BACKGROUND: alpha-Bungarotoxin, decamethonium or lidocaine has a neuromuscular blocking effect. The aim of this study was to evaluate the pharmacodynamic properties of these drugs at the neuromuscular junction and the reversal effects of antagonists in vitro. METHODS: The effects of evoked twitch tension response have been studied on the isolated phrenic nerve hemidiaphragm preparation of the rat, using a single twitch (0.1 Hz) and the train of four (TOF; 2 Hz for 2 s) stimulation. The cumulative concentration effect and TOF ratio at each point of twitch depression after alpha-bungarotoxin, decamethonium or lidocaine were measured mechanomyographically. The EC50 and EC95 of alpha-bungarotoxin, decamethonium or lidocaine were calculated using an inhibitory sigmoid Emax model. The reversal effects of various doses of neostigmine, pyridostigmine or 4-aminopyridine (4-AP) to the partial neuromuscula r block produced by EC50 of alpha-bungarotoxin, decamethonium or lidocaine were determined. RESULTS: The EC50 and EC95 of alpha-bungarotoxin, decamethonium or lidocaine were 0.179 and 0.320 microgram/ml, 17.07 and 26.84 microgram/ml or 76.80 and 105.70 microgram/ml. TOF fade was produced by alpha-bungarotoxin or decamethonium but not by lidocaine. Neostigmine or pyridostigmine did not reverse the partial neuromuscular block induced by alpha-bungarotoxin, decamethonium or lidocaine. However, 4-AP produced a dose-dependent recovery of the twitch response (P < 0.05). CONCLUSIONS: alpha-Bungarotoxin, decamethonium or lidocaine produced different degree of TOF fade, and it means that this may be due to different site of action of these drugs. 4-AP reversed effectively the partial neuromuscular block induced by alpha-bungarotoxin, decamethonium or lidocaine, whereas neostigmine or pyridostigmine did not.