Multi-centre, prospective randomized comparison of three different substrate ablation strategies for persistent atrial fibrillation

Abstract

Aims

The optimal strategy for persistent atrial fibrillation (PerAF) is poorly defined. We conducted a multicentre, randomized, prospective trial to compare the outcomes of different ablation strategies for PerAF.

Methods and results

We enrolled 450 patients and randomly assigned them in a 1:1:1 ratio to undergo pulmonary vein isolation and subsequently undergo the following three different ablation strategies: anatomical guided ablation (ANAT group, n = 150), electrogram guided ablation (EGM group, n = 150), and extensive electro-anatomical guided ablation (EXT group, n = 150). The primary endpoint was freedom from atrial fibrillation (AF) lasting longer than 30 s at 12 months after a single ablation procedure. After 12 months of follow-up, 72% (108) of patients in the EXT group were free from AF recurrence, as compared with the 64% (96) in the EGM group (P = 0.116), and 54% (81) in the ANAT group (P = 0.002). The EXT group showed less AF/atrial tachycardia recurrence than the EGM group (60% vs. 50%, P = 0.064) and the ANAT group (60% vs. 37.3%, P < 0.001). The EXT group showed the highest rate of AF termination (66.7%), followed by 56.7% in the EGM group, and 20.7% in the ANAT group. The AF termination signified less AF recurrence at 12 months compared to patients without AF termination (30.1% vs. 42.7%, P = 0.008). Safety endpoints did not differ significantly between the three groups (P = 0.924).

Conclusions

Electro-anatomical guided ablation achieved the most favourable outcomes among the three ablation strategies. The AF termination is a reliable ablation endpoint.

Graphical Abstract

Graphical Abstract.

Extensive electro-anatomical guided ablation for persistent atrial fibrillation.

What's new?

• Electro-anatomical guided ablation achieved the highest rate of atrial fibrillation (AF) termination and the lowest rate of AF recurrence compared with anatomical guided ablation and electrogram guided ablation for persistent atrial fibrillation (PerAF).

• Ablation aiming at AF termination was associated with less AF recurrence.

• Extensive ablation targeting electro-anatomical mechanisms to pursue AF termination is the optimal strategy for PerAF.

Introduction

The optimal strategy for catheter ablation (CA) of persistent atrial fibrillation (PerAF) has puzzled electrophysiologists for decades. Although additional extra-pulmonary vein (PV) ablation could theoretically improve the outcomes of PerAF, studies seldom yielded positive results when testing the efficacy of different ablation strategies over the past 20 years.^1–4

The current three mainstream ablation strategies for PerAF include: (i) anatomical guided ablation based on linear ablation; (ii) electrogram guided ablation^5–7; (iii) a combination of electro-anatomical guided ablation represented by the ‘stepwise’ approach. Currently, there are still no large-scale, multicentre, randomized studies that directly compare the efficacies of the three ablation strategies, and ablation strategies remain uncertain in the present guidelines.^8,9 Furthermore, the ablation endpoints of different ablation strategies are inconsistent. Whether or not this inconsistency impacts the outcome of ablation remains controversial.¹⁰

Thus, we conducted this multicentre, randomized, prospective study to compare the following three ablation strategies for PerAF: anatomical guided ablation, electrogram guided ablation, and electro-anatomical guided ablation. We also investigated whether the termination of AF can serve as a reliable ablation endpoint.

Methods

Population

The study was designed as a prospective, randomized, multicentre study. All patients enrolled in the study provided written informed consent, and each site was assigned a sequential identification number. From January 2019 to January 2021, at the 10 investigational centres in China, 450 patients (see details of sample size estimation in Supplementary material online, Materials) with PerAF who received CA for the first time were enrolled and randomly assigned to three groups (in a 1:1:1 ratio), including:

• the anatomical guided ablation group (ANAT group)

• the electrogram guided ablation group (EGM group)

• the extensive electro-anatomical guided ablation group (EXT group).

