BACKGROUND: Early, intensive treatment of rheumatoid arthritis (RA) with the combination of (initially high dose) prednisolone, methotrexate and sulfasalazine (COBRA therapy) considerably lowers disease activity and suppresses radiological progression, but is infrequently prescribed in daily practice. Attenuating the COBRA regimen might lessen concerns about side effects, but the efficacy of such strategies is unknown. OBJECTIVE: To compare the 'COBRA-light' strategy with only two drugs, comprising a lower dose of prednisolone (starting at 30 mg/day, tapered to 7.5 mg/day in 9 weeks) and methotrexate (escalated to 25 mg/week in 9 weeks) to COBRA therapy (prednisolone 60 mg/day, tapered to 7.5 mg/day in 6 weeks, methotrexate 7.5 mg/week and sulfasalazine 2 g/day). METHOD: An open, randomised controlled, non-inferiority trial in 164 patients with early active RA, all treated according to a treat to target strategy. RESULTS: At baseline patients had moderately active disease: mean (SD) 44-joint disease activity score (DAS44) 4.13 (0.81) for COBRA and 3.95 (0.9) for COBRA-light. After 6 months, DAS44 significantly decreased in both groups (-2.50 (1.21) for COBRA and -2.18 (1.10) for COBRA-light). The adjusted difference in DAS44 improvement between the groups, 0.21 (95% CI -0.11 to 0.53), was smaller than the predefined clinically relevant difference of 0.5. Minimal disease activity (DAS44 <1.6) was reached in almost half of patients in both groups (49% and 41% in COBRA and COBRA-light, respectively). CONCLUSIONS: At 6 months COBRA-light therapy is most likely non-inferior to COBRA therapy. CLINICAL TRIAL REGISTRATION NUMBER: 55552928.