Background: The relationship between vagal modification and paroxysmal atrial fibrillation (PAF) recurrence after segmental pulmonary vein (PV) isolation (S-PVI) was investigated. Methods and Results: S-PVI was performed in 77 PAF patients using a multielectrode basket or circular catheter to achieve electrical disconnection of all 4 PVs independent of eliminating vagal reflexes. Serial Holter-recordings were obtained at baseline, immediately and 1, 3, 6, and 12 months after S-PVI to analyze the heart rate variability. Fifty-one patients were free from symptomatic PAF (Group A) and 26 had late PAF recurrences (Group B) at 12-month follow-up. Immediately after S-PVI, the root mean square of the successive differences (rMSSD) and high-frequency (HF) power, which reflected parasympathetic nervous activity, were significantly lower in Group A than in Group B (rMSSD: 33.6±26.0 vs 60.6±23.2 ms, P<0.05; ln HF: 8.73±0.84 vs 9.31±0.95 ms2, P<0.05). There were no significant differences in the average heart rate or ratio of the low-frequency to HF powers between the 2 groups. By multivariate analysis, only the HF immediately after S-PVI was an independent predictor of PAF recurrence (hazard ratio 1.707, 95% confidence interval 1.057-2.756, P<0.05). Conclusions: Vagal modification after S-PVI could also help prevent late recurrence of PAF.