An impact of superior vena cava isolation in non‐paroxysmal atrial fibrillation patients with low voltage areas

Abstract Background This study aimed to investigate the correlation between left atrial low‐voltage areas (LVAs) and an arrhythmogenic superior vena cava (SVC) and the impact on the efficacy of an empiric SVC isolation (SVCI) along with a pulmonary vein isolation (PVI) of non‐paroxysmal atrial fibrillation (non‐PAF) with or without LVAs. Methods We retrospectively enrolled 153 consecutive patients with non‐PAF who underwent a PVI alone (n = 51) or empiric PVI plus an SVCI (n = 102). Left atrial voltage maps were constructed during sinus rhythm to identify the LVAs (<0.5 mV). An arrhythmogenic SVC was defined as firing from the SVC and an SVC associated with the maintenance of AF‐like rapid SVC activity. Results An arrhythmogenic SVC and LVAs were identified in 28% and 65% of patients with a PVI alone and 36% and 73% of patients with a PVI plus SVCI, respectively (P = .275 and P = .353). In the multivariate analysis a female gender, higher pulmonary artery systolic pressure (PAPs), and arrhythmogenic SVC were associated with the presence of LVAs. In the PVI plus SVCI strategy, there was no significant difference in the atrial tachyarrhythmia/AF‐free survival between the patients with and without LVAs after initial and multiple sessions (50% vs. 61%; P = .386, 73% vs. 79%; P = .530), however, differences were observed in the PVI alone group (27% vs. 61%; P = .018, 49% vs. 78%; P = .046). Conclusions The presence of LVAs was associated with an arrhythmogenic SVC. An SVCI may have the potential to compensate for an impaired outcome after a PVI in non‐PAF patients with LVAs.


| INTRODUC TI ON
Pulmonary vein (PV) isolation (PVI) is the cornerstone procedure for patients with all types of atrial fibrillation (AF). 1 However, the outcome of a PVI alone is unsatisfactory in patients with nonparoxysmal AF (PAF). To improve the outcome after catheter ablation of non-PAF, various ablation strategies involving substrate modification have been devised. However, an empiric conventional substrate modification, such as left atrial linear ablation and complex fractionated atrial electrogram ablation, added to the PVI have been questioned as to whether they improve the outcome after ablation of persistent AF. 2 The left atrial substrate parameters, such as the voltage and velocity, progress parallel to the progression of the AF type. 3 It was reported that the presence of left atrial lowvoltage areas (LVAs) after the PVI has been shown to be a strong predictor of an AF recurrence, and LVAs could be targeted for substrate modification. [4][5][6] However, LVA ablation in randomized controlled trials also failed to show any advantage over a conventional substrate modification. 7 Therefore, searching for new additional ablation strategies beyond the PVI to improve the outcome after ablation of non-PAF is necessary, especially in patients with an advanced arrhythmogenic substrate. The presence of LVAs is associated with not only a higher age, female gender, larger left atrium (LA), and persistent AF but also with sinoatrial node dysfunction, which suggests the spread of an arrhythmogenic substrate in the right atrium (RA). 8 Although the superior vena cava (SVC) has been described as one of the most frequent non-PV triggers 9 , an empiric SVC isolation (SVCI) in addition to the PVI had limited effect after the initial procedure using a non-contact force (CF) sensing catheter, especially in patients with non-paroxysmal AF 10 -13 . It has been reported that the durability of the PVI and SVCI using a non-CF catheter is low [14][15] , and it is assumed that the use of a CF catheter has recently improved the durability of the PVI and SVCI.
We hypothesized that strict thoracic vein isolation strategies using a CF sensing catheter, which contain a PVI and SVCI, determine the outcome after ablation in patients with non-PAF. The first aim of this study was to investigate the correlation between the LVAs and arrhythmogenicity of the SVC. The second aim of this study was to test the hypothesis that a strict SVCI in addition to the PVI would improve the ablation outcomes in non-PAF patients, with or without the presence of LVAs.

