Early Prediction of Paroxysmal Atrial Fibrillation in Patients With Acute Stroke Using the BANQMR Score

Abstract

Background and Purpose Paroxysmal atrial fibrillation (PAF) can cause embolic stroke but is often asymptomatic and may resolve before it can be detected. We developed a simple, practical scoring system to improve the detection of PAF in stroke patients. Methods Patients with acute ischemic stroke hospitalized in three stroke centers between 2014 and 2021 were retrospectively examined. Multivariate logistic regression analysis was used to identify independent risk factors for PAF in a derivation cohort. Using these factors, a scoring system was developed and validated in a separate cohort. Results The derivation and validation cohorts included 649 and 583 patients, respectively. Median follow-up was 606 and 101 days, respectively. The independent risk factors for PAF were brain natriuretic peptide (BNP) ≥55 pg/mL, atrial premature contractions (APCs), National Institutes of Health Stroke Scale (NIHSS) score ≥11, corrected QT interval (QTc) ≥0.46 seconds, and multiple acute cerebral infarcts (MACIs). The BANQMR score assigns points based on these factors (BNP=3, APC=3, NIHSS score=2, QTc=1, and MACIs=1) and classifies patients into three risk grades based on points: low (0–2), moderate (3–5), and high (6–10). PAF was actually detected in 53.4% of high-risk patients, 15.0% of moderate-risk patients, and 4.6% of low-risk patients. The BANQMR score outperformed other existing scoring systems in terms of PAF prediction. Conclusions The BANQMR score is a simple, accurate tool for predicting PAF in ischemic stroke patients shortly after hospitalization, providing a clinically applicable method for early detection.