CASTLE-AF (Ablation) Trial Delivers Benefits — Was I Critical Enough?

Doctors like me have performed AF ablation for more than a decade without knowing whether the major procedure improves outcomes.

That sounds crazy but it’s true.

Until now, the only evidence we had that doing all these burns or freezes in the heart helped people was that it made people feel better than they did taking AF drugs.

Better quality of life is important but it seems like a procedure this invasive and expensive should do more than make people feel better.

Well, now, we have evidence that AF ablation improves hard outcomes.

The trial is called CASTLE-AF. It took place in 33 centers across the globe. NEJM published it. Led by my friend Dr. Nassir Marrouche from University of Utah, the authors studied AF ablation vs medical management in patients with chronic heart failure.

Patients had to have an ICD, moderate heart muscle weakness and symptomatic AF. It took about 8 years to select 360 people to enroll in the trial.

The results were positive. Patients in the ablation group had much less AF (which can be measured with their ICDs), fewer admissions for heart failure and lower death rates.

The reductions in outcomes were massive, too. Death was reduced by about 50%.

I wrote a column on the CASTLE-AF trial for the heart.org |  Medscape Cardiology. It’s called Nine Things to Remember About CASTLE-AF. 

It was a hard post to get right. I am close to AF ablation. I ablate AF, and though I believe the procedure is massively overused, I still think it has a role in selected patients. I’ve seen AF ablation help people a lot.

The other thing that made this column hard to get right is that the outcomes were so impressive. A 50% reduction in death outstrips any known therapy for heart failure. How could a procedure that was only partially successful in getting rid of AF be that much better than things like beta-blockers, ACE-inhibitors, and ICDs?

In the post I mentioned some of the limitations of the trial–like the fact that investigators knew which treatment arm patients were in. This lack of blinding can bias the results.

My post generated intense criticism from Dr. Milton Packer, who is a very prominent expert in heart failure. His tone in the comment section was patronizing and sanctimonious towards me. He insinuated that my summary was too favorable and that may have been due to the fact that electrophysiologists stand to gain financially from doing more AF ablation.

The personal attack and patronizing tone does not bother me. Not at all.

Seriously, what bothers me was that Dr. Packer brought up many legitimate criticisms. And I am mad at myself for not raising them more clearly in my post.

For instance, CASTLE-AF had large number of patients lost to follow-up. Most troubling was that more were lost to follow-up in the ablation arm than in the medical arm. This is a big deal because the proportion of lost patients to the total number of events was high. Damn it, that should have been included in my summary.

Another criticism Packer noted was that the primary analysis was based on a small number of events. That raises the possibility that the results may not be replicated in future studies.

Packer also complained that randomized patients as well as events following randomization were excluded from the analysis, which he says is an improper technique.

Packer’s take is that this trial is too flawed to be actionable.

I don’t completely agree.

Although, I should have been more rigorous in mentioning these flaws, CASTLE-AF is not an outlier. It goes in the same direction as previous smaller trials, which also suggested benefit for AF ablation over medical therapy in patients with heart failure.

What’s more, CASTLE-AF deals with a super-select group of patients. These were relatively young (64 years) men with only moderate degrees of heart failure. As the eminent Dr. Richard Lehman writes in his weekly journal round-up, it makes sense that doing ablation rather than giving more of a failed medical therapy produced better outcomes.

CASTLE-AF does not apply to the vast majority patients who come to the hospital with AF and heart failure. The typical patient we see with heart failure and AF are older, frailer, often female, and burdened with many other organ issues.

It would be a tragedy if electrophysiologists used CASTLE-AF results to do procedures on these sorts of patients.

I don’t think that will happen.

A different Dr. Packer, Dr. Douglas Packer, from Mayo Clinic, will likely present results of another AF-ablation outcomes trial later this year. It’s called the CABANA trial, and it will look at outcomes in a more typical group of patients who undergo AF ablation. I look forward to these results.

I’ve learned from this experience.

JMM