Back to the Top
Maria and Ladislav,
Thanks for bringing the CTDB package to my attention. I have not yet
had time to explore the CTDB system but I have a couple of comments
based on reading the introduction.
The introduction on http://www.uef.sav.sk/advanced.htm states "the
terms pharmacokinetics and pharmacodynamics had been created by using
the Greek words kinetics and dynamics" but then appears to ignore their
original Greek meaning in attempting to justify the term dynamics to
encompass both pharmacokinetic and pharmacodynamic processes.
"From my limited knowledge of Greek I would say that the word "kinetic"
comes from the Greek word "kinein" meaning "to move" while "dynamic"
comes from the Greek word "dynamis" meaning "power" or "force". E.g.
the moving meaning of kinetic is seen in the word cinematography where
"cine" refers to moving pictures while the force meaning of dynamic is
seen in dyne referring to units of force in the CGS system.
"
Given that a lot of pharmacokinetics can be thought of as describing
the movement of drug molecules into through and out of the body then
the kinetic suffix seems appropriate. Similarly pharmacodynamics
describes the responses to drug concentration, which at least in the
early days of pharmacology were observed by expressing forces generated
by organs, so that dynamics seems a suitable suffix for this class of
processes.
Perhaps you would like to offer an explanation for why you choose to
use the term dynamics to describe both pharmacokinetics and
pharmacodynamics when it seems to me that these terms have commonly
accepted and linguistically separate meanings?
Later on the same page you say "If functions, such as
poly-exponentials, are fitted to the [...] profiles without taking into
consideration the inputs of the drug into the body (which is a common
way in PK methods)...". This seems strange to me. I cannot recall the
use of closed form pharmacokinetic models involving sums of
exponentials that did not explicitly define an input process e.g.
bolus, zero-order, first-order, unless it was using some
semi-parametric method such as splines when the input process was
ill-defined a priori and not readily described by an explicit
parametric input process. Could you provide some justification for your
assertion that polyexponential models are commonly used without
consideration of the drug input?
Nick
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.-a-.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
Nick,
We tried to discuss this issue in the PharmPK in 1998. Unfortunately,
there was no interest in this issue at that time. At the beginning we
can
remind
two studies:
D. M. Fisher. (Almost) Everything you learned about pharmacokinetics was
(somewhat) wrong! Anesth. Analg. 1996, 83: 901-903.
J. M. van Rossum, J. E. de Bie, G. van Lingen, H. W. Teeuwen.
Pharmacokinetics
from a dynamic systems point of view. J. Pharmacokin. Biopharm.
1989, 17: 365-392.
As follows from the title of the first work, there are problems in the
field
of pharmacokinetics. The title of the second study indicates that the
term
"dynamic system" is not new in the given field.
> Given that a lot of pharmacokinetics can be thought of as describing
> the movement of drug molecules into through and out of the body then
> the kinetic suffix seems appropriate.
In practice, such a movement cannot be directly measured. This is the
main
problem if the conception based on the movement of drug molecules is
used
as the basis of pharmacokinetics. We only know what is the cause of
such
a movement, i.e. a drug input into the body, and what is an outcome,
i.e. a resultant drug concentration-time profile in blood.
The relationship between the given cause and outcome is determined by
the drug behavior in the body. The drug behavior in the body is a
dynamic
process that develops over a course of time and that depends on:
the actual drug input into the body, "the body history" (e.g. on
preceding
drug inputs into the body), and on the actual physiological or
pathological
state of the body. It follows then that, time is an essential
independent
variable of the model of the drug behavior in the body. This is what
the adjective "dynamic" in the term "dynamic process" refers to.
The analogous understanding of the term "dynamic process" is
common also in other scientific fields (1). A drug effect, a drug
dissolution,
bioavailability of a drug, e.t.c., also are dynamic processes.
Based on this fact, the drug behavior, drug effect, drug dissolution,
drug bioavailability, and many other dynamic processes can be modeled
in a conceptually, methodologically, and computationally unified way,
employing methods of the dynamic-system theory. Methods for such
modeling are implemented in the package CTDB. Using this package,
optimal structures and point estimates of parameters of models of
diverse
dynamic processes can be determined using a non-iterative method,
i.e. the method which does not require initial estimates of model
parameters.
