BACKGROUND Electrical isolation of the pulmonary veins (PVs) to treat paroxysmal atrial fibrillation (AF) has been described using "entry block" as an endpoint for PV isolation. We describe a new technique for guiding PV isolation, using "exit block" out of the PV after ablation as a criterion for successful isolation. RESULTS A circular mapping catheter was positioned at the os of arrhythmogenic PVs and ablation was performed proximal to the mapping catheter until entry block into the vein was achieved. Pacing was performed from the mapping catheter and from the ablator inside the PV to document exit block out of the PV. In patients in whom cardioversion did not restore sinus rhythm, PV isolation was performed in AF. Entry and exit block were reassessed in ablated veins after a 20-minute waiting period. Ninety-five PVs were ablated in 41 patients. A total of 66 PVs in 34 patients were ablated in sinus rhythm. After entry block was achieved, exit block was present in only 38 (58%) of 66 PVs. A total of 29 PVs in 21 patients were ablated in AF. After cardioversion to sinus rhythm, there was evidence of entry block into the PV in 20 (69%) of 29 PVs and exit block in only 14 (48%) of 29 PVs. There was no significant difference between the total number of lesions applied per vein in sinus rhythm compared with AF (11.6 +/- 8.6 vs 10.3 +/- 6.2; P = NS). There was recovery of conduction after a 20-minute waiting period in 9 (11%) of 84 PVs. CONCLUSIONS Identification of exit block after ostial PV ablation provides a clear endpoint for electrical isolation of the PVs. Isolation of the PVs can be performed during sustained AF without the need to apply excess RF lesions. Applying a 20-minute waiting period after electrical isolation will identify reconnection in approximately 10% of PVs.