Back to the Top
The following message was posted to: PharmPK
Would someone point me to a FDA or EMEA guidance that indicates what
to do when planning a BEQ study and the expected sample size is
unreasonably large e.g. greater than 24 subjects?
Thanks,
Nick
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.-at-.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
Hi Nick,
In the "Guidance for Industry Bioavailability and Bioequivalence
Studies for Orally Administered Drug Products - General Considerations"
It is stated
"Nonreplicate study designs are recommended for BE studies of
immediate-release and modified-release dosage forms. However,
sponsors and/or applicants have the option of using replicate designs
for BE studies for these 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) provide more information about
the intrinsic factors underlying formulation performance, and (4)
reduce the number of subjects needed in the BE study. The recommended
method of analysis of nonreplicate or replicate studies to establish
BE is average bioequivalence, as discussed in section IV. General
recommendations for nonreplicate study designs are provided in
Attachment A. Recommendations for replicate study designs can be
found in the Guidance for Industry Statistical Approaches to
Establishing Bioequivalence (January 2001)."
If you check that Statistical Approaches guidance you will find that
one of the recommendations is replicate studies using treatment
sequences of RTRT and TRTR. This design halves the needed number of
subjects. Of course you have 3 washout periods. So all in all, you
may not save that much more money or time. But since you get repeat
exposure within a subject, it is good for high variable drugs.
Susan Shoaf, PhD.
Otsuka Pharmaceutical Development and Commercialization, Inc
Rockville MD
Back to the Top
Hi Dr. Holford,
If BEQ=bioequivalence, then following guidelines might be helpful
http://www.fda.gov/cder/guidance/5356fnl.htm
http://www.emea.europa.eu/pdfs/human/ewp/140198en.pdf
I dont think there are specific recommendations for higher number of
subjects. Could you provide more information about your concern?
Thanks
Ganesh
Back to the Top
The following message was posted to: PharmPK
Dear Nick!
Most regulators would consider 'unreasonably large study'... to be a
heretic statement
> Would someone point me to a FDA or EMEA guidance that indicates what
> to do when planning a BEQ study and the expected sample size is
> unreasonably large e.g. greater than 24 subjects?
See this pages on concerning HVDs (further links to FDA's and EMEA's
documents):
http://bebac.at/news/2006-10-06.htm
http://bebac.at/news/2006-04-27.htm
Both regulators are considering reference scaled average bioequivalence.
A scientific advisory meeting was planned by the FDA for May 2nd, 2007
but cancelled.
http://bebac.at/forum/mix_entry.php?id=592
EMEA's drafted guideline on HVDs/HVDPs is expected to be released for
consultation in Q2/3 2007.
IMHO New Zealand is the only country setting the maximum number of
subjects in a BE study to 40 (which by chance is the result of a sample
size estimation with: delta=0.05, CVintra=30%, power=80%).
best regards,
Helmut
--
Ing. Helmut Schuetz
BEBAC - Consultancy Services for
Bioequivalence and Bioavailability Studies
Neubaugasse 36/11
1070 Vienna, Austria
e-mail helmut.schuetz.at.bebac.at
web http://bebac.at
forum http://forum.bebac.at
[I think I agree with Nick on this. I find it difficult to justify
exposing > 24, 40 (for example) normal healthy subjects for a BEQ
study as may be suggested by a power analysis. There are many highly
metabolized drugs that exhibit considerable between subject
variability in clearance and therefore AUC etc. which might dictate
large study numbers. However, how variable are these drugs within
subjects? I would expect this variability to be less supporting fewer
subject for a cross-over study. Between - within aside large subject
number studies do become impractical as n increases - db]
Back to the Top
The following message was posted to: PharmPK
Dear all,
The sample size calculations for planning the pivotal bioequivalence
study on the basis of results from pilot bioequivalence study involves
two main factors, one is test by reference ratio (T/R ratio) and
another is intrasubject variability (There is one more factor, the
possibility of outliers, and effect of this on results can be explored
using appropriate simulation method). Although guidelines are not
mentioning the upper limit for sample size (as per my knowledge,
because there are examples where studies were conducted even with more
than 60 subjects in non-replicated design) it is the decision of the
sponsor whether to go for bigger study or not based on risk analysis.
Large sample size can be due to the following three possibilities
(provided you have precise estimates of variability and T/R ratio):
1. Less variability and large deviation of T/R ratio from 100%
2. Large variability and less deviation of T/R ratio from 100%
3. Large variability and large deviation of T/R ratio from 100%
If T/R ratio has considerable deviation from 100% (It doesn't matter
even if T/R ratio lies between acceptance region e.g. for US 80% to
125%) it is always better to improve the formulation because the
possibility of this formulation, coming out to be bioequivalent even
in large sample study is very less.
If variability is large and less deviation of T/R ratio is there, then
(large variability is associated with highly variable drugs) sponsor
can go for a replicated design or non-replicated design based on its
priorities of money and time (it is still better to explore the best
design, replicated or non-replicated, by using appropriate
simulations).
But there is one more situation which is more confusing and risky,
where we don't have precise estimate of variability and T/R ratio
(e.g. getting widely spread sample size estimates from previous
studies). In such a situation the sponsor is unsure whether the
formulations can be made bioequivalent in small sample or large
sample. It is wastage of human testing, time and money if we go for
large sample study when the formulation can be made bioequivalent in
smaller study and at the same time it is equally risky if we go for
small study when formulation was requiring larger study to be
bioequivalent. In such a situation FDA guideline "Statistical
Approaches to Establishing Bioequivalence-2001" suggest to design the
study as "group sequential cross-over" bioequivalence study. These
designs are well known in clinical trials but there are some examples
where such designs were used in bioequivalence studies as well.
Please rectify me where you feel there is a chance of improvement.
Best regards,
Sukhbir Singh
Statistician
Panacea Biotec Ltd.
India
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 "BEQ studies with large sample sizes" as the subject | Support PharmPK by using the |
Copyright 1995-2011 David W. A. Bourne (david@boomer.org)