Catheter Ablation versus Medical Therapy of Atrial Fibrillation in Patients with Heart Failure: An Updated Systematic Review and Meta-Analysis of Randomized Controlled Trials

Background: Atrial fibrillation (AF) and heart failure (HF) often coexist and synergistically contribute to an increased risk of hospitalization, stroke, and mortality. Objective: To compare the efficacy of catheter ablation (CA) versus medical therapy (MT) in HF patients with AF. Methods: Electronic databases were queried for randomized controlled trials (RCTs) of CA versus MT of AF in patients with HF. Risk ratios (RRs), mean differences (MDs), and 95% confidence intervals (CIs) were measured using the Mantel–Haenszel method. Results: A total of nine RCTs enrolling 2155 patients met the inclusion criteria. Compared to MT, CA led to a significant reduction in the composite of all-cause mortality and HF hospitalization (24.6% vs. 37.1%; RR: 0.65 (95% CI: 0.53–0.80); p < 0.0001), all-cause mortality (8.8% vs. 13.6%; RR: 0.65 (95% CI: 0.51–0.82); p = 0.0005), HF hospitalization (15.4% vs. 22.4%; (RR: 0.67 (95% CI: 0.54–0.82); p = 0.0001), AF recurrence (31.8% vs. 77.0%; RR: 0.36 (95% CI: 0.24–0.54); p < 0.0001), and cardiovascular (CV) death (4.9% vs. 8.4%; RR: 0.58 (95% CI: 0.39–0.86); p = 0.007). CA improved the left ventricular ejection fraction (MD:4.76% (95% CI: 2.35–7.18); p = 0.0001), 6 min walk test (MD: 20.48 m (95% CI: 10.83–30.14); p < 0.0001), peak oxygen consumption (MD: 3.1 2mL/kg/min (95% CI: 1.01–5.22); p = 0.004), Minnesota Living with Heart Failure Questionnaire score (MD: −6.98 (95% CI: −12–03, −1.93); p = 0.007), and brain natriuretic peptide levels (MD:−133.94 pg/mL (95% CI: −197.33, −70.55); p < 0.0001). Conclusions: In HF patients, AF catheter ablation was superior to MT in reducing CV and all-cause mortality. Further significant benefits occurred within the ablation group in terms of HF hospitalizations, AF recurrences, the systolic function, exercise capacity, and quality of life.


Introduction
Atrial fibrillation (AF) and heart failure (HF) are closely interlinked by pathophysiological mechanisms synergistically contributing to atrial and ventricular myopathy and dysfunction [1][2][3][4]. 2 of 15 AF prevalence ranges between <10% and 50%, according to the clinical severity of HF [5]. When AF and HF coexist, their natural history is further complicated by an increased risk of hospitalization, stroke, and all-cause and cardiovascular (CV) mortality [6][7][8]. Therefore, treating AF in HF patients poses several challenges due to the complexity of this population [9][10][11].
As earlier trials have provided conflicting results regarding the best treatment strategy for AF, it is still debated whether rhythm control should be preferred over rate control. Recently, the Early Treatment of Atrial Fibrillation for Stroke Prevention Trial (EAST- AFNET 4) showed the significant benefits of a rhythm control strategy compared to usual care among patients with AF diagnosed within 1 year and other concomitant CV conditions [12].
These results, although encouraging, were not corroborated by two other recent trials [13,14] showing no significant clinical advantage provided by catheter ablation (CA), with both trials being terminated early due to their apparent futility. To explain these uncertainties, we performed a systematic review and meta-analysis of the randomized controlled trials (RCTs), aiming to compare CA versus the medical therapy (MT) of AF in HF patients.

Data Sources and Searches
We systematically searched the Medline, Cochrane, Journals@Ovid, and Scopus electronic databases for RCTs published from the time of inception to 30 May 2022 and focusing on CA versus MT in HF patients with AF. Three investigators (A.P, G.V., and M.M.) independently performed searches including the following terms: atrial fibrillation, heart failure, left ventricular dysfunction, and catheter ablation. Detailed information of our literature search strategy is available in the expanded methods.

Study Selection and Outcomes
The preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement for reporting systematic reviews and meta-analyses was used in this study [15].
All titles and full-text versions of the selected relevant RCTs were screened to identify those comparing AF ablation with rate or rhythm control therapy in HF patients, which had at least a 6-month follow-up period, included adults aged 18 years or older, and reported 1 or more clinical outcomes.
Observational studies, nonrandomized trials, editorials, case reports, reviews, expert opinions, and non-English studies were excluded.

