In 15 pigs lidocaine and bupivacaine were injected into the left anterior descending (LAD) coronary artery to investigate the cardiotoxic effects of these drugs. Anesthesia was maintained by a continuous intravenous pentobarbital infusion and ventilation was controlled. Aortic, pulmonary arterial, right atrial, and left ventricular pressures, a standard 12 lead ECG, cardiac output, and great cardiac venous blood flow were recorded. The local anesthetics were administered at body temperature over approximately 10 sec in a random, crossover fashion at the following equipotent anesthetic doses: bupivacaine, 0.25, 0.5, 1, 2, and 4 mg; lidocaine, 1, 2, 4, 8, and 16 mg. The hemodynamic effects were short-lived, peaking about 5 sec after drug infusion. At the highest dose, both drugs decreased left ventricular dP/dT by 28% (P < 0.001) and aortic blood pressure by 12% (lidocaine) and 8% (bupivacaine) (P < 0.001 and P < 0.07). Heart rate, cardiac output, and coronary venous blood flow did not change. Thus, the cardiodepressant ratio between the two drugs was comparable with their local anesthetic potency ratio (bupivacaine I lidocaine, 4:1). Seven animals died in ventricular fibrillation within 1 min after 4 mg bupivacaine dose. All animals given 16 mg lidocaine survived. Ventricular fibrillation was preceded by progressive widening of the QRS complexes recorded over the area perfused by the LAD. The ECG changes after 16 mg lidocaine were of the same magnitude as those recorded after 1 mg bupivacaine. In five of the surviving animals 32 and 64 mg lidocaine were injected intracoronarily after termination of the crossover study. After 64 mg, three animals died in sudden ventricular fibrillation, preceded by similar ECG changes as seen after 4 mg bupivacaine. Thus, the electrophysiologic toxicity ratio between bupivacaine and lidocaine was on the order of 16:1. It is concluded that this animal model is reproducible and allows discrimination between cardiodepressant and electrophysiologic toxicity of local anesthetic agents.