Utility of Routine 24-Hour Electrocardiographic Holter Monitoring for Detecting Atrial Fibrillation in Patients Admitted with Acute Stroke Syndromes

TL;DR

The study finds that routine 24-hour Holter monitoring in stroke patients has low yield for detecting atrial fibrillation, suggesting it should be reserved for high-risk individuals.

Contribution

The study identifies peripheral vascular disease as an independent predictor of AF detection and quantifies the low yield of routine Holter monitoring in acute stroke patients.

Findings

Only 1.7% of patients had new AF detected by Holter monitoring.

Peripheral vascular disease was an independent predictor of AF detection (odds ratio = 17.133).

130 Holter studies are needed to detect one case of treatable AF.

Abstract

This study aimed to assess the operating characteristics of standard 12-lead electrocardiogram (ECG) and 24-hour Holter studies in detecting atrial tachyarrhythmia in patients with acute stroke syndromes. The study's objective was to determine the potential factors that increased the likelihood of detecting atrial fibrillation (AF) and its impact on management.

This retrospective cohort study included adults (aged ≥18 years) admitted at Sultan Qaboos University Hospital from January 2010 to December 2019, with ischaemic stroke or transient ischaemic attack and had undergone a 24-hour Holter study. The hospital stroke registry was used to screen patients.

A total of 1,258 patients were screened and only 460 patients, who had undergone 521 Holter studies, met the inclusion criteria. Among them, AF was detected with baseline ECG in 20 patients (4.4%), while new AF was detected in only eight patients (1.7%). Significant predictors for increased AF detection by Holter included advanced age, peripheral vascular disease (PVD), left ventricular diastolic dysfunction and chronic kidney disease Stage III, with PVD highlighted as an independent predictor (odds ratio = 17.133, 95% confidence interval = 2.139–137.201, P = 0.007). The degree of agreement between the baseline ECG and Holter ECG was substantial (K = 0.649, P < 0.001). A total of 130 Holter needs to be performed to detect one case of treatable AF.

This study showed that a routine ECG Holter study in acute stroke has a low clinical yield in detecting AF, which is in substantial agreement with baseline ECG. Therefore, further investigation of paroxysmal AF in acute stroke should be reserved for high-risk patients.