Atypical antipsychotic drugs and diabetes mellitus in a large outpatient population: a retrospective cohort study.

PURPOSE Previous research has suggested an association between use of atypical antipsychotics and onset of diabetes mellitus. We sought to compare the incidence of new onset diabetes among patients receiving atypical antipsychotics, traditional antipsychotics or antidepressants. METHODS Retrospective cohort study of outpatients with claims for atypical antipsychotics (n = 10 265) compared to controls with claims for traditional antipsychotics (n = 4607), antidepressants (n = 60 856) or antibiotics (n = 59 878) in the administrative claims database of a large pharmaceutical benefit manager between June 2000 and May 2002. Main outcome measures were adjusted and unadjusted incidence rates of diabetes (new cases per 1000 per year) in a 12-month period, as measured using new prescriptions for antidiabetic drugs after a 6-month lead-in period. RESULTS Annual unadjusted incidence rates of diabetes (new cases per 1000 per year) were 7.5 for atypical antipsychotics, 11.3 for traditional antipsychotics, 7.8 for antidepressants and 5.1 for antibiotics. In multivariable analyses, age, male sex and Chronic Disease Score were associated with greater odds of diabetes onset. There were no statistically significant differences in outcome between the atypical antipsychotic, traditional antipsychotic and antidepressant groups. Multivariable comparisons among specific agents showed increased odds of diabetes for clozapine, olanzapine, ziprasidone and thioridazine (relative to risperidone), but these comparisons did not reach statistical significance. CONCLUSIONS In a large prescription claims database, outpatients taking atypical antipsychotics did not have higher rates of diabetes onset, compared to subjects taking traditional antipsychotics or antidepressants.