There are few data regarding postoperative hyperglycaemia in non-diabetic compared with diabetic patients following postoperative nausea and vomiting prophylaxis with dexamethasone. Eighty-five non-diabetic patients and patients with type-2 diabetes were randomly allocated to receive intravenous dexamethasone (8 mg) or ondansetron (4 mg). Blood glucose levels were measured at baseline and then 2, 4 and 24 h following induction of anaesthesia. In non-diabetic patients, the mean (SD) maximum blood glucose was higher in those who received dexamethasone compared with ondansetron (9.1 (2.2) mmol.l(-1) vs. 7.8 (1.4) mmol.l(-1) , p = 0.04). In diabetic patients, the mean (SD) maximum blood glucose was also higher in those who received dexamethasone compared with ondansetron (14.0 (2.5) mmol.l(-1) vs. 10.7 (2.4) mmol.l(-1) , p < 0.01). Multivariate analysis demonstrated that dexamethasone administration was a significant predictor of maximum postoperative blood glucose increase (p < 0.01) after adjusting for potential confounders. There was no interaction between baseline blood glucose level, or presence or absence of diabetes, and dexamethasone administration. We conclude that dexamethasone increases postoperative blood glucose levels in both non-diabetics and diabetics.