Subclinical atrial fibrillation: Implications of recent trials for guideline updates?

Abstract

With the widespread use of cardiac implantable electronic devices and smartwatches, device-detected atrial fibrillation (AF) also referred to as subclinical AF (SCAF) is becoming increasingly common. The incidence of device-detected AF varies between 30% and 60%, depending on the definition and the device used for detection. Multiple studies, such as EMBRACE, CRYSTAL-AF, and FIND-AF, have confirmed higher detection rates of SCAF following prolonged rhythm monitoring using implantable loop recorders or external loop recorders in patients with cryptogenic stroke. The stroke risk associated with SCAF primarily depends on two factors: The baseline CHA₂DS₂-VASc score and the duration of SCAF episodes. Very-short episodes (< 6 minutes) are likely of uncertain significance, whereas episodes lasting > 24 hours increased the risk of stroke/systemic embolism (SE) more than threefold in the ASSERT study. For episodes lasting between 6 minutes and 24 hours, the stroke risk is lower but varies with the baseline CHA₂DS₂-VASc score. Previous randomized trials of direct oral anticoagulants (DOACs) in patients with cryptogenic stroke-NAVIGATE-ESUS (with rivaroxaban) and RE-SPECT ESUS (using dabigatran)-failed to demonstrate superiority over aspirin. More recently, two dedicated studies in SCAF with DOACs have been published: NOAH-AFNET 6 (with edoxaban) and ARTESIA (with apixaban). NOAH-AFNET 6 was terminated early for futility due to slow enrollment and lower-than-expected event rates. In contrast, apixaban reduced the risk of stroke and SE by 37% in the ARTESIA study, albeit with increased bleeding. These differing results may be attributed to differences in the DOAC used, trial design, and enrolled patient populations. Current ACC/AHA guidelines recommend oral anticoagulation (OAC) for SCAF episodes lasting > 24 hours and a baseline CHA₂DS₂-VASc score > 2. For those with episodes lasting between 6 minutes and 24 hours, a higher CHA₂DS₂-VASc score > 3 points towards a benefit of OAC, while a conservative approach-including control of risk factors (e.g., hypertension, thyroid dysfunction, alcohol intake) and periodic follow-up is warranted for the rest. However, considering the positive ARTESIA results, a reevaluation may be needed. Patients with high CHA₂DS₂-VASc score (> 4) and SCAF > 24 hours duration may be ideal candidates for DOAC therapy. Those with prior stroke and vascular disease also have a higher stroke risk in future and may be attractive candidates for OAC too. For those with high bleeding risk, re-evaluation after optimizing modifiable bleeding risk factors (e.g., concomitant medications, blood pressure control) may help determine eligibility for anticoagulation. Ongoing large-scale DOAC trials will further clarify this contentious issue.

Key Words: Atrial high-rate episode; NOAH-AFNET 6; ARTESIA; Apixaban; Bleeding; CHA₂DS₂-VASc score

Core Tip: The risk of stroke in people with subclinical atrial fibrillation (SCAF) depends on: The baseline CHA₂DS₂-VASc score and the duration of the episodes. Short episodes-those lasting less than six minutes-are generally of uncertain significance. However, when the episodes last longer than 24 hours, the risk of stroke or systemic embolism increases more than threefold. The utility of oral anticoagulation, for preventing stroke in those with SCAF is still under debate. According to current ACC/AHA guidelines, oral anticoagulation is recommended for people who have SCAF episodes lasting more than 24 hours and have a high CHA₂DS₂-VASc score. For those with episodes between six minutes and 24 hours, a more cautious approach is advised-this includes managing risk factors along with regular follow-up. However, the positive results from the ARTESIA study suggest these recommendations may need to be reconsidered.

INTRODUCTION

The widespread use of wearable monitors and cardiac implantable electronic devices (CIEDs) has led to the identification of subclinical atrial fibrillation (AF) in a significant proportion of individuals. Emerging evidence now links these asymptomatic arrhythmias to an increased risk of incident stroke. Initiating anticoagulation therapy promptly after detecting subclinical AF (SCAF) may offer an opportunistic window to reduce stroke risk. However, these potential benefits must be weighed against the small but notable risk of bleeding associated with oral anticoagulation (OAC). Consequently, it remains unclear in which populations long-term anticoagulation is warranted. This review evaluates current data on the prevalence, clinical significance, and management of SCAF, while identifying the knowledge gaps and areas of ongoing debate.

SCAF: A primer

AF is one of the most common arrhythmias encountered in clinical practice and carries a significant risk of stroke or systemic embolism (SE)[1,2]. Traditionally, managing clinically apparent AF for primary stroke prevention has been straightforward: Once AF is documented and stroke risk factors are present, OAC is recommended regardless of symptoms or AF classification (paroxysmal, persistent, or permanent). However, the increasing detection of AF via wearable and implantable cardiac devices challenges this approach. These technologies can detect transient and asymptomatic atrial arrhythmias, which are much more prevalent than clinically manifest episodes of AF. The clinical significance of these subclinical episodes remains unclear, and there is no established evidence-based therapeutic strategy[3,4]. Implanted, and wearable devices allow for the detection these of atrial high-rate episodes (AHREs) or SCAF. In contrast, the external electrocardiography (ECG) monitoring has limited sensitivity due to shorter recording durations. When AHRE or SCAF is detected using implanted or wearable devices, it is crucial to review stored electrograms or ECG rhythm strips to exclude artefacts or other causes of false-positive detection[4,5].

