AF-ESUS Score Calculator
Select applicable items — the score is calculated below immediately.
Result
Score: —
Interpretation: —
How to interpret the AF-ESUS score
Because four items subtract points, the score can be negative. The original publication uses a single threshold of 0:
| AF-ESUS score | Share of cohort | New incident AF during follow-up |
|---|---|---|
| ≤ 0 | 42.3% | 1.97% |
| ≥ 1 | 57.7% | 26.9% |
The relative risk of new incident atrial fibrillation (AF) was 13.7 for a score ≥ 1 compared with ≤ 0. A score ≤ 0 had a sensitivity of 94.9% and a negative predictive value of 98.0%. The authors concluded that patients with a score of 1 or more may be better candidates for prolonged automated cardiac monitoring. They stress that the score should not be used to decide on anticoagulation.
What is the AF-ESUS score?
The AF-ESUS score estimates the probability that atrial fibrillation will be newly detected after an embolic stroke of undetermined source (ESUS). Its purpose is to identify patients with a low probability of AF, so that limited resources for prolonged cardiac monitoring can be directed to those more likely to benefit. It was developed by George Ntaios and colleagues and published in 2020.
The score was derived from 839 consecutive ESUS patients in three prospective stroke registries (Lausanne, Athens and Larissa; 43.1% women, median age 67 years). During a median follow-up of 24.3 months, 125 patients (14.9%) had new incident AF. The points were derived from the coefficients of a multivariable regression model. The area under the ROC curve was 84.8%, and calibration was acceptable. The score was internally validated by cross-validation.
Scoring table
| Item | Points |
|---|---|
| Age ≥ 60 years | +3 |
| Arterial hypertension | +2 |
| Left atrial diameter > 40 mm | +2 |
| Any supraventricular extrasystole on the standard 12-lead ECGs during hospitalisation | +1 |
| Left ventricular hypertrophy on echocardiography | −1 |
| Subcortical infarct | −2 |
| Left ventricular ejection fraction < 35% | −3 |
| Non-stenotic carotid plaque (ipsilateral or contralateral) | −3 |
The possible range is −9 to +8 points. In the derivation cohort, the median score was 5 in patients with new incident AF and 1 in patients without.
ESUS was defined according to the Cryptogenic Stroke/ESUS International Working Group as a non-lacunar brain infarct without ≥ 50% stenosis of the supplying extracranial or intracranial artery, without a major-risk cardioembolic source and without another specific cause. All patients had at least 24 hours of continuous cardiac monitoring.
Limitations
- The score has not been externally validated in the original publication. The authors state that it needs validation in independent cohorts, ideally with prolonged automated cardiac monitoring, before routine clinical use.
- Patients were not systematically screened with prolonged cardiac monitoring, so short or subclinical AF episodes may have been missed.
- Left atrial diameter was used because left atrial volume, a better predictor of AF, was not available for many patients.
- Intracranial vessel imaging was not required for the ESUS diagnosis. The analysis was retrospective, and the work-up differed between the participating registries.
- The score is intended to support the selection of patients for prolonged cardiac monitoring, not decisions on anticoagulation.
Frequently asked questions
Why do some items subtract points?
Left ventricular hypertrophy, an ejection fraction < 35% and non-stenotic carotid plaques were associated with a lower probability of AF. The authors explain this by competing causes: these conditions can themselves be the source of an embolic stroke. Subcortical infarcts were also associated with a lower probability, in line with previous evidence that AF-related strokes more often involve the cortex.
Which ECGs count for supraventricular extrasystoles?
In the original study, this item was defined as any supraventricular extrasystole recorded on any of the standard 12-lead ECGs performed during the hospitalisation for the stroke.
Does a score ≤ 0 rule out atrial fibrillation?
No. In the original cohort, 1.97% of patients with a score ≤ 0 still had new incident AF during follow-up. A low score only indicates a lower probability.
How does the AF-ESUS score differ from CHASE-LESS and C2HEST?
The AF-ESUS score was developed specifically in patients with ESUS and includes echocardiographic, ECG, brain imaging and carotid imaging findings. The CHASE-LESS score was developed in unselected patients with ischemic stroke, and the C2HEST score in the general population. Both use only clinical risk factors.
Related calculators
Reference
Ntaios G, Perlepe K, Lambrou D, et al. Identification of patients with embolic stroke of undetermined source and low risk of new incident atrial fibrillation: the AF-ESUS score. Int J Stroke. 2021;16(1):29–38. doi:10.1177/1747493020925281
Written and medically reviewed by Frederic Eberhardt, MD. Last reviewed: 13 October 2026