Midazolam, a new short-acting benzodiazepine with promising premedicant effects, was investigated in a double-blind, randomized clinical trial in 203 patients versus fentanyl/droperidol and placebo. Subjective effects, side-effects, amnesia and overall satisfaction were recorded. Midazolam caused the greatest decrease in anxiety level, and while causing more confusion and somnolence than placebo, caused less confusion and somnolence than fentanyl/droperidol. Half the patients who received midazolam reported anterograde amnesia. No serious side-effects were reported. Patient satisfaction was greater in the midazolam group than in the other groups.