Rockwood & Green’s Fractures in Adults
6th Edition

There remains considerable disagreement concerning the optimal method to rule out lower cervical spine injury (21). The rate of missed injuries has varied from 10% to 48% (22,23,24). Infrequently, this results in neurologic deterioration (22,25). Plain x-rays are useful in detecting and describing lower cervical injuries. They clearly demonstrate the majority of injuries. A standard cervical series includes lateral, anteroposterior (AP), and open-mouth views. Between 83% and 99% of injuries can be detected using these three views (26). Oblique views are frequently obtained; however, their utility in the traumatic setting is limited. They may be somewhat helpful in visualizing fractures of the laminae or articular processes.

An x-ray series that does not allow visualization of the entire cervicothoracic junction should be considered inadequate (27). Pulling traction on the patient’s arms can facilitate imaging this region. However, this may be not feasible with concomitant injuries such a shoulder dislocation or scapulothoracic dislocation. A swimmer’s view entails placing one arm in a fully abducted position, while leaving the other arm at the patient’s side. This somewhat diminishes the obstructing shadows of the deltoid and shoulder joint to allow clearer imaging of the lower cervical and upper thoracic vertebrae.

Various features of the lateral cervical x-ray should be noted. Prevertebral swelling can be detected by assessing the soft tissue shadow thickness anterior to the vertebral bodies (see Fig. 39-1). If this measures more than 7 mm in front of the C2–3 disc space or more than 21 mm at the C6–7 disc space, there is a high
likelihood of a cervical spinal injury (28,29). Of importance, the absence of prevertebral (retropharyngeal) soft tissue swelling does not reliably rule out an occult cervical spine injury, as its reported sensitivity is only 65% (30).

In the awake, cooperative, and alert patient with negative static plain films and no other distracting injuries, lateral flexion-extension views of the cervical spine can be obtained. They can demonstrate instability in up to 8% of patients with negative plain films (31); however, false negatives can occur (21). In a patient with neck pain, they are best delayed until spasm has subsided, which usually takes about 2 weeks. Spasm can mask subtle instability, yielding a false negative study if performed too early after injury (31,32). Flexion-extension views may be of limited utility in detecting instability at the cervicothoracic junction in patients with high-riding shoulders. Additional advanced imaging techniques may be warranted. In the author’s practice, flexion-extension views are obtained by having the patient actively flex and extend the neck while in the seated or standing position. Thirty degrees of excursion is the minimum acceptable amount to be considered an adequate method of ruling out ligamentous injury.

The practitioner must be familiar with the normal anatomic relationships and landmarks of the cervical spine. On a good-quality AP view with the patient looking forward, the spinous processes should be aligned. Displacement of a spinous process to one side can suggest a rotational injury such as a unilateral facet dislocation, facet fracture, or displaced pedicle fracture (33). The distance between each spinous process should also be gauged on the AP view, because widening can suggest posterior ligamentous injury.

The lateral view is the so-called workhorse of injury detection and description (34). Normally unbroken lines include the spinolaminar line, posterior vertebral body line, and anterior vertebral body line (see Fig. 39-1). The spinolaminar line is perhaps the most useful, because it is not usually affected by spondylotic changes such as osteophytes, which may be present along the posterior vertebral body and, to a greater extent, the anterior vertebral body. In older patients with substantial degenerative changes, the morphology of the anterior vertebral body can be quite abnormal, making this landmark/line difficult to assess. The lateral view is useful in judging the interspinous process distances. These can be measured at each level and compared. Substantial widening at one level suggests disruption of the posterior ligamentous complex. The facet joints normally appear as stacked parallelograms on the lateral cervical x-ray. In a perfect lateral view, the joints on either side should overlie each other, facilitating assessment of joint apposition.

Plain cervical x-rays are crucial to the overall description of detected injuries. A number of x-ray injury characteristics can be assessed and measured on a lateral cervical film. Segmental kyphosis can be measured by an end plate (Cobb) method or posterior vertebral body tangent method (Figs. 39-2 and 39-3). The latter may have lower interobserver variability. Kyphosis more than 11 degrees as measured by the end plate method is strongly suggestive of posterior ligamentous injury and potential instability (35). Kyphosis can be the result of posterior widening of the interspinous processes and facet joints about an injury axis of rotation near the anterior longitudinal ligament. Vertebral body fracture with height loss and interspinous gapping suggests an injury axis of rotation about the facet joints.

Anteroposterior translational deformity can be assessed and measured on the lateral view. The distance between the posterior
vertebral body tangent lines at the level of the disc space can be used to measure the absolute distance in millimeters (Fig. 39-4). Lateral translation, though much less common, can be measured on the AP view by drawing longitudinal (vertical) lines along the lateral masses.

