# PharmPK Discussion - WNL user defined model - conc below zero

PharmPK Discussion List Archive Index page
• On 30 Nov 2004 at 14:45:22, =?ISO-8859-1?Q? "Cornelia_Burger" ?= (cbuerger.-a-.zedat.fu-berlin.de) sent the message

Dear all,
I am trying to write a user defined model with multiple iv infusions
(WinNonlin (V4.0), please see example below).
The dosing regimen contains 2 doses of 600 mg at 0 and 12 hours and a
duration of infusion of 0.5 hr.
Problem No. 1:
Simulating the data for 2 doses with first dose of 0mg (no change in
NCON) and second dose 600 mg results in concentrations below zero
during the first 12 hours.
Problem No. 2:
Simulating the data for 1 dose at 12 hours (change in NCON) results in
even lower concentrations during the first 12 hours than with problem
No. 1.
In my opinion, no matter which of the two approaches is used, when the
input at t=0 is zero there should be no change in the concentration and
it should never fall below zero. Has anyone ever observed the same
phenomena? What could be a possible explanation? Any help would be
highly appreciated!
Best regards,
Conny
MODEL
COMM
NPARM 2
NCON 7
PNAMES 'V', 'K10'
NFUNC 1
NDER 1
END
TEMP
ndose = CON(1)
INPUT = 0
T=X
END
START
Z(1) = 0
END
DIFF
j=0
DO i=1 TO ndose
j=j+3
IF T LE CON(j) THEN goto red
ENDIF
NEXT
red:
ndose = i-1
j=0
DO i=1 TO ndose
j=j+3
Dose = CON(j-1)
TI = CON(j+1)
T = X - CON(j)
IF T GE TI THEN INPUT = 0
ELSE INPUT = Dose/TI
ENDIF
DZ(1) = (INPUT - K10*Z(1))
NEXT
END
FUNCTION 1
F = Z(1)/V
END
EOM

Back to the Top

• On 1 Dec 2004 at 09:33:45, David Czock (david.czock.aaa.medizin.uni-ulm.de) sent the message

The following message was posted to: PharmPK

Dear Conny,

I assume that CON(1) is the number of doses,
CON(2) is the dose, CON(3) is the begin of infusion, CON(4) is the
duration of infusion, CON(5) is the next dose, ...

try the following:

1. delete "INPUT=0" from the TEMP section
2. replace the DIFF section by

j=0
INPUT=0
do i=1 to ndose
j=j+3
Dose=CON(j-1)
TI=CON(j+1)
IF T GT CON(j) AND T LE (CON(j)+TI) THEN
INPUT=Dose/TI
ENDIF
NEXT
DZ(1)=INPUT - k10*Z(1)

Best wishes
David

Back to the Top

• On 1 Dec 2004 at 17:30:34, =22Cornelia_Burger=22 (cbuerger.-at-.zedat.fu-berlin.de) sent the message

The following message was posted to: PharmPK

Dear David,
thank you very much for your suggestion.
Unfortunately although the values were closer to zero the problem
persisted.

But changing the function block into F=MAX(0,Z(1))/V solved the problem
:).

Thank you again for your help.

Best regards
Conny

Back to the Top

Want to post a follow-up message on this topic? If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "WNL user defined model - conc below zero" as the subject

Copyright 1995-2010 David W. A. Bourne (david@boomer.org)