(072) Drug Holiday of Oral Bisphosphonates and Mortality following Hip Fracture: A Population-based Cohort Study

Background: Recommendation for drug holiday in patients with extended use of oral bisphosphonates is mainly based on evidence in alendronate. Although risedronate has a much shorter half-life, few studies reported the effect of drug holiday in alendronate users and risedronate users separately. Additionally, most studies reported sustained effects on fracture prevention, but evidence on protection against fracture sequelae remains unknown.

Objectives: To investigate the effect of drug holiday on post-fracture mortality in patients with hip fractures and have had extended use of alendronate and risedronate respectively.

Methods: In this retrospective population-based cohort study of all patients admitted for hip fractures in Victoria, Australia from 2014-18, data was retrieved from our linked administrative databases, namely Victorian Admission Episodes Dataset, Pharmaceutical Benefits Scheme, and National Death Index. We only included patients with extended oral bisphosphonate use, which was defined as ≥80% medication possession ratio of oral alendronate or risedronate for ≥5 years prior to hip fracture. The lengths of drug holidays following the 5-year extended uses were then agnostically determined by group-based trajectory modelling. The hazard ratios (HR) and 95% confidence intervals (CI) were estimated using Cox proportional hazard model to compare mortality of patients with different trajectories of bisphosphonate use to those with continued use.

Results: Among 18,193 patients admitted for hip fractures, 365 patients had extended use of oral alendronate (n=249) or risedronate (n=154) for ≥5 years prior to admission. Most patients (69%) continued to use oral bisphosphonates till admission for hip fracture; 17% had discontinued for one year and 14% had discontinued for two years prior to admission. Compared to patients with continued use, post-fracture mortality was higher in patients who had discontinued risedronate for one year (HR 2.37, 95% CI 1.24-4.53) and two years (HR 3.08, 95% CI 1.48-6.41) prior to hip fracture. No increase or decrease in post-fracture mortality was observed in patients who had discontinued alendronate for one year (HR 0.59, 95% CI 0.29-1.18) or two years (HR 1.05, 95% CI 0.57-1.93) prior to hip fracture.

Conclusions: After at least 5 years of extended use, post-fracture mortality is higher in people who discontinue risedronate, but not alendronate, for 1 or 2 years. The type of bisphosphonate may be a factor to consider when planning drug holidays.