Prognostic value and physiological correlates of heart rate variability in chronic severe mitral regurgitation

A variety of measures of heart rate variability have been devised to measure high-frequency (0.15-0.40 Hz), low-frequency (0.04-0.15 Hz), or ultralow-frequency (< 0.0033 Hz) fluctuations in sinus cycle length. Although measures of low-frequency and ultralow-frequency heart rate variability have been shown to correlate with prognosis in several populations with ischemic heart disease, their relevance to patients with primary valvular heart disease remains to be determined. Thirty-eight patients with nonischemic causes of chronic severe mitral regurgitation who were in sinus rhythm underwent 24-hour ambulatory electrocardiography as part of a prospective study of the natural history of regurgitant valvular heart disease. Patients were followed for as long as 9.2 years, and end points of mortality, progression to mitral valve surgery, and development of chronic atrial fibrillation were tabulated. Time- and frequency-domain measurements of high-frequency, low-frequency, and ultralow-frequency heart rate variability were computed and compared with resting ventricular function by radionuclide cineangiography and outcome. The standard deviation of the 5-minute mean RR intervals (SDANN), a measure of ultralow-frequency heart rate variability, was correlated with left ventricular ejection fraction (r = 0.49, p = 0.002) and right ventricular ejection fraction (r = 0.43, p = 0.007), whereas low-frequency and high-frequency heart rate variabilities were not. Heart rate, ultralow-frequency heart rate variability, and, to a lesser extent, high-frequency heart rate variability exhibited significant diurnal variation, but low-frequency heart rate variability did not. Heart rate and ultralow-frequency, low-frequency, and combined low- and high-frequency heart rate variability predicted mortality and total events. The most powerful predictor of subsequent events was SDANN: Patients with reduced SDANN were significantly more likely to develop end-point events (p < 0.001) with increased progression to mitral valve surgery (p < 0.001) as well as increased early mortality (p = 0.02). In a multivariate proportional hazards model, SDANN retained independent predictive power (p = 0.001). Likewise, SDANN was the only variable that was significantly associated with the subsequent development of atrial fibrillation (relative risk, 3.1; p = 0.03). Ultralow-frequency heart rate variability, as measured by SDANN, correlates with right and left ventricular performance and predicts develo