Regional differences in the predictors of acute electrical reconnection following high‐power pulmonary vein isolation for paroxysmal atrial fibrillation

Abstract Background Acute pulmonary vein reconnection (PVR) is associated with long procedure times and large radiofrequency (RF) energy delivery during pulmonary vein isolation (PVI). Although the efficacy of high‐power PVI (HP‐PVI) has been recently established, the determinants of acute PVR following HP‐PVI remain unclear. Methods We evaluated data on 62 patients with paroxysmal atrial fibrillation undergoing unipolar signal modification (USM)‐guided HP‐PVI. A 50‐W RF wave was applied for 3‐5 seconds after USM. In the segments adjacent to the esophagus (SAEs), the RF time was limited to 5 seconds. Each circle was subdivided into six regions (segments), and the possible predictors of acute PVR, including minimum contact force (CFmin), minimum force‐time integral (FTImin), minimum ablation index (AImin), minimum impedance drop (Imp‐min), and maximum inter‐lesion distance (ILDmax), were assessed in each segment. Results We investigated 1162 ablations in 744 segments (including 124 SAEs). Acute PVR was observed in 21 (17%) SAEs and 43 (7%) other segments (P = .001). The acute PVR segments were characterized by significantly lower CFmin, FTImin, AImin, and Imp‐min values in the segments other than the SAEs and larger ILDmax values in the SAEs. Furthermore, lower Imp‐min and larger ILDmax values independently predicted acute PVR in the segments other than the SAEs and SAEs (odds ratios 0.90 and 1.39 respectively). Acute PVR was not significantly associated with late atrial fibrillation recurrence. Conclusions Avoiding PVR remains a challenge in HP‐PVI cases, but it might be resolved by setting the optimal target impedance drop and lesion distance values.


| INTRODUC TI ON
Pulmonary vein isolation (PVI) is effective in the treatment of atrial fibrillation (AF). High-power short-duration (HPSD) PVI, a recently developed therapeutic option, is associated with a low radiofrequency (RF) energy requirement and short procedure and RF durations. Several studies have demonstrated acceptable acute and late outcomes in association with HPSD-PVI. 1-3 Unipolar signal modification (USM), a sign of transmural lesion creation by RF application, as previously described, 4 has recently been used as a guide for HPSD-PVI with excellent outcomes. 2,5 Despite this accumulated evidence, acute pulmonary vein reconnection (PVR) persists in 10%-13% of patients, 1-3 resulting in longer procedure times and larger RF energy delivery. Even when re-ablation is performed for PVI, patients with acute PVR remain more vulnerable to late recurrence. 6,7 Accordingly, we sought to evaluate the characteristics of acute PVR following HP-PVI under USM guidance and assess the predictors of acute PVR as surrogate markers of durability in patients with paroxysmal atrial fibrillation.

| Study patients
This single-center retrospective observational study comprised 64 consecutive patients with paroxysmal AF who had undergone RF catheter ablation (RFA) using the HP-PVI strategy from October 2018 to June 2019. Of these patients, two who were treated with the dragging technique were excluded. Three-dimensional cardiac computed tomography (3-D CT) with or without a contrast agent and transthoracic echocardiography were performed within a month before the procedure. Transesophageal echocardiography was performed for patients with a CHADS 2 score ≥2 to rule out the presence of thrombi in the left atrium (LA). The use of all antiarrhythmic drugs was stopped for at least five half-lives before the procedure.
The study was approved by the institutional review board of Tokyo Women's Medical University and performed according to the institutional guidelines and in accordance with Declaration of Helsinki.
All patients provided written informed consent.