Thereafter, the models can be refined and interval estimates of
parameters
can be determined, employing iterative procedures. The use of the
non-iterative
method in the package CTDB markedly simplifies and speeds up modeling
procedures.
> I cannot recall the
> use of closed form pharmacokinetic models involving sums of
> exponentials that did not explicitly define an input process
In practice, polyexponential functions (which are solutions of linear
compartment
models for single instantaneous drug inputs into the body, obtained
after
adopting well known simplifying assumptions about the drug behavior in
the
body)
are very frequently fitted to measured drug concentration data in blood.
These procedures yield models of data, because resulting model functions
are fundamentally restricted to the given drug inputs. Additionally,
they
do
not explicitly employ the given inputs in fitting procedures. In
contrast to
this,
models obtained by the package CTBD are mathematical structures,
i.e. differential equations, that after introducing mathematically
described
diverse drug inputs into the body originate profiles describing
resultant
respective concentration-time profiles of drugs in blood.
The use of the models obtained by the package CTBD can be exemplified
as follows: As seen in the figure given at
http://www.uef.sav.sk/Slide1.JPG,
the model of the dynamic system that represented the behavior of FVIII
in
the body was determined using a test-dose of FVIII given in a short-time
infusion.
Thereafter, this model was used to simulate the concentration-time
profile
of FVIII resulting from multiple-bolus doses of FVIII given at
non-equidistantly spaced times. The figure at
http://www.uef.sav.sk/Slide2.JPG
shows that the given model can be also utilized in an "opposite" way,
i.e. to determine a FVIII infusion, necessary to reach and maintain
required levels of FVIII.
1. Ljung, L. System identification - theory for the user. 2nd ed.
PTR Prentice Hall: Upper Saddle River 1999.
Regards,
Maria Durisova
and
Ladislav Dedik
Back to the Top
Maria, Ladislav,
I have read the editorial by Denis Fisher (Fisher DM. (Almost)
everything you learned about pharmacokinetics was (somewhat) wrong!
Anesth Analg 1996;83(5):901-3.). But it sheds no light on why one
should prefer the term dynamics over kinetics. I am aware that others
like to speak of dynamic systems in the context of pharmacokinetics but
this terminology is used by only a minority of authors.
The reason for my comments on your web page was to try to get you to
provide some convincing reason for changing the name from kinetics to
dynamics. I am afraid your response does not convince me that there is
any reason to change the name.
Whatever you call it I think that you and I generally use very similar
models. Can we agree that a rose by any other name would smell as
sweet? I'll stick to pharmacokinetics and pharmacodynamics and you can
lump everything into system dynamics.
Nick
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.aaa.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
Nick,
"
The reason for my comments on your web page was to try to get you to
provide some convincing reason for changing the name from kinetics to
dynamics. I am afraid your response does not convince me that there is
any reason to change the name.
"
Despite this, linear deterministic compartment models commonly
used in the methods called PK methods, i.e. sets of ordinary
differential
equations without time delays are also dynamic models, as far as
the mathematical nature of these models is concerned (1).
"
Whatever you call it I think that you and I generally use very similar
models.
"
Yes, to some extent. All the linear deterministic models can be written
written in the form used in the CTDB package, as exemplified in
study (2) for a two-compartment linear deterministic model. However,
the CTDB package enables to determine many more complex
models than are the linear deterministic compartment models,
e.g. models containing time-delays (2), or shunts (3). See some examples
of models determined by the package CTDB
at http://www.uef.sav.sk/examples.doc .
1. B. Hannon, M. Ruth, Dynamic Modeling, Springer-Verlag, New
York, 1994, 248 pp.
2. Durisova, M., Dedik, L., Batorova, A., Sakalova, A., Hedera, J.
Pharmacokinetics of factor VIII in hemophilia A patients assessed
by frequency response method. Methods Find Exp Clin Pharmacol
1998, 20: 217-26.
3. M. Durisova, Dedik, L., Balan, M. Building a structured model
of a complex pharmacokinetic system with time delays.
Bull Math Biol 1995; 57: 787-808.
4. L. Dedik, M. Durisova, Frequency response method used
in modelling environmental systems - a working example.
Ecol Modelling 1997, 101: 175-184.
Regards,
Maria Durisova
and
Ladislav Dedik
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)