The role of iatrogenic foraminal stenosis from lordotic correction in the development of C5 palsy after posterior laminectomy and fusion

BACKGROUND Post-operative C5 nerve root palsy is a known complication following cervical spine surgery. Although several theories have been proposed, there remains no consensus as to the etiology of the palsies. Multiple pre-operative radiographic measures have been assessed for utility in predicting palsy. The purpose of this study is to evaluate published radiographic parameters as well as specifically evaluate the effect of cervical lordosis in the development of C5 palsy to establish thresholds that reliably predict the incidence. METHODS This study is a retrospective review of 54 consecutive multilevel cervical laminectomy and fusion surgeries performed by a single spine surgeon between June 2007 and February 2014. Pre-operative MRI and pre- and post-operative plain films were assessed to measure anteroposterior diameter (APD) of the spinal cord, cervical laminar angles, anteroposterior foraminal diameters (FD), cervical curvature index (Ishihara), cervical spine angle (C2-7), and C4-5 angle. Univariate analysis through independent t tests was used to compare differences between groups. Stepwise logistic regression was performed to identify pre-operative variables associated with C5 palsy. Receiver operating characteristic curves were created for significant variables to assess predictive accuracy through determining the area under the curve. RESULTS There were 13 (24%) palsies in the 54 patients in the study. All palsies completely resolved within 6 months. Among pre-operative measures, FD and APD were significantly different between the palsy and non-palsy groups. The average post-operative C4-5 angle was significantly different between the groups, though the cervical spine angle and curvature index, as well as the change in these measures from pre-operative measurements, did not differ significantly between groups. CONCLUSIONS Post-operative palsy is likely a result of iatrogenic nerve root compression from a decreased in cross-sectional area of the neuroforamen in a patient with pre-operative narrowing of the foramen. However, spinal cord drift back may also play a role from the combined effect of posterior decompression from laminectomy and relative slack afforded by increased lordosis. Accordingly, increased post-operative lordosis would increase the likelihood of effect from both of these mechanisms. We recommended limited conservative lordotic correction in patients with pre-operative foraminal narrowing.