BackgroundLeft bundle branch pacing (LBBP) and left ventricular septal pacing (LVSP) are referred to as left bundle branch area pacing. ObjectiveThis study investigated whether long-term clinical outcomes differ in patients undergoing LBBP, LVSP, and biventricular pacing (BiVP) for cardiac resynchronization therapy (CRT). MethodsConsecutive patients with reduced left ventricular ejection fraction (LVEF <50%) undergoing CRT were prospectively enrolled if they underwent successful LBBP, LVSP, or BiVP. The primary composite end point was all-cause mortality or heart failure hospitalization. Secondary end points included all-cause mortality, heart failure hospitalization, and echocardiographic measures of reverse remodeling. ResultsA total of 259 patients (68 LBBP, 38 LVSP, and 153 BiVP) were observed for a mean duration of 28.8 ± 15.8 months. LBBP was associated with a significantly reduced risk of the primary end point by 78% compared with both BiVP (7.4% vs 41.2%; adjusted hazard ratio [aHR], 0.22 [0.08–0.57]; P = .002) and LVSP (7.4% vs 47.4%; aHR, 0.22 [0.08–0.63]; P = .004]. The adjusted risk of all-cause mortality was significantly higher in LVSP than in BiVP (31.6% vs 7.2%; aHR, 3.19 [1.38–7.39]; P = .007) but comparable between LBBP and BiVP (2.9% vs 7.2%; aHR, 0.33 [0.07–1.52], P = .155). Propensity score adjustment also obtained similar results. LBBP showed a higher rate of echocardiographic response (ΔLVEF ≥10%: 60.0% vs 36.2% vs 16.1%; P < .001) than BiVP or LVSP. ConclusionLBBP yielded long-term clinical outcomes superior to those of BiVP and LVSP. The role of LVSP for CRT needs to be reevaluated because of its high mortality risk.