This experience reports real-world clinical efficacy of VGLB ablation as part of standard practice for treatment of both paroxysmal and short-standing persistent AF (< 1 year) using PVI-only ablation. Acute electrical isolation could be achieved in 98.2% of PVs. Acute procedural failure happened in 9 patients due to infrequent causes such as technical problems (console error), phrenic nerve paralysis, or inability to achieve permanent isolation of the PV due to refractory conduction in a suspected very thick myocardium. Previous cohort studies with VGLB ablation showed similar acute success rates varying between 97.7 and 100% [9,10,11,12,13,14,15,16, 21].

Following application of a first circumferential ablation lesion, 92.5% of PVs were isolated. Although in earlier studies first-pass success was lower (68 to 83%) [9, 11, 12], more recent publications report similar results (88 to 95%) [13, 14], depending on optimal tissue contact and stable catheter position [9], high-dose (> 8.5W) energy application [10], and growing experience [15]. Our high first-pass PVI success rate may be explained by our ablation approach: rotational angiography was performed in all patients ensuring optimal catheter positioning, optimal tissue contact was pursued in all PV anatomies, and high-dose energy applications (median 10 Watt) were used. These strategies also likely contributed to the long-term durability of ablation lesions. The number of lesions applied per vein was similar to those reported by other authors [9,10,11, 15, 17]. Since all VGLB procedures in our center were performed by only two operators, the operator’s experience advanced over a relatively short period of time, reaching optimal procedure time and radiation dose after approximately 25–50 cases each.

At 1-year follow-up, we report freedom of AF in 88.4% of paroxysmal and 87.7% of persistent AF patients. Previous reports showed 1 year freedom of AF ranging from 60 to 83% for paroxysmal AF [9,10,11,12,13,14,15, 18], and 70 to 75% for persistent AF [16, 17].

Available long-term results for VGLB ablation are limited. Sediva et al.[13] conducted a clinical follow-up of 192 patients with almost exclusively paroxysmal AF, reaching 75% freedom of AF after 4 years, but only 32 paroxysmal AF patients reached 4-year follow-up and none of the persistent AF patients reached 2-year follow-up. Reissmann et al.[21] conducted a follow-up of 90 patients with exclusively paroxysmal AF. Twenty-three patients reached 5-year follow-up, with a reported freedom of paroxysmal AF of 51%. In our current report, 98 patients (including 30 persistent AF patients) reached 4 years follow-up and 20 patients (including 4 persistent AF patients) reached 5-year follow-up. Estimated freedom of AF reached 78.8% for paroxysmal and 68% for persistent AF patients at 4-year follow-up and 78.8% for paroxysmal and 65.3% for persistent AF patients at 5-year follow-up. Our report equals the longest follow-up for paroxysmal AF patients with comparable results and is the first to document long-term follow-up for persistent AF patients treated with VGLB.

As expected, sustained freedom of AF was higher in patients with paroxysmal AF as compared to persistent AF, but not statistically significant (p = 0.108). This could be due to patient selection since VGLB was offered to persistent AF patients without extended arrhythmogenic substrate or extremely dilated atria. The numerical difference, however, was rather large, and with more power, statistical significance would possibly have been met. Subgroup analyses were performed based on the presence of certain risk factors, such as hypertension, obesity, or sleep apnea, but no statistically significant differences in the recurrence of AF were documented. Besides, the effect of these risk factors seems more convincing on the occurrence of AF in patients without previous AF ablation [22]. Although stratification by left atrial dimensions was expected to show more frequent recurrence of AF in patients with severe atrial enlargement, our results do not support this (p = 0.117).

Ease of use of the VGLB catheter resulted in a relatively short learning curve with an initial phase of around 15 procedures and a final phase at 25–50 procedures. Despite no previous experience with VGLB ablation, median procedure duration (129 min, IQR 113–150 min) and fluoroscopy times (15 min, IQR 11–20 min) were low and decreased progressively over time. Similar results are reported by some [9, 10, 16], although most authors report longer procedure durations and fluoroscopy times [11,12,13,14]. Studies comparing first-generation VGLB ablation to RFA [13, 16, 17] or cryoballoon ablation [18, 19] performed by experienced operators, failed to demonstrate different procedure duration.

Redo procedures were performed in 21.1% of cases. In less than half (46.9% of redo procedures, 9.9% of all procedures) PV reconnection was documented. Most redo procedures required additional substrate ablation. The median number of PV reconnections during redo procedures was 0 [IQR 0–2]. When PV reconnection was present, it manifested most frequently in the right inferior PV and left superior PV. This might be explained by difficulties obtaining optimal tissue contact at the ostium of the right inferior PV using the first-generation VGLB catheter. Second- and third-generation VGLB catheters incorporate an ultra-compliant, pear-shaped balloon with adaptable size that in our experience improves the ability to obtain contact in these areas. These second- and third-generation catheters were not used in our study cohort. Left-sided PV reconnection was often located near the appendicular ridge or anterior carina where conduction sleeves are often situated in deeper tissue layers. Right inferior and left-sided reconnections are also common with other ablation techniques, such as RFA and cryoballoon [23]. None of the patients undergoing redo ablation due to AT showed PV reconnection.

Adverse events were limited. Minor vascular complications occurred in 4.6% of patients, possibly due to new experience with a 15F femoral sheath. As a result, we no longer perform femoral invasive arterial monitoring and access points are now closed with a figure-of-8 suture instead of manual compression, largely eliminating vascular complications. Transient phrenic nerve injury occurred in 3.3% and pericardial effusion without tamponade in 1.3% of patients. These results are similar to other reports [10,11,12,13,14,15,16]. Phrenic nerve injury most commonly occurred when the VGLB catheter had to be pushed and advanced slightly inside the ostium of the right superior PV to obtain optimal tissue contact, thereby pushing the target tissue closer towards the phrenic nerve. Second- and third-generation VGLB catheters consisting of a more compliant balloon, allowing more proximal positioning and minimizing the need for pushing might reduce phrenic nerve injury rates. There were no reports of PV stenosis, likely due to the use of 3D rotational angiography and direct endoscopic visualization, allowing for reliable balloon positioning well outside of the PVs.

This retrospective report comprises a real-world experience, posing some limitations. Although routine follow-up after 3 months and 1 year was conducted at our center, further follow-up was left to the referring cardiologist. As such, long-term results are based on the patients’ latest contact with their cardiologist. This limitation is partially overcome by the low rate of patients lost to follow-up. Also notable is the post-procedural continuation of AADs in 23% of patients, due to concomitant arrhythmias (e.g., residual symptomatic atrial ectopy, but mostly ventricular ectopy or non-sustained ventricular tachycardia). Continuation of betablockers and calcium antagonists could be explained by their anti-hypertensive properties.

The presence of left atrial enlargement based on measurement of parasternal long axis dimensions only could be variable. Unfortunately, left atrial volume data were not available for many patients (n = 49/152) and therefore were not included in the dataset. Our report reflects a single-center experience, performing PVI only. Non-PV triggers during the index procedure were not treated. This report only contains results for the first-generation VGLB catheter. During follow-up of this cohort, the system was updated via the second and third generation VGLB systems though the energy source remained unchanged. Third-generation VGLB has been shown to significantly decrease procedure times, maintaining similar acute procedural results [24]. Long-term follow-up results for the third generation VGLB have not been published yet. To further assess clinical efficacy of this technique, larger, randomized studies are required.

Other ablation modalities are not discussed in this report. Previously published long-term results for RFA and cryoballoon PVI are slightly less optimistic [25,26,27]. Long-term results for promising new ablation modalities i.c. pulsed field ablation are not yet available.