Clinical evaluation of a spot-check ECG for the diagnosis of atrial fibrillation

对心电图抽查结果进行临床评估,以诊断房颤

阅读:1

Abstract

BACKGROUND: Opportunistic screening of atrial fibrillation (AF) using wearable spot-check electrocardiogram (ECG) devices has emerged for early detection, yet remains limited by concerns for accuracy and false positives. OBJECTIVE: This study aimed to assess the Masimo W1 wrist-wearable ECG spot-check device in detecting AF and normal sinus rhythm (NSR) while rendering diagnostic ECG waveform quality to support physician overread. METHODS: Participants with and without a previous AF diagnosis were prospectively enrolled across 4 US sites to compare the AF and NSR detection performance between the Masimo W1 ECG and a 12-lead ECG reference, plus concordance of clinically relevant ECG waveform features assessed by blinded reviewers. RESULTS: Among 356 participants (mean age 56 years; 60% female), 169 (47.5%) were in NSR and 165 (46.3%) in AF, whereas 22 were excluded as noise outputs (1.7%) or other rhythms (4.5%). Sensitivity and specificity for automated AF detection were 99.3% (95% confidence interval [CI] 96.3-100) and 100% (95% CI 97.8-100), respectively. The inconclusive rate was 5%. AF and NSR detection probabilities were 89.7% and 99.4%, respectively. Qualitative comparison demonstrated 99% (95% CI 98-100.0) agreement with the reference ECG across reviewers and records. Quantitative analysis of ECG waveform features showed a high correlation between the Masimo W1 ECG and lead I of the 12-lead ECG for diagnosing AF. CONCLUSION: The Masimo W1 ECG demonstrated highly accurate automated AF detection, few noise classifications, and exceptional NSR detection to mitigate concerns over false positives. Qualitative preservation of ECG waveform features supports the requisite physician overread required by practice guidelines.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。