PharmPK Discussion - Predicting the Profile of Multiple Doses?

PharmPK Discussion List Archive Index page

On 13 Mar 1998 at 09:56:15, "Tony Lee" (p2149z.aaa.hotmail.com) sent the message


Dear All:

My current work involves predicting the
concentration-time profile after multiple
dosing using data from single dose.
My method is to fit the data to one or
two compartment model (1CM and 2CM)
and then do the simulation. The problem
is that either one or two compartment
model can fit the data according to the
model selection criteria. (AIC and SC and
R square of 1CM a little less than those of
2CM). As you understand, the profiles
predicted are very different with 1CM or
2CM. Besides the superposition method,
Could you please recommend easy and
fast ways to do this (I mean profiles not
just Cmax, Cmin and Average C at SS.)
Or any software is good at this?
Thank you for your suggestions.

Tony Lee
Emailto:p2149z.-at-.hotmail.com

Back to the Top

On 16 Mar 1998 at 13:45:43, Janusz Byczkowski (byczkowskij.-a-.FALCON.AL.WPAFB.AF.MIL) sent the message


ANSWER:
   You may construct a physiologically based pharmacodynamic model (PBPK).
Apparently you have enough data to calibrate it. The best software to do
this is ACSL (Advanced Continuous Simulation Language). You may download
examples and demo viewer from http://brandywine.net/users/plowchdr/
and/or contact software distributor: http://www.mga.com/
   By the way, do you fit data to the model or rather you fit model to the
data? :-)
             Good luck.
             Janusz Z. Byczkowski, Ph.D., D.Sc.
             ManTech Environmental Technology, Inc.
             Wright Patterson AFB, OH
(937)255-5150 X3121

Back to the Top

On 16 Mar 1998 at 13:48:15, "Thomas Senderovitz" (senderovitz.at.inet.uni2.dk) sent the message


Dear Dr. Lee,

I suggest you do time-concentration plots of your data before anything
else. Look at the plots. Which model would you suggest fitted to the data
best?
If this is not obvious, you must try different scenarios, and also take a
closer look at how your weight your data.
Is it correct that BOTH a 1 and a two compartment model fit the data (or
did neither of them?).
If the first is true, why not chose the least complicated? Ifg the second
is true, then analyze the data again and again. Perhaps a 3 cpompartment
model is better? Perhaps the weighting scheme is wrong?
I want to point out that perhaps you will never be able to make a model fit
the data, if the data are not good enough! I do not think that the ultimate
goal is to make the best fit if this is simply not possible. It is to treat
the data in the most "honest" way.

I can recommend the WinNonlin software, version 1.5. (either the
professional or the standard version, the latter probably would do it).

Hope this is helpfull.

Thomas Senderovitz, MD
Unit of Clinical Pharmacology
Bispebjerg Hospital
University of Copenhagen
E-mail: senderovitz.at.dadlnet.dk

Back to the Top

Want to post a follow-up message on this topic? If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "Predicting the Profile of Multiple Doses?" as the subject

PharmPK Discussion List Archive Index page