Background: Ablation of drivers in persistent atrial fibrillation (AF) has shown controversial results. Purpose: To test the efficacy of a tailored approach for persistent AF ablation which includes pulmonary vein isolation (PVI) plus “subjective” identification and ablation of drivers. Methods: From May 2017 to December 2019, selected patients with persistent AF and ongoing AF at the beginning of the ablation procedure were included. Conventional high-density mapping catheters (PentaRay NAV, IntellaMap Orion or Advisor HD Grid) were used. Drivers were subjectively identified as: a) fractionated continuous (or quasi-continuous) electrograms on 1–2 adjacent bipoles, without dedicated software (Figure 1A, dashed line; PR = PentaRay NAV); and b) sites with spatiotemporal dispersion (i.e. all the cycle length comprised within the mapping catheter) plus non-continuous fractionation on single bipoles (Figure 1B, arrows; in panels A and B: paper speed 200 mm/s; ORB = 24-pole ORBITER Woven catheter, blue bipoles around tricuspid annulus and green bipoles into the coronary sinus). Ablation included PVI + focal or linear ablation targeting sites with drivers. Ablation success was defined as conversion to sinus rhythm or atrial flutter during ablation. Follow-up included visits with 24h Holter ECG at 3–6–12 months. Survival free from atrial arrhythmias lasting >30 seconds was compared between patients ablated with this tailored approach, and all consecutive patients with persistent AF treated with a PVI-only strategy during the same period. Results: 158 Patients received ablation: 35 with the tailored approach (61,7±10,2 years; 29% females) and 123 with only PVI (62,5±9,6 years; 25% females; 89% cryoablation). Basal characteristics were similar (Table 1). In the tailored-approach group, 14 patients (40%) presented 28 detectable sites with continuous fractionated electrograms, 26 on the left atrium and 2 on the right atrium, which was only mapped if ablation of drivers in the left atrium was not successful; 12 (43%) were located within the pulmonary vein antra. 27 patients (77%) showed 103 sites with spatiotemporal dispersion (4 [3–5] per patient). Ablation success was achieved in 17 patients (48%; conversion to sinus rhythm, n=7; conversion to atrial flutter, n=10) in the tailored-approach group and 1 patient (0,8%, sinus rhythm) in the PVI-only group. Excluding a 3-month blanking period, the tailored approach, compared to only PVI, improved one-year freedom from atrial arrhythmias (71% Vs 51%, p=0,05) and mean survival free from atrial arrhythmias (26±3 months; 95% CI 21–32 months Vs 18±2 months; 95% CI 15–22 months) (Figure 1C), at the cost of a longer median procedural time (246 [212–277] vs 108 [81–143] min, p<0,001) and fluoroscopy time (51 [36–76] vs 33 [21–45] min, p<0,001). Conclusion: Subjective identification and ablation of drivers, added to PVI, improved freedom from atrial arrhythmias. Funding Acknowledgement: Type of funding sources: None. Table 1. Basal characteristicsFigure 1