Back to the Top
I would like to know how to model pk/pd data with only one effect
measurement (all patients have one effect at the same time), and if it
can
be done with an indirect response model.
Thanks
Proff. Rosario Calvo
Pharmacology Department
University of the Basque Country
Spain
Rosario Calvo Duo, PhD
Department of Pharmacology
Faculty of Medicine
University of the Basque
Leioa, Vizcaya 48940
Spain
http://www.ehu.es
Back to the Top
Professor Calvo - Anything can be modeled, but not all modeling results
are useful.
Taking the information that you have described (only one data
point/patient at one time) I cannot imagine much usefulness. The
indirect response models have a minimum of 3 parameters (Kin, Kout, and
C50 (or some other measure of drug action)). It would be impossible to
discriminate between these different model components without more
information.
Since you have a belief that there is an indirect response model
underlying the data. You may have sufficient information from other
studies/sources to identify the basic form of the indirect response
model and if you are able to obtain reasonable parameter estimates for
the model, then you should be able to assess the likelihood of your
data given the model. If for example the new data contradicts the
model (given the basic assumption that the model is adequate), that
would be an indication that there is something different about your
patients and you may be able to support changes to the model based on
the difference. Useful? Sounds like it could be.
Some indirect response model basics:
-If you have a baseline point you would at least be able to estimate
the ratio of Kin/Kout
-If your one point is at steady-state and you know Kin/Kout you may be
able to estimate some features of drug potency if you have a range of
doses or plasma concentrations (see Gabrielsson and Weiner's text for a
nice elaboration)
-Indirect response models are typically very difficult to fit without
data on multiple dose levels
-If you know enough to think it is an indirect response model, there is
a good chance that someone, somewhere has at sometime measured the
turnover kinetics of whatever it is you are measuring as a response
thereby giving pretty good information about Kin and Kout.
Regards,
Jeff Wald, PhD
Scientific Consultant
Pharsight Corp.
Cary, NC
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)