AbstractObjectivesCurrent guidelines rank abatacept, rituximab, tocilizumab and TNF-inhibitors (TNFi) as having equal effectiveness for the treatment of RA, at least as second line therapies. These recommendations are mainly based on meta-analysis of randomized controlled trials, with few direct drug–drug comparisons. Our objective was to compare the real-world absolute and relative effectiveness among RA patients starting any of the available biologic DMARDs (bDMARDs).MethodsWe used the Swedish Rheumatology Register to identify patients with RA initiating TNFi, rituximab, abatacept or tocilizumab in 2010–2016 as first bDMARD (n = 9333), or after switch from TNFi as first bDMARD (n = 3941). National Swedish registers provided additional covariates and censoring events. Effectiveness was assessed 3 and 12 months after treatment start, as the proportion remaining on therapy and with EULAR Good Response, HAQ improvement >0.2, zero swollen/tender joints and CDAI remission. Adjusted differences were estimated with multivariable linear regression.ResultsPatients starting non-TNFi (vs TNFi) as first bDMARD had a higher proportion remaining on drug and reaching most response outcomes as first bDMARD (1-year EULAR Good Response/HAQ improvement: TNFi 24.9/25.4%, rituximab 28.6/37.2%, abatacept 31.9/33.7%, tocilizumab 50.9/43.1%). After switch from a first TNFi, rituximab and tocilizumab, but not abatacept, were associated with significantly better response measures than TNFi (1-year EULAR Good Response/HAQ improvement: TNFi 11.6/16.1%, rituximab 24.8/33.2%, abatacept 13.1/17.5%, tocilizumab 34.1/29.4%). Differences remained significant after adjusting for potential confounders.ConclusionTreatment outcomes among RA patients treated in Swedish clinical practice are in line with a superior effectiveness of non-TNFi bDMARDs, in particular tocilizumab and rituximab, compared with TNFi.