Long-term Outcomes of Atrioventricular Node Ablation and Pacemaker Therapy for Rate Control in Atrial Fibrillation: Insights from a 10-year Nationwide Analysis

Effective rate control is crucial in managing patients with persistent or permanent AF, particularly when symptoms are refractory to conventional rate- and rhythm-control therapies (estimated to be approximately one-quarter of patients).^1–3 In patients meeting these criteria, atrioventricular (AV) node ablation (AVNA) combined with permanent pacing – commonly referred to as the ablate-and-pace strategy – represents a well-established and safe approach, promptly improving symptoms, echocardiographic parameters and functional status. Despite these benefits, the optimal pacing strategy following AVNA remains uncertain.^4–8

Permanent pacing from the right ventricle (RV) is historically favoured due to its simplicity, effectiveness and stability. Nevertheless, RV pacing activates the His-Purkinje system in a non-physiological way, potentially causing left ventricular (LV) dyssynchrony, myocardial dysfunction and adverse remodelling, which possibly diminishes the beneficial effects of achieving a regular heart rate.^9,10 CRT, in contrast, provides a more physiological activation of the myocardium, potentially preserving LV function better than RV pacing. Studies comparing RV pacing with CRT after AVNA have shown mixed outcomes, with some indicating clinical superiority of CRT and others demonstrating no substantial benefit beyond structural and functional parameters.^4,11,12 Moreover, most previous studies evaluated surrogate endpoints such as echocardiography and quality-of-life assessments and were underpowered to detect significant differences in outcomes such as long-term survival.^13–15 Therefore, the optimal pacing modality after AVNA remains uncertain to date (as reflected by the class 2a and level of evidence b recommendations in the 2024 European Society of Cardiology [ESC] guidelines for the management of AF).²

To our knowledge, there are no nationwide studies with long-term follow-up (≥10 years) addressing all-cause mortality related to AVNA in patients with AF. Although Estonia’s ablation practices have modernised over the past decade, with AVNA procedures decreasing from 10% to 4.3% (in 2024) of all catheter ablations against a significant increase in AF ablations (from 16% to 45%, respectively), addressing this gap is vital to enhance clinical decision-making, patient choice and risk stratification based on real-world generalisable data.^12,16 Therefore, we conducted this nationwide retrospective cohort study in Estonia to analyse baseline patient characteristics and evaluate long-term survival outcomes following AVNA and permanent pacemaker implantation, with particular emphasis on differences based on pacing strategy and sex.

Methods

Study Design and Population

We conducted a nationwide retrospective observational cohort study of all patients who had received radiofrequency AVNA for AF in Estonia from January 2012 through December 2022. All catheter ablation procedures of the entire country were performed at two tertiary centres (Tartu University Hospital and North Estonia Medical Centre), which were included in the study.

Inclusion criteria were symptomatic persistent or paroxysmal AF with inadequate rate or rhythm control despite optimal pharmacological therapy, in line with ESC guidelines recommending AVNA for AF patients refractory or intolerant to medical therapy and not eligible for AF ablation. Data were extracted from electronic health records, including patient demographics (age, sex), clinical history, medications, baseline echocardiography (left ventricular ejection fraction; LVEF), and ECG parameters (heart rate and QRS duration measured on ECGs performed immediately before and after AVNA), as well as time of death. All patients received permanent pacemaker implantation (either prior to or immediately before AVNA) to maintain ventricular rhythm. The implanted pacemaker type – either conventional RV pacing (single- or dual-chamber devices) or biventricular pacing via CRT – was recorded for each patient. Promptly successful AVNA was defined as permanent AV block achieved by catheter ablation.

Outcomes and Follow-up

The primary endpoint was all-cause mortality, defined as death from any cause. Time of death was obtained from electronic health records to ensure complete follow-up. Survival time was calculated from the date of AVNA until death or censoring at the end of study follow-up (31 December 2024). Patients alive at the end of follow-up were censored on that date.

