Recently, increasing attention has been drawn to the potential diabetogenic effect of novel antipsychotics. Until now, large prospective studies examining the relationship between atypical antipsychotics and impaired glucose metabolism have been lacking. However, the case reports and retrospective studies that we review here suggest an increased risk of developing diabetes mellitus (DM) in patients treated with atypical antipsychotics compared to schizophrenic patients treated with conventional antipsychotics or those without treatment. Although most atypical antipsychotic agents might have a diabetogenic potential, the risk of developing DM might be higher in patients treated with either clozapine or olanzapine than with risperidone, whereas data on quetiapine and ziprasidone is presently limited and needs further attention. Possible mechanisms include the induction of peripheral insulin resistance and the direct influence on pancreatic beta-cell function by 5-HT1A/2A/2C receptor antagonism, by inhibitory effects via alpha 2-adrenergic receptors or by toxic effects. On the other hand, atypical antipsychotics might not be an independent risk factor for the development of DM, but hasten the onset of DM in patients bearing other risk factors. It is suggested that schizophrenic patients should be monitored for the occurrence of glucose metabolism abnormalities before starting atypical antipsychotics, and at a 3-month interval at least during therapy.