This is a retrospective cohort analysis conducted at a tertiary care center in Poland performing approximately 700 ablations per year. A total of 49 procedures of AAFL were performed between November 2019 and January 2023. Four procedures were excluded from this analysis because of second AAFL ablations. Patients were always qualified for the CA after at least a 3-month blanking period post-PVI. Every procedure was performed by experienced operators, who perform > 50 PVIs yearly. The inclusion criteria were: ≥ 18 years of age, history of PVI, and qualification for the first AAFL CA. Patients with an idiopathic AAFL or related to prior surgical procedures were excluded from this analysis. The study was conducted per the Declaration of Helsinki’s ethical principles. The protocol of the investigation was approved by the local Bioethics Committee (approval number: AKBE/127/2022). All patients signed informed consent to the processing of personal data.

The symptoms were assessed with the EHRA scale: EHRA I—‘No symptoms’; EHRA II—‘Mild symptoms’; normal daily activity not affected; EHRA III—‘Severe symptoms’; normal daily activity affected; EHRA IV—‘Disabling symptoms’; normal daily activity discontinued. The majority of PVI procedure was preceded with trans-esophageal echocardiography (TEE) to rule out an intracardiac thrombus and assess the anatomy of the interatrial septum. In some low-risk patients, treated continuously with anticoagulants and with sinus rhythm on the day of the procedure TEE was revoked.

All catheters were inserted under local anesthesia through three vein punctures. During the procedure, unfractionated heparin was infused according to the activated coagulation time (target: > 335 s); the first bolus dose (120 IU/kg) was administered before trans-septal puncture. In all patients with left-sided arrhythmia double trans-septal puncture was performed with subsequent introduction of one unsteerable sheath for PentaRay NAV Eco catheter (Biosense Wenster Inc., CA, USA) and one steerable sheath for ThermoCool Smarttouch Surround Flow (SF) or QDot MICRO catheter (Biosense Wenster Inc., CA, USA). Intra-procedural pain was managed with an opioid (mostly continuous remifentanil infusion). Some patients required sedation with midazolam boluses introduced at an operator’s discretion. In most procedures, a three-dimensional reconstruction of the left atrium (LA) and pulmonary veins (PV) was created using rotational angiography if this has not been performed during previous PVI procedure. Bipolar voltage mapping was generated using a PentaRayTM catheter as well as a CARTO electroanatomic navigating system.

A total of 447 patients underwent AF ablation at the Medical University of Warsaw between November 2019 and January 2023, with 362 first-time sessions and 85 repeat sessions. From the group of patients who underwent first-time AF ablation, AAFL during initial PVI occurred in 7 patients (1.93%). AAFL was induced during maneuvering with catheters in 4 patients and in case of 3 patients AAFL was observed after AF cardioversion. No antazoline and no pacing maneuvers are performed in our center during first-time PVI procedure. On the other hand, of 38 patients with repeat ablation, PV reconnection was observed in 13 patients. After confirming or completing PV isolation, AAFL was induced by administration of antazoline in 4 patients; in 13 patients, AAFL was induced with pacing maneuvers. AAFL spontaneously occurred in the other 21 patients. Then, a total of 49 patients experienced AAFL in repeat sessions. Excluding 4 patients with second AAFL ablation, we examined a total of 45 patients (7 first-time and 38 repeat sessions) who underwent the first-time AAFL ablation. Moreover, it should be emphasized that as a referral center, some procedures were previously performed elsewhere.

