Slowed motor conduction in lumbosacral nerve roots in cauda equina lesions: a new diagnostic technique.

New techniques have been developed for the electrophysiological assessment of patients with suspected cauda equina lesions using transcutaneous spinal stimulation (500-1500 V: time constant 50 microseconds) to measure motor latencies to the external and sphincter and puborectalis muscles from L1 and L4 vertebral levels. These latencies represent motor conduction in the S3 and S4 motor roots of the cauda equina between these levels. Similarly motor latencies can be recorded from spinal stimulation to the anterior tibial muscles (L4 and L5 motor roots). Transrectal stimulation of the pudendal nerves is used to measure the pudendal nerve terminal motor latency. In 32 control subjects, matched for age and sex, mean motor latencies from L1 and L4 spinal stimulation were 5.5 +/- 0.4 ms and 4.4 +/- 0.4 ms (mean + SD). In the 10 patients with cauda equina disease including ependymoma, spinal stenosis, arachnoiditis and trauma, these latencies were 7.2 +/- 0.8 ms and 4.6 +/- 0.9 ms, a significant increase in the L1 latency. The L1/L4 latency ratios to the puborectalis muscle were 1.36 +/- 0.09 in control subjects and 1.72 +/- 0.13 in cauda equina patients. Pudendal nerve terminal motor latencies were normal in eight of the 10 patients with cauda equina disease. The single fibre EMG fibre density in the external and sphincter muscle (normal, 1.5 +/- 0.16) was increased in patients with cauda equina lesions (1.73 +/- 0.28), but was increased more than two standard deviations from the mean only in three patients. This increase in fibre density was not of diagnostic value since it was also found in two of the four patients with low back pain. Slowing of motor conduction in the cauda equina is thus a useful indication of damage to these intraspinal motor roots. These investigations can be used in the selection of patients for myelography, and to follow progress in patients managed conservatively.