BACKGROUND Dexmedetomidine is a new potent and selective alpha 2-agonist that might prove useful as a preanesthetic agent. METHODS A randomized, double-blind study design was used in 192 ASA physical status 1 and 2 patients scheduled for elective abdominal hysterectomy, cholecystectomy, or intraocular surgery under general anesthesia. Intramuscular injection of 2.5 micrograms/kg dexmedetomidine administered 60 min before and intravenous saline placebo 2 min before induction of anesthesia (DEXPLA group, n = 64) was compared with a combination of 0.08 mg/kg intramuscular midazolam 60 min and 1.5 micrograms/kg intravenous fentanyl 2 min before induction (MIDFENT group, n = 64), or a combination of intramuscular dexmedetomidine and intravenous fentanyl (DEXFENT group, n = 64). After thiopental induction, anesthesia was maintained with 70% N2O/O2, and fentanyl was administered according to clinical and cardiovascular criteria. Patients undergoing cholecystectomy received additional enflurane. RESULTS Dexmedetomidine and midazolam induced comparable preoperative sedation and anxiolysis. The DEXFENT combination blunted the increases in blood pressure and heart rate induced by tracheal intubation more efficiently when compared with the DEXPLA and MIDFENT groups, in which approximately 25 mmHg and 15 beats/min greater increases were observed. The intraoperative fentanyl requirements were greater in MIDFENT patients when compared with both dexmedetomidine groups, in which 56% (DEXFENT group) and 31% (DEXPLA group) less fentanyl, respectively, was needed. Intraoperatively, fluids or vasopressors for hypotension and glycopyrrolate for bradycardia were administered more often to patients receiving dexmedetomidine than to those who did not. Postoperatively, there were no differences in oxygen saturation, analgesic, or antiemetic requirements, but dexmedetomidine-induced blood pressure and heart rate reductions were still evident at the end of the 3-h follow-up period. Bradycardia as an adverse event was reported more frequently in dexmedetomidine patients (20% in the DEXPLA and 33% in the DEXFENT groups) than in MIDFENT patients (8%). CONCLUSIONS The results suggest that pretreatment with a single intramuscular injection of 2.5 micrograms/kg dexmedetomidine is efficacious, but significantly increases the incidence of intraoperative hypotension and bradycardia in ASA physical status 1 or 2 patients.