Recent reports have raised the concern that clozapine increases the risk for diabetes mellitus. Accurate pharmacoepidemiologic data on whether such a hazard exists and its magnitude are needed to enable clinicians and patients to make proper treatment decisions about clozapine. The authors performed a case-control study involving 7,227 cases of newly treated diabetes and 6,780 controls, all with psychiatric disorders. Cases and controls were older than 20 years and enrolled in government-sponsored drug benefit programs in New Jersey. The authors measured the use of clozapine or other antipsychotic medications and additional covariates. They developed logistic regression models adjusted for demographic, clinical, and health care use characteristics to identify whether clozapine users were at increased risk to begin treatment for diabetes. Clozapine use was not significantly associated with developing diabetes (adjusted odds ratio [OR], 0.98; 95% confidence interval [CI], 0.74-1.31). There was no suggestion of relationships between larger dosages or longer durations of clozapine use and increasing risks of diabetes. On the other hand, nonclozapine antipsychotic medication use was associated with a modest but significantly increased risk of developing diabetes (adjusted OR, 1.13; 95% CI, 1.05-1.22). Among individual nonclozapine antipsychotics, significantly elevated risks were observed for two phenothiazine agents: chlorpromazine (adjusted OR, 1.31; 95% CI, 1.09-1.56) and perphenazine (adjusted OR, 1.34; 95% CI, 1.11- 1.62). In contrast to earlier reports, these results provide some reassurance that clozapine does not increase the risk of developing diabetes. Additional data from pharmacoepidemiologic studies and randomized controlled trials are needed to exclude the possibility of residual confounding and ensure the appropriate use of this agent.