User Manual Part 4

B-20
CLINICAL STUDY - COMPANION
DATA ANALYSIS AND RESULTS FOR PRIMARY ENDPOINT AND
SECONDARY ALL-CAUSE MORTALITY ENDPOINT
Sequential M onitoring
The COMPANION DSMB me t app rox ima tely every six months to review the
trial’s progress and to review the safety and effectiveness data collected. An
“O’Brien-Fleming” type boundary as implemented by Lan and DeMets was
used in monitoring the trial. The Group sequential procedure ensured that
the total alpha spent across repeated analyses did not exceed the total type I
error, in this case a =0.03.
On November 18, 2002 the DSMB reviewed the study progress for the final
time. The CRT-D arm of the Study had reached the target number of events
for both the combined mortality and hospitalization endpoint as well as the
all-cause mortality endpoint prompting the DSMB to recomm e nd to the Steering
Committee that enrollment be stopped. All effectiveness follow-ups ended
on December 1, 2002.
Results
Primary Endpoint: All-cause Mortality or First Hospitalization
The Kaplan-Me ier curves illustrate the time to all-cau s e mortality or first
hospitalization (Figure B-4 on page B-21). There were 216 primary endpoint
events observed in the OPT arm and 390 in the CRT-D arm (p = 0.010; p =
0.011 after adjustment for interim analyses). The median time to first event was
209 days in the OPT group and 269 days in the CRT-D group. The annual
event rates for OPT and CRT-D, respectively, were 68.0% and 55.9%, with a
hazard ratio of 0.80; 95% CI (0.68, 0.95). This result demonstrated that CRT-D
significantly reduced the relative risk of all-cause mortality or first hospit alization
by 20% when compared to OPT alone.
- DRAFT -