Outcomes following persistent atrial fibrillation ablation using localized sources identified with Ripple map

Abstract Background Ablation of persistent atrial fibrillation (AF) remains challenging. Identification and ablation of localized AF drivers may offer the possibility for improved outcomes. Ripple map is a novel software algorithm that may allow improved localization of possible AF drivers through the whole chamber graphical display of continuously recorded bipolar electrograms. The objective of this study was to determine whether regions of high‐frequency Ripple activation (HFRA) observed on Ripple map provide useful ablation targets in patients with persistent AF. Methods and Results Consecutive patients underwent the first‐time ablation of persistent AF (n = 162) using a standard stepwise (n = 105) or a Ripple map guided approach (n = 57). Ripple map guided patients underwent pulmonary vein antral isolation followed by ablation of HFRA sites. Acute termination of AF was observed in 91.2% of the Ripple‐guided patients vs 52.4% in the stepwise approach, P < .0001. Following a single ablation procedure, after 18 months 98.2% of Ripple map guided patients were free of AF, compared with 81.4% of standard stepwise ablation (P = .005). Freedom from atrial tachycardia (54.4% Ripple map vs 52.4% standard, P = .9) or any atrial arrhythmia (52.6% Ripple map vs 39.0% standard, P = .10) did not differ between the two strategies. In a subset analysis (n = 30 of 56), Ripple map regions corresponded to sites with spatiotemporal dispersion in all atrial locations. No differences were observed in the rate of procedural complications. Conclusions Ablation of HFRA sites identified with Ripple map resulted in a higher rate of acute termination and improved freedom from AF compared to a standard stepwise approach.


| BACKGROUND
Pulmonary vein (PV) isolation remains the cornerstone of ablation for both paroxysmal and persistent atrial fibrillation (AF). 1,2 In paroxysmal AF, PV isolation can achieve approximately 80% 1-year freedom from AF with optimal contact-force radiofrequency or second-generation cryoballoon ablation. [3][4][5] In persistent AF, PV isolation alone has proven less successful likely due to the presence of non-PV drivers. 6,7 Evidence in support of localized drivers maintaining AF include experimental 8,9 and clinical findings. [10][11][12] Ablation of these sources with a procedural endpoint of AF termination may improve freedom from AF recurrence in both patients with persistent AF and the subset of patients with paroxysmal AF driven by non-PV AF sources. 12,13 Ripple map is a novel software feature of the Carto3 (Biosense Webster, Irvine CA) electroanatomic mapping system that graphically displays bipolar electrograms. (Figure 1) This visual marker displays a combination of depolarization frequency, electrogram fractionation, and voltage. The Ripple display corresponds directly to the recorded electrogram, with no interpolation between points or other processing. Using this methodology, it is possible to simultaneously display complete cardiac chamber electrograms recorded continuously over extended time intervals.
We hypothesized that (a) regions on the Ripple map displaying a concentration or cluster of high-frequency Ripple activity (HFRA) would correspond to sites with spatiotemporal electrogram dispersion and (b) ablation of HFRA sites would improve the likelihood for intraprocedure AF termination and freedom from AF in patients undergoing an ablation procedure for persistent AF.

| METHODS
This study was a single-center retrospective evaluation of 162 consecutive patients who underwent first-time ablation for persistent AF (continuous AF with a duration >7 days) from June 2013 to January 2018 using either a standard stepwise approach as developed by Haïssaguerre et al [14][15][16] (n = 105) or a Ripple map guided approach (n = 57). Patients with very severely enlarged left atrial (LA), defined as a Carto measured volume greater than 300 cm 3 , were not included. Study participants were >18 years old undergoing ablation for AF. A minimum 90 days of follow-up after the index ablation was required. There were no other exclusion F I G U R E 1 Example of Ripple Map. This map was obtained during sustained AF after PVI. Color shading represents bipolar voltage displayed on the anatomic map with a range of 0.1 mV (red) to 0.5 mV (purple). White Ripple bars representing the bipolar electrograms are shown perpendicular to the map surface with the size corresponding to the bipolar voltage. A total of 2131 points were acquired for this map with even distribution across the atrium. Some regions show no Ripple bars, indicating there was no depolarization at that location rather than lack of point acquisition. Analysis of this map demonstrated a HFRA site with continuous high-frequency activation on the left atrial septum (dashed line). Ablation of this location resulted in AF termination to NSR. AF, atrial fibrillation; HFRA, high-frequency Ripple activity; NSR, normal sinus rhythm; PVI, pulmonary vein isolation criteria for this study including duration of persistent AF. Institutional Review Board approval was obtained, and all patients consented to research participation.

