To better understand the interaction between cocaine and lidocaine, we studied the cocaine's concentration-effect relationship for action potential duration (APD) and the rate of rise of phase 0 of the action potential (Vmax) of canine papillary muscle in the presence and absence of lidocaine. We measured APD and Vmax during programmed stimulation and superfusion with normal Tyrode's, 30 microM cocaine and 30 microM cocaine + 30 microM lidocaine. Using two microelectrodes, we simultaneously recorded action potentials from two sites during programmed stimulation and measured the conduction velocity and effective refractory period during exposure to normal Tyrode's, cocaine and cocaine + lidocaine. Cocaine with or without lidocaine delayed the plateau of the APD restitution curve. At 1000 msec cycle length, the addition of 30 microM lidocaine to the superfusate containing 30 microM cocaine shortened the time constant for reactivation of Vmax from 514 +/- 63 to 234 +/- 28 msec (P < .01). Lidocaine also improved the conduction velocity decreased by cocaine, but did not significantly change the effective refractory period. The configuration of cocaine concentration-effect curve for APD was biphasic. For cocaine concentrations < 100 microM, APD progressively shortened prolonged with increasing concentrations. As cocaine concentrations increased > 100 microM, APD progressively shortened. The addition of lidocaine to the superfusate with cocaine > 100 microM tended to attenuate the progressive APD shortening due to cocaine. Lidocaine shifted the curve correlating cocaine concentration and reduction of Vmax rightward, but preserved Emax at cocaine concentration > 225 microM. These findings suggest competitive antagonism between cocaine and lidocaine at a single sodium channel receptor. Conclusion: lidocaine displaces cocaine from the sodium channel receptor through competitive binding. Lidocaine may prove to be beneficial in reversing cocaine-induced slowing of ventricular conduction.