Statistical Analysis

Statistical analysis was performed using IBM SPSS Statistics 24 (IBM Corp) and RStudio (PBC). Figures were created using SPSS and Python libraries pandas, seaborn and matplotlib.^17–19 Continuous variables are presented as mean ± SD or median (interquartile range) as appropriate and categorical variables as counts (percentage). Group comparisons were performed using the Student’s t-test or Mann–Whitney U-test for continuous data and χ² or Fisher’s exact test for categorical data. Kaplan–Meier survival curves were used to estimate time-to-event distributions for all-cause mortality. Survival curves were stratified by sex and pacing strategy. A subgroup of patients who underwent AVNA ad hoc to pacemaker implantation (during the same hospital admission) was analysed separately to assess the impact of pacing on outcomes in a more homogeneous context and to reduce temporal confounding. Differences between groups were compared with the log-rank test. Univariable and multivariable Cox proportional hazards regression was applied to identify independent predictors of mortality, after verifying the proportional hazards assumption. Variables included in the multivariable analysis were selected a priori based on clinical relevance and published evidence. Additional candidate variables were considered if clinically justified or if exploratory screening suggested potential confounding. Continuous variables were modelled on their natural scale and model complexity was limited to ensure stable estimates. HRs with 95% CI and p-values are reported. A two-sided p<0.05 was considered statistically significant.

Results

Study Population and Baseline Characteristics

A total of 435 AVNA procedures were performed, acute procedural success was confirmed in 418 (96.1%) and 13 (3%) required a repeat procedure. There were four that did not meet the criteria for acute procedural success, but no repeat procedure was performed. Complete follow-up data were available for 430 patients; five (1.1%) did not have follow-up data available and were excluded from further analysis. Among the 430 patients analysed, the mean age at ablation was 73.2 ± 10.2 years, and 56.5% of the cohort were female (n=243). Women were significantly older than men (76.4 ± 9.3 versus 69.2 ± 9.4 years; p<0.001).

At baseline, men demonstrated more impaired cardiac function than women: they had lower LVEF (31.8 ± 14.4% versus 41.6 ± 15.3%; p<0.001) and longer QRS duration (140.4 ± 38.2 ms versus 107.3 ± 31 ms; p<0.001). In contrast, women had higher resting ventricular rates prior to ablation (125.7 ± 23.4 BPM versus 113.2 ± 27.3 BPM; p<0.001) and lower renal function (estimated glomerular filtration rate [eGFR] 51.7 ± 20 ml/min/1.73 m² versus 60.6 ± 21 ml/min/1.73 m²; p<0.001). All characteristics of the study population are presented in Table 1.

Pacemaker Characteristics

In total, 133 patients (30.9%) received CRT and 297 (69.1%) received RV pacing. The distribution of pacemaker type differed significantly by sex: 16.6% (n=41) of women received CRT devices, whereas 52.5% (n=96) of men received CRT (p<0.001).

Among CRT recipients, 100 (75.2%) had standard CRT indications for heart failure (HF) with reduced ejection fraction (HFrEF), including prolonged QRS duration and left bundle branch block, and had a low biventricular pacing percentage (Table 1). Median time from CRT device implantation to AVNA was 70 days in these patients with standard HFrEF indications.

Another 22% (n=29) received CRT ad hoc with AVNA for rate control only (median time from CRT device implantation to AVNA: 1 day). Seven patients initially managed with RV pacing required an upgrade to CRT, and one patient with CRT-defibrillator device underwent heart transplantation 2.1 years after AVNA.

In the RV pacing group, 148 (49.8%) patients received pacemakers ad hoc followed by AVNA for AF rate control only (median time from pacemaker implantation to AVNA: 1 day), whereas 120 (40.4%) had a pre-existing pacemaker for bradyarrhythmia (median time to AVNA: 2.7 years). Twenty-nine patients had previously received ICDs.

Procedural Complications

The overall procedure-related complication rate was low (5.1%; Table 1 ). Three deaths (0.7%) were attributed to acute HF decompensation within 2 weeks post-AVNA. Among those who underwent ad hoc pace-and-ablate, two patients (0.5% of the entire cohort) required reinterventions: one for lead dislodgement 3 months after implantation, and another for pocket infection 6 months post-implantation. Other complications included puncture-site haematoma (3.3%), unclear infection (0.5%) and pleural effusion (0.2%); however, these did not need reintervention.

