Back to the Top
Dear Kamala:
You asked "We have conducted the BE for a class 2 drug which has an
intra subject CV of 12 %. The study was a two way cross over study in
Fasting and Non-Fasting Conditions. In the study under non-fasting
conditions (30 subjects) we have one subject who is showing
significantly higher T/R ratio.The range for the T/R for the other 29
subjects is 75.29 - 192.64,
while for 1 subject it is 269.9. The FDA Guidance "Statistical
approaches to establishing BE", Section : Outlier consideration: Also
makes a reference to effect of significantly high T/R ratio. In this
situation is it possible to delete / disregard this particular subject
as an OUTLIER.
" Before you throw any subject or treatment you have to have clear cut
rationale based on scientific judgment. First I carefully look at the
individual subject profiles and compare with the other subjects, some
times one or two data points may cause havoc, then look at the medical
history during the study (emesis, dose contamination sample
contamination etc.), if none of these things work out you have to do an
outlier analysis, one procedure you can follow is called Normal
percentile plot technique, this plot will help you to single out a
treatment if it is indeed an outlier or not. After you single out a
treatment as an outlier you have to do a conjectural analysis and
develop a rationale for throwing that treatment out. Typically it is
better to drop the entire subject data than one treatment. Simple
answer is you can not throw any data unless you have a strong
rationale, the problem very well could be a bioanalytical error or
contamination of the plasma samples at some stage.
Hope this explanation helps,
Good luck,
Prasad NV Tata, Ph.D.
Mallinckrodt, Inc.
Saint Louis, MO 64134
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)