Vertebral body height loss is another useful measurement that can be performed on the lateral view. Although it might reflect the axial stability of the anterior column, there are no specific criteria on a critical percentage of vertebral body height loss. The anterior and posterior vertebral body heights should be measured at the injured and adjacent uninjured levels. The distances of the uninjured levels should be averaged, and the percentage height loss for the anterior and posterior vertebral bodies calculated (Fig. 39-5B). The relative amount of comminution should be noted, although this has been more clearly categorized in the thoracolumbar spine. Primary fracture lines, as are present with so-called tear-drop fractures, should be outlined.

Compressive flexion (CF) injuries are divided into five stages (Fig. 39-9). The injury is speculated to occur first by flexion of the spine through the facet joints. The anterior column (vertebral body) becomes increasingly compressed and shortened. Eventually, the posterior ligamentous complex fails, noted by interspinous gapping and local kyphosis. With further energy, the facet joints will fail, leading to translational deformity. The distinguishing x-ray features of each stage should be understood. It should be noted that the stages build upon one another, so that, for example, stage 3 lesions also demonstrate the features described for compressive flexion stages 1 and 2.

CF Stage 1: Blunting of the anterosuperior vertebral body margin.

CF Stage 2: Beak appearance of the anterosuperior vertebral body margin, a sagittal vertebral body split may also be present.

CF Stage 3: Oblique primary fracture line that extends from the anterior vertebral body to the inferior end plate This has been subsequently described by other authors as a so-called tear-drop fracture (41).

CF Stage 4: In addition to stage 3 features, posterior translation of the upper vertebra measuring less than 3 mm.

CF Stage 5: Posterior translation of the upper vertebral measuring 3 mm or greater, facet gapping, indicating anterior and posterior ligamentous injury.

Vertical Compression Vertical compressive (VC) lesions are thought to arise from primarily axial loads to the cervical spine. Although this is the author’s description, the final stage of the injury may result from flexion or extension vectors, which ultimately produce posterior or anterior ligamentous injury, respectively (Fig. 39-10).

VC Stage 1: Central superior or inferior endplate fracture.

VC Stage 2: Superior and inferior end plate fractures, sometimes with vertebral body fracture lines that give the appearance of a quadrangular fracture fragment.

VC Stage 3: Vertebral body comminution, with or without retropulsion of fragments (this has been described by others as a burst-type cervical fracture), with or without kyphotic (late flexion type) or translational (late extension type) deformity.

Distractive flexion (DF) injuries are thought to occur from primarily flexion injury vectors that rotate about an axis anterior to the vertebral body. Thus, distraction and failure of the posterior ligaments can occur without significant vertebral body fracture. In this injury group, increasingly higher stages do not always correspond to increasing amount of instability (Fig. 39-11).

DF Stage 1: Facet subluxation, gapping of the spinous process ligaments, indicating failure of the posterior ligamentous complex, with or without some blunting of anterosuperior vertebral body (like compressive flexion stage 1).

DF Stage 2: Unilateral facet dislocation, usually posterior ligamentous complex is intact, rotational deformity.

DF Stage 3: Bilateral facet dislocations, 50% translation of upper vertebral body on lower one.

DF Stage 4: Close to 100% translation of upper vertebral body on lower one, appearance of a so-called floating vertebra.

Compressive Extension (CE) injuries are divided into five stages (Fig. 39-12). They are postulated to start with compression of the posterior elements without failure of the anterior ligaments. Further injury leads to failure of the anterior/posterior ligaments.

CF Stage 1: Posterior arch fracture that may be facet, pedicle, or lamina fracture, with or without rotation that can result in mild anterior translation (these are more commonly referred to as lateral mass fractures).

CF Stage 2: Bilateral lamina fractures, can be multiple levels.

CF Stage 3: Bilateral lamina, facet, pedicle fractures without vertebral body displacement. Though admittedly “hypothetical…having not been encountered” in their review, the injury may be described as a floating lateral mass fracture.

CF Stage 4: as for CF stage 3, with partial anterior vertebral body displacement.

CF Stage 5: as for CF stage 3, with 100% anterior vertebral body displacement.

Distractive extension (DE) injuries, like DF injuries, demonstrate substantial ligamentous injury in lower stages. Initial failure is through the anterior ligaments (Fig. 39-13).

DF Stage 1: Abnormal widening of the disc space, may or may not be avulsion fractures of the anterior vertebral body margin, no posterior translation.

DF Stage 2: DF stage 1 plus posterior translation.

Lateral flexion (LF) injuries occur by compression on one side of the spine. With further energy, the contralateral side can fail under tension.

LF Stage 1: Unilateral uncovertebral vertebral fracture or asymmetric vertebral body compression.

LF Stage 2: Vertebral body or posterior arch fractures with lateral translation or unilateral facet gapping, coronal angular deformity is noted on an AP x-ray.

