Catheter Ablation of Atrial Fibrillation in Patients With Heart Failure

Abstract

Atrial fibrillation (AF) is associated with thromboembolic stroke, systemic embolism, and decompensated heart failure (HF) requiring hospitalization, thereby imposing a substantial financial burden on the health care industry (13). The prevalence of AF in HF increases with HF severity: 5% in New York Heart Association (NYHA) class I to 50% in class IV HF (4). Together, AF and HF lead to atrial structural and electrical remodeling and progression of paroxysmal to persistent AF, perpetuating a vicious cycle of impaired left ventricular (LV) filling, contractility, and cardiac output (5). Catheter ablation is an established therapeutic strategy for symptomatic, drug-refractory AF (6, 7). However, current guidelines support AF ablation with caution (class IIb recommendation) in patients with LV systolic dysfunction and HF with reduced ejection fraction (HFREF) (6). Recent randomized controlled trials (RCTs) reported clinical improvements in mortality, HF hospitalizations, LV ejection fraction (LVEF), and quality of life in patients with HFREF who have AF ablation (814). We performed a systematic review and meta-analysis of RCTs comparing the safety and efficacy of AF ablation versus drug therapy in patients with HFREF. Methods We developed (on 12 September 2017) and followed a protocol for this systematic review and meta-analysis (Supplement). Supplement. Supplementary Index Data Sources and Searches We searched (without language restrictions) ClinicalTrials.gov, PubMed, Web of Science (Clarivate Analytics), EBSCO Information Services, Cochrane Central Register of Controlled Trials, Google Scholar, and various major scientific conference sessions (held at American Heart Association, American College of Cardiology, European Society of Cardiology Congress, Heart Rhythm Society, and Cardiostim meetings) for abstracts and articles published between 1 January 2005 and 1 October 2018. Two investigators (M.K.T. and J.G.) independently performed searches including the following terms: atrial fibrillation, catheter ablation, heart failure, left ventricular ejection fraction, hospitalizations, functional capacity, peak oxygen consumption, and quality of life. Details regarding search strategies and filters are provided in the Supplement. Study Selection and Outcomes Two investigators (M.K.T. and J.G.) independently screened all titles and full-text versions of selected relevant RCTs to identify those that had at least 6 months of follow-up, included adults aged 18 years or older, compared AF ablation with standard drug therapy (rate or rhythm control medications) in patients with HF, and reported 1 or more clinical outcomes. Retrospective studies, studies with no comparative group, nonrandomized trials, case reports, editorials, reviews, expert opinion, and studies published in a language other than English were excluded. Main outcomes of interest were all-cause mortality and HF hospitalization; change in LVEF; distance on the 6-minute walk test; peak oxygen consumption (Vo 2max); quality of life, assessed by the Minnesota Living with Heart Failure Questionnaire (MLHFQ); and AF recurrence. Harm outcomes were serious adverse events, including access site complications (femoral hematoma or bleeding, pseudoaneurysm, and groin infection), pericardial complications (with and without tamponade), HF, pulmonary vein stenosis, intracranial bleeding, and systemic thromboembolic events. Data Extraction and Quality Appraisal Two independent investigators (M.K.T. and J.G.) extracted data from the individual studies by using a standardized protocol and data extraction form (Supplement). Together, these investigators assessed the quality of individual studies by using the Cochrane Risk of Bias Tool (15). Data Synthesis Because previous studies found no differences in mortality or HF hospitalizations between rate and rhythm control drug therapies, we aggregated both approaches in the standard therapy group. We pooled outcomes of studies regardless of the follow-up length in each trial. We judged this strategy to be appropriate despite differences in study follow-up, because the effect of ablation diminishes over time. Risk ratios (RRs) were used to pool differences in binary events, and mean differences were used to pool differences in continuous outcomes. We pooled within-group recurrence rates for the trials reporting AF-free survival and reported summaries on the basis of FreemanTukey double arcsine transformation (16). For binary outcomes and studies reporting a single proportion, an overall estimate with 95% CI was calculated by using a random-effects model with the PauleMandel method for estimating between-study variance 2. A random-effects model with the restricted maximum likelihood 2 estimator was used to calculate a pooled estimate for studies with continuous outcome