CABANA in Context: A Nuanced Understanding of AF Ablation

Elaine M. Hylek, MD, MPH; Jonathan P. Piccini, MD, MHS, FACC; Christian T. Ruff, MD, MPH


May 30, 2018

Editorial Collaboration

Medscape &

Elaine M. Hylek, MD, MPH: Hello. I'm Elaine Hylek from Boston University. I am here at the Heart Rhythm Society (HRS) 2018 Scientific Sessions in Boston, where the long-awaited CABANA trial on ablation versus medical therapy in patients with atrial fibrillation (AF) was presented. To discuss these results with me, I am happy to be joined by Dr Jon Piccini, an electrophysiologist from Duke University, Duke Medical Center, and Dr Christian Ruff, a cardiologist from the Brigham Women's Hospital and Harvard Medical School.

Jon, you were an investigator in the CABANA trial. Why don't you walk us through the trial design, what was tested, endpoints, and key findings.

CABANA Trial and Results

Jonathan P. Piccini, MD, MHS, FACC: CABANA[1,2] is an international trial of 2200 patients who were 65 years of age or older with AF or younger patients who had cardiovascular risk factors. Those patients were randomized to a therapeutic strategy: catheter ablation or medical therapy.

It was a fairly big effort. The median follow-up was 48 months. The patient followed the longest was followed for 8.1 years. This is pretty unusual in an ablation trial and a testament to the investigators.

The primary hypothesis was that catheter ablation compared with medical therapy would reduce all-cause mortality. Several years into the trial, owing to the blinded event rates being lower than expected and enrollment being lower than expected, the primary and secondary endpoints were flipped. The primary endpoint became the composite of death from any cause, disabling stroke, serious bleeding, or cardiac arrest.

At the conclusion of the trial, there was no difference in the intention-to-treat (ITT) analysis for the primary outcome or the secondary outcome of all-cause mortality. There was a difference in the secondary endpoint of cardiovascular hospitalization or death.

However, one of the key challenges with the study, as might be expected in such a large effort over such a sustained period of time, was that 9% of the patients randomized to ablation never received the assigned treatment ablation and 27% of the medical therapy patients crossed over to ablation. There are some challenges with the strict interpretation based on ITT.

Because of that, a prespecified as-treated analysis was performed, which did suggest a reduction in both the primary endpoint of death, disabling stroke, cardiac arrest, or serious bleeding and a reduction of mortality as well. Of course, those are the as-treated results and not the primary results, which are ITT.

Hylek: As you point out, whenever you run into some of these issues with the ITT, it makes sense to then look at those who got the intervention. It is important to embrace both of those results. We will have Christian walk us through what he feels are some of the other discussion points.

Intention-to-Treat Issues

Christian T. Ruff, MD, MPH: These trial populations are critical. Regulators and guidelines often view ITT as the gold standard because you want the only difference between the two groups to be whatever the randomization is. Every other factor you can imagine about those patients, such as age and medicines, is completely balanced. It is a true testing of A versus B.

Any time you deviate from that, with an as-treated or an on-treatment analysis, there are potential differences because they are not randomized comparisons anymore. There may be differences why a patient either received the therapy or did not that you cannot account for. You always have to acknowledge that there may be some potential bias in your interpretation.

When you look at the robustness of the on-treatment results, although not randomized, it looks like ablation had very significant benefits. When you are trying to figure out how these drugs, devices, or strategies are going to be used in the real world, or who should be on medical therapy versus who should be having an ablation, an on-treatment or as-treated analysis really helps inform whether there is a potential biological benefit that is probably going to lead to improved outcomes.

Hylek: Right. The ITT is going to help you maintain randomization, and hopefully minimize bias chance and confounding. I think that has always been considered the gold standard. There was not really selection criteria as to why patients ended up with one or the other, and the 27% crossover [to ablation] is not trivial, so viewing both of those results together makes a lot of sense. Is there anything other than the ITT? In drug trials, we often do a safety analysis with the on-treatment or per protocol groups, because you want to look at safety in those individuals who are actually taking the drug. There are standard ways of looking at data.

Ruff: Is there a clinical benefit for people who are getting the therapy you are testing, whether a drug or device? I think the as-treated population in the CABANA trial shows that patients who received ablation did better. So there is clearly a clinically important benefit in patients who are having an ablation. For assessing the strength and robustness of trial results, it's important to use both because both are meaningful. They are answering two different questions.

Bundled Endpoint

Hylek: What about the bundled primary endpoint? How do you interpret that?

Ruff: Composite endpoints are very common in clinical trials for drugs or devices. We all do this because if you combine endpoints, you can conduct smaller, more streamlined trials, whereas trials powered for a single endpoint tend to be very large and expensive.

When you look at a composite outcome, it is important to say, "What are the things that you are adding in there? Are they all meaningful?" In this case, mortality, hospitalizations, serious bleeding, disabling stroke, and cardiac arrest are important clinically.

Hylek: It was not embellished with endpoints that we or our patients would not care about.

Ruff: You look to see whether there is a consistent benefit in each of those individual components or whether it driven entirely by one of the outcomes. It looked, at least from the results [so far], like there was consistency in a lot of those individual components, which I think gives more credence that this was a reasonable composite outcome.

Piccini: We are only a couple hours out from the presentation, and I think the trial has already been reduced in many forums to either ITT or per protocol as-treated. I don't think interpretation of the trial can be viewed in either prism alone. It is nice to see the general consistency between the ITT and the as-treated analyses.

Safety and Study Strengths and Limitations

Piccini: The other thing that is really important is safety. Any time there is an interventional procedure, there are questions about safety. Any time we talk about membrane-active antiarrhythmic drugs, there are questions about safety. It was very reassuring to see that the incidence of tamponade was around 1%. In patients randomized to the medical arm, it was very reassuring to see that the risk for ventricular proarrhythmia was 0.8%. Those are important findings that patients often ask about when they are trying to make this decision.

