OBJECTIVE: The associations between glucose intolerance measured at the study entry date and the 23-year incidence of coronary heart disease (CHD), CHD mortality, and total mortality were examined at the Honolulu Heart Program. RESEARCH DESIGN AND METHODS: This prospective study followed a cohort of 8,006 Japanese-American men who were 45-68 years old and living on the island of Oahu, HI, in 1965. Baseline glucose was measured in a nonfasting state 1 h after a 50-g glucose load. History and use of medication for diabetes was obtained during an interview. The cohort was divided into four categories of glucose tolerance: low-normal, high-normal, asymptomatic hyperglycemia, and known diabetes. RESULTS: During the 23 years of follow-up, 864 incident cases of CHD, 384 deaths from CHD, and 2,166 total deaths occurred. The relative risks (RRs) were obtained using Cox proportional hazards modeling, with the low-normal category as a reference. The RRs were adjusted for age only, as well as for age, BMI, hypertension, cholesterol, triglycerides, smoking, alcohol, and a Japanese diet index. The age-adjusted and risk factor-adjusted RRs for all outcomes were significant for the asymptomatic hyperglycemic and known diabetes groups (P<0.05). The age-adjusted RRs for CHD incidence and total mortality were marginally significant in the high-normal group, but the RRs were not significant when adjusted for risk factors. CONCLUSIONS: These results suggest a dose-response relation of glucose intolerance at baseline with CHD incidence, CHD mortality, and total mortality, independent of other risk factors, in this cohort of middle-aged and older Japanese-American men.