The health effects of alcohol consumption remain complex for several reasons: the risks and benefits accrue over many years, assessment of drinking is generally based on self-report, drinking habits change over time and studies estimate average daily drinking disregarding how or when the beverage was consumed. In addition, alcohol consumption is associated with lifestyle factors which may confound relationship with disease. Despite these methodological difficulties, epidemiological studies are surprisingly consistent, showing that light to moderate intake is associated with a lower risk of total mortality compared with those who drink heavily or do not drink at all. Thus there is a J-shaped association of alcohol intake with risk of total mortality whose basis is likely to be the effect of summing the cause-specific effects at the various drinking levels. Studies using a diversity of methods and populations have consistently reported an inverse relationship between coronary heart disease and light to moderate drinking, with the depth and width of the J-shaped mortality curve depending on the underlying risk of coronary heart disease for that population. The higher risk of death at heavy drinking levels is due to increased risk of cancer, liver diseases, cardiomyopathy and stroke. The precise mechanisms behind these effects of alcohol are only now beginning to be understood. The most plausible mechanism by which alcohol reduces the risk of coronary heart disease is by its effects on blood lipids, particularly increases in high density lipoprotein (HDL) cholesterol: about 50% of the risk reduction attributable to alcohol is explained by changes in total HDL. Further support for the HDL hypothesis comes from the lack of a differential effect of alcohol by beverage type. While the association of alcohol and cardiovascular disease is likely to be causal, any public health recommendations must consider the complexity of alcohol's metabolic, physiological and psychological effects.