Acute and Long-term Outcomes of Cryoballoon Ablation in Patients with Atrial Fibrillation. Primary Results of the Spanish Cryoballoon Ablation Registry (RECABA)


 Background

Cryoablation is safe and effective for the treatment of atrial fibrillation (AF) in controlled clinical trials, but contemporary real-world usage and outcomes are limited.
Objectives

The Report of the Spanish Cryoballoon Ablation Registry (RECABA) was designed to evaluate acute and 12-month outcomes of cryoballoon ablation for the treatment of AF in Spain.
Methods

Patients from 27 Spanish centers were prospectively enrolled. Patients were treated with cryoballoon ablation and managed according to standard of care protocols at each center. Theprimary endpoint was ≥ 30 sec freedom from AF at 12-mo after a 3-mo blanking period. Secondary endpoints included a description of patient characteristics, cryoablation procedural strategy and safety, and predictors of efficacy.
Results

In total, 1742 patients (71.4% PAF, 68.8% male, mean age 58.02 ± 10.40 years, 76.1% overweight or obese, CHA2DS2-VASc index 1.40 ± 1.28) were enrolled. Patients received 7.2 ± 2.67 cryoapplications. PV potentials could be detected in 61% of the PVs during ablation, with a mean time to block of 52.9 ± 37.02 seconds. Acute PVI was observed in 97% of PVs with 75.8% isolated with the first cryoapplication. Mean procedural time was 113 ± 41 minutes. Acute complications occurred in 4.4% of the cases21. With follow-up in 1628 patients, AF-free survival was 78.5% (PAF: 80.6% vs PersAF 73.3%; p < 0.001). Left atrium enlargement, female sex, non-PAF, and early recurrence were independent predictors of AF recurrence (p < 0.05).
Conclusions

RECABA provides detailed insight into current dosing practices and demonstrates cryoablation is safe and effective in real-world use.


Introduction
Cryoballoon ablation (CBA) is a well-established technique for the treatment of atrial brillation (AF), that has been demonstrated to achieve similar safety and e cacy as radiofrequency ablation (RFA) with shorter procedure times. 1 While the safety and e cacy of AF ablation has been established, it was only until recently that outcomes of AF ablation were examined within large, prospectively enrolled cohorts of patients with AF. The recent report from the Cryo AF Global Registry described primary outcomes of cryoablation used according to local practice around the world. 2 Country speci c evaluations such as the GWTG-AFIB study, the FREEZE Cohort, and the 1STOP registry describe outcomes of cryoablation in the USA, Germany, and Italy, which have had many years of experience with the cryoablation catheter. [3][4][5] Uniquely, a recent report from Miyazaki et al. described the rst safety experience with cryoablation in Japan. 6 A comprehensive examination of modern cryoablation outcomes in an early, pan-country experience with cryoballoon ablation has yet to be reported.
The Spanish Catheter Ablation Registry is published annually as the o cial report of the Spanish Society of Cardiology Working Group on Electrophysiology and re ects general activity in Spanish electrophysiology units but does not describe AF ablation procedures in detail; therefore, patient selection, procedural techniques, and outcomes in real-world use in Spain have not been reported. 7 . The Spanish Registry of Cryoballoon Ablation (RECABA) prospective multicenter observational study aimed to comprehensively evaluate daily clinical practice and the current state of cryoballoon procedures in Spain.

Study design and population
The Registro Español de CrioAblación con BAlón (RECABA; NCT02785991) is an observational, prospective, national, multicenter (supplementary material online, Table S1) study of patients undergoing CBA for AF in Spanish centers with experience in the technique (at least 10 procedures per year). Patients were enrolled between September 2016 and January 2019. For inclusion in the study, patients were required to 1) be older than 18 years of age, 2) be eligible for CBA for AF, 3) have a life expectancy greater than 1 year, and 4) sign the informed consent. Patients with both rst and repeat procedures were enrolled. AF was classi ed as paroxysmal (PAF) or persistent atrial brillation (PerAF) according to the current ESC guidelines. 8 Clinical data were collected at the baseline procedure and at the annual followup by the coordinator responsible for each center through an electronic data collection system for clinical trials (eCRF). Data collection, management and quality control are detailed in supplementary material online (S2).

