Introduction: An increasing amount of data concerning the outcomes and predictors of recurrence after catheter ablation for atrial fibrillation (AF) is emerging, but the results are seldom in perfect concordance. In order to define the most lucrative combination of parametres to appreciate the eligibility for this procedure, further analysis is necessary. We sought to determine the clinical and paraclinical variables associated with a higher risk of early or late post-ablation arrhythmia relapse.Quantifying the symptomatic and pharmacological impact of the intervention in non-responders was considered a secondary end-point. Methods: 64 patients [56% paroxistic AF (PAF), 44 % persistent AF (PsAF)] were enrolled and followed for a median period of 18.9 months. A total of 50 clinical, ECG, TTE, TEE and intraprocedural variables were registered for each patient. After CARTO 3 mapping, ablation was sequentially targeted to antral segments of the pulmonary veins until bidirectional isolation was achieved. A stepwise approach was adopted for the non-PAF cases, that included complex fractionated atrial electrograms or resultant atrial tachycardias ablation. Clinical and 48 hours Holter ECG monitoring were performed at 1, 3, 6, 12 and 18 months. Results: The success rates were 90% at 6 months and 78% at 18 months for PAF and 55%, respectively 51%, for PsAF, after a mean procedure number of 1.3 ± 0.5 and a blanking period of 3 months. The mean left atrium volume was 72.8 ± 24 ml and AF duration of 4.9 ± 3.4 years. Post-procedural complications consisted only of two mild cases of pericarditis. On bivariate analysis, the predictors of recurrence were: age (p = 0.01), BMI (p = 0.02), CHADS2 score (p = 0.0297), left atrium dimension (p = 0.0046), left atrial appendage outflow velocity (LAAFV, p = 0.008) and the presence of obstructive sleep apnea (p = 0.036). 90% of the group where long term conversion couldn’t be attained reported an important symptomatic amelioration (ΔEHRAscore = -1.5). Antiarrhythmic drugs were discontinued in all but 15.8% of our lot. Conclusions: Our study suggests that satisfactory outcomes can be reached in almost all categories of patients. It confirms some of the recurrence predictors previously postulated and emphasizes the relevance of LAAFV, thus warranting its systematical assessment.