Back to the Top
Hello,
Can someone kindly send across a WinNonlin user defined three
compartment model for an orally administered drug.
Would appreciate it.
Thank you
Anisha M
Back to the Top
Hi Anisha,
Please find attached first-order absorption, 3-compartment disposition
WinNonlin model.
(Usual disclaimers re. applicability and validity apply!)
Regards,
Charlie
[Not attached but as text below. You may need to cut-paste, save with
the correct extensions for this to work. Received as a defanged
attachment ;-) which seemed to be plain text - db]
--
model 20
remark three compartment model - first order input and output
remark defined in terms of a,b,c,alpha,beta,gamma,k01
rema
rema no. parameter constant secondary parm.
rema --- --------- -------- ---------------
rema 1 a dose (strip) alpha half life
rema 2 b # doses beta half life
rema 3 c dose 1 gamma half life
rema 4 k01 time 1, etc k01 half-life
rema 5 alpha
rema 6 beta
rema 7 gamma
rema*************************************************************
comm
nparm 7
nsec 4
pnames 'a', 'b', 'c' 'k01', 'alpha', 'beta' 'gamma'
snames 'alpha_hl', 'beta_hl', 'gamma_hl', 'k01_hl'
end
temp
dose1=con(1)
a1=p(1)/dose1
b1=p(2)/dose1
c1=p(3)/dose1
c2=-1*(a1+b1+c1)
end
func 1
j=2
ndose = con(2)
do i=1 to ndose
j=j+2
if x <= con(j) then goto red
endif
next
red:
ndose=i-1
sum=0
j=2
do i=1 to ndose
j=j+2
t=x - con(j)
d=con(j-1)
amt=a1*dexp(-alpha*t) + b1*dexp(-beta*t)+ c1*dexp(-gamma*t) +c2*dexp(-
k01*t)
sum=sum + amt*d
next
f=sum
end
seco
d=con(3)
d1=d*(a1*(k01-alpha) + b1*(k01-beta) + c1*(k01-gamma))
k01_hl=-dlog(.5)/k01
alpha_hl=-dlog(.5)/alpha
beta_hl=-dlog(.5)/beta
gamma_hl=-dlog(0.5)/gamma
end
eom
Model 21
remark three compartment model - first order input and output
remark defined in terms of a,b,c,alpha,beta,gamma,k01
remark includes a lag time
rema
rema no. parameter constant secondary parm.
rema --- --------- -------- ---------------
rema 1 a dose (strip) k01 half life
rema 2 b # doses alpha half life
rema 3 c dose 1 beta half life
rema 4 k01 time 1, etc gamma half-life
rema 5 alpha
rema 6 beta
rema 7 gamma
rema 8 tlag
rema*************************************************************
comm
nparm 8
nsec 4
pnames 'a', 'b', 'c', 'k01', 'alpha', 'beta', 'gamma', 'tlag'
snames 'k01_hl', &
'alpha_hl', 'beta_hl', 'gamma_hl'
end
temp
dose1=con(1)
a1=p(1)/dose1
b1=p(2)/dose1
c1=p(3)/dose1
c2=-(a1+b1+c1)
end
func1
j=2
ndose=con(2)
do i=1 to ndose
j=j+2
if x <= con(j) then goto red
endif
next
red:
ndose = i-1
sum=0
j=2
do i=1 to ndose
j=j+2
t=x - con(j) - tlag
d=con(j-1)
amt=a1*exp(-alpha*t) + b1*exp(-beta*t) + c1*exp(-gamma*t) + c2*exp(-
k01*t)
sum=sum + max(0,d*amt)
next
f=sum
end
seco
d=con(3)
d1=d*(a1*(k01-alpha) + b1*(k01-beta) + c1*(k01-gamma))
k01_hl=-dlog(.5)/k01
alpha_hl=-dlog(.5)/alpha
beta_hl=-dlog(.5)/beta
gamma_hl=-dlog(.5)/gamma
end
EOM
Back to the Top
The following message was posted to: PharmPK
Dear Anisha,
You wrote:
> Can someone kindly send across a WinNonlin user defined three
compartment
> model for an orally administered drug.
Excuse me for my critical notes, but I wonder why you need this. Do you
really have data that allow the estimation of 7 (or 8, in the case of
a lag
time) parameters from a single curve? This would require at least
15-20 data
points at appropriate time points. But more important, even in the
case that
you can identify these parameters with a reasonable precision, you
cannot
interpret the model. You end up with four exponential coefficients, but
there are always two equally valid solutions, with different values
for the
absorption rate constant (k01 in the code provided by Charlie
Brindley), but
exactly the same concentration - time profile. Usually the software
produces
only one solution, depending on the initial estimates or imposed
constraints
(e.g. k01 > alpha). So you may get one out of two solutions, and there
is no
way to get the right solution. So what is the purpose of such analysis?
best regards,
Hans Proost
Johannes H. Proost
Dept. of Pharmacokinetics and Drug Delivery
University Centre for Pharmacy
Antonius Deusinglaan 1
9713 AV Groningen, The Netherlands
Email: j.h.proost.aaa.rug.nl
Back to the Top
The following message was posted to: PharmPK
Hi Hans,
You are correct that there are different solutions for the model.
However,
if the 3-compartment disposition model is shown to be appropriate using
statistical criteria (AIC, etc) the function can be used to simulate
concentrations (e.g. after multiple dosing). The function can also be
used
as the characteristic response for deconvolution analysis.
Charlie
Back to the Top
Hello Hans,
I think that the data I have would fit a three compartment model
better than a two compartment model (I have a sufficient number of
time points 15 or so).
I know that it would be a more complicated approach, however it would
be a good try. Also as Charlie mentioned I would look out for the AIC
and other plots.
Thank you
Anisha
Back to the Top
The following message was posted to: PharmPK
Dear Anisha and Charles,
Thank you for your replies to my provocative comment. If you have
sufficient data and the 3-compartment model fits better according to
criteria like AIC, then you can use the solution for multiple dose
predictions and as a reference in deconvolution, as stated correctly
by Charles. If you do not interpret the data in a physiological way,
there is no problem. However, a presentation of the data denoting one
of the exponentials as k01 is actually an interpretation that is not
justified by the data.
best regards,
Hans Proost
Johannes H. Proost
Dept. of Pharmacokinetics and Drug Delivery
University Centre for Pharmacy
Antonius Deusinglaan 1
9713 AV Groningen, The Netherlands
Email: j.h.proost.at.rug.nl
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 "WinNonlin User defined three compartment model" as the subject | Support PharmPK by using the |
Copyright 1995-2011 David W. A. Bourne (david@boomer.org)