Background This study was designed to define the appropriate dose of remifentanil hydrochloride alone or combined with midazolam to provide satisfactory comfort and maintain adequate respiration for a monitored anesthesia care setting. Methods One hundred fifty-nine patients scheduled for outpatient surgery participated in this multicenter, double-blind study. Patients were randomly assigned to one of two groups: remifentanil, 1 microgram/kg, given over 30 s followed by a continuous infusion of 0.1 microgram.kg-1.min-1 (remifentanil), remifentanil, 0.5 microgram/kg, given over 30 s followed by a continuous infusion of 0.05 microgram.kg-1.min-1 (remifentanil+midazolam). Five minutes after the start of the infusion, patients received a loading dose of saline placebo (remifentanil) or midazolam, 1 mg, (remifentanil+midazolam). If patients were not oversedated, a second dose of placebo or midazolam, 1 mg, was given. Remifentanil was titrated (in increments of 50% from the initial rate) to limit patient discomfort or pain intraoperatively, and the infusion was terminated at the completion of skin closure. Results At the time of the local anesthetic, most patients in the remifentanil and remifentanil+midazolam groups experienced no pain (66% and 60%, respectively) and no discomfort (66% and 65%, respectively). The final mean (+/-SD) remifentanil infusion rates were 0.12 +/- 0.05 microgram.kg-1.min-1 (remifentanil) and 0.07 +/- 0.03 microgram.kg-1.min-1 (remifentanil+midazolam). Fewer patients in the remifentanil+midazolam group experienced nauses compared with the remifentanil group (16% vs. 36%, respectively; P < 0.05). Four patients (5%) in the remifentanil group and two patients (2%) in the remifentanil+midazolam group experienced brief periods of oxygen desaturation (SpO2 < 90%) and hypoventilation (< 8 breaths/ min). Conclusions Remifentanil alone or combined with midazolam provided adequate analgesia and maintained adequate respiration at the doses reported. The low dose of remifentanil combined with 2 mg midazolam, compared with remifentanil alone, resulted in fewer side effects, slightly greater sedation, and less anxiety.