Objective and Endpoints
The main objective of the RECABA was to report standard clinical practice for PVI using CBA in Spanish hospitals. Standard clinical practice was evaluated by assessing patient characteristics, procedural and CBA techniques, procedure-related complications, the long-term follow-up strategy and e cacy across Spanish centers. The primary endpoint of the study was AF free survival at the 12-month follow-up. Secondary endpoints included the following: (1) a description of baseline characteristics of subjects who underwent CBA, (2) the acute e cacy, safety (the rate of all procedure-related adverse events), and e ciency of the procedure (e.g. electrical PVI, PV potential monitoring, cryoapplications analysis, related complications, and procedural related times), (3) description of the dosing strategies utilized, and (4) identi cation of the factors that predicted AF recurrence during follow-up.

Cryoballoon ablation procedure
General pre-ablation management and the CBA procedure were done in accordance with each center's standard protocol for pre-ablation examination, anesthesia and sedation, auxiliary technology for transseptal puncture, anatomical assessment, and phrenic nerve monitoring during right-sided ablation. Nevertheless, the CBA procedure was similar in all hospitals as cryoballoon-over-the-wire techniques were required to follow the internationally accepted techniques described in detail in recent publications and in the supplementary material online (S3). Of note, the following cryo-dosing parameters followed the usual practice of each hospital including: duration of cryoapplications, bonus/non-bonus-freeze strategy, postablation waiting time, and post ablation acute PVI adenosine testing.

Post-ablation management and follow-up
Oral anticoagulation (OAC) and anti-arrhythmic drug (AAD) management were done in accordance with each center's standard protocol and at the discretion of the cardiologist. Patients could continue taking AAD in the absence of recurrences. Patients were followed with outpatient visits using 24 h to 30-day Holter monitoring systems over the 12-month follow-up. Arrhythmia recurrence could also be monitored via electrograms registered in patients with an implanted device prior to (but not after) enrollment in the study. AF recurrence was de ned as any AF episode that lasted longer than 30-s accordingly documented (on ECG, Holter monitor, event recording systems or implantable devices). A 3-month blanking period was established, during which detected AF episodes were excluded from the primary endpoint.

Statistical analysis
Quantitative variables are expressed as mean, median, standard deviation and interquartile range as appropriate. Differences in quantitative variables were evaluated through the student t-test for independent samples (or analysis of variance depending on the number of groups compared) and between paired variables through the student t-test for related samples or the analysis of variance for repeated measures. Categorical data were compared with the chi-square test or Fisher's exact test, depending on the number of categories. The probability of recurrence was predicted using a logistic regression, and a survival analysis predicted the most likely time of recurrence. Univariate models considering each potential predictor were estimated. Subsequently, a multivariate model considering predictors from the baseline visit and procedural characteristics were estimated. An overall multivariate model including signi cant variables was estimated with non-signi cant variables as covariates. Time to recurrence was studied using Kaplan-Meier estimates and categorical predictive covariates were included as grouping variables and compared using the log-rank statistic. The Cox-proportional hazard regression method was used to estimate the conditional hazard rate. In order to study AAD use retention at follow-up visits, a logistic loglinear model was estimated, including AF type, AAD use after cryoablation procedure and AF recurrence. Effect con dence intervals and model standardized residuals were obtained. Chisquare for independence was used as test statistic for independence. All analyses were carried out using IBM SPSS v26.0 software. A nominal 5% signi cance level was assumed in all analysis and adjustments for multiple comparisons were completed when needed.

