Twenty-five patients (aged 62 +/- 2 years) with stable, moderate to severe ischemic congestive heart failure (CHF) (New York Heart Association class II/III: 15/10; ejection fraction 21.6 +/- 2%; and peak oxygen uptake 13.6 +/- 0.7 ml/kg/min) were studied to evaluate the ability of different methods to characterize autonomic tone in chronic CHF. Sympathovagal balance was assessed by: (1) heart rate variability in the time domain, assessed by the SD of RR intervals; (2) heart rate variability in the frequency domain, assessed by low- (0.03 to 0.14 Hz) and high- (0.18 to 0.40 Hz) frequency components of heart rate variability by autoregressive power spectral analysis; (3) 24-hour, daytime and nighttime heart rate; (4) submaximal heart rate during upright bicycle exercise, with respiratory gas analysis to obtain peak oxygen uptake; and (5) radiolabeled norepinephrine spillover. These methods did not correlate, with the exception of day and nighttime heart rate (r = 0.74; p < 0.001) and the expected inverse correlation between low and high frequency (r = -0.92; p < 0.001). No method correlated significantly with peak oxygen uptake, exercise tolerance or ejection fraction. After 8 weeks of physical training at home, all methods showed improvement in autonomic balance: increases in SD of RR intervals (+21%; p < 0.02) and high frequency (+41%; p < 0.007), and decreases in low frequency (-19%; p < 0.002), low-/high-frequency ratio (-48%; p < 0.03), norepinephrine spillover (-28.9%; p < 0.03), 24-hour heart rate (-2.7%; p < 0.005) and submaximal heart rate (-10.8%; p < 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)