Patients were included in the study if they were aged between 18 and 80 years, and non-responder or intolerant to ≥1 Class I or Class III antiarrhythmic drug and at the first CA. Patients were excluded if any contraindication to the procedure exists, including significant valvular disease and/or prosthetic heart valve(s), significant congenital heart disease, intracavitary thrombus, uncontrolled heart failure, severe pulmonary disease, and if they underwent any cardiac surgery within the past 2 months. The study flowchart is presented in Figure 1. The study complies with the ‘Helsinki Declaration’, and the local ethics committee has approved the protocol at each participating centre.

Figure 1.

Study flowchart. PVI, pulmonary vein isolation; ANAT group, the anatomical guided ablation group; EGM group, the electrogram guided ablation group; EXT group, the extensive electro-anatomical guided ablation group.

Electrophysiological study

A decapolar mapping catheter (Biosense Webster, Diamond Bar, CA, USA) was positioned in the coronary sinus (CS) via left femoral vein. Two SL1-type Swartz sheaths (St. Jude Medical, St. Paul, MN, USA) were advanced into the left atrium after two successful transseptal punctures. After transseptal catheterization, systemic anticoagulation was achieved with intravenous heparin (100 IU/kg) to maintain an activated clotting time between 300 and 350 s. Selective PV venography was performed to identify all PV ostia prior to ablation. PentaRay multispline catheter (Biosense Webster, Diamond Bar, CA, USA) was used as a navigational catheter.

Ablation protocol

Transesophageal echocardiography or intracardiac echocardiography was performed in all patients to rule out LA thrombus. Catheter ablation was performed under the guidance of a three-dimensional non-fluoroscopic mapping system (CARTO, Biosense Webster, CA, USA). THERMOCOOL SMARTTOUCH SF(STSF) catheter (Biosense Webster, Diamond Bar, CA, USA) and THERMOCOOL SMARTTOUCH(ST) catheter (Biosense Webster, Diamond Bar, CA, USA) were applied for ablation. Circumferential pulmonary vein isolation (PVI) was performed in all patients. The radiofrequency energy (RF) was 40–45 W (15–20 mL/min saline perfusion) with an ablation index of 400–450 during PVI. A minimum waiting time of 30 min was observed before the verification of PVI confirmed by both PentaRay catheter (Biosense Webster, Diamond Bar, CA, USA) and ablation catheter. Successful PVI was defined as eliminating all PV potentials and/or atrium-PV potential dissociation.

After PVI, localized AFCL was calculated manually by averaging 10 consecutive beats at three different times in the left atrial appendage (LAA), right atrial appendage (RAA) from bi-polar potentials of mapping catheters or proximal and distal electrodes of the CS catheter.

1. Following PVI, the protocol for the ANAT group patients included:

   1. Linear ablation : Linear lesions of the LA roof and posterior inferior wall were created to link the superior pole of the superior PVs and the inferior aspect of the inferior PVs, respectively, to achieve electrical LA posterior box isolation. Following that, a mitral isthmus line was created from both endocardium in LA (RF: 40–45 W, 15–20 mL/min saline perfusion) and epicardium in CS (RF:25 W, 25 mL/min saline perfusion) to achieve perimetral block.

   2. Vein of Marshall (VOM) ethanol infusion: If perimitral block was not achieved by linear ablation, we made an attempt on VOM ethanol infusion. Coronary sinus venography was performed to confirm the presence of VOM. If present, 1 cc of 98% ethanol was delivered over 2 min through an over-the-wire angioplasty balloon (size 1.5–2.5 mm by 6–8 mm). Following ethanol injection, the ethanol-induced scar was evaluated by a repeat voltage map.

2. The protocol for the EGM group patients included driver ablation.

   The methodology of EGM spatial-temporal analysis for driver mapping was described in our previous work.¹¹ The targeted ablation regions were regions in which EGMs displayed (more details in Supplementary material online, Materials):

   1. spatial-temporal dispersion activation, which spread over AFCL at a minimum of three adjacent bi-pole and associated with local AFCL ≤ mean AFCL

   2. locally short cycle length activity

   3. high-frequency potentials

   4. focal activity.

   The region with the shortest CL was considered as the region of the dominant driver and was ablated first. Point-by-point applications were performed (no dragging). At each RF application location, we sought a complete ‘flattening’ of the bipolar signal amplitude.

3. Patients in the EXT group received anatomical guided ablation following EGM guided ablation. If AF terminates during EGM guided ablation, the anatomical guided ablation should also be performed.