Study population
A total of 1742 patients eligible for CBA for the treatment of paroxysmal (1238; 71.1%) or persistent (504; 28.9%) AF were prospectively included from 27 Spanish Centers. The mean number of patients included per center was 64.5 ± 53.21, (13 centers enrolled less than 50 patients, 9 enrolled between 51-100, 3 between 101-150 and 2 enrolled more than 150 patients). All but one enrolled patient (who opted out of the procedure after singing informed consent) underwent a CBA. Of all procedures, 1665 (96.6%) were rst CBA procedure and 77 (4.4%) were repeat ablation procedures.
Clinical characteristics of the study population are shown in Table 1. Most patients were male (68.8% versus 31.2%; p < 0.001), under 65 years old (70.7%), and were overweight or obese (76.1%). Only 52 patients (3%) were older than 75 years. One patient was younger than 20 years and 4 patients were older than 80 years of age. The most frequently observed cardiovascular risk factors were hypertension (46.3%) and dyslipidemia (34.7%). The population presented a low embolic risk overall (mean CHA 2 DS 2 -VASc index 1.40 ± 1.28), but CHA 2 DS 2 -VASc was ≥ 2 in 39.8% of the cohort. Only 18.9% of patients had structural heart disease (most frequently tachycardia-related cardiomyopathy and coronary artery disease) and left ventricular ejection fraction was preserved in most patients (≥ 50% in nearly 90% of the patients). LA was enlarged (LA > 40 mm or area > 20 cm 2 ) in over half of the patients (most frequently mild). Oral anticoagulation (most frequently direct anticoagulation drugs) and AADs were taken at baseline in 75.5% of patients. In 148 patients (8.5%) CBA was used as a rst-line treatment. Patients had PAF with a mean time since diagnosis of more than 1 year in 85.6% of cases.

Procedural characteristics
Procedure characteristics were analyzed from a total of 1741 CBA procedures (Table 2). In nearly 65% of patients pre-procedural imaging was performed to assess PV anatomy. Most patients (1316 patients; 76.5%) had 4 independent PVs. Only 15% of procedures were performed under general anesthesia. The 28 mm Arctic Front Advance cryo-balloon was used in 93.8% of the cases. A bonus freeze strategy was systematically employed in 553 procedures (33.7%) and was never employed in 679 (41.4%) procedures.
In 407 procedures (24.8%), a bonus application was administered depending on the perception of the quality of the previous application by the operator. Procedure times (cryotherapy, left atrium, total procedure, and x-ray exposure times) are shown in Table 3. Mean number of applications and mean total cryotherapy time was 7.2 ± 2.67 and 21.1 ± 7.80 minutes per patient, respectively. Patients treated with a no-bonus strategy received signi cantly fewer (but longer duration) applications, had a shorter time to effect (TTE), and received signi cantly shorter cryotherapy, LA and uoroscopy exposure times (p < 0.001); however, total procedure time did not reach statistical signi cance between no-bonus and bonus strategies (p = 0.07). Figure 1 shows the distribution of total cryotherapy (A) and procedural times (B).  .085 Quantitative data are expressed as (mean ± SD)

Acute procedural outcomes and applications analysis
A total of 12495 cryoapplications applied to 6715 PVs (277 left and 30 right common trunks) were performed, for a mean of 1.8 applications per vein. At the end of the procedure 97% of PV were isolated, of which 75.8% were isolated with the rst cryoapplication. In 13% of the cases patients received only 1 application per vein for PVI (4 applications per intervention). The rest of the patients (86.2%) received more than 1 application in any of the locations. Global and per vein data analysis are summarized in Table 4. Potentials inside the PV could be monitored during ablation in 61% of PVs with a mean TTE of 52.9 ± 37.02 seconds. Figure 1 shows the histograms of the distribution of TTE (C), temperature at PVI (D) and the minimal temperature reached (E). The most frequent cryoapplications duration was 180 or

