Comparative study of the development of diabetes mellitus in patients taking risperidone and olanzapine

Objectives. A growing body of literature suggests that certain atypical antipsychotics, especially olanzapine and clozapine, may induce glucoregulatory dysfunction. We assessed the differences in risk of developing diabetes mellitus during treatment with olanzapine and risperidone by using patients treated with haloperidol and fluphenazine as control subjects in whom we would not expect to see an increased risk. Methods. We conducted a retrospective analysis of the Veteran's Integrated Service Network 10 Veterans Affairs (VA) database. Data for patients receiving olanzapine, risperidone, haloperidol, or fluphenazine from January 1, 1997-December 31, 2000, were included. Diabetes was defined as any health system encounter associated with the International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis for diabetes (250.xx) or prescription for a hypoglycemic agent. Data of patients with markers for diabetes within 1 year before their index date, female patients, racial groups other than Caucasian or African-American, and patients receiving clozapine were not analyzed. We performed a Cox regression, with antipsychotic therapy as a time-dependent covariate. Other covariates considered for inclusion in the final model were number of days supply of antipsychotic drug, age, race, psychiatric diagnoses, substance abuse, lithium, valproic acid, and other typical or atypical antipsychotic agents. Results. Data for 5837 patients were analyzed. Overall rate of developing diabetes in the study population was 6.3% (368 of 5837 patients). Olanzapine therapy was associated with a significantly higher risk of development of diabetes compared with risperidone (hazard ratio [HR] 1.37, 95% confidence interval 1.06-1.76, p=0.016) while controlling for race, age, diagnosis, substance abuse, lithium, valproic acid, and other atypical antipsychotic agents. No differences in the rate of developing diabetes were detected between fluphenazine and risperidone (HR 1.11, p=0.69), or haloperidol and risperidone (HR 0.89, p=0.41). Conclusions. Olanzapine was associated with a 37% (HR 1.37) increased risk of development of diabetes compared with risperidone in a VA population, even after adjusting for other factors associated with the development of diabetes and temporal exposure to study drug. Because of limitations associated with database research, prospective studies should be conducted to corroborate these findings.