INTRODUCTION: The prevalence of diabetes in schizophrenia is twice that in the general population, but there are few reliable predictors of which individuals will develop glucose dysregulation. OBJECTIVE: To test if abnormal birth weight (either too low or too high) and parental diabetes, both variables that can be ascertained in the clinic, can predict diabetes onset in patients with schizophrenia. RESEARCH DESIGN AND METHODS: Electronic records of a cohort of 190 clozapine-treated patients (37% treated for more than 20 years) and Cox regression survival analysis (with any type of glucose dysregulation as the event) to account for differences in length of treatment before the event and age at clozapine treatment initiation. RESULTS: Age at clozapine initiation (Exp(B)=1.098; p<0.001), family history of diabetes (Exp(B)=2.299; p=0.049) and birth weight2 (Exp(B)=0.999; p=0.013) were significant predictors of glucose dysregulation onset, while gender was not (Exp(B)=0.1.350; p=0.517). Among individuals with 10 years of follow-up, 80% of those with both abnormal birth weight and a family history of diabetes developed diabetes compared with 56% with only abnormal birth weight, 40% with only a family history of diabetes and 20% in those with neither. CONCLUSIONS: Since 48% of cases had at least one risk factor and 6% had both risk factors, there is a substantial proportion of patients for whom preventive strategies could be implemented.