268. Neuroanatomical categorization of motor evoked potential alerts during cervical spine surgery and associated risk for new neurologic deficit: retrospective analysis of 40,919 procedures

Abstract

BACKGROUND CONTEXT Accurate and clinically impactful intraoperative neuromonitoring (IONM) during cervical spine surgery is predicated on multimodality monitoring with motor evoked potentials. A motor evoked potential (MEP) alert is indicative of potential evolving motor dysfunction in a nervous system structure, and in theory the affected muscles should help elucidate the neuroanatomical location of the dysfunction. PURPOSE To associate a neuroanatomical location with an MEP alert based on the pattern of affected muscles. STUDY DESIGN/SETTING A total of 40,919 cervical procedures monitored with MEPs, somatosensory evoked potentials (SSEPs) and spontaneous electromyography (sEMG) from January 2016 to January 2018 were reviewed retrospectively from a multi-institutional IONM database. PATIENT SAMPLE Age 18 and older. OUTCOME MEASURES Clinical outcomes were based on neurologic assessment in the immediate postoperative period. Odds ratios (OR) for new neurologic deficits (NNDs) were calculated according to Altman. METHODS Procedures with MEP alerts were categorized based on patterns of attenuated muscle responses reported to the surgeon during the procedure. MEP alerts involving lower extremity muscles were categorized as CORD alerts and these were subcategorized as CORD-Level 1 (both lower and upper extremity muscles involved) or CORD-Level 2 (just lower extremity muscles involved). Procedures with just upper extremity MEP alerts were categorized as having either a nerve root alert or an arm alert. Nerve root alerts were subcategorized as isolated superior (C5, C6), middle (C7) or inferior (C8, T1) alerts based on the innervation patterns of the affected muscles. Procedures with arm alerts were those with both superior and inferior nerve root alerts. All anatomical categories were further subcategorized based on whether the MEP alerts fully resolved by closure or one or more attenuated muscle responses remained unresolved. RESULTS The rate of NND was 0.63% for all procedures (n=256/40919). For procedures with an MEP alert (n=2505), the rate was 7.5% and the associated risk of a NND, relative to procedures with no MEP alerts, was significantly increased (OR=44.8, CI:[33.9,59.3]). For CORD alerts, the OR was higher for a Level 1 alert (OR=73, CI:[52.9,100.7]) than a Level 2 alert (OR=18.8, CI:[10.1,35.0]). For nerve root alerts, the OR was highest for superior alerts (OR=60.4, CI:[40.9-89.2]), followed by inferior alerts (OR=25.3, CI:[16.3,39.0]) and lastly middle alerts (OR=13.2, CI:[3.2,54.9]). The rate and risk of NND was highest if an alert remained unresolved and in particular for unresolved CORD-Level 1 alerts (15.9%, OR=102.4, CI:[73.8,142.0]) and unresolved superior alerts (15.5%, OR=102.2, CI:[67.9,153.7]). Relative to those procedures with unresolved alerts, the rate and risk of NND was significantly decreased if a CORD-Level 1 alert was resolved (0.94%, OR=0.05, CI:[0.01-0.21]) or if a superior nerve root alert was resolved (1.6%, OR=0.09, CI:[0.03-0.29]). All OR data were associated with p values CONCLUSIONS Categorical MEP alert information helps to localize evolving neural dysfunction during cervical spine surgery and should be communicated to the surgeon using neuroanatomical reference points to facilitate effective interventions. This is critical because the risk of postoperative neurologic dysfunction can be significantly decreased if the alert is resolved by closure. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs.