Abstract

Background:

Long-standing persistent atrial fibrillation (LSPAF) remains a challenge of catheter ablation. The efficiency and optimal procedural endpoints of “2C3L plus” approach—a strategy combining pulmonary vein isolation (PVI) with linear and complex fractionated atrial electrogram (CFAEs) ablation—is unclear.

Methods:

This single-center, retrospective cohort study included 260 consecutive patients with LSPAF (defined as continuous AF lasting > 12 months) who underwent de novo radiofrequency catheter ablation between January 2020 and January 2022. All patients received a standardized “2C3L plus” strategy. The primary endpoint was freedom from any documented atrial tachyarrhythmia lasting >30 s, off antiarrhythmic drugs, at 1-year follow-up. Predictors of recurrence were analyzed using multivariable Cox regression analysis.

Results:

Intraprocedural atrial fibrillation (AF) termination was achieved in 103 of 260 (39.6%) patients and 90 (34.6%) patients converted to atrial flutter (AFL) during ablation. Acute termination of AF directly to sinus rhythm (SR) was not associated with a lower risk of recurrence (adjusted HR: 0.765, 95% CI: 0.410–1.428, P = 0.400). However, intraprocedural conversion from AF to AFL was associated with significantly reduced recurrence risk (Uni: HR: 0.319, 95% CI: 0.142–0.714, P = 0.005; adjusted HR: 0.306, 95% CI: 0.133–0.704, P = 0.005). Further analysis revealed that the sequential intraprocedural conversion of AF-AFL-SR during ablation was a strong and independent predictor of arrhythmia-free survival (Uni: HR: 0.275, 95% CI: 0.108–0.696, P = 0.006; adjusted HR: 0.305, 95% CI: 0.119–0.784, P = 0.014).

Conclusion:

In patients with LSPAF undergoing extensive “2C3L Plus” substrate ablation, the intraprocedural organization of AF-AFL-SR, rather than AF termination itself, emerged as a powerful independent predictor of 1-year arrhythmia-free survival, suggesting its value as a more meaningful prognostic endpoint.

1 Introduction

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, and its natural course often progresses from paroxysmal to persistent, ultimately evolving into long-standing persistent AF (LSPAF) lasting over one year. This progression is accompanied by significant atrial electrical and structural remodeling, creating a complex substrate that sustains the arrhythmia and renders LSPAF particularly challenging to treat (1).

Catheter ablation has become a cornerstone therapy for restoring and maintaining sinus rhythm. Pulmonary vein isolation (PVI), which electrically isolates the pulmonary veins as the primary trigger for AF, is universally recognized as the fundamental element of all AF ablation procedures. However, for LSPAF patients, due to extensive underlying atrial remodeling, PVI alone yields limited success. This has driven clinical exploration of strategies combining PVI with additional substrate modification. These strategies primarily include two approaches: 1) creating linear ablation lines in the left atrium to segment the atria and block macro-reentrant circuits; and 2) ablating complex fractionated atrial electrograms (CFAEs), which are considered key drivers perpetuating AF.

In this context, our center previously proposed the “2C3L” ablation strategy, combining CPVI (“2C”) with three linear ablation lines (“3L”): the left atrial roof line, mitral isthmus line, and tricuspid isthmus line. The recent PROMPT-AF trial, which tested Ethanol Infusion of the Vein of Marshall (EIVOM)-assisted “2C3L” strategy in persistent AF, reported a significantly higher 1-year success rate of 70.7% than PVI alone (2). Given the even more complex substrate in LSPAF, we further developed the “2C3L plus” strategy by adding extensive complex fractionated atrial electrograms (CFAEs) ablation to “2C3L”, aiming to achieve intraprocedural AF termination. However, this “more-is-better” philosophy of extensive substrate modification faced significant challenges from subsequent large randomized clinical trials. The landmark STAR-AF II trial demonstrated that for persistent AF, adding either CFAEs ablation or linear ablation to PVI provided no additional benefit (3). The RASTA trial even found that adding CFAEs ablation resulted in worse outcomes (4). Furthermore, the prognostic value of intraprocedural AF termination—a key procedural endpoint of extensive substrate modification—has become highly debated. A subgroup analysis of STAR-AF II indicated that although additional substrate modification significantly increased the intraprocedural AF termination rate, AF termination itself did not predict long-term success (5).

