319 Predictive Value of Intraoperative Neurophysiological Monitoring During Spine Surgery

Abstract

INTRODUCTION: A single-center study to define the sensitivity and specificity of intraoperative neurophysiological monitoring in detecting new neurological deficits. METHODS: A cohort of 4989 consecutive (2962 male and 2027 female) patients operated on over 3 years was evaluated. Subgroup analysis was performed for patients undergoing posterior cervical intervention (n = 1373) and posterior lumbar fusion surgery (n = 2420). Sensitivity and specificity were determined using Bayesian techniques. Impact of length of surgery and of variables, including age, sex, body mass index, diabetes mellitus, hypertension, coronary artery disease, cerebrovascular disease, and history of smoking, on the development of a new neurodeficit was defined. RESULTS: Of 4489 patients, 426 patients (9.4%) had significant intraoperative somatosensory evoked potential (SSEP) changes. New postoperative neurological deficits occurred in 121 (2.7%) patients. Twenty-five of 426 patients (5.8%) with neuromonitoring changes developed a new neurodeficit. Ninety-six of 4063 patients (2.3%) with no changes developed a new neurodeficit. In posterior cervical interventions, 98 of 1373 (7.1%) developed SSEP changes and 13 of 98 (13.2%) patients developed a new neurodeficit, whereas 49 (3.8%) developed a new neurodeficit without any changes in intraoperative monitoring (IOM). In posterior lumbar fusion SSEPs, electromyographs (EMGs), and pedicle screw stimulation was utilized. Two hundred forty-nine of 2420 (10.2%) patients developed IOM changes and 8 of 249 (3.2%) patients developed a new neurodeficit, while 37 (1.7%) developed a new neurodeficit without changes in IOM. In the entire cohort SSEPs had sensitivity 20.7% and specificity of 90.8%. In the posterior cervical cohort, SSEPs had s sensitivity of 20.9% and a specificity of 93.7%. In the lumbar fusion cohort, multimodality monitoring including SSEPs, EMGs, and pedicle screw stimulation had a sensitivity of 17.8% and a specificity of 88.6%. Neither the length of surgery nor did any analyzed patient-related variable have a statistically significant impact on the development of a new neurodeficit. CONCLUSION: The strength of SSEPs is their negative predictive value. A cost-utility analysis for IOM in spine surgery should be undertaken.