Table 3.

Adjusted HRs for Ventricular Tachycardia and Sudden Cardiac Arrest, Using an Intent-to-Treat Analytical Approach

Group	Comparison	HR (95% CI)
Ventricular tachycardia
HFrEF	HFrEF vs no HF	7.01 (6.11-8.04)
HFpEF	HFpEF vs no HF	1.98 (1.69-2.34)
Taking D or S	Taking D or S vs not taking D or S	2.51 (1.96-3.21)
HFrEF taking D or S	HFrEF taking D or S vs HFrEF not taking D or S	1.35 (0.96-1.90)
HFpEF taking D or S	HFpEF taking D or S vs HFpEF not taking D or S	2.11 (1.08-4.12)
Sudden cardiac arrest
HFrEF	HFrEF vs no HF	5.13 (4.05-6.50)
HFpEF	HFpEF vs no HF	2.54 (1.99-3.25)
Taking D or S	Taking D or S vs not taking D or S	1.24 (0.64-2.40)
HFrEF taking D or S	HFrEF taking D or S vs HFrEF not taking D or S	2.06 (0.76-5.61)
HFpEF taking D or S	HFpEF taking D or S vs HFpEF not taking D or S	0.72 (0.28-1.82)

Abbreviations as in Tables 1 and 2.

The multinomial propensity score method was used to select potential covariates (listed in Supplemental Tables 7 and 8), and any characteristics with SMDs larger than 0.1 (listed in Supplemental Tables 7 and 8) were incorporated into the final model for double adjustment. The inverse probability of treatment weighting was used to stabilize the covariates-induced differences and survival analysis for the recurrent events model was performed. HFrEF, HFpEF, and taking D or S are the main effects, and HFrEF taking D or S and HFpEF taking D or S are the interaction terms in the model.