Figure 3.

Subgroup analyses from individual patient‐level data for the primary outcome of death from any causes (A) and for the secondary outcome of re‐hospitalization (B) after a median follow‐up of 304 days (172 to 725). Odds ratio (OR) and corresponding confidence intervals for patients subgroups were pooled, and interactions were evaluated by random‐effects meta‐analyses. The vertical dashed line on plot represents the point estimate of OR = 1. Asterisk indicates no significant difference. ICD, implantable cardioverter defibrillator; CRT, cardiac resynchronization therapy; EF, ejection fraction; LVED, left ventricle end‐diastolic.