This randomised, double-blind study compared dexmedetomidine and midazolam for intravenous sedation during third molar surgery under local anaesthesia. Sixty patients received either dexmedetomidine (up to 1 microg x kg(-1)) or midazolam (up to 5 mg), which was infused until the Ramsay Sedation Score was four or the maximum dose limit was reached. Intra-operative vital signs, postoperative pain scores and analgesic consumption, amnesia, and satisfaction scores for patients and surgeons, were recorded. Sedation was achieved by median (IQR (range)) doses of 47 microg (39-52 (25-76)) or 0.88 microg x kg(-1) (0.75-1.0 (0.6-1.0)) dexmedetomidine, and 3.6 mg (3.3-4.4 (1.9-5.0)) or 0.07 mg x kg(-1) (0.055-0.085 (0.017-0.12)) midazolam. Heart rate and blood pressure during surgery were lower in dexmedetomidine group. There was no significant difference in satisfaction or pain scores. Midazolam was associated with greater amnesia. Dexmedetomidine produces comparable sedation to midazolam.