Back to the Top
How to do pharmacokinetics in a double-blind study?
Blood would need to be drawn from both subjects, receiving drug or
placebo.
However, it does not make sense to analyse the blood of the subjects
receiving placebo.
I assume it is possible to do the analysis by an unblinded laboratory,
i.e. only using the blood that actually contains drug.
But not using the blood samples of the subjects receiving placebo
(i.e. destroying it without analysis) could be considered unethical.
Does anybody have any experience with this issue or could advise on a
suitable design?
Kind regards,
Joerg Herbst, PhD, ERT, DABT
Director Pharmacology/Toxicology
APOGENIX GmbH
Im Neuenheimer Feld 584
69120 Heidelberg
Germany
Email: joerg.herbst.aaa.apogenix.com
Website: http://www.apogenix.com
Back to the Top
Dear Dr. Herbst,
Regarding analysis of PK samples from double blinded studies, two
common approaches are to either analyze all blood samples from both
active and placebo, or to analyze from all samples from active and
selected samples from placebo (e.g., predose and estimated Cmax). In
either case, the bioanalytical lab becomes unblinded, either by design
or by sample analysis. Therefore the protocol should state that the
bioanalytical lab will be unblinded and lab personnel are prohibited
from discussing actual subject identification with other study
personnel, unless specifically authorized to do so. The randomization
code may thus be provided to the lab, if the lab is a contract lab.
The sponsor contact may also be unblinded in a similar protocol-driven
and limited way and may specify appropriate dilutions for samples
(e.g., those expected to have concentrations above the calibrated
range of the standard curve. If any preliminary data are provided by
the lab during points in the study, then the lab assigns dummy codes
to subjects, so that all study personnel remain unblinded. All this
should be delineated in the study protocol.
I have been associated with such studies many times, either as a
blinded or unblinded bioanalytical point person. As long as the
details are clearly spelled out in the protocol and the protocol is
followed, it works fine. Pre-designation of placebo samples to
analyze, reduces cost when only a few placebo samples per subject are
analyzed. At least per and one post-dose placebo samples should be
analyzed as evidence of proper dosing and absence of assay artifacts.
Many labs feel it is procedurally simpler to analyze all samples than
to try to select. With such processes in place, all such studies I
know have been acceptable to regulatory authorities.
Tom
Thomas L. Tarnowski, Ph.D.
Bioanalytical Development
Elan Pharmaceuticals, Inc.
800 Gateway Boulevard
South San Francisco, CA 94080
thomas.tarnowski.at.elan.com
Back to the Top
Dear Jurg Herbst,
If the PK study is to be done at the end of the dosing period, ie.
after reaching steady-state, for both arms (placebo and drug 'X') you
can discuss with the CRO and the sponsor the option to open the
labeling at the end of the dosing period and to perform a PK study
only on drug 'X' treated patients.
The situation is much more complicated if the study design is
combination of a single dose PK study at the beginning of the dosing
period and ending up with a steady-state PK study after the dosing
period is completed. That kind of design is usually preferred by
sponsors for financial and time saving reasons.
But for strong ethical reasons, as you very correctly admit, you can
offer a slightly modified design beginning with an open single dose
PK study on randomly enrolled patients, after that to begin the true
placebo controlled trial, with or without, enrolling the patients
already assigned to the single dose open PK study. Into the ending up
steady state PK study you will get by chance patients who already have
been taking part in the starting single dose open PK study (in case
you enroll them in the true placebo controlled trial). Their actual
number will depend on the number of all-over assigned patients into
the trial. Every one statistician trained into probability and randon
numbers calculations can easily foresee that number.
For more in dept advices you have to provide the Forum with a little
bit much more info.
Greetings,
Dimiter Terziivanov
--
Dimiter Terziivanov, MD,PhD,DSc, Professor and
Head, Clinic of Clinical Pharmacology and Pharmacokinetics,
Univ Hosp "St. Ivan Rilski",
15 Acad. Ivan Geshov st, 1431 Sofia, Bulgaria
e-mail: dterziivanov.aaa.rilski.com; terziiv.at.yahoo.com
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 "PK in double-blind clinical studies" as the subject | Support PharmPK by using the |
Copyright 1995-2011 David W. A. Bourne (david@boomer.org)