Data Extraction and Quality Appraisal
Three investigators (A.P., G.V., and M.M.) extracted data from each study using standardized protocols and reporting forms and independently assessed the quality items. Disagreements were resolved by consensus. The quality of the individual studies was assessed by three investigators (A.P, G.V., and M.M.) using the Cochrane risk of bias Tool [16], as reported in the Supplementary Materials, Figure S1.

Study Endpoints
The primary endpoint was a composite of all-cause mortality and HF hospitalization. These endpoints were also assessed independently. Other secondary endpoints were CV death, AF recurrence rate, changes in the left ventricular ejection fraction [(LVEF) ∆LVEF], changes in quality of life (assessed via the Minnesota Living with Heart Failure Questionnaire (∆MLHFQ)), changes in peak oxygen consumption (∆VO 2 max), changes in distance walked during a 6-min walk test (∆6MWT), and changes in brain natriuretic peptide (∆BNP) levels.
The safety endpoints were CA-related periprocedural adverse events (e.g., access site complications (femoral bleeding or hematoma), pericardial complications (with and without tamponade), pulmonary vein stenosis, and procedural stroke) and antiarrhyth-mic drug therapy-related side effects (e.g., proarrhythmic effects, pulmonary, liver, and thyroid toxicity).

Statistical Analysis
Descriptive statistics are presented as means and standard deviations (SD) for the continuous variables or a number of cases (n) and percentages (%) for the dichotomous and categorical variables. The Mantel-Haenszel risk ratio (RR) model was used to summarize the data among the treatment arms. Summary estimates and 95% confidence intervals (CI) were reported for the continuous variables as the standardized mean difference. Freeman-Tukey double arcsine transformation was used to establish the variance of raw proportions. We used the Hartung-Knapp-Sidik-Jonkman method with the random effect model to combine the transformed proportions. The heterogeneity across studies was evaluated by using the Chi 2 , Tau 2 , and Higgins-I 2 statistics, while random effects models of DerSimonian and Laird or fixed effects models were used in cases of I 2 > 25% or ≤25%, respectively. Sensitivity analyses comparing CA and drug therapy were performed including patients with reduced LVEF (<50%) or persistent AF or with a pharmacological rhythm control strategy. The publication bias was assessed using the funnel plot and Egger's test. The statistical analysis was performed using Review Manager (RevMan) (computer program) Version 5.4.1, Copenhagen, Denmark: Nordic Cochrane Centre, the Cochrane Collaboration, 2020.
The safety endpoints were CA-related periprocedural adverse events (e.g., access site complications (femoral bleeding or hematoma), pericardial complications (with and without tamponade), pulmonary vein stenosis, and procedural stroke) and antiarrhythmic drug therapy-related side effects (e.g., proarrhythmic effects, pulmonary, liver, and thyroid toxicity).

Statistical Analysis
Descriptive statistics are presented as means and standard deviations (SD) for the continuous variables or a number of cases (n) and percentages (%) for the dichotomous and categorical variables. The Mantel-Haenszel risk ratio (RR) model was used to summarize the data among the treatment arms. Summary estimates and 95% confidence intervals (CI) were reported for the continuous variables as the standardized mean difference. Freeman-Tukey double arcsine transformation was used to establish the variance of raw proportions. We used the Hartung-Knapp-Sidik-Jonkman method with the random effect model to combine the transformed proportions. The heterogeneity across studies was evaluated by using the Chi 2 , Tau 2 , and Higgins-I 2 statistics, while random effects models of DerSimonian and Laird or fixed effects models were used in cases of I 2 > 25% or ≤25%, respectively. Sensitivity analyses comparing CA and drug therapy were performed including patients with reduced LVEF (<50%) or persistent AF or with a pharmacological rhythm control strategy. The publication bias was assessed using the funnel plot and Egger's test. The statistical analysis was performed using Review Manager (RevMan) (computer program) Version 5.4.1, Copenhagen, Denmark: Nordic Cochrane Centre, the Cochrane Collaboration, 2020.

