Tramadol hydrochloride produced dose-related antinociception in mouse abdominal constriction [ED50 = 1.9 (1.2-2.6) mg/kg i.p.], hot-plate [48 degrees C, ED50 = 21.4 (18.4-25.3) mg/kg s.c.; 55 degrees C, ED50 = 33.1 (28.2-39.1) mg/kg s.c.] and tail-flick [ED50 = 22.8 (19.2-30.1) mg/kg s.c.] tests. Tramadol also displayed antinociceptive activity in the rat air-induced abdominal constriction [ED50 = 1.7 (0.7-3.2) mg/kg p.o.] and hot-plate [51 degrees C, ED50 = 19.5 (10.3-27.5) mg/kg i.p.] tests. The antinociceptive activity of tramadol in the mouse tail-flick test was completely antagonized by naloxone, suggesting an opioid mechanism of action. Consistent with this, tramadol bound with modest affinity to opioid mu receptors and with weak affinity to delta and kappa receptors, with Ki values of 2.1, 57.6 and 42.7 microM, respectively. The pA2 value for naloxone obtained with tramadol in the mouse tail-flick test was 7.76 and was not statistically different from that obtained with morphine (7.94). In CXBK mice, tramadol, like morphine, was devoid of antinociceptive activity after intracerebroventricular administration, suggesting that the opioid component of tramadol-induced antinociception is mediated by the mu-opioid receptor. In contrast to the mouse tail-flick test and unlike morphine or codeine, tramadol-induced antinociception in the mouse abdominal constriction, mouse hot-plate (48 degrees or 55 degrees C) or rat hot-plate tests was only partially antagonized by naloxone, implicating a nonopioid component. Further examination of the neurochemical profile of tramadol revealed that, unlike morphine, it also inhibited the uptake of norepinephrine (Ki = 0.79 microM) and serotonin (0.99 microM). The possibility that this additional activity contributes to the antinociceptive activity of tramadol was supported by the finding that systemically administered yohimbine or ritanserin blocked the antinociception produced by intrathecal administration of tramadol, but not morphine, in the rat tail-flick test. These results suggest that tramadol-induced antinociception is mediated by opioid (mu) and nonopioid (inhibition of monoamine uptake) mechanisms. This hypothesis is consistent with the clinical experience of a wide separation between analgesia and typical opioid side effects.