The study results indicate that women with osteoporosis initiated on gastro-resistant risedronate have a lower risk of fracture than those initiated on immediate release risedronate or alendronate. A large proportion of women discontinued all oral bisphosphonate therapies within 1 year of treatment start. Purpose Using a US claims database (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019), we compared risk of fractures between women with osteoporosis initiated on gastro-resistant (GR) risedronate and those initiated on (a) immediate release (IR) risedronate or (b) immediate release alendronate. Methods Women aged ≥ 60 years with osteoporosis who had ≥ 2 oral bisphosphonate prescription fills were followed for ≥ 1 year after the first observed bisphosphonates dispensing (index date). Fracture risk was compared between the GR risedronate and IR risedronate/alendronate cohorts using adjusted incidence rate ratios (aIRRs), both overall and in subgroups with high fracture risk due to older age or comorbidity/medications. Site-specific fractures were identified based on diagnosis codes recorded on medical claims using a claims-based algorithm. Persistence on bisphosphonate therapy was evaluated for all groups. Results aIRRs generally indicated lower fracture risk for GR risedronate than IR risedronate and alendronate. When comparing GR risedronate to IR risedronate, statistically significant aIRRs (p < 0.05) were observed for pelvic fractures in the full cohorts (aIRRs = 0.37), for any fracture and pelvic fractures among women aged ≥ 65 years (aIRRs = 0.63 and 0.41), for any fracture and pelvic fractures among women aged ≥ 70 years (aIRRs = 0.69 and 0.24), and for pelvic fracture among high-risk women due to comorbidity/medications (aIRR = 0.34). When comparing GR risedronate to alendronate, statistically significant aIRRs were observed for pelvic fractures in the full cohorts (aIRR = 0.54), for any fracture and wrist/arm fractures among women aged ≥ 65 years (aIRRs = 0.73 and 0.63), and for any fracture, pelvic, and wrist/arm fractures among women aged ≥ 70 years (aIRRs = 0.72, 0.36, and 0.58). In all cohorts, ~ 40% completely discontinued oral bisphosphonates within 1 year. Conclusions Discontinuation rates of oral bisphosphonate therapy were high. However, women initiated on GR risedronate had a significantly lower risk of fracture for several skeletal sites than women initiated on IR risedronate/alendronate, particularly those aged ≥ 70 years.