A three-decade examination of the prevalence, incidence, secular trends, and prognosis of cardiac failure in the Framingham Study provides insights into its epidemiology. Annual incidence of CHF is observed to increase from 3 to 1000 at ages 35-64, to 10 per 1000 at ages 65-94. There is a slight male predominance, owing to a higher rate of coronary disease, which conferred a fourfold risk of cardiac failure. Most cardiac failure is on the basis of long-standing hypertension or CHD. Silent infarctions were as predisposing for CHF as symptomatic MIs surviving 1 year. Hypertension is a major predisposing factor that at least triples the CHF risk, the systolic component being more predictive than the diastolic component. Correctable predisposing risk factors for CHF include: elevated blood pressure, impaired glucose tolerance, elevated cholesterol, low HDL-cholesterol, obesity, and a high hematocrit. Risk factors reflecting deteriorating cardiac function also were highly predictive, including: an enlarged heart, poor vital capacity, sinus tachycardia, and ECG-LVH. Commonly encountered ECG abnormalities such as intraventricular block, nonspecific repolarization abnormality, and ECG-LVH are all associated with a substantial risk of CHF. ECG-LVH carries a higher risk than x-ray enlargement. Sudden death was a common feature with CHF, occurring at 5 times the general population rate, even excluding those with overt CHD. Using the standard cardiovascular risk factors (age, systolic blood pressure, cholesterol, glucose, cigarettes, and ECG-LVH) jointly, it is possible to identify one tenth of the population from which 40% of CHF events evolve, in the absence of interim CHD or RHD.