| Study population
This study was approved by the institution's ethics committee and followed the "Declaration of Helsinki" and the ethical standards of the responsible committee on human experimentation, and all patients underwent informed consent. Consecutive patients with non-PAF, who underwent catheter ablation using a PVI alone or PVI plus SVCI strategy, were retrospectively enrolled into the current study. The PVI alone strategy was applied in the former half of the cases, and the PVI plus SVCI strategy was applied in the latter half of the cases. The PVI alone or PVI plus SVCI strategies were performed regardless of an arrhythmogenic SVC and the existence of LVAs. The follow-up period was set at 18 months.
Exclusion criteria were (a) an age <20 years, (b) prior surgery of the heart, lungs, or esophagus, (c) radiotherapy due to cancer in the thorax or previously receiving chemotherapy, and (d) a prior catheter ablation.

| Electrophysiological study
Electrocardiogram (ECG)-gated and contrast-enhanced computed tomographic (CT) imaging was performed before the procedure.
One multipolar 6Fr 20-pole catheter (Be eat; Japan Lifeline, Tokyo, Japan) was positioned in the coronary sinus, which covered the high RA and SVC, via the right subclavian vein for pacing, continuous recording, and internal AF cardioversion throughout the procedure.

| Voltage mapping
Sinus rhythm was restored by external or internal cardioversion before the PVI, and then a voltage map was created 10 minutes later. If AF did not convert to sinus rhythm, the PVI was performed during AF. After the right or left PVI, external or internal cardioversion was repeatedly administered aiming at the restoration of sinus rhythm. In such a case, voltage mapping was performed after the PVI. Mapping of the LA was performed during sinus rhythm with a 20-polar Pentaray catheter (Biosense Webster) using the CARTO mapping system and merged with CT integration (CARTOMERGE; Biosense Webster). Five hundred to 1000 LA mapping points per patient were carefully obtained. The band pass filter was set at 30 to 500 Hz. The bipolar peak-to-peak voltage at each acquired point was measured. LVAs were defined as those of <0.5 mV and covering >5% of the LA body surface area according to the published data. [3][4][5][6][7] The CARTO system automatically calculated the surface area from the manually selected points. To exclude LVAs because of insufficient wall contact, the voltages were reconfirmed by the CF catheter introduced through the long sheath for the sites with apparent LVAs.

| Ablation protocol during the initial procedure
All patients underwent a circumferential PVI using irrigated radiofrequency current and an integrated 3D image with CF guidance of more than 10 g. For 30 seconds at each point, irrigated radiofrequency energy was delivered using a target temperature of 43°C, One hundred-two of the patients in this study underwent an SVCI in addition to the PVI. After confirmation of no PV reconduction, the mapping and ablation catheters were withdrawn back into the RA. The geometry of the RA was reconstructed, and the SVC-RA junction was tagged on the geometry based on the SVC angiography. The circular catheter was placed just above the RA-SVC junction. Segmental ablation targeting the earliest RA-SVC junction was applied for the SVCI with CF guidance of more than 10 g. High output pacing (10 mA) was performed before the radiofrequency current delivery at the posterolateral aspect of the SVC. In such sites, ablation was avoided to prevent phrenic nerve injury. Irrigated radiofrequency energy was delivered for 20 seconds using a target temperature of 43°C, maximum power of 20-25 W, and an infusion rate of 17 mL/min. A lower energy (20W) and lower CF (10-15g) were applied on the lateral side as compared to the septal side of the SVC to prevent phrenic nerve palsy. An SVCI was characterized as the disappearance of the SVC potentials or the dissociation of the SVC potentials with RA activity. No patients underwent ablation of linear lesions, complex atrial electrograms, or ablation of non-PV triggers. A cavotricuspid isthmus ablation was only performed in patients with a history of atrial flutter (AFL).

| Identification of an arrhythmogenic superior vena cava
Mapping electrodes were placed in the SVC during the entire procedure (during AF, before and after cardioversion), and detailed mapping was performed when arrhythmogenicity of the SVC was suspected. During an isoproterenol infusion (1-10 μg/min) after the PVI, a circular mapping catheter was placed to find if there was an arrhythmogenic SVC in the initial and repeat ablation procedures.
A 20-mg bolus of ATP was injected with the administration of the isoproterenol. An arrhythmogenic SVC was defined as firing from the SVC with or without triggering AF and an SVC associated with the maintenance of AF-like rapid SVC activity ( Figure 1).

