Background While biologic therapy, typically with infliximab or Roferon, was used for Behçet syndrome after first-line immunosuppressants, no high-quality randomised trials or predictive biomarkers were available. Objective To undertake a randomised controlled clinical trial of infliximab versus Roferon in Behçet syndrome and identify potential biomarkers for response. Design Pragmatic, standard of care, single-masked, randomised, two-arm, parallel head-to-head trial, with exploratory study on potential role of interferon lambda 3 and interferon lambda 4 single nucleotide polymorphisms and urinary metabolomics biomarkers. Setting Three national UK Behçet syndrome centres and allied clinics. Participants Patients with active Behçet syndrome, fulfilling International Study Group 1990 criteria, with inadequate response to or intolerance of first-line treatment. Intervention Randomisation to infliximab (5 mg/kg intravenous infusion) or Roferon (subcutaneous injection), utilising the UK Behçet syndrome drug pathway protocol. Outcomes Primary outcome: modified Behçet’s disease activity index at 12 weeks of therapy. Secondary outcomes: (1) modified Behçet’s disease activity index score at 24 weeks and (2) significant improvement at 12 and 24 weeks from baseline in vitreous haze and best corrected visual acuity change, oral ulcer severity score, number of genital ulcers, arthritis pain, adverse events, reduction in dose of glucocorticoid, quality-of-life scores and Physician’s Global Assessment of disease activity. Sample size Utilising a Bayesian analysis of covariance model (80% credible interval), initial sample size was 45/arm (Bayesian power 90%). With an anticipated 10% dropout rate, 100 patients were to be recruited. Following recommendations to reduce the overall length of the trial, this was revised down to 80 patients (36 in each arm, allowing for 10% dropout): 80% equi-tailed credibility interval, Bayesian power 88%. In total, 79 patients were eventually recruited for the study. Methods Patients with refractory active Behçet syndrome underwent stratified block randomisation, based on randomly permuted blocks with random block sizes of two and four, allocating treatment to either infliximab or Roferon. Follow up with symptom-directed examination at weeks 12 and 24 according to standard of care. Analysis of the primary end point was undertaken using a Bayesian analysis of covariance approach. Informative priors for the anticipated treatment effect were derived from a cohort of six international experts prior to the start of the study. Results In this first prospective head-to-head randomised controlled clinical trial of two biologic drugs in Behçet syndrome, both infliximab and Roferon were equally effective [mean difference (80% credibility interval) = 0.13 (–0.19 to 0.46)], with a trend for minor benefit in favour of infliximab in terms of tolerability and treatment persistence. Genetic data suggested a potential association between patient outcome and carriage of either rs4803221 or rs7248668 variants in the interferon lambda 3 (interleukin 28B) gene locus in the Roferon-treated arm. However, with the relatively small sample size, statistical significance of the association was lost when correcting for multiple tests. Metabolomic analysis identified potential markers of a metabolic response to treatment with infliximab. Limitations Single-masked design. Slow recruitment with fewer patients recruited in total, limiting the strength of analysis for secondary outcomes and mechanistic studies. Conclusion We report clinical efficacy in both infliximab and Roferon in refractory active Behçet syndrome, together with the potential for a novel metabolomic biomarker identifying response to infliximab. Future work Further work will characterise the appropriate metabolite(s) from existing samples to inform future prospective trials to study this in more detail clinically. The efficacy of Roferon in Behçet syndrome may support future manufacture of this drug. Trial registration This trial is registered as EudraCT Number: 2014-005390-36; ISRCTN49793874. Funding This award was funded by the National Institute for Health and Care Research (NIHR) Efficacy and Mechanism Evaluation (EME) programme (NIHR award ref: 12/205/46) and is published in full in Efficacy and Mechanism Evaluation; Vol. 11, No. 17. See the NIHR Funding and Awards website for further award information. Plain language summary Behçet syndrome, a very rare disease in the UK, causes major illness. Yet without high-quality, randomised, controlled trials, choosing treatment is somewhat hit and miss. We set up a randomised controlled clinical trial to study the two most widely used biologic drugs in Behçet syndrome, infliximab and Roferon, head-to-head and searched for potential blood and urinary markers for response. Patients with active Behçet syndrome, in the United Kingdom national Behçet syndrome centres and allied clinics, who had not responded to or could not tolerate first-line treatment were randomised to either infliximab infusions or Roferon injections. The primary outcome was modified Behçet’s disease activity index at 12 weeks of therapy. Secondary outcomes included modified Behçet’s disease activity index at 24 weeks and significant improvements in individual organs, quality of life and Physician’s Global Assessments of activity at 12 and 24 weeks. Initial assessment suggested 100 patients were required for a statistically meaningful outcome but was revised down to 80 following recommendations to shorten the trial. In this first prospective head-to-head randomised controlled clinical trial of two biologics in Behçet syndrome, both drugs worked equally well. There was a non-significant trend for minor benefits of infliximab in terms of tolerability and treatment persistence. Genetic data suggested a potential association between