Periprocedural-related complications
Periprocedural and procedural-related complications occurred in 120 patients (6.9%). Acute intraprocedural complications were reported in 77 patients (4.4%). The most frequent adverse event observed was phrenic nerve injury (49 patients; 3%), of which 36 (73.5%) recovered immediately. All but 3 patients had recovered normal phrenic motility at time of study exit. Other complications occurred in less than 1% of the patients and are summarized in Table 5  Five deaths (0.29%) occurred during follow-up, 2 (0.12%) of which occurred during the rst 30 days after the procedure. One death within 30 days was due to a traumatic cerebral hemorrhage and the other due to a mesenteric embolism (the patient was not taking the correct anticoagulant dosage). Three patients died between 140 and 308 days post procedure, one patient died as a result of lung neoplasia, one due to a complication after a left atrial appendage occlusion, and one due to an asystole documented bay emergency service (no more data were collected  Fig. 2A). The 12-month estimate of freedom from a ≥ 30s recurrence of AF in patients with PAF was superior to those with PerAF (80.6% CI: 78.1%-81.1% vs 73.3% CI: 68.8%-77.8% respectively; p < 0.001; Fig. 2B).
Baseline patient characteristics and procedural variables were analyzed to identify predictors of AF recurrence (Table 6). Univariate analyses identi ed the following baseline characteristics predicted AF recurrence: ≥65 years of age, female gender, non-PAF, CHA 2 DS 2 -VASc ≥ 2, no physical activity, structural heart disease, LVEF ≤ 50%, and LA enlargement. The following predictors were found close to the limit of signi cance: using a bonus strategy (p = 0.098) and duration of AF history > 1 year (p = 0.098). A multivariate model identi ed the following independent predictors of AF recurrence: non-PAF (OR = 1.70, p < 0.001), LA enlargement (OR = 1.35, p = 0.017), and female gender (OR = 1.33, p = 0.039). In total, 276 patients (17%) had AF recurrence during the blanking period, and AF recurrence during the blanking period was a signi cant post-procedural predictor of arrhythmia recurrence (HR = 5.98, p < 0.001; HR = 6.06, p < 0.001 for univariate and multivariate analysis respectively). Figure 3 shows Kaplan-Meier free survival curves for highlighted independent and potential predictors AF recurrence. The AF recurrence rate did not differ based on center volume in this study (p = ns) (Fig. 4). However, procedure time, uoroscopy time, cryoablation time, and percentage of isolated veins was improved in more experienced centers, and a bonus cryoapplication strategy was used more often in experienced centers (all p < 0.001). The adverse event rate was not different between centers grouped by experience (p = 0.1) (S4, Of the 1628 patients who completed follow-up, 1038 were on AADs (63.8%), with a similar proportion in PAF (63%) and PerAF (65.8%). The percentage of patients free of AF recurrence on AADs at 12-month follow-up was also similar between PAF and PerAF groups (29.5% vs 28.9 respectively, p = ns). Patients discharged under AAD treatment were more likely to continue AADs at the end of follow-up, irrespective of AF recurrence (29% versus 12% in the PAF group; p = 0.002). Figure 5 displays the cohort according to type of AF, AAD at discharge, recurrences and AAD at the nal follow-up.

Discussion
The RECABA registry is the largest cryoballoon registry in Spain and re ects the standard clinical practice and clinical characteristics of patients treated by CBA for PVI in Spanish hospitals. AF is the most frequently ablated arrhythmia in Spain (27.8% of all arrhythmias), and last year, 42% of these procedures were performed with cryoballoon. 7 The results of RECABA demonstrated acute and subacute complications associated with real-world usage of CBA for treatment of AF were rare (6.9%), and cryoablation was effective with 79% freedom from AF at 12-mo when used according to standard clinical practice in 27 unique centers in Spain.

Tailored and E cient Cryoballoon Ablation Procedure
Patients in RECABA were mostly young, overweight adult males who presented with PAF without structural heart disease and are consistent with baseline characteristics of patients enrolled in other European, multicenter registries. 5,6,9 . The cryoablation procedure was frequently performed under super cial sedation, which has been reported to result in shorter procedure times without compromising AF recurrences or complication rates. 10 During recent years, cryothermal dosing protocols have also evolved towards shorter application times based on per-PV TTE with or without bonus applications, and have resulted in more e cient procedure times than xed cryoapplication times. [11][12][13] This trend was also observed in RECABA in which two thirds of the centers systematically used dosing protocols without bonus applications that led to signi cantly fewer applications per patient and shorter procedure, cryotherapy, and LA times. Centers with more experience applied the bonus strategy to a greater extent; although, there is no clear explanation for this observation. It is possible these centers perceived a greater need for applications as they treated more patients with PerAF; thus, it is important to attempt PV potential monitoring during ablation to tailor dosing to each PV. In the RECABA registry, PV potentials could be seen during ablation in 61% of the veins and more often in left PVs. These results are similar to contemporary reports, and rates have improved with the introduction of new generations of cryoballoon 11,14 . General procedure times in the RECABA registry, such as cryotherapy, LA and x-ray exposure times were comparable to those obtained in other studies, and they were signi cantly reduced in non-bonus protocols. 5,7,9,11 These data corroborate a trend toward simplifying and maximizing e ciency of AF ablation procedures.
Antiarrhythmic drugs on long-term outcomes There are mixed reports on the in uence of short-term use of AADs on long-term outcomes after AF ablation. [15][16][17] The European survey of AF ablation practice that included 1300 patients from 72 European institutions found that among a patient population that was more than two-thirds PAF, 65% of patients were discharged post-ablation on an antiarrhythmic medication and 49% remained on antiarrhythmic medication at 12 months. 18 The RECABA study observed similar rates of AAD usage after discharge.
Although further randomized data are required, patients discharged on AADs were more likely to continue AAD use throughout study follow-up even in the absence of arrhythmia recurrence. The reasons for this may range from a perception of added e cacy of the AAD or its additional effect on the ablation, but more probably it is due to the nature of a registry and lack of strict control in the follow-up. In the group of PerAF, however, the differences were non-signi cant (Fig. 5).

