A total of 1,875 patients underwent 2,307 consecutive RFCA for AF at Korea University Anam Hospital between June 1998 and May 2014 (Figure 1). Patients with highly symptomatic drug refractory AF were selected for ablation. Among them, 379 patients (20.2%) were women, and 367 patients (19.5%) who underwent first ablation were included in this study. Twelve patients who underwent first ablation in other hospital were excluded. Three hundred sixty-seven men were selected as control to 1:1 match the study group by age (±1 year), type of AF (paroxysmal or non-paroxysmal), and duration of AF before index procedure (±2 years). We considered index procedure era as matching criteria to decrease bias related with ablation technology development.

Duration of AF was calculated by the period between the time of the first diagnosis and index procedure. It was assessed by confirming previous medical record. Definition of clinical type of AF (paroxysmal or non-paroxysmal) was followed the American Heart Association (AHA)/American College of Cardiology (ACC) recommendations.16) Atrial arrhythmia recurrence was defined with documentation of atrial arrhythmia (>30 seconds) including AF or atrial tachycardia (AT). However, we did not consider atrial arrhythmia within blanking period as recurrence. Blanking period was defined as 3 months after procedure. Procedural success was no atrial arrhythmia recurrence during follow-up period with or without anti-arrhythmic drug (AAD). Procedure-related (within 1 week) complication were defined as hematoma requiring transfusion, arteriovenous fistula requiring operation, dissection of major vessel, periprocedural cerebrovascular accident (CVA), tamponade, shock or death, etc. Primary outcome is atrial arrhythmia recurrence. Secondary outcome is thromboembolic event.

AADs were discontinued in more than 5 half-lives and amiodarone was discontinued at 1 month prior to procedure. Anticoagulation was continued for at least 2 months after ablation in all patients regardless of CHA₂DS₂-VASc score. Transesophageal echocardiography was performed in all patients within 24 hours immediately before procedure to rule out the presence of left atrial (LA) thrombi. All the patients underwent electrophysiology study and ablation after written informed consent was obtained.

Our ablation procedure has evolved over the 15-year period of this study. We performed focal ablation at the pulmonary vein (PV) guided by Lasso catheter (Biosense Webster, Inc., Irvine, CA, USA) from 1998 to 2004. Non-contact mapping-guided PV ablation since 2005. We have started to perform circumferential ablation of 4 PVs guided by NavX (St. Jude Medical Inc., St. Paul, MN, USA) since 2007. The end point of PV isolation (PVI) was elimination of PV potentials on the 10-bipole ring-shaped catheter at each PV. Since 2009, subsequent ablation strategy differed according to the clinical type of AF. In case of patient with paroxysmal AF (PAF), non-PV triggers of sustained AF were evaluated and targeted after PVI. To identify triggers initiating AF, multiple electrical cardioversion was performed subsequently under isoproterenol (10–20 µg/min) infusion. Triggers were evaluated if they caused re-initiation of sustained or non-sustained AF within 2 minutes after repetitive cardioversion. After PVI, same protocol was repeated at least 3 times, to identify non-PV foci initiating AF. In case of patient with non-PAF, additional substrate modification was performed after PVI. If AF was sustained after PVI, 3-dimensional complex fractionated atrial electrograms (CFAEs) of LA were mapped during 6 seconds of AF with NavX software. Ablation was performed using the CFAE map until AF termination or organized to an AT, or the fractionated potential in LA was eliminated. The CFAE in the right atrial (RA) was targeted for ablation if AF persisted after LA CFAE ablation. If AF was organized to AT during PVI or CFAEs guided ablation, activation and entrainment mapping guided ablation were performed. If AF/AT was terminated by ablation, inducibility of sustained AF/AT by pacing was evaluated in patients with non-PAF. If SR is not restored after ablation as mentioned above, patient was cardioverted with internal or external direct-current electrical cardioversion. The endpoint of procedure for PAF was to eliminate all triggers including PV. Endpoint of procedure for non-PAF was termination of AF/AT and no induction of sustained atrial arrhythmia by pacing protocol.

Every patient visited outpatient clinics 2 weeks after discharge and every 2–4 weeks during the initial 3 months to monitor the occurrence of complications and early recurrence. Patient visited every 3 months after the blanking period. Electrocardiography was routinely checked during each follow-up at outpatient clinics. 24-hours Holter monitoring or 7-days event recorder was performed at 1, 3, 6, 9, and 12 and then every 6 months after discharge in asymptomatic patients. In case of patients with implanted intracardiac device, the recurrence was checked by interrogation of device. If patients experienced tachycardia-related symptom without documentation, it was confirmed by additional Holter or event recorder. During blanking period, prescription of AADs was at the discretion of physicians. Repeat ablation for recurred AF/AT was not allowed during the blanking period. AAD was discontinued after the end time of the blanking period.

Subgroup analysis was performed in categories defined by clinical AF type and age. We set up criteria of age as 55 to evaluate hormonal effect. It has been known that mean age at natural menopause in Asian woman is 49 years and hormone did not affect human body at age over than 55.17) Clinical outcome was estimated in each 4 subgroups.

