Abstract
The relation between electroencephalographic pattern changes and cerebral ischemia during carotid endarterectomy under general anesthesia is well established. Pattern changes seen on somatosensory evoked potentials under the same conditions are reported to be more sensitive indicators of cerebral ischemia. We estimated the sensitivity and specificity of somatosensory evoked potentials relative to electroencephalography for detecting cerebral ischemia during carotid endarterectomy under general anesthesia. We simultaneously monitored electroencephalographs and somatosensory evoked potentials in 53 carotid endarterectomies performed on 51 patients under general anesthesia, and we determined the extent to which somatosensory evoked potentials detected cerebral ischemia defined by electroencephalographic pattern changes at the time of carotid cross-clamp. Twenty-three of the 53 cases studied had electroencephalographic evidence of ischemia following carotid cross-clamp. Ten of these 23 cases had an increased somatosensory evoked potential latency of 0.1 msec or greater (sensitivity 0.43). One of these 23 patients had a decrease in somatosensory evoked potential amplitude of 50% or greater (sensitivity 0.04). Of the 30 subjects who had no electroencephalographic evidence of ischemia, 13 had either no change or a decrease in somatosensory evoked potential latency (specificity 0.45). None of these 30 cases had a significant decrease in somatosensory evoked potential amplitude (specificity 1.0). If somatosensory evoked potential latencies were a sensitive method for detecting cerebral ischemia (true sensitivity of 0.95 or higher), the probability of only 10 subjects having somatosensory evoked potential latency increases would be less than 0.001. Therefore, our observed sensitivity cannot be attributed to chance. We conclude that measuring somatosensory evoked potentials is not a sensitive method for detecting cerebral ischemia during carotid endarterectomy.