006 Admission glycemia in acute myocardial infarction: incremental prognostic value for mortality over GRACE risk score and left ventricular ejection fraction. Data from the RICO survey

Abstract

High blood glucose (HBG) on admission is a major common metabolic disorder in patients with acute myocardial infarction (MI) and is associated with worse prognosis. However, only few data have examined its predictive performance over established risk score. From a French regional survey for acute MI, we analyzed the relationship between HBG, as defined by admission glycemia >11 mmol/L, and 1 year mortality in patients with acute MI. All multivariate Cox models were adjusted for the Global Registry of Acute Coronary Events (GRACE) risk score, which is a validated 9 variables prediction tool, and left ventricular ejection fraction as assessed by echocardiography <3 days after admission. The additional prognostic information of HBG was tested by comparing the –2log likelihood of the Cox models with vs without HBG (χ 2 ). In the study population (n=3358), both admission glycemia as a continuous variable and HBG were univariately associated with increased mortality (HR(95%CI: 1.06(1.05-1.07) and 2.67(2.17-3.29), respectively). The addition of either admission glycemia as a continuous variable (HR(95%CI): 1.04(1.01-1.06) or of HBG (HR(95%CI): 1.61(1.28-2.03)) significantly improved the risk prediction in the multivariate model (χ 2 : p<0.001). However, in diabetic patients (n=756), HBG failed to independently predict mortality (HR(95%CI):1.17(0.80-1.71)). In contrast, in non diabetic patients (n=2592), HBG remained an independent predictor of death (HR(95%CI): 1.93(1.39-2.67)) and added incremental prognostic value in the model over the GRACE risk score and LVEF (χ 2 : p<0.001). High blood glucose on admission provides incremental prognostic information over established risk score and LVEF, in particular in non diabetic patients. Admission glycemia is not an independent predictive marker in diabetic patients.