From: "Partha Nandy" partha.nandy@bms.com
Subject: [NMusers] Interoccassion variability and Auto-induction model
Date: Fri, September 10, 2004 10:37 am
Hi All,
I have a question for the group, and any help/suggestion that you can
provide is greatly appreciated.
If a drug exhibits auto-induction, how feasible is it to estimate
Interoccassion variability, and how reliable are the estimates in such a
situation? I am not sure whether one can estimate IOV in this case.
Does anyone have prior experience with such a situation?
Thanks
Partha
Partha Nandy
Clinical Discovery-BMS
Interoccassion variability and Auto-induction model
From: "Gordi, Toufigh" Toufigh.Gordi@cvt.com
Subject: RE: [NMusers] Interoccassion variability and Auto-induction model
Date: Fri, September 10, 2004 1:54 pm
Dear Partha,
I guess the answer depends on how you model the auto-induction and what data you
have. We have applied a semi-physiological model to data on the auto-induction of
artemisinin given orally. The model includes estimations of IOV on ka and time-lag.
The induction is modeled as increase in the enzyme amounts, which increase the
extraction ratio of the compound, thereby affecting its F and CL. The manuscript is
sent to BJCP and is under review. I would be more than happy to provide you with the
control stream.
Best wishes,
Toufigh Gordi
From: Gastonguay, Marc" marcg@metrumrg.com
Subject: RE: [NMusers] Interoccassion variability and Auto-induction model
Date: Mon, September 13, 2004 9:57 am
Hello Partha,
Given an adequate experimental design and an appropriate auto-induction
model, you should be able to separate the systematic time-dependent
increases in clearance due to auto-induction, from the non-systematic random
fluctuations in clearance across sampling occasions.
You can check the precision (reliability) of the estimates by obtaining some
measure of estimation uncertainty (NONMEM standard errors, bootstrap CI,
etc.). The reliability of the model for your intended purposes could be
explored by performing a predictive check (Monte Carlo simulations of the
original data design) using models with/without IOV and comparing the
predictive performance based on a characteristic of interest from the
observed data (e.g. AUC, Cmin, Cavg etc. on each occasion).
Marc
==========================
Marc R. Gastonguay, Ph.D.
Metrum Research Group LLC
15 Ensign Drive
Avon, CT 06001
T: 860.670.0744
F: 860.760.6014
E: marcg@metrumrg.com
www.metrumrg.com
_______________________________________________________