Patients with multifocal motor neuropathy may complain of muscle fatigue, even though the degree of conduction block assessed at rest has improved with treatment. To explore the mechanism involved, we examined changes in muscle force during maximum voluntary contraction (MVC) and monitored conduction block before and after MVC in five patients with multifocal motor neuropathy. The results were compared with those for the contralateral unaffected homonymous muscles. For one patient, who had bilateral involvement, a normal subject of a similar age and stature served as the control. Results of conduction studies were also compared with those from six patients with amyotrophic lateral sclerosis (ALS) with similar compound muscle action potential (CMAP) amplitudes after proximal stimulation. During MVC for 60 s, the affected muscles developed prominent fatigue; the force at the end of contraction compared with the initial force was significantly lower for the affected muscles [42 +/- 19% (mean +/- standard deviation) of the initial force] than for the control muscles (94 +/- 9%; P = 0.01). After MVC, the amplitude ratio of CMAPs after proximal versus distal nerve stimulation transiently decreased to 19 +/- 14% of that before MVC in the affected muscles, but not in the control muscles (94 +/- 3.8% of that before MVC) and in patients with ALS (95 +/- 6.7%). In one patient with a focal lesion in the forearm, nerve excitability was monitored at the lesion site before and after MVC for 120 s. There were significant increases in axonal threshold (approximately 48%) and supernormality (approximately 135%) immediately after MVC, suggesting that the axonal membrane had undergone hyperpolarization and, by extrapolation, that this had precipitated the conduction block. This study is the first to show that activity-dependent conduction block plays a role in human disease by causing muscle fatigue.