Preoperative Clinical and Radiographic Variables Predict Postoperative C5 Palsy

Abstract

Postoperative C5 palsy affects 8% of patients undergoing posterior cervical decompression. It is associated with a period functional disability that may exceed 12 months and increase direct care costs > $2000. All patients undergoing posterior cervical decompression at a single tertiary-care facility for degenerative conditions were evaluated for preoperative imaging, clinical presentation, surgical operation, and postoperative course. We sought to identify those variables predictive of postoperative C5 palsy. Of 221 included patients (mean age, 63 years; 54% male), 12.2% experienced C5 palsy. On univariate analysis, C5 palsy was associated with foraminal diameter (P= 0.0005), spinal cord cross-sectional area (P= 0.11), number of levels undergoing laminectomy (P= 0.14), and clinical presentation of dropping objects (P= 0.07), hand clumsiness (P= 0.13), or paresthesias in the upper extremities (P= 0.08). Foraminal diameter (odds ratio, 0.31 per mm increase; 95% confidence interval, 0.16-0.60; P < 0.001) and patient report of gait disturbance (odds ratio, 0.53; 95% confidence interval, 0.33-0.85; P= 0.008) were the only significant predictors on multivariate analysis. A foraminal diameter <2 mm had 2-fold greater odds of postoperative C5 palsy. A receiver operating curve for the multivariate logistic model had an associated C-statistic of 0.7818. The absolute error of this model was 9.3% on internal validation. Foraminal stenosis most strongly predicted postoperative C5 palsy. A proof-of-concept model incorporating foramen size, as well as clinical complaints of paresthesias, hand clumsiness, and gait abnormality, successfully predicts the occurrence of postoperative palsy with an overall accuracy of 78%.