Much remains to be clarified in the apparently protective effect of moderate alcohol use against coronary heart disease risk. Insulin levels are positively associated with coronary heart disease risk, so recent reports of decreased insulin sensitivity among nondrinkers and lower fasting insulin levels with increasing alcohol intake suggest the possibility that insulin may play a role. Between 1987 and 1991, the authors examined fasting insulin concentrations and the empiric fasting insulin resistance index in relation to reported alcohol intake (mean, 15.3 g/day; standard deviation, 19.6; range, 0–120.6) and potential confounders. The latter included age, obesity, fat distribution, smoking, energy, saturated fat intake, antihypertensive medication, and physical activity. Participants in this cross-sectional analysis were 938 nondiabetic men from the Boston, Massachusetts, area who were part of the Normative Aging Study. Unadjusted fasting insulin levels were significantly different (p = 0.008) between categories of alcohol intake, as were fasting insulin resistance index values (p = 0.01). After adjustment for potential confounders, analysis revealed that subjects consuming moderate amounts of alcohol had the lowest fasting insulin and fasting insulin resistance index values. Compared with values from moderate drinkers, fasting insulin resistance index values were higher in those subjects reporting no alcohol intake (p = 0.011), low intake (p = 0.004), and high intake (p = 0.04). A similar pattern was observed for fasting insulin values. Among this sample of nondiabetic men, moderate drinkers had the lowest levels of fasting insulin resistance index and fasting insulin, consistent with lower levels of insulin resistance and thus lower risk for coronary heart disease. These findings suggest the possibility that the coronary heart disease-protective effects of moderate alcohol use are at least partially mediated by insulin. Am J Epidemiol 1997; 145: 909–16.