# Re: Sequential PK PD

From: Indranil Bhattacharya <ibhattacharya>
Date: Wed, 10 Feb 2010 09:50:40 -0500

Leonid, the code when fitting the PK data is as follows
\$PK
TVCL=THETA(1)
CL=TVCL*EXP(ETA(1))
TVV2=THETA(2)
V2=TVV2*EXP(ETA(2))
TVQ=THETA(3)
Q=TVQ
TVV3=THETA(4)
V3=TVV3
TVKA=THETA(5)
KA=TVKA
TLAG1=THETA(6)
ALAG1=TLAG1
TLAG2=THETA(7)
ALAG2=TLAG2+TLAG1
TVD2=THETA(8)
D2=TVD2
F1=THETA(9)
F2=1-F1
S2=V2/1000

\$ERROR
IPRED = -5
IF (F.GT.0) IPRED = LOG(F) ;log transforming predicition
IRES=DV-IPRED
W=1
IWRES=IRES/W ;Uniform Weighting
Y = IPRED + ERR(1)

\$EST METHOD=1 INTERACTION PRINT=6 MAX=9999 SIG=3 MSFO=6.msf

There was no need for \$DES here and the data was set up as follows
DV EVID MDV CMT AMT RATE . 1 1 1 100 0 . 1 1 2 100 -2 5.556828 0 0 2 . .
5.968708 0 0 2 . .

you should keep K0=0 before the time ALAG1, and then switch K0 to zero again
as soon as A(1)=0, how it is done in your code? Is it outside of what we
see, in the definition of KK0 in the input file?
When using ADVAN4 TRANS4 and setting up the data set the switching was done
BY NONMEM based on ALAGS, correct? That one extravascular Dose given at time
=0, after ALAG1 KA, comes into play and after ALAG2 K0 comes into play?

When using the \$DES for sequential PK/PD, then should I write the following
IF(T.LE.ALAG1) THEN
IF (T.LE.ALAG2) THEN
ELSE
END IF

Thanks again for your feedback.

Neil
On Wed, Feb 10, 2010 at 9:20 AM, Leonid Gibiansky <LGibiansky
> wrote:

> Nail,
>
> Your PD model should be OK as long as your PK model is working properly,
> but I cannot figure out how PK part works in your code. I do not see any
> control over drug amount in the depot compartment: you should keep K0=0
> before the time ALAG1, and then switch K0 to zero again as soon as A(1)=0,
> how it is done in your code? Is it outside of what we see, in the definition
> of KK0 in the input file? Also, the role of this K0 is not clear: you
> separate the dose into the "first order" part that goes to the first
> compartment, "zero order" part that goes to the second compartment, and then
> use K0 to move more drug from the first compartment to the second one, was
> it intentional?
>
> Regarding your specific questions:
> Q1: According to your code, K0=KK0, and the KK0 should be provided in the
> \$INPUT
> Q2: ALAG1 and ALG2 are delays of the dose to the first and the second
> compartment, respectively, and nonmem will interpret it correctly
>
> My guess is that you should set K0=0, and then everything will work as you
> intended (the PK model should be identical to the PK part of the PD model).
>
> Thanks
> Leonid
>
> --------------------------------------
> Leonid Gibiansky, Ph.D.
> President, QuantPharm LLC
> web: www.quantpharm.com
> e-mail: LGibiansky at quantpharm.com
> tel: (301) 767 5566
>
>
>
>
>
> Indranil Bhattacharya wrote:
>
>> Hi, I have a question regarding how NONMEM reads the code when doing a
>> sequential PK/PD run.
>> I fit the PK data (extravascular admin) using a KOKA model with sequential
>> absorption. The data was based described by this sequential model where K0
>> process starts after Ka. To fit the data I used ADVAN4 TRANS4 (2 COMP
>> model).
>> Now the next step is to fit the PD data and here is where I am confused.
>> For the PK parameters I designate them under \$PK as shown in the code
>> provided. CL and V2 had individual values which were included in the data
>> set. Also in the PD data was set up to show zero and first order absorption
>> (2 dosing lines, RATE=0 or RATE=-2, COMP#)
>> Question -1- Does NONMEM understand that K0 values are to be calculated
>> from F2.DOSE/D2 or should I have to provide it in the data set ?
>> Question-2- For ALAG1 and ALAG2 again, does NONMEM understand how to use
>> them or I have to specify that under \$DES using IF (T.LE.ALAG1) etc.?
>> \$MODEL
>> NCOMP=5
>> COMP=(DEPOT,DEFDOSE)
>> COMP=(CENTRAL)
>> COMP=(TISSUE)
>> COMP=(RESPONSE)
>> COMP=(TISSUE)
>> \$PK
>> CL=CCL
>> V2=VV2
>> Q=0.01
>> V3=2.94
>> KA=0.004
>> K0=KK0
>> TLAG1=1.17
>> ALAG1=TLAG1
>> TLAG2=4.7
>> ALAG2=TLAG1+TLAG2
>> F1=0.7
>> F2=0.3
>> D2=108.96
>> S2=V2/1000
>> ;PD MODEL
>> E0= THETA(1)
>> KIN=(THETA(2))*(EXP(ETA(1)))
>> KILL=THETA(3)*(EXP(ETA(2)))
>> K45=THETA(4)
>> K54=THETA(5)
>> KOUT=KIN/E0
>> A_0(4)=E0
>> A_0(5)=K45*E0/K54
>>
>> \$DES
>> CONC=A(2)/V2
>> S=CONC
>> DADT(4)= KIN- KOUT*A(4)-(KILL)*S*A(4)-K45*A(4)+K54*A(5)
>>
>> \$ERROR
>> EFF=LOG(A(4))
>> IPRED=EFF
>> Y=EFF+ERR(1)
>> Regards
>> Neil
>>
>> --
>> Indranil Bhattacharya
>>
>>

--
Indranil Bhattacharya

Received on Wed Feb 10 2010 - 09:50:40 EST

The NONMEM Users Network is maintained by ICON plc. Requests to subscribe to the network should be sent to: nmusers-request@iconplc.com.

Once subscribed, you may contribute to the discussion by emailing: nmusers@globomaxnm.com.