NONMEM Users Network Archive

Hosted by Cognigen

[NMusers] DEFDOSE/DEFOBS Required for Many Compartments?

From: Bill Denney <>
Date: Mon, 1 May 2017 12:33:17 -0400


I have a model with many compartments (42 different compartments are
defined). It is currently setup this way to future-proof the dataset in
case we want to combine analyses of many drugs (about 20 different drugs
with central/peripheral compartments) and several endpoints.

In the current model, I'm only fitting the PK of one drug, so only two
of the compartments are enabled: compartments 39 and 42. After the
IGNORE statement, the data is relatively small with 400 records between
16 subjects, and all the data are in compartment 39 (the central
compartment for the current drug).

CMT, EVID, and MDV (along with everything else) is correctly specified.
When I run the model with CMT 39 defined by:

COMP=(PKCENT) ; 39 central PK compartment (ug/mL)

the model fails with the error "COMPARTMENT ASSOCIATED WITH THE
PREDICTION IS OFF" noted for all the dosing rows, but when I specify
that compartment 39 is the default observation and dose compartment (due
to IV dosing), the model works.

COMP=(PKCENT,DEFOBS,DEFDOSE) ; 39 central PK compartment (ug/mL)

I thought that when CMT and EVID are specified, DEFOBS and DEFDOSE are
not important.

Does anyone know why adding DEFOBS and DEFDOSE fixed my model?



Human Predictions Logo <>*William S.
Denney, PhD*
Chief Scientist, Human Predictions LLC

This e-mail communication is confidential and is intended only for the
individual(s) or entity named above and others who have been
specifically authorized to receive it. If you are not the intended
recipient, please do not read, copy, use or disclose the contents of
this communication to others. Please notify the sender that you have
received this e-mail in error by replying to the e-mail or by calling
+1-617-899-8123. Please then delete the e-mail and any copies of it.
Thank you.

Received on Mon May 01 2017 - 12:33:17 EDT

The NONMEM Users Network is maintained by ICON plc. Requests to subscribe to the network should be sent to: Once subscribed, you may contribute to the discussion by emailing: