Impact of previous left atrial ablation procedures on the mechanism of left atrial flutter: A single‐centre experience

Left atrial flutter predominantly occurs after surgical or ablation procedures but this entity has also been recently reported in individuals without previous interventions. The use of high‐density electroanatomical mapping‐systems (HDM) has improved the understanding of underlying mechanisms beyond entrainment maneuvers and substrate analyses. We aimed to evaluate the mechanism of left atrial (LA) flutters in preablated vs ablation‐naïve individuals and sought to assess the efficacy of empiric ablations sets in these groups.


| INTRODUCTION
Left atrial (LA) flutter is believed to be mainly attributed to cardiac interventions, like pulmonary vein isolation (PVI) or surgical valve repair. Less frequent, LA flutter occurs in patients without previous cardiac interventions, in those cases commonly associated with structural heart disease due to arterial hypertension, diabetes, coronary artery disease, or heart failure. 1 Despite macroreentry as the suspected mechanism in most cases of LA flutters it is unclear whether there are differences in arrhythmia mechanism in patients with and without a history of previous cardiac procedures. Recent advances in three-dimensional (3D) high-density mapping (HDM) systems enabling high-resolution maps of activation circuits have led to a profound understanding of the underlying arrhythmia mechanism beyond entrainment-mapping manoeuvers. 2 Therefore, with the increasing number of patients presenting with LA flutters we sought to identify the mechanism of the flutter form depending on a history of previous LA ablation procedures. As HDM is still not the widespread standard for flutter ablation procedures, we aimed to evaluate whether there are specific patients preferably being treated by HDM guidance for ablation of LA flutter.

| METHODS
We conducted a retrospective cohort study at the West German Heart and Vascular Center Essen, Germany. We identified 55 consecutive patients with any type of macroreentrant LA flutter admitted for radiofrequency ablation (RF) with an HDM system be-

| Procedure and mapping protocol
The details on procedure preparation and mapping are provided in the online supporting infomation. In brief, a 3D electroanatomical map of the LA with simultaneous local activation mapping was performed using the HDM system by two experienced operators (RW and JS) with an experience of more than 350 LA procedures each. The components of this HDM system have been described previously. 3,4 There was an intention to not perform entrainment mapping to prevent a potential early termination of the tachycardia or conversion into a different form during the pacing manoeuvers.
The goal was to assess the mechanism and potential reentry circuit by HDM mapping to consecutively target the critical isthmus by RF ablation to proof the suspected arrhythmia mechanism.

| Analysis of interobserver and intraobserver variability
For assessment of interobserver variability, the generated maps were reanalyzed by an independent physician (NV) with extensive experience in 3D mapping on an offline panel blinded to the initial result of the mapping process. Intraobserver variability was reassessed by the initial analyst (SK) by reanalyzing the HD maps after a mean period of 585 ± 408 days with respect to the dominant mechanism, also blinded to the initial result.

| Methodology of flutter classification
The first LA flutter form (present at the beginning of the procedure) was categorized into (a) fixed anatomical structure-related, including (i) perimitral flutter, (ii) roof-dependent flutter, (iii) flutter within the PVs, and (b) non-fixed anatomical structure-related flutter forms, comprising localized reentries involving low-voltage/slow conduction areas. All flutter forms occurring spontaneously or converting during RF delivery following the predefined ablation approach were documented. Each RF ablation was limited to 240 seconds, delivered by a 500 kHz ablation unit (Stockert EP shuttle; Biosense Webster, Inc, Baldwin Park, CA). The periprocedural anticoagluation regime has been previously described. 5 After analysis and identification of the mechanism of atrial flutter, RF energy was delivered according to the predefined protocol: After conversion, a detailed remap (see detailed mapping protocol, online Supporting Information) and reablation were performed until termination in sinus rhythm could be achieved or the procedure had to be stopped at the discretion of the physician (long duration, safety concerns). If rhythm degenerated into atrial fibrillation (AF) during RF delivery, electrical cardioversion was performed. Empiric PVI was only performed if there was clear documentation of AF before the procedure. In preablated patients, all PVs and in the ablation-naïve group only PVs that were involved in the tachycardia mechanism (eg, in case of within-PV flutter) were checked for complete isolation and reisolated in case of persistent atrio-PV conduction.

