OBJECTIVE: To determine whether hyperglycemia in the acute stroke period is associated with worse survival and functional outcome after accounting for acute stress response and chronic hyperglycemia. DESIGN: Prospective, county-wide, multicenter cohort study. SETTING: A community hospital, a university hospital, and a Veterans Affairs hospital. PATIENTS: A cohort of 146 patients hospitalized with new atherothrombotic stroke. MEASUREMENTS: Admission blood glucose concentration, demographic characteristics of patients, 24-hour urinary catecholamine, serum cortisol, and glycosylated hemoglobin levels; outcomes included mortality and functional outcome (Barthel index and Fugl-Meyer score) at 5, 30, 90, and 180 days after stroke. RESULTS: Of the 996 patients with possible acute stroke who were screened, 146 (15%) were eligible for and consented to participate in the study; in most cases, exclusion from study was based on the absence of acute, atherothrombotic stroke. Overall, no evidence was found of a significant univariate association between admission blood glucose level and survival (relative risk, 1.02; 95% Cl, 0.94 to 1.09) or functional outcome (univariate regression coefficient for adjusted Fugl-Meyer score at day 30, - 0.36; Cl, - 1.08 to 0.27). This absence of an association persisted after adjustment for significant predictors of outcome in a multivariate model. CONCLUSIONS: These data do not support an association between level of glycemia and outcome from acute stroke.