METHODS AND RESULTS: In this multicentre, open-label study, we randomly assigned 203 participants to undergo one additional 24-h Holter monitoring (control group, n = 98) vs. 30-day smartphone ECG monitoring (intervention group, n = 105) using KardiaMobile (AliveCor®, Mountain View, CA, USA). Major inclusion criteria included age ≥55 years old, without known AF, and ischaemic stroke or transient ischaemic attack (TIA) within the preceding 12 months. Baseline characteristics were similar between the two groups. The index event was ischaemic stroke in 88.5% in the intervention group and 88.8% in the control group (P = 0.852). AF lasting ≥30 s was detected in 10 of 105 patients in the intervention group and 2 of 98 patients in the control group (9.5% vs. 2.0%; absolute difference 7.5%; P = 0.024). The number needed to screen to detect one AF was 13. After the 30-day smartphone monitoring, there was a significantly higher proportion of patients on oral anticoagulation therapy at 3 months compared with baseline in the intervention group (9.5% vs. 0%, P = 0.002).
CONCLUSIONS: Among patients ≥55 years of age with a recent cryptogenic stroke or TIA, 30-day smartphone ECG recording significantly improved the detection of AF when compared with the standard repeat 24-h Holter monitoring.
METHODS: Patients with cryptogenic stroke referred between August 2014 and February 2017 had ILRs implanted. Episodes of AF >2 minutes duration were recorded using proprietary algorithms within the ILRs, whereupon clinicians and patients were alerted via remote monitoring. All AF episodes were adjudicated using recorded electrograms. Once AF was detected, patients were counseled for anticoagulation.
RESULTS: Seventy-one patients with cryptogenic stroke, (age 61.9 ± 13.5 years, 77.5% male, mean CHA2DS2VASc score of 4.2 ± 1.3) had ILRs implanted. Time from stroke to the ILR implant was a median of 66 days. Duration of ILR monitoring was 345 ± 229 days. The primary endpoint of AF detection at 6 months was 12.9%; and at 12 months it was 15.2%. Median time to detection of AF was 50 days. The AF episodes were all asymptomatic and lasted a mean of 77 minutes (± 118.9). Anticoagulation was initiated in all but 1 patient found to have AF.
CONCLUSIONS: The incidence of occult AF is high in Asian patients with cryptogenic stroke and comparable to western cohorts. The combination of ILR and remote monitoring is a highly automated, technologically driven, and clinically effective technique to screen for AF.
METHODS: A convolutional auto-encoder (CAE) based nonlinear compression structure is implemented to reduce the signal size of arrhythmic beats. Long-short term memory (LSTM) classifiers are employed to automatically recognize arrhythmias using ECG features, which are deeply coded with the CAE network.
RESULTS: Based upon the coded ECG signals, both storage requirement and classification time were considerably reduced. In experimental studies conducted with the MIT-BIH arrhythmia database, ECG signals were compressed by an average 0.70% percentage root mean square difference (PRD) rate, and an accuracy of over 99.0% was observed.
CONCLUSIONS: One of the significant contributions of this study is that the proposed approach can significantly reduce time duration when using LSTM networks for data analysis. Thus, a novel and effective approach was proposed for both ECG signal compression, and their high-performance automatic recognition, with very low computational cost.