Subclinical AF | Pocketcardio

Device-detected subclinical atrial fibrillation – thromboembolic risk and the need for anticoagulation

Introduction

Atrial fibrillation has long been recognised as a risk factor for thromboembolic events, and the benefit of anticoagulation in clinical atrial fibrillation for the prevention of ischaemic stroke and systemic embolism is undisputed. Nevertheless, atrial fibrillation continues to account for a large proportion of severe cerebrovascular events. One proposed solution has been earlier detection of atrial fibrillation, allowing preventive anticoagulation to be initiated in a timely manner.

Atrial fibrillation can now be detected with increasing sensitivity using smartphones, smartwatches, long-term ECG monitoring, and cardiac implantable devices capable of rhythm monitoring, such as pacemakers. This has raised a new clinical question: what should be done when a short, device-detected atrial arrhythmia compatible with atrial fibrillation is found incidentally during long-term rhythm monitoring?

The key practical issue is whether lifelong anticoagulation should be initiated on the basis of a short device-detected AF/AHRE episode. The available evidence suggests that the answer is not unequivocally yes.

Concepts: AHRE, device-detected AF, and clinical AF

Cardiac rhythm management devices, such as bradycardia pacemakers, cardiac resynchronisation devices, implantable cardioverter-defibrillators, and implantable cardiac monitors, continuously monitor cardiac rhythm. They detect occult atrial fibrillation considerably more sensitively than a 12-lead ECG, Holter recording, or symptom-triggered ECG.

Device-detected subclinical atrial fibrillation refers to asymptomatic atrial fibrillation episodes detected by devices providing continuous rhythm monitoring. This category includes both implantable rhythm management devices and consumer wearable devices, provided that the finding is reliably documented.

AHRE, or atrial high-rate episode, usually refers to an atrial tachyarrhythmia detected by an implanted device, with an atrial rate typically >175/min and a duration >5 minutes. However, not all AHRE findings represent true atrial fibrillation. False-positive findings are relatively often caused by artefacts, noise, and far-field R-wave sensing. These non-true false positives are often episodes lasting less than 5 minutes. The nature and duration of the longest or clinically most relevant episode should always be confirmed by visual review of the recording by a physician.

Clinical atrial fibrillation, in contrast, refers to symptomatic or asymptomatic atrial fibrillation documented on surface ECG.

On single-lead recording or a device-stored electrogram, the diagnosis of atrial fibrillation requires an arrhythmia lasting at least 30 seconds that, on visual review, is compatible with atrial fibrillation. In clinical practice, however, atrial fibrillation detected on surface ECG is considered sufficient if atrial fibrillation is present throughout a standard ECG recording or is documented during symptoms.

Why is device-detected subclinical AF important?

Subclinical atrial fibrillation is not a harmless finding. Numerous studies have shown that device-detected subclinical atrial fibrillation predicts:

subsequently diagnosed clinical atrial fibrillation
long episodes lasting >24 hours
an increased risk of thromboembolic events

At the same time, it is important to recognise that the thromboembolic risk associated with subclinical atrial fibrillation is lower than the risk associated with clinical atrial fibrillation documented on surface ECG. In studies of device-detected subclinical AF, the risk of stroke has been lower than in patients with clinical atrial fibrillation and a comparable risk profile.

AHRE/SCAF episodes are common, particularly in elderly, multimorbid patients and in those with higher thromboembolic risk. In a meta-analysis of 54 studies, the pooled prevalence of SCAF, or subclinical atrial fibrillation, was 28.1% among patients with CIEDs, or cardiac implantable electronic devices. Device-detected AF episodes lasting more than 6 minutes are found in approximately 20% of patients with an implanted pacemaker, implantable cardioverter-defibrillator, or long-term rhythm monitor. In observational studies, up to 30% of patients have AHRE episodes lasting 5–6 minutes during 1–3 years of follow-up. This high prevalence of short atrial fibrillation episodes has also led to consideration of whether such short atrial fibrillation episodes may to some extent be physiological.

AF burden, episode duration, and thromboembolic risk

Based on current evidence, stroke risk is related to both the duration and the burden of subclinical atrial fibrillation. The longer or more frequent the device-detected atrial fibrillation, the higher the thromboembolic risk.

