Background: Rectal diazepam and buccal midazolam are used for emergency treatment of acute febrile and afebrile (epileptic) seizures in children. We aimed to compare the safety and efficacy of these drugs. Methods: A multicentre, randomised controlled trial was undertaken to compare buccal midazolam with rectal diazepam for emergency-room treatment of children aged 6 months and older presenting to hospital with active seizures and without intravenous access. The dose varied according to age from 2.5 to 10 mg. The primary endpoint was therapeutic success: cessation of seizures within 10 min and for at least 1 hour, without respiratory depression requiring intervention. Analysis was per protocol. Findings: Consent was obtained for 219 separate episodes involving 177 patients, who had a median age of 3 years (IQR 1-5) at initial episode. Therapeutic success was 56%(61 of 109) for buccal midazolam and 27%(30 of 110) for rectal diazepam (percentage difference 29%, 95%CI 16-41). Analysing only initial episodes revealed a similar result. The rate of respiratory depression did not differ between groups. When centre, age, known diagnosis of epilepsy, use of antiepileptic drugs, prior treatment, and length of seizure before treatment were adjusted for with logistic regression, buccal midazolam was more effective than rectal diazepam. Interpretation: Buccal midazolam was more effective than rectal diazepam for children presenting to hospital with acute seizures and was not associated with an increased incidence of respiratory depression. Background: Rectal diazepam and buccal midazolam are used for emergency treatment of acute febrile and afebrile (epileptic) seizures in children. We aimed to compare the safety and efficacy of these drugs. Methods: A multicentre, randomized controlled trial was undertaken to compare buccal midazolam with rectal diazepam for emergency-room treatment of children aged 6 months and older presenting to hospital with active seizures and without intravenous access. The dose varies according to age from 2.5 to 10 mg. The primary endpoint was therapeutic success: cessation of seizures within 10 min and for at least 1 hour, without respiratory depression requiring intervention. Analysis was per protocol. Findings: Consent was obtained for 219 separate episodes involving 177 patients, who had a median age of 3 years (IQR 1-5) at initial episode. Therapeutic success was 56% (61 of 109) for buccal midazolam and 27% (30 of 110) for rectal diazepam (percentage difference 29%, 95% CI 16-41). Analysing on The rate of respiratory depression did not differ between groups. When center, age, known diagnosis of epilepsy, use of antiepileptic drugs, prior treatment, and length of seizure before treatment were adjusted for with logistic regression, Buccal midazolam was more effective than rectal diazepam. Interpretation: Buccal midazolam was more effective than rectal diazepam for children presenting to hospital with acute seizures and was not associated with an increased incidence of respiratory depression.