Baseline Characteristics
Baseline clinical characteristics are reported in Table 1. The nine RCTs enrolled 2155 patients. Among them, 1077 were assigned to CA and 1078 to drug therapy. Overall, 69.9% (n = 1507) of patients were male, with an average age of 63.5 years (95% CI: 62.1-64.9), and the mean LVEF was 37.9% (95% CI: 32.3-43.6). Further details on the baseline clinical characteristics are reported in Table 1 and Supplementary Materials, Table S1. Note: 6MWT: 6 min walk test; AATAC: ablation versus amiodarone for treatment of atrial fibrillation in patients with congestive heart failure and an implanted ICD; AF: atrial fibrillation; AMICA: atrial fibrillation management in congestive heart failure with ablation; ARC-HF: a randomized trial to assess catheter ablation versus rate control in the management of persistent atrial fibrillation in chronic heart failure; BNP: brain natriuretic peptide; CABANA: catheter ablation vs. antiarrhythmic drug therapy for atrial fibrillation; CAMERA-MRI: catheter ablation versus medical rate control in atrial fibrillation and systolic dysfunction; CAMTAF: a randomized controlled trial of catheter ablation versus medical treatment of atrial fibrillation in heart failure; CASTLE-AF: catheter ablation versus standard conventional therapy in patients with left ventricular dysfunction and atrial fibrillation; HF: heart failure; IQR: interquartile range; LVEF: left ventricular ejection fraction; MLHFQ: Minnesota Living with Heart Failure Questionnaires; mo: months; MRI: magnetic resonance imaging; NA: not available; NYHA: New York Heart Association; RAFT: randomized ablation-Based rhythm control versus rate control; RNVG: radionuclide ventriculography; SD: standard deviation; VO 2 max: peak oxygen consumption.
Among the patients randomized to MT, a rate control strategy was pursued in five studies [14,[20][21][22][23] and rhythm was pursued control therapy in one, while a combined treatment of both strategies was pursued in three trials (the AMICA (atrial fibrillation management in congestive heart failure with ablation) [13], CASTLE-AF [18], and CABANA trials [17] had 38%, 30%, and 29% of patients on rhythm control, respectively).

Composite Endpoint, All-Cause Mortality, HF Hospitalizations
All trials reported data on all-cause mortality and/or HF hospitalization. AF ablation led to a significant reduction in the composite endpoint (24.