| Ablation protocol during repeat procedures
Repeated electrophysiological procedures were undertaken for any atrial tachyarrhythmias (ATs) lasting for 30 seconds, which included PAF, atrial tachycardia, and AFL and episodes of AF lasting for seven days (persistent AF). The initial strategy was an assessment of the PV reconduction during sinus rhythm after cardioversion, followed by the closure of all PV conduction gaps and an electrical re-isolation. In patients with a PVI plus an SVCI strategy, SVC reconduction during sinus rhythm was also assessed, followed by the closure of all SVC conduction gaps and an electrical re-isolation.
F I G U R E 1 A circular mapping catheter is placed in the SVC. A, AF initiated from the SVC after cardioversion. Note that the SVC potentials are recorded following the RA potentials during sinus rhythm. B, At the beginning of the procedure, very rapid activity was observed inside the SVC during AF. CS; coronary sinus; HRA, high right atrium; SVC, superior vena cava

| Post-ablation management and clinical followup
In all patients the anti-arrhythmic drugs were discontinued after the procedure. For 3 days after the procedure, ECG monitoring was performed.
The ECG and Holter ECG recordings were repeated at 1, 3, 6, 12, and 18 months after the procedure. If the patients complained of symptoms suggestive of an arrhythmia recurrence, an event monitor was provided. Long-term success was defined as the absence of any sustained (>30 sec) ATs. Three-month of blanking period was applied. The indication for a repeat procedure was left to the discretion of the physician.

| Statistical analysis
Categorical variables are expressed as absolute and relative frequencies.
Continuous variables are expressed as the mean ±SD or median and interquartile range as appropriate. Unpaired t-tests, using a Wilcoxon rank-sum test where appropriate, were used for the comparisons between groups. Categorical variables were compared using a Fisher's exact test. To test for predictors of LVAs we used a multivariate binary logistic regression. Baseline variables that were significant (P < .05) in the univariate analysis entered into the multivariate analysis. The eventfree rate was estimated by the Kaplan-Meier analysis using a log-rank test. A two-tailed probability value of <.05 was considered significant.  (Figure 3).

| Predictors of LVAS
The comparison of the patient characteristics between the patients with and without LVAs is shown in Table 2. LVAs were more frequently found in the patients with a higher age (66.0 ± 9.2 vs.

| D ISCUSS I ON
Our study of patients undergoing catheter ablation using a PVI alone or PVI plus SVCI strategy for non-PAF produced the following results: (a) the presence of LVAs was associated with a female sex, the estimated PAPs, and an arrhythmogenic SVC; (b) in the PVI plus SVCI strategy, there were no significant differences in the ATs recurrence between the patients with and without LVAs, however, a difference was observed for the PVI alone strategy.

| Low voltage areas as the substrate of AF
Low-voltage electrograms are mainly due to atrial fibrosis. AF also promotes atrial fibrosis caused by tissue remodeling and perpetuates the maintenance of AF. Electrophysiologically, atrial fibrosis produces lower amplitude electrograms, fractionated electrograms, and a conduction heterogeneity that can be identified using electroanatomic mapping during sinus rhythm. Therefore, fibrotic remodeling tissue acts as an AF substrate because it exhibits slow conduction and a short action potential duration, which facilitates reentry. [16][17] It is reported that the presence of LVAs is associated with a higher age, female gender, larger LA, sinoatrial node dysfunction, and persistent AF. 8 LVAs are observed in patients with persistent AF, with a prevalence of 35%-84%. [5][6][7][8][18][19]

