OBJECTIVES: To assess the incidence of diabetes for individuals with schizophrenia treated with antipsychotics. METHODS: Retrospective analysis of a large, geographically diverse claims database of insured individuals identified 815 enrollees aged 18 to 64 who: (1) were diagnosed with schizophrenia; (2) were initiated on typical (n=353) or atypical (n=462) antipsychotics between October 1, 1996 and December 31,1998; (3) had no use of any antipsychotics six-month prior-initiation; and (4) had no diagnosis of diabetes and/or no use of antidiabetics in the year prior. New onset diabetes was defined as either two diagnoses for diabetes (ICD9 250.xx) or prescription for antidiabetics in the year post-initiation. Logistic regressions were used to compare the odds of incidence of diabetes, controlling for demographics and prior-medical comorbidities. RESULTS: The probability of becoming diabetic was not significantly different for atypical cohort compared to typical cohort (odds ratio=2.533; p=0.088) or for olanzapine cohort versus typical cohort (odds ratio=1.093, p=0.900). Risperidone-treated patients had significantly higher incidence of diabetes compared to those treated with typicals (odds ratio=4.362, p=0.016). Olanzapine compared to risperidone cohort was associated with a significantly lower incidence of diabetes (odds ratio=0.277, p=0.050). CONCLUSIONS: The incidence of diabetes was similar for individuals with schizophrenia receiving treatment with atypical compared to typical antipsychotic agents. Additionally, individuals receiving treatment with olanzapine compared to risperidone had a lower incidence of diabetes.