Surgical_Treatment_of_Spinal_Injury.doc

Surgical Treatment of Spinal Injury

Daniel R. Fassett

James S. Harrop

Traumatic injuries to the spinal column are common events, with more than 50,000 fractures to the spinal column occurring annually in the United States (1). Spinal injury remains a heterogeneous group of injuries and therefore various strategies are employed in their treatment. Multiple clinical variables must be addressed, including the degree of ligamentous and bony injury, the presence of neurologic deficits, perceived patient compliance, and overall health status; these factors are used to determine how the injuries are treated. Treatment can range from simple limitation in activity to external orthosis to open reduction and internal fixation with spinal instrumentation. The goal of treating these injuries is to utilize the least invasive surgical technique to stabilize the injured segment while limiting the potential for subsequent catastrophic neurologic injury, progression of a deformity, and chronic pain conditions. These surgical goals are also tempered by other medical management issues that focus on minimizing hospitalization and immobilization and maximizing the benefits of early and aggressive rehabilitation.

Historical Perspective of Spinal Injury Treatment

Treatment of traumatic spinal injuries was first recorded by Hippocrates (460-370 BCE) who used traction devices to obtain spinal reduction and advocated external stabilization and immobilization. Surgery was not considered a viable option at this time because of the high mortality of surgical techniques, and the presence of neurologic deficits in the setting of spinal trauma was deemed universally fatal. Surgical decompression for the treatment of traumatic spinal cord injury was initially popularized by Paulus of Aegina (625-690 CE) but was not universally accepted because of very poor surgical outcomes at the time. In 1646, Fabricius Hildanus performed the first documented open reduction of a spinal fracture (2, 3, 4, 5, 6 and 7).

It was not until the advent of spinal instrumentation in the 1950s that a more aggressive surgical approach was favored in the treatment of spinal column injuries. Before the development of spinal instrumentation, there was a bias toward conservative treatment, which often involved

P.40

long periods of immobilization (4 to 8 weeks commonly) typically with traction to restore the spinal alignment and allow the fractures time to heal (8). These long periods of immobilization were associated with significant medical complications including pneumonia, deep vein thrombosis, and decubitus ulcers. The use of spinal instrumentation provided surgeons the ability to restore immediate stability to the spinal column, thus allowing for earlier mobilization and fewer complications from prolonged immobilization. In addition, spinal instrumentation theoretically improved fusion rates by providing a stable environment of bone healing, thus reducing the risks of late neurologic deterioration due to spinal instability, progressive spinal deformity, and associated axial back pain syndromes. Even with improvements in instrumentation, it was realized that all instrumentation will fail eventually unless a bony fusion is achieved and, therefore, arthrodesis remains a critical part of all spinal stabilization surgeries (4,6).

Clinical and Radiographic Evaluation of the Trauma Patient

The treatment of spinal trauma consists of an assessment of the traumatic injury through a detailed neurologic examination, physical examination, and then a radiographic evaluation. Radiographic evaluation often begins with plain radiographs followed by supplemental imaging of questionable areas of injury. Although modern imaging techniques have greatly aided in the diagnosis of fractures, determination of ligamentous instability with imaging alone is still unproven even with techniques designed to evaluate the soft tissues such as magnetic resonance imaging (MRI) (9).

Cervical Spine Evaluation

Any trauma patient should immediately be placed in cervical spine immobilization when assessed by emergency medical services (EMS) in the field. Any nonintoxicated patient without neck pain, neurologic deficits, and distracting injuries (injuries to other portions of the body that could potentially mask the pain associated with spinal injury) can be cleared of cervical spine injury with a normal clinical examination alone (i.e., showing no neck pain over a full range of motion of the cervical spine) (10). Neurologically intact patients with neck pain or tenderness are usually assessed with three view (anteroposterior [AP], lateral, and open-mouth odontoid views) plain radiographs as initial assessment (11). If these plain radiographs are normal, these patients are often kept in cervical collar immobilization for 1 to 2 weeks and then should have delayed passive cervical flexion and extension imaging to assess for potential occult ligamentous injury. Although the prevalence of occult ligamentous injury in the setting of normal radiographs is small, the delay in the follow-up flexion/extension imaging can minimize false negative results by allowing muscle spasm to subside. In the neurologically intact patient with severe neck pain and normal plain radiographs, computed tomography (CT), and possibly MRI should be considered to rule out an occult fracture or herniated disc not seen on the plain radiographs (11).

In comatose, obtunded, or intoxicated/sedated patients, where an adequate neurologic examination cannot be obtained, plain radiographs or CT scan are standard in most trauma protocols. With the increase in speed and resolution of multidetector helical CT scanning, this modality is becoming more popular for evaluating multitrauma patients in a time-efficient manner. If these patients remain comatose, dynamic flexion/extension studies with fluoroscopic guidance or a normal cervical spine MRI within 48 hours of injury is sometimes performed for cervical spine clearance, although the inherent value of either method for the exclusion of occult soft tissue injury is questionable (9,11).