Given this background and controversies, this study aims to retrospectively analyze the outcomes of the “2C3L plus” strategy in LSPAF patients. Our primary objectives are: 1) to describe the one-year freedom from AF recurrence off antiarrhythmic drugs (AADs) and the intraprocedural AF termination rate in LSPAF patients treated with this strategy; 2) to re-evaluate whether intraprocedural AF termination predicts AF recurrence specifically under this strategy; and 3) to explore the relationship between different modes of AF termination (direct conversion to sinus rhythm vs. conversion through AFL/AT transitional rhythm.

2 Materials and methods

This was a single-center, retrospective observational study. A total of 260 inpatients with LSPAF who underwent de novo radiofrequency ablation were consecutively enrolled in this study from January 2020 to January 2022 at Anzhen Hospital. LSPAF was defined as continuous atrial fibrillation (AF) persisting for more than 12 months, despite attempts to restore sinus rhythm. The major exclusion criteria were paroxysmal atrial fibrillation, history of AF ablation, AF lasting less than 12 months, rheumatic heart disease, and loss to follow-up. This study was performed following the Helsinki Declaration of Human Rights and was approved by the institutional review board of Beijing Anzhen Hospital, Capital Medical University (Ethics Number: 2026033X).

Data regarding the demographic characteristics, clinical features and laboratory examinations were collected for all subjects including sex, age, past medical history (hypertension, diabetes mellitus), smoking status, body mass index, blood pressure, heart rate (HR), systolic blood pressure (SBP), left atrial diameter (LAD), left ventricular ejection fraction (LVEF), and albumin.

Before ablation, oral anticoagulation was given for at least four weeks. Transesophageal echocardiography or intracardiac echocardiography was performed to exclude left atrial thrombus. During RFCA, heparin was administered to maintain an activated clotting time of more than 300 s. After RFCA, anti-arrhythmic drugs and oral anticoagulation were given for at least three months. Under a CARTO mapping system (CARTO; Biosense Webster, Inc., Irvine, California), RFCA was performed at a maximum temperature of 45 °C, maximum power of 50 W, and flow rate of ≥15 mL/min. Ablation index targets will be set with 500–550 for anterior wall; 350–400 for posterior wall; 450–550 for the LA roof and CTI; 550–600 for MI. First, the CPVI, roofline, mitral isthmus, and cavotricuspid isthmus were ablated one by one. Then, the bottom-line, left atrial CFAE and right atrial CFAE were ablated one by one. However, if termination of AF occurred during this linear ablation phase, the following CFAE ablation steps were not performed. All CFAEs defined as electrograms with continuous activity or complex fractionated electrogram mean detected by the system of <80 ms were eliminated.

Whenever AF organized into atrial flutter (AFL) documented by a sudden change to regular tachycardia with fixed cycle length and consistent activation pattern, detailed activation mapping was immediately performed. The re-entry circuit was mapped with high-density multipolar catheters. Targeted RF applications were delivered at the critical isthmus until complete conduction block across the circuit was achieved and AFL terminated. If AFL could not be terminated after meticulous mapping and ablation, or if AF persisted without transitioning to AFL, external direct-current cardioversion was performed under deep intravenous sedation to restore sinus rhythm.

Linear block are validated under sinus rhythm or atrial pacing: Roofline bidirectional block will be confirmed by: 1) caudal-cranial activation pattern in the posterior wall under sinus rhythm or pacing anterior to the ablation line (LAA) and 2) caudal-cranial activation in the anterior wall under pacing posterior to the ablation line. MI bidirectional block is confirmed by: 1) proximal-to-distal CS activation pattern when pacing at the LAA and left lateral ridge; 2) the activation of the anterior mitral annulus is from the septum to the LA lateral wall and then to ablation line when pacing at the distal CS. CTI bidirectional block is confirmed by: 1) the activation is from the septum to the ablation line when pacing at RA free wall; 2) the activation is from the RA lateral wall to the ablation line when pacing at the proximal CS.

Patients were followed-up by telephone and outpatient clinic visit. Twenty-four-hour Holter monitoring was checked at 1, 2, 3, 6, 9 and 12 months after discharge. In addition, if patients felt serious symptoms of arrhythmia, they were required to undergo an electrocardiogram at the closest hospital. AF recurrence was defined as documented atrial tachycardia, atrial flutter or AF for at least 30 s during the 1-year follow-up. The final diagnosis of recurrence was reviewed by two cardiologists.

