Risk of hospitalized infection among rheumatoid arthritis patients concurrently treated with a biologic agent and denosumab.

Abstract

Denosumab is a biologic agent used to treat osteoporosis. Its safety profile given concurrently with biologic drugs for rheumatoid arthritis (RA) has not been well studied. We evaluated hospitalized infections among patients treated with biologic agents for RA who initiated denosumab or zoledronic acid (ZA), a parenteral bisphosphonate without known associations with infection. We hypothesized that the rate of hospitalized infection with denosumab would be noninferior to ZA. We identified RA patients enrolled in Medicare in 2006-2012 treated with biologic agents who initiated denosumab or ZA. Cox proportional hazards models compared the risk for hospitalized infection, comparing denosumab users to ZA users and adjusting for potentially confounding factors. A noninferiority margin was specified a priori to demonstrate that denosumab had no greater infection risk than ZA if the upper bound of the 95% confidence interval (95% CI) of the hazard ratio (HR) was <1.5. Eligible RA patients receiving biologic agents initiated denosumab (n = 1,354) or ZA (n = 4,460). Characteristics of the denosumab users were as follows: mean ± SD age 73.0 ± 8.9, 98.2% women, with a majority receiving infliximab (35.7%) or abatacept (18.6%). Denosumab users had a higher prevalence of prior infections (11.5% hospitalized and 48.3% outpatient) and infection-related risk factors. The crude rate of hospitalized infections for denosumab (14.9/100 person-years [95% CI 12.2-18.1]) was comparable to that for ZA (13.9/100 person-years [95% CI 12.5-15.4]). After adjustment, the HR of hospitalized infection for denosumab users was noninferior to that for ZA users (HR 0.89 [95% CI 0.69-1.15]). The rate of hospitalized infection among RA patients receiving denosumab concurrently with biologic agents for RA was not increased compared to those receiving zoledronate.