The high mortality of heart failure is associated with hemodynamic abnormalities, depressed cardiac function, and reduced exercise capacity. That these factors can be modified by drug treatment is of potential prognostic significance. Hemodynamic variables are related to survival, and long-term prognosis is better in patients with only midly abnormal cardiac output or ventricular filling pressures. Indexes of left ventricular function such as ejection or shortening fraction tend to be higher in patients who survive for longer periods. The relation between exercise capacity and survival, however, is unclear. Those patients with severe exercise intolerance (maximal oxygen uptake below 10 ml/min/kg) or with severe symptoms are at great risk of dying. However, exercise capacity and functional class are not related to prognosis when all classes of patients are considered together, especially if class IV patients are excluded. Most of the available data derive from retrospective analyses of trials involving heterogeneous patient populations and aimed at improving left ventricular performance or functional capacity. Large prospective trials aimed primarily at affecting mortality in a broad spectrum of patients are needed to learn more about determinants of survival in heart failure.