Back to the Top
The following message was posted to: PharmPK
Dear all,
I wish to conduct a two way, two period and two treatment
crossover bioequivalence study on a drug having
1) Long elimination half life (16 days)
2) High intrasubject variability (> 40% for different PK
variables)
If I dont want to conduct a pilot study, what should be the
last samplng point (At what point the study is to be
truncated?). In my knowledge , the study cannot be
truncated at 72 hr as the drug is having high intrasubject
variability.
What should be the wash out period for the same?
Pl throw a light on this issue.
Regards,
Dr Suhas Khandave
Back to the Top
The following message was posted to: PharmPK
Hi Dr. Khandave,
In general, you want to cover 3 half-lives of the drug so if
truncation is
not an option your last sampling time will be on day 48. Washout periods
can (debateably) range anywhere from 7-11 half-lives, giving you a
washout
of about, say, 11 weeks. Why is your half-life so long? Is it because of
the route of entry? Is this an IM injection? If so, 72 hour
truncation does
not apply as it does not cover the absorption period. It's too bad
the ISV
of your drug is so high, because otherwise you could consider a parallel
study. What is your country of submission? For some FDA studies, I have
seen the truncated AUC applied regardless of the high PK variability.
You
may wish to consult CDER for their recommendations if the study is
intended
for FDA submission, as they may recommend a truncated AUC despite
your high
ISV.
Hope this helps,
Dr. Dave Dubins
Back to the Top
The following message was posted to: PharmPK
Dear Dr Dave Dubins,
Thanks a lot for your prompt reply.
My formulation is oral solid and the sponsor does not wish
to conduct a parallel study. This is for USFDA submission.
The pharmacokinetic behavior of the drug under consideration
states that the absorption is very rapid and longer
elimination half life is due to its larger volume of
distribution.
As per your suggestion I would like to consult CDER .
Are there certain criteria to consider crossover BE studies
on such types of drugs where keeping subjects under
controlled conditions is practically difficult due to longer
duration of the study ?
Regards,
Dr Suhas Khandave
Back to the Top
The following message was posted to: PharmPK
Dear Dr Dave Dubins,
Thanks a lot for your prompt reply.
My formulation is oral solid and the sponsor does not wish
to conduct a parallel study. This is for USFDA submission.
The pharmacokinetic behavior of the drug under consideration
states that the absorption is very rapid and longer
elimination half life is due to its larger volume of
distribution.
As per your suggestion I would like to consult CDER .
Are there certain criteria to consider crossover BE studies
on such types of drugs where keeping subjects under
controlled conditions is practically difficult due to longer
duration of the study ?
Regards,
Dr Suhas Khandave
Back to the Top
Dr Suhas Khandave,
Dave's comments are correct. However I have to wonder if your half-
life is really a terminal elimination half-life. There are often
measurable levels of drug in the body that 'sit' at very low
concentrations for extended periods of time (i.e. drugs leaching form
the bone etc...). This might be producing your very long half-life.
The FDA guidance states that if pre-dose values are less than 5% of
the respective Cmax, you can conduct your analysis without
manipulation of the data. If the concentration of this drug falls
below 5% of the maximum concentration in a reasonable period of time,
then you should conduct the study with that washout and anticipate
that there will be pre-dose levels in subsequent periods. Just
ensure that your washout will be enough time to ensure that only 5%
or less of the drug remains in the body. You should be able to tell
this from the data that you already have.
Robert Lepage, M.Sc. CCRP
Pharmacokineticist & Assistant Study Director
Pharma Medica Research Inc.
Back to the Top
The following message was posted to: PharmPK
Dear Dr Robert Lepage,
Thanks a lot for your fruitful comments.
I came across a problem of designing a protocol for the drug
for which the literature is scanty and one cannot draw a
conclusion whether the mentioned half life is really a
terminal elimination half life. As per the guidelines the
washout period should be sufficient to make the predose
conc. less than the 5 % of Cmax. If the half life ( not
terminal elimination half life ) of a drug is 16 days,
probably 64-80 days are required for a conc. to fall to 5 %
of the Cmax. We are finding it very dificult to maintain the
protocol complaince after discharge of the subjects from the
clinical facility if such a long washout period is kept. I
guess others also must be having these types of problems in
a crossover study. Parallel study design is a good solution
to this issue but again has its own limitations. I would
like to hear how others are handling this issue and
appreciate your expert comments.
