Considerable progress has been made in recent years in our understanding of atherogenesis and, in particular, of how it relates to lipoprotein metabolism. In this conference, we attempt to re-evaluate the data on the relation between alcohol intake and coronary heart disease, emphasizing the effects of alcohol on lipoprotein metabolism. Epidemiologic data generally show an inverse correlation between coronary heart disease risk and moderate alcohol intake (variously defined but generally corresponding to 2 to 4 drinks per day). The potentially drastic effects of excessive alcohol intake on health, however, preclude any recommendation that patients increase their alcohol consumption. Equally, there may be no basis for proscribing moderate alcohol intake but, even here, the data are far from complete. The mechanism by which moderate alcohol intake "protects" remains unclear. Perhaps the best available hypothesis relates to the increased concentration of high-density lipoprotein (HDL) cholesterol associated with moderate alcohol intake. However, it should be stressed that we are still uncertain about the mechanisms linking a high HDL level to protection against coronary heart disease. If a high HDL level is only a marker (and not directly protective), raising HDL levels need not confer protection. Alcoholism, on the other hand, is associated with marked elevation of triglyceride-rich lipoproteins and is among the most common causes of hypertriglyceridemia. Moreover, once hepatic damage occurs, plasma HDL levels may actually be lower than normal. The determining point is that high alcohol intakes are associated with increased overall mortality. Until we know more about the metabolic and behavioral effects of alcohol and about its linkage to atherosclerosis, we have no basis for recommending either that patients increase their alcohol intake or that they start to drink if they do not already.