BACKGROUNDData from head-to-head trials of immunomodulatory therapies for Behçet's disease are scarce. We aimed to compare the efficacy and safety of ciclosporin, interferon alfa-2a, and adalimumab, each combined with corticosteroids, in preventing uveitis relapse in patients with severe Behçet's disease.METHODSWe did a randomised, open-label, assessor-masked, head-to-head trial at a large, specialised uveitis centre in Chongqing, China. Patients aged 18 years or older with severe Behçet's disease uveitis on corticosteroids and naive to anti-TNF therapy were eligible. Patients were randomly assigned in a 1:1:1 ratio to ciclosporin (2-5 mg/kg per day orally), interferon alfa-2a (3 million IU per day subcutaneously), or adalimumab (40 mg every 2 weeks subcutaneously), each combined with a tapering dose of corticosteroids with subsequent dose adjustments. The primary outcome was the annualised relapse rate of uveitis, assessed in the full analysis set (all randomly assigned patients with at least one post-baseline assessment). The non-inferiority margin of difference between the interferon alfa-2a and adalimumab groups was set to 1·0 for the primary outcome. Safety was assessed in all patients who received at least one dose of trial drugs. Individuals with lived experience of Behçet's disease uveitis were involved in the trial design and implementation. This study is registered with Chinese Clinical Trial Registry, ChiCTR2000031637. The trial is ongoing, but is closed to new participants.FINDINGSBetween May 12, 2020, and Feb 22, 2022, a total of 270 patients (mean age 38·1 years [SD 9·8]; 213 [79%] men, 57 [21%] women; 270 [100%] east Asian ethnicity) were randomly assigned to ciclosporin, interferon alfa-2a, or adalimumab (n=90 in each group); 261 patients were included in the full analysis set. For the primary outcome, the least-squares mean was 1·84 (95% CI 1·40 to 2·44) with ciclosporin, 1·44 (1·10 to 1·89) with interferon alfa-2a, and 0·95 (0·64 to 1·40) with adalimumab. The annualised relapse rate was significantly higher in patients receiving ciclosporin than in those receiving adalimumab (least-squares mean difference 0·90 [95% CI 0·27 to 1·53]; p=0·0054 for superiority). The least-squares mean difference between interferon alfa-2a and adalimumab was 0·50 (-0·04 to 1·04), which did not meet non-inferiority criteria (p=0·034 for non-inferiority). The primary outcome did not differ substantially between interferon alfa-2a and ciclosporin (least-squares mean difference -0·40 [-1·05 to 0·25]; p=0·23 for superiority). Serious adverse events were reported in 12 (13%) of 90 patients on ciclosporin plus corticosteroids, eight (9%) of 90 patients on interferon alfa-2a plus corticosteroids, and seven (8%) of 90 patients on adalimumab plus corticosteroids. There were no treatment-related deaths.INTERPRETATIONAdalimumab plus corticosteroids was superior to ciclosporin plus corticosteroids with respect to uveitis relapse rate in patients with severe Behçet's disease naive to anti-TNF therapy, and interferon alfa-2a plus corticosteroids was not found to be non-inferior to adalimumab plus corticosteroids or superior to ciclosporin plus corticosteroids.FUNDINGNational Natural Science Foundation of China Key Program, Major Program of Medical Science and Technology Project of Health Commission of Henan Province, Chongqing Key Laboratory of Ophthalmology, and China National Postdoctoral Program for Innovative Talents.