Time to benefit and the long-term persistence of new users of oral bisphosphonates.

Abstract

This study aimed to examine long-term persistence in new users of oral bisphosphonates in a population-wide cohort in Manitoba, Canada. A longitudinal observational study was conducted using administrative health data characterizing long-term bisphosphonate persistence in those who started treatment between 1997 and 2018. Treatment discontinuation was evaluated using Kaplan-Meier methods. Cox regression was used to examine associations between discontinuation and osteoporosis diagnosis, previous fractures, and age. A sub-analysis of users with FRAX scores examined the relationship between 10-year fracture risk estimations and discontinuation. Of 42,249 new bisphosphonate users, median age was 71years, with 88.6% being female. Median duration of bisphosphonate use was 0.95years (IQR 0.25, 3.9years). Overall, 47.9% of incident users persisted up to 1year, 25.0% persisted up to 3years, and 14.1% up to 5years. Presence of an indication for bisphosphonate use was associated with decreased discontinuation risk. Persistence generally increased with age. Having a BMD test performed was a predictor of lower discontinuation. The strongest predictor was having an osteoporosis diagnosis [HR for discontinuation = 0.68 (95% CI 0.66, 0.70)]. In users with FRAX scores (n = 14,114), moderate-risk [HR = 0.86 (95% CI 0.77, 0.96)] and high-risk users [HR = 0.77 (95% CI 0.69, 0.85)] were less likely to discontinue compared to lower-risk users. A rapid decline in bisphosphonate persistence was shown. Almost half of users would not be expected to achieve clinically relevant benefits with a persistence of less than 1year. Allowing informed choice in high-risk patients may be the best way to focus on those likely to benefit and persist with treatment.