More details of electrophysiological study and ablation protocol were shown in the Supplementary material online, Materials.

The endpoint of ablation was AF termination, defined as conversion to sinus rhythm (SR) or a stable atrial flutter/tachycardia (AFL/AT). If AF converted to AT during the procedure, the arrhythmia was mapped and ablated until conversion to SR. If the AF cannot be terminated by ablation, 150–200 J direct current conversion was performed.

For redo procedure, the strategy applied in the second ablation was the same as that used in the first ablation.

Follow-up

Following the ablation procedure, all patients were hospitalized for at least 3 days, and cardiac rhythm was continuously monitored during the first 48 h. Anti-arrhythmic medications (amiodarone, dronedarone, or propafenone) were prescribed for 1–2 months after ablation and were discontinued five half-lives before the end of the first 3 months (blanking period). Any atrial arrhythmia occurring during the blanking period was not considered a recurrence. Outpatient visits and 48-h Holter monitoring were scheduled at 1, 3, 6, 9, and 12 months, and every 6 months thereafter if the patient remained asymptomatic. Monthly telephone interviews were also done. All patients were asked to undergo additional ECGs and 7-day Holter recordings when their symptoms were suggestive of tachycardia. A ‘recurrence’ of atrial arrhythmia was considered any episode lasting 30 s (symptomatic or asymptomatic) detected by ECG and/or Holter.

Study endpoint

The primary endpoint was defined as freedom from any AF episodes lasting more than 30 s after the blanking period without anti-arrhythmic drugs at 12 months after a single procedure.

Secondary outcomes included: (i) freedom from any documented AF/AT episode lasting more than 30 s after the blanking period without anti-arrhythmic drug treatment at 12 months after a single procedure; (ii) any documented AT episode lasting more than 30 s after the blanking period without anti-arrhythmic drug treatment at 12 months after a single procedure; (iii) freedom from any AF episodes lasting more than 30 s after the blanking period without anti-arrhythmic drugs after two ablation procedures; and (iv) periprocedural adverse events.

Statistical analysis

Continuous variables were expressed as mean ± standard deviation or median [interquartile range (IQR)] and compared using the Kruskal–Wallis test or ANOVA as appropriate, whereas categorical variables were expressed as percentages and compared using Pearson’s chi-square test. Survival curves were performed using the Kaplan–Meier method, and comparisons among the three groups were performed using the log-rank test with two degrees of freedom. All tests of significance were two-sided, with a probability <0.05 considered significant. All statistical analyses were performed using SAS version 9.1 (SAS Institute, Inc., Cary, NC) and GraphPad Prism7.0 software (GraphPad Software Inc, San Diego, CA, USA). More details are found in the Supplementary material online, Materials.

Results

Baseline characteristics

A total of 450 patients were enrolled in the study from January 2019 to January 2020. Nine patients dropped out before the 12-month follow-up, which included two patients in the ANAT group, four in the EGM group, and three in the EXT group (Figure 1). Four patients did not end up ablation because of severe adverse events during the procedure, including one pericardial effusion in the ANAT group, one acute heart failure in the EGM group, and two pericardial effusion in the EXT group. Therefore, a total of 437 patients were available for as-treated analysis, which included 147 patients in the ANAT group, 145 in the EGM group, and 145 in the EXT group (see Supplementary material online, Table S1).

The baseline clinical characteristics of the study population (intention-to-treat analysis) are listed in Table 1. Two hundred and twenty-four (50.8%) patients were presented as longstanding AF. There were no significant differences among the three groups.

Table 1.

Baseline characteristics (ITT)