patient outcome and carriage of either rs4803221 or rs7248668 variants in the interferon lambda 3 (interleukin 28B) gene locus in the Roferon arm, but statistical significance was lost with the relatively small sample size. Metabolomics analysis identified potential markers of a metabolic response to infliximab. The limitations of the study included the single-masked design: patients (but not clinicians) were aware of their treatment, and fewer patients were studied than planned. This limited the strength of analysis for secondary outcomes and mechanistic studies. We now plan to characterise the metabolite(s) from existing samples to design future trials to study if there can be effective targeting of treatment in Behçet syndrome. Scientific summary Background Behçet syndrome (BS), a multisystem inflammatory vasculitis, is infrequent in the UK, but it has the potential to cause significant morbidity and mortality. The evidence base supporting biologic treatment, which is used for active disease after failure of standard immunosuppression or when prognosis is poor, is largely based on uncontrolled studies. At the time of the trial, the UK National guideline for therapy of Behçet’s indicated that either the tumour necrosis factor alpha inhibitor infliximab or the interferon alpha drug Roferon could be employed as treatment for patients following failure of first-line treatment with standard immunomodulators. The Bio Behçet’s trial was conceived to exploit the opportunity of the three UK National Centres of Excellence for Behçet’s and associated satellite centres to undertake the first randomised clinical trial to compare infliximab and Roferon as treatment for BS, together with an exploratory analysis of potential genomic and metabolomic biomarkers of therapeutic response. Methods The Bio Behçet’s trial is a pragmatic, standard of care, single-masked, randomised, two-arm, parallel trial comparing the efficacies of infliximab and Roferon employed after failure of first-line therapy in BS. Patients with BS, diagnosed according to the International Study Group 1990 criteria, with active disease who had failed to respond to, or were intolerant of, first-line treatment of BS with topical steroids or small-molecule immunosuppressants were randomised to treatment with either infliximab (Remicade) 5 mg/kg intravenous infusions or Roferon subcutaneous injection (in variable dose), utilising the treatment protocol for each of these drugs in normal clinical care as detailed in the BS drugs pathway for England. The trial utilised a Bayesian design utilising priors informed by a small survey of international experts in BS. Utilising a Bayesian analysis of covariance model, with an 80% credible interval, a sample size of 45 patients per arm was deemed appropriate and gave a Bayesian power of 90%. Allowing for an anticipated 10% dropout rate, 100 patients were planned to be recruited but reduced to 80 following recommendations to reduce the overall length of the trial. Allowing for a 10% dropout rate, estimates of study power based on 72 evaluable patients (36 on each arm) and an 80% credible interval a Bayesian power of 88% was obtained. Between June 2016 and February 2020, 161 patients were screened, and 79 patients were randomised. The intention-to-treat analysis was restricted to 37 subjects allocated to infliximab and 37 to Roferon. Based on previous work with hepatitis C infection and response to interferon therapy and given the role of the innate immune system in the pathogenesis of BS, we examined interferon lambda 3 (IFNL3) and interferon lambda 4 (IFNL4) single nucleotide polymorphisms (SNPs) as biomarkers of response to treatment with alpha interferon and/or infliximab in BS. We also examined the potential for urine metabolomics to act as biomarkers for drug response. The primary outcome was a modified version of the Behçet’s disease activity index (mBDAI) after 3 months of therapy. Secondary outcomes comprised mBDAI score after 6 months of therapy and significant improvement in organ systems after 3 and 6 months (week 12 and week 24 visits) assessed by: reduction in vitreous haze using the SUN consensus group grading scale and best corrected visual acuity change [using the logarithm of the minimal angle of resolution (LogMAR) chart at 4 m] from baseline; change in oral ulcer severity score; change in number of genital ulcers; arthritis pain (10 cm Likert scale); adverse events (AEs) in each group; reduction in dose of prednisolone (or equivalent glucocorticoid) at 3 months (week 12 visit); reduction in dose of prednisolone (or equivalent glucocorticoid) at 6 months (week 24 visit); quality-of-life (QoL) scores at 3 and 6 months (week 12 and week 24 visits) and Physician’s Global Assessment of disease activity at 3 and 6 months (week 12 and week 24 visits). Results Baseline characteristics of the two treatment arms did not differ significantly by sex, ethnic profile, baseline disease characteristics and steroid use. For the primary outcome measure, change in mBDAI between baseline and 3 months (and as a secondary outcome between baseline and 6 months) did not differ significantly between the two treatments [mean difference (80% CrI) = 0.13 (–0.19 to 0.46)]. A significantly higher proportion of patients randomised to Roferon swapped away from their randomised treatment compared to those randomised to infliximab treatment (Roferon 11 of 37, infliximab 3 of 37; p = 0.0104). The clinician’s overall perception of disease activity indicated a reduction in disease activity for most patients between baseline and 3 months and between baseline and 6 months, with a median reduction of −2.0 (infliximab) and −1.0 (Roferon) at 3 months and −3.0 (infliximab) and −2.0 (alpha interferon) at 6 months. There was a small but significant difference in favour of infliximab compared to Roferon at both 3 and 6 months (p