| Superior vena cava as the arrhythmogenic origin of AF
The SVC, one of the most common sites of non-PV foci, has been established as an important source of AF. 9 However, the incidence of AF originating from the SVC is unknown. In general, an arrhythmogenic SVC is defined as an SVC triggering AF, SVC initiating AF, and/or SVC associated with the maintenance of AF. An isoproterenol infusion, ATP bolus injection, and cardioversion of reinitiated AF are the techniques for the provocation of non-PV triggers. It has been reported that arrhythmogenic SVCs are identified in 2.2% to 19.4% of PAF patients using various provocation techniques. [20][21][22] An arrhythmogenic SVC is rarely identified in patients with persistent or long-standing persistent AF. [23][24] The mechanism why the frequency of an SVC origin of AF is lower in non-PAF than PAF is unclear. In this study, an arrhythmogenic SVC was identified in 51 cases (33.3%) by observation during AF, cardioversion and provocation maneuvers with an isoproterenol infusion, and an ATP bolus injection. The discrepancy in the incidence of arrhythmogenic SVCs between the several reports and our results may be due to the different definition of an arrhythmogenic SVC and provocation maneuvers.
We defined an arrhythmogenic SVC as not only SVC firing triggering AF but also SVC firing not triggering AF and rapid SVC activity during AF.
In fact, SVC firing and rapid SVC activity during AF were observed in 13.7% and 24.8% in this study, respectively.

| The effects of a superior vena CAVA isolation in non-PAF patients with low voltage areas
The SVC has been described as one of the most frequent non-PV triggers. 9 However, an empiric SVC isolation (SVCI) in addition to the PVI had a limited effect, especially in patients with non-PAF. [10][11][12][13] The differences in the efficacy of the SVCI depending on the arrhythmogenic substrate of AF, which is suggested in this study, may explain the reason why the effects of an empirical SVCI are controversial.
In general, the procedure added to the PVI for AF with LVAs is substrate modification for LVAs. Some papers reported that a

| Safety and limitations of a contact force guided empiric superior vena CAVA isolation
Right phrenic nerve injury is a major concern during an SVCI. In addition, a high CF may increase the risk of phrenic nerve injury. In this study, no phrenic nerve injury was detected after the SVCI. Therefore, we could not comment on what the predictors of phrenic nerve palsy were in our case series. Applying a lower energy and CF on the lateral side and avoiding applications at sites with phrenic nerve capture during high output pacing might protect against phrenic nerve injury. On the other hand, such an ablation setting might lead to a high RA-SVC reconnection rate in patients with repeat procedures. Recently, the ablation index has been developed as a novel marker of the lesion quality incorporating the CF, time, and power in a weighted formula, and has been shown to improve the one-year outcome and prevent reconnections by using an appropriate ablation index value. 27 Understanding the appropriate ablation settings to suppress reconnections after an SVCI without complications is desired to improve the outcome after the initial procedure with a PVI plus SVCI strategy.

F I G U R E 4
Comparison of the freedom from ATs after the initial and last procedure between the patients with and without LVAs. The graph shows the Kaplan-Meier estimates of the freedom from documented ATs after the initial (A) and last (B) procedure in the patients with the PVI alone strategy and after the initial (C) and last (D) procedure in the patients with the PVI plus SVCI strategy. ATs, atrial tachyarrhythmias; LVAs, low voltage areas; Pts, patients [Colour figure can be viewed at wileyonlinelibrary.com]

| S TUDY LIMITATI ON
This study has several limitations. First, this is a retrospective singlecenter study with a relatively small sample size. Thus, larger multicenter prospective studies are needed to confirm our findings.
Second, the recurrence rate of ATs/AF might have been underestimated because asymptomatic ATs/AF episode might have been undetected by using 24-hour ambulatory monitoring as compared with implantable loop recorders. Third, the confidence module of the CARTO 3 system was not used in this case series because the confidence module had not been introduced at the time of the first procedure. Poor contact of the mapping catheter was one of the most important limitations for detecting the accurate LVAs. Thus, our findings need to be confirmed using a new mapping system in order to detect more accurate LVAs.

| CON CLUS IONS
The presence of LVAs was associated with a female gender, the estimated PAPs, and an arrhythmogenic SVC. In the PVI plus SVCI strategy, pre-existence of LVAs as detected by LA voltage mapping has not been shown to be a predictor of an arrhythmia recurrence after an AF catheter ablation. These findings suggest that a strict thoracic vein isolation is the one of the options for catheter ablation in non-PAF patients with LVAs.

ACK N OWLED G EM ENT
The authors thank Mr John Martin for his linguistic assistance with this article.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets analyzed in this study are available from the corresponding author on reasonable request.