A Comprehensive Model to Predict Atrial Fibrillation in Cryptogenic Stroke: The Decryptoring Score

Area of Science:

• Cardiology
• Neurology
• Medical Diagnostics

Background:

• Cryptogenic stroke (CS) accounts for up to 30% of ischemic strokes (IS).
• Atrial fibrillation (AF) is detected in up to 30% of CS cases, necessitating effective detection methods for secondary prevention.
• Current diagnostic tools lack comprehensive predictive capabilities for AF in CS.

Purpose of the Study:

• To develop and validate the Decryptoring score, a novel predictive tool for AF detection in patients with CS.
• To integrate clinical conditions, biomarkers, and left atrial strain (LAS) into a single predictive score.
• To guide secondary prevention strategies in cryptogenic stroke patients.

Main Methods:

• Prospective recruitment of 63 patients with IS or transient ischemic attack (TIA) and ABCD2 score ≥ 4.
• Collection of clinical, laboratory, and echocardiographic data, including LAS.
• 15-day wearable Holter-ECG monitoring for AF detection.
• Development of the Decryptoring score using univariate and multivariant analyses.

Main Results:

• Atrial fibrillation was detected in 24% of patients.
• Key predictors for AF included age > 75, hypertension, elevated Troponin T and NTproBNP, and reduced LAS reservoir and conductivity.
• The Decryptoring score demonstrated high predictive accuracy (AUC 0.94), significantly outperforming the AF-ESUS score (AUC 0.65).
• Patients with scores < 10 had 0% AF detection, while those with scores > 35 had 80% AF detection.

Conclusions:

• The novel Decryptoring score accurately predicts AF in patients with cryptogenic stroke.
• This score has the potential to guide secondary prevention strategies.
• Further validation in an independent cohort is required before clinical implementation.