The statistical computations were performed using R version 4.5.0 (The R Project for Statistical Computing, Vienna, Austria). Continuous variables are reported as the means ± standard deviations for normally distributed data or medians and interquartile for non-normally distributed data. They were compared using Student's t-test if normally distributed or the Mann–Whitney U-test if nonnormally distributed. We used the Kolmogorov‒Smirnov test to check their normality. Discrete variables are expressed as frequencies and percentages and were compared using the chi-square test. Multivariable Cox regression analysis was performed to detect the independent risk factor for AF recurrence with the adjustment of age, sex, smoking, drinking, LAD, history of hypertension, diabetes and LVEF. The Kaplan–Meier method was used to estimate the relationship between freedom from atrial tachyarrhythmias and ablation termination or AFL. A two-sided p value <0.05 was considered statistically significant.

3 Results

Baseline characteristics of the participants are summarized in Table 1. The average age was 55.8 ± 7.3 years, and 195 (75.0%) patients were male. Patients in the recurrence group had a significantly larger LAD compared to those in the no-recurrence group. There were no significant differences in the incidence of hypertension, diabetes, severe mitral regurgitation, or severe tricuspid regurgitation between the two groups (all P > 0.05).

VariablesAll RecurrenceNo recurrenceP Valuen = 260n = 45n = 215Male, n (%)195 (75.0)37 (82.2)158 (73.5)0.219Age, years55.8 ± 7.356.4 ± 8.555.6 ± 7.00.511AF duration, month51.6 ± 47.071.9 ± 49.747.4 ± 45.4<0.001Left atrial diameter, mm42.8 ± 4.246.6 ± 4.742.0 ± 43.6<0.001LVEF, %51.8 ± 3.351.8 ± 3.551.8 ± 3.20.892BMI, Kg/m225.4 ± 1.626.3 ± 2.025.2 ± 1.4<0.001Severe mitral valve regurgitation, n (%)27 (10.4)8 (17.8)19 (8.8)0.074Severe tricuspid valve regurgitation, n (%)36 (13.8)12 (26.7)24 (11.2)0.066Hypertension, n (%)124 (47.7)26 (57.8)98 (45.6)0.136Diabetes mellitus, n (%)82 (31.5)17 (37.8)65 (30.2)0.322Systolic blood pressure, mmHg127.6 ± 8.0128.2 ± 10.1127.4 ± 7.50.560Diastolic blood pressure, mmHg79.6 ± 6.680.1 ± 6.879.5 ± 6.60.550Fasting plasma glucose, mmol/L5.46 ± 1.035.91 ± 0.875.36 ± 1.040.001Smoking, n (%)95 (36.5)14 (31.1)81 (37.7)0.406Alcohol consumption, n (%)76 (29.2)18 (40.0)58 (27.0)0.081PVI ablation, n (%)260 (100)45 (100)215 (100)-Roofline ablation, n (%)260 (100)45 (100)215 (100)-Mitral isthmus ablation, n (%)260 (100)45 (100)215 (100)-Mitral isthmus block achieved, n (%)230 (88.5)39 (86.7)191 (88.8)0.679Cavotricuspid isthmus ablation, n (%)260 (100)45 (100)215 (100)-Cavotricuspid isthmus block achieved, n (%)252 (96.9)44 (97.8)208 (96.7)0.715Bottom line, n (%)237 (91.2)37 (82.2)200 (93.0)0.119Left atrial CFAE, n (%)222 (85.4)41 (91.1)181 (84.2)0.232Right atrial CFAE, n (%)195 (75.0)36 (80.0)159 (74.0)0.394Superior vena cava, n (%)42 (16.2)8 (17.8)34 (15.8)0.745Coronary sinus, n (%)125 (48.1)24 (53.3)101 (47.0)0.438Ethanol ablation of the vein of Marshall, n (%)37 (14.2)5 (11.1)32 (14.9)0.510Switched to AFL, n (%)90 (34.6)7 (15.6)83 (38.6)0.003AFL ablation termination, n (%)76 (29.2)5 (11.1)71 (33.0)0.003Terminate AF, n (%)103(39.6)15(33.3)88(40.9)0.404

Baseline characteristics of the recurrence and no recurrence.

AF, atrial fibrillation; LVEF, left ventricular ejection fractions; BMI, body mass index; PVI, pulmonary vein isolation; CFAE, complex fractionated atrial electrograms; AFL, atrial flutter.

