Back to the Top
Dear all
I would appreciate very much any advice regarding some considerations
on the design of a Bioequivalence study for Paroxetine.
We are planing to conduct a "replicate design study", since this drug
seems to be highly variable as we could find out from a previous
nonreplicate design study conducted in our Centre. From the ANOVA
analysis we found out a MS error of 0.210 for Cmax and 0.229 for
AUClast, which means a 45.8% and 47.8% within-subject variability,
respectively (from the MSE square root). Unfortunately, we did find
in the literature references concerning the between-subject
variability of Paroxetine, but nothing regarding the within-subject
variability. So, our data is the only information so far available
for us.
Therefore we are discussing with our staff on the advantages of
performing a replicate design study, as proposed by the FDA (Guidance
for Industry - Statistical Approaches to establishing Bioequivalence
- January 2001), in order to access the average bioequivalence.
However, although I agree with this approach, there are some concerns
in accepting it, based on the argument that it should only be used
when a high within-subject variability has been well established (and
we don't see that from the information in the literature).
As we can find on page 7 of the FDA Guidance - Bioavailability and
Bioequivalence Studies for Orally Administered Drug Products -
General Considerations - October 2000 - "Nonreplicate study designs
are recommended for BE studies of most orally administered,
immediate-release dosage forms. However, sponsors and/or applicants
have the option of using replicate designs for BE studies of these
drug products." ... "Replicate study designs are recommended for BE
studies of modified-release dosage forms and highly variable drug
products (within-subject coefficient of variation >= 30%), including
those that are immediate release, modified-release, and other orally
administered drug products." ... "Replicate study designs offer
several scientific advantages compared to nonreplicate designs. The
advantages of replicate study designs are that they (1) allow
comparisons of within-subject variances for the test and reference
products; (2) indicate whether a test product exhibits higher or
lower within-subject variability in the bioavailability measures when
compared to the reference product; (3) suggest whether a
subject-by-formulation (S*F) interaction may be present; (4) provide
more information about factors underlying formulation performance;
and (5) reduce the number of subjects needed in the BE study."
To my understanding from the above, even if the drug is not highly
variable, there are advantages in conducting a replicate design study
and these kind of studies should be welcomed. This seems to be a
cost-benefit decision, since doing so will not "ease" the conclusion
for bioequivalence, provided the sample size for the replicate and
"classical" approach is equivalent (i.e. greater number of volunteers
in the classical approach to obtain the same power).
Therefore, my question is: From the practical point of view, is the
FDA avoiding to accept replicate designs unless clearly defined
conditions are met ... or no restrictions are imposed to anyone
wishing to conduct such a study ?
Thank you in advance for your comments,
Jaime Ilha
Cartesius Analytical Unit
ilha.-a-.lexxa.com.br
Back to the Top
No restrictions are imposed to anyone wishing to conduct a replicate design
bioequivalence study. In fact, replicate designs are encouraged from a
scientific point of view.
Mei-Ling Chen, Ph.D.
FDA
Back to the Top
Dear Ilha,
Your understanding about the problem is correct and FDA is not against the
replicate designs and as you have mentioned it describes about the
advantages of replicate designs. Neverthelss, it would be better if you
get the prior approval of the protocol by FDA. There are two approaches
you can opt when doing replicate designs, average BE and Individual BE
criteria.
As the data indicates your drug qualifies for replicate design as
intrasubject CV% is greater than 30%.
thanks,
Ramakrishna Bangaru
Biopharmaceutics Research
Dr. Reddy's Laboratories, Generics
P.O.Box: 15, Kukatpally Post
Hyderabad-500 072, (India)
Back to the Top
Dear Mei-Ling:
Is there any way to see the correct statistic symbols used in the new FDA BE
guidance?
I tried both HTML and pdf files. I can not see the statistic symbols in the
documents.
Best regards,
Sam Liao, Ph.D.
PharMax Research
270 Kerry Lane,
Blue Bell, PA 19422
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)