Statistical analysis
The data are presented as a number (percentage) and mean ± standard deviation. Clinical characteristics between the study groups were compared by a t-test for continuous variables and chi-squared test for categorical variables. The first occurrence of RR-NSVT during the follow-up period was considered as a time-dependent covariate. For cluster RR-NSVT, the time interval until the occurrence of the third RR-NSVT in the first cluster episode was taken into account. The first appropriate ICD therapy was credited to the (cluster or frequent) RR-NSVT only if it occurred after the first episode to determine whether each form of RR-NSVT was associated with the following VTAs. Cox regression analyses were performed to identify associations with the occurrence of VTAs and were expressed as hazard ratios (HRs) and 95% confidence intervals (CIs). In the multivariate models, adjustment was performed for significant outcome confounders identified by stepwise selection from the baseline variables with the model entry significance limit set at 0.100. The cumulative incidence of VTA occurrence and admission for HF and cardiovascular mortality was estimated using the Kaplan-Meier method and compared using a log-rank test between groups. All statistical analyses were performed using statistical software (SPSS for Windows, version 22). A two-sided p-value ≤0.05 was considered significant.