Comparative statistical analysis entailed calculating
the relative risks plus the 95% confidence
intervals for all dichotomous outcomes, the mean
differences plus 95% confidence intervals for normally
distributed, continuous outcomes (using
analysis of covariance where appropriate), and the
median differences plus 95% confidence intervals
for skewed continuous variables. In addition, ordered
categorical variables were examined with the
use of the chi-square test for trend. An adjusted
analysis of the primary outcome was performed
to investigate the effect of known prognostic factors.
All statistical tests were two-sided and were
not adjusted for multiple comparisons.