Risk of hip fracture after bisphosphonate discontinuation: implications for a drug holiday

Abstract

Based upon interest in a bisphosphonate drug holiday, we evaluate the risk for hip fracture after bisphosphonate discontinuation. Among women compliant with bisphosphonates for > or = 2 years, the risk of hip fracture was increased after discontinuation, although with higher compliance and a longer duration of preceding bisphosphonate therapy, this risk was attenuated. Recent data suggest that hip fracture risk was not significantly increased among women receiving 5 years of bisphosphonate therapy who were subsequently randomized to placebo. We studied older women compliant with bisphosphonates > or = 2 years to evaluate the risk of hip fracture after bisphosphonate discontinuation. Using administrative databases from a large U.S. healthcare organization, we identified women initiating bisphosphonate therapy compliant (Medication Possession Ratio, MPR > or = 66%) for 2 years. We examined the rate of hip fracture among women who discontinued bisphosphonates versus those who remained on therapy. At 2 years, 9,063 women were eligible for analysis. Hip fracture incidence among women who discontinued bisphosphonates versus those who did not was 8.43 versus 4.67 per 1000 person years (p = 0.016). The adjusted hazard ratio of hip fracture per 90 days following discontinuation was 1.2 (1.1-1.3). For women with higher compliance at 2 years (MPR > or = 80%) or compliant for 3 years, there were no significant differences in risk associated with discontinuation. The rate of hip fracture was increased among women compliant with bisphosphonate therapy for 2 years who subsequently discontinued, suggesting that discontinuation is not advisable under these conditions. This association was attenuated with higher compliance and a longer duration of previous bisphosphonate therapy.