Back to the Top
The following message was posted to: PharmPK
Dear All,
We have a BE study based on AUC[0-7days]. The sampling time is pre-
infusion, immediately following infusion, and 15 minutes, 2 hours, 4
hours, 8 hours, 24 hours, 48 hours, 120 hours (5 days) and 168 hours
(7 days) after the infusion. The study will be conducted in
outpatient patients, not healthy volunteers.
For PK sample on Days 5 and 7, it is difficult to draw the samples on
the exact nominal time (120 hours and 168 hours). Would it ok to
allow for some time deviations (the difference between the actual
time and the nominal time). If so, how much difference can normally
be allowed? Does the last blood draw at 168 hours have to be within 1
or 2 hours of the scheduled time?
Thanks,
CQ Deng, PhD
http://www.talecris.com
Back to the Top
The following message was posted to: PharmPK
Dear Dr Deng,
If you can record the actual time the sample was taken, you
should use this time for you analysis and any model building. As
long as you have accurate record of the actual time a sample was
taken it shouldn't really matter that it is not the Nominal
(scheduled time) since it will still be in the 'right' place in your
time-concentration profile. To compare an equivalent AUC[0-7days]
across subjects with differing actual sample times for the 168h
sample, you could use the predicted value that most pharmacokinetic
software will be able to calculate for you as a 'partial area' of
0-168h.
When you come to summarising data by nominal time e.g. for a
tables and mean curves then these deviations may be more pertinent;
in my experience a Statistical Analysis Plan stating the rules to
applied will be appended to, or included in, the protocol before the
trial commences. I would discuss the specifics of this with your
Stats group, however a typical rule might be that it is OK to
summarise all samples that were taken within 5% of the scheduled
time, and for samples within the first hour of dosing you might say a
deviation of upto 5 minutes is acceptable.
Applying the now standard caveat that I am not a statistician I
would be interested to know what other contributors would consider an
acceptable sample time deviation to be included in a summary, or
indeed in this case a simple comparison of AUClast.
Best regards,
Simon.
--
Simon Davis
Senior Scientific Consultant
Mobile : +44 7980 832 666
Facsimile : +1 801 991 7145
www.pharsight.com
Back to the Top
The following message was posted to: PharmPK
Dear Dr. Deng,
I agree with Simon in this case. Most of the facilities for PK
studies are
equipped with synchronized clocks with which we can measure the time
up to
last second. Software used for PK analysis can handle separate time
points
for each volunteer. Your approach in this case should involve taking
ACTUAL
timings of blood collection and analyzing the data.
(Though the term "actual sample time" is often debatable. Some people
prefer
the time of initiation of sample while some prefer completion of it.
With a
trained phlebotomist, it is a difference of 6-7 sec though! ) As far as
"actual times" are concerned, it's most advisable to clarify this
term in
your protocol and adhere to it.
Rhishikesh MANDKE
Back to the Top
The following message was posted to: PharmPK
Dear Simon,
Thank you for your response.
We are going to use the actual time (instead of the norminal time) in
AUC calculation. We will use the nominal time only for summary of
concentrations.
But I was questioned that AUC would be overestimated or
underestimated if a subject had his/her last blood sampling time
several hours behind the scheduled time or several hours before the
scheduled time. Therefore they would like to impose a very small time
deviation on the last sample time (168 +/- 1 hours), which is very
difficult to implement in the practice.
Are you saying that this will be adjusted when AUC is calculated (for
example using WIN NONLIN)?
Thanks,
CQ Deng, PhD
http://www.talecris.com
[As long as you extrapolate from the actual time for each subject you
should be OK. You could get the slope for extrapolation from curve
fitting all the data or semi-log regression with the terminal (log-
linear) points. Is there a guidance on this? Sounds like you should
at least describe your planned approach before starting - db]
Back to the Top
The following message was posted to: PharmPK
Dear Deng
In case of ambulatory samples, we allow one hr time deviation. Even
if it goes beyond one hr, we collect samples documenting the actual
reason. This time deviation is to be incorporated, while computation
of data.
Hope, this would serve your purpose.
Regards
Mitesh Gandhi
Back to the Top
The following message was posted to: PharmPK
Dear Deng
Yes it's a practical problem. It is difficult to collect samples at
exact times during ambulatory visits/ outpatients. In our studies we
collect samples within one hour of the actual time (of course not a
rule!) and use actual times in the calculation of pharmacokinetic
parameters. Such minor differences will not affect significantly the
final results of bioequivalence (AUC calculations).
regards
Vinay Ph D
Back to the Top
Hi CQ,
I think David has answered your question in general terms
already. I personally do not believe it would be necessary to
enforce such a tight deviation on these last samples since it is easy
to interpolate or extrapolate to a common end time.
To confirm, yes WinNonlin can account for these potentially differing
sample times; to do this look at the Partial Areas tab after you have
selected the appropriate NCA model, partial areas can be computed per
profile, by setting start and end times for each. The start and end
times need not appear in the data set, and can be after the last
observed time, if lz is estimable. If you need more information
about this please refer to the online help, WinNonlin manual or
contact support.-a-.pharsight.com; you may also wish to consider one of
our trainng courses:
http://www.pharsight.com/training/.
Best regards,
Simon.
--
Simon Davis
Senior Scientific Consultant
Facsimile : +1 801 991 7145
Home Office : +44 113 274 1198.
Back to the Top
there are no permissible sample time deviation limits. however, in
our lab, for walk-in samples, we consider the samples collected after
one hour of the scheduled time as having effect on AUC, rate of
elimination etc.
regds,
renu
Back to the Top
The following message was posted to: PharmPK
Dr. Renu
If the ambulatory sample falls in the terminal
elimination phase, the Kel remains the same at
whatever time it is collected. The AUC also will
remain the same. Only thing of concern is that whether
your protocol specifies any acceptable limits for time
deviation during ambulatory samples. The time
deviation should be incorporated in the PK analysis
accordingly.
Regards
V.Rajesh
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)