Background The optimal endpoint for catheter ablation of persistent atrial fibrillation (AF) remains ambiguous.This study investigated the impact of AF termination as a procedural endpoint and the termination mode on long-term clinical outcome.Methods Two hundred and ninety-three patients who underwent stepwise ablation for persistent AF were categorized into the AF termination by ablation group and into the electrical cardioversion (CV) group.Subgroups were also analyzed based on different termination modes.Follow-up assessment included early recurrence and sinus rhythm (SR) maintenance.Results During initial ablation,33 patients (11.3％) were directly converted to SR,166 patients (56.7％) were converted to atrial tachycardia (AT) that subsequently restored SR with further ablation in 98 patients (33.4％),and a total of 162 patients (55.3％) underwent cardioversion due to persistent atrial arrhythmias.Comparison between termination by ablation and termination by cardioversion in patients exhibiting AF or AT revealed that no significant difference was observed in early recurrence (38.2％ vs.43.8％,P=0.328) and SR maintenance (67.2％ vs.59.8％,P=0.198) during the (23±7) months follow-up.Even after repeat ablation,the SR maintenance continued to exhibit no statistical difference in above two groups (72.5％ vs.70.4％,P=0.686).Further analysis of subgroups,however,demonstrated that patients with AF terminated directly to SR experienced better clinical outcomes than other subgroups (P ＜0.05).Furthermore,atrial arrhythmias present during ablation have been implicated in prediction of recurrence mode:AF or AT (P ＜0.05).Conclusions Termination as a procedural endpoint is not associated with favorable long-term SR maintenance in persistent AF.AF methods that convert arrhythmia directly to SR have,however,been linked with improved clinical outcomes,although conversions to AT may not be correlated.Atrial arrhythmias observed during the ablation may be used to predict the recurrence mode.