OBJECTIVES: Certain atypical antipsychotic medications have been shown to increase diabetes and CHD risk in patients by exacerbating lipid, weight, and glucose risk factors comprising the metabolic syndrome. We compared events and costs associated with the use of olanzapine versus aripiprazole from a double blind, 16 week randomized trial of schizophrenia patients. METHODS: A total of 173 overweight patients with body mass index greater than 27 kg/m2 and schizophrenia who were previously treated with olanzapine were randomized to double-blind treatment with aripiprazole or continued olanzapine monotherapy for 16-weeks following a 2-week, open-label, observation period during which subjects continued to receive olanzapine. The rate of ATP-III metabolic syndrome, and the long term risk for diabetes and CHD was estimated and compared between treatment arms. Brazilian costs and health resource estimates were applied to each event. Established risk functions were used to estimate diabetes and CHD risk. RESULTS: Among all patients, the baseline rate of metabolic syndrome was 89%. At 16 weeks, 80.3% of olanzapine versus 60% of aripiprazole patients exhibited metabolic syndrome (RR:0.75; 95%CI:0.61-0.92,p=.006). Diabetes risk increased by 1.3% among aripiprazole patients and 6.4% among olanzapine patients (risk difference = 5.1%). CHD risk decreased by 0.6% among aripiprazole patients and increased by 0.3% among olanzapine patients (risk difference = 0.9%). Among 1000 patients, treatment with aripiprazole versus olanzapine would avert 203 metabolic syndrome events, 51 diabetes events, and 9 CHD events at a cost of R$ 1.520,63, R$ 1.856, and R$ 4.122,31, per event avoided respectively. The total cost difference was R$968.585,72. Risk differences were driven primarily by significant weight and lipid changes between agents, favoring aripiprazole. CONCLUSION: Antipsychotic-related metabolic adverse events and the risk of diabetes and CHD can add substantively to health care costs among schizophrenia patients in Brazil. Health care providers should consider these risks in the selection of antipsychotic agents.