General Session: Pediatric Spine - Hall F
Presented by: A.K. Sestokas
A.K. Sestokas(1), E.A. Tesdahl(1), J.M. Flynn(2), S.A. Shah(3), T.S. Oswald(4), M. Beckish(5), W.B. Wilent(1), C.R. Wiggins(1), A.J. Franco(1), J. Cohen(1)
(1) SpecialtyCare, Nashville, TN, United States
(2) The Children's Hospital of Philadelphia, Philadelphia, PA, United States
(3) Nemours / A.I. duPont Hospital for Children, Wilmington, DE, United States
(4) Pediatric Orthopaedic Associates, Atlanta, GA, United States
(5) Greenville Health System, Greenville, SC, United States
Introduction: Intraoperative neurophysiological monitoring (IONM) detects negative changes in neurophysiologic function during surgery, affording the surgical team opportunity to possibly reverse evolving neurologic injury. The effectiveness of IONM in reducing postoperative neurologic sequelae depends both on accurate identification of significant neurophysiologic change and timely intervention by the surgical team to resolve neuromonitoring alerts. This study sought to quantify the impact of alert resolution on neurologic outcome in pediatric spine surgery by comparing outcomes when alerts were fully resolved, partially resolved or unresolved prior to the end of surgery.
Methods: This study is based on a retrospective review of a multi-institutional database of 6,564 pediatric spine surgeries performed with IONM between 2013 and 2016. For each procedure, we extracted information about presence of IONM alerts, alert resolution and clinical outcome based on neurologic examination in the immediate postoperative period. Patients were assigned to one of four cohorts: no neuromonitoring alerts during surgery, fully resolved alerts, partially resolved alerts, unresolved alerts. Differences in rate of new postoperative neurologic deficit across groups were evaluated by chi-squared tests & logistic regression with the post-hoc Tukey HSD test for multiple comparisons of rates between groups.
Results: Intraoperative neuromonitoring alerts occurred in 512 of 6564 (7.8%) cases, resolving fully in 357 (69.7%), partially in 84 (16.4%), and not resolving in 71 (13.9%). New neurologic deficits were noted following 34 (0.52%) cases: 5 of 6,052 (0.08%) with no alerts, 2 of 357 (0.6%) with fully resolved alerts, 11 of 84 (13.1%) with partial resolution and 16 of 71 (22.5%) with no resolution. Logistic regression & post-hoc test of multiple comparisons showed no significant difference in rate of postoperative deficit between patients with no alerts and those with fully resolved alerts (p=0.095). Compared to patients with no alerts, those with a partially resolved alert were much more likely to experience a postoperative deficit (OR=182.2, 95%CI [64.6, 590.64], p< 0.001), as were those with an unresolved alert (OR=351.8, 95%CI [132.7, 1106.64], p< 0.001). Summary and
Conclusion: The majority of intraoperative neuromonitoring alerts in the present series of pediatric spine surgeries were resolved successfully. Patients with fully resolved neuromonitoring alerts had neurologic outcomes comparable to those with no neuromonitoring changes during surgery. Risk of postoperative neurologic complication increased progressively for patients with partially resolved and unresolved alerts, respectively. Accurate detection of negative changes in neurophysiologic function by the IONM team allows for timely intraoperative intervention that can mitigate postoperative neurologic complications in pediatric spine surgery.