Hylek: I would add a little caveat that not every center in the world is going to be like your centers. There is some expertise in these centers.

Ruff: That is true. We face this in drug trials, where centers running trials may not represent global practice. Experience of the center and the proceduralist is critical when studying a procedure. It is reassuring that in experienced centers, it looks like ablation is relatively safe. You can offer ablation to patients and reasonably expect that the serious complication rates are really low. This is an important option for patients and with very reassuring safety data.

Hylek: Christian, can you summarize what you think were the strengths and limitations of the study?

Ruff: As Jonathan said, this is a large study, and it offers some very important data in that it was fairly inclusive. Lots of trials around ablation have addressed a specific population, but this used a broad strategy with not a lot of exclusion criteria. One nice thing about this trial is the generalizability, that in a large group of patients with AF, ablation is a reasonable strategy and certainly very safe. Actually, patients who received ablation had significant reductions in clinical events. In that sense, it is an important advance. This is really one of the strongest and largest ablation trials we have.

One of the limitations, shown with the primary endpoints in the ITT population, is that technically it's a neutral study. I don't think we can interpret this as saying that everybody who has AF should receive an ablation, certainly not as first-line therapy. We need to understand the nuances of the trial—populations that seemed to benefit and outcomes that were particularly impacted. It is a more nuanced understanding of who we should we be doing ablations in up front and those who potentially may not benefit as much in clinical practice.

Subgroup Findings

Hylek: When you look at the subgroup analyses, with all the caveats about looking at subgroup analyses, there was a suggestion that it might be a little better individuals with NYHA class II or greater heart failure, potentially individuals younger than 65 years (even though those P values for interaction were not statistically significant). Any comment, Jon?

Piccini: You summarized the subgroup findings perfectly. There is an important clinical message: Several clinical trials have demonstrated improvement in cardiovascular outcomes in patients with symptomatic heart failure who have AF. The AATAC study[3] compared amiodarone with catheter ablation, CASTLE-AF[4] compared catheter ablation with medical therapy, and now we have CABANA, whose results suggest that restoring and maintaining sinus rhythm in patients with heart failure appears to yield significant benefits. It is not a story that is completed. In fact, rather than the beginning of the end, we may be at the end of the beginning.

Regarding age, I think it's tough to know. Maybe the data are pointing us to a hypothesis where younger patients who have more of their natural history in front of them benefit from better maintenance of sinus rhythm? It's important to note that catheter ablation was associated with a 47% lower risk for recurrent AF. In the ablation arm, maintenance of sinus rhythm was a little under 60% at 4 years, which is pretty amazing.

I don't think we should be ageist. If there is a 75-year-old patient who is active, and healthy with a single risk factor or two, I don't think that we should not offer that patient ablation. On the other hand, if there is a 78-year-old patient who has multiple comorbidities, it's probably very reasonable to offer that patient medical therapy.

Hylek: There is a debate about stopping anticoagulation, and I am not sure that data were presented. Does the low stroke rate perhaps reflect continued anticoagulation?

Piccini: That is a hugely important point. Many trials have shown that after catheter ablation, patients with multiple risk factors for stroke do better when their anticoagulation is continued. Dr Packer pointed out last night at the investigators committee meeting that the rates of stroke were very low because there was careful attention to anticoagulation. That is also a significant message from the trial.

Ruff: By decreasing the burden of AF, are you potentially improving clinical outcomes? These are not cures for AF; we are not eliminating AF. Even the link between the rhythm itself and the mechanism of stroke is still not perfectly understood. At this time, you may choose ablation for improving quality of life and potentially improving outcomes, particularly in the subgroups that have more robust benefits for ablation. Regardless of that decision, the totality of the evidence says if a patient qualifies for anticoagulation, no matter which medicine or procedure you do, you should anticoagulate that patient because that is the only reliable way we know to reduce ischemic stroke.

Will CABANA Change Practice?

Hylek: Will practice change? Will there be more, less, or the same number of ablations after CABANA?

Ruff: In some ways, I think it will reinforce what we are already starting to do in clinical practice, which is moving ablation more up front for many of our patients, particularly younger patients who have very symptomatic AF looking at a lifetime of AF. Intervening early may improve long-term outcomes. In patients with heart failure with reduced ejection fraction, there may be real and robust benefits—even for mortality and certainly for heart failure hospitalization.

Because ablation looks to be a very safe procedure, I think that people will think of ablation earlier than they would have initially. We used to say that a patient had to fail a whole host of medical regimens, particularly antiarrhythmics, before we defaulted to ablation. Now it is a reasonable alternative early for many patients, but I still don't think we are at the point where we are doing it on everybody.

Hylek: The other issue was repeat procedures. I was under the impression that people were getting all these repeat procedures. But it didn't seem to be that high [in this trial]. This trial had a fairly long follow-up.

Piccini: Yes, that was partially influenced by the protocol, where the intention was a single ablation. Again, that makes the maintenance of sinus rhythm and long-term follow-up all the more impressive. I agree with Christian. The primary indication for catheter ablation in 2018 is for the treatment of symptomatic AF in patients who have failed medical therapy. This trial, in combination with CASTLE-AF and others, is showing us that although it may not be superior, catheter ablation may be better for several important patient groups, and one of them is patients with heart failure —perhaps younger patients as well.

Hylek: Thank you both for joining me today. We have been waiting forever for CABANA and for your balanced assessment of the trial. I hope our audience also enjoyed this discussion of the long-awaited CABANA trial. Thank you very much for your attention.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.