Alcohol consumption is associated with a reduced risk of coronary heart disease (CHD) but an increased risk of other causes of morbidity and mortality. It remains unclear whether there is an upper limit to a protective effect of alcohol intake on CHD risk. Whether there is a U- or an L-shaped relation between alcohol consumption and CHD incidence (hospitalization and mortality due to ischemic heart disease: International Classification of Diseases codes 410-414) is examined using the National Health and Nutrition Examination Survey I. Baseline data were collected in 1971-1975. Follow-up data through 1987 (14.6 years mean follow-up) were analyzed for 6,788 European-American males (n = 2,960) and females (n = 3,828) aged 40-75 years at baseline. Cox regression was used to assess the association between alcohol consumption and incidence of CHD. For females, an increased risk was found above 28 drinks per week relative to abstainers (relative risk = 2.6, 95% confidence interval 1.2-5.5), which was significant, but was based on small numbers. For males, no upturn in risk was found at higher intake. Mortality data supported these results. Sex differences should be explored further, since they are relevant to understanding causal mechanisms and public policy and prevention.