| Ripple map guided protocol and mapping
Patients who were in normal sinus rhythm (NSR) at baseline underwent AF induction with atrial pacing from the coronary sinus.
Before PV isolation, the PV and LA anatomy was constructed using a Pentaray catheter (Biosense Webster, 2-6-2 mm electrode spacing). After completion of the high-density AF anatomic and point map,  Figure   1, Video S1) HFRA locations with the highest relative frequency by the visual analysis were targeted first, followed by progressively slower HFRA regions until termination. The highest frequency was defined specifically in relation to the atrium being mapped rather than an arbitrary absolute frequency. The largest anatomic regions as measured on the mapping system (typically greater than~1.0 cm diameter) with HFRA were targeted for ablation first, followed by smaller regions. Isolated small segments, arbitrarily defined as less than 0.5 cm in diameter, were not ablated as these often represented isolated complex fractionated atrial electrogram (CFAI) electrograms.
Ablation proceeded with PV antral isolation followed by sequential ablation of the largest HFRA regions. Ablation of HFRA regions was performed in a cluster format and progressively expanded until local cycle length was observed to slow and there was loss of localized CFAE and dispersion. Ablation of HFRA sites was continued until AF termination occurred. HFRA sites not required for AF termination were not ablated as these sites did not appear critical for maintenance of AF. If AF remained after ablation of all the HFRA sites observed on the initial Ripple map, up to two additional Ripple maps of the LA were performed, followed by a Ripple map of the right atrium (RA). If AF persisted despite ablation of all recognized HFRA regions in either chamber or if the remaining HFRA region was adjacent to a critical structure (eg, the AV node or sinus node), cardioversion was performed to restore NSR. Linear ablation of the LA roof was performed for roof dependent macro reentrant atrial tachycardia (AT) or if extensive LA roof ablation was performed for HFRA in this location. Other linear or focal ablation was not performed unless a sustained AT was observed.

| Comparison stepwise ablation protocol
A comparison historical cohort consisted of sequential patients who underwent the first-time ablation for persistent AF using a standard stepwise approach as described by Haïssaguerre et al. 14 For patients in the stepwise cohort, AF was induced with atrial pacing if NSR was present at baseline. Ablation was performed in the following sequence until AF termination occurred: PV antral isolation, posterior LA ablation, LA roof ablation, CFAE ablation, and lateral mitral isthmus ablation. In the CFAE step, ablation was performed for CFAE in both the LA and RA. If AF remained after the final step, cardioversion was performed to restore NSR.

| All patients ablation protocols
All patients in both cohorts underwent radiofrequency ablation using an externally irrigated, contact-force sensing catheter (SmartTouch and SmartTouch SF; Biosense Webster). A minimum of 5g contact was required before initiation of radiofrequency ablation. Power was fixed at 35 W and increased rarely to a maximum of 45 W if LA roof or lateral mitral isthmus block could not be achieved. All linear lines performed were tested for conduction block. PV entrance block was reconfirmed 30 minutes after initial isolation was observed. Esophageal temperature was monitored during posterior wall ablation.
After the restoration of NSR through either ablation or cardioversion, an attempt at AF induction was performed with atrial pacing for up to 5 seconds in duration at the atrial effective refractory period or a minimum of 220 ms. Repeat ablation procedures for AF were performed using the same strategy as the first ablation (stepwise or Ripple guided).
The primary study outcome was 18-month-single-procedure freedom from AF lasting longer than 30 seconds assessed by all available means including serial ECGs, Holter, extended outpatient monitoring, implanted device interrogation, and symptoms consistent with sustained tachyarrhythmia. Secondary outcomes included 18-month-single-procedure freedom from AT, and any atrial arrhythmia, 18-month freedom from AF, AT or any atrial arrhythmia after two procedures, acute AF termination during the index procedure, the use of antiarrhythmic medication, index procedure time, and complications.

