Possible mechanisms of the ability of clonidine to correct bupivacaine-induced ventricular electrophysiologic impairment were evaluated in an electrophysiologic model on closed-chest dogs. Nine groups (n = 6) of pentobarbital-anesthetized dogs were given atropine, 0.2 mg/kg intravenously (i.v.), and bupivacaine, 4 mg/kg i.v., over a 10-s period. Group 1 was then given only saline solution. Group 2 was given clonidine, 0.01 mg/kg i.v., over a 1-min period. Group 3 was given clonidine followed by i.v. administration of yohimbine, 1 mg/kg, an alpha 2-antagonist. Group 4 was given carbachol, 1 mg/kg i.v., a long-lasting cholinergic agonist, over a 1-min period. Group 5 was given electrical stimulation of the left vagus nerve. Group 6 was given physostigmine, 0.1 mg/kg i.v., known to inhibit cholinesterase degradation, 5 min before bupivacaine administration, and Group 7 received a combination of physostigmine pretreatment and electrical vagal stimulation. Group 8 was given physostigmine, 0.1 mg/kg i.v., and pancuronium bromide, 1 mg/kg i.v., known to inhibit nicotinic receptors, 5 min before bupivacaine administration. Then electrical stimulation of the left vagus nerve was performed. Group 9 was given nicotine, 0.1 mg/kg i.v., 1 min after bupivacaine injection over 1 min. Bupivacaine induced bradycardia, markedly increased the His-Purkinje conduction time (HV interval) and QRS duration. Bupivacaine decreased the peak of first derivative of left ventricle pressure (LVdP/dtmax) and increased left ventricular end-diastolic pressure (LVEDP). Clonidine improved QRS duration and HV interval. Yohimbine did not modify the effects of clonidine. QRS duration and HV interval were significantly improved in Groups 4-7. In Group 8, pancuronium pretreatment inhibited the beneficial effects of the combination of physostigmine pretreatment and electrical vagal stimulation. In contrast, in Group 9, like clonidine, nicotine improved QRS duration and HV interval. Three other groups of anesthetized dogs (n = 6) were then studied. All dogs were given hexamethonium, 10 mg/kg i.v. Then, Group 10 was given only saline solution; Group 11 was given bupivacaine, 4 mg/kg, and Group 12 was given bupivacaine and nicotine as in Group 9. In Group 11, bupivacaine induced its usual alterations. In contrast, nicotine did not modify the cardiotoxic profile of bupivacaine after hexamethonium pretreatment. We conclude that the beneficial effect of clonidine on the variables of ventricular conduction altered by bupivacaine 1) is not mediated by central alpha 2-activation, 2) is mediated by the activation of parasympathetic pathways, and 3) is indirect and not mediated by acetylcholine release but is mediated by the activation of parasympathetic ganglionic nicotinic receptors.(ABSTRACT TRUNCATED AT 400 WORDS)