Apixaban Cuts Stroke but Ups Bleeding in Subclinical AF: ARTESIA

In patients with subclinical atrial fibrillation (AF) detected by implanted devices such as pacemakers or loop recorders, oral anticoagulation with apixaban resulted in a lower risk of stroke or systemic embolism than aspirin, but a higher risk of major bleeding in the ARTESIA study.

The results appear to contrast somewhat with the recently reported NOAH-AFNET 6 trial, which failed to show a reduction in stroke with the anticoagulant edoxaban vs placebo in a similar patient group, but that trial was stopped early and so was underpowered.

However, the lead investigators of both trials say the studies actually show consistent results — both found a lower rate of stroke than expected in this population, but the confidence intervals for stroke reduction with anticoagulation overlap, suggesting there is likely some effect, albeit less than that in clinical AF.

The big question is whether the reduction in stroke with anticoagulation outweighs the increase in major bleeding.

A new meta-analysis of the two trials showed that "oral anticoagulation with edoxaban or apixaban reduces the risk of ischemic stroke by approximately one-third and increases major bleeding by roughly double."

In absolute numbers, there were three fewer ischemic strokes per 1000 patient years with anticoagulation in the two trials combined, at the cost of seven more major bleeds.

The lead investigators of the two trials have somewhat different opinions on how these findings may translate into clinical practice.

Jeff Healey, MD, Population Health Research Institute, McMaster University, Hamilton, Ontario, Canada, lead investigator of the ARTESIA trial, believes that the risks and benefits need to be assessed in individual patients, but there should be some patient groups that will benefit from anticoagulation treatment.

"In patients with pacemakers or implantable loop recorders with continuous monitoring, subclinical AF is detected in about one third of patients, so this is extremely common," he told theheart.org | Medscape Cardiology. "The question is whether this is just a normal feature of getting older or is this like AF that we see in the clinic which increases stroke risk, and I think we can conclude from ARTESIA that this subclinical AF is associated with an increased risk of stroke, although that is lower than the risk with clinical AF, and that it can be reduced by anticoagulation."

Until recently it hasn't been possible to quantify the risk associated with subclinical AF, he noted. "But now we have a rich dataset to use to see if we can tease out some specifics on this. Future analyses of this dataset will help define patients where the benefits outweigh the risks of bleeding. For now, I think we can look at the data in a qualitative way and consider the totality of risk factors in each patient — their bleeding risk, stroke risk, how much AF they have, and make a decision as to whether to give anticoagulation or not."

But Paulus Kirchhof, MD, University Heart and Vascular Center Hamburg, Germany, lead investigator of the NOAH-AFNET 6 trial said, "Both trials showed the stroke rate is low in these patients — about 1% per year — and that anticoagulation can reduce it a bit further at the expense of increasing major bleeding. I don't believe the AF episodes picked up on these devices constitute a sufficient stroke risk to warrant anticoagulation, given the bleeding risk."

Kirchhof suggests an alternate approach of performing further traditional AF monitoring on these patients.

"I think going forward in my practice, when we come across this device-detected AF, we will do further investigations with an established method for detecting AF involving surface ECG monitoring — maybe a 3-day or 7-day Holter. If that shows AF, then we will be on firm ground to start anticoagulation. If that doesn't show AF, we will probably not use anticoagulation."

The ARTESIA trial and the meta-analysis of the two trials were both presented today at the 2023 American Heart Association (AHA) Scientific Sessions, being held in Philadelphia, Pennsylvania. Both studies were also simultaneously published online — ARTESIA in the New England Journal of Medicine and the meta-analysis in Circulation.

ARTESIA

For the ARTESIA study, 4012 patients with device-detected AF and other clinical risk factors for stroke were randomly assigned to treatment with apixaban (5 mg twice daily) or aspirin (81 mg daily).

After a mean follow-up of 3.5 years, the primary endpoint — stroke or systemic embolism — occurred in 55 patients in the apixaban group (0.78% per patient-year) compared with 86 patients in the aspirin group (1.24% per patient-year), giving a hazard ratio (HR) of 0.63 (95% CI, 0.45 - 0.88; P = .007).