Discussion
Herein, we performed a systematic review and meta-analysis to provide a comprehensive overview of the outcomes of HF patients undergoing CA versus MT of AF.
We observed that CA was associated with a 35% reduction in the composite endpoint (all-cause mortality and HF hospitalization) compared to MT. A significant reduction was observed when the endpoints above were assessed separately (all-cause mortality: −35%; HF hospitalization: −33%).
Regarding other secondary outcomes, CA led to a lower rate of AF recurrence and CV death and promoted a significant improvement in the left ventricular systolic function (∆LVEF), exercise capacity (∆6MWT), cardiorespiratory fitness (∆VO 2 max), quality of life (∆MLHFQ), and HF severity (∆BNP).
To the best of our knowledge, our systematic review and meta-analysis is the most updated and comprehensive analysis of RCTs thus far. Compared to prior studies [28,29], we included data from the two recent AMICA [13] and RAFT-AF [14] trials, which together account for approximately one third of our study population, and performed additional endpoint/subpopulation analyses.
CA has been demonstrated to be the most effective rhythm control strategy, with several recent trials showing its superiority compared to MT as an early AF therapy in the general population [30]. Nonetheless, it is far more challenging to achieve rhythm control in patients with a high burden of comorbidities (e.g., HF, obstructive sleep apnea, coronary artery disease) [3,26,[31][32][33] due to their more diseased atrial substrate. These patients are also more prone to drug-related, as well as ablation-related, complications.
AF and HF often coexist and share several risk factors, which may contribute to arrhythmia progression, as well as worse clinical outcomes, including stroke, hospitalization, and overall and cardiovascular mortality. RCTs investigating the success rate of CA compared to MT in patients with AF and HF have provided conflicting results. Among them, data from the two recent RCTs, AMICA [13] and RAFT [14], were published after the enrolment was closed early due to their apparent futility [13,14].
The RAFT-AF [14] trial enrolled 411 patients with high-burden paroxysmal or persistent AF. Although the initial goal was to recruit 600 patients, a determination of apparent futility by the Data Safety Monitoring Committee after an interim analysis led to the trial's early termination. A non-significant trend towards improved outcomes with CA was reported, with the HR value of 0.71 (95% CI: 0.49-1.03; p = 0.066). Notably, CA led to a significantly greater improvement in LVEF, exercise capacity, AF and HF symptoms, and quality of life.
The AMICA [13] trial enrolled patients with persistent/longstanding persistent AF and LVEF ≤35%, aiming to assess any LVEF changes after 1 year. The study revealed a similar improvement of the LVEF with CA and MT; however, the comparison between the two treatment arms showed no statistically significant difference. The only significant difference was observed in the device-recorded AF burden, which was significantly higher in the MT group. It can be speculated that the trial included patients with a higher comorbidity rate compared to other similar trials (e.g., CASTLE-AF) [18], who were potentially too sick to show any significant clinical and functional advantages from CA and the resulting better sinus rhythm control. Similar findings were reported in the subpopulation of CASTLE-AF patients with New York Heart Association (NYHA) functional class III symptoms or an LVEF of <25%, who did not show any significant benefits from CA. In the CAMERA-MRI [24] trial, the improvement in LVEF promoted by CA was significantly better, with systolic function normalization being achieved in 58% of patients compared to 9% of those on MT (p = 0.0002). Nonetheless, the absolute LVEF improvement was significantly greater among patients without late gadolinium enhancement [3]. Thus, advanced cardiac remodeling may limit the recovery of the LVEF and also increase the risk of AF relapse in the long term [26,32]. These observations highlight the importance of an early AF ablation strategy.
CA, as a first line approach, might be particularly important for HF patients, since arrhythmic recurrences promote worsening cardiomyopathy, arrhythmia progression, and poor outcomes. Once the vicious cycle of AF and HF begins, the success rate of PVI is significantly lower and other sources of triggers outside the PVs may contribute to arrhythmia initiation/relapse [4]. Therefore, early ablation is critical for HF patients, as it is associated with better ablation success. Mechanistic and clinical studies have also highlighted the dynamic interplay between sinus rhythm restoration, cardiac function, symptoms, and clinical outcomes. Specifically, successful rhythm control may subsequently promote positive atrial and ventricular remodeling, which may result in better clinical outcomes. From this perspective, the benefits of CA on the cardiac function may manifest themselves later, and outcome improvements may occur even later. From a functional standpoint, the PABA-CHF [34] trial compared PVI versus atrioventricular-node ablation combined with biventricular pacing in HF patients with EF < 40%. Patients were followedup over 6 months, showing a steady improvement in the mean LVEF in the CA group, who continued to recover until the end of the study period. From a clinical standpoint, the mortality curves separated only after 2 years in the CASTLE-AF [18] trial. Further studies are necessary in order to understand the best AF ablation strategy for HF patients, as well as the pathophysiological basis of CA-mediated functional and clinical benefits.
Another finding worth highlighting is the safety profile of CA in a population such as the HF population, characterized by a high comorbid profile. Notably, CA appeared to be safe and showed a similar risk of adverse events as MT. The prevalence of procedural complications was 5.02% (95% CI: 3.44-6.81), with access-site complications accounting for approximately half of the events [28,35,36].

Limitations
Our study has several limitations that need to be acknowledged. (1) Patients selected for a randomized catheter ablation trial may be healthier than those in real-world situations.
(2) The number of patients enrolled in each RCTs was highly variable and may account for the imbalance in the results and heterogeneity. (3) The high heterogeneity regarding the arrhythmia detection techniques during follow-up should be considered. (4) Additional ablation outside the PVs, performed in some RCTs, could have affected the clinical outcomes. (5) The RCTs included here enrolled patients from 2011 to 2022, involving temporal changes in both CA and drug therapy. (6) Because patients and physicians were not blinded to the treatment assignment, it is possible that the post-ablation medical management differed between RCTs.

Conclusions
AF ablation was superior to conventional drug therapy in improving the composite of all-cause mortality and HF hospitalization, the all-cause mortality, HF hospitalization, AF recurrence, cardiovascular death, LVEF, 6 min walk test distance, VO 2 max, and quality of life.  Figure S1: Methodological evaluation of the included studies. Figure S2: Any Adverse Events and Thromboembolic Events. Figure S3: Catheter Ablation vs Medical Rate Control. Figure S4: Catheter Ablation vs Drug Therapy in patients with LVEF ≤ 50%. Figure  Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Biosense Webster, St. Jude Medical, and Janssen. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.