	All patients (n = 450)	ANAT group (n = 150)	EGM group (n = 150)	EXT group (n = 150)	P value
Age (years)	65.8 ± 9.4	65.2 ± 9.1	66.1 ± 9.2	66.1 ± 9.9	0.664
Male, n (%)	310 (68.9)	105 (70.0)	98 (65.8)	107 (71.3)	0.499
AF duration, months	12 (1–48)	12 (2–60)	7.5 (1–39)	12 (1–39)	0.513*
Longstanding AF	229 (50.9)	80 (53.3)	73 (48.7)	76 (50.7)	0.345
CHADS2-VAS score	2.2 ± 1.4	2.2 ± 1.5	2.1 ± 1.4	2.4 ± 1.4	0.272
Comorbidity, n (%)
Hypertension	246 (54.7)	80 (53.3)	78 (52.0)	88 (58.7)	0.471
Diabetes mellitus	74 (16.4)	27 (18.0)	20 (13.3)	27 (18.0)	0.453
Prior stroke	61 (13.6)	22 (14.7)	17 (11.3)	22 (14.7)	0.622
Coronary artery disease	54 (12.0)	18 (12.0)	16 (10.7)	20 (13.3)	0.777
Heart failure	48 (10.7)	14 (9.3)	15 (10.0)	19 (12.7)	0.613
Concomitant medications, n (%)
Amiodarone	419 (93.1)	141 (94.0)	137 (91.3)	141 (94.0)	0.574
Beta-blocker	166 (36.9)	53 (35.3)	61 (41.7)	52 (34.7)	0.498
Echocardiography parameters
LAD (mm)	44.0 ± 5.6	44.4 ± 5.8	43.5 ± 5.7	44.1 ± 5.3	0.323
LVEF (%)	60.9 ± 8.5	60.3 ± 9.5	62.0 ± 7.6	60.4 ± 8.2	0.150

AF, atrial fibrillation; BMI, body mass index; LAD, left atrial diameter; LVEF, left ventricle ejection fraction; *Kruskal–Wallis test.

Procedural characteristics

The PVI was successfully performed in all patients. For the ANAT group, all patients received LA posterior box isolation and mitral isthmus linear ablation. The rate of complete LA posterior wall electrical isolation was 78% (117/150) and the rate of successful bidirectional block of the mitral line was 56% (84/150). Sixty-four (42.7%) patients successfully completed VOM ethanol infusion, and the rate of successful bidirectional block of the mitral line increased to 78% (117/150) after VOM ethanol infusion.

All patients in the EGM group and the EXT group completed EGM guided ablation. A total of 996 (3.4 regions per patient) targeted EGM regions were ablated, which were most commonly located on the LA roof (23.5%), followed by 19.6% EGM regions in the bottom, 15.1% in the posterior LA wall, 11.1% in the anterior wall, 9.3% in the LA appendage basal, 8.4% in the left septum, and 13% in the right atrial free wall and right septum (Figure 2). Targeted regions with EGMs that displayed short cycle length activity were the most frequent (45%), followed by high-frequency potentials (24%), spatio-temporal potentials (20%), and focal activity (11%). The mean areas of targeted EGM ablation regions were 19.6 ± 9.0%, and were comparable between the EGM group and the EXT group (17.8 ± 9.5% vs. 18.5 ± 8.4%, P = 0.527).

Figure 2.

Distribution of targeted EGM regions in the EGM group and the EXT group. LA, left atrium; RA, right atrium; EGM, electrogram.

In the EXT group, complete LA posterior wall electrical isolation was achieved in 75.3% (113/150) patients, and bidirectional block of the mitral line was achieved in 57.3% (86/150) patients. Sixty-two (41.3%) patients received VOM ethanol infusion, and the rate of bidirectional block of the mitral line increased to 80% (120/150) after VOM ethanol infusion.

Acute AF termination during the procedure was observed in 216(48%) patients, among them, 80 (37%) converted to SR directly and 136 (63%) converted to AFL/AT. Both the EXT group (66.7% vs. 20.7%, P < 0.001) and the EGM group (56.7% vs. 20.7%, P < 0.001) showed a significantly higher rate of AF termination compared to the ANAT group. There was also a trend for a higher rate of AF termination in the EXT group compared with the EGM group (P = 0.075, Table 2).

Table 2.

