Although lidocaine block of cardiac Na+ current (INa) has been extensively studied in animal tissues, very little is known about its actions on human cardiac INa. We studied the effects of lidocaine (0.01-10 mM) on human atrial INa in single myocytes using whole-cell patch-clamp techniques. The dose-response relationship for lidocaine block at a low frequency (0.2 Hz, "tonic" block) indicated that lidocaine blocked Na+ channels by one-to-one binding with an apparent Kd of 291 microM. Lidocaine (200 microM) shifted the steady-state INa availability curve by -11 mV, but did not change the slope factor (n = 5). Lidocaine also induced use-dependent block that increased directly with increases in drug concentration (0.01-1 mM) and pulse duration (3-100 msec) and inversely with interpulse interval (2-0.33 sec). The time constant for onset of lidocaine (200 microM) block of INa displayed both a fast (tau f = 3.6 +/- 0.4 msec) and a slow (tau s = 168 +/- 21 msec) exponential component (n = 10). In addition, lidocaine slowed the rate of INa recovery after a 1-sec conditioning pulse to -20 mV, recovery was biexponential at a low drug concentration (20 microM), but had only a single slow phase at a high drug concentration (200 microM). These characteristics of lidocaine block suggest that lidocaine binds to both inactivated and activated Na+ channels in human atrial cells and that use-dependent block of INa by lidocaine is dependent on drug concentration, interpulse interval and pulse duration, findings similar to those reported for other mammalian species. The similarity of these results to those obtained from atrial as well as ventricular cells from other species suggests that some source other than differential drug action on atrial and ventricular INa underlies differential drug efficacy against supraventricular and ventricular dysrhythmias.