Short episodes are associated with lower risk than longer episodes. Long episodes lasting at least 24 hours are clearly associated with an increased risk of stroke or systemic embolism. By contrast, very short episodes lasting less than 5 minutes are generally not associated with clinical events, and some represent another arrhythmia or a technical false alarm.

Risk increases according to AF burden and the CHA₂DS₂-VASc risk profile. Nevertheless, no precise threshold for the duration or burden of a single episode is known beyond which thromboembolic risk increases sufficiently to justify initiation of permanent anticoagulation. In observational studies, proposed thresholds have varied widely: from 5–6 minutes to 1 hour, 5.5 hours, or 24 hours.

It should be noted that most DDAF studies have specifically examined the duration of the longest episode and classified patients accordingly, rather than according to AF burden.

Progression to clinical atrial fibrillation

Device-detected subclinical AF is often an early phase of atrial fibrillation. On average, approximately one in five patients progresses to clinical atrial fibrillation or to AHRE/SCAF lasting more than 24 hours during approximately 2 years of follow-up. This is consistent with the estimate that progression to clinical atrial fibrillation occurs in approximately 6–9% of patients per year.

In the NOAH-AFNET 6 dataset, progression to clinical atrial fibrillation was more common when device-detected AHRE lasted more than 24 hours. In this case, the rate of progression to clinical AF was approximately twice that seen with shorter episodes, at approximately 17% per patient-year.

Conversely, in a subgroup analysis of the LOOP study, approximately half of patients in whom a short subclinical atrial fibrillation episode was detected had no recurrence during approximately 3 years of follow-up. This supports a cautious approach to initiating permanent lifelong anticoagulation on the basis of a single short incidental finding.

Randomised trials: NOAH-AFNET 6 and ARTESIA

The anticoagulation benefit demonstrated in clinical atrial fibrillation cannot be directly extrapolated to device-detected subclinical AF. This question was examined in two large randomised trials.

NOAH-AFNET 6

In the NOAH-AFNET 6 trial, 2536 patients, mainly at high thromboembolic risk, were randomised. All had at least one subclinical atrial fibrillation episode lasting more than 6 minutes stored in the pacemaker memory. The median CHA₂DS₂-VASc score was approximately 4. The mean follow-up was 1.8 years. The control group received placebo, but approximately 50% used aspirin for another indication.

The trial was terminated prematurely because of safety concerns and futility. Edoxaban did not reduce the composite endpoint of cardiovascular death, stroke, or systemic embolism, but increased bleeding risk. In addition, the composite of death or major bleeding was more common in the edoxaban group.

ARTESIA

In the ARTESIA trial, 4012 similar patients, mainly at high thromboembolic risk, were randomised. They had at least one device-stored subclinical AF episode lasting more than 6 minutes but no more than 24 hours. Mean follow-up was 3.5 years. All patients in the control group were started on aspirin.

ARTESIA showed that apixaban reduced the risk of stroke or systemic embolism compared with aspirin. At the same time, apixaban increased the risk of major bleeding. There was no difference between treatment groups in the incidence of fatal bleeding.

An important observation was that, in ARTESIA, 43% of strokes occurring in the aspirin group led to significant disability or death. Conversely, most major bleeds were managed conservatively, often with transfusions, and no increase in fatal bleeding or all-cause mortality was observed.

When interpreting the results of ARTESIA, it is important to note that initiation of aspirin in all control patients does not correspond to the usual clinical situation, and this may affect both bleeding and efficacy comparisons.

Summary of the studies

In both studies, patients were predominantly elderly and belonged to a high thromboembolic-risk group; in meta-analytic summaries, the mean age was approximately 77 years and the mean CHA₂DS₂-VASc score approximately 4. In both trials, the burden of subclinical AF was relatively low: the median duration was approximately 1.5 hours in ARTESIA and approximately 2.8 hours in NOAH-AFNET 6. During follow-up, clinically documented AF or subclinical AF lasting more than 24 hours was found in approximately 20% of patients.

According to a meta-analysis of the studies, DOAC therapy:

reduced the relative risk of ischaemic stroke by approximately 30%
increased the relative risk of major bleeding by approximately 60%
did not show a clear mortality benefit

In absolute terms, the effect was small. In one summary, DOAC therapy reduced the absolute risk of ischaemic stroke by approximately 0.3% per year, i.e. approximately 3 fewer ischaemic strokes per 1000 patient-years, but increased major bleeding by approximately 0.7% per year, i.e. approximately 7 more major bleeds per 1000 patient-years. Expressed another way, the intention-to-treat analysis of ARTESIA suggested approximately 4.6 fewer stroke/systemic embolism events and approximately 4.1 more major bleeds per 1000 patient-years.

