Patch-clamp recording techniques have permitted measurement of the fast Na+current (INa) in isolated cardiac cells from a number of species in recent years. However, there is still only very little information concerning human cardiac INa. The purpose of this study was to describe the kinetics of INain normal-appearing, Ca2+-tolerant, enzymatically isolated human atrial myocytes using whole-cell voltageclamp techniques. Atrial specimens were obtained from 46 patients undergoing open heart surgery. Cs+was substituted for K+in both pipette and external solutions and F- was added to the former. The reversal potential of the rapid inward current varied approximately 57 mV at 17±1°C with a 10-fold change in [Na+]., and the current was completely blocked by 100 μM tetrodotoxin, findings typical of the fast cardiac Na+current. The tetrodotoxin dose-response curve was best fitted by an equation describing binding to high- and low-affinity sites. INawas activated at a voltage threshold of −70 to −60 mV, and peak inward current was obtained at ≈ −30 mV (holding potential, −140 mV). The inactivation time course was voltage dependent and was fitted best by the sum of two exponentials. The relation between voltage and steady-state availability (h∞) was sigmoidal with the half-inactivation at −95.8±0.9 mV and a slope factor of 5.3±0.1 mV (n=46), and we did not observe a significant difference with disease and age. The overlap of the h∞ and activation curves suggested the presence of a Na+“window” current. Recovery from inactivation also was voltage dependent and best fitted by a model describing the sum of two exponentials. Recovery occurred after an initial delay at potentials positive to −140 mV, suggesting that inactivation of human atrial INais a multistate process. We conclude that INaof normal-appearing, Ca2+-tolerant human atrial myocytes is similar to that of other mammalian cardiac cells with the possible exception of having two tetrodotoxin binding sites.