Incidence of fractures of the femur, including subtrochanteric, up to 8 years since initiation of oral bisphosphonate therapy: a register-based cohort study using the US MarketScan claims databases.

UNLABELLED: In a cohort study of users of bisphosphonates, we evaluated the incidence of fragility fractures at all sites on the femur following for up to 8 years of therapy with alendronate or risedronate. We did not find evidence for a reversal of fracture protection with long-term use of bisphosphonates. INTRODUCTION: Few studies have acquired adequate data with prolonged follow-up on bisphosphonate users in the general population to evaluate their long-term effects on the risk of hip fractures including those in the subtrochanteric region. METHODS: This cohort study utilizes a large USA database (January 1, 2000 to June 30, 2009). We compared patients with higher versus lower degrees of compliance [medication possession ratio, MPR <1/3 (the reference), 1/3-<2/3, or ≥ 2/3]. Radiographic adjudication of fracture site and features were not performed. Hazard ratios (HR) for fracture were estimated using time-dependent Cox models. Restricted cubic splines (RCS) were used to plot HRs for fracture against duration of therapy. RESULTS: There were 3,655 incident cases of femoral fracture (764 subtrochanteric/shaft, 2,769 hip) identified during 917,741 person-years of follow-up (median  =  3 years) on 287,099 patients (267,374 were women) from the date when they initiated oral bisphosphonate therapy. The corresponding HRs (95% confidence interval, CI) for overall femoral fractures associated with each additional year of therapy were 0.93 (0.86-1.01) within 5 years, and 0.89 (0.77-1.03) beyond 5 years for risedronate and 0.86 (0.81-0.91) and 0.95 (0.84-1.07) for alendronate, respectively. The corresponding estimates for subtrochanteric/shaft fractures were 1.05 (0.87-1.26) and 0.89 (0.60-1.33) for risedronate and 0.99 (0.92-1.05) and 1.05 (0.92-1.20) for alendronate, respectively. The HRs (95% CI) for overall femoral fractures associated with each additional year of alendronate or risedronate therapy within 5 and beyond 5 years were not significantly different. CONCLUSION: Our study showed persistence of overall hip fracture protection with long-term use of alendronate or risedronate.