Dilated left atrium as a predictor of late outcome after pulmonary vein isolation concomitant with aortic valve replacement and/or coronary artery bypass grafting†.

Objectives: Left atrial (LA) dimension can predict atrial fibrillation (AF) recurrence after catheter-based or surgical ablation. Pulmonary vein isolation (PVI) may be a surgical option during aortic valve replacement (AVR) and/or coronary artery bypass grafting (CABG), though consensus regarding patient selection and late outcome is lacking.

Methods: We studied 160 patients (mean age 70 ± 9 years) with paroxysmal AF who underwent radiofrequency-based PVI during AVR and/or CABG, and were followed up postoperatively for at least 6 months. Mean preoperative LA dimension was 44 ± 7 mm. Serial echocardiography was performed to evaluate left ventricular (LV) and LA dimensions, E/e', estimated systolic pulmonary artery (PA) pressure and degree of valvular regurgitation. Follow-up was completed with a mean duration of 47 ± 25 months.

Results: At the latest follow-up, 133 patients (83%) remained in sinus rhythm. Preoperative LA dimension was independently associated with increased risk of AF recurrence at 6 months after surgery [adjusted odds ratio 1.3 per 1-mm increase in LA dimension, 95% confidence interval (CI) 1.1-1.6, P < 0.001]. Receiver-operating characteristic curve analysis demonstrated an optimal cut-off value for preoperative LA dimension of 45 mm to predict sinus rhythm restoration (98% for <45 mm vs 55% for ≥45 mm, P < 0.001). Patients with LA dimension ≥45 mm had a significantly lower 5-year survival rate (62 ± 7 vs 82 ± 7%, P = 0.025) and freedom from adverse events defined as cerebral infarction/haemorrhage, admission for heart failure, catheter ablation and permanent pacemaker implantation (58 ± 7 vs 91 ± 4%, P < 0.001). Multivariate analysis showed that preoperative LA dimension ≥45 mm was independently associated with adverse events (adjusted hazards ratio 2.4, 95% CI 1.2-5.1, P = 0.019). Serial echocardiography demonstrated improvement in LV systolic function irrespective of LA dimension, whereas patients with LA dimension ≥45 mm showed less improvement in LA dimension and systolic PA pressure (interaction effect P < 0.001) and persistent higher E/e' (group effect P < 0.001), along with aggravated tricuspid regurgitation.

Conclusions: In patients with paroxysmal AF related to aortic valve disease and/or coronary artery disease, a dilated left atrium (≥45 mm) was associated with inferior AF- and event-free survival after PVI, accompanied by persistent abnormalities in cardiac and haemodynamic function. These findings may assist patient selection for PVI during AVR and/or CABG.