Continuous variables are expressed as a mean standard deviation. Categorical variables are reported as counts with percentage. An independent Student's t-test was used to compare continuous variables between women and men groups. Categorical variables were compared by a χ² test or Fisher's exact test. Time to AF/AT recurrence was calculated using the Kaplan-Meier analysis with comparisons between both genders performed by log-rank statistics. To analyze independent predictors of AF recurrence after ablation and clinical outcomes, multivariable Cox regression model was used. Confounders adjusted in the multivariable model included age, sex, LA diameter index, and history of heart failure (HF). The hazard ratio (HR) and 95% confidence interval (CI) were presented in the results. A p value <0.05 was considered statistically significant. Statistical analyses were performed using SPSS 20.0 software package (SPSS Inc., Chicago, IL, USA).

Patient characteristics and echocardiographic parameters according to gender are shown in Table 1. Mean age of both groups was 59±10 years. All study population were Asian. PAF was present in 69% of overall patients and 31% had non-PAF. Women were more likely to have history of thyroid disease (7.9% vs. 3.8%; p=0.013) and sick sinus syndrome (5.2% vs. 1.1%; p=0.002) compared to men. Women were more likely on class III AAD than men. Women tended to have more prior CVA history, but no significant difference was found. No significant difference was found between groups in terms of other cardiovascular comorbidities including the presence of coronary artery disease. CHA₂DS₂-VASc score (calculated excluding 1 point of female gender) was not different between both groups. Mean duration of AF and the number of previously ineffective AAD were similar between both groups. Mean LVEF (55% vs. 56%; p=0.613) and non-adjusting mean LA anteroposterior diameter (40.9 vs. 41.0 mm; p=0.761) did not differ between both groups. However, LA diameter index (26±4 vs. 23±4 mm/m²) and measured E/E′ (11±4 vs. 10±3) were significantly higher in women than men (both p<0.001).

The mean procedure time, fluoroscopic time and total radiofrequency (RF) delivery time were not significantly different (Table 2). Non-PV trigger revealed by mentioned protocol was noted in 10% of women and 12% of men (p=0.724).

The number of patients who underwent additional substrate modification to PVI was not different (31% vs. 35%; p=0.201). Successful block of peri-mitral isthmus (13% vs. 10%; p=0.098), roof (14% vs. 13%; p=0.297), and cavotricuspid isthmus (49% vs. 51%, p=0.262) were achieved without difference between both groups.

Procedure related complication rate was significantly higher in women group (9.2% vs. 4.9%; p=0.030) (Figure 2). The most common complication is cardiac tamponade. Its occurrence rate was significantly higher in women (6.5% vs. 2.7%; p=0.021). All these patients were treated by percutaneous drainage or surgical repair. Second common complication was vascular access related ones including dissection, arteriovenous fistula and hematoma requiring transfusion. It did not significantly differ between women and men (1.6% vs. 0.8%; p=0.253). The occurrence rates of bradycardia requiring temporary or permanent pacing (n=4), peri-procedural (within 30 days after procedure) CVA (n=3), gastrointestinal paralysis (n=1), and phrenic nerve palsy (n=1) were not related to gender.

Mean follow-up period after index ablation was 55±38 months. The patients without recurrence of atrial arrhythmia was 405 (79.4%) after ablation for PAF and 133 (59.3%) after ablation for non-PAF during follow-up period. Successful ablation rate without recurrence during follow-up was not significantly different between both groups. The 76% of men and 71% of women were arrhythmia-free without AAD (HR, 1.12; 95% CI, 0.82–1.53; p=0.478). Kaplan-Meier curve of arrhythmia-free survival in overall women and men group is shown in Figure 3.

An estimated recurrence-free survival in all subgroup is shown in Figure 4. The women with non-PAF or age ≤55 years were more likely to suffer from a recurrence than men (log-rank p=0.019 in non-PAF subgroup, p=0.005 in age ≤55 subgroup). Among the patient with PAF, freedom from atrial arrhythmia recurrence rate was not different regardless of age (p=0.943). There was no gender-related difference in subgroup of age >55 years (p=0.912). The women with PAF showed no significant difference of AF/AT recurrence compared to men in age >55 years subgroup (p=0.197). In age >55 subgroup with PAF, outcome tended to be worse in women but not statistically different (p=0.052). The women had worse clinical outcome than men in age ≤55 years with non-PAF subgroup (p=0.019), but not in age >55 years subgroup (p=0.176).

The multivariable Cox regression analysis showed that female gender was the independent predictor for recurrence after catheter ablation in age ≤55 years subgroup with non-PAF (HR, 2.539; 95% CI, 1.112–5.801; p=0.027). In contrast, it was not the independent predictor for recurrence in age >55 subgroup with non-PAF (HR, 0.734; 95% CI, 0.458–1.176; p=0.196), age ≤55 subgroup with PAF (HR, 1.974; 95% CI, 0.934–4.170; p=0.075) and >55 subgroup with PAF (HR, 0.733; 95% CI, 0.431–1.246; p=0.251). We analyzed characteristics of non-PAF subgroup with age ≤55 years. Women in that subgroup have higher left atrial diameter (LAD) index (27±4 vs. 23±3; p<0.001) and E of E′ ratio (11±4 vs. 8±2; p=0.002) compared to men (Table 3).

In total follow-up period regardless of recurrence, several ischemic strokes occurred. Three were women and one was man (0.8% vs. 0.3%; p=0.127). Two women and man were on recurrence status but one woman was not. Recurrence of atrial arrhythmia was not documented in one woman at the time of stroke event.