BACKGROUND Epidural clonidine produces effective postoperative analgesia in humans. Observed side effects include hypotension, bradycardia, sedation, and dryness of the mouth. A recent clinical study demonstrated that 150 micrograms intrathecal clonidine administered postoperatively as the sole analgesic agent was effective but produced hypotension and sedation. Animal studies have provided evidence of a biphasic effect on blood pressure after intrathecal clonidine administration, but no data concerning this effect in humans currently exist. This study was performed to evaluate the dose-response hemodynamic and analgesic profiles of intrathecal clonidine administered after a standard surgical intervention, without perioperative administration of additional analgesics, local anesthetics, or tranquilizers. METHODS In a randomized prospective double-blind study, 30 women who underwent elective cesarean section during general anesthesia with thiopental, nitrous oxide, and halothane were studied. Forty-five minutes after tracheal extubation, a lumbar intrathecal puncture was performed, and the patients received 150 (group 1), 300 (group 2), or 450 (group 3) micrograms clonidine. Postoperative analgesia was assessed on a visual analog scale at rest and after deep cough at standard time points up to 24 h. At the same time points, blood pressure, heart rate, sedation, and respiratory rate also were recorded. RESULTS Intrathecal clonidine decreased pain in all three groups both at rest and with coughing very shortly after injection, in a dose-dependent fashion. Clonidine 450 and 300 micrograms reduced pain scores significantly earlier (3rd and 6th min after intrathecal injection respectively), compared with 150 micrograms clonidine. Pain relief, defined as the time to first request for supplemental analgesic by patients, lasted 402 +/- 75 min in group 1, 570 +/- 76 min in group 2, and 864 +/- 80 min in group 3; significant differences among all groups; P < 0.01-0.001). Clonidine reduced mean arterial pressure compared with baseline only in group 1 (21 +/- 13%, P < 0.05). Delayed hypotension or bradycardia were not encountered after any of the three dose studies. Sedation was evident in all groups, but group 3 patients were significantly more sedated than group 1 and 2 patients. Respiratory rate and motor activity of the lower extremities were unaffected in all three groups (differences not significant). CONCLUSIONS These results demonstrate dose-dependent analgesia after intrathecal clonidine at doses as great as 450 micrograms. The nearly immediate analgesic effect observed after intrathecal injection of 300 and 450 micrograms clonidine strongly argues for a spinal rather than a systemic site of action of this alpha 2-adrenergic agonist. After 300 and 450 micrograms intrathecal clonidine a relative hemodynamic stability is observed, suggesting a pressor effect at peripheral sites.