From: Miguel Alexandre Soares <miguel.a.soares@netc.pt>

Subject: problem advan4 trans4

Date: Thu, 18 May 2000 13:33:09 +0200

Dear group

I am using the input file bellow to model the clinical data of a drug that is known to be well described by a two compartment model with oral absorption. After analyze the output file, I noticed that NONMEM didn't change V3, Ka, F2 and ALAG2. I tried other models but was unable to solve the problem. Could someone help me to find where the problem is?

$PROBLEM CCA

$INPUT ID,DAT1=DROP,TIME,Dose=AMT,Cp=DV,Sex,Wt,Ht,SCr,Age,TrA,BSA,CrCL,LBW

$DATA ..\temp\CCA1.CSV WIDE IGNORE=I

$SUBROUTINE ADVAN4 TRANS4

$PK

TCL=THETA(1)

TV2=THETA(2)

d#

TQ=THETA(3)

TV3=THETA(4)

TKA=THETA(5)

CL=TCL+ETA(1)

V2=TV2+ETA(2)

Q=TQ

V3=TV3

KA=TKA+ETA(3)

K=CL/V2

K23=Q/V2

K32=Q/V3

S2=V2

S3=V3

F2=THETA(7)

ALAG2=THETA(6)

$ERROR

Y=F+EPS(1)

$THETA

(0,10,) (0,90,) (0,25,) (0,170,) (0,2,) (0,0.1,) (0,.6,1)

$OMEGA DIAGONAL(3) 0.30 0.67 0.25

$SIGMA DIAGONAL(1) 0.40

$ESTIMATION MAXEVAL=9000 POSTHOC SIGDIGITS=3 PRINT=5

$COVARIANCE

$TABLE ID CL Q V2 V3 KA ETA(1) ETA(2) FILE=TABELA.TXT NOPRINT ONEHEADER

$SCAT DV VS PRED UNIT

$SCAT DV VS RES ORD0

~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Dr. Miguel Alexandre Soares

Instituto Superior Ciências da Saúde - Sul

Campus Universitário

Quinta da Granja

2825 Monte da Caparica - Portugal

e-mail: miguel.a.soares@netc.pt

~~~~~~~~~~~~~~~~~~~~~~~~~~~~

From: "Bachman, William" <bachmanw@globomax.com>

Subject: RE: problem advan4 trans4

Date: Thu, 18 May 2000 08:48:52 -0400

I believe you really want to estimate F1 and ALAG1 (not F2 and ALAG2). (The depot compartment is 1.) Try that.

William J. Bachman, Ph.D.

GloboMax LLC

Senior Scientist

7250 Parkway Drive, Suite 430

Hanover, MD 21076

Voice (410) 782-2212

FAX (410) 712-0737

bachmanw@globomax.com

From: "Sreenivasa Rao Vanapalli" <rao@pharm.cpb.uokhsc.edu>

Subject: RE: problem advan4 trans4

Date: Thu, 18 May 2000 10:07:10 -0700

I don't know how far I'm right about my suggestion. But i had the same problem with two compartment first order oral abosrption. I did not calculate CL as parameter so I used ADVAN4. With my meagre knowledge I can add my own words to this problem. As you mentioned you had problem with V3, KA, F2, ALAG2. The problem is with declaring your parameters in the control file. Why did you chose to reparameterize V3 with S3 (S3=V3). Just leave V3 as it is. Next, if you check the NONMEM manual I think volume V, you will find what parameters can be calculated with ADVAN4 TRANS4. Because this does not support F2 and ALAG2 (so does't calculate). Next with KA, try reparameterizing KA=KA and see if that solves your problem. I'm a graduate student learnng alphabets in NONMEM. Please advice me if I'm wrong.

Sreenivasa Rao Vanapalli

College of Pharmacy

University of Oklahoma health Sciences center

Oklahoma city, OK-73106

From: ABoeckmann <alison@c255.ucsf.edu>

Subject: RE: problem advan4 trans4

Date: Thu, 18 May 2000 08:48:29 -0700 (PDT)

Dr. Soares writes:

> I am using the input file bellow to model the clinical data of a drug

> thatis known to be well described by a two compartment model with oral

> absorption. After analyze the output file, I noticed that NONMEM

> didn't change V3, Ka, F2 and ALAG2. I tried other models but was

> unable to solve the problem. Could someone help me to find where the

> problem is?

Bill Bachmann responded:

> I believe you really want to estimate F1 and ALAG1 (not F2 and ALAG2).

Sreenivasa Rao Vanapalli responded:

> Why did you chose to reparameterize V3 with S3 (S3=V3). Just leave V3

> as it is. Next, if you check the NONMEM manual I think volume V, you will

> find what parameters can be calculated with ADVAN4 TRANS4. Because this does

> not support F2 and ALAG2 (so does't calculate). Next with KA, try

> reparameterizing KA=KA and see if that solves your problem.

They are correct about F2 and ALAG2. With no CMT data item, there can be no doses to compartment 2, hence F2 and ALAG2 have no effect on the model. Similarly, there can be there can be no observations from compartment 3, hence S3 cannot have any effect. However, V3 does affect the model via K32, so THETA(3) should be capable of estimation. Similarly, KA (THETA(5)) is capable of estimation. KA=KA is not needed.

I think one would have to see the data and the output to know what is going wrong. Are the gradients for thetas 3 and 5 always zero? Does NONMEM reach a minimum? Are the predictions (with and without $ESTIM) reasonable? Are there any observations close enough to the dose time (during absorption) to allow KA to be estimated?

Also, the following statements are not needed with TRANS4, because they are what TRANS4 does. They do no harm, but can be deleted. Or, use TRANS1:

K=CL/V2

K23=Q/V2

K32=Q/V3

Alison Boeckmann

===========================

$PROBLEM CCA

$INPUT ID,DAT1=DROP,TIME,Dose=AMT,Cp=DV,Sex,Wt,Ht,SCr,Age,TrA,BSA,CrCL,LBW

$DATA ..\temp\CCA1.CSV WIDE IGNORE=I

$SUBROUTINE ADVAN4 TRANS4

$PK

TCL=THETA(1)

TV2=THETA(2)

TQ=THETA(3)

TV3=THETA(4)

TKA=THETA(5)

CL=TCL+ETA(1)

V2=TV2+ETA(2)

Q=TQ

V3=TV3

KA=TKA+ETA(3)

K=CL/V2

K23=Q/V2

K32=Q/V3

S2=V2

S3=V3

F2=THETA(7)

ALAG2=THETA(6)

$ERROR

Y=F+EPS(1)

$THETA

(0,10,) (0,90,) (0,25,) (0,170,) (0,2,) (0,0.1,) (0,.6,1)

$OMEGA DIAGONAL(3) 0.30 0.67 0.25

$SIGMA DIAGONAL(1) 0.40

$ESTIMATION MAXEVAL=9000 POSTHOC SIGDIGITS=3 PRINT=5

$COVARIANCE

$TABLE ID CL Q V2 V3 KA ETA(1) ETA(2) FILE=TABELA.TXT NOPRINT ONEHEADER

$SCAT DV VS PRED UNIT

$SCAT DV VS RES ORD0

~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Dr. Miguel Alexandre Soares