Comparative performance of wearable ECG devices for heart rhythm monitoring in endurance athletes

Abstract Background Athletes frequently present with palpitations, yet initial assessments may fail to detect transient arrhythmias, necessitating prolonged monitoring. While many wearables track heart rate, few capture ECG, limiting arrhythmia detection. Factors like sweating and movement further challenge accurate ECG signal acquisition during exercise. Purpose Reliable rhythm monitoring during exercise remains a significant challenge in athletes. This study compared the signal quality of wearable ECG modalities during exercise and prolonged monitoring in endurance athletes. Methods Fifteen healthy recreational cyclists (60% male; mean age 35.7 ± 1.9 years; VO2max 61.1 ± 2.7 mL/kg/min) underwent two supervised three-hour cycling sessions to compare four ECG monitoring modalities: a single-lead (SL) patch, a SL chest strap, and two multi-lead (ML) patches (linear geometry in session one; non-linear L-shaped in session two). Additionally, athletes followed a four-week protocol to evaluate these modalities in real-life conditions. Participants were instructed to wear each device for seven days, aiming to maximise usage and capture the highest number of sports sessions. Some devices allowed user-initiated detachment when preferred. For each modality, participants completed a 5-point Likert questionnaire on comfort, fit, and preference. ECG quality was assessed by an experienced cardiologist. Results The SL chest-strap achieved the highest percentage of interpretable ECG recordings during the three-hour cycling sessions (100.0% 99.69–100.0), outperforming the SL patch (98.1% 92.1–100.0; p = 0.021), both the Linear ML patch (38.8% 18.48–82.34; p = 0.024) and Non-linear ML patch (77.07% 30.09–97.86; p = 0.009) (Figure 1). Over 7 days, total recording time differed significantly (p 0.001). The SL patch recorded the longest time (10,080 ± 0 min), exceeding the chest-strap (2,533 ± 1,798 min; p 0.001), Linear ML patch (8,225 ± 1,696 min; p = 0.028) and Non-linear ML patch (6,767 ± 1,861 min; p 0.001). Despite shorter total recording time, the chest-strap captured the highest proportion of time during sports (33 ± 27% vs 10% for all other modalities; p 0.001) and maintained the highest percentage of interpretable ECG recordings during monitoring (99% 93–100 vs SL patch 93% 50–99, Linear ML patch 74% 33–98, Non-linear ML patch 79% 7–99; p 0.001) (Figure 2). Conclusion The SL chest-strap provided superior signal quality during exercise and was preferred by most. Patch-based systems, while less accurate during exercise, enabled substantially longer recording times and were more comfortable during daily life activities. The choice of modality should be tailored to the athlete's needs: While chest-strap ECGs may be preferable for rhythm monitoring during sports, patch-ECGs may be better suited for prolonged rhythm monitoring during daily life activities.For image description, please refer to the figure legend and surrounding text. For image description, please refer to the figure legend and surrounding text.