| Objectives and endpoints
The aim of the present study was to assess the impact of a LA preablation history on the dominant mechanism in LA flutters. Flutter forms were divided into "classical" fixed anatomic structure-related (e.g., mitral isthmus or roof-dependent) and non-anatomic structurerelated flutters (involving localized scar and low-voltage/slow conduction areas). In a second step, we sought to empirically assess the efficacy of predefined ablation sets in the two groups. Third, we analyzed the midterm arrhythmia recurrence between preablated and ablation-naïve patients which was defined as the occurrence of any atrial tachycardia during 720 days after the index procedure after a 90-day blanking interval.

| Evaluation of potential empirical linear ablation sets
Empirical linear ablation is sometimes used as an additional tool in AF ablation or in ablation of LA flutters in case of nonavailability of the HDM system, failed activation mapping and/or complex arrhythmia mechanism. 7,8 On the basis of the underlying mechanism of the respective tachycardia, we sought to assess whether one of three

| Clinical outcome
According to our local clinical routine, clinical follow-ups (FUs) were performed at 3, 6, and 12 months after PVI, and additionally in a 6-month interval if applicable, including physical examination, 12-lead ECG and a 7-day Holter-ECG. Any symptomatic atrial tachycardia greater than 30 seconds (AF, atrial flutter, focal tachycardia) on a Holter-ECG recording or on a 12-lead ECG was considered a recurrence. We applied the established 90-day blanking period with respect to early recurrences.
Palpitations without ECG documentation of AF were not considered a recurrence. Antiarrhythmic drugs were discontinued after 90 days of the blanking period. FU was censored at 2 years.

| Statistical analysis
Continuous variables are expressed as mean ± standard deviation. Normally distributed data were compared using the unpaired Student t test and nonparametric variables using the Mann-Whitney U test, respectively. Descriptive statistics assessing the accordance between the raters was performed for interobserver and intraobserver variability. A P < .05 was considered statistically significant. Statistical analyses were performed using SPSS for Windows (version 20.0; SPSS Inc, Chicago, IL).  Table 1 in detail.

| Procedural data including details of LA mapping process
A total of 55 activation maps was assessed for the initial flutter form.
Procedural parameters including procedure duration, LA (mapping) time, and cumulative RF time were not significantly different between preablated and ablation-naïve individuals (see Table 1).
3.2 | Mechanism of macroreentry depending on preexisting LA procedure  Our analysis revealed a significant difference between the two groups with respect to the underlying arrhythmia mechanism (P = .03). Looking at the reentries in detail, we found that perimitral reentry was the most common mechanism in preablated patients (14 of 39 patients, 35.9%), while the localized non-anatomical reentry was the most common form in non-preablated patients (9 of 16, 56.3%) ( Figure 2B). Video S1 (online Supporting Information) visualizes perimitral flutter in a preablated patient in a counter-clockwise rotation direction. Our results demonstrate a trend for a different pattern of the underlying reentry mechanism between the groups with respect to a preexisting history of a prior LA ablation ( Figure 2B; P = .07).
In ablation-naïve individuals, the majority of reentries (9 of 16, 56.3%) were not related to anatomically fixed structures, predominantly involving low-voltage areas at the anterior or posterior LA wall (Video S2, online Supporting Information).

