The perioperative effects of administering sedative and analgesic drugs prior to outpatient surgery were evaluated. One hundred fifty adult outpatients were randomly assigned to one of six study groups according to a double-blind protocol design. Patients were given placebo (saline) or midazolam (5 mg im) 30-60 min prior to surgery, and then either placebo, oxymorphone (1 mg iv), or fentanyl (100 micrograms iv) 3-5 min prior to a standardized anesthetic technique. Preoperatively, midazolam premedication was associated with a significantly lower anxiety level (37 +/- 29 mm vs. 50 +/- 32 mm, P less than 0.05), higher sedation level (254 +/- 136 mm vs. 145 +/- 109 mm, P less than 0.01), worsening of psychomotor skill (5 +/- 5 vs. 2 +/- 2 dots missed, P less than 0.01; midazolam vs. placebo), and impaired recall abilities. In addition, use of midazolam did not prolong the discharge time. Compared to control patients, those who received fentanyl had a decreased incidence of intraoperative airway difficulties such as coughing (28% vs. 0%, P less than 0.01). Although use of opioids increased the incidence of postoperative nausea (42% vs. 18%, P less than 0.01) and vomiting (23% vs. 2%, P less than 0.01; opioid vs. no opioid), average recovery times were not affected by opioid administration. Oxymorphone use was associated with a lower incidence of pain at home compared with that following fentanyl (46% vs. 74%, P less than 0.05). Finally, preoperative administration of both midazolam and fentanyl or oxymorphone prior to a standardized methohexital-nitrous oxide anesthetic technique did not adversely affect recovery after outpatient surgery.