Hyperglycemia as a predictor of in-hospital mortality in elderly patients without diabetes mellitus admitted to a sub-intensive care unit

OBJECTIVES: To investigate the association between hyperglycemia and in-hospital and 45-day mortality in acutely ill elderly patients. DESIGN: Retrospective cohort. SETTING: Hospital medical patients admitted to a sub-intensive care unit (sub-ICU) for elderly patients, which is a level of care between ordinary wards and intensive care. PARTICIPANTS: One thousand two hundred twenty-nine patients (mean age 79.6 +/- 8.4) admitted to the sub-ICU from January 2003 to January 2006. Forty patients with acute myocardial infarction and 34 patients with extreme fasting glucose values (< 60 or > 500 mg/dL) were excluded. Eight hundred twenty-two patients without a history of diabetes mellitus (DM) and 333 patients with a diagnosis of DM were selected and subdivided into three categories according to serum fasting blood glucose: 60 to 126 mg/dL (Group A), 127 to 180 mg/dL (Group B), and 181 to 500 mg/dL (Group C). MEASUREMENTS: Age, sex, mental and functional status, Acute Physiology Score, comorbid conditions, serum albumin, serum cholesterol, fasting serum glucose, and length of stay. In-hospital mortality was the primary outcome, and 45-day mortality was the secondary outcome. RESULTS: Total in-hospital mortality was 14.5%. In patients with and without DM, mortality was 8.8% and 11.3%, respectively, in Group A; 13.6% and 17.3% in Group B, and 12.6% and 34.3% in Group C. After controlling for confounders, newly recognized hyperglycemia (> 181 mg/dL) was independently associated with in-hospital mortality (adjusted odds ratio=2.7, 95% confidence interval=1.6-4.8). Forty-five-day mortality in newly recognized hyperglycemic patients was 17.5%, 25.7%, and 42% in Groups A, B, and C, respectively, whereas it was 21.2% in patients with DM. CONCLUSION: In elderly patients, newly recognized hyperglycemia was associated with a higher mortality rate than in those with a prior history of DM. These data suggest that further randomized clinical trials are needed to assess the efficacy and the risk of a target glucose of greater than 180 mg/dL.