SCAF: Definition and incidence

Asymptomatic AF episodes detected by CIEDs are referred to as SCAF. The terms SCAF, AHREs, and device-detected AF are often used interchangeably. Recent research indicates that among individuals with stroke risk factors but no history of AF, SCAF is detected in approximately 30%-40% over a 3-year follow-up period using implanted cardiac monitors[5]. SCAF refers specifically to asymptomatic AF identified by cardiac devices. These episodes are generally short-lived. However, the term “AHRE” may be more accurate, but henceforth, the term “SCAF” will be utilized throughout the manuscript for clarity.

SCAF: Risk of stroke

The threshold duration of arrhythmic episodes that confers stroke risk in SCAF varies across studies. In the MOST study, AHREs lasting > 5 minutes were associated with increased stroke risk. In the ASSERT study, the threshold was > 6 minutes; in SOS-AF, it was 1 hour; and in the TRENDS and IMPACT studies, a total arrhythmic burden exceeding 5.5 hours was significant[5-12]. Conversely, in the Finnish Cardio Version Trial, stroke risk was higher in patients with episodes exceeding 12 or 24 hours[13]. The relationship between a SCAF episode duration and stroke risk has been extensively studied. In the ASSERT trial, the risk of stroke was lowest (< 1.68% per patient-year) in episodes lasting < 6 minutes, intermediate (approximately 1.7%-2.5% per patient-year) for episodes of 6 minutes-24 hours, and highest (> 4% per patient-year) for episodes > 24 hours[5,6].

Beyond these thresholds, recent studies have introduced the concepts of ‘micro-AF’ and ‘ultra-short AF’, referring to atrial arrhythmias of < 30 seconds’ duration[14,15]. Although these very brief episodes do not meet diagnostic criteria for AF, they are increasingly recognised as risk markers for future AF development and possibly thromboembolic events. Prospective studies have demonstrated that micro-AF predicts incident AF, and reports on ultra-short AF suggest potential prognostic significance, highlighting the need for improved screening strategies[16,17].

In the ASSERT study, 2580 adults aged > 65 years with hypertension, no prior AF, and recently implanted CIEDs (St. Jude Medical, MN, United States) for sinus or AV node dysfunction were included[5]. Patients requiring OAC for other reasons or with a prior history of AF/flutter > 5 minutes were excluded. After a 3-month observation period, patients were divided into two groups: Those who had SCAF (AHRE, defined as an atrial rate ≥ 190 beats per minute for > 6 minutes) and those who did not. The primary outcome, either a stroke or a SE, was monitored over a mean of 30 months, with follow-up conducted every 6 months. During the observation period, SCAF was detected in 10.1% of patients. Those with SCAF had a five-fold higher risk of developing clinical AF [hazard ratio (HR) 5.56; 95% confidence interval (CI): 3.78-8.17; P < 0.001] and a 2.5-fold higher risk of stroke or SE (HR 2.49; 95%CI: 1.28-4.85; P = 0.007; Figure 1). Even after adjusting for stroke risk factors, SCAF remained predictive of embolic events. Although the stroke risk is not as high as with clinical AF (4-5 times baseline risk), it is still two to three times higher than in the general population. Notably, atrial overdrive pacing did not prevent the future development of AF. A major limitation of ASSERT is that SCAF detection was based on only a 3-month monitoring window, during which device lead implantation itself may have provoked arrhythmias. A follow-up analysis examined the impact of episode duration on future risk[6]. At 2.5 years, SCAF episode durations were distributed as follows: > 24 hours (10.7%), 6-24 hours (6.9%), and 6 minutes to 6 hours (18.8%; Figure 1). Only episodes > 24 hours significantly predicted stroke/SE (HR 3.24; P = 0.003). Shorter episodes were not statistically significant after adjustment (HR 0.75; P = 0.562, and HR 1.32; P = 0.646, respectively). Thus, patients with episodes lasting 6 minutes -24 hours represent a gray area in clinical management. To address this uncertainty, two major trials were launched: NOAH-AFNET 6 and ARTESIA.

Figure 1 The risk of future embolic events in patients with subclinical atrial fibrillation with a recent pacemaker implantation from the ASSERT study. The upper panel depicts data from the original ASSERT study, where 10% patients developed Atrial high-rate episode (AHRE) within 3 months of a pacemaker implantation. The presence of AHRE was predictive of future risk of atrial fibrillation (AF) & ischemic events at 2.5-year follow-up. The lower panel stratifies patients based on subclinical AF duration and risk of future AF & ischemic events. AHRE: Atrial high-rate episode; SE: Systemic embolism; SCAF: Subclinical atrial fibrillation; FU: Follow up; HR: Hazard ratio.

