Prevalence, predictors, and outcomes of steroid-induced hyperglycemia in hospitalized patients with hematologic malignancies.

To determine prevalence, predictors, and outcomes of steroid-induced hyperglycemia in hospitalized patients with hematologic malignancies METHODS: We performed retrospective analysis of patients who were hospitalized on any of the four hematology services or bone marrow transplant service and who received systemic steroids during a 2-month period. Mean glucoses for days 1-4 and maximum glucose were calculated and the total daily steroid dose was converted to equivalent dose of dexamethasone. Hyperglycemia was defined as any glucose >9.917 mmol/l during days 1-4. We examined associations between variables using Spearman's correlations and multivariable linear regression RESULTS: 168 patients were included in the analysis. Mean age was 57.1 ± 14.4 years with 59% males and 90% Caucasians. The prevalence of hyperglycemia was 39%. Eight patients received intravenous insulin. In patients without diabetes, steroid dose equivalent to >12 mg dexamethasone and longer acting steroids caused greater degree of hyperglycemia compared to a dose <12 mg. Maximum glucose was a predictor of hospital length of stay among patients without diabetes (but not those with diabetes), and with acute leukemia or stem cell transplant (but not other hematologic malignancies). There were no significant differences in in mortality or other outcomes between the groups with and without hyperglycemia CONCLUSIONS: Hyperglycemia is common in patients with hematologic malignancies, which require frequent corticosteroids. Higher steroid dose and long-acting steroids caused a greater degree of hyperglycemia. Maximum glucose was a predictor of length of stay but this was only significant for patients without diabetes, those with acute hematologic malignancy or stem cell transplant.