Real-world evidence of raloxifene versus alendronate in preventing non-vertebral fractures in Japanese women with osteoporosis: retrospective analysis of a hospital claims database.

We conducted a retrospective cohort study using a de-identified hospital administrative claims database to assess the risk of non-vertebral fracture in Japanese women with osteoporosis treated with raloxifene compared with alendronate. The study included Japanese women ≥50 years of age with newly initiated alendronate or raloxifene treatment between July 2008 and March 2013 (index date was defined as the day of first prescription for alendronate or raloxifene), and had any claim with an osteoporosis definition during the study period. A total of 37,056 patients in the database initiated treatment, and there were 4802 and 1250 patients included in the alendronate and raloxifene analysis groups, respectively. The mean observation period in the alendronate group (529.2 days) was significantly longer than that for the raloxifene group (473.5 days, P < 0.001). Non-vertebral fractures accumulated linearly, at a similar rate, for both study drugs: incidence at 1 year was 2.83 and 2.64% for the alendronate and raloxifene groups, respectively. For the relative risk of non-vertebral fracture within 1 year, the adjusted hazard ratio was 0.933 for raloxifene versus alendronate, indicating that the relative risk of non-vertebral fracture was similar for the two drugs. The effectiveness of raloxifene in preventing non-vertebral fractures in Japanese women with osteoporosis was similar to that of alendronate. Therefore, raloxifene may be worthy of consideration as an alternative treatment.