| Catheter ablation protocol
Details of the catheter ablation protocol employed have been previously published. 8 Briefly, all patients underwent PVI and superior vena cava isolation and were deeply sedated with a 10-minute continuous administration of dexmedetomidine (6 µg/kg/h), followed by a continuous infusion (0.3-0.7 µg/kg/h). Four well-trained operators performed these procedures using geometric information, as obtained from a reconstructed 3-D CT imaging system (CARTO 3; Biosense Webster, Inc). To monitor the esophageal temperature, a multielectrode esophageal temperature-monitoring probe was employed (Esophaster; Japan Lifeline).
F I G U R E 1 Unipolar signal modification and impedance changes. Example of a change in the unipolar electrogram result with dynamic changes in the generator impedance value. The S-wave of unipolar signal disappears 3 s after radiofrequency catheter ablation (RFA), according to significant impedance decreases, and an R-pattern is achieved 5 s after RFA. Following this, the impedance value gradually decreases from 133 Ω and finally reaches 120 Ω, while a notch in the positive unipolar electrogram completely disappears. RF

| Regional assessment of acute PVR
Acute PVR was defined as the presence of spontaneous or Webster, Inc) were as previously described 5 ; however, VisiTag was not utilized as a guide for the RF delivery but as one of the anatomical pieces of information on the potential site of a residual LA-PV electrical connection after circumferential RFA around the ipsilateral PVs.

| Follow-up after the procedure
All patients were followed up in the outpatient clinic at 1, 3, 6, 9, and 12 months after the procedure and every 6 months thereafter.
Atrial tachyarrhythmia (ATA) recurrence was evaluated according to the patients' symptoms and 24-hour ambulatory monitoring (3, 6, 9, and 12 months after the ablation and every 6 months, thereafter). Patients with palpitations were encouraged to use a portable electrocardiographic monitoring device (HCG-801R; Omron).
Recurrence was defined as the presence of recurrent symptoms and/or detection of ATAs using the aforementioned modalities or data provided by the cardiac implantable electrical devices (ATAs lasting >30 seconds) after a 2-month blanking period, without the use of any anti-arrhythmic drugs.

| Statistical analysis
Continuous variables are expressed as the mean ± standard deviation or median with interquartile range. Student's t-test was used for the comparison of the characteristics of the segments with acute PVR and those without it, and the chi-square test was used to evaluate statistical differences in the categorical variables. Regional The tests were considered statistically significant at P < .05. All statistical analyses were performed with JMP ® 13 (SAS Institute Inc). Table 1 presents the 62 patients' baseline and procedural characteristics.

| Background characteristics and clinical outcomes
Structural heart disease (SHD) was observed in 11 (15%) patients, and the median AF history duration was 8 months. The average procedure duration (from catheter insertion to removal) was 125 ± 46 minutes, whereas the RF duration and energy for PVI were 10 ± 3 minutes and 28 ± 8 kJ respectively. Acute PVR was observed in 44 (70%) patients; these were eliminated in the same session. All patients had undergone superior vena cava isolation after PVI. Complications occurred in only one patient; an acute right phrenic nerve injury was identified following the completion of the right-side PVI before superior vena cava isolation.
This patient showed partial recovery at the 6-month follow-up.

| Region-specific analysis of acute PVR
Of the 1162 ablations observed in 620 segments other than the 124 SAEs, acute PVR was observed in 43 (7%) and 21 (17%) segments (P = .001) respectively. Acute PVR included 18 spontaneous/ isoproterenol-induced PVRs and 5 ATP-induced DCs in the SAEs and 39 spontaneous/isoproterenol-induced PVRs and 6 ATP-induced DCs in segments other than SAEs. Figure 2 shows the distribution of the acute PVR sites, and Figure 3 shows the relevant indices in each segment. The degree of regional variance in these indices was significant (P < .0001). Spontaneous/isoproterenol-induced PVR was more prevalent in the SAEs than in the segments other than the SAEs (15% vs 6%, P = .005); and for the ATP-induced DCs, similar findings were also observed (4% vs 1%, P = .02). In the SAEs, significant differences were observed only in the ILD max between the segments with and without acute PVR; in the segments other than the SAEs, significant differences were noted in the CF min , FTI min , AI min , and Imp-min ( Table 2). This difference in these indices based on the detailed classification (spontaneous/isoproterenol-induced PVR or ATP-induced DC) is described in Tables S1 and S2. In the ROC curve analysis, AUCs for the prediction of acute PVR absence in the SAEs and segments other than the SAEs are indicated in Figure S1. The relatively low sensitivity of each variable was evident in the discrimination of durable lesions at each known cut-off value, 11-13 both in the SAEs and segments other than the SAEs (Table 3). Meanwhile, durable segments were characterized by an Imp-min of 6.5 Ω in the segments other than the SAEs, ILD max of 4.8 mm in the SAEs, and AI min of 405 au in the segments other than the SAEs, with a specificity of 90% (Table S3). Table S4 shows the odds ratio of acute PVR in association with the various 3-D mapping-related indices. In the multivariate analysis, ILD max and Imp-min were found to be the sole independent predictors of acute PVR in the SAEs and segments other than the SAEs, respectively, after adjustment for the confounders of AI min , ILD max , and Imp-min.

