P-wave evidence as a method for improving algorithm to detect atrial fibrillation in insertable cardiac monitors

BACKGROUND Frequent premature atrial contractions and sick sinus syndrome are primary causes of inappropriate atrial fibrillation (AF) detection in insertable cardiac monitors (ICMs). OBJECTIVE The study aimed to validate an algorithm designed to reduce inappropriate AF detection on the basis of the identification of a single P wave during the cardiac cycle. METHODS The original detection algorithm looks for evidence of AF based on differences in the pattern of R-R intervals over a 2-minute period. The improved algorithm reduces evidence for AF detection if P waves are detected. The algorithm was validated by using Hater data, which collected 2 leads of surface electrocardiogram and continuously uplinked ICM electrocardiogram over a 46-hour period. ICM detections were compared with Holter annotations to compute episode and duration detection performance. RESULTS Valid Holler recordings (8442 hours) were analyzed from 206 patients. True AF was observed in 76 patients, yielding 482 true AF episodes >= 2 minutes in duration and 1191 hours of AF. The algorithm correctly identified 97.8% of the total AF duration and 99.3% of the total sinus or non-AF rhythm duration. The algorithm detected 85% (90% per-patient average) of all AF episodes >= 2 minutes in duration, and 55% (78% per-patient average) of the detected episodes had AF. AF was found in 95 % of the detected episodes >1 hour. The improved algorithm reduced inappropriate episodes and duration by 46% and 55%, respectively, while also reducing appropriate episodes and duration by 2% and 0.1 %, respectively. CONCLUSION An improvement in the ICM algorithm for AF detection incorporating P-wave information substantially reduced inappropriately detected episodes and duration, with minimal reduction in sensitivity for detecting AF.