Midazolam is a new parenteral benzodiazepine premedication for endoscopy. Consecutive patients were randomized to receive either intravenous midazolam or diazepam as premedication for outpatient total colonoscopy by one endoscopist. Fifty-five patients received diazepam (0.15 mg/kg) and 50 received midazolam (0.07 mg/kg). Both patient and endoscopist were blind to the study drug used. The two groups were similar with respect to age, sex, and indication for colonoscopy. Patients were rated by the endoscopist for degree of cooperation, sedation, and pain during examination. There was significantly more oversedation in the midazolam group than in the diazepam group (p less than 0.05). Immediate procedure recall was less in midazolam patients (p less than 0.005), but on repeat interview the next day there was no difference between the two groups concerning recall of the endoscopy. There was no significant difference between the two groups in the incidence of arm pain. We conclude that in a clinical setting, midazolam does not appear to offer any significant advantage over diazepam, except for cost. Midazolam carries an increased risk of oversedation when it is administered on a milligram per kilogram basis and should instead be titrated individually.