On the basis of the results of the SOAP study,3
which showed a rate of death of 43% among patients
receiving dopamine and a rate of 36% among
patients receiving norepinephrine, we estimated
that with 765 patients in each group, the study
would have 80% power to show a 15% relative difThe
New England Journal of Medicine
Downloaded from nejm.org on November 25, 2015. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
The new england journal o f medicine
782 n engl j med 362;9 nejm.org march 4, 2010
ference in the rate of death at 28 days, at a twosided
alpha level of 0.05.
Since the magnitude of the effect derived from
observational studies can be misleading, we opted
for a sequential trial design with two-sided alternatives20;
the trial design called for analyses to be
performed after inclusion of the first 50 and 100
patients, and then after inclusion of each additional
100 patients, and allowed for the discontinuation
of the trial according to the following
predefined boundaries: superiority of norepinephrine
over dopamine, superiority of dopamine over
norepinephrine, or no difference between the two.
An independent statistician who is also a physician
monitored the efficacy analyses and the adverse
events; on October 6, 2007, after analysis of the
outcome in the first 1600 patients showed that
one of the three predefined boundaries had been
crossed, the statistician advised that the trial be
stopped.
All data were analyzed according to the intention-to-treat
principle. Differences in the primary
outcome were analyzed with the use of an unadjusted
chi-square test. Results are presented as
absolute and relative risks and 95% confidence
intervals. Kaplan–Meier curves for estimated survival
were compared with the use of a log-rank
test. A Cox proportional-hazards regression model
was used to evaluate the influence of potential
confounding factors on the outcome (factors were
selected if the P value in the univariate analysis
was