Utility of intraoperative neuromonitoring and outcomes of neurological complication in lower cervical and upper thoracic posterior-based three-column osteotomies for cervical deformity.

Abstract

For severe and rigid adult cervical deformity, posterior-based three-column osteotomies (3COs) are warranted, but neurological complications are relatively high with such procedures. The performance measures of intraoperative neuromonitoring (IONM) during cervicothoracic 3CO have yet to be studied, and there remains a paucity of literature regarding the topic. Therefore, the authors of this study examined the performance of IONM in predicting new neurological weakness following lower cervical and upper thoracic 3CO. In addition, they report the 6-month, 1-year, and 2-year outcomes of patients who experienced new postoperative weakness. The authors performed a retrospective review of a single surgeon's experience from 2011 to 2018 with all patients who had undergone posterior-based 3CO in the lower cervical (C7) or upper thoracic (T1-4) spine. Medical and neuromonitoring records were independently reviewed. A total of 56 patients were included in the analysis, 38 of whom had undergone pedicle subtraction osteotomy and 18 of whom had undergone vertebral column resection. The mean age was 61.6 years, and 41.1% of the patients were male. Among the study cohort, 66.1% were myelopathic and 33.9% had preoperative weakness. Mean blood loss was 1565.0 ml, and length of surgery was 315.9 minutes. Preoperative and postoperative measures assessed were cervical sagittal vertical axis (6.5 and 3.8 cm, respectively; p < 0.001), cervical lordosis (2.3° and -6.7°, p = 0.042), and T1 slope (48.6° and 35.8°, p < 0.001). The complication rate was 49.0%, and the new neurological deficit rate was 17.9%. When stratifying by osteotomy level, there were significantly higher rates of neurological deficits at C7 and T1: C7 (37.5%), T1 (44.4%), T2 (16.7%), T3 (14.3%), and T4 (0.0%; p = 0.042). Most new neurological weakness was the nerve root pattern rather than the spinal cord pattern. Overall, there were 16 IONM changes at any threshold: 14 at 50%, 8 at 75%, and 13 if only counting patients who did not return to baseline (RTB). Performance measures for the various thresholds were accuracy (73.2% to 77.8%), positive predictive value (25.0% to 46.2%), negative predictive value (81.3% to 88.1%), sensitivity (18.2% to 54.5%), and specificity (77.8% to 86.7%). Sensitivity to detect a spinal cord pattern of weakness was 100% and 28.6% for a nerve root pattern of weakness. In patients with a new postoperative deficit, 22.2% were unchanged, 44.4% improved, and 33.3% had a RTB at the 2-year follow-up. Complication rates are high following posterior 3CO for cervical deformity. 3CO at C7 and T1 has the highest rates of neurological deficit. Current IONM modalities have modest performance in predicting postoperative deficits, especially for nerve root neuropraxia. A large prospective multicenter study is warranted.