BACKGROUND: Kawasaki disease, the most common cause of acquired heart disease in developed countries, is a self-limited vasculitis that is treated with high doses of intravenous immunoglobulin. Resistance to intravenous immunoglobulin in Kawasaki disease increases the risk of coronary artery aneurysms. We assessed whether the addition of infliximab to standard therapy (intravenous immunoglobulin and aspirin) in acute Kawasaki disease reduces the rate of treatment resistance. METHODS: We undertook a phase 3, randomised, double-blind, placebo-controlled trial in two children's hospitals in the USA to assess the addition of infliximab (5 mg per kg) to standard therapy. Eligible participants were children aged 4 weeks-17 years who had a fever (temperature ≥38·0°C) for 3-10 days and met American Heart Association criteria for Kawasaki disease. Participants were randomly allocated in 1:1 ratio to two treatment groups: infliximab 5 mg/kg at 1 mg/mL intravenously over 2 h or placebo (normal saline 5 mL/kg, administered intravenously). Randomisation was based on a randomly permuted block design (block sizes 2 and 4), stratified by age, sex, and centre. Patients, treating physicians and staff, study team members, and echocardiographers were all masked to treament assignment. The primary outcome was the difference between the groups in treatment resistance defined as a temperature of 38·0°C or higher at 36 h to 7 days after completion of the infusion of intravenous immunoglobulin. Analysis was by intention to treat. This trial is registered with ClinicalTrials.gov, NCT00760435. FINDINGS: 196 patients were enrolled and randomised: 98 to the infliximab group and 98 to placebo. One patient in the placebo group was withdrawn from the study because of hypotension before receiving treatment. Treatment resistance rate did not differ significantly (11 [11·2%] for infliximab and 11 [11·3%] for placebo; p=0·81). Compared with the placebo group, participants given infliximab had fewer days of fever (median 1 day for infliximab vs 2 days for placebo; p<0·0001). At week 2, infliximab-treated patients had greater mean reductions in erythrocyte sedimentation rate (p=0·009) and a two-fold greater decrease in Z score of the left anterior descending artery (p=0·045) than did those in the placebo group, but this difference was not significant at week 5. Participants in the infliximab group had a greater mean reduction in C-reactive protein concentration (p=0·0003) and in absolute neutrophil count (p=0·024) at 24 h after treatment than did those given placebo, but by week 2 this difference was not significant. At week 5, none of the laboratory values differed significantly compared with baseline. No significant differences were recorded between the two groups at any timepoint in proximal right coronary artery Z scores, age-adjusted haemoglobin values, duration of hospital stay, or any other laboratory markers of inflammation measured. No reactions to intravenous immunoglobulin infusion occurred in patients treated with infliximab compared with 13 (13·4%) patients given placebo (p<0·0001). No serious adverse events were directly attributable to infliximab infusion. INTERPRETATION: The addition of infliximab to primary treatment in acute Kawasaki disease did not reduce treatment resistance. However, it was safe and well tolerated and reduced fever duration, some markers of inflammation, left anterior descending coronary artery Z score, and intravenous immunoglobulin reaction rates. FUNDING: US Food and Drug Administration, Robert Wood Johnson Foundation, and Janssen Biotech.