Pivotal Trials Testing the Efficacy of Novel Anticoagulants for Preventing Emerging Cardioembolic Risk Entities

Atrial fibrillation (AF) increases the risk of ischemic stroke, accounts for about one quarter of all ischemic strokes worldwide, and is associated with a substantial morbidity and mortality. ¹ Oral anticoagulation can reduce the stroke risk in nonvalvular AF by two thirds by using vitamin K antagonists or non‐vitamin K‐dependent oral anticoagulants (NOACs), which are now preferentially used, as NOACs were superior to warfarin in reducing the risk of intracranial bleeding. ¹ Whether NOAC use is beneficial in stroke prevention in other patient populations is discussed in this viewpoint, based on ongoing or published randomized controlled trials (RCTs).

Anticoagulation in Patients With Newly Detected AF After Stroke

Patients included in randomized anticoagulation trials mostly had persistent/permanent AF or at least 2 documented episodes of paroxysmal AF. Notably, patients with (paroxysmal) AF detected after stroke (AFDAS) have not been studied in such trials ² and may have a lower stroke risk compared with clinical AF, depending on the mode of detection ³ and AFDAS burden (number of episodes, duration of the longest episode, time spent in AF during a defined period). Although numerous randomized trials have shown that screening for AF in patients with stroke by different methods increases the rate of AFDAS and subsequent anticoagulation rate, there is no proven prevention of recurrent ischemic stroke by AF screening. ⁴ , ⁵ To date, patients with AFDAS are anticoagulated in daily practice, but RCT‐based evidence is lacking ⁴ and this patient subgroup is not discussed in detail in recent AF guidelines. ¹ , ⁶

Randomizing patients with AFDAS to oral anticoagulation versus antiplatelet therapy would provide profound evidence. However, it is deemed impossible to convince treating physicians, as oral anticoagulation is an effective treatment in clinical AF. Another possibility is to comprehensively screen for AFDAS in only 1 arm of a randomized trial. This approach is applied in the FIND‐AF 2 (Finding Atrial Fibrillation 2) study. This investigator‐initiated study randomizes 5200 German patients (aged ≥60 years with recent ischemic stroke but without known AF) to prolonged and intensified ECG monitoring or usual care diagnostics. The primary end point is recurrent ischemic stroke or systemic embolism. The trial has nearly finished recruitment and will report results in 2026. ⁷

Anticoagulation in Patients With Atrial High‐Rate Episodes/Subclinical AF

Implantable devices (pacemakers, defibrillator, cardiac monitors) allow continuous ECG monitoring. They ensure the detection of even short episodes of cardiac arrhythmias or pauses. Atrial high‐rate episodes (AHRE) are defined as any atrial tachycardia with a frequency of 175 bpm or higher in patients in whom clinical AF is not previously detected. There is a large overlap with the term subclinical AF, which includes AHRE confirmed to be AF, atrial flutter, or AF episodes detected by insertable cardiac monitors or wearable monitors and confirmed by visually reviewed intracardiac electrograms or ECG‐recorded rhythm. ¹ As patients with subclinical AF had a 2.5‐fold increased risk of ischemic stroke or systemic embolism in the ASSERT (Asymptomatic Atrial Fibrillation and Stroke Evaluation in Pacemaker Patients and the Atrial Fibrillation Reduction Atrial Pacing Trial) study, present European Society of Cardiology guidelines recommend that “the use of oral anticoagulation may be considered in selected patients with AHRE/subclinical AF (lasting ≥24 hours) and an estimated high individual risk of stroke.” ¹ Notably, the relationship of AHRE and stroke is temporally rather low. ⁸ , ⁹

