To Anticoagulate or Not in Subclinical AF: ARTESiA

Recorded November 12, 2023. This transcript has been edited for clarity.

John M. Mandrola, MD: Hi everyone. This is John Mandrola from theheart.org | Medscape Cardiology. I'm here at the American Heart Association meeting in Philadelphia, and I'm excited to have Professor Jeff Healey from McMaster University, who was a primary investigator of the ARTESIA trial. ARTESIA is a really important trial because it addresses the question of what to do with patients who have short-duration subclinical atrial fibrillation (AF).

Jeff, welcome, and thanks for being with us.

Jeff S. Healey, MD, MSc: Thanks for the invite, John.

Mandrola: Tell us about what you found right off the bat, and then we'll go into the details.

Healey: In patients with pacemakers and implantable cardioverter-defibrillators, we found that among those with 6 minutes to 24 hours of subclinical AF, apixaban reduced the risk for stroke or embolism by approximately one third, with about a 50% reduction in large strokes that were disabling or fatal. To balance that, there was also an increase in the risk for major bleeding, which was a hazard ratio of 1.8 in an on-treatment analysis.

Aspirin as Control

Mandrola: Just a few weeks ago, we heard about NOAH-AFNET 6, which was edoxaban vs placebo, but 50% of that control arm was on aspirin. Why did you pick aspirin as the control? How does that influence the interpretation of these lower stroke rates and higher bleeding rates?

Healey: Aspirin was based on the patients themselves. At the beginning of the trial, prior to randomization, two thirds of our patients took an antiplatelet agent: aspirin, clopidogrel, or something else. This is the baseline and why we chose this as a comparator.

Also, as you recall, there's a difference between NOAH and ARTESIA here, and it's borne out of the difference between the American and the European approach to AF from 10 to 15 years ago. For many low-risk individuals in North America, we used aspirin. There was this expectation when we approached sites to participate that we have to try something with our patients who have this. There was this culture of using low-dose aspirin here in America that wasn't present in Europe, as they had abandoned it from their guidelines much earlier. I think that's why there is the difference between the two trials.

In terms of the effect, aspirin will possibly modestly reduce the risk for stroke. The meta-analysis suggests that at most, it's a 20% reduction, so not a big reduction, but of course, it is associated with a risk for bleeding and so would increase it compared with placebo. One third to one half of individuals were different between the trials where the control arm wasn't active therapy in NOAH, where it was aspirin across the board with ARTESIA.

Mandrola: What kind of patients were in ARTESIA?

Healey: First of all, this is a very common problem. You and I are both electrophysiologists. If you take those older than 65 years with hypertension, roughly one third will have this. This is common. Episodes had to be 6 minutes to 24 hours. We didn't allow any longer episodes, unlike NOAH. Furthermore, if during the course of the trial patients developed longer episodes, they were taken off study medication and treated with an anticoagulant.

We were looking for an enriched population, right? Because this is lower risk than clinical AF, we went for a higher-risk group, so a minimum CHA2DS2-VASc score of 3. We did allow patients with prior stroke or who are older than 75 years as a sole risk factor because those are such important risk factors. That's the population that we studied.

Mandrola: The CHA2DS2-VASc entry criterion was a score of 3, but it was actually higher in enrolled patients.

Healey: Minimum 3. There were some patients with very high CHA2DS2-VASc scores.

Mandrola: What was the median duration of the episodes? How short?

Healey: It's 1.5 hours, so that's shorter than NOAH, which speaks to the difference in the upper cutoff that we used that they didn't use, and shorter than in the observational trials like ASSERT, which probably speaks to some degree of selection bias that we had here.

Stroke Risk in Subclinical AF vs Clinical AF

Mandrola: We've all been discussing in our groups that this is a really low stroke rate. We saw that in NOAH, and of course, the hazard ratio for NOAH was definitely favoring the direct-acting oral anticoagulant (DOAC), but it didn't reach statistical significance because the stroke rates were low. You also found low stroke rates, even with high CHA2DS2-VASc scores. There's something different about this subclinical AF.

