From Sun May 4 10:14:15 1997
Subject: 3 compartment model TRANS and ADVAN

We are conducting an study with NONMEM at NC State University in which we are trying to fit datapoints obtained both from blood and a peripheral compartment using a three compartment structural model. All the rate constants are first order and elimination is from the central compartment. We have a data rich situation for the blood compartment and a sparse data situation for the observations from the peripheral compartment.

What advan and trans subroutines should be used?.

Is there a model already built-up in NONMEM to handle this, and if not, what kind of user-defined code should we input?.



From alison Mon May 5 09:52:39 1997
Subject: Three compartment model

In NONMEM Users Guide Part VIII (Help Guide) p. 242 there is an example of the three compartment model with first order absorption, using ADVAN7 and TRANS1.

I suggest looking at this example first, and at the suggestions for modifications (e.g., the DEPOT compartment may be omitted).

For observations from the peripheral compartment, record those observations in the DV column. Set the value of CMT to 3 (for the first peripheral compartment) for these observations. Include in the $PK a model for S3 (scale for first peripheral compartment).

The $ERROR model may also test CMT, since the intraindividual variability may be quite different for the two kinds of observations. See Users Guide Part V (Introductory Guide) p. 82 (Chapt8 3.6), Two Different Types of Observations, for an example.

Alison Boeckmann
NONMEM Project Group