Two recent RCTs investigated the use of NOACs in AHRE/subclinical AF. First, the double‐blind placebo‐controlled NOAH‐AFNET 6 (Non–Vitamin K Antagonist Oral Anticoagulants in Patients With Atrial High Rate Episodes) trial investigated anticoagulation with edoxaban versus placebo (or aspirin if indicated) in 2536 patients aged ≥65 years with AHRE detected by an implanted device (Table). ¹⁰ Notably, the trial was halted early for safety reasons, as the rate of major bleeding was increased in the edoxaban arm (2.1% versus 1.0% per patient‐year). The primary end point—a composite of cardiovascular death, stroke, or systemic embolism (including myocardial infarction and pulmonary embolism)—was not different (3.2% using edoxaban versus 4.0% per patient‐year), with no difference regarding stroke (1.1% using edoxaban versus 0.9% per patient‐year) in a population with 10% prior stroke/transient ischemic attack at baseline (Table). In addition to premature termination compromising power, the primary end point of NOAH‐AFNET 6 includes components that are less responsive to oral anticoagulation treatment (eg myocardial infarction or cardiovascular death), driving the study result toward equipoise. A published subanalysis of NOAH‐AFNET 6 revealed that AHRE lasting ≥24 hours did not interact with the efficacy or safety of oral anticoagulation, ¹¹ but this subgroup of 259 patients was small and there were only a few events during follow‐up. Second, the double‐blind placebo‐controlled ARTESIA (Apixaban for the Reduction of Thrombo‐Embolism in Patients With Device‐Detected Sub‐Clinical Atrial Fibrillation) trial randomized 4012 patients (aged ≥55 years, CHA₂DS₂‐VASc score≥3) with device‐detected subclinical AF to apixaban or aspirin. The trial was terminated early because of problems in resupply of medication, slow recruitment, and an event rate lower than expected. After a mean follow‐up of 3.5 years, the rate of stroke and systemic embolism was lower with apixaban versus aspirin (0.78% versus 1.24% per patient‐year) in a population with 9% prior stroke/transient ischemic attack at randomization. Although there was a relative risk reduction of 37% regarding the primary end point in the intention‐to‐treat population (Table), the relative risk of major bleeding was increased by 80% in the apixaban arm (1.7% versus 0.94% per patient‐year) in the on‐treatment population. ¹²

Table .

Summary of Recently Published Randomized Trials Testing the Efficacy of Novel Anticoagulants for Preventing Emerging Cardioembolic Risk Entities

	ARTESIA	NOAH‐AFNET 6	ATTICUS	ARCADIA
Patient population	Implanted cardiac devices	Implanted cardiac devices	Patients with stroke with increased risk to develop AF	Patients with stroke with left atrial cardiomyopathy
Design	1:1 randomized double‐blind	1:1 randomized double‐blind	1:1 randomized blinded end point assessment	1:1 randomized double‐blind
Study intervention	Apixaban vs aspirin	Edoxaban vs placebo (or aspirin if indicated)	Apixaban vs aspirin	Apixaban vs aspirin
Number of patients	4012	2536	352	1015
Study termination	Early for slow recruitment, low event rate	Early for harm	Early for futility	Early for futility
Age, y, mean±SD	77±8	78±7	68±10	68±11
Female sex	36%	37%	49%	54%
CHA2DS2‐VASc score (points)	3.9±1.1	4 (IQR 3–5)	5 (IQR 4–5.5)	4.7±1.3
Prior stroke/transient ischemic attack	9%	10%	100%	100%
Follow‐up duration	3.5 y (mean)	21 mo (median)	12 mo (mean)	1.8 y (mean)
End points
Primary end point	Stroke or systemic embolism	Cardiovascular death, stroke, systemic embolism	Any new ischemic lesion on brain magnetic resonance imaging	Recurrent stroke
Event rate for the primary end point (% per year)	0.8% apixaban, 1.2% aspirin	3.2% edoxaban, 4.0% placebo/aspirin	13.6% apixaban, 16.0% aspirin	4.4% apixaban, 4.4% aspirin
HR or OR primary end point, P value	HR, 0.63, P=0.007	HR, 0.81, P=0.15	OR, 0.79, P=0.57	HR, 1.00, no P value provided
Ischemic stroke (n, % per patient year)	116, 0.8%	49, 1.0%	23, 5.8%	80, 4.4%
Major bleeding (n, % per patient year)	184, 1.3%	78, 1.5%	2, 0.6%	10, 0.8%
Major bleeding by study arm (% per patient year)	1.5% apixaban, 1.1% aspirin	2.1% edoxaban, 1.0% placebo/aspirin	0.6% apixaban, 0.6% aspirin	0.7% apixaban, 0.8% aspirin (excluding intracranial hemorrhage)
HR or OR for major bleeding	HR, 1.36, P=0.04	HR, 2.10, P=0.002	OR, 0.95, no P value reported	HR, 1.02, no P value reported
New atrial fibrillation (n, %)	Not applicable	Not applicable	89, 25.3% in 12 mo	149, 14.7% in 1.8 y

ARCADIA indicates Atrial Cardiopathy and Antithrombotic Drugs in Prevention After Cryptogenic Stroke; ARTESIA, Apixaban for the Reduction of Thrombo‐Embolism in Patients With Device‐Detected Sub‐Clinical Atrial Fibrillation; ATTICUS, Apixaban for Treatment of Embolic Stroke of Undetermined Source; HR, hazard ratio; NOAH‐AFNET 6, Non–Vitamin K Antagonist Oral Anticoagulants in Patients With Atrial High Rate Episodes; and OR, odds ratio.

