Role of SEP in identifying patients requiring temporary shunt during carotid endarterectomy

EEGs and short-latency somatosensory evoked potentials (SEPs) to median nerve stimulation were recorded during 151 carotid endarterectomies, performe under general anaesthesia. Carotid occlusion did not affect either EEG or SEP in 120 cases (group A). In 31 cases the EEG showed "ischaemic" abnormalities (group B). A temporary shunt was inserted only in 16 B patients showing also severely depressed cortical SEPs within 2 min after carotid occlusion (group B shunt). In 15 B patients in whom SEPs were less affected, the operation was completed without shunt (group B no shunt). One intraoperative stroke occurred in group A and two in group B shunt. No neurological complications occurred in group B no shunt. Overall stroke rate was 2%. On retrospective analysis, latency and amplitude of N20 and P25 waves proved to be uninfluenced by carotid occlusion in group A, but were significantly affected in groep B shunt. P25 amplitude alone was reduced in B no shunt. An arbitrary index (need-for-shunt index, NSI) was made in order to rate changes of P25 latency and amplitude. Its mean values were significantly different in the 3 groups. A threshold value is suggested above which shunt is required, as a useful adjuntt to EEG, in order to balance prevention of brain ischaemia against the risks of shunt.