Statistical Analysis.
For the primary endpoint analysis, the following inequality hypothesis
were tested: H0: Proportion of responders at 6 months with the MPP
feature activated at baseline ≤ 57%; H1: Proportion of responders at 6
months with the MPP feature activated at baseline > 57%.
The MPP responder rate was compared to 57%, which is the responder rate
obtained from literature13 for patients without the
MPP feature activated.
The sample size calculation was based on the primary endpoint,
proportion of CRT responders at 6 months who have the MPP feature
activated at baseline. Using a one-sided, one sample Exact Binomial Test
and a significance level of 2.5%, the study required 74 patients to
reach a power of 90%. With the MPP feature activated at baseline, it
was assumed that the proportion of responders at 6 months was 75%. In
order to detect the 18% difference between the two responder rates, a
total of 74 patients was needed to have analyzable endpoint data at 6
months. Assuming that, some patients would drop out of the study, the
sample size increased to 103 subjects.
Mean ± standard deviation was reported for continuous data and
frequencies and percentage for categorical data. The comparison of
numerical variables was performed using Student’s t test or Wilcoxon’s
rank-sum test, depending on the distribution of the variables. The Chi
squared test, Fisher’s exact test or Wald’s test was used to compare
qualitative variables as appropriate. Kaplan-Meier estimates with
log-rank tests and multivariable Cox proportional hazards models were
used to compare the QUARTO III study participants with the control
cohort (Quarto II patients) for patient mortality, hospitalization, and
the combined endpoint of mortality and/or hospitalization.