| Statistical analysis
Outcome analyses were performed for all patients who had a followup. If both AF and AT occurred in the same patient during follow-up, only one event was counted in the freedom from any atrial arrhythmia analysis. Survival analyses were performed to compare study groups for the primary and secondary time to event outcomes.
Survival curves were generated using multivariable Cox proportional hazard modeling to adjust for baseline characteristics. For secondary categorical variables, the χ 2 test was performed with a statistical significance preset at 0.05. Continuous variables were compared using the Student t test.

| Patients
The patient population comprised a total of 162 consecutive patients who underwent a first-time ablation for persistent AF using either a Ripple map guided (n = 57) or standard stepwise (n = 105) approach.  Table 1. There were more males in the standard-map group (86% vs 67%, P = .009). No other significant differences were observed. The average duration of continuously persistent AF before the procedure was not different between the two groups (5.0 ± 5.7 months in the Ripple group vs 7.3 ± 10.8 months in the stepwise group, P = .14), nor was there any difference in the number of patients in continuous AF for more than 6 months (23% in the Ripple group vs 29% in the standard group, P = .18). There was no difference between in the groups in baseline left ventricular ejection fraction or LA size as measured by 2D echocardiogram. The survival curves for freedom from AF, AT, and any atrial arrhythmia are shown in Figures 2, 3, and 4. The survival curves were adjusted for sex due to its associations with the ablation approach; however, there was no evidence that sex confounded the relationship between ablation approach and freedom from arrhythmias.

| Acute procedure characteristics and outcomes
Acute AF termination during the index ablation procedure was observed in 91.1% of Ripple map guided patients, vs 52.4% of standard stepwise patients (P = < .0001; Table 3) The mode of termination was not different between the two groups (21% to NSR and 67% to AT/atrial flutter (AFL) in the Ripple group vs 24% to NSR and 69% in the standard group). A mean of 4.3 ± 1.9 (range 1-9) Structural heart disease* (%) 8 (14) 18 (17) .61 Left atrial diameter, mm 43 ± 5 45 ± 6 .062 Ejection fraction 53 ± 11 55 ± 11 .28   This study supports the pioneering work of other investigators including Narayan et al, 11 Seitz et al, 12 and Nadamanee et al 17 who previously demonstrated a high rate of freedom from AF through ablation of identified AF drivers through alternative but supporting techniques. As was the case with these authors, we found an electrogram approach to mapping and ablation of AF drivers resulted in a higher rate of freedom from AF than a conventional anatomic approach. Our study differed from these prior investigations in so far as we only evaluated persistent AF rather than a mixed cohort of paroxysmal and persistent AF.

| LIMITATIONS
This study is limited by its single-center retrospective nature.
Freedom from AF and AT is likely overestimated as systematic extended monitoring was not performed in all patients to reveal asymptotic paroxysmal tachycardias. The identification of critical HFRA sites are not strictly objective or quantified in this study methodology, therefore training is necessary to reduce observation variability among operators. This is not substantially different though from other methods such as spatiotemporal dispersion upon which these findings corroborate. Larger, randomized multicenter investigations are necessary to confirm the effectiveness of HFRA ablation.

| CONCLUSION
In this study, we evaluated a novel ablation target for persistent AF.
Ablation of sites which we have termed HFRA resulted in frequent termination of persistent AF and a higher rate of single-procedure freedom from AF than was observed in a historical cohort in which a conventional sequential ablation strategy was employed. This