Back to the Top
Dear all
I would appreciate any help regarding the consequences of finding a
sequence effect in the ANOVA tests (subject within sequence term)
when performing a bioequivalence study (conventional single dose, 2
treatment, 2 period, 2 sequence cross-over, balanced design with 24
subjects for Captopril).
My impression is that the conclusion of bioequivalence can not be
ruled out, provided one gets the 90%CIs within the proposed limits
(in fact I got ratios like 100.7% (CI 90.7-111.8%) for Cmax and
101.2% (CI 93.9-109.1) for AUC.)
Just as an additional information, figures are almost the same when
taking the mean ratio and 90%CIs of Period 2 against Period 1.
Sequence t-r against sequence r-t shows, however, a ratio around 80%.
ANOVA
Effect SS df MS f Probability
Ln(Cmax) Seq 0.632 1 0.631992 14.33 0.0010
Ln(Cmax) Subject(Seq) 2.537 22 0.115321 2.62 0.0143
Ln(Cmax) Period 0.0097 1 0.009699 0.22 0.6437
Ln(Cmax) FormName 0.0006 1 0.000634 0.01 0.9056
Ln(Cmax) Error 0.9701 22 0.044100
Ln(AUC) Seq 0.3116 1 0.311565 13.67 0.0013
Ln(AUC) Subject(Seq) 2.2398 22 0.101809 4.47 0.0004
Ln(AUC) Period 0.0017 1 0.001677 0.07 0.7887
Ln(AUC) FormName 0.0018 1 0.001789 0.08 0.7819
Ln(AUC) Error 0.5015 22 0.022700
Thanks in advance
Jaime Ilha
Cartesius Analytical Unit
ilha.at.lexxa.com.br
Back to the Top
Dear Jaime;
In a situation such as yours, where there are sequence effects but no
associated period effects (and the CIs are acceptable), it generally is okay
to claim bioequivalence. Sequence effects by themselves are usually not
worrisome in that they reflect subject to subject variation. I'm assuming
the sequence groups had no designed differences (such as male vs female,
etc.). On the other hand, if you experience significant sequence effects in
conjunction with significant period effects, that is suggestive of a
treatment by period interaction which is more problematic.
Dan
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)