Mitral isthmus block is associated with favorable outcomes after reablation for long‐standing persistent atrial fibrillation

Abstract Background Mitral isthmus (MI) ablation was limited due to technical challenges in the index ablation for long‐standing persistent atrial fibrillation (LPeAF). The role of adjunctive MI ablation was controversial. Hypothesis MI block could be achieved in most patients undergoing repeat LPeAF ablation and was associated with favorable clinical outcomes. Methods Of 87 consecutively patients undergoing reablation for recurrent atrial tachyarrhythmias (ATa), 41 patients with residual MI conduction but without pulmonary vein reconnection or left atrial roof conduction were enrolled to treat recurrent atrial flutter (AFL) (n = 20) and AF (n = 21). After AFL ablation and AF cardioversion, MI conduction gaps (CGs) were mapped and closed. Results MI line was successfully blocked in 37 (90.2%) of 41 patients after closing 1.4 ± 0.5 CGs (31 endocardial CGs and 16 epicardial ones) in the initial MI lines. CGs were more often located at the endocardial sites close to the lateral ridge between left atrial appendage and left‐sided PVs, midportion of MI and at the epicardial breakthroughs within coronary sinus. At the end of 16.0 ± 1.9 months' follow‐up, 31 (83.8%) of 37 patients with MI block and 1 of 4 patients without MI block were free of further recurrence of ATa off anti‐arrhythmic drugs. MI block was positively associated with ATa‐free survival by Cox's regression analysis (hazard ratio [HR]: 0.012, 95% confidence interval [CI]: 0.000‐0.456, P = .02). Conclusions MI block could be achieved in the majority of patients during repeat ablation for LPeAF. MI block was associated with favorable clinical outcomes after LPeAF reablation.

atrial electrograms (CFAEs) ablation, and rotor ablation in LPeAF ablation is controversial. [2][3][4][5] The Cox Maze III procedure, by creating multiple incisions in both atria, is effective for treating lone AF or AF concomitant with valve replacement and other cardiac surgery. 6,7 It seems a "gap" existed between the efficacy of surgical incision lines and percutaneous ablation lines. One resonable explanation was the relatively low rate of line block in percutaneous ablation, since incomplete linear ablation is proarrhythmic. 8,9 The efficacy of catheter ablation for LPeAF might be improved if the lesion lines could be blocked.
It might be difficult to achieve a higher than usual rate of line block in the index procedures since aggressive or epicardial ablation might increase the risks of severe complications. However, in reablation it might be easier to achieve line block by closing gaps in initial ablation lines. In this study we hypothesized that mitral isthmus (MI) block could be achieved in the majority of redo LPeAF cases and MI block was associated with favorable outcomes.

| Patient population
From Jan 2018 to August 2018 patients undergoing reblation for recurrent atrial tachyarrhythmias (ATa) after initial catheter ablation of LPeAF were included retrospectively. LPeAF was defined as AF lasting for ≥1 year. The index ablation approach was circumferential PV isolation (CPVI) and linear ablation at left atrial (LA) roof and MI. Patients were eligible for enrollment in this study if they met the inclusion criteria: recurrent ATa after 3 months postablation; failure or intolerance to at least one antiarrhythmic drug (AAD); presence of residual MI conduction in reablation. The exclusion criteria were: LA thrombus detected by transesophageal echocardiography (TEE); PV reconnection and/or residual LA roof conduction identified in reablation. This study was in accordance with the Declaration of Helsinki and was approved by the Institutional Ethnics Committee. All patients provided written informed consent.

| Electrophysiological study
The reablation procedure was performed under conscious analgesia and sedation with continuous infusion of fentanyl and midazolam. All AADs were withdrawn for at least five half-lives except amiodarone.

| AFL ablation and AF cardioversion
In patients with AFL, activation mapping and multiple pacing entrainment was performed to determine the reentrant circuit and the key isthmus. The entrainment site with the postpacing interval (PPI) < 20 ms longer than the AFL cycle length was deemed within the reentrant circuit. 9 Linear ablation was applied at the isthmus to achieve AFL termination. After AFL termination, pacing maneuvers were applied to evaluate the integrity of the ablation lines.
In patients with AF, direct current cardioversion was performed to restore normal sinus rhythm (NSR). from CSd pacing to the ablation catheter was longer than that from CS medium bipole pacing to the ablation catheter 10 ( Figure 1D).

| Exclusion of PV reconnection and LA roof conduction recovery
PV reconnection was ruled out by absence of PV potentials or dissociation of PV potentials with the atrial electrograms 11 in ongoing ATa, and was further excluded by both PV entrance and exit block in NSR. In NSR, LA roof conduction recovery was excluded if: (a) the activation sequence at the left atrial posterior wall (LAPW) was from caudal to cranial in NSR and during left atrial appendage (LAA) pacing; (b) the interval from LAA pacing to high LAPW was longer than that from LAA pacing to medium and low LAPW; (c) the interval from high LAPW pacing to LAA was longer than that from medium and low LAPW pacing to LAA. 12,13 2.6 | Identification of conduction gaps in MI line  Figure 1F).

| RF ablation settings
All ablation lesions were created by the contact-force sensing salineirrigated catheter. For endocardial MI or LA roof ablation, radiofrequency (RF) energy was delivered at 35 to 40 W, 43 C and saline irrigation speed 20 mL/min, with the target ablation index (AI) 500 to 550. For epicardial ablation within the CS, RF energy was delivered at 25 W, 43 C and saline irrigation speed of 30 to 40 mL/min, with the duration of 30 seconds for each lesion.

