-days time course, it took 7 hours and for the 30-days, it took 10 min. NONMEM slowed down dramatically when the no. of dosing and observation increased. I used a PC with 1.5 GHz Pentium 4 with 780 KB RAM to run the job. I found this method is not feasible to use because it will probably take over 6 months to run the NONMEM simulation. The other method I used (see http://www.cognigencorp.com/nonmem/nm/98feb192002.html) only take few minutes to run NONMEM. Best regards, Sam Liao, Ph.D. PharMax Research 270 Kerry Lane, Blue Bell, PA 19422 phone: 215-6541151 efax: 1-720-2946783 ******* From: Mats Karlsson Subject: Re: [NMusers] Capturing AUC between observation time intervals Date: Thu, 04 Apr 2002 06:21:22 +0200 Hi Chuanpu, I think it would work if you used the code where DADT(x)=C gets the cumulative AUC and then "empty" compartment X and the suitable times by turning the CMT off and back on again (using CMT=-x and CMT=x with EVID=2 data items). This doesn't store the old value, but would work for the problem you described. Best regards, Mats -- Mats Karlsson, PhD Professor of Pharmacometrics Div. of Pharmacokinetics and Drug Therapy Dept. of Pharmaceutical Biosciences Faculty of Pharmacy Uppsala University Box 591 SE-751 24 Uppsala Sweden phone +46 18 471 4105 fax +46 18 471 4003 mats.karlsson@farmbio.uu.se ******* From: "Hu, Chuanpu" Subject: RE: [NMusers] Capturing AUC between observation time intervals Date: Thu, 4 Apr 2002 12:46:53 -0500 Mats, Your suggestion makes a lot of sense. Thanks a lot! Chuanpu