Importantly, the STROKESTOP trial provided randomized evidence that systematic screening for AF in an elderly population (aged 75-76 years) with intermittent ECG monitoring was feasible and safe, and although the reduction in stroke or SE did not reach conventional statistical significance, the intervention group demonstrated a favorable trend towards improved clinical outcomes, including fewer strokes and lower mortality[18]. Together, these findings suggest that AF risk is not only dependent on episode duration but also on early detection of even very short atrial arrhythmias, and that population-level screening strategies may help refine preventive anticoagulation approaches.

Anticoagulation for cryptogenic stroke: Lessons learned

Patients with ischemic stroke without any underlying cause are classified as either cryptogenic stroke or embolic stroke of unknown source (ESUS) and cause up to 1/4^th of all ischemic strokes[19]. Detecting covert AF is crucial for two reasons: Its presence increases stroke recurrence risk, and its management relies more on anticoagulation than on antiplatelet therapy. The EMBRACE and CRYSTAL-AF trials demonstrated that prolonged monitoring with implanted devices in cryptogenic stroke patients often reveals asymptomatic SCAF[20,21]. In the EMBRACE trial, eligible participants were ≥ 55 years, had experienced ESUS in the previous 6 months, and had no identifiable cause of stroke after comprehensive evaluation[20]. They were randomized to either 24-hour Holter monitoring or 30-day ambulatory ECG using an event-triggered loop recorder. AF lasting ≥ 30 seconds was detected in 16.1% of the 30-day group compared with 3.2% in the Holter group (absolute difference 12.9%; 95%CI: 8.0-17.6; P < 0.001; number needed to screen, 8). In CRYSTAL-AF, patients aged ≥ 40 years who experienced ESUS within the past 3 months and had no known cause after full work-up were randomized to standard follow-up or implantable loop recorders (ILRs) (REVEAL XT)[21]. At 12 months, AF was detected in 12.4% of the ILR group vs 2.0% in the control group (HR 7.3; 95%CI: 2.6-20.8; P < 0.001). Similarly, in the retrospective CRYPTON-ICM registry from Japan, AF was detected in 33% of cryptogenic stroke patients monitored with ILRs[22]. Of these, 52% had a high AF burden (> 0.1%).

Two randomized trials assessed empiric anticoagulation for all patients with cryptogenic stroke: Randomised, Double-Blind, Evaluation in Secondary Stroke Prevention Comparing the Efficacy and Safety of the Oral Thrombin Inhibitor Dabigatran Etexilate vs Acetylsalicylic Acid in Patients with Embolic Stroke of Undetermined Source (RESPECT-ESUS) and New Approach Rivaroxaban Inhibition of Factor Xa in a Global Trial vs ASA to Prevent Embolism in Embolic Stroke of Undetermined Source (NAVIGATE-ESUS; Figure 2)[23,24].

Figure 2 The four pivotal direct oral anticoagulant trials in patients with cryptogenic stroke & subclinical atrial fibrillation. The NAVIGATE ESUS & RE-SPECT embolic stroke of unknown source evaluated the role of direct oral anticoagulant therapy in cryptogenic stroke, while NOAH-AFNET-6 & ARTESIA enrolled patients with subclinical atrial fibrillation. ESUS: Embolic stroke of unknown source.

NAVIGATE-ESUS

As previously stated, up to 25% of ischemic strokes are ESUS, with a high recurrence rate. Rivaroxaban, an oral factor Xa inhibitor, was hypothesized to reduce the risk of recurrent stroke when used in place of aspirin. The NAVIGATE-ESUS trial evaluated the safety and efficacy of 15 mg rivaroxaban once daily vs 100 mg aspirin once daily in patients who had recently experienced an ischemic stroke presumed to be embolic in nature, but without identifiable arterial stenosis, lacunar infarction, or a known cardioembolic source (Figure 2)[23]. The primary efficacy outcome was the time to first recurrence of ischemic or hemorrhagic stroke or SE. The primary safety outcome was the rate of major bleeding. A total of 7213 participants were enrolled across 459 sites-3609 were randomized to aspirin and 3604 to rivaroxaban. The trial was terminated early due to increased bleeding associated with rivaroxaban and lack of efficacy in reducing stroke recurrence. Over a median follow-up of 11 months, 172 patients in the rivaroxaban group (annualized rate: 5.1%) and 160 in the aspirin group (annualized rate: 4.8%) experienced the primary efficacy outcome. Major bleeding occurred in 62 patients (1.8% per year) in the rivaroxaban group and 23 patients (0.7% per year) in the aspirin group. Thus, rivaroxaban did not outperform aspirin in preventing recurrent stroke in patients with ESUS and was associated with a higher risk of bleeding.

