From: "Ralph Quadflieg" <quadflieg@uni-bonn.de>

Subject: 4Comp.mod.

Date: 28 Sep 1998 14:16:27 -0400

Dear NONMEM users,

I am a new NONMEM user and I have some difficulties in writing a control file for my pk problem:

I would like to model 3 data sets (3 different settings) with a 4 compartment model (depot, central, peripher1, peripher2).

a) the depot comp. should be the dosing compartment with a first order process to the central comp. (is not known) a bioavailability factor (is not known, oral appl.) and a lagtime parameter (is not known)

b) the central comp. should be the dosing comp. with two different constant rates (are known) and a bioavailability factor = 1 (iv infusion)

I thought that this should be a problem for ADVAN6 and TRANS1 with a MODEL record like this:

MODEL

NCOMPARTMENTS = 4

COMP = (DEPOT)

COMP = (CENTRAL, DEFOBS)

COMP = (PER1)

COMP = (PER2)

LINK DEPOT TO CENTRAL = 1

LINK CENTRAL AND PER1 = 3

LINK CENTRAL AND PER2 = 5

LINK CENTRAL TO OUTPUT = 7

PK

KA = THETA(1) * EXP(ETA(1))

V2 = THETA(2) * EXP(ETA(2))

CL23 = THETA(3) * EXP(ETA(3))

V3 = THETA(4) * EXP(ETA(4))

CL24 = THETA(5) * EXP(ETA(5))

V4 = THETA(6) * EXP(ETA(6))

CL20 = THETA(7) * EXP(ETA(7))

F1 = THETA(8) * EXP(ETA(8))

ALAG1 = THETA(9) * EXP(ETA(9))

S1 = V1

k23 = CL23/V2

k32 = CL23/V3

k24 = CL24/V2

k42 = CL24/V4

k20 = CL20/V2

and a corresponding data file with RATE CMT EVID PCMT data items.

But there are problems due to the MODEL block, which I cannot solve (especially the LINK statement).

I am a bit confused about the different possibilities of working with the LINK statements within the MODEL block or within th PK block, I should say.

The MODEL is established with differential equations, but with this control file the run takes a lot of time. (around one hour).

I would be pleased if there was somebody who can help me.

Thank you in advance

Ralph Quadflieg

Dept.Pharm.Techn.Biopharm.

University of Bonn

An der Immenburg 4

D-53121 Bonn

Germany

quadflieg@uni-bonn

http://www.pharma.uni-bonn.de/pharmtech/

Tel.: ++49 228 218997

Fax.:++49 228 735268

****

From: alison@c255.ucsf.EDU (ABoeckmann)

Subject: Re: 4Comp.mod.

Date: 28 Sep 1998 16:13:20 -0400

Ralph Quadflieg asked about a 4 compartment model.

The error messages arise because LINK clauses in $MODEL are used only with a user-written PK routine *and* with either ADVAN5 or ADVAN7.

With $PK, he should omit the LINK clauses.

1) With ADVAN6 8 or 9, use a $DES block to define the differential equations, using the K's defined in $PK.

E.g.,

$DES

DADT(1)=-KA*A(1)

etc.

2) Since the kinetics are linear, ADVAN5 or ADVAN7 can be used. NMTRAN will understand that these statments in $PK imply the links between compartments, providing that the lower case k's are changed to upper case K's:

K23 = CL23/V2

K32 = CL23/V3

K24 = CL24/V2

K42 = CL24/V4

K20 = CL20/V2

Note that the following are also needed:

K12=KA and a model for V1.

Here is a better approach.

With NONMEM V, the analytic solution to the 4 compartment problem has been supplied as ADVAN12. This will give faster run times than the general ADVAN routines 5,6,7,8,9.

Omit the $MODEL block. Use the original $PK block (with uppercase k's as noted above).

With ADVAN12 TRANS1, the statement

K12=KA

is not needed, but

K=K20

is needed.

It is also possible to use ADVAN12 with TRANS4. This is a very similar parameterization, with slightly different notation.

See Users Guide VIII, ADVAN12 and TRANS4 entries (or nmhelp advan12, nmhelp trans4), for the details.