Opportunistic Screening for Atrial Fibrillation With Continuous ECG Monitoring in the Emergency Department

Abstract

Study objectives: To determine the prevalence of undiagnosed atrial fibrillation detectable through continuous ECG monitoring in adult emergency department (ED) patients, and to estimate the stroke risk of ED patients discharged with undiagnosed atrial fibrillation.

Methods: Retrospective cohort study of 65,244 unique consecutive adult patients who received continuous ECG monitoring in an academic ED from August 23, 2020, to January 16, 2024. Primary outcome was the proportion of patients with at least 30 seconds of atrial fibrillation on continuous monitoring who were discharged without documented atrial fibrillation diagnosis or anticoagulation. Secondary outcomes included atrial fibrillation burden, proportion of cases suitable for anticoagulation based on CHA₂DS₂-VASc score, and incidence of ischemic stroke through August 31, 2024, by atrial fibrillation status on discharge (no atrial fibrillation, known atrial fibrillation, and undiagnosed atrial fibrillation). We compared demographic and insurance characteristics of patients with undiagnosed versus known atrial fibrillation. We estimated Cox proportional hazard models for ischemic stroke after ED visit, stratified by atrial fibrillation status and adjusting for CHA₂DS₂-VASc score.

Results: Of 65,244 monitored patients, 1,945 (3.0%) were discharged with undiagnosed atrial fibrillation, with 1,385 (71.2%) meeting criteria for anticoagulation. Compared to patients with known atrial fibrillation, patients with undiagnosed atrial fibrillation were younger (median age 72 [interquartile range, 50 to 84] versus 79 [interquartile range, 69 to 87] years), more likely to be women (49.6% versus 43.7%), on Medicaid (16.9% versus 7.0%), lack a primary care provider (22.5% versus 12.4%), and identify as Black (7.0% versus 3.7%) or Hispanic/Latino (17.1% versus 9.5%). Patients with undiagnosed atrial fibrillation had 2.6 ischemic strokes per 100 person-years of follow-up, with an adjusted hazard ratio for stroke of 3.00 (95% confidence interval, 2.39 to 3.76) compared to patients without atrial fibrillation on monitor, and 1.32 (95% confidence interval, 1.02 to 1.71) compared to patients with known atrial fibrillation.

Conclusion: Analysis of continuous ECG monitoring in the ED identified undiagnosed atrial fibrillation in 3% of patients, predominantly from underserved populations, and at high risk of stroke. Opportunistic screening in the ED could facilitate earlier diagnosis and anticoagulation to prevent stroke.

Keywords: Atrial fibrillation; Continuous monitoring; Opportunistic screening.