With data from the Lipid Research Clinics Prevalence Study used as a baseline, an average of 8.4 years of mortality follow-up has been completed for 8,825 male and female participants who were at least 30 years of age at the time of the study. The participants were examined at 10 North American sites during 1972-1976. Univariate findings for high density lipoprotein (HDL) cholesterol in the 7,569 persons initially free of coronary heart disease indicate an inverse relation with cardiovascular disease mortality for men and women. Controlling for age, low density lipoprotein (LDL) cholesterol, triglycerides, body mass index, systolic blood pressure, and smoking, multivariate analysis results indicate inverse relations between HDL cholesterol and cardiovascular disease mortality. In men, the cardiovascular disease mortality rate ratio for HDL cholesterol increments of 10 mg/dl was similar to the cardiovascular disease mortality rate ratio for LDL cholesterol increments of 30 mg/dl. In women, HDL cholesterol is more closely related to cardiovascular disease than is LDL cholesterol. HDL cholesterol is inversely related to both coronary heart disease and other cardiovascular disease mortality in both sexes, though the relation of HDL cholesterol with coronary heart disease mortality in women approached statistical significance only for the combined end point of suspect and definite cases. No relation was observed in either sex group between HDL cholesterol and noncardiovascular disease mortality. This study, only the second large prospective study evaluating the role of HDL cholesterol in women, confirms both the importance and the independence in both men and women of HDL cholesterol in the epidemiology of coronary heart disease and other cardiovascular disease mortality.