This paper reviews the epidemiological evidence on the association of alcohol consumption with the major cardiovascular diseases (hypertension, stroke and coronary heart disease), and all causes of death. The focus is on light and moderate consumption and several important methodological issues are apparent with the epidemiological evidence on alcohol and mortality. The epidemiological data justify the following recommendations on alcohol consumption. The evidence does not support the unqualified claim that light and moderate drinking confers overall health benefits. However, in persons over 35 years of age, there is no consistent evidence that daily consumption of up to 2-3 drinks in men or up to 1-2 drinks in women increases the risk of dying. Non-drinkers should not be encouraged to change their drinking status. The consumption of more than 2-3 drinks per day in men and more than 1-2 drinks per day in women should be actively discouraged. Further research on the effects of light and moderate alcohol consumption on cardiovascular disease and all causes of death are required, particularly in young people, women and the elderly.