RESPECT-ESUS

The RESPECT-ESUS trial was a multicenter, randomized, double-blind study evaluating the efficacy of dabigatran vs aspirin in patients who had experienced an ESUS. Participants received either 150 mg or 110 mg of dabigatran twice daily (depending on age and renal function) or 100 mg aspirin once daily. The primary efficacy outcome was recurrent stroke[24]. The primary safety outcome was major bleeding. A total of 5390 patients were randomized equally: 2695 to dabigatran and 2695 to aspirin. Over a median follow-up of 19 months, recurrent strokes occurred in 177 patients (6.6%) in the dabigatran group (4.1% per year) and in 207 patients (7.7%) in the aspirin group (4.8% per year). Major bleeding occurred in 77 patients (1.7% per year) in the dabigatran group and 64 patients (1.4% per year) in the aspirin group. Dabigatran did not significantly reduce the risk of recurrent stroke compared to aspirin. Although there were more clinically relevant non-major bleeding events in the dabigatran group, the rate of major bleeding was not significantly higher.

In the CRYPTON-ICM registry from Japan, OAC was initiated in 91% of cryptogenic stroke patients after AF detection[22]. The rate of ischemic stroke was not significantly different between those in whom AF was detected and treated using anticoagulants (5.8%) and in those whom AF was not detected (4%; P = 0.45). The timing of AF detection (early vs late) also did not affect stroke risk, suggesting that AF detection alone, regardless of timing, may not independently raise stroke risk in this population.

A NAVIGATE-ESUS sub study explored whether rivaroxaban offered benefit in ESUS patients with high risk for underlying AF[25]. Predictors of future AF included left atrial (LA) enlargement, high HAVOC scores, and frequent premature atrial contractions. In patients with LA diameter > 4.6 cm, rivaroxaban significantly reduced stroke risk compared with aspirin (1.7% vs 6.5%; HR 0.26; P for interaction = 0.02). While the primary NAVIGATE-ESUS trial was negative, these subgroup findings emphasize the importance of identifying high-risk ESUS patients who may benefit from anticoagulation.

The ARCADIA trial is currently investigating whether apixaban is superior to aspirin for stroke prevention in patients with ESUS and atrial cardiomyopathy-defined using ECG, biomarker, and echocardiographic features predictive of AF risk[26]. The ATTICUS study is similarly evaluating the role of apixaban vs aspirin in cryptogenic stroke or ESUS patients with high-risk features for embolism, including enlarged LA, spontaneous echo contrast, and low LA appendage velocity[27].

Anticoagulation in SCAF

Patients with device-detected AF episodes lasting between 6 minutes and 24 hours remain a particularly challenging group with respect to anticoagulation decisions. The uncertainty surrounding stroke risk in this range- especially in patients with longer episodes-prompted the design of two landmark trials: NOAH-AFNET 6 and ARTESIA.

ARTESIA

There is a 2.5-fold increase in stroke risk associated with SCAF. It was soon recognized that even brief, asymptomatic AF episodes were common in patients without a prior clinical diagnosis of AF[28]. Individuals with SCAF had an absolute increase in stroke risk of approximately 1% per year-roughly half the risk observed in patients clinically diagnosed with AF[29]. Building on this premise, the ARTESIA trial enrolled patients aged > 55 years with at least one episode of device-detected SCAF lasting > 6 minutes but no episode exceeding 24 hours, as detected by an implanted pacemaker, defibrillator, or cardiac monitor[30]. Eligible patients also had a CHA₂DS₂-VASc score ≥ 3. The key exclusion criteria included uncorrected severe bleeding in the past 6 months, ongoing need for OAC, a history of clinical AF, and a creatinine clearance < 25 mL/min. Patients were randomized in a double-blind, double-dummy design to receive either 81 mg of aspirin once daily or 5 mg of apixaban twice daily. The primary efficacy outcome-stroke or SE-was assessed in the intention-to-treat population. The primary safety outcome was major bleeding, as defined by the International Society on Thrombosis and Haemostasis (ISTH). A total of 4012 patients were randomized: 2015 to the apixaban group and 1997 to the aspirin group. Notably, nearly a quarter of patients in each group discontinued the study medication due to the recurrence of longer AF episodes (24.3% in the apixaban arm vs 23.8% in the aspirin arm). Additionally, around 34% of patients in both arms discontinued trial medication for unspecified reasons.

