Background In advanced chronic heart failure (CHF), correlation between heart rate variability (HRV) and parameters of disease severity is still unclear. A reduced HRV has been related to left but not to right ventricular function parameters. Moreover, the prognostic role of spectral measures is not fully defined. We sought to assess HRV by using a short electrocardiographic recording in ambulatory patients with severe CHF to investigate the relation of HRV with clinical neurohormonal and hemodynamic parameters and to determine its predictive prognostic power. Methods and Results HRV was obtained from 5-minute electrocardiographic recordings in 75 ambulatory patients with CHF referred for heart transplantation screening. Standard frequency-domain parameters (total power, low-frequency power, and high-frequency power) were calculated. Prognostic value of these autonomic markers and their correlation with clinical and instrumental parameters were also assessed. A low low-frequency/high-frequency ratio was an independent predictor of cardiac events (P =.015). No correlation was found between New York Heart Association class and HRV, whereas significant correlations were identified between norepinephrine plasma levels, several hemodynamic parameters, and spectral measures (P ≤.03). A reduced HRV, particularly a low-frequency power reduction (P =.000), was highly related to indexes of right ventricular dysfunction. Conclusions The current data indicate that spectral analysis of HRV, calculated from short electrocardiographic recordings, may represent a simple but effective means contributing to risk stratification of patients with severe CHF. Autonomic information obtained from this analysis suggests that right ventricular dysfunction may be a critical element determining autonomic imbalance in patients with severe CHF. (Am Heart J 2000;139:1088-95.)