For patients in sinus rhythm, stimulation maneuvers and isoprenaline infusion were applied for arrhythmia induction. For patients presenting in AAFL at the beginning of the procedure, the coronary sinus (CS) catheter activation was analyzed. In case of a distal to proximal activation, left atrial access was obtained and the left atrium (LA) was mapped first. In patients with a proximal to distal CS activation or whenever a right atrium (RA) AAFL was suspected, mapping of the RA was performed first. However, the trans-septal puncture was also performed even in right-sided ablations to ensure the durability of previously performed PVI. In patients with AF at the beginning of the procedure, antazoline was administered. Up to 400 mg during 10 min, infusion of antazoline has been administered. In the case of arrhythmia conversion from AF to AAFl, patients were included in this study. Local activation time mapping was performed with high-density point collection. Windows of interest were set to 90% of the total cycle length. CS signals were used for local activation point annotation. Both isochronal and propagation maps were generated at the discretion of the operator. Slow-conducting zones were defined by electrogram (EGM) of low-amplitude and isochronal crowding. Efforts were made to map the entire AAFL circuit. Voltage mapping was generally performed using < 0.05 mV as low voltage or scar and > 0.5 mV as healthy tissue. Moreover, entrainment was minimized and used in case of the inability to identify the AAFL circuit and critical isthmus with high-density mapping. Following beat, acceptance criteria were applied during automated point acquisition: (1) tachycardia cycle length stability ± 10 ms); (2) time stability of a reference EGM from the coronary sinus catheter (± 5 ms); (3) beat-to-beat EGM consistency; (4) EGMs acquired during the expiration phase; (5) EGMs acquired during stable position of the catheter (< 1 mm catheter movement during the acquisition window); and (6) tracking quality of the mapping catheter to ensure correct location of the mapping catheter. Manual re-annotations of doubtful EGMs were performed in a small subset of cases. In all patients, AP and RL diameters of LA were measured based on the electroanatomical map. Areas with voltage ≥ 0.3 mV during AF/AFL and > 0.5 mV during sinus rhythm have been qualified as low voltage. Patients were divided into two groups of low voltage (at least 25% of LA area with ow voltage) and regular voltage (the remaining part). Voltage areas were analyzed only if at least half of the expected chamber volume has been mapped.

AI-guided ablations with the use of Thermocool Smarttouch SF or QDot MICRO catheter were performed by adapting the CLOSE protocol. Power output was 35 W with a target AI of > 400 at the superior, posterior, and inferior wall of the left atrium and > 550 at the remaining sites. Other settings included: aimed CF of 10–30 g, irrigation rate of 15 mL/min, a maximum inter-lesion distance of 6 mm, and maximum temperature cut-off of 40 °C. Ablation of AAFL aimed to restore sinus rhythm and make a durable ablation line in the atrium, which would prevent arrhythmia recurrence. A two-sided conduction block was confirmed after the last application. Additionally, all patients had pulmonary veins checked and re-isolated in case of reconnection. Procedural success was confirmed when an entrance block to all pulmonary veins was proved. In each case, 15-min waiting period was ordered, when both stimulation maneuvers and isoprenaline infusion took place. If no arrhythmia was induced, the procedure was finished. Pacing maneuvers were made only in patients with history of AAFL. In case of arrhythmia induction, additional ablations were performed. If the arrhythmia was constantly inducible, it was left to the operator whether to continue procedure considering potential gain and risk of next ablations. However, in each case, cardioversion has been made before procedure completion. In 7/45 (15.6%) patients, AAFL occurred during the first PVI procedure. Based on the workflow used in our center, in case of registration of regular, sustained atrial arrhythmia, we perform the remap of the left atrium and ablate the arrhythmia substrate. In 13/45 (28.9%) patients, we performed re-isolation of pulmonary veins because of reconnections, which was always the first step of the procedure (before linear ablations).

Pericardial effusion and other intracardiac complications were excluded by transthoracic echocardiography performed twice: immediately after the procedure and in the morning of the following day, before being discharged from the hospital. Subsequent antiarrhythmic treatment was prescribed at the discretion of the operator.

The primary endpoint was freedom from any > 30 s, of atrial arrhythmia recorded on ECG after the procedure. The secondary aim of this study was to establish the predictors of arrhythmia recurrence. Additionally, patient satisfaction and symptoms after the CA were analyzed.

All aforementioned information was extracted from the medical records. Clinical follow-up data were gathered during a routine post-discharge appointment at the outpatient clinic. Each patient had a 24-h Holter scheduled at 3 and 12 months after ablation. Then patients were contacted by telephone for additional information including any post-ablation symptoms of arrhythmia, documentation of potential recurrence, and precise time when it occurred. No blanking period was established.

Distributions of continuous variables were assessed with the Shapiro–Wilk test. For variables with non-normal distribution, the results are presented as median and interquartile range, while variables with normal distribution are presented as mean ± standard deviation. Categorical variables are presented as percentages. Fisher exact test was used for comparing categorical variables, and Mann–Whitney U test was used for continuous variables. To assess the risk of atrial arrhythmia recurrences, univariate and multivariate stepwise Cox proportional hazards regressions were performed, assuming the p entry value as 0.2 and the p stay value as 0.1. To assess the influence of measured parameters on impedance level, Spearman correlations were calculated. Kaplan–Meier survival curves were plotted for an analysis of the atrial arrhythmia recurrences. A p value of < 0.05 was considered statistically significant. The statistical analysis was performed using Statistical Analysis Software (Cary, NC, USA), version 9.4.