OBJECTIVE: To analyze expenditure on treatment for postmenopausal osteoporosis and associated factors on mean per capita expenditure. METHODS: A probabilistic-deterministic linkage between the database of Authorizations for Highly Complex Procedures and the mortality information system was constructed, resulting in a historical cohort of patients using high-cost medications for the treatment of postmenopausal osteoporosis, between 2000-2006. Mean monthly spending on medicines was stratified by age group and described according to demographic and clinical characteristics and the type of drug used. A linear regression model was used to assess the impact of demographic and clinical characteristics on per capita mean monthly expenditure on medicines. RESULTS: We identified 72,265 women who received drugs for the treatment of postmenopausal osteoporosis. The average monthly expenditure per capita in the first year of treatment was $ 54.02 (sd $ 86.72). The population was predominantly composed of women aged 60-69 years old, who had started treatment in 2000, resident in the Southeast of Brazil, who had previously suffered osteoporotic fractures, and Alendronate sodium was the drug most commonly used at baseline. For most of the patients, the same active ingredient remained in use throughout the treatment period. During the program, 6,429 deaths were identified among participants. More than a third of women remained in treatment for up to 12 months. Raloxifen and calcitonin were the therapeutic alternatives with the greatest impact on the average monthly expenditure on medicine using alendronate sodium as a reference standard. CONCLUSIONS: Due to the high impact of the type of drug used on expenditure on medication, it is recommended that criteria for prescribing and dispensing be established by prioritizing those with lower costs and greater effectiveness in order to optimize the process of pharmaceutical care and provide the population with a greater number of pharmaceutical units.