Abstract 3631: Elevated In-hospital Glucose Level, and not Known History of Diabetes Mellitus, Predicts 30-day Mortality Following Acute ST-elevation Myocardial Infarction

Introduction: According to clinical risk assessment guidelines, a history of diabetes mellitus (DM) portends poor outcomes following acute MI. Elevated in-hospital glucose levels also predict early mortality in acute MI patients, but the degree to which glucose levels and diabetic history independently predict post-MI mortality is unclear. Methods and Hypothesis: We analyzed data from the combined cohort of the CREATE-ECLA and OASIS-6 randomized trials that evaluated the impact of glucose-insulin-potassium (GIK) infusion versus no infusion on 30-day mortality in 22,943 patients hospitalized with acute ST-elevation MI. We calculated the average in-hospital glucose level for each patient (mean of the admission, 6-hour, and 24-hour glucose levels). Logistic regression was performed to determine whether average glucose level and history of DM remained significant mortality predictors after adjusting for age, sex, and GIK allocation. Results: Glucose data were recorded in 22,860 (99.6%) patients; 10,050 (44%) had an average in-hospital glucose level ≥ 8 mmol/L (144 mg/dL), of whom 65% did not have known prior DM. Among patients with glucose >8 mmol/L, 30-day mortality rates were similar in patients with and without known DM (Figure ). In-hospital glucose, but not history of DM, was a significant multivariable predictor of mortality (Table). Conclusions: By considering only history of DM and not in-hospital glucose levels, risk assessment guidelines for acute MI overlook a large proportion of patients at high risk for early death. Therefore, clinicians should emphasize elevated glucose levels in addition to history of DM as a risk marker in patients with acute MI.