Carbamazepine was tested on high-threshold TTX-resistant Na+ currents (TTX-R-currents), evoked from acutely isolated rat dorsal root ganglion (DRG) cells. Under control conditions, the TTX-R-currents recorded from different DRG cells varied greatly regarding use-dependent inactivation (TTX-R-current UDI), measured as the percent decrease in current amplitude induced by changing the current activation rate from 0.1 Hz to 1.0 Hz. Also, when TTX-R-currents were evoked at 0.1 Hz from a holding potential (hp) of -60 mV, a larger fraction of TTX-R-channels resided tonically in a slow inactivation state in DRG cells with more TTX-R-current UDI versus those with less TTX-R-current UDI. The block of TTX-R-currents evoked from hp -60 mV by 100-microM carbamazepine and the EC50 for carbamazepine block was positively correlated with TTX-R-current UDI. The slope factors estimated for the concentration-response curves averaged 0.68, suggesting the presence of low and high affinity sites. Fitting the data with a two-site binding isotherm gave estimates of 30 microM and 760 microM for the EC50s of the high and low affinity sites, respectively. The fraction of the total fit attributed to the high affinity site was positively correlated with TTX-R-current UDI. Carbamazepine increased the fast and slow time constants for recovery from inactivation and the fraction of the fit attributed to the slow time constant. These data suggest that carbamazepine interacts with a slow inactivation state of TTX-R-channels. This particular mechanism might be exploited in future research aimed at developing pain medications that selectively block Na(V)1.8 channels or Na+ channels in general.