| Postablation management
All patients were kept on therapeutic oral anticoagulation for at least 3 months. NOACs (dabigatran or rivaroxaban) were preferred to warfarin unless they were contraindicated. Anticoagulation therapy were continued in patients experiencing further recurrence or in those free of recurrence but with high risk of thromboembolism (CHA2DS2VASc score >1 for male, or >2 for female), and could be withdrawn in patients free of recurrence and with low risk of thromboembolism.
Amiodarone was continued for 3 months for all patients and was withdrawn if no further recurrence was detected, but could be continued otherwise. Oral proton-pump inhibitors were administered for 2 months to reduce the probability of esophagus injury.

| Follow-up
The first 3 months postablation were defined as the blanking period of this study. Recurrences within the blanking period were treated by AADs or DC cardioversion. After the blanking period, the patients were followed up at the out-patient clinic at 3, 6, 12, 18, and 22 months.
During each clinic visit, ECG and 24-hours Holter monitoring were performed to detect events of recurrence. The patients were asked to record their ECGs when there were symptoms indicating a recurrence.
After the blanking period, the clinical success was defined as freedom from documented ATa recurrence (>30s) off AADs.

| Data expression and statistical analysis
Continuous data were expressed as mean ± SD (SD) and category data as counts or proportions (%). Continuous data were compared by   Table 1.

| Comparison between patients with recurrent AFL and AF
There was no significant difference in proportion of male, LAD, LV dimension, and LVEF between patients with AFL and AF recurrence.
However, the average age was older and LAA blood flow velocity was faster in patients with AFL recurrence than in patients with AF recurrence (Table S1).    Table 2. The anatomic distribution of MI CGs was illustrated in Figure 3A. CGs were more often located at the F I G U R E 2 Patients' enrollment in this study

| Complications
No major complications (ie, tamponade, stroke, or atrio-esophagus fistula) occurred. Coronary angiography following CS ablation showed no evidence of coronary artery injury.  , which is shown in gray color. B, Comparison of ATa-free survival probability following reablation between the patients with AFL and AF recurrence after the index procedure (Log rank test, P = .72). Abbreviations seen in Figure 1 After reablation, there was no significant difference in ATa-free survival probability between patients with recurrent AFL and those with recurrent AF (80.0% vs 76.2%, Log rank test, P = .72, Figure 3B).

| Follow-up data
Of the variables including age, AF duration, MI block and ultrasound parameters (LAD, LVEDD, LVEF, and LAA velocity), MI block was the only factor positively associated with ATa-free survival (Hazard ratio: 0.012, 95% confidence interval: 0.000 to 0.456, P = .02, Table S2).

| DISCUSSION
The major findings of our study were: (a) MI block could be achieved in the majority of redo cases despite block failure in initial ablation; (b) There was anatomic propensity of the conduction gaps in initial MI lines: the endocardial sites close to the lateral ridge between LAA and LPV, midportion of MI line and epicardial breakthroughs within CS; (c) MI block was associated with favorable clinical outcomes of LPeAF reablation.

| Lessons learned from surgical Maze procedure
In the classic Maze III procedure, the "cut and sew" technique was utilized to create multiple incision lines, including circumferential PV incision, incision to the mitral annulus, LAA excision, incision from left-sided PV to LAA, right atrial appendage excision, right atrial anterior incision, posterior incision to tricuspid annulus (TA), lateral incision from superior vena cava to inferior vena cava and adjunctive cryolesion from mitral annulus (MA) to CS. 6 The Maze incisions were devised to create atrial compartmentation preventing multiple wavelet reentries, and yielded a good long-term success rate. 6,7 Bi-atrial Maze was more effective than stand-alone PVI, 15 indicating the importance of AF substrate modification by linear incisions.
Because the "cut and sew" technique was associated with elevated mortality, alternative energies (RF, cryo-ablation, and microwave) were applied. However, non-"cut and sew" Maze was significantly less effective for durable block lines creation as well as NSR maintenance compared with "cut and sew" Maze, 16

| Improving MI line block by closing conduction gaps in reablation
Incomplete ablation usually resulted in tissue edema and swelling. 19 The increased tissue thickness might prevent further energy penetration and resulted in failure of MI block. In the areas with incomplete ablation, tissue edema and swelling subsided after 2 to 3 months, 19 and the tissue surviving the index ablation might form the CGs.

| The impact of MI block on clinical outcomes of LPeAF ablation
The impact of MI block on clinical outcomes remained debatable.
Willems and his colleagues 2 found that linear ablation improved the success rate of persistent AF ablation. While linear ablation did not improve AF-free survival in STAR AF II trial. 3 To be noted, the success rate of linear ablation in the latter study was less than 80%. Due to the proarrhythmic property of incomplete ablation, 9 the beneficial effect of linear ablation might be neutralized by its proarrhythmic effect. The poorer effectiveness of non-"cut and sew" Maze in contrast to "cut and sew" Maze also supported the importance of line integrity. 15,16 In our study, we found that MI block was exclusively and positively associated with ATa-free survival after analysis of multiple variables, which indicating that MI block was associated with favorable clinical outcomes after reablation of LPeAF.

| Limitations
This study was a retrospective, preliminary clinical study and enrolled