Related Experiment Video
Updated: Jun 24, 2026
Echocardiographic Evaluation of Atrial Communications before Transcatheter Closure
Published on: February 8, 2022
External Validation of Published Scoring Systems to Predict Incident Atrial Fibrillation After Cryptogenic
Gaetano Pinnacchio1, Roberto Scacciavillani1, Maria Lucia Narducci1
1Department of Cardiovascular Sciences Fondazione Policlinico Universitario A. Gemelli IRCCS Rome Italy.
Insights
Three scoring systems, C2HEST, Brown ESUS-AF, and AS5F, reliably predicted atrial fibrillation (AF) in patients with cryptogenic stroke (CS) or embolic stroke of undetermined source (ESUS). These tools can identify low-risk individuals, guiding targeted diagnostic and prevention strategies.
Area of Science:
- Cardiology
- Neurology
- Medical Diagnostics
Background:
- Atrial fibrillation (AF) is a significant, often undetected, cause of cryptogenic stroke (CS) and embolic stroke of undetermined source (ESUS).
- Implantable cardiac monitors improve AF detection but face limitations in cost and accessibility.
- Existing AF prediction scoring systems for CS/ESUS patients often lack external validation.
Purpose of the Study:
- To externally validate seven established scoring systems for predicting incident atrial fibrillation (AF) in patients diagnosed with CS or ESUS.
- To assess the discriminatory ability and threshold performance of these scoring systems.
Main Methods:
- An observational study involving 132 patients with CS or transient ischemic attack who underwent implantable cardiac monitoring.
- Application of seven prediction models: C2HEST, Brown ESUS-AF, ESUS-AF, HAVOC, ACTEL, AS5F, and CHASE-LESS.
- Primary outcome: discriminatory ability (Area Under the Curve - AUC) for AF detection; Secondary outcomes: threshold performance and specificity.
Main Results:
- Atrial fibrillation (AF) was detected in 30.3% of patients during a mean follow-up of 13 months.
- The C2HEST (AUC 0.704), Brown ESUS-AF (AUC 0.755), and AS5F (AUC 0.726) scores demonstrated acceptable predictive performance (AUC > 0.7).
- Threshold criteria derived from these three scores achieved high specificity and positive predictive value for AF prediction.
Conclusions:
- The C2HEST, Brown ESUS-AF, and AS5F scoring systems reliably predict incident AF in patients with CS/ESUS.
- These validated scores can identify patients at low risk for AF, enabling more focused use of diagnostic technologies.
- Findings support optimized stroke prevention strategies by guiding the application of cardiac monitoring.
Background:
Atrial fibrillation (AF) is a common but often undetected cause of cryptogenic stroke (CS) and embolic stroke of undetermined source (ESUS). Implantable cardiac monitors aid in detecting subclinical AF, but their widespread use is limited by cost and logistical challenges. Several scoring systems exist to predict AF in patients with CS/ESUS, though few have been externally validated. This study was designed to externally validate 7 scoring systems for predicting incident AF in patients with CS/ESUS.
Methods:
In this observational study, we analyzed 132 patients with CS or transient ischemic attack who received implantable cardiac monitors. After a comprehensive literature search, 7 prediction models (Coronary Artery Disease or Chronic Obstructive Pulmonary Disease, Hypertension, Elderly, Systolic Heart Failure, Thyroid Disease [C2HEST], Brown ESUS-AF, ESUS-AF, Hypertension, Age, Valvular Heart Disease, Peripheral Vascular Disease, Obesity, Congestive Heart Failure, Coronary Artery Disease [HAVOC], Age, Hypercholesterolemia, Tricuspid Regurgitation, Left Ventricular End-Diastolic Volume, Left Atrium [ACTEL], Age, Stroke Severity, National Institutes of Health Stroke Scale >5 to Find Atrial Fibrillation [AS5F], and Coronary, Heart Failure, Age, Stroke Severity, Lipidemia, Sugar, Prior Stroke [CHASE-LESS]) were applied. The primary outcome was the discriminatory ability to detect AF. Secondary outcomes included evaluating the threshold scores and the performance of such threshold-based criteria for predicting nonoccurrence of AF with high specificity.
Results:
AF was detected in 40 patients (30.3%) over a mean follow-up of 13±12 months. Area under the receiver operating characteristic curves were C2HEST (0.704), Brown ESUS-AF (0.755), AS5F (0.726), ESUS-AF (0.607), HAVOC (0.661), ACTEL (0.650), and CHASE-LESS (0.671). C2HEST, Brown ESUS-AF, and AS5F showed acceptable predictive performance (area under the receiver operating characteristic curve >0.7). Threshold-based criteria using these scores achieved high specificity and positive predictive value. Combining thresholds across scores improved identification of patients unlikely to develop AF.
Conclusions:
C2HEST, Brown ESUS-AF, and AS5F scores reliably predicted AF in patients with CS/ESUS and may help identify low-risk individuals, supporting more targeted use of diagnostic tools and guiding stroke prevention strategies.
Related Concept Videos
01:25Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT
01:30Acute Coronary Syndrome III: Diagnostic Studies
01:20Venous Thrombosis II: Clinical Manifestations and Diagnostic Studies