BACKGROUND: Case series suggest that some antipsychotics may induce or exacerbate type 2 diabetes. This study measured the association of antipsychotic treatments with diabetes at a population level. METHOD: Claims data for psychosis patients (ICD-CM-9 290.xx-299.xx) within health plans encompassing 2.5 million individuals were analyzed. Patients reporting preexisting type 2 diabetes up to 8 months prior to observation were excluded. The frequency of newly reported type 2 diabetes in untreated patients and among patients treated with antipsychotics from 5 categories (risperidone, olanzapine, clozapine, and high-potency and low-potency conventionals) was compared. Logistic regression models compared the odds of diabetes based on exposure to each of the antipsychotic categories. RESULTS: Based on 12 months of exposure, the odds of type 2 diabetes for risperidone-treated patients (odds ratio = 0.88, 95% CI = 0.372 to 2.070) was not significantly different from that for untreated patients, whereas patients receiving other antipsychotics had a significantly greater risk of diabetes than untreated patients (p < .05): olanzapine, 3.10 (95% CI = 1.620 to 5.934); clozapine, 7.44 (95% CI = 0.603 to 34.751); high-potency conventionals, 2.13 (95% CI = 1.097 to 4.134); and low-potency conventionals, 3.46 (95% CI = 1.522 to 7.785). Older age and greater use of non-antipsychotic psychotropic medications also contributed to risk of type 2 diabetes. Olanzapine also showed significantly higher (p < .01) odds of diabetes associated with increasing dose. CONCLUSION: Consistent with previously published literature, these data suggest that olanzapine, clozapine, and some conventional antipsychotics appear to increase the risk of acquiring or exacerbating type 2 diabetes and that the effect may vary by drug. In contrast to these agents, risperidone was not associated with an increased risk of type 2 diabetes.
BACKGROUND: Case series suggest that some antipsychotics may induce or exacerbate type 2 diabetes. This study measured the association of antipsychotic treatments with diabetes at a population level. METHOD: Claims data for psychosispatients (ICD-CM-9 290.xx-299.xx) within health plans encompassing 2.5 million individuals were analyzed. Patients reporting preexisting type 2 diabetes up to 8 months prior to observation were excluded. The frequency of newly reported type 2 diabetes in untreated patients and among patients treated with antipsychotics from 5 categories (risperidone, olanzapine, clozapine, and high-potency and low-potency conventionals) was compared. Logistic regression models compared the odds of diabetes based on exposure to each of the antipsychotic categories. RESULTS: Based on 12 months of exposure, the odds of type 2 diabetes for risperidone-treated patients (odds ratio = 0.88, 95% CI = 0.372 to 2.070) was not significantly different from that for untreated patients, whereas patients receiving other antipsychotics had a significantly greater risk of diabetes than untreated patients (p < .05): olanzapine, 3.10 (95% CI = 1.620 to 5.934); clozapine, 7.44 (95% CI = 0.603 to 34.751); high-potency conventionals, 2.13 (95% CI = 1.097 to 4.134); and low-potency conventionals, 3.46 (95% CI = 1.522 to 7.785). Older age and greater use of non-antipsychotic psychotropic medications also contributed to risk of type 2 diabetes. Olanzapine also showed significantly higher (p < .01) odds of diabetes associated with increasing dose. CONCLUSION: Consistent with previously published literature, these data suggest that olanzapine, clozapine, and some conventional antipsychotics appear to increase the risk of acquiring or exacerbating type 2 diabetes and that the effect may vary by drug. In contrast to these agents, risperidone was not associated with an increased risk of type 2 diabetes.