Rheumatoid arthritis (RA), the most common autoimmune disease in the UK, is a chronic systemic inflammatory arthritis that affects 0.8% of the UK population. To determine whether or not an alternative class of biologic disease-modifying antirheumatic drugs (bDMARDs) are comparable to rituximab in terms of efficacy and safety outcomes in patients with RA in whom initial tumour necrosis factor inhibitor (TNFi) bDMARD and methotrexate (MTX) therapy failed because of inefficacy. Multicentre, Phase III, open-label, parallel-group, three-arm, non-inferiority randomised controlled trial comparing the clinical and cost-effectiveness of alternative TNFi and abatacept with that of rituximab (and background MTX therapy). Eligible consenting patients were randomised in a 1 : 1 : 1 ratio using minimisation incorporating a random element. Minimisation factors were centre, disease duration, non-response category and seropositive/seronegative status. UK outpatient rheumatology departments. Patients aged ≥ 18 years who were diagnosed with RA and were receiving MTX, but had not responded to two or more conventional synthetic disease-modifying antirheumatic drug therapies and had shown an inadequate treatment response to a first TNFi. Alternative TNFi, abatacept or rituximab (and continued background MTX). The primary outcome was absolute reduction in the Disease Activity Score of 28 joints (DAS28) at 24 weeks post randomisation. Secondary outcome measures over 48 weeks were additional measures of disease activity, quality of life, cost-effectiveness, radiographic measures, safety and toxicity. Owing to third-party contractual issues, commissioning challenges delaying centre set-up and thus slower than expected recruitment, the funders terminated the trial early. Between July 2012 and December 2014, 149 patients in 35 centres were registered, of whom 122 were randomised to treatment (alternative TNFi, <i>n</i> = 41; abatacept, <i>n</i> = 41; rituximab, <i>n</i> = 40). The numbers, as specified, were analysed in each group [in line with the intention-to-treat (ITT) principle]. Comparing alternative TNFi with rituximab, the difference in mean reduction in DAS28 at 24 weeks post randomisation was 0.3 [95% confidence interval (CI) -0.45 to 1.05] in the ITT patient population and -0.58 (95% CI -1.72 to 0.55) in the per protocol (PP) population. Corresponding results for the abatacept and rituximab comparison were 0.04 (95% CI -0.72 to 0.79) in the ITT population and -0.15 (95% CI -1.27 to 0.98) in the PP population. General improvement in the Health Assessment Questionnaire Disability Index, Rheumatoid Arthritis Quality of Life and the patients' general health was apparent over time, with no notable differences between treatment groups. There was a marked initial improvement in the patients' global assessment of pain and arthritis at 12 weeks across all three treatment groups. Switching to alternative TNFi may be cost-effective compared with rituximab [incremental cost-effectiveness ratio (ICER) £5332.02 per quality-adjusted life-year gained]; however, switching to abatacept compared with switching to alternative TNFi is unlikely to be cost-effective (ICER £253,967.96), but there was substantial uncertainty in the decisions. The value of information analysis indicated that further research would be highly valuable to the NHS. Ten serious adverse events in nine patients were reported; none were suspected unexpected serious adverse reactions. Two patients died and 10 experienced toxicity. The results will add to the randomised evidence base and could be included in future meta-analyses. How to manage first-line TNFi treatment failures remains unresolved. Had the trial recruited to target, more credible evidence on whether or not either of the interventions were non-inferior to rituximab may have been provided, although this remains speculative. Current Controlled Trials ISRCTN89222125 and ClinicalTrials.gov NCT01295151. This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in <i>Health Technology Assessment</i>; Vol. 22, No. 34. See the NIHR Journals Library website for further project information.