Bone and joint infections (BJIs) are treated with intravenous antibiotics, which are burdensome and costly. No randomised controlled studies have compared if initial oral antibiotics are as effective as intravenous therapy. We aimed to investigate the efficacy and safety of initial oral antibiotics compared with initial intravenous antibiotics followed by oral antibiotics in children and adolescents with uncomplicated BJIs. From Sept 15, 2020, to June 30, 2023, this nationwide, randomised, non-inferiority trial included patients aged 3 months to 17 years with BJIs who presented to one of the 18 paediatric hospital departments in Denmark. Exclusion criteria were severe infection (ie, septic shock, the need for acute surgery, or substantial soft tissue involvement), prosthetic material, comorbidity, previous BJIs, or antibiotic therapy for longer than 24 h before inclusion. Patients were randomly assigned (1:1), stratified by C-reactive protein concentration (<35 mg/L vs ≥35 mg/L), to initially receive either high-dose oral antibiotics or intravenous ceftriaxone (100 mg/kg per day in one dose). High-dose oral antibiotics were coformulated amoxicillin (100 mg/kg per day) and clavulanic acid (12·5 mg/kg per day) in three doses for patients younger than 5 years or dicloxacillin (200 mg/kg per day) in four doses for patients aged 5 years or older. After a minimum of 3 days, and upon clinical improvement and decrease in C-reactive protein, patients in both groups received oral antibiotics in standard doses. The primary outcome was sequelae after 6 months in patients with BJIs, defined as any atypical mobility or function of the affected bone or joint, assessed blindly, in all randomised patients who were not terminated early due to an alternative diagnosis (ie, not BJI) and who attended the primary outcome assessment. A risk difference in sequelae after 6 months of less than 5% implied non-inferiority of the oral treatment. Safety outcomes were serious complications, the need for surgery after initiation of antibiotics, and treatment-related adverse events in the as-randomised population. This trial was registered with ClinicalTrials.gov, NCT04563325. 248 children and adolescents with suspected BJIs were randomly assigned to initial oral antibiotics (n=123) or initial intravenous antibiotics (n=125). After exclusion of patients without BJIs (n=54) or consent withdrawal (n=2), 101 patients randomised to oral treatment and 91 patients randomised to intravenous treatment were included. Ten patients did not attend the primary outcome evaluation. Sequelae after 6 months occurred in none of 98 patients with BJIs in the oral group and none of 84 patients with BJIs in the intravenous group (risk difference 0, one-sided 97·5% CI 0·0 to 3·8, pnon-inferiority=0·012). Surgery after randomisation was done in 12 (9·8%) of 123 patients in the oral group compared with seven (5·6%) of 125 patients in the intravenous group (risk difference 4·2%, 95% CI -2·7 to 11·5). We observed no serious complications. Rates of adverse events were similar across both treatment groups. In children and adolescents with uncomplicated BJIs, initial oral antibiotic treatment was non-inferior to initial intravenous antibiotics followed by oral therapy. The results are promising for oral treatment of uncomplicated BJIs, precluding the need for intravenous catheters and aligning with the principles of antimicrobial stewardship. Innovation Fund Denmark and Rigshospitalets Forskningsfond.