Thus, the overall direction is that stroke risk decreases, bleeding risk increases, and the net benefit is borderline.

What does the current evidence mean in practice?

Current evidence does not support the idea that all patients with device-detected subclinical atrial fibrillation should routinely be started on permanent anticoagulation. At the same time, device-detected subclinical AF should not be regarded as an entirely benign finding.

In practical decision-making, three key issues should be considered:

the thromboembolic risk associated with device-detected subclinical AF is lower than that associated with clinical AF
a DOAC may reduce stroke risk but increases the risk of major bleeding
the patient group is heterogeneous, and the optimal burden or duration threshold is unknown

It has been proposed that, in clinical atrial fibrillation, the net benefit of anticoagulation becomes clearly favourable when the annual thrombotic risk exceeds approximately 1%. In both NOAH-AFNET 6 and ARTESIA, the thrombotic risk in the control group was approximately 1% per year, which is considerably lower than would be expected on the basis of CHA₂DS₂-VA/CHA₂DS₂-VASc risk scoring, and therefore exactly at the threshold where the net effect of anticoagulation becomes uncertain. This helps explain why the established benefit in clinical AF does not directly translate to this patient population.

Guideline recommendations

ESC 2024

The European guideline is cautious on this issue. According to it, DOAC therapy may be considered in patients with device-detected subclinical AF and elevated thromboembolic risk, provided that bleeding risk is not high. The recommendation class is IIb and the level of evidence is B.

The ESC does not provide a precise burden or duration threshold above which anticoagulation should be initiated. The guideline emphasises that the duration or burden threshold of subclinical AF at which the benefits of oral anticoagulation clearly begin to outweigh the harms remains uncertain.

The ESC’s practical interpretation of “elevated thromboembolic risk” relies primarily on the CHA₂DS₂-VA/CHA₂DS₂-VASc risk profile. “Bleeding risk is not high” does not refer to a single absolute numerical value, but to an overall assessment. The HAS-BLED score helps identify patients at high bleeding risk, but does not alone determine the treatment decision.

A HAS-BLED score ≥3 is, in practice, a warning marker for bleeding risk and supports closer follow-up and active correction of bleeding risk factors. In the context of subclinical AF, however, this should be assessed particularly carefully, because the benefit of DOAC therapy is more uncertain than in clinical atrial fibrillation.

ACC/AHA/HRS 2023

The US guideline is more prescriptive and uses duration-based thresholds. According to it:

if AHRE lasts ≥24 hours and CHA₂DS₂-VASc is ≥2, initiation of OAC is reasonable
if AHRE lasts from 5 minutes to 24 hours and CHA₂DS₂-VASc is ≥3, OAC may be reasonable
if AHRE lasts <5 minutes and there is no other indication, OAC should not be initiated

The guideline emphasises that the duration and nature of the arrhythmia should be confirmed and that the decision should be made according to the principle of shared decision-making.

It is important to note that the 2023 ACC/AHA/HRS guideline was published before the final results of NOAH-AFNET 6 and ARTESIA. Since then, US expert discussion has become more cautious, although the actual guideline text has not been formally changed.

When ARTESIA and the NOAH+ARTESIA meta-analysis were incorporated, the ACC journal scan stated that these two trials together suggest that DOACs reduce ischaemic stroke but increase major bleeding and do not reduce mortality. In its practical perspective, the ACC assessed that the target group for treatment might be particularly patients with a longer subclinical AF burden, >2–4 h, and a higher CHA₂DS₂-VASc score, ≥4.

The ACC highlighted that in patients with CHA₂DS₂-VASc >4, apixaban prevented more stroke/SE events than it caused major bleeding, and the article stated that these patients “should generally be treated with an OAC”; in the lower-risk group, the benefit was less clear.