Healey: From our ASSERT trial, we knew that we're dealing with a different entity in terms of relative and absolute risks. This is why we defined this 24-hour cutoff where they started to behave the same as clinical AF. In this trial, we really focused on this pre-AF stage. However, as we found out during the trial, about 1 in 4 patients within the first 1.5 years progresses to clinical AF or longer episodes. We really focused on this early, early window.

Yes, it is a lower risk for stroke: Across the board, it is around 1.24% per year on aspirin. Again, I think it's a very interesting place to be and where we'll really need to understand the clinical risk factors, the features of the arrhythmia — duration, burden — that influence risk, and maybe even some of the echo factors that we collected. CHA2DS2-VASc for AF was built around identifying low-risk patients under 1% per year where we didn't have to give an anticoagulant. That treatment threshold, if you will, is somewhere in that 1%-1.5% per year, and this lines up quite in the middle.

I think what we will find — without revealing any data, which are just currently being analyzed — is that we all know that there are many good stratifying tools, such as age, prior history of stroke, CHA2DS2-VASc, and maybe new ones that we're exploring here. If you're getting an average risk in that range on aspirin, I think it's highly probable that we will define significant subgroups where the risk will be even higher, and the treatment risk-benefit ratio weighing will be a little more obvious for clinicians.

To Anticoagulate or Not?

Mandrola: I can't tell you how excited we will be to learn that there's better risk stratification coming than CHA2DS2-VASc. You have a clear stroke reduction, statistically significant, but yet, there is a tradeoff with higher bleeding. When my partner calls me up and ask me what we should do with a 2-hour episode of AF, what do you think?

Healey: It's a very complex problem, hence the question, and there are many factors to consider, like patient preference, which generally favors stroke avoidance rather than bleeds because of the significance. We've tried to limit the paper to just the facts in the studies so far, but there are these other factors, such as guideline committees and health economics; that all has to be weighed in.

What we looked at in the trial was assessing not just a major bleed and a stroke but rather what did they look like. The modified Rankin score is a very nice way to break down whether these were small strokes that people walked away and said, "Hey, no big deal." What we found, which was surprising to me, was that about 45% of the strokes were in the Rankin 3-6 score. This is permanent, moderate to severe disability, or death. These are not trivial strokes.

On the other hand, on the bleeding side, about 85% of the bleeding in the apixaban arm was resolved with just observation or simple interventions, including crystalloid and blood transfusion. In the very severe patients, where you needed surgery or the bleed led to death, this was less than 15% and numerically similar in both groups, as were fatal bleeds and intracranial bleeds.

Again, you're weighing a problem that occurs for a few days vs a problem that gives lifelong impact in strokes. It's not going to be a flat comparison one to one. This is going to require further analysis.

Mandrola: Jeff, I have to say that now, I'm confused because I was at the European Society for Cardiology Congress when Professor Kirchhoff stood up and said, "This is practice-changing. We don't want to treat these patients with anticoagulants." It's a different interpretation, I think. Is this true?

Healey: It's a different conclusion, but I think the data speak the same message. At the sessions, we will see results of a meta-analysis. Anyone who's read both papers, which are now out there, can see that if you look at the confidence intervals of all the outcomes — stroke, bleeding — they overlap quite a bit.

We may be dealing with a simple issue of statistical power between the trials, a difference in choice of primary outcome, with ARTESIA including a very narrow focus on stroke and embolism, and NOAH have a broader focus, including things like cardiovascular death, which we know to be less responsive to anticoagulants.

Mandrola: It's not going to be an algorithmic answer. We're going to have to think in clinic about stroke risk reduction, bleeding increase, and mixing that into what patients value.

Healey: Of course, that's why we have doctors.

Mandrola: I think there was a famous doctor, who came from your area, who talked about best evidence and mixing that in.

Excellent. This is such an important trial, and I really appreciate you taking the time to be with us. Thank you.

Healey: Thanks for having me.

Follow theheart.org | Medscape Cardiology on X (formerly known as Twitter)

Follow Medscape on Facebook, X (formerly known as Twitter), Instagram, and YouTube