From: "Gordi, Toufigh" Toufigh.Gordi@cvt.com
Subject: [NMusers] Modeling of iv and po data
Date: Wed, August 11, 2004 4:31 pm

Dear all,

I am trying to model data from a study including both iv and po data. The
bioavailability is quite low (~15%). The highest po doses given were 10 times higher
than the highest iv dose.

When modeling the iv data alone, a non-linear CL gives the best fit. Such a model,
however, does not converge when iv and po data are combined.  The major problem, I
believe, is the estimation of ka.

Any suggestion on how to proceed? And in general, if one of the characteristics of
the model including iv data (let's say the non-linear CL) is masked by the PK
profile after po, what would be the best approach to model the data? Model the data
separately or apply a simpler model to the combined data set?

best wishes,

Toufigh Gordi 
_______________________________________________________

From: Nick Holford n.holford@auckland.ac.nz
Subject: RE:[NMusers] Modeling of iv and po data
Date: Wed, August 11, 2004 4:54 pm

Toufigh,

If you have mixed order elimination detectable from the IV dose then it is quite
possible that you will have rate dependent first pass extraction with the oral dose
e.g. see Holford NHG. Complex PK/PD models - an alcoholic experience. International
Journal of Clinical Pharmacology and Therapeutics 1997;35(10):465-468.

I would not worry about lack of convergence until you have got a good handle on the
structural model. In the end the visual goodness of fit of the observed and
predicted values and perhap a simple predictive check (simulation from your final
model) are more meaningful than NONMEM's semi-arbitrary decision to declare
'convergence'.

Nick
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
email:n.holford@auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
_______________________________________________________