The mechanism of interaction of lidocaine with cardiac sodium channels during use-dependent block is not well denned. We examined the blockade of single cardiac sodium channels by lidocaine and its hydrophobic derivative RAD-242 in rabbit ventricular myocytes. Experiments were performed in cell-attached and inside-out patches. Use-dependent block was assessed with trains of ten 200 -msec pulses with interpulse intervals of 500 msec and test potentials of -60 to -40 mV. Single-channel kinetics sometimes showed time-dependent change in the absence of drug. During exposure to 80 μM lidocaine, use-dependent block during the trains was associated with a decrease in the average number of openings per step. At -60 mV, mean open time was not significantly changed (control, 1.4 ± 0.6 msec; lidocaine, 1.2 ± 03 msec, P > 0.05). Greater block developed during trains of 200 -msec pulses compared with trains of 20-msec pulses at the same interpulse interval at test potentials during which openings were uncommon later than 20 msec (−50 and -40 mV). Prolonged bursts of channels showing slow-gating kinetics were observed both in control and the presence of 80 μM lidocaine. However, lidocaine may decrease the late sodium current by altering the kinetics of slow gating. The hydrophobic lidocaine derivative RAD-242, which has a 10-fold greater lipid solubility than lidocaine, decreased the peak averaged current during pulse train stimulation by 60% without a change in the mean open time. Our results suggest that the major effect of lidocaine during use-dependent block involves the interaction with a nonconducting state of the sodium channel followed by a failure to open during subsequent depolarization. {Circulation Research1989;65:1247-1262)