During follow-up, death occurred in 457 patients (22.7%) in the apixaban group and 438 patients (21.9%) in the aspirin group. The mean (± SD) follow-up duration was 3.5 ± 1.8 years. The mean patient age was 76.8 ± 7.6 years, with 36.1% of patients being women. Compared with prior AF trials, baseline demographics in ARTESIA were similar, with two notable differences: Underrepresentation of women and patients with a history of stroke. The median duration of the longest SCAF episode in the 6 months before enrollment was 1.47 hours, and the mean CHA₂DS₂-VASc score was 3.9 ± 1.1. The primary outcome occurred in 55 patients in the apixaban group and 86 in the aspirin group (HR 0.63; 95%CI: 0.45-0.88; P = 0.007). Hence, a 37% reduction in stroke and SE was seen with apixaban therapy (Figure 3). The rate of disabling or fatal stroke (Modified Rankin Scale score -3-6) was also 49% lower with apixaban. Similarly, ischemic strokes and strokes from any cause were also lower in the apixaban arm. Interestingly, hemorrhagic stroke was numerically lower with apixaban, too. However, the risk of the composite outcome (stroke, SE, or cardiovascular death) was similar between groups. The rate of major bleeding was 1.71% per patient-year with apixaban and 0.94% per patient-year with aspirin (HR 1.80; 95%CI: 1.26-2.57; P = 0.001). Thus, apixaban significantly reduced the risk of stroke/ SE but increased the risk of major bleeding compared to aspirin in patients with SCAF (Table 1 and Figure 3). Interestingly, in the control arm, all patients received aspirin therapy while only 61% were on it at baseline.

Figure 3 Comparison of results of two recently published large direct oral anticoagulant randomized controlled trials in patients with subclinical AF-NOAH-AFNET with edoxaban & ARTESIA with apixaban, respectively. A: The primary efficacy outcome; B: The primary safety endpoint. The primary efficacy endpoint in the ARTESIA study was a composite of stroke/systemic embolism (SE), while in the NOAH AFNET-6 was cardiovascular death, stroke, or SE. The y axis represents the event rate measured in person-years. Similarly, major bleeding was the primary safety endpoint in ARTESIA, while in the NOAH-AFNET study, it was a combination of all-cause death or major bleeding. DOAC: Direct oral anticoagulant.

Table 1 Comparison of key attributes of two recent randomized controlled trials on use of oral anticoagulation in subclinical atrial fibrillation.

	ARTESIA	NOAH AFNET 6
Population size	4012	2536
SCAF duration	6 minutes-24 hours	> 6 minutes
Longest duration of SCAF (median)	1.5 hours	2.8 hours
Drug used	Apixaban 5 mg BD	Edoxaban 60 mg OD
Comparator	Aspirin	Aspirin/placebo
Mean CHA2DS2VaSc score	3.9 ± 1.1	4
Mean age (years)	76 ± 7.6	78
Females (%)	36.1	37.4
Follow up (years)	3.5 ± 1.8	1.8
Primary efficacy end point	Stroke or SE (intention-to-treat analysis)	Cardiovascular death, stroke, or SE (time-to-event analysis)
Primary safety end point	Major bleeding	Death from any cause or major bleeding
Trial completion	Completed	Premature termination

SE: Atrial fibrillation; SCAF: Subclinical atrial fibrillation.

A subgroup analysis of the ARTESIA study explored the impact of baseline CHA₂DS₂-VASc scores on the primary outcome events[31]. Approximately 27% of patients enrolled in the ARTESIA trial with SCAF exhibited a CHA₂DS₂-VASc score > 4, one-third had a score of 4, and the remaining 40% had a risk score of < 4. Among patients with a CHA₂DS₂-VASc score > 4, the advantages of apixaban therapy in mitigating stroke/SE outweighed the potential risks compared with aspirin. In this subgroup, while apixaban prevented 1.3 strokes/SE per 100 patient-years, it concomitantly led to 0.68 bleeding events/100 patient-years. The corresponding HRs compared with aspirin for stroke/SE and major bleeding were 0.44 and 1.48, respectively. In contrast, in patients with CHA₂DS₂-VASc score < 4 and 4 there were 0.12 and 0.33 events per 100 patient-years, respectively. The stroke prevention benefit by apixaban was minimal, with HRs for stroke/SE of 0.66 and 0.87, respectively. Interesting, although HRs for stroke prevention reduced drastically with reduced CHA₂DS₂-VASc scores, the HRs for bleeding in both groups were not much attenuated, at 1.27 and 1.31, respectively. Conversely, patients with a CHA₂DS₂-VASc score < 4 experienced a lower risk profile, making apixaban treatment less beneficial. In patients who present with an intermediate -₂DS₂-VASc score of 4, individual preferences should guide treatment choices.

Another ARTESIA subgroup analysis stratified patients based on their history of prior stroke, with a history of prior stroke/transient ischemic attack (TIA) present in only 8.5%[32]. In these patients, apixaban provided a 7% absolute risk reduction in stroke/SE over 3.5 years, compared with just 1% in those without a history of stroke. Among prior stroke patients, the annualised stroke/SE rate with aspirin was quite high at 3.1%-well above the 1% threshold recommended in the ESC guidelines for initiating direct oral anticoagulants (DOACs)[1,33,34]. The 2023 ACC/AHA guidelines recommend OAC for patients with an annual stroke risk > 2%[2]. Using data from > 13000 patients in the ATRIA study, Eckman et al[33] developed a Markov decision model to predict the stroke rate above which anticoagulation is preferable. Warfarin was found to be suitable beyond a stroke rate of 1.7%/year, whereas DOACs were preferred at > 0.9%/year. Given that the overall stroke rate in ATRESIA was low (< 1.25%/year), demonstrating significant benefit from DOACs was inherently challenging. However, the subgroup with prior stroke represents a high-risk population where apixaban offers a clear net clinical benefit.