Predictors of outcomes after cryoablation
Independent predictors for AF recurrence were female sex, non-PAF, LA enlargement, and early recurrence in blanking period. Non-PAF and LA enlargement are a well-stablished predictors of AF recurrence 19 . In the recent CRYO4PERSISTENT AF Trial 20 , a controlled multicenter study, a single CBA for treatment of PerAF demonstrated 61% success at 12 months and improved quality of life which is in accordance with the data of the current study. A sub-analysis of the Fire and Ice Trial showed that female sex was associated with an almost 40% increase in the risk of AF recurrence and cardiovascular rehospitalization after PV isolation 21 . In the RECABA study we found early recurrence in blanking period increased the risk for late AF recurrence 6-fold. Lack of recurrence in the blanking period, however, had less predictive power in the PerAF group. In accordance with this nding, early AF recurrence in blanking period (classically considered as non-valuable and unrelated with late AF recurrence, especially in radiofrequency PV ablations) has been found as a strong predictor of long-term AF recurrence in several recent studies 22 , especially when it occurs in the second or third month after ablation. These data align with some authors suggestion that recurrences during the blanking period should be rede ned. 23 Interestingly, the independent predictors of AF recurrence within the RECABA registry were in general alignment with the ndings of large-randomized trials, as well as the safety and e cacy data. Consequently, the RECABA results demonstrate the value of using real-world registries to further validate study ndings from randomized controlled trials.

Study Limitations
The RECABA registry was a multicenter prospective observational registry; therefore, acknowledged limitations include potential bias in patient selection, varied patient management, and the lack of a control group. Nevertheless, possible biases are mitigated by the fact that data were collected prospectively, and research endpoints were pre-speci ed. As per nature of this project, a standardizedcryoballoon dosing protocol was not implemented. Few patients were implanted with an internal looprecorder; therefore, asymptomatic episodes may have occurred unnoticed, and our success rate may have been over-estimated. Complications that occurred during the ablation procedure may be registered adequately in a registry, but the voluntary nature of data provision may have led to underreporting of procedural complications after the ablation. No recommendations were provided to the participating centers in terms of pharmacological treatment following CBA. Thus, the exact temporal sequence of AAD management could not be established, and it is unknown whether AADs were simply continued after the blanking period as per center practice. Despite limitations, this prospective research may provide a representative analysis of the real-world outcomes of CBA in a broad cohort of patients with AF in a large number of centers with different levels of experience.

Conclusions
Cryoablation was effective (79% freedom from AF at 12-mo) and safe (6.9% adverse event rate) when used according to standard clinical practice during the early cryoballoon ablation experience at 27 unique centers in Spain. LA enlargement, non-PAF, female sex and early AF recurrence in the rst 3 months independently predicted late AF recurrence.

Clinical competencies
Cryoablation for the treatment of AF performed according to standard clinical practice in 27 centers was safe and effective.
Cryoballoon ablation is commonly delivered with a tailored, e cient approach.
Female sex, non-PAF, LA enlargement, and early recurrence in blanking period predicted recurrence during 12-mo follow-up.

Translational Outlook
Cryoballoon ablation is used to treat a broad group of patients with atrial brillation and results in consistent outcomes. Distribution of Observed Procedural Characteristics.

Figure 2
Freedom from atrial brillation over study follow-up.

Figure 3
Analysis of different potential predictors of AF recurrence.

Figure 4
AF recurrence and no-recurrence at 12 months of follow-up (absolute number of patients).