We sought, in this observer-blinded study, to determine the lethal dose for each of the local anesthetics levobupivacaine (L), racemic bupivacaine (B), and ropivacaine (R), and to compare their respective effects on the QRS interval of the precordial electrocardiograph after intracoronary; injection. Anesthetized swine were instrumented with; a left anterior descending artery coronary angiography catheter and injected with increasing doses of L, B, or R according to a randomized protocol. The doses administered were 0.375, 0.75, 1.5, 3.0, and 4.0 mg, with further doses increasing in 1-mg increments until death occurred. Plotting the mean maximum QRS interval as a function of the log(10) mmol dose allowed the following cardiotoxicity potency ratios to be determined for a doubling of QRS duration-B:L: R = 2.1:1.4:1. The lethal doses in millimoles (median/range) for L and R were (0.028/0.024-0.031) and (0.032/0.013-0.032), respectively, and were significantly higher than for B (0.015/0.012-0.019) - (P < 0.05, n = 7 for all groups). The lethal dose did not differ between R and L. Thus, the cardiotoxicity potency ratios for the three anesthetics based on lethal dose were: 2.1:1.2:1. If the anesthetic potencies for B and L, are similar, the latter should have less potential for cardiotoxicity in the clinical situation.