2024 American Academy of Neurology Abstract Website

Objective:

Develop a score to evaluate or predict underlying atrial fibrillation (AF) risk in patients with cryptogenic stroke.

Background:

Paroxysmal AF is difficult to diagnose and can lead to the underutilization of anticoagulants in stroke prevention. Detection of AF remains crucial in patients with cryptogenic stroke. We present a novel scoring tool termed ACE2L to predict AF as shown in this table.

ACE2L Score
A	Age > 60	1
C	Cortical or embolic ischemic stroke	1
E	ECG p-wave terminal force > 5,000 µVms	1
2	Echo evidence of left atrial dilatation	1
L	Lack of extracranial Large vessel hemodynamic stenosis on imaging	1

Design/Methods:

We retrospectively identified patients who had acute ischemic stroke and classified them as AF or non-AF by two years after the initial stroke. Patients who had a pre-existing diagnosis of AF were excluded.

Results:

Data was collected on 100 patients, of which 49 were included in the pAF group and 51 in the non-pAF group. With a cut-off score of > 3, the sensitivity and specificity are 85% (95% CI: 0.6935-1.0000) and 65% (95% CI: 0.4926-0.7074), respectively and the PPV and NPV are 34.7% (95% CI: 0.2137-0.4802) and 94.1% (95% CI: 0.8766-1.0000), respectively. The ACE2L score can also be utilized at a slightly higher cutoff score of > 4, with a corresponding sensitivity of 62.7% (95% CI: 0.5037-0.7505), specificity of 70.7% (95% CI: 0.5680-0.8466), PPV of 75.5% (95% CI: 0.6347-0.8755), and NPV of 56.9% (95% CI: 0.4327-0.7046).

Conclusions:

The ACE2L scoring tool has been shown to have high sensitivity, specificity, PPV, and NPV in predicting AF in patients with cryptogenic stroke by two years after initial ischemic stroke. Future prospective studies are needed to clearly establish the utility and role of this novel scoring tool in the management of cryptogenic stroke.