Intraprocedural AF termination by ablation was achieved in 103 patients (39.6%). During ablation, 90 patients (34.6%) converted to AFL; of these, 38 (42.2%) occurred AFL after completion of the bottom line, 27 (30.0%) after left atrial CFAE ablation, 15 (16.6%) after right atrial CFAE ablation, 7 (7.8%) after MI line ablation and 3 (3.3%) after PVI ablation. Subsequent mapping during AFL confirmed MI-dependent AFL in 42 (46.7%), CTI-dependent AFL in 30 (33.3%), roofline-dependent AFL in 7 (7.8%), and other forms of AFL in 11 (12.2%) (Table 2). 76 were subsequently converted to sinus rhythm with further ablation, while 14 required cardioversion (CV). A comparison of clinical and procedural characteristics between patients who achieved AF termination via ablation (termination group, n = 103) and those who required CV (CV group, n = 157) is presented in Table 2. Patients in the termination group had a significantly smaller LAD and shorter AF duration compared to the CV group.

VariablesTerminationCardioversionP Valuen = 103n = 157Male, n (%)80 (77.7)115 (73.2)0.421Age, years55.4 ± 6.556.0 ± 7.80.516AF duration, month37.8 ± 26.860.7 ± 54.60.004Left atrial diameter, mm41.6 ± 3.943.6 ± 4.2<0.001LVEF, %51.8 ± 3.251.8 ± 3.40.989BMI, Kg/m225.2 ± 1.325.5 ± 1.70.205Severe mitral regurgitation, n (%)12 (11.7)15 (9.6)0.588Severe tricuspid regurgitation, n (%)15 (14.6)21 (13.4)0.786Hypertension, n (%)43 (41.7)81 (51.6)0.120Diabetes mellitus, n (%)34 (33.0)48 (30.6)0.679Systolic blood pressure, mmHg128.4 ± 8.4127.1 ± 7.60.194Diastolic blood pressure, mmHg79.4 ± 25.979.7 ± 7.10.676Fasting plasma glucose, mmol/L5.60 ± 1.055.36 ± 1.000.073Smoking, n (%)42 (40.8)53 (33.8)0.250Alcohol consumption, n (%)31 (30.1)45 (28.7)0.804Mitral isthmus block achieved, n (%)92 (89.3)138 (87.9)0.726Cavotricuspid isthmus block achieved, n (%)100 (97.1)152 (96.8)0.901Bottom line, n (%)90 (87.4)147 (93.6)0.083Left atrial CFAE, n (%)75 (72.8)147 (93.6)<0.001Right atrial CFAE, n (%)54 (52.4)141 (89.8)<0.001Superior vena cava, n (%)19 (18.4)23 (14.6)0.417Coronary sinus, n (%)45 (43.7)80 (51.0)0.252Ethanol ablation of the vein of Marshall, n (%)14 (13.6)23 (14.6)0.812Switched to AFL, n (%)76(73.8)14(8.9)<0.001

Clinical and procedural characteristics of patients.

AF, atrial fibrillation; LVEF, left ventricular ejection fractions; BMI, body mass index; PVI, pulmonary vein isolation; CFAE, complex fractionated atrial electrograms; AFL, atrial flutter.

All enrolled patients underwent PVI, roofline, MI and CTI ablation, where the block rate were 100%, 100%, 88.5% and 96.9% respectively. We completed MI bidirectional conduction block in 92 (89.3%) for termination group and 138 (87.9%) patients for CV group. The CTI was blocked in 100 (97.1%) patients in the termination group, and 152 (96.8%) patients achieved bidirectional conduction block in the CV group. In the termination group, bottom-line ablation was performed in 87.4% of patients, left atrial CFAE ablation was performed in 72.8%, and right atrial CFAE ablation was performed in 52.4%. EIVOM was done in 14 (13.6%) patients in the termination group and 23 (14.6%) in the CV group.

After a 12-month follow-up period, arrhythmia recurrence was confirmed in 45 patients (17.3%). The recurrent rate was 14.6% (15 of 103) in the termination group and 19.1% (30 of 157) in the CV group; this difference was not statistically significant (Table 3). The most common type of recurrent arrhythmia was AFL, observed in 34 of the 45 patients (75.6%) with recurrence. Among the patients with recurrence, 21 underwent a second ablation procedure. Mapping during the repeat procedure identified the recurrent arrhythmia as AF in 4 patients (19.0%), CTI-dependent AFL in 2 (9.5%), MI-dependent AFL in 10 (47.6%), roofline-dependent AFL in 2 (9.5%), and other forms of AFL in 3 (14.3%). A significant reduction in mean LAD was observed at the 1-year follow-up compared to baseline (39.57 mm vs. 42.82 mm; P < 0.001). post hoc analysis showed LA diameter were decreased in both termination and CV group (P < 0.001 for all).

