Back to the Top
Dear All,
I’m trying to model local (at the injection site) and systemic (plasma) drug concentrations
simultaneously using a two parallel absorptions model (two first order) from local compartment to
the systemic. Local drug levels decline in biphasic pattern.
I came across the topic: fast and slow absorption in user archive
http://www.cognigencorp.com/nonmem/nm/98nov152006.html
They suggested in this discussion to use 2 Depot compartments (local) linked by first-order rate
constants to the Central compartment (Compartment 3).
It is assumed that DEPOT1 has availability F1 and that DEPOT2 has availability F2=1-F1
$SUBR ADVAN6
$MODEL
COMP (FAST)
COMP (SLOW)
COMP (CENTRAL)
$PK
F1=THETA(1)
F2=1-THETA(1)
etc...
$DES
DFAST=A(1)
DSLOW=A(2)
DCP=A(3)/V3
DADT(1)= -KA1*DFAST
DADT(2)= -KA2*DSLOW
DADT(3)= IN - DCP*CL
IN= KA1*DFAST + KA2*DSLOW
DADT(3)= IN - DCP*CL
The problem in my case that I want to fit the local concentrations also (remaining amount in the
depot compartment), the model suggested has two depot compartments while I measure the local
concentration in only one compartment. Do you have an idea how I can modify this model to fit local
and systemic concentrations simultaneously?
Thank you in advance for any suggestion.
Fady Thomas
Back to the Top
This question actually belongs to the nmusers group list
nmusers
You can use
Y = (A(1)+A(2))/VDEP + error part
Leonid
Back to the Top
Hi Fady, if you want to try a much easier way of building your model I would suggest you try Phoenix
with it’s large library of built-in models that can be further modified graphically or in text mode.
At all the times the code used is visible so it’s very easy to learn the syntax.
Here you see how to add your second observation at a compartment of choosing, e.g. the dosing
site;
(I realise those on the PharmPK list won’t get the inline images but you can find many other
examples of using the graphical model builder on the Phoenix forums e.g.
https://support.certara.com/forums/topic/825-2-compartment-extravascular-parent-metabolite-modeling/
?p=3389 )
The full model text being in the tab circled in red to review at anytime, you may well find this a
more intuitive way to develop your models. We also have free licenses for teaching as well as
attractive academic research licensing.
test(){
# some example code easily annotated
deriv(A1 = - (Cl * C) + (Aa * Ka)- (Cl2 * (C - C2)))
urinecpt(A0 = (Cl * C)) # indenting and line spaces permitted
deriv(Aa = - (Aa * Ka))
deriv(A2 = (Cl2 * (C - C2)))
C = A1 / V
dosepoint(Aa) # flexible dosing statements with optiosn of tlag, rate, bioavail, duration etc
C2 = A2 / V2
error(CEps = 1)
observe(CObs = C + CEps) # can be continuous, categorical, occasion etc.
error(DepotCEps = 1)
observe(DepotObs = Aa * (1 + DepotCEps))
stparm(V = tvV) # with NLME license can also assess randome effects
stparm(Cl = tvCl)
stparm(Ka = tvKa)
stparm(V2 = tvV2)
stparm(Cl2 = tvCl2)
fixef(tvV = c(, 1, ))
fixef(tvCl = c(, 1, ))
fixef(tvKa = c(, 1, ))
fixef(tvV2 = c(, 1, ))
fixef(tvCl2 = c(, 1, ))
}
Best regards,
Simon.
Get tips and discuss Phoenix with other users; www.certara.com/forums
--
Simon.Davis.aaa.certara.com
Senior Scientific Consultant
Pharsight- A Certara™ Company
Back to the Top
Hi Leonid,
Thanks for your reply, I tried to apply your suggestion but I’m not sure what should I put in the
CMT column in the dataset when I have an observation in the depot compartment (CMT=1 or 2). And how
should I identify VDEP in the $PK section (S1 or S2).
Thanks again.
Fady Thomas
Back to the Top
One rule that is very helpful is NEVER use F in the error block (like
Y=F +...). Then, you are free to use any notations. In your case, there
should be a column in the data set that would distinguish blood and
depot observations. It could be CMT, then you can use CMT=1 (or 2, it
really does no matter) for absorption and CMT=3 for central
$ERROR
CDEPOT = (A(1)+A(2))/VDEP (where VDEP=THETA()*EXP(ETA()) is defined in
the PK block)
CBLOOD=A(3)/VC (where VC=THETA()*EXP(ETA()) is defined in the PK block)
Y=1 ;(so that nonmem does not complain if CMT is not equal to 1 or 3)
IF(CMT.EQ.1) Y = CDEPOT*(1+EPS(1))...
IF(CMT.EQ.3) Y = CBLOOD*(1+EPS(2))...
VDEP and VC play role of Sx, so you should be careful with scaling. For
example, if you used S3=VC/1000 before, then you should do the same here
CBLOOD=A(3)/S3
if you like to scale VDEP similarly, you can define SDEP=VDEP/1000 and use
CDEPOT = (A(1)+A(2))/SDEP
Let me know if this is not clear
Leonid
| PharmPK Discussion List Archive Index page |
Copyright 1995-2014 David W. A. Bourne (david@boomer.org)