Back to the Top
Dear all,
An article entitled "Limitations of Noncompartmental Pharmacokinetic
Analysis of Biotech Drugs" presented in the textbook of
'Pharmacokinetics and Pharmacodynamics of Biotech Drugs: Principles and
Case Studies in Drug Development' (edited by Dr. Bernd Meibohm,
WILEY-VCH, 2006) stressed that the calculation of Vss for protein drugs
from plasma concentrations using NCA will result in errors, since the
elimination may occur not only in the sampling compartment
(plasma/blood), but also in the tissue compartment. An elimination rate
constant k20 is thus introduced for representing the metabolism of a
protein drug in the tissue compartment in a two-compartment open model.
What I am confused about is the statement: 'For the case where tissue
elimination exists, it is possible to code into the model the existence
of a K20, but the convergence process will not be able to resolve the
appropriate micro rate constant.'
Does that mean there is no solution for the potential errors? I cannot
find the corresponding model in WinNonlin (version 5.2.1). Yet I am
wondering why the program cannot converge with the additional k20? If
the statement is true, should we avoid using the NCA when the drugs are
peptide or proteins?
Any comment is greatly appreciated.
Sincerely,
Ralf
--
Development Center for Biotechnology
101, Lane 169, Kangning St.,
Xizhi District 22180, New Taipei City
Taiwan, R.O.C.
Back to the Top
The following message was posted to: PharmPK
Ralf
I think the problem is not the drug (protein versus not protein) but the
model (linear versus nonlinear).. Biologics just happen to be nonlinear
much more often that small-molecules drugs (in the therapeutic range of
doses). For nonlinear drugs, one cannot use NCA-type analysis (that is
based on the linear-model assumptions).
The text that you mentioned probably referred to the fact that one
cannot separate elimination from the central and peripheral compartments
(at least, for drugs with linear kinetics).
The best way to get parameters of the drugs with the non-linear kinetics
is to use modeling.
Regards,
Leonid
--
Leonid Gibiansky, Ph.D.
President, QuantPharm LLC
web: www.quantpharm.com
e-mail: LGibiansky at quantpharm.com
Back to the Top
The following message was posted to: PharmPK
Ralf
I think the problem is not the drug (protein versus not protein) but the
model (linear versus nonlinear).. Biologics just happen to be nonlinear
much more often that small-molecules drugs (in the therapeutic range of
doses). For nonlinear drugs, one cannot use NCA-type analysis (that is
based on the linear-model assumptions).
The text that you mentioned probably referred to the fact that one
cannot separate elimination from the central and peripheral compartments
(at least, for drugs with linear kinetics).
The best way to get parameters of the drugs with the non-linear kinetics
is to use modeling.
Regards,
Leonid
--
Leonid Gibiansky, Ph.D.
President, QuantPharm LLC
web: www.quantpharm.com
e-mail: LGibiansky at quantpharm.com
Back to the Top
The following message was posted to: PharmPK
Dear Ralf,
If you delete clearance from the central compartment and allow clearance from peripheral
compartment, the 2 compartment model can be solved, but if you have a model with clearance
from both central and peripheral compartments, AND if you only have plasma concentration
values, there is an infinite number of solutions. If you could measure the "peripheral"
concentration, you could solve this, but usually this is not possible. Another approach is
to use a semi-physiological model. For this you need to know the volume of interstitial
fluid in the peripheral compartment; the flow rate or rate constant of lymphatic drainage
return of drug from ISF to plasma; and the rate of transfer of drug from plasma into the
ISF. Unfortunately, this last is also very difficult to measure.
Although NCA may be technically wrong in these circumstances, it still provides parameters
which can be used in various ways (eg interspecies scaling) for estimates of plasma
concentration. But it is a very poor starting point for estimating tissue concentrations.
As written, these very much depend on where degradation of a protein is taking place.
Interesting topic, but one which will be overtaken by PBPK models soon.
Best regards.
Ted
Back to the Top
Hi Ralf,
what Leonid and Ted described is generally called the structural
unidentifibility of a model.
WinNonlin doesn't have the 2-comp model, with elimination from both
central and peripheral compartment if you measure in plasma only, for
exactly this reason.
See for more details for example Bonate 2011
(Pharmacokinetic-Pharmacodynamic Modeling and Simulation) section
'Identifiability of Compartmental Models', p. 29-30. There is lots of
very technical literature about this issue (e.g. Godfrey, Compartmental
models and their application, 1983) but Bonate makes it easy
understandable using simple examples.
Estimating few parameters in a 2-comp model seems like a trivial job but
such unidentifible models create pitfalls in that you cannot trust an
estimated parameter because is not unique (and this independent from the
optimisation algorithm you are applying).
Cheers, Maciej
[See also http://www.boomer.org/c/p4/c22/c20/c2001.html
and subsequent pages - db]
|
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 "Limitations of NCA of biotech drugs" as the subject | Support PharmPK by using the |
Copyright 1995-2011 David W. A. Bourne (david@boomer.org)