Problems in measuring heart rate variability of patients with congestive heart failure

Abstract Heart rate variability (HRV) has become an important noninvasive measure of the integrity of the autonomic nervous system in various disease states. The power spectrum of HRV is a means to separate the instability oscillations of the various feedback mechanisms that contribute to cardiovascular homeostasis. The reliability of HRV data is largely unexplored. The day-to-day correlations in the low and mid-frequency components of HRV spectra average 91%, and that of the high-frequency component averages 81%. The correlations among spectral and nonspectral measures of HRV (SD) for the same data segment average 50–60%, suggesting that they encode similar information. Heart rate variability spectra exhibit diurnal variation consistent with physiologic expectation: respiratory sinus arrhythmia (thought to be mediated by parasympathetic tone) and to a lesser extent, the low-frequency spectral component (thought to be of mixed sympathetic-parasympathetic origin) are higher at night than in the daytime; the mid-frequency component (associated with the baroreflex, which is more excited when the patient is upright) is slightly higher during the daytime. Increased frequency of ectopic beats, such as occurs in congestive heart failure, reduces the reliability of the power spectrum since the number of usable data segments falls off rapidly with even small increases in rate of ectopy, and the variance of the estimate (in the method of averaged periodograms) is inversely proportional to the square root of the number of data segments. Using shorter data segments increases the number of segments available, but reduces resolution. Interpolation over ectopic beats (by either linear or cubic splint interpolation) increases the apparent power in low frequencies. Recommendations are made as to the design of HRV-based clinical and experimental protocols.