Introduction Up to 40% of orthopaedic injuries in children require a closed reduction, almost always necessitating procedural sedation. Intravenous ketamine is the most commonly used sedative agent. However, intravenous insertion is painful and can be technically difficult in children. We hypothesise that a combination of intranasal dexmedetomidine plus intranasal ketamine (Ketodex) will be non-inferior to intravenous ketamine for effective sedation in children undergoing a closed reduction. Methods and analysis This is a six-centre, four-arm, adaptive, randomised, blinded, controlled, non-inferiority trial. We will include children 4-17 years with a simple upper limb fracture or dislocation that requires sedation for a closed reduction. Participants will be randomised to receive either intranasal Ketodex (one of three dexmedetomidine and ketamine combinations) or intravenous ketamine. The primary outcome is adequate sedation as measured using the Paediatric Sedation State Scale. Secondary outcomes include length of stay, time to wakening and adverse effects. The results of both per protocol and intention-to-treat analyses will be reported for the primary outcome. All inferential analyses will be undertaken using a response-adaptive Bayesian design. Logistic regression will be used to model the dose-response relationship for the combinations of intranasal Ketodex. Using the Average Length Criterion for Bayesian sample size estimation, a survey-informed non-inferiority margin of 17.8% and priors from historical data, a sample size of 410 participants will be required. Simulations estimate a type II error rate of 0.08 and a type I error rate of 0.047. Ethics and dissemination Ethics approval was obtained from Clinical Trials Ontario for London Health Sciences Centre and McMaster Research Ethics Board. Other sites have yet to receive approval from their institutions. Informed consent will be obtained from guardians of all participants in addition to assent from participants. Study data will be submitted for publication regardless of results. Trial registration number NCT0419525.