Comparison of heart rate variability in survivors and nonsurvivors of sudden cardiac arrest

Abstract

Imbalances in autonomic nervous system function have been posed as a possible mechanism that produces ventricular fibrillation and sudden cardiac arrest in patients with cardiovascular disease. Heart rate (HR) variability is described in survivors and nonsurvivors of sudden cardiac arrest within 48 hours after resuscitation using time and frequency domain analytic approaches. HR data were collected using 24-hour ambulatory electrocardiograms in 16 survivors and 5 nonsurvivors of sudden cardiac arrest, and 5 control subjects. Survivors of sudden cardiac arrest were followed for 1 year, with recurrent cardiac events occuring in 4 patients who died within that year. Analysis of 24-hour electrocardiograms demonstrated that control subjects had the highest HR variability (standard deviation of all RR intervals = 155.2 ± 54 ms), with nonsurvivors demonstrating the lowest HR variability (standard deviation of all RR intervals = 52.3 ± 6.1 ms) and survivors of sudden cardiac arrest falling between the other 2 groups (standard deviation of all RR intervals = 78 ± 25.5 ms, p ≤0.0000). Two other indexes of HR variability (mean number of beat to beat differences in RR intervals >50 ms/hour and root-mean-square of successive differences in RR intervals) did not demonstrate the expected pattern in this sample, indicating that perhaps patterns of HR variability differ between groups of patients with cardiovascular disorders. Spectral analytic methods demonstrated that survivors of sudden cardiac arrest had reduced low- and high-frequency spectral power, whereas nonsurvivors demonstrated a loss of both low- and high-frequency spectral power. Left ventricular ejection fraction was significantly correlated with standard deviation of all RR intervals and standard deviation of the mean of RR intervals for 5-minute segments. One-year mortality in the survivor group of cardiac arrest was inversely related to several variables including age, standard deviation of all RR intervals, and low frequency power.