"The risk of stroke or systemic embolism was lower by 37% with apixaban than with aspirin, and the risk of disabling or fatal stroke was lower by 49%," Healey reported.

In the "on-treatment" population, the rate of major bleeding was 1.71% per patient-year in the apixaban group and 0.94% per patient-year in the aspirin group (HR, 1.80; 95% CI, 1.26 - 2.57; P = .001).

Fatal bleeding occurred in five patients in the apixaban group and eight patients in the aspirin group. Symptomatic intracranial hemorrhage occurred in 12 patients with apixaban and 15 patients with aspirin.

One of the main findings of the trial is the lower-than-expected risk of ischemic stroke in this population — about 1% per year in the aspirin group, which was reduced to 0.64% per year in the apixaban group.

The authors note that, "Simply counting strokes as compared with bleeding events might suggest a neutral overall effect. With apixaban as compared with aspirin, 31 fewer cases of stroke or systemic embolism were seen in the intention-to-treat analysis, as compared with 39 more major bleeding events in the on-treatment analysis."

However, they point out that strokes involve permanent loss of brain tissue, whereas major bleeding is usually reversible, with most patients having complete recovery, which was the case in this study.

"Thus, on the basis of the considerably greater severity of the stroke events prevented than the bleeding events caused, we believe that these findings favor consideration of the use of oral anticoagulation for patients with risk factors for stroke in whom subclinical atrial fibrillation develops," they conclude.

First Well-Powered Trial Addressing This Question

Discussing the ARTESIA trial at an AHA press conference, Christine Albert, MD, Cedars-Sinai Medical Center, Los Angeles, California, said; "I want to emphasize how important this trial is."

She explained that current guidelines do not recommend any treatment for patients with device-detected AF that is not shown on ECG, even though it is known this confers some excess risk of stroke.

"ARTESIA is the first well-powered, long-term trial looking at this question," she said. "It found a clear reduction in the risk of stroke/systemic embolism with apixaban vs aspirin, but there was also a significant amount of bleeding — about an 80% increase. The question is whether the benefit on stroke is worth it given the bleeding risk."

Albert highlighted the low absolute risk of stroke in this study population of around 1.2%, pointing out that even with the 37% relative reduction with anticoagulation, stroke is only reduced in absolute terms by 0.4%.

"We are going to have to take this back to committees and guidelines and look at the balance between the benefit on stroke and the increase in bleeding," she concluded.

Noting that observational studies have shown that the duration of AF impacts the risk of stroke, Albert suggested that patients with longer durations AF may benefit from anticoagulation to a greater extent; and given that the bleeding seen in ARTESIA was mainly GI bleeding, it might be possible to screen out patients at high risk of GI bleeding.

She also pointed out that a lot of patients discontinued anticoagulation treatment in both ARTESIA and NOAH-AFNET 6, showing that this is not an easy strategy for elderly patients.

In an NEJM editorial accompanying publication of the ARTESIA trial, Emma Svennberg, MD, Karolinska Institute, Stockholm, Sweden, also concludes that: "Going forward, we must balance the increased bleeding risks with the risk for disabling strokes," and that, "Future sub-studies and meta-analyses may provide further insights regarding treatment benefits in specific subgroups."

NOAH-AFNET 6: New Subgroup Analysis

The previously reported NOAH-AFNET 6 study randomly assigned 2538 patients with subclinical AF and additional risk factors for stroke to anticoagulation with edoxaban or placebo. The trial was stopped early, so it was underpowered — but it found no difference between groups in the incidence of the composite endpoint of stroke, systemic embolism, or death from cardiovascular causes or in the incidence of stroke, although there was higher risk of major bleeding.

Again, there was a low rate of stroke in this trial with just 49 strokes in total in the whole study. The NOAH-AFNET-6 investigators concluded that these patients should not receive anticoagulation because the risk of bleeding outweighed any potential benefits.

