LAUSR

Background Given FDA warnings, drug holidays (temporary or permanent discontinuation) of bisphosphonates (BPs) after long-term (3–5 years) continuous therapy is becoming increasingly common in the United States (US). However, the benefits and risks of stopping BPs, and the optimal timing to restart, remain unclear. Objectives We conducted a population-based cohort study of women on long term BP therapy to evaluate the rate of hip fracture following a drug holiday. Methods We used Medicare data (2006–2014) to identify all women with medical and pharmacy coverage who initiated a BP and were at least 80% adherent for ≥3 years (‘baseline’), at which follow-up time began. Patients using other bone therapies (e.g. denosumab, oestrogen, teriparatide, calcitonin) were excluded or censored if they started after follow-up began. We calculated crude rates of hip fracture for continuing BP therapy and among those who discontinued, for categories of time since discontinuing (i.e., length of drug holiday), extending up to 3 years. We used Cox proportional hazards models to evaluate the risk of discontinuing per the length of the drug holiday, using age as the time axis and controlling for potentially confounding factors, with and without adjusting for death as a competing risk. Results We identified 1 56 236 women who were highly adherent, long-term BP users. The mean (SD) age was 78.5 (7.5) years. The most commonly used BPs were alendronate (71.7% ever use, 52% exclusive use) and zoledronic acid (16.2% ever use, 8.9% exclusive use). During a median (IQR) followup of 2.1 (1.0, 3.3) years, 62 676 (40.1%) of women stopped BP therapy for at least 6 months or more. Among these women, 7947 (12.7%) subsequently restarted any BP. Overall, 16 904 (10.8%) died. A total of 3745 hip fractures occurred during follow-up. Hip fracture rates were lowest among women who were current users, and gradually increased as the length of the drug holiday increased, achieving their maximum with a drug holiday >2 years (table 1).Abstract OP0017 – Table 1 Hip fracture rate by duration of BP drug holiday, adjusting for competing risk of death Time since Bisphosphonate Discontinuation (yrs) Number of hip fractures, n Crude Incidence Rate per 1000 person-years Adjusted* Hazard Ratio(95% CI) 0 (i.e. current use) 1958 9.6 (9.2–10.1) 1.0 (reference) >0 to≤3 months 530 13.1 (12.0–14.3) 1.29 (1.17–1.42) >3 months≤1 year 539 12.0 (11.0–13.1) 1.12 (1.02–1.24) >1 to≤2 years 422 13.3 (12.0–14.6) 1.21 (1.09–1.35) >2 to≤3 years 235 15.7 (13.7–17.8) 1.39 (1.21–1.59) *adjusted for age, region, race, rural or urban, median income, calendar year, comorbidity(fragility fracture, charlson comorbidity index score), DXA, number of physician visits, care by a rheumatologist or endocrinologist, long term care residence, vitamin D deficiency, glucocorticoids, and proton pump inhibitors Conclusions In a large cohort of U.S. women, a BP drug holiday greater than 2 years was associated with a significantly increased risk for hip fracture of up to 39% compared to continued BP use. Disclosure of Interest J. Curtis Grant/research support from: AbbVie, Amgen, BMS, Corrona, Janssen, Lilly, Myriad, Pfizer, Roche/Genentech, UCB, Consultant for: AbbVie, Amgen, BMS, Corrona, Janssen, Lilly, Myriad, Pfizer, Roche/Genentech, UCB, R. Chen Grant/research support from: Amgen, Z. Li Grant/research support from: Amgen, T. Arora Grant/research support from: Amgen, K. Saag Grant/research support from: Amgen, Merck, Consultant for: Amgen, Merck, Radius, N. Wright: None declared, S. Daigle: None declared, M. Kilgore Grant/research support from: Amgen, E. Delzell: None declared