Back to the Top
Dear members,
I have a question about mouse plasma PK. I obtained mouse plasma
samples at 5min, 15min, 30min, 1hr, 2hr, 4hr and 6hr post-dosing. Due
to limited blood volume of mice, only two or three samples were
obtained from one mouse. However, the experiment was designed that at
least three plasma samples were obtained for each time point. My
question is, should I (1) calculate the average concentration at each
time point, then enter the average concentration into Winnonlin to
derive PK parameters, OR (2) enter all concentrations directly into
Winnonlin to derive PK parameters? Please give your suggestions. Thanks.
Best Regards,
Yulan QI
[I would think choice 2 or NONMEM - db]
Back to the Top
I thing that the best option for not lost information is use the
sparse analysis with all data.
Best regards
Benjamin Santos Lobo PhD
Head of Departament
Pharmacokinetics and Metabolism
R&D Center. Ferrer Internacional S.A.
08028 Juan de Sada 32 Barcelona
bsantos-research.-at-.ferrergrupo.com
www.ferrergrupo.com
Back to the Top
The following message was posted to: PharmPK
Dear Yulan,
You can find more hints on this webpage:
http://www.accp1.org/pharmacometrics/theory_popmeth.htm
Naive pooled data approach (NPD) is your method 2.
Naive averaged data approach (NAD) is your method 1.
The best choice is NONMEM. :)
Good luck
Guangli
Back to the Top
The following message was posted to: PharmPK
Dear Yulan,
WinNonlin NCA has a sparse sampling option that can be used to pool
sparse data for analysis. To access this option, you can check the
sparse sampling checkbox when selecting your NCA model. The model will
calculate PK parameters from the mean curve of the pooled data,
without requiring the user to create a mean curve separately. It will
also give the standard errors for CMax and AUC, based upon methods
described in Nedelman and Jia (1998), using a modifica\0xADtion in
Holder (2001).
Sincerely,
Christopher Mehl
Pharsight Software Support
Email: support.at.pharsight.com
Web: http://www.pharsight.com/support/support_sflogin.php
Back to the Top
The following message was posted to: PharmPK
Dear Dr QI
One of the new features in versions of WinNonlin since 5.0
(current version is 5.2) was support for Sparse Sample data such as you
describe. Note this is not a population analysis such as would be
provided by NONMEM but will provide some extra useful information
compared to a standard NCA.
When you apply one of the NCA models in WinNonlin note that
there is a check box "Sparse sampling" if you click this option on; then
it will ask you in the "Data variables" selections (the "XY" button); to
identify the subject variable by mapping it to the extra box
"Subject_ID". Multiple groups are still selected by using the sort keys
as usual.
If you choose this method then WinNonlin will calculate standard
error for the mean concentration curve's maximum value (Cmax), and for
the area under the mean concentration curve from dose time through the
final observed time.
Standard error of the mean Cmax will be calculated as the sample
standard deviation of the y-values at time Tmax divided by the square
root of the number of observations at Tmax.
Standard error of the mean AUC will be calculated as described in
Nedelman and Jia (1998), using a modification in Holder (2001), and will
account for any correlations in the data resulting from repeated
sampling of individual animals.
Nedelman and Jia (1998). An extension of Satterthwaite's approximation
applied to pharmacokinetics. J Biopharm Stat 8(2):317-28.
Holder (2001). Comments on Nedelman and Jia's extension of
Satterthwaite's approximation applied to pharmacokinetics. J Biopharm
Stat 11(1-2):75-9.
Best regards,
Simon.
--
Simon Davis
Senior Scientific Consultant
Back to the Top
Dear Yulan
The best option is to analyze the data through Non linear mixed effect
modeling.
Here you have an example conducted with NONMEM to explore the
valspodar effect on methadone PK.
It was also conducted in rodents (rat) and only 2-3 samples were
available per animal, very similar to your scheme.
Ortega I. Modeling methadone PK in rats in presence of P-glycoprotein
inhibitor valspodar. Pharm res 2007
Let me know if you have any problem to access to the reference.
As Benjamin suggested if mixed effect modeling is not possible I would
go for sparse analysis with all data
Back to the Top
The following message was posted to: PharmPK
Thanks to everyone for the contributions on this question. I had missed
the "sparse sampling" button on WinNonlin 5, and processed some data of
exactly this sort just yesterday by hand. However, I found a
difference. WinNonlin uses arithmetic means of the timepoints where I
had used geometric means. What do others think about this? Am I wrong
to be a compulsive geometric meaner? When I look at the distribution of
data, they look log-normal to me.
As for the other thread - I think I will have to bite the bullet and
enrol on a NONMEM course.
Best regards.
Ted
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 "Mouse pharmacokinetics" as the subject | Support PharmPK by using the |
Copyright 1995-2011 David W. A. Bourne (david@boomer.org)