| Follow-up and late outcomes
Over the median follow-up duration of 12. (20%) (treated as a recurrence) patients. The ATA-free rates were 81% and 74% at 6 months and 1 year after the procedure, respectively; the exclusion of patients with SHD increased these values to 90% and 86% at 6 months and 1 year after the procedure respectively. There was no significant difference in the occurrence of late outcomes between the patients with and without acute PVR (P = .64; Figure S2). Of patients with late recurrence, 12 underwent a redo procedure a median of 203 days after the first session. Ten patients had segments with late PVR (Figure 2)  PVR. There was no significant difference in the 3-D mappingrelated indices between the segments with late PVR and those without it. Only the Imp-min tended to be higher in the segments other than the SAEs without late PVRs than in those with them

| D ISCUSS I ON
In the present study, HP-PVI was achieved under USM guidance with relatively short procedure and isolation durations, and acceptable late outcomes, among the patients with paroxysmal AF.
The AI min value required for acute PVR avoidance was remarkably lower than that used in clinical practice 1 ; in addition, higher ILD max and lower Imp-min values were shown to be independently predictive of acute PVR during the procedure in the SAEs and segments other than the SAEs, respectively, after adjusting for AI min . Acute PVR was not significantly associated with late AF recurrence.
Recent studies have reported the superior efficacy and safety of HPSD-PVI compared to conventional PVI, 14 Additionally, the acute and late reconnection sites were not equivalent. 22 Nevertheless, converse advocation has been noted in the past.  Note: Values are expressed as mean ± standard deviation.
Abbreviations: AI min , minimum ablation index; CF min , minimum contact force; FTI min , minimum force-time integral; ILD max , maximum inter-lesion distance; Imp-min, minimum impedance drop; PVR, pulmonary vein reconnection; SAE, segment adjacent to the esophagus.

TA B L E 2
Differences in the threedimensional mapping-related indices in the segments with and without acute PVR final application. 6 Anter et al concluded that acute PVR is associated with six times the risk of late AF recurrence even when appropri-    There are certain limitations to this study. First, our study had a relatively small sample size. Although this may have led to the underpowering of the statistical analysis, in terms of per-segments analysis, 744 regions were considered, which, in turn, could improve the level of statistical confidence. Second, Imp-min and late recurrence had an associative tendency, but this association was not significant in the segments other than the SAEs, which may be derived from a small number of participants for statistics. Lastly, the acute PVRs were distributed heterogeneously. Future studies with larger sample sizes should assess each segment for the clarification of regional differences in the characteristics of acute PVR.

| CON CLUS IONS
Our findings primarily suggest that in association with USM-guided HP-PVI, lesion distance and impedance drop are better predictors of acute PVR in the SAEs and segments other than the SAEs, respectively, as compared with other 3-D mapping-related indices.

ACK N OWLED G EM ENTS
We thank Editage (www.edita ge.jp) for English language editing.

CO N FLI C T O F I NTE R E S T
Authors declare no conflict of interests for this article.