This study retrospectively investigated the incidence of spontaneous conversion to SR and its predicting factors in patients who underwent attempted AVNA and CSP for symptomatic AF between June 2018 and December 2022 at University Medical Centre Ljubljana. Consecutive patients with symptomatic AF refractory to rhythm and pharmacological rate control therapy were included. Patients with severe valvular disease were excluded. SR restoration was defined as the spontaneous conversion to SR documented on 12-lead electrocardiogram (ECG) in the device clinic during follow-up after AVNA and CSP.

All patients were asked for written informed consent before data collection. The study design was approved by the Republic of Slovenia National Medical Ethics Committee.

Pacing device implantations were always performed first, followed by AVNA, preferably during the same hospitalization. All device implantations were performed by two experienced operators.

2.2.1 His Bundle Pacing

The HBP procedures were conducted as previously reported [13, 19, 20]. His bundle (HB) potential mapping was performed using a continuous recording of intracardiac electrograms with the electrophysiological recording system (EP-TRACER 2 Portable CardioTek B.V., Sittard, The Netherlands or LAB system Pro, BARD Boston Scientific, Lowell, MA, USA). The tricuspid valve annulus ring was imaged by contrast angiography to facilitate HB localization. The sheath and the pacing lead were advanced to the HB area, where larger ventricular and smaller atrial signals were detected (ventricular to atrial electrogram ratio at least 3:1). The pacing lead was screwed into position, and threshold measurement was performed at the pulse width of 1 ms. HBP threshold $\le$2.5 V at a pulse width of 1.0 ms was considered appropriate. Additional backup pacing or implantable cardioverter-defibrillator (ICD) lead was implanted and connected to the ventricular port of the pacing device or IS-1 port of a DF-1 ICD device, while HBP lead was connected to the atrial port of the dual-chamber device. Schematic illustration and fluoroscopic view of the pacemaker lead position are presented in Fig. 1A.

Fig. 1.

Schematic representation of pacemaker (PM) lead positions (1) and their relation to the ablation catheter under fluoroscopy (2). (A) His bundle pacing (HBP). (B) Left bundle branch pacing (LBBP). For clarity, the backup pacing lead is not illustrated.

Twelve-lead ECGs were assessed at baseline and each outpatient follow-up visit: at 1 month, 6 months, and every 6 months thereafter. Assessment of clinical outcomes (echocardiographic parameters, laboratory parameters, and symptomatic evaluation) was performed at baseline and at 6 months or immediately after SR on the ECG strip was detected.

Categorical variables are reported as frequencies and percentages and were compared using Chi-square and Fisher exact test as appropriate. Continuous variables are expressed as mean $\pm$ standard deviation or as median (interquartile range), according to the distribution. Kolmogorov–Smirnov test was used to test the normality of distribution. Intra- and intergroup differences were compared with the use of independent or paired sample Student t-test, Wilcoxon rank-sum test, and Wilcoxon signed-rank test as appropriate. The multivariate logistic regression model was used to analyze predictors of spontaneous conversion to sinus rhythm. All hypotheses were two-tailed, and a p-value $\le$ 0.05 was considered significant. Statistical analysis was performed in IBM SPSS Statistics for Windows, version 25.0 (IBM Corp., Armonk, NY, USA).

Baseline patient characteristics according to the occurrence of spontaneous conversion to SR are presented in Table 1. Sixty-eight consecutive patients undergoing CSP combined with AVNA were included. The mean age of the patients was 71 $\pm$ 8 years, and 29 (42.6%) were male. The baseline QRS width was 128 ms (110–140), and twelve patients (17.6%) had atypical atrial flutter. Patients were highly symptomatic with the median New York Heart Association (NYHA) class 3. The mean left ventricular ejection fraction (LVEF) was reduced to 40 $\pm$ 15%, and the atria were enlarged with the median left atrial volume index (LAVI) of 59 mL/m${}^2$ (51–72). Thirty-seven patients (54.4%) received HBP and 31 (45.6%) LBBP.

The median follow-up time was 16 months (6–27). Spontaneous conversion to SR during follow-up was registered in 6 patients (8.8%); 3 in the HBP group and 3 in the LBBP group. In patients who converted to SR, baseline LAVI was smaller (45 mL/m${}^2$ (41–51) vs. 60 mL/m${}^2$ (52–75); p = 0.002) (Table 1). Due to cardiogenic shock, one of the SR patients required temporal circulatory support with veno-arterial (VA) extracorporeal membrane oxygenation (ECMO), which was applied 24 hours before pacemaker implantation and AVNA. SR on ECG was first detected after the median follow-up of 4.5 months (2–24).

To further clarify the predictors of spontaneous conversion to SR after AVNA, we performed a multiple regression analysis (Table 2). Covariates that were considered clinically relevant were age, gender, baseline LVEF, LAVI, and indexed left ventricular end-diastolic volume index (LVEDVi). Even after consideration of these covariates, LAVI remained a significant predictor for conversion to SR (odds ratio 1.273, 95% confidence interval [1.027, 1.578], p = 0.028).

