Intraoperative neuromonitoring alerts in a pediatric deformity center

OBJECTIVE Intraoperative neuromonitoring (IONM) involves the use of somatosensory evoked potentials (SSEPs) and transcranial electric motor evoked potentials (TceMEPs). In this retrospective study the authors examined the sensitivity and specificity of both SSEPs and TceMEPs during pediatric spinal deformity surgeries. METHODS The authors performed a retrospective quantitative analysis of data obtained in 806 patients (197 males and 609 females) treated from December 2011 until October 2015. All patients were diagnosed with scoliosis that was classified as one of the following: adolescent idiopathic scoliosis (AIS) (38%), congenital scoliosis (22%), or syndromic scoliosis (40%). Also, 53 patients underwent vertebral column resection (VCR). All surgeries were monitored by high-level neuromonitoring specialists and were performed with total intravenous anesthesia. Alerts were described as a decrease in amplitude by 50% or greater (bilateral or unilateral) in SSEPs, TceMEPs, or both. RESULTS True-positive alerts for TceMEPs were observed in 60 of the 806 patients (7.4%). True-positive alerts for SSEPs were observed in 7 of the 806 patients (0.9%). In contrast, there were no false-positive or false-negative outcomes. Only 1 case (0.1%) was reported with a permanent postoperative deficit. No reported false negatives or false positives were observed, and thus sensitivity was 100% and specificity was 93%–100% for TceMEPs. The rate of sensitivity was 13.2% and the rate of specificity was 100% for SSEPs. The breakdown of total alert was as follows: 6.6% in AIS cases, 24.5% in congenital scoliosis cases, and 10.2% in syndromic scoliosis cases. Neurological injury rates were significantly lower than in previous studies, as there were 0% for AIS cases (p = 0.12), 0.6% for congenital scoliosis cases (p = 0.17), and 0% for syndromic scoliosis cases (p = 0.07). One injury in a patient with congenital scoliosis occurred during a VCR procedure, which brought the injury rate to 1.9% (p < 0.005). IONM alerts occurred during 34% of rod/correction cases, 25% of thoracic screw placements, 20% of the osteotomies, 17% of the resections, 3% of the cage insertions, and 2% of the sublaminar wiring procedures. CONCLUSIONS The authors hypothesize that the results of this study will support the necessity, as a standard of care, of multimodality neuromonitoring during high-risk pediatric spinal deformity surgery because of the decrease in postoperative deficits. Their data suggest that the TceMEPs are more sensitive than SSEPs, but when used in combination, they offer the patient a level of safety that would otherwise not exist. Last, these findings support the notion that better outcomes are achieved with high-level IONM professionals.