The use of propofol alone or with alfentanil in the day-case anaesthesia for abortion was compared with that of ketamine with midazolam. Two hundred young women were assigned to two successive series of two groups each. The four groups were: group 1 (2 mg . kg-1 propofol only); group II (0.5 mg . kg-1 ketamine with 0.25 mg . kg-1 midazolam); group III (2 mg . kg-1 propofol with 4 micrograms . kg-1 alfentanil); group IV (1 mg . kg-1 ketamine with 0.1 mg . kg-1 midazolam). All the patients were premedicated one hour before anaesthesia with 0.25 mg . kg-1 midazolam orally. All the patients were asleep at the end of the propofol injection (60 s), and 10 to 15 s later for the ketamine-midazolam groups. The haemodynamic parameters did not vary much during induction with ketamine-midazolam. In the propofol groups, the heart rate remained steady, with an 8 to 12% fall in blood pressure. A fall of the mandible was seen in 40 and 84% of the patients in the propofol groups, with a short apnoea in 32 and 48% of these same patients. Clinical recovery was very quick, less than 12 min for all groups. The four psychomotor and sensory tests were carried out at the 30th min by 95% of the patients in the propofol groups, whereas only 50% of those in the ketamine-midazolam groups did so. Speed and quality were significantly better in the propofol groups. The most frequent adverse effect of propofol was pain during injection in 32 and 14% of patients.(ABSTRACT TRUNCATED AT 250 WORDS)