OBJECTIVES To perform a cost-effectiveness analysis of thrombolysis with alteplase within 3 hours after acute ischemic stroke versus conservative treatment under the Brazilian public health care system perspective. METHODS A Markov model was developed simulating acute stroke treatment with conservative treatment or alteplase. Three-monthly cycles were considered, during which, patients might transit between five post-stroke disability states, based on the modified Rankin Scale. The probability of presenting intracerebral hemorrhage after alteplase treatment and transition probabilities in the first year were obtained from NINDS trial. For subsequent years, 1-year cycles were considered, to account for patients' mortality. The outcomes were expressed as life years gained (LYG). Both direct costs and indirect costs were considered in the analysis. Costs and outcomes were discounted at 5% per year. Results were segmented by gender and calculated for different time horizons, ranging from 1 to 30 years. Unit costs for drugs were obtained from the Brazilian Health Prices Database, hospitalization and procedure costs were extracted from the National Database of Hospital Costs (SIH/DATASUS). RESULTS In one year, incremental LYG were 0.0324 for both genders, with incremental costs of BRL608 for men and BRL363 for women. The incremental cost-effectiveness ratio in 1 year was BRL18,765/LYG (US$13,404) for men and BRL11,204/LYG (US$8,003) for women. After the second year, alteplase became cost-saving. The shorter length of hospitalization and reductions in rehabilitation needs and productivity losses overweighed incremental drug costs. (2005 PPP 1USD=1.4BRL) CONCLUSIONS Thrombolytic therapy with alteplase in the first three hours after stroke changes the natural history of disease, reducing both direct and indirect costs. In the long term, there are large scale resource savings for the Brazilian public health care system.