Monitoring evoked potentials during spinal surgery in one institution.

PURPOSE To review the experience of one tertiary care institution with somatosensory evoked potential (SSEP) monitoring during spinal surgery in order to assess the ability to monitor and predict neurological outcome effectively. METHODS Records of all patients undergoing spinal surgery during 18 mo were retrospectively reviewed. Information from the patient chart included preoperative neurological status, surgical procedure, anaesthetic management, and postoperative neurological outcome. Information regarding the techniques used and interpretation of all SSEP tracings were obtained from evoked potential data sheets completed for each patient. The incidences of clinically important SSEP changes and new postoperative neurological deficits were analysed. RESULTS Somatosensory evoked potential monitoring of the lower and upper extremities with non invasive techniques was used in 309 patients undergoing surgery on the cervical (88), thoracic (52), and lumbar spine (169). Thirty seven patients (11%) did not have suitable tracings for interpretation and 17 (5.5%) had baseline tracings described as poor. An intraoperative SSEP change occurred in 16 patients (6%) with SSEP and seven (2.6%) had a new neurological deficit postoperatively. Three persistent deficits were predicted by permanent SSEP change, and one transient deficit by a transient SSEP change. False positive results occurred in 12 patients (4.4%) and false negative results occurred in three (1.1%), with a sensitivity of 57% and a specificity of 95%. The incidence of SSEP changes was greater in the thoracic (18%) than in the cervical (1.2%) or lumbar (5.4%) groups (P < 0.05). CONCLUSION Effective SSEP monitoring was possible despite the many factors which may have interfered with monitoring. More improvements in the techniques and conditions of monitoring are needed to decrease the incidence of false positive and negative results.