The relationship between alcohol consumption and three cardiovascular diseases (ischemic stroke, hemorrhagic stroke, and heart failure) was examined in a population sample of middle-aged U.S. adults who were participants in the Atherosclerosis Risk in Communities (ARIC) cohort. Alcohol intake was selfreported at two time periods. Stroke events were ascertained by contacting study participants and reviewing hospitalization and death certificate data. Heart failure (HF) events were identified from hospital discharge diagnoses and death certificates. To assess the association between alcohol consumption and each cardiovascular disease outcome separately, age-adjusted incidence rates by level of alcohol intake were calculated. Alcohol intake was categorized based on American Heart Association guidelines. Poisson regression was used to estimate incidence rate ratios (RRs) comparing levels of alcohol intake with never drinkers. Results were stratified by race. The crude RRs for all stroke showed an inverse association for occasional drinkers (RR=0.57, 95%CI:0.42-0.77) and light/moderate drinkers (RR=0.74, 95%CI:0.59-0.94) as compared with never drinkers. After adjustment for age, race, sex, and socioeconomic status the RRs were attenuated (comparing occasional with never drinkers RR=0.91, 95%CI:0.67-1.25 and light/moderate with never drinkers iv (RR=0.99, 95%CI:0.77-1.28). Results for ischemic stroke were similar to those for all stroke. The crude RRs for hemorrhagic stroke were 1.16 (95%CI:0.51-2.63) for former drinkers, 1.13 (95%CI:0.48-2.69) for occasional drinkers, 0.74 (95%CI:0.32- 1.71) for light/moderate drinkers, and 1.67 (95%CI:0.66-4.25) for heavy drinkers, as compared with never drinkers. Adjustment for age, race, sex, and socioeconomic status resulted in an increase in RRs. The RRs comparing each level of alcohol intake with never drinkers for HF incidence were 1.12 (95%CI:0.95-1.31) for former drinkers, 0.67 (95%CI:0.54-0.82) for occasional drinkers, 0.65 (95%CI:0.54-0.78) for light/moderate drinkers, and 0.75 (95%CI:0.59-0.95) for heavy drinkers. Similar patterns were seen for blacks and whites. We found no compelling evidence that occasional or light/moderate alcohol intake reduces stroke incidence rates. Adjusted RRs suggest that any level of alcohol intake increases hemorrhagic stroke incidence rates. We found a positive association for former drinking and an inverse association for current drinking for HF incidence. While the association between former drinkers and HF incidence was evident for whites, the evidence among blacks was less strong.