Risk of respiratory tract infections and serious infections in psoriasis patients treated with biologics: Results from the BioCAPTURE registry

Abstract

Background: Limited real-world studies are available comparing infection risk between biologics for psoriasis.Objectives: The primary aim was to determine the differential effect of currently available biologics on the risk of respiratory tract infections (RTI) among psoriasis patients in a real-world setting. Secondary aims were to explore the differential risk of all types of serious infections (SI) between biologics and to provide an early overview of SARS-CoV-2 infections during the pre-vaccine era. Methods: Crude incidence rates of RTI and SI were calculated per 100 patient-years (PY) per biologic using prospective BioCAPTURE data. Negative Binomial Regression modeling was used to explore the risk of RTI. Frailty Cox proportional hazards modeling was used to estimate hazard ratios for the risk of the first SI. Confounders adjusted for both models were selected by a directed acyclic graph. A post hoc exploratory analysis of SARS-CoV-2 infection incidence rates during 2020 was performed. Results: We included 714 patients with 1325 treatment episodes (3607.7PY between 2005 and 2020), in which 2224 RTI and 63 SI occurred. Among RTI, 1.3% were serious. The crude incidence rates were 61.7 (95% confidence interval [CI]: 59.1-64.3) per 100PY for RTI, and 1.8 (95% CI: 1.4-2.2) per 100PY for SI. Confounder adjusted analyses showed no differential risk of RTI between adalimumab, etanercept, infliximab, ustekinumab, secukinumab, ixekizumab and guselkumab. For SI, no differential risk was found between biologics either. Extended single-center data showed 3.8 (95% CI: 2.2-6.1) SARS-CoV-2 infections per 100PY in 2020. Conclusions: Confounder adjusted analyses showed no differential risks of RTI or SI between included biologics (adalimumab, etanercept, infliximab, ustekinumab, secukinumab, ixekizumab and guselkumab) in a prospective psoriasis patients cohort. In general, absolute numbers of all types of SI were low.