At first glance, results seem different, but they share consistent findings, as demonstrated in a meta‐analysis including both RCTs. ¹³ A common finding in NOAH‐AFNET 6 and ARTESIA was a low rate for ischemic strokes (of about 1% per year) in nonanticoagulated patients with subclinical AF, which is about 4% per year in patients with clinical AF and a similar CHA₂DS₂‐VASc score of 4. In both trials, there were numerically lower rates of stroke/systemic embolism with oral anticoagulation compared with the nonanticoagulation regimen. The combined analysis showed a significant 32% risk reduction for ischemic stroke and a 62% increase of major bleeding by using a NOAC. Notably, aspirin use in all study patients in ARTESIA, of which (only) 61% had an antiplatelet before randomization, may have affected trial results.

Anticoagulation in Embolic Stroke of Undetermined Source

Ischemic stroke is a heterogeneous disease regarding cause, which remains unknown (so called “cryptogenic”) in about 20% to 30% of all patients after (in‐hospital) diagnostic workup. Ten years ago, the clinical construct of embolic stroke of undetermined source (ESUS) was introduced. ¹⁴ It aimed to define a subset of patients with cryptogenic stroke who might benefit from oral anticoagulation instead of using aspirin. ESUS diagnostic criteria require patients with nonlacunar brain infarcts without proximal arterial luminal stenosis (of ≥50%) or apparent cardioembolic source according to transthoracic or transesophageal echocardiography and at least 24 hours of ECG monitoring to exclude AF exceeding ≥6 minutes duration. Two large RCTs were not able to show a reduction in ischemic or hemorrhagic stroke or systemic embolism using rivaroxaban versus aspirin (NAVIGATE‐ESUS [Rivaroxaban Versus Aspirin in Secondary Prevention of Stroke and Prevention of Systemic Embolism in Patients With Recent Embolic Stroke of Undetermined Source]) ¹⁵ or ischemic stroke using dabigatran versus aspirin (RESPECT‐ESUS [Randomized, Double‐Blind, Evaluation in Secondary Stroke Prevention Comparing the Efficacy and Safety of the Oral Thrombin Inhibitor Dabigatran Etexilate Versus Acetylsalicylic Acid in Patients With Embolic Stroke of Undetermined Source]). ¹⁶ Furthermore, major bleeding rate was elevated in patients receiving anticoagulants in NAVIGATE‐ESUS. The failure of oral anticoagulation showing efficacy in patients with ESUS is likely explained by the unspecific definition chosen, including a too heterogeneous mixture of potential stroke causes besides (undetected) AF, like patent foramen ovale, valvular heart disease, nonstenotic atherosclerotic plaques, or undiagnosed cancer.

Recently, results from the third ESUS trial were published. ¹⁷ The ATTICUS (Apixaban for Treatment of Embolic Stroke of Undetermined Source) trial is different in design, end point, and sample size to the 2 former studies (Table). It included patients with ESUS with at least 1 predictive factor for AF (eg, CHA₂DS₂‐VASc score≥4, left atrium size >45 mm, spontaneous echo contrast or flow velocity of ≤0.2 m/s in the left atrial appendage, atrial runs of <30 seconds) or a patent foramen ovale. The trial was halted for futility after a planned interim analysis, resulting in 352 (of the targeted 500) patients randomized 1:1 to standard dose apixaban or aspirin 100 mg OD. Nearly all patients received an implantable loop recorder. If any episode of clinical AF or subclinical AF lasting >2 minutes was detected, patients on aspirin were switched to apixaban. The primary end point—at least 1 new magnetic resonance imaging‐detected ischemic brain lesion—was not different (13.6% using apixaban versus 16.0% using aspirin) during 1 year of follow‐up. Recurrent ischemic stroke (5.7% versus 5.9%) or major bleeding (0.6% versus 0.6%) rates were similar. Notably, 26% of all study patients and >40% of those aged ≥75 years were switched to apixaban following AF detection, minimizing power to show a difference between study arms. As similarly shown in RESPECT‐ESUS, ¹⁶ there was a trend toward a benefit of apixaban in patients aged ≥75 years, ¹⁷ which needs confirmation in a future RCT. Taken together, aspirin but not oral anticoagulation is indicated for long‐term stroke prevention in patients with ESUS.

