The Association of Oral Bisphosphonate Use With Mortality Risk Following a Major Osteoporotic Fracture in the United Kingdom: Population-Based Cohort Study

OBJECTIVES: Bisphosphonates (BPs) might have extra benefits in reducing mortality because of their anti-atherosclerotic effects, but studies reported conflicting results. We investigated the association between oral BP use and mortality risk following a major osteoporotic fracture (MOF) in the United Kingdom.

DESIGN: This was a population-based cohort study.

SETTING AND PARTICIPANTS: In total, 163,273 adults aged 50 years and older with an MOF between 2000 and 2018 from the Clinical Practice Research Datalink in the United Kingdom.

METHODS: Cox proportional hazards models were used to estimate the risk of all-cause mortality in current (0‒6 months), recent (7‒12 months), and past (>1 year) exposures to oral BPs after nonhip MOF and hip fracture. In addition, stratification by sex, BP type, and duration of follow-up was performed.

RESULTS: Compared with never users of oral BPs, current BP use was associated with a 7% higher all-cause mortality risk after nonhip MOF, whereas a 28% lower all-cause mortality risk was observed after hip fracture. Past BP exposure was associated with a 14% and 42% lower risk after nonhip MOF and hip fracture, respectively. When considering only the first 5 years of follow-up, mortality risk associated with current BP use was significantly lower for both fracture groups, and the greatest reduction in mortality risk was observed within the first year. Women had slightly lower risk compared with men.

CONCLUSIONS AND IMPLICATIONS: We found a slight increased risk of all-cause mortality with current BP exposure after a nonhip MOF; however, a protective effect was observed following a hip fracture. Both the timing and the effect size of an association based on the anti-atherosclerotic hypothesis of BPs are not supported by our results. The decreasing trend of the mortality risk with shorter durations of follow-up suggests that the observed association is likely due to unknown distortion or unknown pleiotropic properties of BPs.