A new subanalysis of the 259 patients who had durations of subclinical AF of 24 hours or longer in the NOAH-AFNET 6 trial was presented at the AHA meeting, and simultaneously published online in the European Heart Journal.

This showed that the rate of stroke also appeared low in patients with these long durations of subclinical AF, and that there was no interaction between the duration of subclinical AF and the efficacy and safety of oral anticoagulation.

But with such a low number of events in the study as a whole and in the long duration subclinical AF subgroup (in which there were just two strokes in each treatment group), this analysis was unlikely to show a difference, Kirchhof commented to theheart.org | Medscape Cardiology.

The subgroup analysis did, however, show that patients experiencing subclinical AF durations of 24 hours or more were more likely to develop clinical AF over time than those with shorter durations, suggesting the need for regular ECGs in these patients.

Kirchhof said better methods are needed to detect patients with subclinical AF at high risk of stroke. "I don't think our clinical stroke risk factor scores such as CHA2DS2-VASc are sufficient to detect high-risk patients. Patients in both NOAH-AFNET 6 and ARTESIA had a median CHA2DS2-VASc score of 4, but they had a stroke rate of just 1% per year," he noted.

The meta-analysis of the two trials showed that the results from both are consistent, with an overall reduction in ischemic stroke with oral anticoagulation (relative risk [RR], 0.68). Oral anticoagulation also reduced a composite of cardiovascular death, all-cause stroke, peripheral arterial embolism, myocardial infarction, or pulmonary embolism (RR, 0.85).

There was no significant difference in cardiovascular death (RR, 0.95) or all-cause mortality (RR, 1.08), but anticoagulation significantly increased major bleeding (RR, 1.62).

Aspirin Use Complicates Results

Healey said further analyses of the ARTESIA data will try to tease out the effect of concomitant aspirin use in the trial.

He explained that patients in this trial were allowed to take a single antiplatelet agent on top of study therapy.

"It is difficult to work out the exact use of antiplatelet therapy as it changed throughout the study," he said. "About two thirds were taking antiplatelet agents at the time of enrollment into the trial, but this decreased throughout the study. Many clinicians stopped open label antiplatelet therapy during the trial when new evidence came out to suggest that there was no added benefit of adding aspirin on top of anticoagulants.

"We need to look carefully as to what impact that may have had," Healey added. "We know from other studies that adding an antiplatelet on top of an anticoagulant doesn't do much to thromboembolic events, but it approximately doubles the risk of major bleeding."

In contrast, the NOAH-AFNET trial did not allow aspirin use in the anticoagulation group and aspirin was taken by around half the patients in the placebo group who had an indication for its use.

The authors of the meta-analysis point out that the omission of aspirin in nearly half of the control patients in NOAH-AFNET 6 and the early termination of the trial may have led to a slightly higher estimate for excess major bleeding with anticoagulation.

The ARTESIA study was supported by the Canadian Institutes for Health Research, the Bristol Meyers Squibb-Pfizer Alliance, the Heart and Stroke Foundation of Canada, the Canadian Stroke Prevention Intervention Network, Hamilton Health Sciences, the Advancing Clinical Trials Network and the Population Health Research Institute. Healey reported research grants and speaking fees from BMS/Pfizer Alliance, Servier, Novartis, Boston Scientific, Medtronic; and acts as a consultant to Bayer, Servier and Boston Scientific. The NOAH-AFNET 6 trial was an investigator-initiated trial funded by the German Center for Cardiovascular Research and Daiichi Sankyo Europe. Kirchhof reported research support from several drug and device companies active in AF. He is also listed as an inventor on two patents held by the University of Hamburg on AF therapy and AF markers.

American Heart Association (AHA) Scientific Sessions 2023. Presented November 12.

N Eng J Med. Published online November 12, 2023. Abstract, Editorial

Circulation. Published online November 12, 2023. Abstract

Eur Heart J. 2023;ehad771. Abstract