Regards,
Dr Suhas Khandave
Back to the Top
The following message was posted to: PharmPK
Dear Dr Suhas Khandave,
Did you ever consider to use a semisimultaneous study design ? See
Karlsson & Bredberg, Pharm. Res. 6: 817 - 821 (1989) and J. Pharmacokin.
Biopharm. 18(4): 293 - 311 (1990).
Don't know how your sponsor and the FDA feel about this, but this design
reduces intraindividual variability, the number of samples and can also
save a lot of (study) time.
Regards Bert Meijer
Back to the Top
Hello,
Just a comment on your crossover design. The major statistical
premise of this design is the belief that the patient population
remains essentially the same between one period and the next. For
most washout periods this is reasonable to believe, but not for 64-80
days. It is easier to believe that each subject is different after
this long wait, and your statistical analysis may very well be
questioned by the regulatory authorities.
Much better to convince your client to spend a little more money on a
parallel design and thereby save time, than to take a chance on a
design that is inherently flawed.
However, if you should happen to get a reviewer that doesn't
appreciate this subtlety, you may get away with it - good luck.
Edmond B. Edwards, Ph.D.
EDIT Research
Ottawa, Ontario, Canada
Back to the Top
The following message was posted to: PharmPK
Edmond / Dr Suhas Khandave,
I agree that a parallel design sounds to be the more logical choice
initially. However, the number of subjects required for a parallel
study is based on the INTERsubject variability. Assuming the
INTERsubject variability is larger then INTRAsubject variability that
Dr. Suhas has reported, the number of required subjects would be very
large.
Based on a multiplicative model with 40% CV and 80% power - 134 in
each group or 268 total. This is assuming that the INTERsubject
variability is only 40% (not a very high INTERsubject variability).
I don't agree that subjects will be significantly different after a
couple of months. BE studies do not require the subject to be
exactly the same between two periods. With simple demographic
verification you can assume that subjects will be pretty much 'the
same' several months after starting the study. The chances of a
healthy subject absorbing and metabolizing a drug completely
different after a couple of months is slim.
I am assuming that the data you currently have is based on a
crossover type study (since you are reporting a intrasubject CV).
You must then know the intersubject variability? If your data is not
from a study and just gathered from the literature (scanty at best as
you mention) then you should definitely convince the sponsor that a
pilot study is best before jumping into a 200+ subject expensive
parallel study. You could learn a lot from a 18 or 24 person pilot.
The real question is, did you calculate the time that the
concentration would be less than 5% using a standard equation or are
you using actual data? Without actual data and concentration time
profiles, the limited literature is not very helpful.
Robert Lepage, M.Sc. CCRP
Pharmacokineticist & Assistant Study Director
Pharma Medica Research Inc.
Back to the Top
The following message was posted to: PharmPK
Dear Dr Robert Lepage,
Thanks a lot for sharing your knowledge and valuable time
dedicated to throw a light on this issue.
As you recommended I will try to convince my client to go
for a pilot study which is preferable on all grounds.
Regards & Thanks
Dr Suhas Khandave
--
The following message was posted to: PharmPK
Dear Meijer,
Sincere thanks for providing the reference.
Regards,
Dr Suhas
Back to the Top
The following message was posted to: PharmPK
The appropriate pivotal study design can be deviced by means of pilot
study
data. Considering all the elaborated factors by Mr. Lepage, affecting
the
study.
Dr.Manoj Bose, MD (Pharmacology)
Principal Investigator
Associate VP- Clinical Research
Vimta Labs Ltd
Back to the Top
Re the truncated BE study correspondence, Edmond Edwards wrote that a
washout period of 64-80 days is unacceptable but he gave no data to
justify this assertion. I can't see why we we should necessarily
assume that a gap of 10 weeks is so problematic and it raises several
questions I think.
What changes could there be that we cannot detect by questioning (eg:
diarrhoea) or medical examination? Why would the differences
neccesarily be larger than at 2 weeks? Why would they bias the
outcome if the randomisation has been done properly? They could just
as easily cancel out if they were to exist in such numbers. How
would the differences be greater than we see between subjects in an
average study or in every day life come to that?
The compound in question had a long half life so there is an
unusually wide potential for inter-subject variation which might well
justify a crossover design by reducing the numbers required. It
would also minimise the effect of any change that does occur due to
the 10 week gap.
Andrew Sutton
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)