| Abstract: | Background: In heart rate variability (HRV) studies, the duration of the electrocardiographic recordings may vary between 20 seconds and 24 hours. Little is known about the reproducibility or comparability of HRV measures calculated over observation periods of different duration. Methods: To assess the reproducibility and comparability of HRV measures computed from short recordings under standardized conditions and from ambulatory 24-hour recordings, 15 healthy young males were studied on three occasions. Results: The reliability coefficient (between subject variation divided by summed between and within subject variation) for 20-second intervals was low for all measures. For 5-minute intervals in supine position, but not standing position, the reliability coefficients of the standard deviation, root of mean squared successive differences, proportion of successive differences > 50 ms, and the proportion low frequency power were about 70%. The reliability coefficients of the 24-hour HRV measures were > 80% for all measures. Short- and long-term measures of heart rate and HRV appeared to be correlated, implying that these measures result in similar ranking of subjects. Conclusions: The results indicate that HRV measures based on 24-hour monitoring during regular activities are better reproducible than short-term measures under standardized conditions. A single 5-minute recording may suffice to characterize a group of persons in a population study, but, because of considerable day-to-day variation, it is not adequate for individual characterization. |
