To examine the relative efficacy of spinally administered alpha-2 adrenergic agonists [dexmedetomidine, clonidine and ST-91 (2-[2,6- diethylphenylamino]-2-imidazole)] on the 52.5 degrees C hot plate in rats, EEDQ (N-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline), an irreversible alpha-2 antagonist, was administered intrathecally (in doses of 0, 8.1, 81 and 810 nmol) 24 hr before the administration of these agonists. EEDQ alone results in an acute mild hyperalgesic effect. Intrathecal dexmedetomidine, clonidine and ST-91 result in a dose-dependent increase in the hot plate response latency: ED50 (95% confidence interval) = 13 (2.1-73), 120 (52-310) and 21 (1.3-240) nmol, respectively. Pretreatment with intrathecal EEDQ (8.1-810 nmol) caused a rightward shift of dose-response curve and reduction of maximal effect of alpha-2 agonists used. For example, 24 hr after 81 nmol of EEDQ, the degree of the rightward shift and the depression of slope was ST-91 greater than clonidine greater than dexmedetomidine. Analysis of the double reciprocal plot indicated that at the ED50, intrathecal dexmedetomidine, clonidine and ST-91 required an apparent occupancy of 10, 36 and 30%, respectively. These results suggest that dexmedetomidine has a higher intrinsic efficacy than clonidine or ST-91.