Biological DMARD survival in rheumatoid arthritis patients in clinical practice: a METEOR registry data analysis.

To assess treatment survival of all available biologic disease modifying anti-rheumatic drugs (bDMARD) in patients with rheumatoid arthritis (RA) in a clinical practice scenario. Data were selected from the METEOR registry, an international registry capturing daily practice data of patients with a clinical diagnosis of RA. We selected patients ≥ 16 years old on bDMARDs, with complete data on start/stop date of medication and available data on disease activity measures. Time on treatment was calculated for the first treatment course of each bDMARD (infliximab, certolizumab, adalimumab, golimumab, etanercept, rituximab, tocilizumab and abatacept). Cox proportional hazards regression analysis was used to compare time to stop treatment between the different bDMARDs. For the current analysis, 9516 patients were available of which 49% used etanercept, 34% used adalimumab, 18% used infliximab, 14% used tocilizumab, 9% used abatacept, 8% used rituximab and 1% used golimumab. Median treatment duration per bDMARD was longest for infliximab (median 24.5 months), adalimumab (22.7 months) and etanercept (22.0 months). Fully adjusted Cox regression analyses showed that compared to infliximab, patients using abatacept, certolizumab, rituximab or tocilizumab had a significantly higher hazard to stop treatment, and thus had a shorter bDMARD treatment survival. Adalimumab, etanercept and golimumab had a similar hazard to stop treatment as infliximab. In this analysis of daily practice data, most TNF inhibitors (infliximab, adalimumab, etanercept and golimumab) had similar treatment survival in fully adjusted models. Patients using abatacept, certolizumab, rituximab or tocilizumab had a lower treatment survival compared to infliximab.