By Joshua D. Moss, MD

Associate Professor of Clinical Medicine, Cardiac Electrophysiology, Division of Cardiology, University of California, San Francisco

Dr. Moss reports he is a consultant for Biosense Webster and Abbott.

SYNOPSIS: Ablation for atrial fibrillation did not reduce the composite endpoint of death, disabling stroke, serious bleeding, or cardiac arrest compared with drug therapy, although adverse events were infrequent and arrhythmia recurrence was reduced significantly.

SOURCE: Packer DL, Mark DB, Robb RA, for the CABANA Investigators. Catheter Ablation versus Antiarrhythmic Drug Therapy for Atrial Fibrillation (CABANA) trial: Study rationale and design. Presented at the 39th Heart Rhythm Scientific Sessions, Boston, May 10, 2018. Results published in Am Heart J 2018;199:192-199.

The long-awaited results of the Catheter Ablation vs. Antiarrhythmic Drug Therapy in Atrial Fibrillation (CABANA) trial were presented as a late-breaking session at this year’s Heart Rhythm Scientific Sessions. The goal of the randomized study, which began enrolling patients in 2009, was to compare catheter ablation to state-of-the-art drug therapy for patients with new onset or “undertreated” paroxysmal or persistent atrial fibrillation (AF) that warranted therapy. Patients < 65 years of age also were required to exhibit at least one cardiovascular or stroke risk factor to be eligible. Patients in the ablation arm underwent pulmonary vein isolation at a minimum, and patients in the drug therapy arm could be treated with rate or rhythm control. Originally, the primary endpoint was defined as mortality. However, partway through the trial, investigators changed this to a composite endpoint of all-cause mortality, disabling stroke, serious bleeding, or cardiac arrest due to difficulty with enrollment and low mortality.

A total of 1,108 patients were randomized to ablation therapy, 1,006 of whom underwent ablation and 102 of whom crossed over to the drug therapy arm. A total of 1,096 patients were randomized to drug therapy, 301 of whom crossed over to receive catheter ablation. Approximately one-third of patients in both arms were < 65 years of age, and approximately 37% were female. Baseline cardiomyopathy was slightly more prevalent in the drug therapy arm, although rates of congestive heart failure were similar. About 10% of patients had a history of prior stroke or transient ischemic attack. Overall, about 90% of patients completed a mean follow-up around four years. Complications in the ablation group were rare; hematoma at the catheter insertion site (2.3%) was most common. There were eight episodes of perforation and cardiac tamponade, but no atrioesophageal fistulas. In the drug therapy arm, hypo- or hyperthyroidism occurred in 1.6% of patients.

Over 60 months of follow-up, there was no statistically significant difference in the primary endpoint by an intention-to-treat analysis, although rates remained slightly lower in the ablation group throughout. Mortality alone also was not significantly different. There was a significant reduction in combined death or cardiovascular hospitalization in the ablation group, a secondary outcome, with a hazard ratio (HR) of 0.83. A subgroup analysis suggested a possibility of improved outcomes with ablation in younger patients and minorities. Freedom from recurrent AF was about 65% at one year in the ablation arm vs. only about 40% in the drug therapy arm. Rates of recurrent AF remained significantly lower in the ablation arm throughout follow-up (HR, 0.53; 95% confidence interval, 0.46-0.61). When groups were compared based on treatment received, there were significant reductions in the primary combined endpoint (HR, 0.67) and all-cause mortality (HR, 0.60) in the ablation arm compared with drug therapy.


The presentation of CABANA was highly anticipated, and electrophysiologists and cardiologists alike continue hotly debating the meaning and results. Many held up the “negative” result, with no significant difference in the combined primary endpoint, as proof that catheter ablation was overused compared with antiarrhythmic drugs. One prominent commentator even suggested that any alternate interpretations were evidence that electrophysiologists must be hopelessly conflicted by the financial rewards of a complex procedure. By contrast, many others pointed to the dramatic reduction in recurrent AF in the patients randomized to ablation, with minimal cost in terms of procedural risk and no cost in terms of stroke, mortality, or bleeding.

I drew the following conclusions from the available data:

  1. Ablation for atrial fibrillation is at least as good as medical therapy. Many electrophysiologists hoped to see superior results with ablation. They were either disappointed in the actual results, or eager to focus on the on-treatment analysis, which is scientifically dubious. Researchers use randomization for a reason, and we do not know how patients who crossed over to ablation were different from other patients in the medical therapy arm. However, many of us were pleased to receive confirmation that outcomes such as mortality, disabling stroke, and serious bleeding were, at worst, equivalent in an intention-to-treat analysis. In general, crossover in a randomized trial tends to bias outcomes toward the null-hypothesis (i.e., toward no difference in treatment effect). Ablation proved to be as good as medical therapy despite that crossover.
  2. Clearly, ablation is more effective than medical therapy at preventing AF. Reduction in arrhythmia burden was not in the primary combined endpoint, and it often was overlooked during ensuing debates about intention-to-treat vs. on-treatment results. However, the likelihood of recurrent AF was reduced almost 50% with ablation compared with medical therapy — again, despite the crossover. Cardiovascular hospitalization also declined. Our primary objective in treating AF in patients without heart failure always has been to reduce symptoms and improve quality of life. It will be interesting to see the full impact of ablation in the final published results.
  3. Ablation is safe. Overall rates of complications were low, a finding that was important to confirm for a complex procedure performed electively.
  4. A double-blinded trial of ablation is likely to follow. Despite the clear reduction in AF with ablation, and the undisputable correlation between symptoms and arrhythmia in some patients, the true vs. placebo effect of ablation on symptoms in the broader population of patients with AF is unknown. Particularly after the successful completion of the ORBITA trial and its intriguing results, we probably can expect a similar trial of ablation with a sham procedure control in the near future. The ethics and comparability of such a sham procedure, depending on how it is performed, undoubtedly will be the subject of more debate.