NOAH-AFNET 6

The NOAH-AFNET 6 study evaluated the role of edoxaban in elderly patients (> 65 years) with device-detected AF or SCAF[35]. Eligible participants were aged 65 years or older, had AHREs detected by implanted devices, and had no prior ECG-confirmed AF. AHREs had to persist for at least 2 months after device implantation, with an atrial rate > 170 beats per minute and duration ≥ 6 minutes. Additional inclusion criteria required one or more stroke risk factors: Heart failure, hypertension, diabetes mellitus, vascular disease (e.g., aortic plaque, prior myocardial infarction, or peripheral/carotid/cerebral artery disease), age ≥ 75 years, or a history of stroke or TIA. Participants were randomly assigned 1:1 to receive edoxaban (anticoagulation) or placebo in a double-blind, double-dummy design. Edoxaban, a factor Xa inhibitor approved for stroke prevention in clinical AF, was administered at 60 mg once daily. A dose reduction to 30 mg was allowed for patients meeting any of the following: Body weight ≤ 60 kg, creatinine clearance 15-50 mL/min, or use of potent P-glycoprotein inhibitors. The placebo group received either an inactive placebo or acetylsalicylic acid 100 mg daily. A time-to-event analysis assessed the primary efficacy endpoint-a composite of cardiovascular mortality, stroke, or SE. The primary safety endpoint was a composite of major bleeding or all-cause death, as defined by ISTH criteria.

Of the total 2536 patients enrolled, 1270 were assigned to edoxaban and 1266 to placebo. The median AHRE duration was 2.8 hours; mean age was 78 years; women comprised 37.4% of participants; and the mean CHA₂DS₂-VASc score was 4. The trial was terminated early after a median follow-up of 21 months due to safety concerns and an independent assessment indicating futility[36]. A primary efficacy event occurred in 83 patients in the edoxaban group (3.2% per patient-year) vs 101 in the placebo group (4.0% per patient-year; HR 0.81; 95%CI: 0.60-1.08; P = 0.15). Although the stroke incidence was low at approximately 1% per patient-year in both groups, ischemic stroke, SE and the composite of stroke/SE were numerically lower in the edoxaban arm but did not reach statistical significance. Cardiovascular death was numerically lower in the edoxaban arm too. Safety events occurred in 149 patients (5.9% per patient-year) in the edoxaban group vs 114 (4.5% per patient-year) in the placebo group (HR 1.31; 95%CI: 1.02-1.67; P = 0.03). Specifically, major bleeding events were more frequent in the edoxaban arm (2.1% vs 1%; HR -2.1; P = 0.002). Clinical AF was diagnosed via ECG in 462 patients (18.2% overall; 8.7% per patient-year). Thus, in patients with AHREs detected by implanted devices, edoxaban did not significantly reduce the incidence of cardiovascular mortality, stroke, or SE compared to placebo, but it did increase the risk of death or major bleeding (Table 1 and Figure 3). As discussed earlier, patients with longer-duration SCAF episodes may have higher stroke risk and could derive greater benefit from anticoagulation. A sub-study of NOAH-AFNET 6 examined patients with SCAF episodes > 24 hours[37]. About 11% of participants had AHREs lasting > 24 hours (median duration of the longest episode was 53 hours). Duration of AHRE did not influence the primary efficacy or safety outcomes. However, patients with episodes > 24 hours had a 2.2-fold increased risk of developing ECG-confirmed AF, with 29% eventually developing AF during follow-up. Notably, CHA₂DS₂-VASc scores and ischemic stroke rates were similar between groups. The shorter follow-up duration compared to ARTESIA may explain the lower event rates (Table 1).

Key difference between NOAH-AFNET 6 and ARTESIA trials

The stark difference between results of these two trials largely emanates from lower stroke risk in general and shorter follow up duration of NOAH-AFNET 6 trial. These factors rendered the trial underpowered to detect any positive efficacy outcomes. Whether longer follow-up would reveal a significant benefit from OAC in patients with prolonged AHREs remains speculative. Another major difference was inclusion of cardiovascular deaths as component of primary efficacy outcome which may have mitigated the effect of edoxaban on reduction in stroke and SE. Variation in treatment assigned to control arm may also have contributed to difference in safety outcomes. All control arm patients received aspirin in ATRESIA trial while only 54% of control arm patients in NOAH-AFNET 6 trial received aspirin, which may have resulted in perception of reduced bleeding events with apixaban. Choice of study drug may also affect the differing outcomes. Although heterogenous, a few observational studies have indicated lower bleeding risk with apixaban[38]. Table 1 depicts the important differences between these two major trials.

