It is still controversial whether fasting blood glucose (FBG) is associated with the risk of vascular outcomes among nondiabetic subjects. We sought to determine whether FBG is associated with vascular outcomes and whether this association differs among various racial or ethnic groups. In the Northern Manhattan Study, a total of 2,372 subjects (mean age 68.8 +/- 10.7 years, 36% men) without a history of diabetes mellitus, stroke, or myocardial infarction (MI) were followed for an average of 7.5 years for ischemic stroke, MI, and combined vascular events defined as either ischemic stroke, MI, or vascular death. Cox proportional-hazards models were used to calculate hazard ratios and 95% confidence intervals (CIs) of FBG-associated risk for vascular outcomes after adjusting for age, gender, race/ethnicity, education, body mass index, hypertension, current smoking, previous coronary artery disease, low-density lipoprotein cholesterol, alcohol intake, and physical activity. The incidences of MI, ischemic stroke, and combined vascular events were 5.5, 6.3, and 20.0 per 1,000 person-years, respectively. Each SD increase of FBG (27 mg/dl) was associated with statistically significantly increased risks of combined vascular events (hazard ratio 1.20, 95% CI 1.09 to 1.31) and MI (hazard ratio 1.21, 95% CI 1.02 to 1.44), but the effect was weaker, evident for ischemic stroke (hazard ratio 1.13, 95% CI 0.95 to 1.34). FBG was significantly associated with incident ischemic stroke among African-American subjects (hazard ratio 1.38, 95% CI 1.09 to 1.74) and incident MI among Hispanic subjects (hazard ratio 1.24, 95% CI 0.99 to 1.55). In conclusion, FBG was an independent predictor for vascular outcomes among individuals without history of diabetes from this multiethnic cohort. The effects were more apparent for MI than for ischemic stroke; however, FBG was a strong predictor of ischemic stroke among African-American subjects.