Time- and voltage-dependent local anesthetic effects on sodium (Na) currents are generally interpreted using modulated receptor models that require formation of drug-associated nonconducting states with high affinity for the inactivated channel. The availability of inactivation-deficient Na channels has enabled us to test this traditional view of the drug-channel interaction. Rat skeletal muscle Na channels were mutated in the III-IV linker to disable fast inactivation (F1304Q: FQ). Lidocaine accelerated the decay of whole-cell FQ currents in Xenopus oocytes, reestablishing the wild-type phenotype; peak inward current at -20 mV was blocked with an IC50 of 513 microM, while plateau current was blocked with an IC50 of only 74 microM (P < 0.005 vs. peak). In single-channel experiments, mean open time was unaltered and unitary current was only reduced at higher drug concentrations, suggesting that open-channel block does not explain the effect of lidocaine on FQ plateau current. We considered a simple model in which lidocaine reduced the free energy for inactivation, causing altered coupling between activation and inactivation. This model readily simulated macroscopic Na current kinetics over a range of lidocaine concentrations. Traditional modulated receptor models which did not modify coupling between gating processes could not reproduce the effects of lidocaine with rate constants constrained by single-channel data. Our results support a reinterpretation of local anesthetic action whereby lidocaine functions as an allosteric effector to enhance Na channel inactivation.