P studies from all parts of the world have consistently reported that a low plasma level of high-density lipoprotein (HDL) cholesterol is a strong predictor of coronary heart disease (CHD). Multiple lines of evidence suggest that HDL has a direct beneficial effect on the arterial wall. Metabolically, HDL particles induce the removal of cholesterol from cells, including those in atherosclerotic plaques. Apolipoprotein A-I (apoA-I), the major protein in HDL, activates the mobilization of cholesterol ester stores in macrophages, leading to the reduction of the cholesterol content of this major cell type in atherosclerosis. At least 2 specific receptors participate in mediating this protective effect. Intravenous infusion of HDL in rabbits prevents atherosclerosis, and the introduction and expression of the human apoA-I gene in mice stimulates the regression of preexisting atherosclerosis. HDL appears to deliver cholesterol to the liver for excretion, and this function is termed, “reverse cholesterol transport.” HDL, in the form of recombinant human apoA-I liposomes, infused into hypercholesterolemic humans, produced net cholesterol excretion from the body, directly demonstrating the stimulation of reverse cholesterol transport. HDL also has other functions that may contribute to its ability to protect against CHD. Examples include anti-inflammatory and antioxidant properties. Experiments demonstrating the direct antiatherosclerotic potential of HDL strongly support the findings from epidemiology and clinical trials, that the apparent clinical benefit of raising HDL cholesterol is not secondary to its relation to other coronary risk factors, and justify renewed evaluation of HDL cholesterol as a target in the prevention and treatment of CHD. Many patients with low levels of HDL cholesterol also have high triglyceride levels, and elevated concentrations of highly atherogenic triglyceride-rich lipoprotein remnants (derived from chylomicrons and very low-density lipoproteins [VLDLs]). The PROCAM study demonstrates the high risk associated with this particular lipid profile. Such patients may have predominantly abdominal obesity, and have a high prevalence of insulin resistance, characterized by relatively high levels of fasting glucose and insulin. Increased VLDL production and impaired VLDL lipolysis link a high triglyceride level with a low HDL cholesterol concentration. A high triglyceride concentration may indicate a high concentration of atherogenic chylomicron and VLDL remnants, and these triglyceride-rich remnants, which possess apolipoprotein C-III, are associated with an increased risk for CHD. Insulin resistance causes increased VLDL production. This constellation of metabolically linked factors is often termed “the metabolic syndrome.” The relation between HDL cholesterol and the incidence of CHD is curvilinear and mirrors that of the corresponding curve for LDL cholesterol as seen in the Framingham Heart Study. Low HDL cholesterol predicts CHD in both sexes (Figure 1), but the strength of the relation may be greater in women than in men. A meta-analysis of 4 large prospective epidemiologic studies has defined the relation between differences in HDL cholesterol and shifts in cardiovascular risk. An increase in HDL cholesterol of 1 mg/dl (0.026 mmol/L) is equated with an independent relative risk reduction in the incidence of coronary events by 2% in men and 3% in women. In other words, depressed levels of HDL cholesterol herald increased risk for CHD at all levels of LDL cholesterol and triglycerides, in persons with diabetes and in nondiabetics, in men and women, and in persons with and without known CHD. The US National Health and Nutrition Examination Survey III (1988 to 1991) shows that HDL cholesterol levels 40 mg/dl ( 1 mmol/L) occur in about 35% of adult men and 15% of adult women. However, a higher percentage of men and women who develop CHD have HDL cholesterol 40 mg/dl. In the Framingham Heart Study, 57% of men who developed CHD had an HDL cholesterol 40 mg/dl, and in the Cholesterol and Recurrent Events (CARE) trial, about 40% of the women who had had a myocardial infarction had HDL cholesterol 40 mg/dl. Statin treatment reduces coronary events in patients whose HDL cholesterol concentrations are low. Nonetheless, the strong relation between baseline HDL cholesterol and subsequent event rates was present in statin-treated patients in 5 of the 6 large trials. For example, in the Scandinavian Simvastatin Survival Study (4S), and in the CARE and Long term Intervention with Pravastatin in IschFrom the Harvard School of Public Health, Boston, Massachusetts. Manuscript received January 14, 2002; revised manuscript received and accepted March 14, 2002. Address for reprints: Frank M. Sacks, MD, Harvard School of Public Health, 665 Huntington Ave, Boston, Massachusetts 02115. E-mail: fsacks@hsph.harvard.edu.