Patients with neurologic deficits that are clinically attributable to a spinal cord injury deserve rapid radiographic assessment possibly including plain films, CT scanning, and MRI. In the setting of an obvious cervical spine deformity with neurologic deficits, some surgeons may immediately institute reduction measures such as cervical traction. Other surgeons may insist upon further evaluation with CT and MRI before initiating any reduction measures. The extent of radiographic workup in the setting of spinal cord injury will depend on the preferences of the individual surgeon, the unique characteristics of the fracture being evaluated, and the character of neurologic examination. Patients with incomplete spinal cord injuries, where there is some neurologic function below the level of the spinal cord injury, may warrant an emergent MRI examination to assess integrity of the spinal canal and rule out herniated discs as an explanation for the neurologic deficits. The patient with a progressive, incomplete neurologic deficit requires immediate assessment and treatment as these patients have the greatest potential to permanently lose function with treatment delay.

Thoracic and Lumbar Spine Evaluation

Awake, neurologically intact patients can have thoracic and lumbar spine precautions discontinued if they do not have any pain suggestive of spinal injury and do not have distracting injuries. Neurologically intact patients that complain of pain localizing to the spine or who harbor a distracting injury should be evaluated radiographically with a minimum of AP and lateral plain radiographs.

P.41

Depending upon the severity of their symptoms, CT or MRI imaging may be warranted. Comatose, obtunded, or sedated/intoxicated patients should always be evaluated with plain films or CT scanning. Multisystem trauma patients often require routine CT imaging of the chest, abdomen, and pelvis. It has been suggested that limited resolution imaging of the thoracic and lumbar spine can be extracted from these data sets and used as a substitute for radiographs of these areas (12).

In patients with neurologic deficits where there is a high suspicion for spinal injury, CT scans with coronal and sagittal reconstructions are often the initial imaging modality to improve the sensitivity for diagnosis of spinal injury and also provide better anatomic details about the specific fracture. A patient with a persistent neurologic deficit and a “normal” CT scan warrants performance of an emergent MRI both to visualize the spinal cord and cauda equina and to rule out soft tissue etiologies of spinal column compromise such as herniated discs or epidural hematoma that may be not visualized with CT scanning. Some surgeons may wish to obtain emergent MRI in patients with obvious fractures diagnosed with CT, since the MRI can help locate the level of the conus medullaris, assess the integrity of the intervertebral discs, and better appreciate the extent of ligamentous injury. All of these factors may impact the treatment of the patient by providing the surgeon with a better appreciation of the anatomy of the spinal injury.

Current Treatment Options

External Orthosis

Numerous external orthosis (spinal braces) options are available for the treatment of spinal injuries. The principle of bracing is to reduce motion at the injured spinal area in order to improve the likelihood of healing and reduce the potential for neurologic injury as a result of spinal instability. In general it is felt that maximal reduction in motion will result in better healing of the injured spinal segment, but literature is lacking in regard to how much motion is “too much” when considering bracing. Indications for external orthosis following spinal injury can vary significantly among individual surgeons since there are limited guidelines in the surgical literature for this type of treatment. Some fractures may not require any bracing as they are deemed to be very low risk for spinal instability and other fractures may be stabilized surgically, thus eliminating the need for external orthosis.

For the cervical spine, options ranging from least to most restrictive are soft and hard cervical collars (Philadelphia, Aspen, Miami J), cervical bracing with the addition of a thoracic vest (SOMI and Minerva braces), and halo-vest immobilization (Fig. 3-1). A cervical collar is the least cumbersome of the cervical spine orthosis options; however, this comes at the cost of it offering the least support in terms of limiting range of motion. Studies have shown that cervical hard collars allow for over 30 degrees of flexion-extension motion in the cervical spine and provide minimal support at the lower cervical spine (13). Braces that add a thoracic vest immobilize the cervical spine and cervicothoracic junction better but still allow for significant motion at the craniocervical junction (Fig. 3-1B) (14,15). Halo-vest immobilization (Fig. 3-1C) accomplishes the most rigid immobilization by fixating a halo-ring around the head (pins into the skull) and securing the halo-ring to a thoracic vest by rods. Although halo immobilization provides the most support and may improve fusion rates, it may be associated with complications ranging from pin loosening, pin site infections, to swallowing dysfunction, reduced immobilization, and cerebral abscesses attributable to intracranial penetration of fixation pins. Halo immobilization also tends to limit motion of the upper cervical spine with greater efficiency than the middle and lower cervical spine. Even with halo immobilization, studies have shown that 2 to 10 degrees of motion can take place at the craniocervical junction, and the lower cervical spine and cervicothoracic junction may not be adequately mobilized (14). In addition, immobilization in a halo can cause limited motion at the ends of the spine (craniocervical and cervicothoracic) with exaggerated motion in the subaxial spine, referred to as snaking (14).

