Ambulatory Heart Rate Variability Monitoring: Comparisons Between the Empatica E4 Wristband and Holter Electrocardiogram

OBJECTIVE: Heart rate variability (HRV) is a useful index of psychological and physiological stress. Although several wristband devices have purported to measure HRV, none have demonstrated reliability when compared with the criterion-standard Holter monitor. We evaluated the reliability of HRV readings from the Empatica E4 wristband compared with a Holter monitor over a 24-hour period of simultaneous monitoring. METHODS: Agreement between the monitors was assessed by examining correlations and intraclass correlations (ICCs) for fixed sets in 13 individuals in a treatment trial for posttraumatic stress disorder (4 women; mean [standard deviation] age = 51.92 [6.17] years). Agreement was calculated at 1-second and 5-minute intervals for interbeat intervals (IBIs) and for 5-minute intervals of the root mean square of successive differences between normal heartbeats (RMSSD) and standard deviation of all normal R-R intervals (SDNN). Agreement across the entire 24-hour observation period was also measured. Frequency-domain measures of HRV could not be calculated because of too much missing data from the E4. RESULTS: Although high interdevice correlations and ICCs were observed between the E4 and Holter monitors for IBIs at 1-second (median r = 0.88; median ICC = 0.87) and 5-minute (median r = 0.94; median ICC = 0.94) intervals, reliabilities for 5-minute RMSSD (median r = -0.09; median ICC = -0.05) and 5-minute SDNN (median r = 0.48; median ICC = 0.47) were poor. Agreement between the devices on 24-hour measures of HRV was satisfactory (IBI: r = 0.97, ICC = 0.97; RMSSD: r = 0.77, IBI = 0.76; SDNN: r = 0.92, IBI = 0.89). CONCLUSIONS: Findings suggest that the low reliability of Empatica E4 as compared with the Holter monitor does not justify its use in ambulatory research for the measurement of HRV over time periods of 5 minutes or less.