LAUSR

BACKGROUND To compare the risks of major osteoporotic, vertebral, and non-vertebral fractures between patients who discontinued anti-osteoporosis medications. METHODS We conducted a comparative effectiveness study with a nationwide population-based cohort study design. Patients aged ≥50 years admitted between 2012 and 2015 for incident hip fractures and receiving denosumab or bisphosphonates with sufficient compliance for at least one year were included. Patients were categorized into persistent or non-persistent denosumab or bisphosphonates users based on their subsequent use pattern. The main outcomes were subsequent hospitalizations for a major osteoporotic, vertebral or non-vertebral fracture. Multivariate, time-varying Cox proportional hazards model was used to evaluate the risk of major outcomes. RESULTS Compared with persistent denosumab users, non-persistent denosumab users had a significantly higher risk of major osteoporotic fractures (Hazard ratio (HR) = 1.60; 95% confidence interval (CI), 1.20 to 2.14),vertebral fractures (HR = 2.18; 95% CI, 1.46 to 3.24) and death (HR = 3.57; 95%CI, 2.63-4.84). However, the increased risk of fracture was not found in both persistent and non-persistent bisphosphonates users. Noteworthy, the increased risk of vertebral fractures in non-persistent denosumab users was more pronounced within one-year post discontinuation (HR = 2.90, 95% CI, 1.77 to 4.74) and among patients who discontinued from 2-year denosumab therapy (HR = 3.58; 95% CI, 1.74 to7.40). DISCUSSION Discontinuation of denosumab resulted in an increased risk of major osteoporotic fractures, especially vertebral fractures. The increased risk tends to reveal within one-year post discontinuation and be greater after a longer treatment duration. Notably, only fracture with hospitalization was identified as our research outcome, the real risk of osteoporotic fracture post discontinuation is believed to be higher, especially for vertebral fracture. This article is protected by copyright. All rights reserved.