OBJECTIVE: Elevation of blood glucose is a common metabolic disorder among patients with acute myocardial infarction (AMI) and is associated with adverse prognosis. However, few data are available concerning the long-term prognostic value of elevated fasting glucose during the acute phase of infarction. RESEARCH DESIGN AND METHODS: We prospectively studied the relationship between fasting glucose and long-term mortality in patients with AMI. Fasting glucose was determined after an >/=8 h fast within 24 h of admission. The median duration of follow-up was 24 months (range 6-48). All multivariable Cox models were adjusted for the Global Registry of Acute Coronary Events (GRACE) risk score. RESULTS: In nondiabetic patients (n = 1,101), compared with patients with normal fasting glucose (<100 mg/dl), the adjusted hazard ratio for mortality progressively increased with higher tertiles of elevated fasting glucose (first tertile 1.5 [95% CI 0.8-2.9], P = 0.19; second tertile 3.2 [1.9-5.5], P < 0.0001; third tertile 5.7 [3.5-9.3], P < 0.0001). The c statistic of the model containing the GRACE risk score increased when fasting glucose data were added (0.8 +/- 0.02-0.85 +/- 0.02, P = 0.004). Fasting glucose remained an independent predictor of mortality after further adjustment for ejection fraction. Elevated fasting glucose did not predict mortality in patients with diabetes (n = 462). CONCLUSIONS: Fasting glucose is a simple robust tool for predicting long-term mortality in nondiabetic patients with AMI. Fasting glucose provides incremental prognostic information when added to the GRACE risk score and left ventricular ejection fraction. Fasting glucose is not a useful prognostic marker in patients with diabetes.