Additional Clarity Provided Through the Lens of CAMERA-MRI

Why Patients With Heart Failure and Atrial Fibrillation May Benefit From Catheter Ablation

Corresponding Author

• Download figure
• Open in new tab
• Download powerpoint

• atrial fibrillation
• cardiac MRI
• catheter ablation
• diffuse fibrosis
• heart failure
• reduced ejection fraction
• systolic dysfunction

Multiple, randomized clinical trials have demonstrated that patients with heart failure with reduced ejection fraction (HFrEF) benefit from catheter ablation (CA) for management of atrial fibrillation (AF) compared with medical therapy (1,2). In both the AATAC (Ablation Versus Amiodarone for Treatment of Persistent Atrial Fibrillation in Patients With Congestive Heart Failure and an Implanted Device) and the CASTLE-AF (Catheter Ablation Versus Standard Conventional Therapy in Patients With Left Ventricular Dysfunction and Atrial Fibrillation) trials, patients with HFrEF (left ventricular ejection fraction [LVEF] <40% or ≤35%, respectively) and AF randomized to CA or medical therapy had significantly reduced AF burdens as well as important improvements in hard endpoints. In AATAC, CA treatment was associated with a 49% relative risk reduction in recurrent AF, a 45% reduced risk of unplanned hospitalization and a 56% relative risk reduction in mortality over 2-year follow-up, compared with amiodarone, emphasizing that the mode, and not just the intent, of achieving rhythm control is important in the prognosis of these patients (1). Similarly, CA-treated patients in CASTLE-AF, who were 3 times more likely to be in sinus rhythm after 5 years of follow-up, were 38% less likely to experience a primary-outcome event, specifically heart failure hospitalization or death (2). Such dramatic differences in outcomes are somewhat surprising when considering that the absolute change in LVEF in the CA arms of each of the studies was relatively modest (∼8%), even if statistically significant. Finally, long-anticipated results from the CABANA (Catheter Ablation Vs Anti-arrhythmic Drug Therapy for Atrial Fibrillation) trial were publicly reported among Late-Breaking Clinical Trials at the 2018 Heart Rhythm Society Scientific Sessions in Boston in May 2018. Although intention-to-treat analysis did not show a significant difference among the >2,000 patients randomized to CA versus medical therapy, those with congestive heart failure undergoing CA demonstrated a significant 49% risk reduction in primary-outcome events (all-cause mortality, disabling stroke, serious bleeding, and cardiac arrest). It is (somewhat vexingly) notable that improved survival has been consistently demonstrated among HFrEF patients undergoing CA for management of AF when the same benefit has not been observed among HFrEF patients undergoing CA for ventricular tachycardia (VT), even when VT recurrence rates following ablation are similarly reduced compared with medical therapy, and VT ablation is, in many cases, acutely life-saving (3–6). Why has there been such a paradoxical difference observed, with improved longer-term survival observed in association with a treatment that has largely been considered to be elective and not acutely life-saving? CAMERA-MRI (Catheter Ablation versus Medical Rate Control in Atrial Fibrillation and Systolic Dysfunction-an MRI-Guided Multi-centre Randomised Controlled Trial) provides important insights that may help to answer this question as well as to impress the importance of effective rhythm control with CA for optimization in therapy of HFrEF patients with AF.

CAMERA-MRI was a multi-center, randomized clinical trial in which patients with persistent AF, New York Heart Association functional class II to IV status and LVEF ≤45% were randomized to CA versus medical rate control (MRC) (7). All subjects (n = 66) underwent cardiac magnetic resonance imaging (CMR) prior to randomization and 6 months following treatment. In the primary analysis, those undergoing CA not only had significant improvement in LVEF (by 18 ± 13%, compared with 4 ± 13% among those undergoing MRC), but more patients undergoing CA normalized left ventricular (LV) function (LVEF ≥50%) compared with those treated with MRC (58% vs. 9%; p = 0.0002). Importantly, and lending credence to the biological plausibility of benefit from effective CA, those in the CA arm who had late gadolinium enhancement (LGE) (i.e., presence of ostensibly irreversible fibrosis) at baseline experienced the least amount of improvement in subsequent LV systolic function following CA (mean 22.3% vs. 11.6%; p = 0.0069). Although much contemporary attention regarding CMR-detectable AF changes has focused on atrial fibrosis (8), this study was the first to formally evaluate the impact of AF CA on ventricular function and fibrosis with use of LGE. LGE identified on CMR has importantly been associated with increased risk for sudden cardiac death and ventricular arrhythmias, even in the absence of overt LV systolic dysfunction, and among heterogeneous disease subsets (9).

