This ten-year follow-up study examined the prevalence and the most relevant baseline predictors of suicide in schizophrenic patients. In 1993, 3470 patients meeting the ICD-10 criteria for schizophrenia were assessed. We used national death certificate data to identify patients that had died by suicide for each year included in the study. In this way, we calculated standardized mortality ratios, adjusting for age and sex relative to the general population. We used Cox's proportional hazards models to investigate potential sociodemographic and clinical risk factors. There were 141 suicides in the cohort during the follow-up period, corresponding to a risk of suicide that was approximately 16 times higher than that of the general population. Women had slightly higher standardized mortality ratios than men. Suicide was the cause of death in more than half (53.9%) of deaths occurring during the first year of follow-up and nearly one-third (31.8%) of those occurring in the ten-year period of the study. There were four significant baseline predictors of suicide remaining in the final logistic regression model: male gender, drug abuse, previous suicide attempts, and short duration of illness. Sex, age, history of suicide attempt should be particularly considered in the assessment of suicide risk in schizophrenic patients. Our findings also emphasize the need for detection and effective management of associated comorbid drug abuse.