Ablation vs. Meds in HFrEF: One More Trial, or Time to Move On?

© 2026 HMP Global. All Rights Reserved.

Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of EP Lab Digest or HMP Global, their employees, and affiliates. 

Featured is the presentation entitled "Ablation vs. Meds in HFrEF: One More Trial, or Time to Move On?" by Jason Andrade, MD, from Session 6 at the 2026 Western AF Symposium.

Thank you for having me back. This is somewhat of a minefield that I’m about to step into, but here we go. 

Atrial fibrillation (AF) and heart failure are colliding epidemics. You’ve likely heard a great deal about this over the past few days, if not the past several years. We know these conditions worsen one another. A patient with heart failure who develops AF can experience poorer quality of life, faster disease progression, a higher risk of ICD shocks, ventricular arrhythmias, stroke, and increased mortality.

When considering AF management in patients with heart failure, we often return to the AF-CHF study. This trial, conducted by my mentor nearly 20 years ago, enrolled approximately 1400 patients with persistent AF and reduced ejection fraction. Patients were randomized to drug-based rhythm control or drug-based rate control. Ultimately, there was no difference in overall mortality or worsening heart failure, keeping in mind that amiodarone was essentially the rhythm-control strategy at that time.

As a result, guidelines historically did not explicitly favor or discourage either rhythm control or rate control. Instead, they emphasized shared decision-making, with rhythm-control recommendations tailored to individual patients.

Since AF-CHF, multiple trials have compared catheter ablation with other interventions. These studies have suggested improvements in LVEF, as well as benefits in functional measures such as 6-minute walk distance.

The problem is that many of these outcomes are unblinded and therefore susceptible to bias. While improvements in EF may be real, there are reasons to question how robust those findings are. As a community, we have therefore focused on hard-outcome trials.

Fortunately, several such trials have been completed, and you’ve likely heard about them repeatedly over the years and during this meeting.

The AATAC trial enrolled 203 patients with persistent AF and reduced ejection fraction, randomizing them to amiodarone or ablation. Ablation resulted in substantially fewer AF recurrences. As secondary endpoints, there were fewer unplanned hospitalizations and fewer deaths—approximately a 50% reduction, which was a very impressive result.
The CASTLE-AF trial enrolled 363 patients with NYHA class II–III heart failure, most with an EF below 25%. Again, investigators observed lower mortality and lower rates of worsening heart failure, with reductions of roughly 50% following ablation.

More recently, the CASTLE-HTx trial examined a pre-transplant population of about 200 patients with approximately 18 months of follow-up. Most participants had class II–III heart failure and EF below 35%. Once again, favorable mortality results were observed.

At this point, we have 3 randomized trials showing similar findings. As you’ve heard, the ACC guidelines now give ablation a Class I recommendation. It would seem that the issue is settled.

Except that other major guidelines have not gone that far.

More recent guidelines from HRS, LAHRS, and EHRA assign a Class IIa recommendation for heart failure. Tachycardia-induced cardiomyopathy is a different matter—in that setting, ablation is clearly indicated. But for AF in the presence of structural heart disease, the recommendation remains Class IIa. The European AF guidelines, American heart failure guidelines, and European heart failure guidelines all similarly provide Class IIa recommendations.
So why hasn’t this become a universal Class I recommendation? Why don’t all societies agree?

Part of the answer may be that not all trials have shown consistent results.

RAFT-AF enrolled approximately 400 patients. Not all had HFrEF, which is an important caveat, but the study did not demonstrate a significant difference in mortality or heart failure outcomes. The results were neutral.

If you aggregate all of these studies, you end up with roughly 1400 patients across 9 trials. However, those 1400 patients are spread across studies ranging from 40 patients to 400 patients.

As Milton Packer has pointed out, none of these trials would qualify as a phase II trial for a heart failure drug.

Looking at outcomes across all 9 trials, there were about 200 deaths and approximately 285 heart failure events.

For comparison, AF-CHF enrolled a similar number of patients but observed twice as many deaths and 3 times as many heart failure events. One could argue that this reflects a different era, but the broader point remains: when event rates are low, the risk of misinterpretation increases.