All device implantations and subsequent AVNAs were successful without any acute adverse events. Apart from two patients in whom an existing ICD device was upgraded with a CSP lead, all other procedures were de-novo implantations. The median pacemaker implantation fluoroscopy time was 6 minutes (4.2–8.1). An ICD device for primary prevention was used in 3 (8%) patients who received HBP and 7 (22%) patients who underwent LBBP. Additional atrial lead was implanted in only one patient. Periprocedural increase in HBP threshold after AVNA was documented in 1 patient. However, the HB capture was maintained at higher outputs, and the lead revision was not required.

At baseline, NYHA class of patients who converted to SR (SR group) did not differ from patients who remained in atrial arrhythmia (non-sinus rhythm (NSR) group) (p = 0.082). As reported in Table 3, NYHA class improvement was registered regardless of SR restoration (p = 0.026 for SR; p 0.001 for NSR). In the SR group, 2 patients improved for 1 NYHA class, 3 patients improved for 2 NYHA classes, and 1 patient improved for 3 NYHA classes. In the NSR group, 27 patients improved for 1 class, 19 patients improved for 2 classes, and 1 patient improved for 3 NYHA classes. No change in NYHA class was detected in 15 patients. No patient deteriorated in any group. In all patients, digoxin and amiodarone were discontinued, and dosages of beta-blockers were reduced. In the NSR group, diuretics were discontinued in 15 of 39 patients who were receiving them at baseline (28.0%). In the SR group, they were discontinued in 1 of 3 patients (33.3%); however, the difference did not reach statistical significance. Other HF therapy did not change during follow-up.

At baseline, median N-terminal pro-b-type natriuretic peptide (NT-proBNP) was 2969 (1569–3635) pg/mL and did not differ between both groups (p = 0.339). At follow-up, NT-proBNP decreased in both groups, although the decrease in the SR group, despite being relatively higher, did not reach statistical significance due to the smaller sample size.

Four patients in the NSR group (3 with HBP and one with LBBP) died during follow-up. While the death in the LBBP group was associated with progressive HF, the other 3 deaths were determined as non-cardiac.

At baseline, 6 patients had left bundle branch block (LBBB), and 2 patients had right bundle branch block (RBBB). While none of these 8 patients converted to sinus rhythm, there was no difference in baseline QRS width between SR (108 ms (94–119)) and NSR group (109 ms (95–120)), p = 0.792. Post-procedural QRS width was similar to baseline QRS (p = 0.109 for SR; p = 0.08 for NSR).

While there were no significant differences in baseline LVEF (p = 1), LVEDVi (p = 0.214), and indexed LV systolic volume (LVESVi) (p = 0.311), baseline LAVI was, as previously described, significantly smaller in patients who converted to SR (p = 0.002) (Table 4). At follow-up, LAVI did not change in any group. The increase in LVEF was numerically comparable in both groups, although, in the SR group, it did not reach statistical significance due to the smaller sample size. Similarly, LVESVi decreased in both groups; however, in the SR group, the decrease did not reach statistical significance due to both smaller sample size and smaller initial volumes. As for LVESVi, we observed consistent changes in LVEDVi. A comparison of mean changes in echocardiographic parameters is presented in Fig. 2. The follow-up electrocardiogram and echocardiographic mitral inflow pattern of the patient who converted to SR after LBBP with subsequent AVNA are presented in Fig. 3.

Fig. 2.

Comparison of the mean ($\mathrm{\pm }$SD) changes in echocardiographic left ventricular ejection fraction and volumes between baseline and follow-up according to conversion to sinus rhythm. Legend: LV, left ventricular; SR, sinus rhythm; NSR, non-sinus rhythm; EDVi, end-diastolic volume index; ESVi, end-systolic volume index.

Fig. 3.

12-lead ECG (A) and mitral inflow pattern (B) of the patient who converted to sinus rhythm after left bundle branch pacing and atrioventricular node ablation. Note the VVI pacing mode and atrioventricular dissociation due to the lack of atrial lead. ECG, electrocardiogram.

Larger studies are warranted to clarify the predicting variables of SR restoration in patients scheduled for an “ablate and pace” strategy with CSP. The incidence of SR restoration is clinically important as these patients may require an upgrade to a dual-chamber device to ensure AV synchrony that could further maintain SR and prevent potential pacemaker syndrome. On the contrary, initial atrial lead implantation in all patients may imply an unnecessary increase in lead burden. Therefore, the ability to identify which patients are expected to experience SR during follow-up is certainly important to optimally manage these patients.

The retrospective design of this single-center study with a relatively small number of included patients limits the strength of our findings. Furthermore, a smaller sample size in the SR group might have affected the outcomes compared to NSR. The potential exclusion of patients with severe valvular disease might have had an impact on the results of the study, while the probability of SR restoration in these patients is very low. However, none of the patients were, in fact, excluded due to this exclusion criterion. The study only recorded SR restoration in the device clinic during regular follow-ups that were documented on 12-lead ECG, while potential intermittent conversions to sinus rhythm were not detected. However, persistent sinus rhythm was not achieved in the NSR group. The data on the duration of AF before the ‘pace and ablate’ intervention could not be obtained in several patients. As some of the patients who converted to SR were urgently admitted to the hospital due to acute decompensated HF, a possible shorter duration of AF in these patients could have resulted in less structural and electrical left atrial changes and increased the likelihood of the SR restoration [1, 2, 3]. Therefore, our findings should be interpreted with caution and need to be confirmed in larger studies with longer follow-ups.