It is known that other than well known distal peripheral nervous system complication, diabetes can affect, though rarely, the central nervous system. There have been a number of studies concerned with somatosensory evoked potentials (SEPs) in patients after stroke. Most of these studies have, however, paid attention to the relationship between anatomical location of the lesion and SEPs. Only a few have discussed the relationship among conduction delays in distinct parts of somatosensory pathways, especially in the central nervous system for diabetic patients with stroke.

The purpose of this study was to compare the latencies and amplitudes of SEPs between 2 groups of stroke patients with and without diabetes mellitus and to evaluate the mechanism of showing abnormal SEPs in stroke patients with diabetes mellitus.

Fifty stroke patients who had no history or findings suggestive of diabetes mellitus, thirty diabetic stroke patients, twenty diabetics, and fourteen age-matched controls were studied by SEPs with stimulation of the median nerve at the wrist and recorded from Erb's point, second cervical vertebra, and scalp referenced to FPz. In stroke patients who were diabetic, N9, N13, and N20 latencies were very likely to be increased and N9-N20 conduction time was also prolonged. In stroke patients with the lesion involving somatosensory pathway whether or not diabetic, scalp recorded N20 potentials were of low amplitude even though the amplitude is more reduced in diabetic patients.

These findings suggests that prolonged peripheral and central conduction time in diabetic stroke patients is most likely due to diabetes mellitus rather than the stroke itself, even though the scalp recorded potential of low amplitude can be shown also.