Minimally Invasive Epicardial Surgical Ablation Alone Versus Hybrid Ablation for Atrial Fibrillation: A Systematic Review and Meta-Analysis.

Abstract

Maintaining sinus rhythm in patients with non-paroxysmal AF is an elusive goal. Some suggest that hybrid ablation, combining minimally invasive epicardial surgical ablation with endocardial catheter ablation, may be more effective than either modality alone. However, randomised trials are lacking. We investigated whether hybrid ablation is more effective than epicardial ablation alone at preventing recurrent AF by performing a systematic review and meta-analysis. The review was prospectively registered with PROSPERO (CRD42016043389). MEDLINE and EMBASE were searched for studies of standalone minimally invasive epicardial ablation of AF and/or hybrid ablation, identifying 41 non-overlapping studies comprising 2737 patients. A random-effects meta-analysis, meta-regression and sensitivity analysis were performed. Single-procedure survival free from atrial arrhythmias without antiarrhythmic drugs was similar between epicardial-alone and hybrid approaches at 12 months (epicardial alone 71.5 %; [95 % CI 66.1-76.9], hybrid 63.2 %; [95 % CI 51.5-75.0]) and 24 months (epicardial alone 68.5 %; [95 % CI 57.7-79.3], hybrid 57.0 %; [95 % CI 33.6-80.4]). Freedom from atrial arrhythmias with AADs and rates of unplanned additional catheter ablations were also similar between groups. Major complications occurred more often with hybrid ablation (epicardial alone 2.9 %; [95 % CI 1.9-3.9], hybrid 7.3 %; [95 % CI 4.2-10.5]). Meta-regression suggested that bipolar radiofrequency energy and thoracoscopic access were associated with greater efficacy, but adjusting for these factors did not unmask any difference between epicardial-alone and hybrid ablation. Hybrid and epicardial ablation alone appear to be equally effective treatments for AF, although hybrid ablation may be associated with higher complication rates. These data derived from observational studies should be verified with randomised data.