Intraoperative Spinal Cord and Nerve Root Injuries

Anterior Cervical Spine Surgery

Since Robinson and Smith first described the approach to the anterior cervical spine in 1955, anterior cervical surgery has become the preferred choice in the treatment of most cervical pathologies ( Fig. 95.1 ). During the approach, the recurrent laryngeal nerve (RLN) is at risk in its path running between the esophagus and trachea, with injury resulting in unilateral vocal cord paralysis. The reported rates of RLN injury have ranged from 0.8% to as high as 11%. It has been heavily debated regarding which is the appropriate side of neck to approach the cervical spine. Various reasons for a surgeon to choose one side over another include surgeon comfort and previous neck surgery, while some have cited anatomic considerations. Classic anatomic teaching has shown that the left recurrent laryngeal nerve traverses in a more midline course and is less likely to be encountered in a left-sided approach compared to a right-sided approach, where it travels in a more lateral course on its path to the tracheoesophageal groove. Beutler et al. compared a consecutive series of 328 anterior cervical spine fusion procedures, 173 procedures completed from the right side and 155 completed from the left. The overall incidence of RLN injury was 2.1%. The incidence of RLN injury with reoperative anterior cervical fusion surgery was 9.5%. They reported no difference in incidence of RLN injury with the side of surgical approach and surmised that the surgeon may safely approach the cervical spine from the side of personal preference and experience.

Another neurologic structure at risk during the anterior approach to the cervical spine is the sympathetic chain. It is found between the posterior lateral pharyngeal and carotid regions. The cervical sympathetic chain comprises three ganglia. The superior ganglion is at C2–C3; the middle ganglion is adjacent to C6; and the stellate ganglion lies in the costotransversopleural fossa. Damage to the sympathetic chain, the stellate ganglion in particular, can result in Horner syndrome. Horner syndrome is characterized by the classic triad of miosis, partial ptosis, and loss of hemifacial sweating. It has a reported incidence between 0.2% and 4%. Resolution typically occurs over several weeks to 1 month. Raising the medial edge of the longus colli muscle without excessive traction is the key to prevention.

The most dreaded complication in cervical spine surgery is injury to the spinal cord itself. Fortunately, the reported incidence is quite low, with a reported rate of less than 1%. In a 1989 survey of members of the Scoliosis Research Society, rates of spinal cord injury following anterior cervical surgery were lower than those reported with posterior procedures. Preoperative neurologic status and reason for surgery play a clear role in risk of injury. Myelopathic patients appear to be the most at risk, not surprisingly, as there is little room for error during canal decompression. In a study of over 1000 patients, Fountas et al. reported a worsening of myelopathy in 3 patients (0.2%).

The most common reason to perform anterior cervical surgery is recalcitrant radicular pain. Unfortunately, worsening radiculopathy is a known complication following anterior cervical surgery. In a series of 450 consecutive cases of cervical discectomy without fusion, a worsening radiculopathy was noticed in 1.3% of patients. In a single surgeon's experience of over 1500 cases of discectomy with fusion, a worsening radiculopathy at the index level of operation occurred in 14 patients (0.88%).

The most common worsening radiculopathy following cervical surgery is a C5 palsy. More often associated with posterior cervical surgeries, there are reports of C5 palsy following anterior cervical surgery, most commonly following cervical corpectomy. Saunders et al. reported one case of C5 palsy in a retrospective review of 40 patients undergoing multilevel corpectomies. Wada et al. found a slightly higher rate of 4% with an average of 2.5 levels of corpectomy. In a retrospective series of 563 anterior cervical surgeries, Ikenaga and colleagues reported a 3.25% incidence of C5 palsy, all occurring in patients with opacification of posterior longitudinal ligament or suffering from cervical spondylotic myelopathy.

Posterior Cervical Spine Surgery

There are several types of posterior cervical surgery. Some of the more commonly used posterior cervical procedures include posterior foraminotomy, laminoplasty, laminectomy, and laminectomy with fusion and instrumentation.

The rate of neurologic injury following posterior foraminotomy is very low. In a retrospective review of 95 operative levels, Skovrj et al. reported one case of radiculitis. Jagannathan et al. report a 1.2% incidence of neurologic injury, all cases being a C5 palsy following a C4–C5 foraminotomy. Grieve et al. reported a slightly higher incidence of worsening radiculopathy after surgery (7%) in 77 patients undergoing one-level posterior foraminotomies. There were no cases of spinal cord injury reported in these studies. The complication profile of this procedure will be at the forefront of discussions due to renewed interest in the technique given the development of minimally invasive access and concern for adjacent-level disease with anterior cervical discectomy and fusion.

