Prediction of Pulmonary vein isolation responders using morphology-based ECG criteria

Background Pulmonary vein isolation (PVI) is the cornerstone of invasive treatment for atrial fibrillation (AF), yet recurrence remains common. Identifying patients who respond to a PVI-only strategy versus those with atrial fibrosis or extra-pulmonary vein triggers remains clinically challenging. We recently proposed four ECG criteria predicting atrial fibrosis defined as low-voltage areas. Purpose This study evaluates whether these ECG criteria can be used to predict PVI responders. Methods We retrospectively analyzed 12-lead ECGs of 313 patients who underwent first-time ablation by PVI. A PVI responder was defined as absence of atrial arrhythmia recurrence after a PVI with allowance of one redo ablation for a PV reconnection only, with a blanking period of 60 days. Follow-up was based on standard clinical practice with additional rhythm documentation as clinically indicated. Four ECG criteria were evaluated: (1) history of atypical atrial flutter, (2) negative atrial vector attenuation in V1, (3) presence of typical or atypical interatrial block (types I–IV), and (4) poor terminal P-wave progression in V4-V6 (figure 1). Performance was assessed using sensitivity, specificity, positive predictive value, negative predictive value and accuracy. Uni- and multivariable survival analyses at 2 years were performed and compared with established clinical risk scores. Results Of the 313 patients, 78 patients had AF recurrence within 2 years after PVI with isolated PVs on redo ablation or in which a conservative non-invasive strategy was chosen. The proposed ECG criteria showed good diagnostic performance optimal with a general accuracy of 84.7%, with a high specificity (93%) and negative predictive value (88%). Univariable survival analysis showed a significant difference between patients with negative ECG criteria versus those with positive criteria (2-year AF-free survival 87.6% vs 26.6%; p<0.001). The proposed ECG criteria outperformed the current division in paroxysmal and persistent AF (82.8% vs 60.9%) and the DR-FLASH score (83.8% vs 67.3%) (figure 2). Multivariable analysis, correcting for AF characteristics, comorbidities and inter-operator differences, confirmed the ECG-criteria as a strong independent risk factor for AF recurrence after PVI (HR 10.41, CI 5.67-19.11; p<0.001). Conclusions A combination of 4 ECG-based criteria for atrial fibrosis may be used to predict a PVI-only response. These non-invasive markers may improve identification of patients who are less likely to benefit from a PVI-only strategy. Further prospective multicentric studies are needed to confirm these findings in larger cohorts.