Performance of a new atrial fibrillation detection algorithm in a miniaturized insertable cardiac monitor: Results from the Reveal LINQ Usability Study

Abstract

For clinicians, confidence in atrial fibrillation (AF) episode classification is an important consideration when electing to use insertable cardiac monitors (ICMs). The purpose of this study was to report on the improved AF detection algorithm in the Reveal LINQ ICM. The Reveal LINQ Usability Study is a nonrandomized, prospective, multicenter trial. The ICM has been miniaturized, uses wireless telemetry for remote patient monitoring, and its AF algorithm includes a new p-wave filter. At 1 month post-device insertion, Holter monitor data were collected and annotated for true AF episodes ≥2 minutes, and performance metrics were evaluated by comparing Holter annotations with ICM detections. The study enrolled 151 patients (age 56.6 ± 12.1, male 67%). Reasons for monitoring included AF ablation or AF management in 81.5% (n = 123), syncope in 12.6% (n = 19), and other indications in 5.9% (n = 9) of patients. Of the 138 patients with an analyzable Holter recording, a total of 112 true AF episodes were identified in 38 patients (27.5%). The overall accuracy of the ICM to detect durations of AF or non-AF episodes was 99.4%, and the AF burden measured by the ICM was highly correlated with the Holter (Pearson coefficient 0.995). The new AF detection algorithm in the Reveal LINQ ICM accurately detects the presence or absence of AF. Additionally, it showed high sensitivity in detecting AF duration in patients with a history of intermittent and symptomatic AF.