AGREEMENT BETWEEN A WRISTBANDWEARABLEDEVICE AND ELECTROCARDIOGRAPHY FOR ULTRA-SHORT HEART RATE VARIABILITY ASSESSMENT

PURPOSE: Wristbands enable remote assessment of resting heart rate (RHR) and HR variability (HRV), which enable cardiac-autonomic status monitoring in patients, athletes, and research participants. The Kairos device (Biostrap USA LLC, Los Angeles, California) offers manual sampling for rapid assessment and feedback, but validation studies are lacking. Thus, we aimed to evaluate the agreement between the Kairos wristband and electrocardiography (ECG) for determining ultra-short RHR and HRV METHODS: Data from 32 young adults (34% male, 21.7 ± 3.2 years, 169.5 ± 10.3 cm, 76.0 ± 14.9 kg)met inclusion criteria for analysis (0 ectopic beats on ECG). Simultaneous 2-min ECG (1000 Hz) and Kairos samples were obtained in the supine and then seated position, each preceded by 1 min for stabilization. Kairos-derived RHR and root-mean square of successive differences (RMSSD) were automatically computed via the accompanying mobile application and retrieved from a downloadable report. ECG-derived values were computed via specialized software. Comparison statistics included paired t-tests and Bland-Altman analyses. Relative and absolute agreement were assessed with Lin’s concordance correlation coefficient (CCC) and the coefficient of variation (CV), respectively RESULTS: Supine and seated RHR for ECG (71.1 ± 9.7 and 75.8 ± 9.7 beats·min-1, respectively) and Kairos (71.1 ± 9.5 and 75.9 ± 9.7 beats·min-1, respectively) demonstrated minimal mean biases (0.0 ± 0.7 and 0.1 ± 0.6 beats·min-1, respectively, Ps ≥ 0.53), tight 95% limits of agreement (-1.3 - 1.3 and -1.1 - 1.2 beats·min-1, respectively), and strong relative (CCC ≥ 0.99) and absolute agreement (CV ≤ 0.7%). Comparatively, supine and seated RMSSD for ECG (56.3 ± 28.2 and 45.1 ± 29.7 ms, respectively) and Kairos (53.3 ± 22.3 and 46.3 ± 24.6 ms, respectively) demonstrated greater mean biases (-2.9 ± 7.7 and 1.2 ± 7.8 ms, respectively, P = 0.04 and 0.41, respectively), wider 95% limits of agreement (-18.1 - 12.3 and -14.2 - 16.6 ms, respectively), and weaker relative (CCC ≥ 0.95) and absolute agreement (CV ≤ 13.6%). CONCLUSIONS: The Kairos wristband provided valid assessment of RHR during 2-min sampling. The error displayed for HRV can be used to improve RMSSD assessment in future software updates.