The CABANA Trial and Heart Failure: Some Sun and Some Shade

Lynda E. Rosenfeld, MD; Alan D. Enriquez, MD


Circulation. 2021;143(14):1391-1393. 

Trials such as CASTLE-AF (Catheter Ablation Versus Standard Conventional Therapy in Patients With Left Ventricular Dysfunction and Atrial Fibrillation), AATAC (Ablation Versus Amiodarone for Treatment of Atrial Fibrillation in Patients With Congestive Heart Failure and an Implanted ICD/CRTD), and AMICA (Atrial Fibrillation Management in Congestive Heart Failure With Ablation), comparing catheter ablation (CA) with medical therapy in patients with atrial fibrillation (AF) and heart failure with reduced ejection fraction (HFrEF), have found significant reductions in all-cause mortality, heart failure (HF) admissions, and AF recurrence, as well as improvements in measures of quality of life in those patients randomized to an ablation strategy.[1–3] This has led to recent guideline changes suggesting that CA may be considered as first-line therapy in patients with AF and HFrEF (class IIA[4] or class IIB[5] recommendation). Although CA appears to be more effective than antiarrhythmic agents in maintaining sinus rhythm in patients with HF, whether with HFrEF or with a preserved or minimally reduced ejection fraction (HFpEF),[6] to date, no randomized controlled trials have compared CA with medical therapy in patients with HFpEF.

HFpEF and AF are intimately related. It is often debated whether AF exacerbates HF or HF precipitates AF…or both. Each condition has similar risk factors, including advanced age, hypertension, obesity, sleep apnea, and diastolic dysfunction. The prevalence of HFpEF in patients with AF ranges from 8% to 24%,[7–9] whereas several studies have found AF to be present in ≈20% to 30% of patients with HFpEF.[10] We know that HFpEF worsens the prognosis of AF, as does the identification of AF in patients with HFpEF.[7,11] We also know that patients with HFpEF and AF are often very symptomatic[7,10,11] and that no treatment has been clearly proven to improve their survival.

In this issue of Circulation, Packer et al[12] report their findings for a prespecified subgroup analysis of 778 of the 2204 patients randomized in the CABANA (Catheter Ablation Versus Antiarrhythmic Drug Therapy for Atrial Fibrillation) trial[13] who had HF defined as New York Heart Association (NYHA) class II or worse symptoms. Although this substudy was not specifically designed to evaluate patients with HFpEF, ≈91% of the patients with a clinical diagnosis of HF participating in CABANA for whom such data were available had a left ventricular ejection fraction of ≥40%, and fully 79% had a left ventricular ejection fraction ≥50%. Too few patients (9.3%) had significantly reduced left ventricular ejection fractions to draw separate conclusions about this group. Thus, this is the first large randomized trial to look at the relative effectiveness of CA versus drug therapy in patients believed to have HFpEF.

In the patients in the CABANA HF substudy, in contrast with the overall CABANA population, a strategy of CA compared with drug therapy reduced the primary composite end point of death, disabling stroke, serious bleeding, or cardiac arrest (hazard ratio, 0.64 [95% CI, 0.41–0.99]), as well as all-cause mortality (hazard ratio, 0.57 [95% CI, 0.33–0.96]). As in the larger study, patients undergoing CA had a lower AF burden and improved quality of life as measured by the AFEQT (Atrial Fibrillation Effect on Quality of Life) and the MAFSI (Mayo Atrial Fibrillation–Specific Symptom Inventory). These findings are important and deserve significant consideration but must be viewed in the context of a number of caveats, including the limitations associated with the larger CABANA study and the HF substudy in particular.

The diagnosis of HFpEF is largely clinical, and the heterogeneity of this patient population makes classification challenging. In addition, many of the tools we use to make the diagnosis of HFpEF are less helpful in the setting of AF. Although we can still measure global ventricular function and left ventricular wall thickness, overall measures of diastolic dysfunction are less sensitive, and left atrial size may be increased because of a left atrial myopathy independent of filling pressures. Similarly, NT-proBNP (N-terminal pro-B-type natruretic peptide) levels, frequently used to confirm the diagnosis of HF, may be increased in the presence of AF.[14] Therefore, we often rely on NYHA classification, which is a functional classification categorizing the degree of breathlessness and effort intolerance, to define and grade HF. Patients in the CABANA HF substudy were distinguished only by being judged as having NYHA class II or greater symptoms by their enrolling physicians. In parallel, the CCS-SAF (Canadian Cardiovascular Society Severity of AF) scale was also used by the CABANA investigators to characterize the severity of trial participants' AF, and it, too, uses the degree of breathlessness, fatigue, and exercise intolerance among the components to quantify the severity of AF.

Although the assignment of NYHA class was supported by measures of patient-reported functional status, AF may have contributed to some of these symptoms. It is important to note that 83.2% of the patients with HF were graded as having AF of CCS-SAF scale class II or higher, which is associated with symptoms that may be similar to those leading to at least an NYHA class II assignment. Although the HF and non-HF groups were fairly balanced, when compared with the non-HF patients, those with HF included somewhat more women and more patients ≥75 years old (both demographics known to be associated with more severe AF symptoms) and had somewhat higher CCS-SAF scores overall.

Furthermore, left ventricular ejection fraction data were not known in 27% of the patients with HF to help confirm the diagnosis of HFpEF. Of interest as well is that only 29.5% of the HF substudy patients had a previous known congestive heart failure history compared with 7.4% of the patients not included in the substudy. No additional explanation of this history is provided, and we are also given no information about diuretic use or responsiveness in the patients with HF, which some believe may help distinguish symptoms of HFpEF from those of AF.[10]

Thus, what we know is that the patients included in the CABANA HF substudy were functionally impaired. We are not given the tools to tease out how much of their functional impairment was attributed to HFpEF or the AF itself. In addition, an ablation strategy did not reduce death from cardiovascular causes or HF hospitalizations, and deaths because of HF were so few in both groups that they could not be evaluated. Thus, we are also not given a clear insight into the mechanisms by which an ablation strategy might have logically improved the composite outcome of death, disabling stroke, serious bleeding, or cardiac arrest or reduced overall mortality specifically in patients with functional impairment and a clinical phenotype of HF.

It is important to note that both the patients in the overall CABANA study and those in the HF substudy were generally enrolled within a year of their diagnosis of AF. The recent EAST-AFNET 4 (Early Treatment of Atrial Fibrillation for Stroke Prevention Trial) has emphasized that aggressive rhythm control (with either CA or drugs) early in the course of AF reduced adverse cardiovascular outcomes (death from cardiovascular causes, stroke, or hospitalization with worsening HF or acute coronary syndrome) compared with usual care.[15] Because ablation is more effective than drug therapy in maintaining sinus rhythm, this relatively early intervention may be especially important in these functionally impaired patients, although the exact mechanism(s) remain uncertain.

Thus, it is not clear that the CABANA HF substudy allows us to extend the conclusions of studies such as CASTLE AF, AATAC, and AMICA and make CA a first-line therapy in all patients with AF and HF, including those with HFpEF, but the CABANA HF substudy does suggest that we should consider CA sooner in patients with AF with significant functional impairment and a clinical phenotype of HFpEF. It also emphasizes the need for better diagnostic tools to assess the contributions of AF and the condition known as HFpEF to patient symptoms and outcomes. Such tools would further empower the additional rigorous trials that are needed to support the tantalizing, but as yet preliminary, findings of the CABANA HF substudy and hopefully confirm that there is a therapy that improves survival and quality of life in patients with HFpEF—if they also have AF.