data (17). Because fewer than 10 studies were analyzed, the HartungKnapp small-sample adjustment was used (18). Treatment-group continuity correction was performed for studies with zero cell frequency. Between-study heterogeneity was assessed by using the 2 and I 2 statistics. Sensitivity analyses were performed to assess the contribution of each study to the pooled estimate. Further sensitivity analysis included the PABA-CHF (Pulmonary Vein Antrum Isolation vs. AV Node Ablation With Biventricular Pacing for Treatment of Atrial Fibrillation in Patients With Congestive Heart Failure) study, which compared catheter ablation versus AV node ablation and cardiac resynchronization therapy in patients with AF and HF (13). The meta packages metabin, metacont, and metaprop were used to pool RRs, mean differences, and recurrence rates, respectively. All statistical analyses were performed with R, version 3.5.1 (The R Foundation for Statistical Computing). Role of the Funding Source The study received no funding. Results Overview of Trials Among 24960 records identified, 6 trials met the inclusion criteria (Figure 1 and Supplement Figures 1 to 7) (812, 14). An additional trial, PABA-CHF, compared ablation with rate control, but the latter was achieved with AV junction ablation plus biventricular pacing, not medications; this trial was included only in the sensitivity analysis (13). We judged trials as having low risk of selection and attrition biases, as well as moderate risks of performance and detection biases due to open-label designs (Supplement Table 1). We judged risk of reporting bias as low for mortality and HF hospitalization outcomes and moderate for other outcomes. Figure 1. Evidence search and selection. PABA-CHF= Pulmonary Vein Antrum Isolation vs. AV Node Ablation With Biventricular Pacing for Treatment of Atrial Fibrillation in Patients With Congestive Heart Failure; RCT= randomized controlled trial. * ClinicalTrials.gov, PubMed, Web of Science (Clarivate Analytics), EBSCO Information Services, Cochrane Central Register of Controlled Trials, Google Scholar, and various scientific conference sessions from 1 January 2005 to 1 October 2018. PABA-CHF was excluded because it did not meet the exclusion criteria. The Table summarizes the study characteristics, and Supplement Table 2 shows each trial s inclusion and exclusion criteria. The 6 trials enrolled 775 patients; 388 were randomly assigned to the AF ablation group and 387 to the standard therapy group. The mean age of patients in the trials ranged from 55 to 64 years. Ischemic cardiomyopathy was the predominant cause of HF. Five trials included patients with persistent AF (812). In CASTLE-AF (Catheter Ablation versus Standard Conventional Therapy in Patients with Left Ventricular Dysfunction and Atrial Fibrillation), 52% of patients had paroxysmal AF (14). The mean duration of persistent AF was longer than 12 months in all trials except AATAC (Ablation versus Amiodarone for Treatment of Atrial Fibrillation in Patients with Congestive Heart Failure and an Implanted ICD), in which it was 8.23.6 months. The mean LVEF among the study populations was 28.3%. Four trials included patients with NYHA class II to III HF; CAMERA-MRI (Catheter Ablation Versus Medical Rate Control in Atrial Fibrillation and Systolic Dysfunction) and CASTLE-AF also included a minority of patients (1.3%) with NYHA class IV HF. Study follow-up ranged from 6 to 60 months. Table. Study Characteristics Radiofrequency ablation catheters were used in all trials. Pulmonary vein isolation was the key ablation strategy in all studies, but in most patients, further ablation was done outside the pulmonary veins. In 4 trials, standard drug therapy consisted of rate control. Rhythm control with amiodarone was used exclusively in the AATAC trial, whereas CASTLE-AF used rate control in two thirds and rhythm control in one third of the cohort; most patients (approximately 90%) in the latter group received amiodarone. Outcomes All-cause mortality and HF hospitalization data were available in 5 and 4 trials, respectively. Compared with standard drug therapy, AF ablation was associated with a greater reduction in all-cause mortality (9.0% vs. 17.6%; RR, 0.52 [95% CI, 0.33 to 0.81]) and HF hospitalizations (16.4% vs. 27.6%; RR, 0.60 [CI, 0.39 to 0.93]) (Figure 2). No statistically significant heterogeneity was observed (I 2= 0%). Figure 2. Forest plot of the decrease in all-cause mortality and HF hospitalization with catheter ablation compared with medical treatment for atrial fibrillation. AATAC= Ablation versus Amiodarone for Treatment of Atrial Fibrillation in Patients with Congestive Heart Failure and an Implanted ICD; ARC-HF= A Randomized Trial to Assess Catheter Ablation Versus Rate Control in the Management of Persistent Atrial Fibrillation in Chronic Heart Failure; 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 Thera