Looking at the totality of evidence: Meta-analysis of OAC in SCAF

With a positive ATESIA study and a prematurely terminated NOAH-AFNET 6 study, meta-analyses were poised to throw additional light on the issue. A study-level meta-analysis by McIntyre et al[39] evaluated the role of OACs in device-detected AF. It included > 6500 patients from the NOAH-AFNET 6 and ARTESIA trials. OAC use was associated with a 32% reduction in ischemic stroke (HR 0.68; 95%CI: 0.50-0.92), with no heterogeneity between the two studies, indicating high data quality. There was also a concomitant 35% reduction in all-cause stroke or SE. A 15% reduction was also observed in the composite outcome of cardiovascular death, all-cause stroke, peripheral arterial embolism, myocardial infarction, or pulmonary embolism (95%CI: 0.73-0.99). However, OAC did not impact cardiovascular or all-cause mortality. Major bleeding events increased by 62% with OAC, though there was no increase in fatal bleeding.

Another meta-analysis examined the time to benefit with OAC in patients with SCAF, again using data from over 6500 patients enrolled in the NOAH-AFNET 6 and ARTESIA[40]. This analysis corroborated a roughly one-third reduction in stroke risk and a 57% increase in major bleeding. Simulation modelling indicated that it would take 2.67 years to prevent one stroke per 1000 DOAC-treated patients, while only 1.67 years were needed to observe one major bleeding event. These findings suggest a delayed ischemic benefit from OAC in SCAF, whereas bleeding complications manifest earlier. These findings may help to explain the failure of edoxaban therapy in producing a statistically significant benefit due to premature termination of the NOAH-AFNET 6 trial (< 2.67 years).

The impact of prior vascular disease-such as stroke/TIA, coronary artery disease, or peripheral arterial disease-on DOAC efficacy was assessed in another meta-analysis[41]. Approximately half of the patients in the two trials had pre-existing vascular disease (56% and 46%). Both thromboembolic and bleeding events were more frequent in these patients compared to those without vascular disease. Among patients with prior vascular disease, the magnitude of benefit from anticoagulation was greater: For the combined outcome of stroke, SE, myocardial infarction, pulmonary embolism, and cardiovascular death, the incidence rate ratio (IRR) was 0.75 (95%CI: 0.61-0.92), compared to 1.01 (95%CI: 0.78-1.30) in those without vascular disease. A similar trend was seen for the composite of stroke and SE (IRR: 0.55 vs 0.87, respectively), although the differences did not reach statistical significance. These findings suggest that the absence of prior vascular disease may identify a subgroup in which DOAC therapy may be unnecessary. Conversely, SCAF patients with established vascular disease demonstrated higher stroke/ SE rates (1.8%-2.2%), exceeding the 1% annual risk threshold recommended by current AF guidelines for initiating DOAC therapy[1,2].

In a nutshell, the meta-analyses of two SCAF randomized controlled trials reveal a clear and consistent reduction of ischemic endpoints with OAC in SCAF. More so, SCAF patients with prior vascular diseases derive higher benefit from OAC therapy. However, these ischemic benefits appear late after therapy initiation, which mandates a prolonged duration of the treatment.

Choosing the right patient

Based on the results from the positive ATRESIA study and the McIntyre meta-analysis, we can infer that OAC in SCAF leads to an unequivocal reduction in ischemic stroke/SE by roughly 1/3rd (32%-37%). Nonetheless, these benefits are accrued at the cost of an increase in major bleeding events. However, it should be noted here that bleeding related to DOAC therapy has a lower in-hospital mortality and adverse consequences compared to warfarin therapy[42,43]. Even among patients with OAC-related ICH (which is the most feared complication), DOAC-related ICH patients had less ICH volume, less in-hospital mortality, coupled with better neurological outcomes[44,45]. In the extracranial sites, gastrointestinal (GI) bleeding is a concern regarding DOAC use compared to warfarin[46]. But in a sub-analysis of ENGAGE-AF TIMI-48 study, only 10% of major GI bleeds were fatal. Moreover, even though edoxaban in higher doses produced more GI bleeding than warfarin, the rates of fatal GI bleeding were not different[47]. Hence, fear of bleeding should not deter a clinician from initiating DOAC in SCAF whenever indicated.

Even when one disregarded the bleeding risk, the overall incident stroke rates were also low in the SCAF population of both trials. Hence, the net clinical benefit might not favour the universal use of DOAC in SCAF. However, if one were to segregate the SCAF sub-population with high stroke risk, we might get a clear net clinical benefit in favor of DOAC. Patients with high CHA₂DS₂VASc score (> 4), SCAF duration > 24 hours, those with prior stroke, and established vascular disease are the four evidence-based subgroups with high stroke rate[31,32,37,41]. These patients represent an important subgroup in whom the risk–benefit ratio of anticoagulation may be favorable (Figure 4).

