7.2.2.3.6.4
CGI-S
The
LOCF analyses showed p values of 0.05 or less favoring drug over placebo at
three and six through eight weeks for 100 mg/d and at no other times or
dosages. The OC analysis showed p
values of less than 0.05 favoring drug over placebo at three weeks for 100 and
300 mg/d, at seven weeks for 300 mg/d and at no other times or dosages. Had the sponsor corrected their analysis for
multiple-comparisons, no significant differences would have been demonstrated at
any time point.
7.2.2.3.6.5
CGI-I
The
LOCF analyses showed p values less than 0.05 favoring drug over placebo at
seven and eight weeks for 100 mg/d and at six and seven week for 200 mg/d and
at no times for higher dosages. The OC analysis
showed p values of less than 0.05 favoring drug over placebo at no time for 100
mg/d, at seven weeks for 200 mg/d, at three, six, and seven weeks for 300 mg/d,
and at seven weeks for 500 mg/d. Had
the sponsor corrected their analysis for multiple-comparisons, no significant
differences would have been demonstrated at any time point. Sixty of 112 (54%) 100 mg/d patients, 57 of
114 (50%) 200 mg/d patients, 49 of 110 (45%) 300 mg/d patients, 49 of 111 (44%)
400 mg/d patients, and 48 of 116 (41%) placebo patients were considered
treatment responders at time of discontinuation. Pearson chi-square analysis comparing the proportion of
responders across the treatment groups showed no significant differences in response
rates between the groups (X2=4.44, df=4, p=0.35).
7.2.2.4
Conclusions
The
parametric analyses show inconsistent results across outcome measures, no clear
evidence of an expected dose-response relationship, nor a consistency between
LOCF and OC analyses. Had there been
statistical correction for multiple-time-point or multiple-dose testing, then
the data in favor of efficacy of any dose of bupropion sustained-release tested
would be even weaker. The
non-parametric analyses are not consistent across outcome measures, do not
provide clear evidence of a dose-response relationship, and do not provide
statistically significant evidence of the effect of bupropion
sustained-release. On balance, although
no specific standard of efficacy was established in the design of this study,
the data from protocol 205 fail to show convincing evidence of the efficacy of
any of the doses of bupropion sustained-release that were tested.
7.2.3
Other Studies
Protocol
208 was an open multi-center trial of bupropion sustained-release in 3100
patients with the diagnosis of depression.
For eight weeks patients were treated with the higher dose tolerable up
to 300 mg/day. Beyond that time point,
patients who choose to remain in the study were permitted to remain at any
clinically appropriated dose. Clinical
response was assessed by the CGI-S and the CGI-I, both of which demonstrated
statistically significant improvement over baseline. Being that this was an uncontrolled study, these results cannot
adequately demonstrate the efficacy of bupropion sustained-release.
7.3
Summary of Data Pertinent to Important Clinical Issues
7.3.1
Predictors of Response
The sponsor did not report any particular predictors of response.
7.3.2
Size of Treatment Effect
The
relative effect of bupropion sustained-release as compared with placebo was
slight. In both protocols 203 and 205
the mean decrease from baseline in 17-item HAMD was 10 for bupropion
sustained-release and 8 for placebo at endpoint. The baseline score of study 203 was 23 and in study 205 the
baseline was 24.
Bupropion
Sustained-Release Clinical Review 23
Back a Page
Next Page
Back to Wellbutrin SRŪ NDA Index Page
Back to Main
Index