Warning Thresholds on the Basis of Origin of Amplitude Changes in Transcranial Electrical Motor-Evoked Potential Monitoring for Cervical Compression Myelopathy

Abstract

A retrospective analysis of prospectively collected data from consecutive patients undergoing transcranial electrical motor-evoked potential (TCE-MEP: compound muscle action potentials) monitoring during cervical spine surgery. OBJECTIVE.: To divide the warning threshold of TCE-MEP amplitude changes on the basis of origin into the spinal tract and spinal segments and decide warning thresholds for each. The parameter commonly used for the warning threshold in TCE-MEP monitoring is wave amplitude, but amplitude changes have not been examined by anatomical origin. Intraoperative TCE-MEP amplitude changes were reviewed for 357 patients with cervical myelopathy. Most of the patients were monitored by transcranial electrical stimulated spinal-evoked potential combined with TCE-MEP. The warning threshold of TCE-MEP was taken as waveform disappearance. For each patient, amplitude changes were separated, according to origin, into the spinal tract and spinal segments and compared with clinical outcome. Assessable TCE-MEP waves were obtained in 350 cases. Disappearance of TCE-MEP waves, which were innervated by the spinal levels exposed to the surgical invasion, was seen in 11 cases. Disappearance of TCE-MEPs, which were innervated by the spinal levels inferior to them, was seen in 43 cases. There was no postoperative motor deficit in those cases. However, such deficits caused by spinal segment injury were seen in 2 cases, which showed that intraoperative amplitude decreased to 4.5% and 27%. If we had established the warning threshold as 30% of the control amplitude, we would likely have prevented both cases of postoperative motor deficits, but 106 (30.3%) cases would have become positive cases. If we had established the warning threshold separately as wave disappearance for the spinal tract and 30% of the control amplitude for the spinal segments, sensitivity and specificity would have been 100% and 83.7%, respectively. Dividing the warning threshold on the basis of origin of amplitude changes could reduce false-positive cases and prevent intraoperative injuries.