AIMS: The aim of this retrospective cohort study was to assess the influence of early recurrence (ER) after pulmonary vein isolation (PVI) for paroxysmal atrial fibrillation (AF) on long-term outcomes and to identify clinical variables associated with ER. METHODS: We retrospectively collected clinical and procedural data from 1285 patients with paroxysmal AF who underwent PVI from 2011 to 2016. Kaplan-Meier, receiver operating characteristic (ROC) curve, logistic and Cox regression analyses were performed to analyze the influence of ER on long-term outcomes. RESULTS: ER was observed in 13% of patients. Kaplan-Meier analyses showed significantly different outcomes in 1285 patients with and without ER (49% vs 74%, log rank P < .01) and in 286 patients in the subgroup that underwent reablation (44% vs 79%, log rank P < .01). The hazard ratio (HR) of ER was 1.7 within 48 hours (5% of patients), 2.7 within 1 month (5%), 3.0 within 2 months (2%), and 6.4 within 3 months (1%) for late recurrence (LR), P < .01. ROC analysis (area under the curve [AUC] = 0.79) resulted in 70.3% sensitivity and 74.2% specificity for a 14-day blanking period, and 53.1% sensitivity and 85.5% specificity for a 30-day blanking period. Female patients (odds ratio [OR] 1.69, P < .01) and those with diabetes (OR 1.95, P = .01) were at higher risk for ER. CONCLUSIONS: ER is observed in a substantial number of patients with paroxysmal AF after PVI and has a continuous direct effect on LR according to the timing of ER. Randomized trials are required to assess the safety and effects of reablations in a shortened blanking period on long-term outcomes.
AIMS: The aim of this retrospective cohort study was to assess the influence of early recurrence (ER) after pulmonary vein isolation (PVI) for paroxysmal atrial fibrillation (AF) on long-term outcomes and to identify clinical variables associated with ER. METHODS: We retrospectively collected clinical and procedural data from 1285 patients with paroxysmal AF who underwent PVI from 2011 to 2016. Kaplan-Meier, receiver operating characteristic (ROC) curve, logistic and Cox regression analyses were performed to analyze the influence of ER on long-term outcomes. RESULTS: ER was observed in 13% of patients. Kaplan-Meier analyses showed significantly different outcomes in 1285 patients with and without ER (49% vs 74%, log rank P < .01) and in 286 patients in the subgroup that underwent reablation (44% vs 79%, log rank P < .01). The hazard ratio (HR) of ER was 1.7 within 48 hours (5% of patients), 2.7 within 1 month (5%), 3.0 within 2 months (2%), and 6.4 within 3 months (1%) for late recurrence (LR), P < .01. ROC analysis (area under the curve [AUC] = 0.79) resulted in 70.3% sensitivity and 74.2% specificity for a 14-day blanking period, and 53.1% sensitivity and 85.5% specificity for a 30-day blanking period. Female patients (odds ratio [OR] 1.69, P < .01) and those with diabetes (OR 1.95, P = .01) were at higher risk for ER. CONCLUSIONS: ER is observed in a substantial number of patients with paroxysmal AF after PVI and has a continuous direct effect on LR according to the timing of ER. Randomized trials are required to assess the safety and effects of reablations in a shortened blanking period on long-term outcomes.