The reported hazard ratios are dramatic—often suggesting 50% or even 70% reductions in risk. Ablation appears highly effective.

If you haven’t read Stuart Pocock’s papers on interpreting positive and negative trials, I highly recommend them. One key point is that when a small trial reports a positive result, limited statistical power can lead to exaggerated treatment effects and increase the likelihood of false-positive findings.
Examples from the heart failure literature illustrate this concern. In one drug comparison, a smaller trial with only 46 deaths showed a substantial reduction in mortality. However, a larger trial with 10 times as many events found the opposite result—an increased risk.

The concern is that effects observed in small trials may be exaggerated and may not hold up in larger studies. As a result, we may be overestimating the true treatment benefit because we continue to see small studies producing dramatic results. Those findings may not necessarily reflect the underlying truth, particularly when compared with larger, adequately powered trials.

Additional criticisms include the lack of contemporary guideline-directed medical therapy (GDMT), reliance on subjective outcomes, crossover between treatment groups, missing data, loss to follow-up, and, in the case of RAFT-AF, early termination for futility.
When examining GDMT, studies such as CASTLE-AF reported beta-blocker use above 90%. More contemporary trials, such as CASTLE-HTx and others, included more modern therapies and reasonable medication utilization.

However, RAFT-AF, despite being published relatively recently, included no SGLT2 inhibitors and very low rates of mineralocorticoid receptor antagonist use.

This raises an important question: even if the benefits of ablation observed in these trials are real, would they still be seen in patients receiving contemporary 4-pillar GDMT for heart failure?

Within the electrophysiology community, most of us likely believe the answer is yes. When patients with AF and structural heart disease are referred to us, we generally proceed with ablation.

However, outside the EP community, referral patterns tell a different story. Heart failure clinics often do not refer these patients because many clinicians remain unconvinced. Around the world, there are groups that question these trial results for all the reasons discussed.

As a result, there have been calls for large international multicenter trials.
This topic was discussed extensively at the Global CVCT Forum this past December. There was an entire session devoted to it. What we need are comparative trials focused on hard clinical outcomes. Whenever subjective endpoints are used, there are always calls for sham controls. However, in this high-risk population—with so much to gain if our interventions truly work—hard clinical outcomes should remain the primary focus.

You’ve already heard about HFpEF trials, but if we want heart failure physicians to refer these patients, we must design these studies as heart failure trials conducted within heart failure populations.

If heart failure specialists have buy-in—as discussed in the previous session regarding stakeholder engagement in AF clinics—they will trust the results and act on them. But as long as they perceive these studies as niche ablation trials driven solely by electrophysiologists, skepticism will remain.

Currently, 3 large HFrEF trials are underway.

DANABLATE is enrolling approximately 1600 patients with EF below 50% and receiving GDMT. Patients are randomized to early intervention versus usual care. Usual care may include pulmonary vein isolation (PVI) later—it is not prohibited, just not expedited. The primary outcomes are cardiovascular death and heart failure hospitalization.

CRAAFT-HF is primarily UK-based but has recently expanded into Canada and is expected to expand further into Australia and parts of Europe. The study will enroll 1200 patients with AF and EF below 50%, comparing PVI with no PVI. It is largely driven by heart failure physicians, with Pier Lambiase serving as the EP principal investigator.
Rod and I first conceptualized OPTIMAL AF-HF many years ago at this meeting when it was held in Park City. This study is essentially the North American counterpart. We have now enrolled our first patient. The trial compares PVI with either no ablation or AV node ablation, again focusing primarily on hard clinical outcomes.

Regarding that final point, I pulled this slide together just before this talk, so I’ll move quickly.

One area where we have almost no contemporary evidence—and I mention this because Ken is here—is the role of AV node ablation in patients with AF and heart failure, particularly in the era of conduction system pacing.

If you believe the ablation trials, then these data suggest all-cause mortality may be reduced by as much as 70%. That raises the possibility that AV node ablation could potentially be even more effective than current approaches.

Again, the same limitations apply: this was a study of only about 100 patients, so the findings may not be definitive. Nevertheless, it is an important area to consider.

Thank you.