In the thoracic spine, the rib cage provides some natural support for thoracic spine fractures. The upper thoracic region (T5 and above) is a very difficult region to immobilize with external orthosis, unless the patient is immobilized with a halo orthosis with a long thoracic vest. Spinal fractures from T6 to L2 are typically braced with a custom molded, hard-shell orthosis (thoracolumbar-sacral orthosis [TLSO]) or with more versatile, adjustable-fit braces (e.g., Jewitt, Aspen) (Fig. 3-2A) or clamshell brace (Fig. 3-2B). Below L3, a lumbosacral orthosis is used for support. In addition, to increase the immobilization at the lumbosacral junction, a leg extension can be fitted to the orthosis to assist in limiting motion across the pelvis. Casting (Fig. 3-2C) is another option for lumbar and thoracolumbar fractures and can provide better support and eliminate concerns of noncompliance.

Surgical Options for Traumatic Spinal Injuries

Controversy persists in the surgical community regarding the optimal treatment of many traumatic spinal injuries, especially regarding timing of surgical intervention and type of surgical approach. Surgical intervention is often

P.42

advocated to (a) decompress the neural elements in cases of neurologic deficit; (b) prevent possible late neurologic injury in unstable fractures; (c) correct and prevent deformity that could result in chronic axial (back) pain or neurologic loss; and (d) provide for early mobilization, thus avoiding the complications of prolonged bed rest.

Anterior (ventral), posterior (dorsal), and combined anterior and posterior approaches can be used to treat traumatic spinal instability. The surgical approach selected may depend on the fracture pattern, the neurologic status of the patient, and the individual preference of the surgeon. Anterior approaches may be favored in situations

P.43

where a herniated disc or bone fragment is causing ventral compression on the spinal cord. In addition, fracture patterns where the integrity of the anterior column of the spine is significantly compromised (unstable spine) may be best addressed by an anterior approach to restore the structural stability of the anterior spinal column. In either case, the surgical approach also includes some form of instrumentation. Spinal instrumentation is a method of straightening and stabilizing the spine after spinal fusion, by surgically attaching hooks, rods, and wire to the spine in a way that redistributes the stresses on the bones and keeps them in proper alignment.

Posterior surgical approaches and instrumentation typically allow for better reduction when deformities are present and may benefit in restoring the posterior tension band in distraction-type injuries where there is disruption of the posterior ligamentous structures. The posterior ligamentous structures (ligamentum flavum, interspinous ligaments, supraspinous ligaments, and so forth) serve to hold the spine in normal alignment and since

P.44

they are under tension in most parts of the spine they are referred to collectively as the posterior tension band. Injury to these ligamentous structures can allow the spine to deform into a more kyphotic posture. With posterior instrumentation, there is restoration of the biomechanical forces needed to hold the spine in normal alignment. In terms of restoration of alignment, posterior instrumentation (lateral mass screws) typically provides better fixation and mechanical advantage that can be used in spinal reduction maneuvers to better restore spinal alignment.

In translation injuries (fracture-dislocations), when there is severe, circumferential disruption of the spinal column, combined anterior-posterior instrumentation procedures may be used to maximize stability of the spinal column and increase the fusion rates. Circumferential spinal instrumentation (anterior and posterior combined operations) is more commonly utilized in areas of high biomechanical stress, such as the cervicothoracic junction and thoracolumbar junction, where the biomechanical forces on the spine are greater and make these areas more prone to failure of stabilization procedures.

There is no single preferred approach to many types of spinal fractures; frequently the preferences of the individual surgeon take precedence. Despite the maturation of surgical techniques and development of sophisticated spinal instrumentation devices, there is a lack of good guidelines for the treatment of many fractures. In general, posterior approaches to the thoracic and lumbar spine are often favored because of the ease and familiarity of approach. Anterior approaches to the thoracic and lumbar spine tend to be more technically challenging (mobilizing the lung, viscera, and great vessels) and may require the assistance of a general or thoracic surgeon to aid with the approach to the spine.

