OBJECTIVES: To assess the risk of diabetes among patients undergoing treatment with risperidone vs olanzapine. A series of case reports had previously associated olanzapine use with the development of hyperglycemia, diabetes, and diabetic ketoacidosis. METHODS: Two cohorts totaling 34,713 patients were identified from the Quebec Medicare database between January 1997 and 31st December 1999. One cohort consisted of patients who had at least one prescription for olanzapine during that period (n=19,779) and the other of patients receiving risperidone but not olanzapine (n=14,934). In either case, patients with a diagnosis of diabetes (defined as either a recorded ICD9 250.0 to 250.93 or a prescription for insulin or an oral hypoglycemic agent) before beginning anti-psychotic therapy were excluded. New diabetes diagnoses after the first antipsychotic prescription were tabulated until 31st December 1999; censoring occurred at this date, or the diagnosis of diabetes or at the last service date if there was no record of using any services during the last six months. Crude hazard ratios and proportional hazards analyses were carried out. RESULTS: 336 patients developed diabetes after being prescribed olanzapine, and 220 after risperidone, a crude hazard ratio of 1.10 (95% CI 0.90 to 1.35). When correcting for severe imbalances in age and gender using proportional hazards analysis, the hazard ratio increased to 1.22 (95% CI 1.02 to 1.45). After also adjusting for the duration of olanzapine exposure, the first three months of olanzapine treatment was associated with the highest risk of diabetes, hazard ratio 1.91 (95% CI 1.42 to 2.58). CONCLUSIONS: Olanzapine was associated with an increased risk of developing diabetes compared to risperidone. More studies are required to confirm the association of olanzapine with hyperglycemia and to identify the patient groups at highest risk of developing diabetes.