Propofol anaesthesia is frequently associated with movement responses in non-paralysed patients. Opioids decrease the probability of movement during noxious stimulation. Heart rate variability and frontal electromyography (EMG), which are related to subcortical functions, may be more closely related than surface electroencephalography (EEG) to movement responses to noxious stimulation. Eighty-two patients scheduled for uterine dilatation and curettage were randomized to receive at the first intra-operative movement either a supplemental alfentanil bolus, 0.5 mg intravenously, or a supplemental propofol bolus, 0.7 mg/kg intravenously. The incidences of recurring movement during the procedure were compared between the two groups. The associations of a measure of heart rate variability (Anemon index), heart rate, EMG and two EEG variables with movement responses were evaluated. The incidences of recurring movement were 73% and 38% in the alfentanil and propofol groups, respectively [difference, 35%; 95% confidence interval, 9-56%; P= 0.014 between the groups). The Anemon index, heart rate, EMG and surface EEG variables displayed mainly reactive associations with movement responses. During uterine curettage under propofol-alfentanil-nitrous oxide anaesthesia, a propofol bolus of 0.7 mg/kg was more effective in preventing the recurrence of movement responses than an alfentanil bolus of 0.5 mg. Several physiological variables may be used to track significant arousal reactions, but not to predict them.