Treatment of Cervical Spine Injuries

Occipital Condyle Fractures

Occiptial condyle fracture is an uncommon injury occurring in less than 3% of patients with blunt craniovertebral trauma (16,17). CT is required to diagnose this injury as there is less than 3% diagnostic sensitivity with plain radiographs (18). These fractures were first classified by Anderson and Montesano (19) into (a) Type I—comminuted due to axial compression, (b) Type II—extension of a basilar skull fracture through the occipital condyle, and (c) Type III—an avulsion of the occipital condyle likely due to a rotational force that avulses a portion of the occipital condyle with the alar ligament (Fig. 3-3). There is a lack of adequate studies to determine the optimum treatment strategy for these fractures. Most surgeons consider type I and II fractures stable injuries and will recommend cervical collar immobilization alone as an option to reduce pain associated with this injury.

Type III occipital condyle fractures are considered to be mechanically unstable and have been associated with development of lower cranial nerve deficits if untreated. Translation ≤ 1 mm between the occipital condyles and lateral masses of C1 at the occipital-C1 joint is considered abnormal. Most unilateral type III fractures are treated with cervical collar immobilization, but some surgeons advocate halo immobilization for fractures that have features of instability such as marked fracture displacement or abnormal craniocervical alignment. There are no specific guidelines or measurements that predict which unilateral type III fractures are at risk for long-term instability. After a period of immobilization, unilateral fractures can be evaluated in follow-up with CT scanning to assess for the extent of bone union across the fractured segment, and flexion/extension radiographs can be useful to assess for stability at the occipitocervical junction. Gross instability at the occipitocervical junction is presumed for the rare bilateral type III occipital condyle fractures, and atlanto-occipital dislocation (AOD) can be a component of this injury. When the features of AOD are present, an occipital cervical fusion is the preferred method of treatment or in any patient that continues to have instability despite conservative therapy with external immobilization. (20,21).

Atlanto-Occipital Dislocation

AOD has a significantly high fatality rate as a result of the significant forces required to create this injury. AOD is commonly associated with significant intracranial injury as well as vertebral artery injuries as a result of this distraction injury across the craniocervical junction. With improvements in the early recognition and stabilization of spinal injuries by EMS, more patients are surviving this injury. As a result of the tremendous distractive forces associated with the AOD, the tectorial membrane, posterior ligamentous structures, and facet capsules between the atlas and occipital condyles are injured, yet surprisingly these injuries can be difficult to detect on radiography and a high degree of vigilance is required. Several diagnostic criteria exist to help diagnose this injury on lateral radiographs including (a) the Powers ratio (22), (b) basion-dens distances, (such as Harris's rule of 12) (23, 24 and 25), (c) distances from posterior mandible to anterior arch of C1 or dens (Dublin method) (26), and (d) Lee's X-line method (27) (Table 3-1). Of these diagnostic options, Harris's rule of 12 appears to be the most sensitive means of diagnosing this injury on plain films or sagittal reformatted CT images (Fig. 3-4A). MRI potentially can also be

P.45

very beneficial by showing the ligamentous disruption at the craniocervical junction.

AOD is considered highly unstable because of the extent of ligamentous injury and requires surgical stabilization with occipitocervical fusion procedures that instrument bridge across the occiput and upper cervical spine via a posterior approach (Fig. 3-4B).

Jefferson Fracture

Bilateral fractures through the ring of C1 (classic Jefferson fracture) (Fig. 3-5A) and other fractures of C1 can typically be treated with conservative measures (collar or halo immobilization) because of the high rate of spontaneous fusion and limited ligamentous instability. Integrity of the transverse ligament is used as a determinant of stability and the need for possible surgical stabilization. The most common means of evaluating the integrity of the transverse ligament is with an open-mouth odontoid view radiograph to assess the alignment of the lateral masses of C1 and C2 using the rule of Spence (28). Greater than 7 mm of combined lateral overhang of the lateral masses of C1 on C2 constitutes violation of the rule of Spence and suggests likely transverse ligament rupture (Fig. 3-5B). The transverse ligament may also be evaluated on MRI, but the application of MRI in detecting transverse ligament rupture is unproven (29). Flexion-extension plain films can also be used to assess for possible C1-2 instability. In the presence of C1-2 instability from transverse ligament rupture, C1-2 arthrodesis is recommended via wiring techniques, transarticular screws, or other C1 and C2 screw techniques (Fig. 3-6). Various rods, plates, and wire loop (Fig. 3-6A) constructs are available to stabilize the craniocervical junction. These systems generally provide screw fixation into the posterior occiput at the cephalad end. For fixation at the caudal end, a variety of devices can be used, including atlantoaxial transarticular screws (screws placed through the C2 pars interarticularis, across the C1-2 lateral mass articulation, and into the lateral mass of C1) (Fig. 3-6B), C2 pars interarticularis or

P.46

pedicle screws, and C2 laminar screws (Fig. 3-6C). Extension of the construct to the subaxial spine with lateral mass screws can provide improved fixation in some cases where bone quality or poor screw purchase is a concern.