Long-term prognostic value of left ventricular echocardiographic parameters for atrial fibrillation recurrence following pulmonary vein isolation
A Castro Pinto, B Lage Garcia, E Mata, F Castro, L Pinheiro, M Castro, D Goncalves, J Gomes, L Calvo, S Ribeiro, F Cordeiro, O Azevedo, A LourencoAbstract
Background
Left ventricular (LV) remodeling and systolic performance are key determinants of cardiovascular outcomes. Conventional echocardiographic parameters such as ejection fraction (EF) are widely used, but their value in predicting atrial fibrillation (AF) recurrence after pulmonary vein isolation (PVI) remains uncertain.
Purpose
To evaluate the predictive performance of pre-procedural LV echocardiographic parameters for AF recurrence at 2 and 3 years following first-time PVI.
Methods
This retrospective single-center study included consecutive patients with AF who underwent initial PVI between 2015 and 2024 and had pre-procedural echocardiography. AF recurrence was confirmed by 12-lead ECG or Holter monitoring. Time-dependent ROC analyses using the nearest-neighbor method assessed the discriminatory ability of LV ejection fraction (LVEF), global longitudinal strain (GLS), indexed LV mass, and LV end-diastolic volume for predicting recurrence at 2 and 3 years. Areas under the curve (AUC) with bootstrapped 95% confidence intervals (CI), Youden indices, and optimal cut-offs were calculated.
Results
Seventy-two patients were analyzed. Median time from echocardiography to ablation was 312 days (IQR 120–610), and median follow-up was 335.5 days (IQR 158.5–765.2). Annualized AF recurrence, from Kaplan-Meier analysis, was 19.9% per patient-year. At 2 years, LVEF showed an AUC of 0.522 [0.406–0.629] (optimal cut-off ≥59%, 71.9% with sensitivity and 36.1% specificity). LV end-diastolic volume (AUC=0.432), LV mass (AUC=0.387) and GLS (AUC=0.473) demonstrated poor separation. Similar trend was observed at 3 years, LVEF achieved the highest discrimination with an AUC of 0.563 [0.422–0.756] (optimal cut-off ≥53% with 92.4% sensitivity and 26.9% specificity). The performance of LV mass was slightly higher (AUC = 0.534 [0.390–0.696]), whereas LV end-diastolic volume and GLS again failed to demonstrate meaningful separation.
Conclusion
In this cohort, both LV indices demonstrated limited prognostic accuracy for AF recurrence after PVI. LVEF showed a modest trend toward higher discrimination, but overall predictive capacity was poor. These findings underscore that recurrence after ablation is likely influenced by multifactorial mechanisms beyond isolated LV remodeling. The small sample size, variable timing of echocardiography, and potential selection bias limit generalizability. Larger prospective studies integrating LV mechanics, atrial parameters, and myocardial tissue characterization are warranted to refine recurrence risk stratification.For image description, please refer to the figure legend and surrounding text.