Anticoagulation in ESUS Patients With Left Atrial Cardiopathy

It has been speculated that patients with ESUS with a diseased left atrium might benefit from oral anticoagulation, based on post hoc analyses from the WARSS (Warfarin‐Aspirin Recurrent Stroke Study) trial (reduced risk of ischemic stroke or death with warfarin versus aspirin in patients with cryptogenic stroke with NT‐proBNP [N‐terminal pro‐B‐type natriuretic peptide]‐plasma levels >750 pg/m) ¹⁸ and the NAVIGATE‐ESUS trial (reduced risk of ischemic stroke or systemic embolism in patients with left atrial diameter>4.6 cm receiving anticoagulants). ¹⁹ Very recently, the double‐blind ARCADIA (Atrial Cardiopathy and Antithrombotic Drugs in Prevention After Cryptogenic Stroke) trial randomized 1015 (of the targeted 1100) patients with ESUS and signs of left atrial cardiopathy (NT‐proBNP plasma level ≥250 pg/mL or P‐wave terminal force >5000 μV×ms in V1 or left atrial diameter index ≥3 cm/m²) 1:1 to standard dose apixaban or aspirin 81 mg OD (Table). The trial was halted early for futility after a preplanned interim analysis. The primary end point recurrent ischemic stroke was similar in both arms (4.4% per patient‐year) and the rate of symptomatic intracranial (0.0% per patient‐year versus 1.1% using aspirin) or major hemorrhage excluding intracranial hemorrhage (0.7% per patient‐year versus 0.8% using aspirin) was not different. ²⁰ The trial results indicate that anticoagulation in patients with ESUS with evidence of left atrial cardiopathy (but without AF) is not better but also no worse than aspirin. The most likely explanations for this result are that either the chosen criteria for left atrial cardiopathy were not specific enough to identify high‐risk patients or that oral anticoagulation for left atrial cardiopathy is just not efficacious. Importantly, the median delay of randomization after stroke of 50 days excluded the period with the highest stroke recurrence risk (first 30 days across different stroke causes). However, data from NAVIGATE‐ESUS and RESPECT‐ESUS did not show differences in stroke recurrences between direct anticoagulation and aspirin, even in the first 30 days after stroke. ¹⁶ , ¹⁷ Secondary analyses of ARCADIA are pending at present, but we doubt that they will change the present view substantially.

Further Methodological Considerations

Some methodology aspects of the recent RCTs discussed have to be pointed out. First, all trials were stopped prematurely limiting power (Table). Second, the follow‐up of ATTICUS and ARCADIA was maybe too short to uncover a benefit of oral anticoagulation over aspirin. Such a trend was seen in the RESPECT‐ESUS trial, which demonstrated a benefit of dabigatran over aspirin in a post hoc landmark analysis starting 1 year after stroke ¹⁶ Third, it is obvious that recurrent stroke rates in ATTICUS and ARCADIA were much higher than the reported stroke rate in ARTESIA and NOAH‐AFNET 6 (Table), including patients with a rather low rate of prior stroke/transient ischemic attack. Whether secondary stroke prevention in patients with AHRE is more effective using oral anticoagulation remains to be established.

Summary

Beyond stroke prevention in AF, oral anticoagulation may conceptually be efficacious to reduce the risk of ischemic stroke in other cardioembolic risk entities. On the basis of randomized trials, AHRE/subclinical AF is the only entity at present in which oral anticoagulation may be considered. As the risk of ischemic stroke in the majority of patients with subclinical AF is low and anticoagulation is associated with increased bleeding risk, shared decision‐making is needed in clinical practice also reflecting the patient's individual stroke and bleeding risk.

Sources of Funding

None.

Disclosures

Rolf Wachter reports speaker's honoraria or consulting fees from AstraZeneca, Bayer, BMS, Boehringer Ingelheim, CVRx, Daiichi Sankyo, Medtronic, Novartis, Pfizer, Pharmacosmos, Sciarc, Servier, and Vifor, His research is supported by grants from Bundesministerium für Bildung und Forschung, Deutsche Forschungsgemeinschaft, European Union, Deutsches Zentrum für Herz‐/Kreislaufforschung, and Medtronic. Karl Georg Haeusler reports speaker's honoraria, consulting fees, lecture honoraria, /or study grants from Abbott, Alexion, Amarin, AstraZeneca, Bayer Healthcare, Biotronik, Boehringer Ingelheim, Boston Scientific, Bristol‐Myers Squibb, Daiichi Sankyo, Edwards Lifesciences, Medronic, Novartis, Pfizer, Portola, Premier Research, Sanofi, SUN Pharma, and W. L. Gore and Associates.