VariablesAllTerminationCVP Valuen = 260n = 103n = 157Free from AF, n (%)215 (82.7)88 (85.4)127 (80.9)0.343Recurrence with AFL, n (%)34 (75.6)12 (80.0)22 (73.3)0.624Recurrence with paroxysmal AF, n (%)7 (15.6)2 (13.3)5 (16.7)0.771Recurrence with persistent AF, n (%)4 (8.9)1 (6.7)3 (10.0)0.711Repeat ablation, n (%)21 (46.7)9 (60.0)12 (40.0)0.205 AF4 (19.1)1 (11.1)3 (25.0)0.422 Cavotricuspid isthmus AFL2 (9.5)1 (11.1)1 (8.3)0.830 Mitral isthmus AFL10 (47.6)4 (44.4)6 (50.0)0.801 Roofline AFL2 (9.5)1 (11.1)1 (8.3)0.830 Others AFL3 (14.3)2 (22.2)1 (8.3)0.368Left atrial diameter before RFCA, mm42.8 ± 4.241.6 ± 3.943.6 ± 4.2<0.001Left atrial diameter after RFCA 1 year, mm39.6 ± 4.1*38.2 ± 3.7*40.5 ± 4.1*<0.001

AF, atrial fibrillation; AFL, atrial flutter.

Univariable and multivariable Cox regression analyses were performed to identify predictors of arrhythmia-free survival (Table 4). In the multivariable analysis, AF termination by ablation alone was not significantly associated with arrhythmia-free survival (Uni: HR: 0.739, 95% CI: 0.397–1.373, P = 0.754; adjusted HR: 0.765, 95% CI: 0.410–1.428, P = 0.400). However, intraprocedural conversion to AFL was identified as an independent protective factor against recurrence (Uni: HR: 0.319, 95% CI: 0.142–0.714, P = 0.005; adjusted HR: 0.306, 95% CI: 0.133–0.704, P = 0.005). Further analysis revealed that among patients whose arrhythmia terminated during ablation, those who first converted to AFL had a significantly lower risk of recurrence compared to those who converted directly to sinus rhythm (Uni: HR: 0.275, 95% CI: 0.108–0.696, P = 0.006; adjusted HR: 0.305, 95% CI: 0.119–0.784, P = 0.014).

VariableUnivariate analysisMultivariate analysisHR (95% CI)P-valueHR (95% CI)P-valueMale, n (%)1.57 (0.73–3.38)0.245Age, years1.01 (0.97–1.05)0.772AF duration, month1.007 (1.003–1.012)0.0021.007 (1.002–1.013)0.011Left atrial diameter, mm1.22 (1.15–1.29)<0.0011.24 (1.15–1.34)<0.001LVEF, %1.00 (0.91–1.09)0.932BMI, Kg/m21.33 (1.15–1.54)0.0001.33 (1.12–1.58)0.001Severe mitral valve regurgitation, n (%)2.18 (1.02–4.68)0.0462.05 (0.74–5.66)0.169Severe tricuspid valve regurgitation, n (%)2.39 (1.24–4.63)0.0102.57 (0.89–5.69)0.112Hypertension, n (%)1.55 (0.86–2.80)0.148Diabetes mellitus, n (%)1.08 (0.58–2.01)0.807Systolic blood pressure, mmHg1.00 (0.96–1.04)0.929Diastolic blood pressure, mmHg1.02 (0.97–1.07)0.399Fasting plasma glucose, mmol/L1.36 (1.04–1.79)0.0261.54 (0.97–2.13)0.066Smoking, n (%)0.79 (0.42–1.49)0.473Alcohol consumption, n (%)1.67 (0.92–3.03)0.093Mitral isthmus block achieved, n (%)1.01 (0.40–2.57)0.978Cavotricuspid isthmus block achieved, n (%)0.65 (0.16–2.68)0.550Left atrial CFAE, n (%)1.13 (0.48–2.66)0.784Right atrial CFAE, n (%)1.17 (0.58–2.36)0.664Superior vena cava, n (%)1.12 (0.52–2.40)0.775Coronary sinus, n (%)1.14 (0.63–2.04)0.665Ethanol ablation of the vein of Marshall, n (%)0.58 (0.21–1.61)0.295Switched to AFL, n (%)0.32 (0.14–0.71)0.0050.31 (0.13–0.70)0.005AFL ablation termination, n (%)0.28 (0.11–0.70)0.0060.31 (0.12–0.78)0.014Terminate AF, n (%)0.