Hip and other fragility fracture incidence in real-world teriparatide-treated patients in the United States

SUMMARY: This study demonstrates real-world effectiveness of teriparatide in reducing the risk of hip and other fragility fractures. Fracture incidence significantly decreased as adherence and persistence increased for any clinical, vertebral, nonvertebral, and hip fractures among patients who were observed for 2 years after teriparatide initiation. INTRODUCTION: Examine the relationship of treatment adherence and persistence to teriparatide with hip and other fractures. METHODS: Truven MarketScan Research Databases, 2004 through 2014, provided teriparatide users ≥18 years old with continuous coverage 12 months pre- and 24 months post-teriparatide prescription. Adherence (medication possession ratio, MPR) groups were defined as high (≥0.80), medium (0.50 ≤ MPR < 0.80), and low (<0.50). Persistence, allowing for ≤90-day gaps between prescriptions, was defined as 1–6, 7–12, 13–18, and 19–24 months. Fracture incidence was summarized and compared by using ANOVA and logistic regression models; the effects of adherence were examined with Cox proportional hazard models with time-dependent covariates for teriparatide exposure. RESULTS: Among 14,284 teriparatide subjects, mean age was 68.4 years, 89.8% were female, and 29.6% had a fracture in the previous year; these characteristics were similar across MPR and persistence groups. The effects of adherence and persistence to teriparatide were statistically significant (P < .001) for all fracture types except wrist (P ≥ .125). By logistic regression, high vs low adherence was associated with reduced risk for any (OR = 0.67; P < .001); vertebral (OR = 0.64; P < .001); nonvertebral (OR = 0.71; P < .001); and hip fractures (OR = 0.52; P < .001) and longer (19–24 months) vs shorter persistence (1–6 months) was associated with reduced risk for any (OR = 0.63, P < .001); vertebral (OR = 0.56, P < .001); nonvertebral (OR = 0.69, P < .001); and hip fractures (OR = 0.48, P < .001). Cox models revealed a significantly reduced risk between high and low adherence for any (OR = 0.69, P < .001); vertebral (OR = 0.60, P < .001); nonvertebral (OR = 0.77, P < .001); and hip fractures (OR = 0.55, P < .001). CONCLUSION: Fracture incidence significantly decreased as persistence and adherence to teriparatide increased. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1007/s00198-016-3888-9) contains supplementary material, which is available to authorized users.