Ventricular fibrillation (VF) is a cause of death in bupivacaine-induced cardiovascular toxicity. We have examined the therapeutic effects of lidocaine on the threshold for bupivacaine-induced VF in in situ beating swine hearts. Twenty-four animals were allocated to one of three groups: 0.25% bupivacaine, 1% lidocaine or 0.25% bupivacaine with 1% lidocaine were infused into the left anterior descending coronary artery in increasing doses of 0, 1, 2, 4, 8 and 16 ml h-1 for 15 min, respectively. ECG and haemodynamic variables were monitored continuously during infusion. Regional myocardial function in the area supplied by the left anterior descending coronary artery was assessed using the sonomicrometry technique. VF did not occur in the lidocaine group. VF developed at higher infusion rates in animals given bupivacaine with lidocaine (in one animal at an infusion rate of 8 ml h-1 and in seven at 16 ml h-1) compared with animals given bupivacaine alone (in one at an infusion rate of 4 and in seven at 8 ml h-1). Although regional myocardial function decreased with increases in the infusion rate in each group, the depressant effects of the bupivacaine solution (medial inhibitory infusion rate of systolic shortening: IR50 = 2.43 (0.43) ml h-1) were significantly greater than those of the lidocaine solution (IR50 = 5.83 (0.87) ml h-1), but did not differ from those of the bupivacaine with lidocaine solution (IR50 = 3.54 (0.56) ml h-1). This study indicates that a combination of lidocaine and bupivacaine increased the threshold for bupivacaine-induced VF without further depressing myocardial contractility.