After emergence from anesthesia, the incidence and severity of adverse airway effects caused by the laryngeal mask airway (LMA) can vary, depending on when the device was removed; nonetheless, reports differ regarding the exact optimal timing of LMA removal. The purpose of this study was to compare the rate of adverse events between 2 groups: those whose LMA was removed under general anesthesia (“deep” group) or under target-controlled infusion (TCI) of propofol (“awake” group).Institutional Review Board approval and written informed consent were obtained; 124 patients were then randomly allocated into either the “awake” group or the “deep” group. Anesthesia was induced and maintained using TCI of propofol, as well as intravenous fentanyl. In the “deep” group, the LMA was removed after surgery while the patients were deeply anesthetized using a target effect-site propofol concentration of 2 μg/mL, whereas in the “awake” group, the device was removed while the patients followed verbal instructions. The incidence of the following adverse events was recorded: coughing, straining, bronchospasm, laryngospasm, clenching, breath holding, gross purposeful movement, airway obstruction, retching, vomiting, and oxygen desaturation. If any such event occurred, the LMA removal was considered a failure. Airway hyperreactivity was recorded and graded – based on the severity of cough, breath holding, and oxygen desaturation.The failure rate was higher in the “awake” group (15/61; 24.6%) than in the “deep” group (5/60; 8.3%). Airway hyperreactivity was mild (score, <3) in both groups.Removal of the LMA under deep anesthesia using a target-controlled, effect-site propofol concentration of 2 μg/mL may be safer and more successful than removal when patients are fully awake after surgery.