Survival Outcomes

The median follow-up duration was 4.2 years (2 days to 12.9 years). By study end, 168 patients (39%) had died. Overall survival was similar between sexes: the primary endpoint of all-cause mortality occurred in 100 women (41.2%) and 68 (36.4%) men (HR 1.14; 95% CI [0.84–1.54]; p=0.411). The median survival time was 7.6 years (95% CI [6.7–8.4]) for women and 7.9 years (95% CI [5.5–10.4]) for men (Figure 1 ). The estimated mortality rates at 1 year were 8.2% (95% CI [4.1–12.1]) and 11.5% (95% CI [7.3–15.5]), at 5 years 30.6% (95% CI [22.8–37.6]) and 35.1% (95% CI [28.0–41.5]), and at 10 years 56.1% (95% CI [41.0–67.3]) and 60.7% (95% CI [49.8–69.2]) for men and women, respectively.

Predictors of Mortality

In univariable analysis, age was significantly associated with all-cause mortality (HR 1.06; 95% CI [1.04–1.07]; p<0.001; Supplementary Figures 1 and 2). In a multivariable Cox proportional hazards model including age, sex and LVEF, both age (HR 1.07; 95% CI [1.04–1.09]; p<0.001) and LVEF (HR 0.98; 95% CI [0.97–0.99]; p=0.035) were independent predictors of all-cause mortality. However, sex did not independently predict mortality (HR 0.81; 95% CI [0.58–1.14]; p=0.232) and neither did pacing strategy (CRT versus RV pacing HR 0.93; 95% CI [0.67–1.3]; p=0.683).

LVEF-stratified Survival

When stratified by baseline LVEF, survival did not differ significantly between pacing strategies. Among patients with LVEF ≤35%, 5-year mortality was 38.1% (95% CI [28.2–48.0]) with CRT (n=121) versus 38.4% (95% CI [27.7–49.1]) with RV pacing (n=99) (log-rank p=0.386). At 10 years, mortality reached 59% (95% CI [43.7–74.4]) versus 63.2% (95% CI [48.9–77.6]) (log-rank p=0.345), respectively. In patients with LVEF >35%, 5-year mortality was 8.3% (95% CI [0.0–24.0]) with CRT (n=12) versus 27.6% (95% CI [20.9–34.3]) with RV pacing (n=198) (log-rank p=0.272). At 10 years, mortality was 8.3% (95% CI [0.0–24.0]) with CRT versus 57.5% (95% CI [46.6–68.4]) with RV pacing (log-rank p=0.101), with one and 12 patients at risk, respectively. Findings were consistent in a Cox proportional hazards model including a CRT × LVEF ≤35% interaction, which showed no evidence of effect modification (HR 4.18; 95% CI [0.55–31.39]; p=0.164).

Atrioventricular Node Ablation Ad Hoc with Pacemaker Implantation

We analysed the subgroup of 177 patients who underwent AVNA at the time of pacemaker implantation purely for AF rate control (ad hoc ablate-and-pace). Baseline differences were observed between those who received CRT (n=29) versus RV pacing (n=148) in this setting: patients who received CRT were more often male (65.5% versus 29.7%; p<0.001) and younger (67.5 ± 10.1 years versus 75.6 ± 9.5 years; p<0.001), with much lower LVEF (24.2% versus 42.6%; p<0.001) (Table 2).

Despite these differences, survival in the ad hoc subgroup did not differ by pacing strategy: 5-year survival was ~65% in both CRT and RV groups (log-rank p=0.890; Figure 2). Median survival was 7.3 years (95% CI [2.8–11.8]) for CRT versus 7.5 years (95% CI [5.9–9.2]) for the RV pacing group. In a multivariable Cox proportional hazards model adjusting for sex, age and LVEF, pacing strategy was not significantly associated with mortality (HR 1.91; 95% CI [0.91–4.03]; p=0.089). Furthermore, patients who underwent ad hoc ablate-and-pace had similar survival to those whose devices were implanted earlier for other indications (log-rank p=0.208).

Timing of Atrioventricular Node Ablation after Pacemaker Implantation

We next examined whether the timing of AVNA relative to pacemaker implantation influenced outcomes, stratified by pacing strategy. In CRT recipients, those who underwent AVNA within 6 months (‘early AVNA’) of device implantation had significantly improved survival compared with those ablated >6 months (‘late AVNA’) post-implantation (log-rank p=0.008; Figure 3A). Cox regression analysis confirmed a substantially reduced hazard of death in the early AVNA group (HR 0.47; 95% CI [0.27–0.83]; p=0.010). On multivariable Cox analysis adjusting for potential confounders (including pre-AVNA QRS, post-AVNA QRS, LVEF and renal function), early AVNA remained strongly associated with improved survival (adjusted HR 0.48; 95% CI [0.24–0.96]; p=0.038; Table 3).

