Background In real-life world settings, failure of a first TNF-inhibitor (TNFi) put rheumatologists against a crossroad to choose a further TNFi or a biologic drug with a different mechanism of action.
Objectives Aim of this study was to assess effectiveness of golimumab (GOL) as second line drug after failure of a first TNFi as treatment of patients affected with rheumatoid arthritis (RA), psoriatic arthritis (PsA), or axial-spondyloarthritis (Ax-SpA).
Methods GOAREL is a prospective cohort of patients starting GOL in community-based care rheumatology centers in Apulia (south Italy) since 2013. Of 494, we selected 368 patients (RA n.73, PsA n.168, and n. Ax-SpA 127) commencing GOL as first ever biological (n.206, 56%) or second treatment (n.161, 44%) after inadequate response to a first TNFi (adalimumab (ADA, n.51), etanercept (ETA, n.81), or infliximab (IFX, n.29). Three patients failing certolizumab were excluded. Primary endpoint was to compare the drug retention rate of biological naïve and TNFi inadequate responders (TNFi-IR) patients on treatment with GOL. In addition, drug retention of second line GOL patients according to the firs TNFi was assessed. Drug survival was estimated by Kaplan-Meier life table analysis. Estimates hazard ratios (HRs) of 2- years drug discontinuation adjusted for demographics, type of disease, disease characteristics, body mass index, co-therapy with glucocorticoids or methotrexate, and prior TNFi were computed by backward selection Cox-regression model.
Results 2-year drug survival on GOL of naïve and TNF-IR patients was not significantly different (Figure 1). Mean survival time was 20.0 months (95% CI 18.9–21.0) for naïve and 20.4 months (95% CI 19.3–21.6) for TNFi-IR patients. Likewise, drug retention rates on GOL of TNFi-IR patients subdivided by previous TNFi was not significantly different and similar to naïve patients (Figure 1). Mean survival time was 19.0 months (95% CI 16.7–21.4) for prior-ADA, 20.5 months (95% CI 18.9–22.1) for prior-ETA, and 22.0 months (95% CI 19.6–24.3) for prior-IFX. Multiple Cox-regression model showed gender female as the only independent factor positively associated to the risk of GOL discontinuation (HR 2.5, 95% CI 1.5–4.2, p 0.0001).
Conclusions In real-life setting effectiveness of GOL seems to be similar in naïve or TNFi-IR patients with RA, PsA, and Ax-Spa. Furthermore, clinical outcomes were not influenced by the previous TNFi.
Disclosure of Interest None declared