OBJECTIVES: To model the long-term metabolic costs and consequences of antipsychotic therapy for chronic schizophrenia. METHODS: We developed a Markov model to simulate long-term treatment courses of 10,000 chronic schizophrenia patients using annual cycles over a 10-year period. The model was developed with psychiatrists, an endocrinologist, and a cardiologist and simulated individual patients. Patient health status (comorbidities, lipid levels, body mass index, and blood pressure) could change as a function of treatment-related events. Probabilities were determined from published pooled analyses of clinical trial data. Costs and resource utilization patterns were obtained from published data and standard costing sources. The model estimated costs of long-term consequences of weight gain, such as diabetes and coronary heart disease. We used data from the Framingham Heart Study and Nurses’ Health study to assess risks for weight gain–induced diabetes and cardiovascular events. Outcomes included total number of patients with metabolic events and associated costs. RESULTS: Fewer patients treated with ziprasidone developed diabetes compared with other atypical antipsychotics. The greatest difference was observed between ziprasidone and olanzapine: costs for the entire cohort were $12.3 million vs $32.9 million for ziprasidone vs olanzapine, respectively. Increased diabetes risk and lipid levels translated into higher risk for CHD. At 10 years, an additional 235 olanzapine-treated patients developed CHD vs ziprasidone. Costs were again lowest for ziprasidone compared with other atypical antipsychotics. CONCLUSIONS: Adverse events common to some atypical antipsychotics (ie, weight gain) can have a deleterious long-term metabolic impact and should be considered when prescribing an agent for the treatment of schizophrenia.