Agreement of two different methods for measurement of heart rate variability

Abstract.

Background: The widespread use of affordable devices with sufficient precision for measurement of heart rate variability (HRV) might lead to early detection of abnormalities in a large number of high-risk patients and athletes. The purpose of this study was to determine the limits of agreement of two devices for measuring HRV parameters differing in price and assumed precision. Subjects and methods: 36 healthy subjects (22 men and 14 women) with a mean age of 27.4 (SD 11.1) years were included. The two devices used for comparison were PowerLab^® with Chart^® software as the reference golden standard, and Polar^® Transmitter^®/Advantage^® with Precision Performance^® software, respectively. Measurements included the following heart rate variability parameters: heart rate, range of R-R-interval duration, SDNN, rMSSD, total Power, VLF power, LF power, and HF power. Measurements were taken during metronomic respiration over a total period of 3 minutes. Statistical analysis was performed according to Bland and Altman and by means of scatterplots and Spearman correlation coefficients. Results: Good agreement was found for heart rate (95 % CI of limits of agreement: −0.7–0.6 bpm; r = 0.999), range of duration of R-R-intervals (95 % CI: −18.9–17.0 ms; r = 0.997), rMSSD (95 % CI: −1.5–2.5 ms; r = 0.999), and SDNN (95 % CI: −3.0–3.1 ms; r = 0.997). Correlation of measurements was high for the variables total Power, VLF power, LF power, and HF power. Analysis of method agreement for frequency domain variables was statistically not feasible. Conclusion: The level of agreement for the analyzed time domain variables between the reference golden standard and the inexpensive device is sufficient to permit initial screening by family doctors, and self-administration by high-risk patients and athletes.