Clinical and echocardiographic outcomes of conduction system pacing: a bicentric retrospective analysis

Introduction Conduction System Pacing (CSP) has emerged as a physiological pacing alternative designed to preserve ventricular synchrony and reduce the risk of pacing-induced cardiomyopathy. The two main CSP modalities—Left Bundle Branch Pacing (LBBP) and Deep Septal Pacing (DSP)—differ in their anatomical targets and capture selectivity. However, comparative real-world data on their clinical and instrumental outcomes remain limited. Purpose To evaluate and compare the electrical, echocardiographic, and clinical outcomes of LBBP and DSP. Methods This bicentric, retrospective observational study included consecutive adult patients who underwent CSP between 2022 and 2025. Clinical, ECG, and echocardiographic data were collected at baseline and at ≥6 months of follow-up. Only patients with ≥80% ventricular pacing and complete documentation were analyzed. Procedural parameters, pacing thresholds, QRS characteristics, and echocardiographic measures of left ventricular function and mechanical dispersion were assessed. The primary endpoint was the change in left ventricular ejection fraction (LVEF) from baseline to follow-up. Secondary endpoints included electrical activation, ventricular synchrony, and pacing stability. Results A total of 124 CSP patients were analyzed (LBBP = 67, DSP = 57). Baseline characteristics, comorbidities, and LVEF were comparable between groups. The median follow-up duration was 9 months, with a ventricular pacing burden exceeding 90% in both groups. Procedural and fluoroscopy times were similar. Electrically, LBBP achieved more physiological activation with a narrower paced QRS duration (126 ± 18 ms vs 136 ± 24 ms; p = 0.03). QRS duration remained stable over time in both groups. LBBP showed greater systolic recovery, with higher follow-up LVEF (47.7 ± 12.1% vs 42.0 ± 12.4%; p = 0.01), greater ΔLVEF (+5.3 vs +0.9 pp), higher stroke volume (75 ± 20 vs 63 ± 17 ml; p = 0.04), and higher cardiac output (5.1 ± 1.1 vs 4.3 ± 0.9 L/min; p = 0.03). Mechanical synchrony also favored LBBP, with lower TDI s’–s’ dispersion (4CH: 49 vs 77 ms; p = 0.04; 3CH: 41 vs 66 ms; p < 0.05). Univariable analysis revealed a positive correlation between pacing type and follow-up LVEF, and in multivariable analysis, LBBP independently predicted greater LVEF improvement (+5 pp; p = 0.01). In subgroup analysis, LBBP significantly improved LVEF in both native LBBB and non-LBBB patients, whereas DSP did not. Device performance and complication rates were similar between groups. Conclusion In this study, LBBP provided more physiological activation, tighter mechanical synchrony, and greater systolic improvement than DSP, with comparable safety and electrical stability. DSP remains a reliable fallback option when LBBP is not feasible.