Comparisons of study endpoints (ITT)

	ANAT group (n = 150, n (%)	EGM group (n = 150), n (%)	EXT group (n = 150), n (%)	P value	EXT vs. ANAT	EXT vs. EGM	EGM vs. ANAT
Primary endpoint
Freedom from AF recurrence at 12 months	81 (54)	96 (64)	108 (72)	Log-rank	0.002	0.116	0.136
				Absolute percentage difference (95% CI)	0.18 (0.071–0.283)	0.08 (−0.026–0.183)	0.1 (−0.012–0.208)
				Hazard rate (95% CI)	0.562 (0.387–0.815)	0.734 (0.492–1.096)	0.774 (0.543–1.101)
Secondary endpoints
Freedom from AF/AT recurrence at 12 months	56 (37.3)	75 (50)	90 (60)	Log-rank	<0.001	0.064	0.049
				Absolute percentage difference (95% CI)	0.227 (0.114–0.332)	0.1 (−0.012–0.209)	0.127 (0.015–0.235)
				Hazard rate (95% CI)	0.557 (0.406–0.764)	0.739 (0.528–1.036)	0.755 (0.559–1.021)
Documented AT recurrence at 12 months	27 (18)	25 (16.7)	21 (14)	Log-rank	0.33	0.524	0.728
				Absolute percentage difference (95% CI)	0.04 (−0.044–0.123)	0.027 (−0.056–0.109)	0.013 (−0.073–0.099)
				Hazard rate (95% CI)	0.759 (0.431–1.336)	0.831 (0.466–1.482)	0.910 (0.529–1.568)
Freedom from AF recurrence after second procedure	90 (60)	109 (72.6)	120 (80)	Log-rank	<0.001	0.112	0.035
				Absolute percentage difference (95% CI)	0.2 (0.097–0.298)	0.073 (−0.023–0.168)	0.127 (0.020–0.230)
				Hazard rate (95% CI)	0.453 (0.300–0.685)	0.692 (0.435–1.102)	0.665 (0.451–0.983)
Procedure endpoint
Conversion to AT or SR	31 (20.7)	85 (56.7)	100 (66.7)	Log-rank	<0.001	0.075	<0.001
				Absolute percentage difference (95% CI)	0.46 (0.354–0.551)	0.1 (−0.010–0.207)	0.36 (0.253–0.456)
				Hazard rate (95% CI)	2.380 (1.871–3.027)	1.300 (0.972–1.739)	1.831 (1.498–2.237)
Safety endpoint
Adverse events	4 (2.7)	4 (2.7)	5 (3.3)	Log-rank	1	1	1
				Absolute percentage difference (95% CI)	0.007 (−0.038–0.052)	0.007 (−0.038–0.052)	0 (−0.043–0.043)
				Hazard rate (95% CI)	1.007 (0.968–1.048)	1.007 (0.968–1.048)	1.000 (0.963–1.038)

AF, atrial fibrillation; ANAT group, the anatomical guided ablation group; EGM group, the electrogram guided ablation group; EXT group, the extensive electro-anatomical guided ablation group.

The mean total procedure duration and fluoroscopy time was significantly shorter in the EGM group compared with the ANAT group (183.8 ± 35.1 vs. 200.4 ± 24.7 min, P < 0.001; 7.4 ± 1.3 vs. 9.9 ± 4.1 min, P < 0.001) and the EXT group (183.8 ± 35.1 vs. 207.6 ± 29.6 min, P < 0.001; 7.4 ± 1.3 vs. 10.1 ± 4.1 min, P < 0.001). The EXT group had a longer procedural time than the ANAT group (P = 0.023), while had a similar fluoroscopy time to the ANAT group (P = 0.693).

Study outcomes

The results of the study outcomes (intention-to-treat analysis) were presented in Table 2. At the end of the 12-month follow-up, freedom from any AF episodes lasting more than 30 s after the blanking period without anti-arrhythmic drugs was achieved in 81 of 150 patients (54%) in the ANAT group, 96 of 150 (64%) in the EGM group, and 108 of 150 (72%) in the EXT group, respectively.

As shown in Figure 3A, the Kaplan–Meier survival analysis revealed that the EXT group exhibited less AF recurrence at the end of 12 months of follow-up than the ANAT group [log-rank P = 0.002; HR 0.562 (95% CI, 0.387–0.815), with an absolute difference of 18% (95% CI, 7.1–28.3%)]. The EGM group also showed less AF recurrence compared with the ANAT group, but the result did not reach a statistical difference (P = 0.136). There also existed a trend for less AF recurrence in the EXT group compared with the EGM group (P = 0.116).

Figure 3.