Figure 4 Flow chart regarding anticoagulation approach in subclinical atrial fibrillation. The green colour indicated factors favouring direct oral anticoagulation (DOAC) use in subclinical atrial fibrillation, while the red colour represents factors arguing against DOAC therapy. The CHA₂DS₂-VASc score is used to predict the risk of ischemic stroke among patients with atrial fibrillation and ranges from 0 to 9, with higher scores indicating a greater risk of stroke. SCAF: Subclinical atrial fibrillation; CAD: Coronary artery disease; PAD: Peripheral artery disease; DOAC: Direct oral anticoagulation; CHA₂DS₂-VASc score: Congestive heart failure, hypertension, age ≥ 75 years (2 points), diabetes mellitus, prior stroke or TIA or thromboembolism (2 points), vascular disease, age 65-74 years, sex category.

Inflammation and SCAF

A prior study demonstrated an association between elevated post–cardiac surgery white blood cell (WBC) count and an increased risk of postoperative AF[48]. This finding highlights a link between inflammation and the onset or recurrence of atrial arrhythmias, suggesting a shared pathophysiological mechanism between AHREs and AF. Activated WBC release proinflammatory mediators, including cytokines and reactive oxygen species, which interact with atrial cardiomyocytes, promoting electrical remodeling and fibrosis[49-51]. These mediators also contribute to a prothrombotic state, which may explain the elevated risk of thromboembolism and cardiovascular mortality in patients with AHREs. The relationship between systemic inflammation and cardiovascular outcomes in this population requires further investigation. Additionally, advanced age, elevated C-reactive protein (CRP), higher WBC counts, and a history of AF have been independently associated with the occurrence of AHREs[52]. Peri-implantation levels of inflammatory markers (CRP and WBC count) correlate with the development of AHREs over time. Oxidative stress has also been postulated to be involved in the genesis and perpetuation of AF[53]. The damage to atrial myocardium due to oxidative stress is a contributor to the remodeling of atrial walls. Collectively, these findings suggest that SCAF carries an increased stroke risk, albeit lower than that observed in patients with clinical AF.

Guidelines

Current clinical guidelines stratify the management of SCAF based on the duration of AHREs and the CHA₂DS₂VASc score[1,2]. The 2023 ACC/AHA guideline for AF management advocates that patients with SCAF < 6 minutes duration do not need any OAC. For patients with SCAF duration between 6 minutes to 24 hours, a CHA₂DS₂VASc score > 3 would indicate the need for OAC. For those with SCAF duration > 24 hours, a CHA₂DS₂VASc score > 2 would indicate the need for OAC. However, the guidelines acknowledge the lack of randomized data and give emphasis to shared decision making (Figure 5)[2]. It stratifies patients according to their duration of SCAF and baseline ischemic risk (CHA₂DS₂VASc score). The red zones indicate areas of low risk and obviate the need for any OAC therapy. The green zones demarcate areas of high risk and need for OAC based on contemporary evidence. The orange zones are a grey area and need further research.

Figure 5 The contemporary 2023 ACC/AHA guideline on the use of oral anticoagulation in subclinical atrial fibrillation. The orange zones indicate areas of low risk and obviate the need for oral anticoagulation. The green zones demarcate areas of high risk and need for oral anticoagulation. The yellow zones are a grey area and need further research.

The 2020 ESC guidelines for management of AF advise a complete workup of the patient with SCAF, including an ECG, comorbidity assessment & CHA₂DS₂VASc score[1]. It recommends regular follow-up for most patients with remote monitoring to detect progression into clinical AF or SCAF burden > 24 hours. OAC is advised only in patients with SCAF duration > 24 hours and a high risk of stroke (e.g., prior stroke, age ≥ 75 years, CHA₂DS₂VASc score ≥ 3). The 2024 ESC guidelines fail to give an unequivocal recommendation despite being published after ATRESIA and NOAH-AFNET studies[54]. The guidelines say that DOACs can be used in patients with SCAF with high stroke risk and low bleeding risk (Class IIb recommendation). Unlike the 2023 ACC/AHA guideline document, they acknowledge that the duration of SCAF, which would justify OAC use, is uncertain. However, owing to the high risk of future AF in this scenario, they advise emphasis on all the aspects of the “CARE” pathway for routine AF management (C represents comorbidity and risk factor assessment, A refers to avoiding stroke and thromboembolism, R denotes reduction of symptoms by attention to rate and finally E represents evaluation and dynamic assessment).

CONCLUSION

SCAF is being detected at an increased frequency, and can be demonstrated in up to 30% patients with CIED device implantation. Although advanced monitoring tools, such as ILRs, smartwatches, etc., are refining risk assessment and stroke prevention strategies, many grey areas still remain[55]. Anticoagulation reduces the stroke risk but simultaneously increases bleeding risk, making personalised, shared decision-making essential. The positive results of the ARTESIA study should be interpreted in the light of increased bleeding risk with OAC. SCAF > 24 hours should be managed as clinical AF, especially with a High CHA₂DS₂-VASc score (> 4) and antecedent stroke /vascular disease, while shorter episodes require CHA₂DS₂-VASc based stratification. Patients with high CHA₂DS₂-VASc scores and low bleeding risk are optimal candidates for anticoagulation; those at high bleeding risk need risk factor optimisation. Future research should focus on predictors such as AF burden, episode duration, structural changes, and biomarkers (e.g., pro-BNP).