In the present, pre-specified, secondary analysis of CAMERA-MRI in this issue of JACC: Clinical Electrophysiology, T1 mapping, which has been previously validated with histology to correlate with potentially reversible, diffuse interstitial fibrosis, was used to quantify ventricular interstitial fibrosis, as opposed to LGE-based irreversible replacement fibrosis, among the same population of patients with persistent AF, HFrEF and idiopathic cardiomyopathy enrolled in the original trial and in whom T1 mapping images were of sufficient quality (10). T1 maps were generated using investigational software, after sequential, post-contrast measurements were acquired during varying inversion times from mid-septal LV short-axis slices. Of the original 66 patients in the CAMERA-MRI Trial, only 36 had images suitable for inclusion in this analysis. There were no differences noted in baseline characteristics between patients treated with CA and MRC (n = 18 each), including heart failure (HF) medications, LVEF, baseline heart rates, proportion with LGE present (42% in each group) or T1 times.

At 6 months, those in whom LVEF improved most substantially (≥15%) tended to have the greatest reductions (i.e., improvements) in T1 times regardless of treatment modality (p = 0.09). However, subjects treated with CA appeared to derive the most benefit from treatment. Subjects treated with CA had significantly reduced T1 times compared with MRC (−124 ms; 95% confidence interval: 23 to −225 ms; p = 0.017), in the setting of effective rhythm control (AF burden 0.8 ± 2.9%, with 89% having burden of <0.1%), all with a single CA procedure and without any treatment crossovers. The CA group also demonstrated significant improvement in LVEF (absolute improvement from baseline of +14 ± 11%, vs. +1.5 ± 11% for MRC; p = 0.0004), as well as in most other measured echocardiographic parameters, serum brain natriuretic peptide levels and quality-of-life and standardized functional assessments, compared with the MRC cohort. Notably, T1 times also improved significantly among the LGE-positive subjects undergoing CA but not among those undergoing MRC. This effect may explain the improvement in LVEF observed in the original CAMERA-MRI cohort with LGE; LVEF in LGE-positive patients did not improve as much in LGE-negative patients undergoing CA, but it still improved, and to a magnitude comparable with what had been observed among the HFrEF patients experiencing mortality benefits in AATAC and CASTLE-AF.

This analysis of the CAMERA-MRI Trial provides incremental insights about a potential mechanism underlying mortality benefit observed among patients with HFrEF and AF in whom rhythm control is able to be effectively maintained with CA. The findings suggest that: 1) there may be an early, potentially reversible ventricular fibrosis involved in AF in patients with HFrEF, which can be non-invasively identified; 2) this adverse remodeling may have important impact on survival; and 3) adverse remodeling may be reversible with catheter ablation. The findings also add evidence that the wide-spread, laissez-faire practice of rate control for patients with AF should likely not be accepted for AF patients with heart failure (11). Rate control has previously been demonstrated to be ineffective in improving symptoms or in improving systolic function among patients with cardiomyopathy, even in the absence of overt ventricular fibrosis, and, in the present trial, patients treated with MRC similarly experienced no benefits in any measured parameter despite maintenance of adequate rate control throughout the treatment and follow-up period (10–12). This study also provides some insights into why patients with HFrEF and VT have not observed similar benefits despite similar reductions in VT burden following ablation–the substrate for VT is relatively fixed, whereas the ventricular substrate produced with AF may not always be and, in fact, may be able to be reversed with ablation.

Despite the strengths of this study, there are important limitations to acknowledge. First, follow-up time was short and sample size was small, thus limiting its generalizability. In fact, only just over half of the original trial cohort had adequate images for T1 analysis, which speaks to the greater technical and time demands of the imaging protocol, not to mention greater demands for patients (long breath holds, prolonged imaging in relatively confined space and supine position are difficult for many, even in the absence of HFrEF). Second, CMR is not widely available outside of specialized medical institutions for many reasons, including cost, use of specific imaging protocols that may be difficult to uniformly reproduce as well as artifact caused by presence of cardiovascular implantable electronic devices (CIEDs) that may render images uninterpretable. Finally, there continues to be extremely slow acceptance of the practice of performing such studies among patients with CIEDs that are not “Food and Drug Administration–labeled” for CMR use, even at many specialized medical centers, even though the safety of doing so has been demonstrated in most (13).

Despite weaknesses, Dr. Prabhu and colleagues should be commended for building on their cumulative previous efforts with this one, in an overall attempt to not just rely on reporting of outcomes, as has become rote tendency in contemporary, “big-data” medical research, but to help us to try to understand why outcomes are as they have been. Although the intent of this study was not to focus on efficacy of ablation in maintaining sinus rhythm, the fact that CA was so effective in doing so in these patients makes the weight of the authors’ argument even stronger regarding the potential reverse ventricular remodeling that may be observed with effective, non-pharmacologic AF rhythm control in patients with HFrEF. A natural extension illustrating the true utility of this technique, of course, would be to use it among a larger group of patients with longer follow-up, and to assess the potential impact on mortality, but there is now a biologically plausible basis provided for why doing so may be important. In fact, this analysis of the CAMERA-MRI Trial has brought into clearer focus why rhythm control with catheter ablation should be included among optimal goal-directed therapies for patients with AF and HF.

• 2018 American College of Cardiology Foundation