Kaplan–Meier (K–M) survival curves of study outcomes among the three groups. (A) Comparison of primary outcome among the three groups. (B) Comparison of secondary outcome among the three groups. AF, atrial fibrillation; ANAT group, the anatomical guided ablation group; EGM group, the electrogram guided ablation group; EXT group, the extensive electro-anatomical guided ablation group.

After excluding four patients who did not end up ablation and nine patients who were lost of follow up (as-treated analysis), the primary outcome was reached in 81 of 147 patients (55.1%) in the ANAT group, 96 of 145(66.2%) in the EGM group, and 108 of 145(74.5%) in the EXT group, respectively. Similarly, the EXT group also showed better SR maintenance than both the ANAT group [log-rank P = 0.001; HR 0.517 (95% CI, 0.351–0.761)] and the EGM group [log-rank P = 0.09; HR 0.708 (95% CI, 0.464–1.025)] (see Supplementary material online, Table S1).

For the secondary endpoint, the proportion of any documented AT recurrence did not differ significantly among the three groups both in the intention-to-treat analysis (P = 0.633) and the as-treated analysis (P = 0.458). However, both the EXT group [log-rank P < 0.001; HR 0.557 (95% CI, 0.406–0.764)] and the EGM group [log-rank P = 0.049; HR 0.755 (95% CI, 0.559–1.021)] had less AF/AT recurrence at 12 months after a single procedure than the ANAT group (Figure 3B). Similar results were also observed in the as-treated analysis (P < 0.001 for EXT vs. ANAT, P = 0.033 for EGM vs. ANAT) (see Supplementary material online, Table S1).

Among 156 patients who experience AF recurrence, 55 patients underwent a redo procedure (16 in the ANAT group, 20 in the EGM group, and 19 in the EXT group). The PV reconnection was documented in 26 (47%) patients, 35 patients received linear ablation, 14 (40%) had roofline reconnection, and 9 (25.7%) had mitral isthmus reconnection. After a second procedure, both the EXT group (20% vs. 40%, P < 0.001) and the EGM group (27.4% vs. 40%, P = 0.035) showed less AF recurrence than the ANAT group.

Relationship between termination and clinical outcomes

The comparison of primary and secondary outcomes between patients with and without AF termination was presented in Table 3. For the entire study cohort, patients with AF termination exhibited less AF recurrence [30.1% vs. 42.7%, P = 0.008, HR 0.669 (95% CI, 0.493–0.908), with an absolute difference of 12.6% (95% CI, 3.7–21.2%)] and less AF/AT recurrence [42.6% vs. 58.5%, P = 0.001, HR 0.657 (95% CI, 0.507–0.852), with an absolute difference of 16% (95% CI, 6.7–24.8%)] at the end of 12 months follow-up after a single ablation than patients without AF termination (Figure4A and B). There was no difference for AT recurrence [13.9% vs. 18.4%, P = 0.179, HR 0.733 (95% CI, 0.463–1.159)] between patients with AF termination and patients without AF termination. Similar results were also observed in the as-treated analysis (see Supplementary material online, Table S2).

Table 3.

Termination vs. not termination study outcome (ITT)

	AF termination (n = 216)	Without AF termination (n = 234)	Absolute percentage difference (95% CI)	Hazard rate (95% CI)	P value
AF recurrence at 12 months	65 (30.1)	100 (42.7)	0.126 (0.037–0.212)	0.669 (0.493–0.908)	0.008
AF/AT recurrence at 12 months	92 (42.6)	137 (58.5)	0.160 (0.067–0.248)	0.657 (0.507–0.852)	0.001
Documented AT recurrence at 12 months	30 (13.9)	43 (18.4)	0.045 (−0.024–0.113)	0.733 (0.463–1.159)	0.179

AF, atrial fibrillation; AT, atrial tachycardia; ANAT group, the anatomical guided ablation group; EGM group, the electrogram guided ablation group; EXT group, the extensive electro-anatomical guided ablation group.

Figure 4.

Kaplan–Meier (K–M) survival curves of study outcomes between patients with and without AF termination. (A) Comparison of primary outcome between patients with and without AF termination. (B) Comparison of secondary outcome between patients with and without AF termination. AF, atrial fibrillation; AT, atrial tachycardia.