Baseline comparisons within the CRT group indicated that patients receiving early AVNA had a somewhat more favourable profile: narrower QRS durations pre- and post-ablation (pre-AVNA QRS 146.9 ± 35.2 ms versus 164.9 ± 31.3 ms; p=0.005; post-AVNA QRS 153.8 ± 32.5 ms versus 167.9 ± 29.1 ms; p=0.017) and better renal function (eGFR 62.8 ± 19.6 versus 50.5 ± 17.1 ml/min/1.73 m²; p=0.001) compared with those who had AVNA more than 6 months after pacemaker implantation. Nevertheless, none of these factors predicted mortality on their own in univariable models, and adjusting for them did not diminish the survival benefit of early AVNA (as described above).

By contrast, in RV-paced patients, the timing of AVNA in relation to pacemaker implantation had no statistically significant association with survival (log-rank p=0.254; Figure 3B). Baseline characteristics were similar between those who underwent early versus late AVNA in the RV pacing group, and neither univariable nor multivariable models showed a significant effect of AVNA timing on mortality (adjusted HR 1.29; 95% CI [0.84–1.98]; p=0.245). Within the RV pacing group, the strongest predictor of mortality was eGFR (p<0.001) in both univariable and multivariable analysis (Table 3).

Discussion

In this nationwide 10-year study, we evaluated long-term mortality after AVNA in patients with AF refractory to medical rate control. Although AVNA is often considered as a last resort due to concerns about permanent pacemaker dependence and potential procedural complications, our cohort achieved a median post-AVNA survival of ~7.8 years.^8,20,21 This is substantial for such a high-risk population and similar to the mean life expectancy of the Estonian population, although still shorter than age- and sex-matched remaining life expectancy in the general Estonian population, assuming unchanged mortality.²²

A central aim of the study was whether pacing type (RV pacing versus CRT) influences survival outcomes after AVNA. In both the entire cohort and the subgroup who received AVNA at the time of pacemaker implantation purely for rate control (ad hoc ablate-and-pace), we observed no significant difference in long-term survival between patients receiving CRT and those receiving RV pacing. However, it is important to note that the study was not designed to directly compare equivalent populations: CRT was mostly selected for patients with severe LV systolic dysfunction and other guideline-directed indications, whereas patients with preserved or only mildly reduced LVEF often received RV pacing. Thus, the observed similarity in survival reflects real-world, indication-driven device selection rather than showing that the two pacing modalities would perform the same in patients with identical baseline characteristics.

Despite their higher-risk profile, CRT recipients did not exhibit worse mortality than RV-paced patients, suggesting that biventricular pacing may have compensated for some of the excess risk associated with LV dysfunction. These results align with prior trials and meta-analyses showing that, although CRT often yields superior surrogate outcomes (e.g. LVEF, 6-minute walking distance, quality of life), it has not been shown to improve all-cause mortality compared with RV pacing post-AVNA.^5,23,24 However, it is noteworthy, that the APAF-CRT mortality trial showed that AVNA with CRT compared to pharmacological rate control resulted in a 74% relative reduction in the hazard of all-cause mortality (HR 0.26; 95% CI [0.10–0.65]; p=0.004).²⁵

These results solidify the benefit of synchronised biventricular pacing once the AV node is ablated, compared with medical therapy alone. But our real-world analysis, which included patients with a heterogeneous baseline risk, suggests that after rapid AF is eliminated, RV pacing can also provide the essential rate control that likely drives the survival benefit, particularly in patients without marked dyssynchrony or other clear CRT indications. Thus, our cohort’s comparable survival with RV pacing versus CRT suggests that, once rapid AF is eliminated, selection of pacing strategy may be less determinative than achieving reliable rate control in appropriately selected patients.