| Clinical outcome
Recurrence rates were comparable between preablated and ablationnaïve individuals (35.9% vs 25%; P = .72) after 2 years of FU. Figure 3 depicts the Kaplan-Meier survival estimation for the two cohorts

| Safety
With respect to safety, one major stroke occurred due to throm-

| Underlying causes of atypical LA flutter
The incidence of LA flutters has spread exponentially which can be attributed to the increase in cardiac interventions (especially LA F I G U R E 3 Clinical outcome in preablated vs ablation-naïve patients: Midterm outcome was comparable between preablated and ablationnaïve individuals, with recurrence rates of 35.9% and 25% in preablated and non-preablated individuals, respectively (P = .72) F I G U R E 4 Hypothetical acute conversion/ termination rates by predefined ablation sets, stratified according to previous LA ablation procedure: Hypothetical acute conversion/ termination rates by predefined ablation sets, stratified according to previous LA ablation procedure. HDM, high-density mapping; other abbreviations as in Figure 1 ablations) in the last years. [10][11][12] With respect to AF ablation and its development of ablation strategies the reported incidence of LA flutters after AF ablation increased up to 30% 6 and even up to 50% in case of additional substrate modification for persistent AF. 13 In our present experience, we observed that nearly one-third (16 of 55; 29%) of patients with LA flutter did not undergo any LA ablation in the past. This is a high number when compared to previous reports where the rate of ablation-naïve patients in LA flutter cohorts was around 10%. 14 As reentrant flutter forms always require anatomic barriers with a protected isthmus region favoring slow conduction, critical sites for reentrant tachycardias 15 are suggested to be associated with atrial remodeling. Besides iatrogenic creation of those anatomic barriers during ablation or surgical procedures, they are especially found in atria of individuals with structural heart disease, like coronary or rheumatic heart disease, hypertrophic or dilated cardiomyopathy, or even in the absence of any heart disease. 10,16,17 This could be linked to a recently reported phenomenon called "atrial myopathy" which is In contrast to the preablated individuals, localized flutter forms unrelated to fixed anatomical reentries were much more common in

| Empirical linear ablation
As our acute termination rates were high, we assume that we actually correctly identified the critical isthmus in most patients. As we identified different mechanisms of LA flutters in preablated and ablation-naïve patients, we analyzed whether hypothetical empiric ablation lines would be able to terminate the distinct flutter form.
We postulate significant lower (hypothetical) success rates with linear ablation approaches in ablation-naïve patients as the mechanism in those patients was less likely macroreentry involving fixed anatomical structures. Interestingly, when deploying PVI, anterior and roof line, a termination/conversion success rate of more than 87% could be achieved in the preablated individuals, while only 63% of flutters would have been terminated/converted in the ablation-naïve patients.

| Relevance and potential clinical implications
On the basis of our results we hypothesize that linear ablation sets at standardized locations can terminate many LA flutter forms especially in preablated patients, but still do not reach success rates of SIEBERMAIR ET AL.
| 1637 HDM systems. In ablation-naïve patients a HDM-guided approach with identification of the individual tachycardia mechanism should be the preferred strategy since arrhythmia mechanisms are distinct and complex in these patients. As we did not perform empiric PVI in the ablation-naïve cohort except for PVs involved in the reentrant mechanism, there is a need for studies assessing the impact of an empiric PVI in those patients presenting with atrial flutter without a history of AF.

| LIMITATIONS
Although we present one of the largest studies on ablation outcomes in LA flutters the patient number is low, especially in the subgroup (n = 16) of ablation-naïve individuals. Additional verification of the arrhythmia mechanism (eg, by entrainment maneuvers) was not performed. Further, midterm outcome was evaluated by regular in-office FUs and Holter-ECG recordings. No continuous 24 hours telemetric ECG surveillance was performed to rule out asymptomatic arrhythmia recurrence.

| CONCLUSION
Our study provides evidence for distinct mechanisms of LA flutters in preablated vs ablation-naïve patients showing significantly higher rates of tachycardias involving low-voltage and slowconducting areas in ablation-naïve patients. Use of HDM-guided ablation resulted in very high acute success rates in both groups.
Nevertheless, based on different dominant mechanisms, linear ablation approaches seem to be still reasonable in preablated patients, while LA flutters in ablation-naïve individuals are probably less accessible to this approach highlighting the need to identify the specific individual tachycardia mechanism in these patients.