BackgroundOur previous study reported a modified endoscopic procedure for nonvalvular atrial fibrillation (AF) that requires only 3 ports in the left chest wall.HypothesisCertain preoperative variables might be predictive risk factors for AF recurrence among patients who underwent this procedure.MethodsFrom October 2010 to April 2014, 114 patients with either paroxysmal AF (PAF) or nonparoxysmal AF (non‐PAF) underwent the procedure and completed postoperative cardiac‐rhythm measurement via electrocardiography and Holter monitoring. Univariate and multivariate analyses of the possible AF‐related risk factors were conducted.ResultsDuring 2‐year follow‐up, 99 of 114 patients (86.8%) were free from atrial tachyarrhythmia. Results from univariate analyses showed that AF duration, left atrial diameter (LAD), left atrial minimum volume, left atrial empty fraction, left atrial expansion index, and left atrial active empty fraction (LAAEF) were significantly associated with postoperative AF recurrence. Results from multivariate analyses showed that AF duration (odds ratio [OR]: 1.194, 95% CI: 1.063‐1.340, P = 0.003), LAD (OR: 1.101, 95% CI: 1.005‐1.205, P = 0.039), and LAAEF (OR: 0.490, 95% CI: 0.277‐0.865, P = 0.014) were independent risk factors. There was no difference in AF recurrence between patients with PAF and non‐PAF (P = 0.250).ConclusionsOur 2‐year follow‐up study suggested that low LAAEF, long AF duration, and large LAD might be potential predictive risk factors for AF recurrence. Patients with PAF and non‐PAF had a similar AF recurrence rate after modified endoscopic ablation.