Traumatic Injury Clinical Trial Evaluating Tranexamic Acid in Children: An Efficacy Study

The TIC-TOC efficacy trial is a multicenter, adaptive allocation, randomized controlled trial of children younger than 18 years with hemorrhagic injuries to the torso and/or brain to evaluate the efficacy of TXA on functional outcome as measured by the PedsQL. Children will be randomized to one of three arms: 1) TXA 15 mg/kg bolus over 30 minutes, followed by a 2 mg/kg/hr infusion over 8 hours), 2) TXA 30 mg/kg bolus over 30 minutes, followed by a 4 mg/kg/hr infusion over 8 hours), and 3) normal saline placebo. A third TXA dose (45 mg/kg bolus dose over 30 minutes, followed by a 6 mg/kg/hr infusion over 8 hours) may be added later in the trial if a dose effect based on accumulating data is noted. The trial will be conducted in the Pediatric Emergency Care Applied Research Network (PECARN) across 40 sites over 4 years of enrollment for a maximum sample size of 2000 patients. A Bayesian adaptive randomization design will be used to evaluate the efficacy of TXA in children with hemorrhagic brain and/or torso injuries. Because different types of injury have different pathophysiology and potential response to TXA, three different injury strata will be evaluated: isolated hemorrhagic brain injury, isolated hemorrhagic torso injury, and both hemorrhagic brain and torso injuries. The efficacy of TXA will be analyzed across all enrolled children as well as across each type of injury. The Bayesian adaptive trial design also efficiently evaluates the effectiveness of TXA across different TXA doses. The trial will randomize the first 500 patients to two doses of TXA and placebo at a fixed 1:1:1 ratio. Interim analyses will be conducted when 500, 750, 1000, 1250, 1500, and 1750 patients have been enrolled. At each interim analysis, randomization probabilities will be adjusted in order to preferentially allocate patients to better performing doses, while allocation to the placebo arm will stay fixed. The adaptive randomization will be based entirely on pre-planned rules using accumulating data. A Bayesian hierarchical model will be used to estimate the treatment effect for each of the injury types to be informed by the data accumulated from all injury types. At interim analyses, if a dose effect is noted towards the higher dose of TXA (30 mg/kg bolus then a 4 mg/kg/hr infusion) being more efficacious using pre-specified criteria, then a higher dose study arm (TXA 45 mg/kg bolus then a 6 mg/kg/hr infusion) will be opened later in the trial. If the dose response curve is flat, suggesting that TXA is ineffective, then futility stopping rules can end the trial early.