Baseline characteristics were summarized by counts and percentages for categorical variables and by means ± SD for continuous variables. Comparisons of continuous variables and quartiles of beta blocker doses were done using one‐way analysis of variance. Comparisons of categorical variables were performed using χ2 tests. The unadjusted relationships between beta blocker dose at baseline and patient characteristics and outcome endpoints were explored with Cox and linear regression models. For the three endpoints (cardiac death, all cause death, and arrhythmic events), predictive models were developed with the four variables of interest: heart rate (continuous variable), plasma norepinephrine (continuous variable, available in 744 beta blocker and 72 non‐beta blocker patients), beta blocker dose (expressed as carvedilol‐equivalent dose, continuous variable), and H/M ratio (continuous variable). Cox proportional hazards modelling was used to assess the relationship between outcomes and beta blocker dose as a continuous variable before and after adjustment for the variables found to be significantly associated with each endpoint. Kaplan–Meier survival analyses were performed to estimate 2‐year event probabilities, with differences between groups assessed using the log‐rank test.
A P value <0.05 was considered statistically significant for all comparisons. All analyses were performed using Medcalc v12.6‐13.0 (Medcalc Software, Ostend, Belgium).
A P value <0.05 was considered statistically significant for all comparisons. All analyses were performed using Medcalc v12.6‐13.0 (Medcalc Software, Ostend, Belgium).