Laminoplasty and laminectomy are both widely used procedures to address multilevel spinal stenosis and myelopathy. The most commonly encountered neurologic complication is a C5 palsy. A large systematic review of over 103 studies and over more than 8500 patients looking specifically at the rate of C5 palsy during laminoplasty found that 16% of the studies reported a C5 palsy rate of greater than 10% (534 patients), 41% reported a rate of 5% to 10% (1006 patients), 23% reported a rate of 1% to 5% (857 patients), and 12.5% reported a rate of 0% (168 patients). Another review specifically addressing the rate of C5 palsy with various techniques of posterior cervical surgery showed an overall incidence of 5.8%. The incidence after open-door laminoplasty, double-door laminoplasty, and laminectomy was 4.5%, 3.1%, and 11.3%, respectively. In a systematic review comparing laminectomy versus laminoplasty for cervical spondylotic myelopathy, C5 paresis was found in 9 of 176 (5.11%) patients treated with laminoplasty and 12 of 157 (7.64%) patients treated with laminectomy with no significant difference found between the rates of injury.

Despite the well-known risk of C5 palsy following cervical decompression, there is no definitive explanation for why it occurs. Its presentation after surgery can appear acutely or in a delayed fashion, occurring up to 2 months after the index procedure. It will manifest clinically as deltoid weakness and potential biceps weakness with sensory loss. It typically presents unilaterally; on rare instances, however, bilateral paresis is apparent. Aside from actual iatrogenic mechanical injury to the nerve during surgery, several other causes have been reported in the literature. Tethering of the nerve results from shifting of the spinal cord, spinal cord ischemia, and reperfusion injury of the spinal cord. Intraoperative spinal cord monitoring (discussed in further detail elsewhere in this text) has been shown to be a valuable tool for intraoperative detection of impending C5 palsy. However, not all cases of C5 palsy are detected intraoperatively with monitoring, which supports the theory that some cases of C5 palsy occur after surgery is complete.

The addition of instrumentation to spine surgery increases the rate of a number of complications. Theoretically, too long a screw placed through an anterior cervical plate could result in neurologic injury by direct spinal cord compression; no literature could be found describing this complication as yet.

However, the use of lateral mass fixation in posterior cervical surgery has been reported in multiple studies as a cause of iatrogenic neurologic injury ( Fig. 95.2 ). Lateral mass screws are frequently used to augment posterior fusion techniques. Although not currently Food and Drug Administration (FDA)-approved for this application, lateral mass screws have been widely used over the last several decades. The original technique was first described by Roy-Camille in the late 1970s. Since that description, various authors have attempted to modify the technique in attempt to improve screw purchase and reduce complications related to screw malposition. Al Barbarawi et al. reviewed a total of 2500 screws placed in 430 patients with a range of pathologies. They report a total of five screws requiring removal for worsening radiculopathy. They also reported 20 patients overall with a worsening C5 radiculopathy following surgery that resolved over time, although two patients did undergo screw revision at the C5 level. In a 2013 systematic review of the complications associated with lateral mass screw fixation, Coe et al. reviewed 20 articles from the literature and found the overall rate of neurologic injury to be 3.9%. The incidence was inclusive of C5 palsy and radiculopathy. The rate of injury attributed to screw malposition was 1.0%. These authors further reported the rate of nerve root injury per screw placement (N = 5771) to be 0.17%. Not surprisingly, given the trajectory of lateral mass screws, there were no reported cases of spinal cord injury.

A less commonly used technique of instrumentation within the cervical spine, at least in North America, is pedicle screw fixation. The concerns with this technique are the small pedicle diameters and risks associated with medial and lateral perforation. With medial or caudal perforation, there is risk to the spinal cord and nerve roots; with lateral perforation, there is risk of vertebral artery injury. In light of these risks, the interest in transpedicular screw fixation centers on improved biomechanical properties compared to lateral mass screws. Transpedicular screw fixation has been shown to have far superior pullout strength than lateral mass screws. Despite its technical demands, this use of this technique has not been shown to have a significantly greater risk of neurologic injury compared to other posterior fixation techniques. Abumi et al. reported on 180 patients and 712 screws. These authors attributed two cases of nerve root injury to screw malposition. One case of nerve root injury resolved with observation, whereas one case of muscle weakness required screw removal, with subsequent improvement in strength. In another study evaluating placement of cervical pedicle screws, 94 screws were implanted in 26 patients. Eight screws (9%) had a critical breach based on postoperative computed tomography (CT) evaluation. Three screws passed through the intervertebral foramen, causing temporary paresis in one case and a new sensory loss in another. In the latter patient, revision surgery was performed. Most screw malposition complications occurred in screws placed from C3 to C5. This is not surprising, as anatomic studies have shown smaller pedicles at C3 and C4 compared to C5 and C6. In a systematic review comparing 10 studies of lateral mass screw fixation and 12 studies of pedicle screw fixation, there was a 0.19% incidence of nerve injury per lateral mass screw placed and 0.31% per pedicle screw placed. In the largest study to date, a review of 283 patients and more than 1000 pedicle screws revealed nerve root irritation attributed to pedicle screw insertion in three patients, all of whom required screw removal. There were no cases of spinal cord injury reported.