Safety outcomes

Adverse events occurred in four (2.7%) patients in the ANAT group, four (2.7%) patients in the EGM group, and five (3.3%) patients in the EXT group (Table 4). There was no difference with regard to safety outcomes among the three groups (P = 0.924). The most common adverse events were vascular complications including four events of haematoma at the access site, one arteriovenous fistula, and one pseudo-aneurysm. Serious adverse events include one transient ischaemic attack, one stroke, one acute heart failure, and four pericardial effusions (one after ablation) and were treated conventionally with no long-term sequelae. No death, PV stenosis, or atrial esophageal fistula occurred.

Table 4.

Adverse events

	ANAT group (n = 150)	EGM group (n = 150)	EXT group (n = 150)
Overall adverse events	4 (2.7)	4 (2.7)	5 (3.3)
Pericardial effusion	1	1	2
acute heart failure	0	1	0
Death	0	0	0
Pericarditis	0	0	0
Pneumonia <1 month after ablation	0	0	0
Transient ischaemic attack or stroke	1	0	1
Vascular complications	2	2	2

There were no significant differences among the three groups.

ANAT group, the anatomical guided ablation group; EGM group, the electrogram guided ablation group; EXT group, the extensive electro-anatomical guided ablation group.

Discussion

Main findings

In this multicentre, prospective, randomized study, we classified ablation methods into anatomical guided ablation, EGM guided ablation, and a combination of electro-anatomical guided ablation. We found that, among the three strategies, electro-anatomical guided ablation achieved the highest rate of AF termination and the lowest rate of AF recurrence despite a longer procedure time being required. Furthermore, AF termination during ablation was associated with less AF recurrence. Based on our findings, we concluded that extensive ablation targeting electro-anatomical mechanisms to pursue AF termination is the optimal strategy for PerAF.

Anatomical strategies for persistent atrial fibrillation

The aim of anatomical guided ablation is to target anatomical structures that are critical to the fibrillatory process. Although the STAR AF II study failed to demonstrate the superiority of PVI + linear ablation over PVI alone, in the VENUS study, additional VOM ethanol infusion increased the success rate of single ablation by 11.2%, which might be related to more durable peri-mitral block and atrial denervation.⁴ In our study, the success rate of the ANAT group was higher than that in the STAR AF II study, and lower than that in the VENUS study. This may be due to the beneficial effect of the VOM ethanol infusion, as only 42.6% of patients in the ANAT group received the VOM ethanol infusion. These results indicate that enhanced anatomical ablation can provide long-term benefits. Although simplified linear ablation approach has also achieved a good prognosis in some small sample studies,^12,13 these studies also targeted VOM for ablation. The efficacy of simple anatomical ablation was still inferior to that of the ‘stepwise’ approach. There are several limitations of anatomical ablation, which include (i) its inability to address the maintenance mechanism of PerAF; (ii) the fact that it cannot be patient-tailored; and (iii) its unsatisfactory efficacy for patients who underwent repeat procedures.^1,14

Targeting substrate sustaining persistent atrial fibrillation

Recently, the EGM guided driver ablation serves as an alternative procedural method for PerAF, since basic research has developed the new concept that AF is mechanistically maintained by several localized drivers.^7,15 The targeted EGM regions displayed spatial-temporal dispersion activation, high-frequency potentials, or locally short cycle lengths activity, which are the hallmark of critical driver regions that sustain AF.^5,6 Nademanee et al.¹⁶ reported the first attempt to apply EGM guided driver ablation in 2004. They targeted regions characterized by complex fractionated atrial electrograms (CAFEs). However, subsequent studies did not consistently show the effectiveness of CAFEs ablation,⁵ which could be because the majority of CAFEs resulted from wavefront collisions rather than from true drivers of AF.

