Comparison of novel wearable cardiac monitors to yield accurate exercising ECG and heart rate data in horses

Abstract

Exercise associated deaths (EADs) in horses are a problem for the equestrian industry. Sudden death (SD) is responsible for approximately 20% of EADs. The underlying cause of SD is suspected to be cardiovascular disease but often cannot be determined post-mortem. User-friendly cardiac monitors are needed for large scale investigations of arrhythmias associated with SD in horses. We hypothesised that novel wearable devices would provide exercising electrocardiograms (ECGs) of sufficient diagnostic quality for this purpose. Diagnostic quality of ECGs generated by two wearable devices (W2nd^™ and Polar Equine^™) were compared to simultaneous recordings with a telemetry unit (Televet^™) in 5 Thoroughbreds completing 43 separate submaximal exercise tests on a high-speed treadmill. Maximal heart rate (HR_max) generated by mobile applications (HR_maxapp), HR_max after manual correction (HR_maxcorr), percentage of diagnostic ECGs (%diag) at the gallop, and overall quality assessed by visual analogue scale (VAS) were assessed by a blinded observer. HR_maxcorr did not differ significantly between groups. HR_maxapp was significantly lower for W2nd (166.8/min, 95% confidence interval (CI): 160.5-173.1/min) but did not differ significantly between Televet (178.8/min 95% CI: 165.8-191.1/min) and Polar (181.3/min, 95% CI: 174.5-188.1/min). HR_maxcorr was accurate and precise in all runs. HR_maxapp was within a priori limits of agreement in 16/23 W2nd and 18/19 Polar recordings. %diag was significantly lower (77.1%, 95% CI: 67.4-86.8) for W2nd than Polar (100%, 95% CI: 89.9-110.3). VAS was lower for W2nd (46.2, 95% CI: 35.5-57.0) than Polar (90.6, 25% CI: 79.4-101.9). In conclusion, wearable devices appear to be promising tools for investigation of equine exercising arrhythmias in large-scale studies.