APHRS 4S-DDAF approach

In the 4S-DDAF approach of the Asia Pacific Heart Rhythm Society, device-detected AF is assessed as follows:

strip documentation and longest AF duration
symptoms
stroke history
score

In this model:

an episode lasting more than 24 hours supports management similar to clinical AF, including OAC
if the longest episode is ≤24 hours, symptoms are assessed
symptomatic patients may be managed as having clinical AF
in asymptomatic patients, prior ischaemic stroke or TIA favours OAC
if there is no prior stroke/TIA, CHA₂DS₂-VASc and vascular disease are assessed
if CHA₂DS₂-VASc is ≥4 or vascular disease is present, OAC may be considered
in lower-risk patients, follow-up, management of risk factors, and an individualised plan are emphasised

NHS Highland 2025, Management of Device-Detected Atrial Fibrillation

If device-detected atrial fibrillation (DDAF) lasting from 6 minutes to 24 hours occurs in a patient without documented vascular disease, such as prior stroke or TIA, coronary artery disease, or peripheral arterial disease, initiating DOAC therapy is not recommended.

If these episodes last more than 24 hours, an individual risk assessment is advised, for example using the CHADSVA score, and it is recommended to consider apixaban. These guidelines state that “the available weak evidence has not shown a meaningful difference in benefit between DOAC use for episodes lasting more than 24 hours and those lasting less than 24 hours”.

If DDAF lasting more than 6 minutes occurs in a patient who is already taking an antiplatelet drug for documented vascular disease, such as aspirin or clopidogrel, switching from the antiplatelet agent to a DOAC may reduce the annual stroke risk by 1%, at the cost of a 0.6% higher annual risk of major bleeding. A switch should be considered.

If DDAF lasts less than 6 minutes, continued monitoring is recommended.

DGK, German Cardiac Society position paper

The DGK position paper argues that DDAF should be defined only for episodes lasting at least 5 to 6 minutes, because shorter episodes are often artefacts. It also states that the AF burden threshold at which anticoagulation is clearly beneficial cannot yet be defined.

According to the DGK, a CHA₂DS₂-VASc score above 4 alone is still insufficient to support a general recommendation for anticoagulation, and in patients with scores in the 2 to 4 range, benefit is considered highly unlikely.

In patients without vascular disease, the paper generally does not support anticoagulation for DDAF. If vascular disease is present and there is an indication for aspirin, switching to a DOAC may be considered. Prior stroke or TIA is regarded as the group in which anticoagulation can most readily be considered. The paper also emphasises that only AF documented on surface ECG provides a clear conventional indication for anticoagulation.

How should “vascular disease” be understood?

Some analyses have suggested that the benefit of DOAC therapy may be concentrated particularly among patients with vascular disease. This generally refers to documented atherosclerotic vascular disease, such as:

coronary artery disease
prior myocardial infarction
prior PCI or CABG
symptomatic or angiographically documented significant coronary artery disease
peripheral arterial disease
aortic plaque

In some contexts, prior stroke/TIA has also been included under the term vascular disease.

Always treat risk factors and arrange follow-up

Irrespective of whether OAC is initiated, the patient requires a comprehensive assessment:

treatment of hypertension
treatment of diabetes
weight management
reduction of alcohol intake
identification of sleep apnoea
physical activity and lifestyle management
rhythm monitoring to detect progression

When should anticoagulation generally not be initiated?

Caution is particularly justified if:

there is a single short incidental finding
the episode is less than 5 minutes
thromboembolic risk is low
bleeding risk is high
the episode does not recur during follow-up
the episode is an AHRE that has not been confirmed as true AF

Particular restraint is justified if permanent lifelong anticoagulation would be based solely on one short device-detected atrial fibrillation episode.

Summary

Device-detected subclinical atrial fibrillation is a common finding, particularly in elderly and multimorbid patients. It is not entirely harmless, because it is associated with an increased risk of subsequent clinical atrial fibrillation and thromboembolic events. However, the risk is clearly lower than in clinical atrial fibrillation documented on surface ECG.

Current evidence does not support routine anticoagulation for all patients on the basis of a single short device-detected AF/AHRE episode. At the same time, the evidence also does not support disregarding the finding entirely. DOAC therapy may be justified in selected patients with high thromboembolic risk and low bleeding risk, but the decision should be individualised on the basis of burden, risk profile, bleeding risk, and the patient’s own values.

In practice, device-detected subclinical AF is above all a signal for more detailed assessment, follow-up, and management of risk factors. In some patients, it also leads to anticoagulation.