Résumé : Induction, emergence and recovery characteristics were compared during sevoflurane or halothane anaesthetic in a large (428) multicentre, international study of children undergoing elective inpatient surgical procedures. Two hundred and fourteen children in each group underwent inhalation induction with nitrous oxide/oxygen and sevoflurane or halothane. Incremental doses of either study drug were added until loss of eyelash reflex was achieved. Steady state concentrations of anaesthesia were maintained until the end of surgery when anaesthetic agents were terminated simultaneously. Time variables were recorded for induction, emergence and the first need for analgesia in the recovery room. In addition, in 86 of the children in both groups, venous blood samples were drawn for plasma fluoride levels during and after surgery. There was a trend toward smoother induction (induction of anaesthesia without coughing, breath holding, excitement laryngospasm, bronchospasm, increased secretion, and vomiting) in the sevoflurane group with faster induction (2.1 min vs 2.9 min, P = 0.037) and rapid emergence times (10.3 min vs 13.9 min, P = 0.003). Among the children given sevoflurane, 2% developed bradycardia compared with 11% in the halothane group. Postoperatively, 46% of the children in the halothane group developed nausea and or vomiting versus 31% in the sevoflurane group (P = 0.002). Two children in the halothane group developed cardiac dysrhythmia and were dropped from the study. In addition, a child in the halothane group developed malignant hyperthermia, received dantrolene, and had an uneventful recovery. Mean maximum inorganic fluoride concentration was 18.3 μM·1 1. The fluoride concentrations peaked within one h of termination of sevoflurane anaesthetic and returned rapidly to baseline within 48 h. This study suggests that sevoflurane may be the drug of choice for the anaesthetic management of children.