Neurophysiological Detection of Impending Spinal Cord Injury During Scoliosis Surgery

Background: Despite the many reports attesting to the efficacy of intraoperative somatosensory evoked potential monitoring in reducing the prevalence of iatrogenic spinal cord injury during corrective scoliosis surgery, these afferent neurophysiological signals can provide only indirect evidence of injury to the motor tracts since they monitor posterior column function. Early reports on the use of transcranial electric motor evoked potentials to monitor the corticospinal motor tracts directly suggested that the method holds great promise for improving detection of emerging spinal cord injury. We sought to compare the efficacy of these two methods of monitoring to detect impending iatrogenic neural injury during scoliosis surgery. Methods: We reviewed the intraoperative neurophysiological monitoring records of 1121 consecutive patients (834 female and 287 male) with adolescent idiopathic scoliosis (mean age, 13.9 years) treated between 2000 and 2004 at four pediatric spine centers. The same group of experienced surgical neurophysiologists monitored spinal cord function in all patients with use of a standardized multimodality technique with the patient under total intravenous anesthesia. A relevant neurophysiological change (an alert) was defined as a reduction in amplitude (unilateral or bilateral) of at least 50% for somatosensory evoked potentials and at least 65% for transcranial electric motor evoked potentials compared with baseline. Results: Thirty-eight (3.4%) of the 1121 patients had recordings that met the criteria for a relevant signal change (i.e., an alert). Of those thirty-eight patients, seventeen showed suppression of the amplitude of transcranial electric motor evoked potentials in excess of 65% without any evidence of changes in somatosensory evoked potentials. In nine of the thirty-eight patients, the signal change was related to hypotension and was corrected with augmentation of the blood pressure. The remaining twenty-nine patients had an alert that was related directly to a surgical maneuver. Three alerts occurred following segmental vessel clamping, and the remaining twenty-six were related to posterior instrumentation and correction. Nine (35%) of these twenty-six patients with an instrumentation-related alert, or 0.8% of the cohort, awoke with a transient motor and/or sensory deficit. Seven of these nine patients presented solely with a motor deficit, which was detected by intraoperative monitoring of transcranial electric motor evoked potentials in all cases, and two patients had only sensory symptoms. Somatosensory evoked potential monitoring failed to identify a motor deficit in four of the seven patients with a confirmed motor deficit. Furthermore, when changes in somatosensory evoked potentials occurred, they lagged behind the changes in transcranial electric motor evoked potentials by an average of approximately five minutes. With an appropriate response to the alert, the motor or sensory deficit resolved in all nine patients within one to ninety days. Conclusions: This study underscores the advantage of monitoring the spinal cord motor tracts directly by recording transcranial electric motor evoked potentials in addition to somatosensory evoked potentials. Transcranial electric motor evoked potentials are exquisitely sensitive to altered spinal cord blood flow due to either hypotension or a vascular insult. Moreover, changes in transcranial electric motor evoked potentials are detected earlier than are changes in somatosensory evoked potentials, thereby facilitating more rapid identification of impending spinal cord injury. Level of Evidence: Diagnostic Level I. See Instructions to Authors for a complete description of levels of evidence.