Most of the results from epidemiologic studies support the general idea that high density lipoproteins (HDL) cholesterol is inversely related to coronary heart disease (CHD) incidence. Results from the literature and from a large cohort study in Belgium (the BIRNH study) are used to describe the distribution and the major determinants of HDL cholesterol. HDL cholesterol is influenced by a variety of biologic, environmental and behavioral characteristics. Results of a 10-year mortality follow-up of the BIRNH study are presented and compared to those observed in other large cohort studies. The inverse relationship between HDL cholesterol and CHD is confirmed, although the strength of the association varies between studies and is weakened after adjustment for other coronary risk factors. The results from the BIRNH study also suggest that the relation between HDL cholesterol and CVD mortality is curvilinear. At present, only indirect evidence is available to support the idea that raising HDL cholesterol is useful in primary and secondary prevention of CHD.