The timing of AVNA relative to pacemaker implantation emerged as significant for CRT recipients: patients who underwent AVNA within 6 months of CRT implantation had ~50% reduction in mortality risk compared with delayed AVNA beyond 6 months. We believe this survival benefit reflects the more rapid attainment of near-complete biventricular pacing, which is an essential requirement for CRT efficacy. Studies have shown that even modest reductions in delivered biventricular pacing (<95%) are linked to worse outcomes, whereas capture rates above 98% confer a clear survival advantage.^11,26–29 In our cohort, pre-AVNA biventricular pacing averaged only 68.9%. This suboptimal pacing likely contributes to AF-related cardiomyopathy and helps explain why earlier AVNA translates into improved survival by restoring effective biventricular capture sooner. Emerging conduction system pacing techniques, such as His-bundle pacing and left bundle branch area pacing (LBBAP), show promise as simpler alternatives, potentially offering even more physiological pacing than CRT.²⁸ Since the close of our study window, LBBAP has been rapidly embedded into routine practice in Estonia. While our inclusion criteria and CRT indications reflected contemporary guidelines during 2012 to 2022, today the adoption of LBBAP is altering device selection in this patient group. Future analyses should re-examine outcomes as LBBAP becomes standard in routine care.

Baseline LVEF significantly predicted mortality, highlighting ventricular function as a critical determinant of long-term prognosis, corroborating previous studies.^30,31 A recent meta-analysis by Baudo et al. found that among patients with AVNA and pacing, those with lower ejection fraction (EF) had higher cardiac mortality, particularly in RV-paced patients.³² In our study, many patients with low EF probably had tachycardia-induced cardiomyopathy from uncontrolled AF. While RV pacing in this context has historically been viewed as suboptimal, we observed no difference in

long-term survival between the RV-paced and CRT-paced subgroups. Ahluwalia et al. recently reported that smaller atrial size and absence of fibrosis predict LVEF recovery after AF ablation, which supports the notion that restoration of regular ventricular activation may promote reverse remodelling and contribute to the survival benefit seen after AVNA.³³ This suggests, once again, that after rapid AF is eliminated, improvement in ventricular function rather than the choice of pacing may drive outcomes. Encouragingly, eliminating the rapid rates via AVNA can allow some EF recovery (~4–5% absolute improvement in EF on average, as shown in prospective studies).³⁴ While modest, this benefit reinforces that preserving or improving EF is crucial, and therapies that mitigate rapid AF-induced cardiomyopathy (whether through timely AVNA, CRT or appropriately programmed RV pacing) can be meaningful for survival.

Regarding sex differences, we observed no survival disparities between men and women after adjusting for baseline demographic and clinical differences. Unlike some studies suggesting worse survival in men, our results indicate both sexes derive similar benefit from AVNA.³⁵

Study Limitations

The current study has several limitations. First, its retrospective design introduces potential selection bias, because pacemaker type and timing of AVNA were not randomised: clinicians selected CRT for patients with more advanced disease and RV pacing for others, which may confound outcome comparisons, so the apparent similarity in survival should not be interpreted as CRT and RV pacing being equivalent in a randomised population. Survival times were right-censored at the study end and some misclassification of baseline covariates cannot be fully excluded. Likewise, LVEF was abstracted from the echocardiogram performed before AVNA, thus patients were imaged during rapid AF that increases beat-to-beat variability and restricts diastolic filling, thus potentially underestimating LVEF. We attempted to adjust for measurable differences, but unmeasured confounders may remain.

Second, we did not capture non-fatal outcomes (HF hospitalisations, symptoms, quality of life) or post-AVNA medical therapy (specifically guideline-directed HF medications and anticoagulation), which are relevant when comparing pacing strategies, especially CRT.

Third, the cohort size, while substantial for a single-country nationwide study, limits the power to detect subtle differences in survival between pacing strategies. This was especially true in the LVEF >35% stratum, where CRT numbers were very small (n=12; only one patient at risk at 10 syears), resulting in wide CIs and underpowered comparisons; findings in this subgroup should therefore be interpreted with caution.

Conclusion

In this nationwide 10-year observational study, AVNA combined with permanent pacing was associated with survival comparable with the general Estonian population. There was no significant difference in overall survival between RV pacing and CRT groups, suggesting that RV pacing may remain a suitable option for many patients. Notably, CRT recipients who underwent AVNA within 6 months after device implantation showed an association with improved survival compared with those with delayed AVNA. These findings support the hypothesis that earlier AVNA might optimise pacing effectiveness and outcomes in select patient populations. Prospective randomised studies are warranted to confirm these observations and guide clinical recommendations.

Click here to view Supplementary Material.