After that, algorithm and mapping technology that advances continuously optimized the accuracy of targeted EGM identification. Convincing data have indicated that the ablation of targeted EGM regions acutely terminated AF and/or prolonged AFCL, suggesting an intervention of areas that are critical to sustaining AF.^5,17 Therefore, EGM guided ablation is an accurate strategy with fewer ablation lesions. However, EGM guided ablation also has several limitations, which include the absence of a uniform definition of targeted EGM and a standard mapping method, a steep learning curve, and AT recurrence. Clinical rotor identification might be substantially influenced by (i) non-identical surface activation patterns, which resulted from a diverse 3D form of scroll wave, and (ii) inadequate resolution of mapping techniques¹⁸ and the targeted EGM regions are usually operator-dependent. Despite these problems, the methodology of targeted EGM identification has been validated in prior works done by us and other groups.¹¹

However, recent small and single-centre studies that applied novel mapping technologies have confirmed the superiority of EGM guided ablation over anatomical guided ablation.^6,19 These studies are subjected to small sample sizes or limitations of single-centre studies. Therefore, we conducted this large-scale, prospective, multicentre study to provide evidence of the incremental value of EGM guided ablation. In this study, we found that EGM guided driver ablation had a shorter procedural duration while achieving a higher rate of AF termination and better outcomes than anatomical guided ablation. It is noteworthy that rotational drivers had the propensity to be anchored to regions of fibrosis, suggesting that EGM guided ablation may also modify structural substrates.¹⁵ This may also explain, to some extent, that persistent AF patients who underwent low-voltage area substrate modification could improve freedom from arrhythmia.²⁰

Electro-anatomical guided ablation strategy

The ‘stepwise’ approach, as a classic electro-anatomical guided ablation strategy, has a commendable arrhythmia-free survival rate (90% at 1 year after multiple ablations) and procedural AF termination rate (up to 80%), suggesting the necessity of extended ablation. However, the ‘stepwise’ approach may create unnecessary proarrhythmic scars, resulting in a high risk of AT recurrence (up to 53%).

Compared with the ‘stepwise’ approach, the electro-anatomical guided ablation applied in our study simplified the steps of linear ablation.²¹

In our study, the acute AF termination rate was 67.3% in the EXT group, and 73.6% of patients were free of AF after a single procedure at 12 months of follow-up. These numbers were the highest among the three ablation strategies, and there was difference in AT recurrence between the three groups. We also found that extensive ablation was not associated with an increased risk of periprocedural adverse events.

In a recent study by Natale’s group, the extensive ablation strategy showed a significantly higher success rate than the limited ablation strategy, and the atrial scar resulting from extensive ablation did not increase the risk of stroke.²² Therefore, we concluded that the ablation strategy addressing both electrical and anatomical substrates is currently the optimal procedure method for PerAF.

Termination of atrial fibrillation as optimal procedure endpoint

As early as 2009, the optimal clinical endpoint of AF termination was confirmed via a stepwise procedure. Haissaguerre et al. indicated that termination of AF via a stepwise ablation strategy is associated with less AF recurrence during a 5-year follow-up period.²³ From a clinical point of view, the termination of AF should be the hard endpoint of clinical treatment and the result of AF substrate elimination, which is reflected in more and more studies on EGM guided ablation. In the present study, the termination of AF led to favourable outcomes. The EGM guided ablation achieved 41% of acute AF termination, and this proportion increased to 51% when anatomical guided ablation was combined with EGM guided ablation. These results suggested that the elimination of the AF substrate by extensive ablation was the most effective strategy. Similarly, the recent TARGET-AF 1 trial provided new confirmation of the strong relationship between AF termination and SR maintenance, when the localized driver was ablated.⁶ Since the localized driver is critical to AF maintenance, it is plausible that only by eliminating the maintenance mechanism of PerAF can AF termination be translated into a desirable outcome.

Study limitations

This study has several limitations. The follow-up period was short, only 12 months; therefore, we could not identify late AF/AT recurrences. We might underestimate some asymptomatic AF episodes without the use of implantable event recorders.^24,25 We did not set up a control group in which PVI was the only therapy, since, in our experience, PVI alone could not adequately eliminate PerAF.

Conclusions

An ablation strategy addressing electro-anatomical substrates is the current optimal ablation method for PerAF, despite the longer procedural time involved. The AF termination was a reliable ablation endpoint. (Optimization of intervention strategies for persistent atrial fibrillation: ChiCTR2200060075).

Supplementary material

Supplementary material is available at Europace online.

Funding

This research was supported by National Natural Science Foundation of China Grants (Grant No: 81770324) and National Key Research and Development Project (Grant Number 2018YFC1312503).

Data availability

The data underlying this article will be shared on a reasonable request to the corresponding author.