EARLY AF and STOP AF Put Electrophysiology at a Crossroads

John Mandrola, MD


November 17, 2020

Two studies published in the New England Journal of Medicine compare ablation to antiarrhythmic drugs early in the course of atrial fibrillation (AF). Their "positive" results scare me; AF ablation is already overused.

Both the U.S. STOP AF trial and the Canadian EARLY-AF study compared the Medtronic cryoballoon system for pulmonary vein isolation vs antiarrhythmic drugs in patients who had not previously tried rhythm control.

For the primary endpoint, STOP AF used a composite of treatment success which included freedom from AF, crossover to ablation, cardioversion, or use of antiarrhythmic drugs (ablation group). EARLY AF used recurrence of any atrial tachyarrhythmia. Both enrolled relatively young patients (about 60 years) who did not have significant structural heart disease or left atrial enlargement.

STOP AF and EARLY-AF both found that ablation reduced the endpoint of AF recurrences compared with drugs. In STOP AF, most treatment failures in the drug therapy arm occurred while patients were taking very low drug doses; in 13% of cases the patients were not taking an antiarrhythmic drug.

EARLY-AF had a more rigorous design. While STOP AF assessed for arrythmias with intermittent 24-hour electrocardiography monitoring, EARLY-AF used implantable loop recorders—which are far more sensitive. EARLY-AF also used drug-specific algorithms designed to maximize dosing in the drug arm, and it measured quality of life (QOL).

Standardized questionnaires revealed that patients enrolled in EARLY-AF were highly symptomatic at baseline. Both treatment strategies markedly improved QOL—though there were significantly larger gains in the ablation arm.

STOP AF reported no difference in safety events but obfuscated the results by including many mundane adverse events, such as heart rate increases and fluid overload. The study even included AF, the efficacy outcome, in the final tally of adverse events. Table S9 in the supplement reveals that concerning ablation complications were not rare: 6% of patients had significant complications from the procedure, including phrenic nerve paralysis and pericardial effusion.

EARLY-AF reported serious adverse events in five patients (3.2%) in the ablation arm, including three patients with phrenic nerve injuries and two who had pacemakers placed for bradycardia. In the drug arm, six patients (4%) had serious adverse events: Two patients had wide-complex tachycardia, two had pacemakers, one had heart failure, and one had syncope.

The messages from these two papers will likely be that early ablation should be considered to arrest the progression of AF and that AF ablation with cryoballoon is safe and more effective than medicines.

My Comments

Since I have opined a lot on this topic, I invited some other electrophysiologists to comment via email.

Dhiraj Gupta, MD, from the University of Liverpool in the United Kingdom, pointed me to this sentence from the STOP AF report: "Medtronic sponsored the trial, was responsible (with steering committee oversight) for development of the final protocol and statistical analysis and reviewed the manuscript that was submitted."

I would emphasize that this does not nullify the results, but it must be considered among the trial's limitations.

These include woefully inadequate dosing in the drug arm. Gupta also noted that crossover from drugs to ablation was counted as a component of the endpoint of "treatment success." Slightly more than 15% of patients crossed over to ablation, mostly for drug side effects—which I find peculiar given the low doses. Given the recent news on nocebo effects, Gupta makes an important point because in an unblinded trial done at ablation centers, having crossover count as an endpoint biases against the drug arm.

Piotr Futyma, MD, from St Joseph's in Rzeszow, Poland, reiterated my concerns about what counted as adverse events in STOP AF. He wrote that "influenza, osteoarthritis, rotator cuff syndrome, and spinal stenosis do not sound so scary, but they drive the total percentage up to 14% of serious adverse events in AAD [antiarrhythmic drug] arm, which can be misleading".

J. Rod Gimbel, MD, from Fort Wayne, Indiana, focused on the lack of a proper control arm in these trials. "It's been over two decades since AF ablation was introduced and it appears impossible to do a trial with a true control group, where simply observation is performed." This comment is especially relevant because both trials enrolled patients with minimal structural heart disease who had relatively low-burden AF.

Gimbel cited two studies showing that we should not be afraid to observe AF initially. The RACE V investigators followed 200 patients with intermittent AF and found widely variable patterns of AF over time. Nearly two-thirds of patients remained stable, 22% progressed to longer episodes, and 16% regressed. Only 3% developed persistent AF. In the AF-RISK study of nearly 400 patients with AF, the progression rate was a mere 13%.

Exposing low-risk patients, many of whom have AF that will not progress, to an invasive procedure is unwise. In EARLY-AF, the median AF burden was 0% in the ablation arm and 0.13% in the drug arm. What might it have been in an arm that included the reassurance of a calm clinician who attended to known risk factors?

I can already hear the proponents of early intervention citing the recent EAST-AFNET 4 trial, which found benefit for early rhythm control. To this I would say, not so fast, read the fine print; that trial is complicated.

Finally, there is the eminent Paul Dorian, MD, from the University of Toronto in Canada, who wrote that "AF in and of itself is at least in part a cosmetic outcome." In other words, AF episodes are a lousy surrogate marker of atrial disease. He added that when symptoms are considered, treatment expectations will be influenced by the type of therapy, a bias that favors ablation over drugs.

Dorian is right; expectations clearly play a role in interpreting the QOL signals in EARLY-AF. You have median AF burden approaching 0% in both groups, but patients in the ablation arm feel a lot better. The caring signal of an ablation crushes that of drugs.

Clinical Implications

I wrote to the primary investigator of EARLY-AF, Jason Andrade, MD, from the University of British Columbia in Canada about my concerns that two positive trials published in the New England Journal of Medicine might be used to expand cryoballoon ablation to centers with less expertise—perhaps even to interventional cardiology.

Andrade wrote that he strongly disagrees with the idea that "cryoballoon PVI [pulmonary vein isolation] is an IC [interventional cardiology] procedure to be done willy-nilly. We should not be pushing cryoballoon ablation out of the hands of experienced EP labs." He directed me to the subgroup analysis of EARLY-AF, which found a strong trend toward better outcomes in more experienced labs.


The results of these trials should surprise no one. Oodles of unblinded trials show that AF ablation beats drugs when you measure AF recurrence or quality of life. The question is not whether we can ablate AF but whether we should do so in the patient in front of us.

Given the above critiques of the evidence, I strongly oppose the use of these 1-year trials to pressure patients into an early invasive approach to their AF. We must remain aware of the severe asymmetry of power in the doctor-patient relationship.

I believe AF ablation has a role, but a slow, calm, medically conservative approach remains the right approach in lower-risk patients. This procedure can cause significant harm.

The field of electrophysiology is at a crossroads. I hope it chooses the right path.

John Mandrola practices cardiac electrophysiology in Louisville, Kentucky, and is a writer and podcaster for Medscape. He espouses a conservative approach to medical practice. He participates in clinical research and writes often about the state of medical evidence. 

Follow John Mandrola on Twitter

Follow | Medscape Cardiology on Twitter

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


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.