Sadly, ARTESIA Doesn't Answer a Common Question in Cardiology

John M. Mandrola, MD


November 15, 2023

What a super study. A great effort.

ARTESIA investigators randomized more than 4000 patients with short-duration asymptomatic atrial fibrillation (AF) to receive either apixaban or aspirin. They measured an important low-bias endpoint: stroke and systemic embolism. Trial conduct was excellent.

Yet, even when combined with the similar NOAH-AFNET 6 trial, the results do not provide a clear answer on how best to treat patients with these common arrhythmias.

There will be no algorithms or quality measures for treating patients with subclinical AF. Despite two large trials, clinicians will have to stop, think, listen to their patients, and align care with each patient's specific goals and preferences — just as the late David Sackett, MD, laid out when he pioneered modern-day evidence-based medicine.


Patients with device-detected asymptomatic short-duration AF were recruited from nearly 250 sites in 16 European and North American locations.

The average age was 77 years, and slightly more than one third of the patients were female. The mean CHA2DS2-VASc score was 3.9. The median duration of AF was 1.5 hours. Only 20% of patients had AF duration longer than 6 hours.

A primary endpoint occurred in 55 of 2015 patients in the apixaban arm vs 86 of 1997 patients in the aspirin arm (0.78% vs 1.24% per patient-year). This difference of only 31 stroke events in a trial of more than 4000 patients yielded a statistically significant relative risk reduction of 37% (HR, 0.63; 95% CI, 0.45- 0.88; P = .007).

The authors assessed stroke severity and reported that 18 of 55 strokes in the apixaban group were disabling or fatal vs 36 of 84 strokes in the aspirin group (33% vs 43%, respectively). The relative risk reduction for these severe strokes, with a modified Rankin scale score of 3 to 6, was greater at 49%.

The safety endpoint of major bleeding occurred in 86 of 1989 patients in the apixaban arm and 47 of 1972 patients in the aspirin arm. This difference of 39 excess major bleeding events yielded a relative risk increase of 80% (HR, 1.80; 95% CI, 1.26-2.57; P = .001).

The investigators also subdivided the bleeding events, noting that most bleeding events were handled with supportive care. Fatal bleeding and intracranial bleeding events were actually numerically lower in the apixaban arm.

They properly reported the efficacy results of stroke and systemic embolism using intention-to-treat analysis and the safety endpoint of bleeding with an on-treatment analysis.

The Authors' Conclusions

The main conclusion in The New England Journal of Medicine simply is that "apixaban resulted in a lower risk of stroke or systemic embolism than aspirin but a higher risk of major bleeding."

But in the manuscript and at the AHA meeting, the message was more focused. The authors encouraged clinicians to consider both the benefits and risks of apixaban therapy.

"Simply counting strokes as compared with bleeding events might suggest a neutral overall effect," they write, noting that "strokes involve permanent loss of brain tissue whereas major bleeding is usually reversible with most patients having a complete recovery."

My Comments on NOAH-AFNET 6 and ARTESIA

The results of both NOAH-AFNET 6 and ARTESIA align well with the known risk-benefit thresholds we use today for clinical AF. The CHA2DS2-VASc score estimates (albeit imperfectly) the untreated stroke risk. If it is low, we withhold anticoagulants, not because the drugs don't decrease stroke risk, but because the degree of stroke risk reduction is not outweighed by the increased rate of bleeding.

That is exactly what happened in both the trials. Take ARTESIA. The average duration of AF was only 1.5 hours. Most patients had AF duration less than 6 hours. We know from the TRENDS observational study of patients with cardiac devices and short-duration AF that 5.5 hours of AF was the threshold for increased stroke hazard. And in the ASSERT study, the AF duration threshold for elevated stroke risk was closer to 24 hours.

The issue, therefore, for both trials is that subclinical AF does not confer a high enough yearly stroke risk to warrant anticoagulation. In ARTESIA, the yearly stroke risk in the aspirin arm was only 1.24% — well below the standard estimate of 2.2% for a CHA2DS2-VASc score of 2. In AVERROES, a trial that found superiority of apixaban over aspirin in patients with clinical AF, the rate of stroke in the aspirin arm was 3.7% per year.

The ARTESIA authors argue that strokes are worse than bleeds, so we should favor oral anticoagulation. I understand that sentiment, but I don't completely agree with it.

Here is why: The low absolute stroke rates in ARTESIA means that the 37% reduction in stroke translates to an absolute risk decrease of only 1.6% over 3.5 years. That's a number needed to treat (NNT) of 62. That small reduction is balanced not only by a similar increased rate of major bleeding events but also the burden and costs of taking a twice-daily pill for years.

This is why Sackett's view of evidence-based medicine, in which we use the best evidence and align it with patient's preferences, is critical. I am certain that tomorrow in clinic, I could have two similar patients with subclinical AF who would choose opposite therapies.

One big upside of having two big randomized trials is what we will soon learn. There is already a trial-level meta-analysis, but I expect that the authors will share and pool their patient-level data. That will allow estimates between AF duration and treatment effect. I look forward to the graph that plots AF duration with stroke risk reduction. It still won't be perfect, but we will have a better sense about how much AF is required to gain a net benefit from anticoagulation.

In sum, these two trials give me pride in my field of electrophysiology. We now have evidence where there once was none. That it still requires clinicians to stop and think is fine. Kudos to the trialists.

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.


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.