The association of diabetes mellitus and mental illness, in particular, schizophrenia, has been remarked upon for over a century. Recent epidemiological studies have shown the age- and sex-matched prevalence of diabetes in patients with schizophrenia to be 1.5-2 times those in the general population. This difference is particularly noticeable in younger patients. The explanation for this finding probably resides in both environmental and biological factors. Patients with schizophrenia tend to be sedentary and have a poor diet, which are both known risk factors for diabetes. However, familial studies have indicated a heritable component to the risk of diabetes in patients with schizophrenia. A number of biological mechanisms have been proposed to explain this, including neuroendocrine changes and neurodevelopmental anomalies, but none are entirely satisfactory. In addition, it has been suggested that treatment with antipsychotic medication may potentially increase the risk of diabetes and account for some of the increased prevalence seen in patients with schizophrenia. It has been suggested that different antipsychotic drugs may differ in their ability to facilitate the emergence of poor glycaemic control in patients with schizophrenia. However, the situation is far from clear and more work is required to accurately assess the potential risk associated with different antipsychotic drugs.