OBJECTIVES: The objective was to estimate direct medical costs for non-vertebral fractures in the first year of therapy among risedronate, alendronate, and nasal calcitonin patients utilizing an integrated administrative, medical and pharmacy claims database. METHODS: A retrospective cohort study was conducted among 5024 women and men (aged 45 +) with a new prescription for risedronate (5mg/day, or 30 mg/week), alendronate (5mg/day, 10mg/day, 35mg/week or 70mg/week), or nasal calcitonin (200I.U./day) between July 1, 2000, and June 30, 2001. Non-vertebral fracture-related direct medical costs (inpatient and outpatient care) were assessed for a 12-month period following initiation of the bisphosphonate or nasal calcitonin therapy using 2003 Medicare fee schedule payments. Sites at which a patient had a clinical fracture in the 6 months prior to initiation of therapy were excluded from the analysis. RESULTS: During the capture period patients were treated with alendronate (74%), risedronate (13%), and nasal calcitonin (13%). There were no baseline differences in age, gender or number of concomitant medications between risedronate and alendronate patients. Nasal calcitonin patients, however, were significantly older and had higher concomitant medications use than both risedronate and alendronate patients. Alendronate patients incurred almost three times the fracture-related costs of risedronate patients ($124 vs. $45, p = 0.019). Nasal calcitonin patients incurred more than four times the fracture-related costs of risedronate patients ($198 vs. $45, p = 0.028) in an adjusted model. Alendronate patients did not have significantly different fracture-related costs than nasal calcitonin patients. CONCLUSIONS: Observed fracture-related costs among patients initiating risedronate therapy were lower than fracture-related costs among patients of similar characteristics who initiated alendronate therapy.