The system of Allen et al (40) is the most frequently cited and used classification for subaxial cervical spine injuries. Despite this, it has not been validated, retrospectively or prospectively, since its publication in 1982. Intraobserver and interobserver reliability has not been tested, to the author’s knowledge, and the significance of the injury groups on treatment decision making is not yet clear.

Subsequent studies have demonstrated that a wide spectrum of injury patterns can occur from a single mechanism. Torg et al (42,43) found facet dislocations, tear-drop vertebral body fractures, and anterior translational injuries at the C3–4 segment in football players injured during a witnessed axial loading mechanism to the head and cervical spine. According to Allen et al (40), these would have had to occur from distractive flexion, compressive flexion, and either compressive extension or vertical compressive mechanisms, respectively. The disparity between these findings underscores the very complex three-dimensional biomechanical response of the cervical spine to even simple, unidirectional force vectors. Postulating the direction or mechanism of injury, though a common academic exercise,
may not result in accurate or useful information. Perhaps what is more important is determination of the integrity of the ligamentous structures. This is a major, if not the primary, determinant of cervical spine stability (35).

Regardless of the mechanism of injury, vertebral body fractures are readily detected by plain films and CT. Fractures may be simple wedge types, also known as compression fractures, in which there is primarily anterior height loss and no posterior vertebral involvement. Teardrop fractures, as described by Allen and Ferguson as CF stage 3 injuries, demonstrate a characteristic primary fracture that extends obliquely from the anterosuperior vertebral body to the inferior end plate. These can involve varying percentages of the end plate, which may influence the decision to perform a discectomy or corpectomy if surgical treatment is planned. Burst fractures, much like their thoracic and lumbar counterparts, demonstrate extensive vertebral body comminution, varying degrees of height loss, but, most importantly, posterior vertebral body involvement with fragment retropulsion. One area of confusion is the term teardrop burst fractures. Teardrop fractures often have a midsagittal split in addition to posterior translation, which many describe as a burst fracture. Quadrangular burst fractures, similar to that described as VC stage 2 injury by Allen and Ferguson, are sometimes distinguished in the literature from other vertebral body fractures. The clinical significance of this distinction is unknown, because the treatments are often similar. With any of these vertebral body fractures, the posterior ligamentous complex can be disrupted, either by translational, flexion, or rotational forces.

Facet injuries are extremely common. Although Allen and Ferguson have suggested that they occur primarily through distractive flexion mechanisms, it is clear that rotational forces, axial compressive forces, and various other vectors can be involved. Facet fractures can be associated with dislocations or other posterior arch fractures. Reports of facet fractures in the literature generally refer to isolated, unilateral, minimally displaced fractures of varying sizes. Often thought to be benign fractures, they can be associated with ligamentous injury, leading to subluxation and instability. Because of this potential, there is significant controversy regarding their initial treatment. Facet subluxations are the result of facet capsule and posterior ligament disruption. By definition, some portion of the articular surfaces of the involved levels is still in apposition. Facet dislocations can be unilateral or bilateral. These can be described in a number ways, such as perched or locked. By definition, the articular surfaces are no longer in opposition. Cadaveric sectioning studies, as well as intraoperative observations, have indicated that unilateral dislocations can be present without complete posterior ligamentous complex disruption and in many cases may be mechanically stable injuries. There may be some utility in distinguishing facet dislocations from facet fracture-dislocations, in which the facet joint is dislocated, unilaterally or bilaterally, and fractured. With large facet fracture fragments, reduction can be difficult to achieve or maintain by closed methods. In addition, extensive articular process fractures can preclude lateral mass screw placement.

Isolated, unilateral pedicle fractures can suggest rotational instability, which can present late as in the case of isolated facet fractures. For this reason, pedicle and facet fractures are often referred to in the literature as lateral mass fractures. A concomitant lamina fracture and pedicle fracture effectively negates the contribution of the adjacent facet joint to overall cervical stability. This can be described as a floating lateral mass fracture. Such fractures can potentially destabilize two motion segments.

The anterior longitudinal ligament and intervertebral disc can fail in tension (see Fig. 39-41). Without speculating about the mechanism of injury, widening of the intervertebral disc space is a highly suggestive sign that the anterior tension band has been disrupted and that significant spinal instability is present. Small avulsion injuries of the vertebral body can also result in teardrop-shaped fragments and are commonly referred to as extension-type teardrop fractures. The mechanism of injury is most likely extension, with failure of the anterior ligaments being more likely than posterior ligamentous complex disruption. This fracture is typically more common in elderly patients, and should be considered more likely in this age-group, in particular if there is no kyphotic deformity at the level of injury.

The majority of cervical spine fractures can be treated nonoperatively. The most common method of nonoperative treatment is immobilization in a cervical orthosis. In reality, orthoses decrease motion rather than effect true immobilization. Orthoses work through padded contact areas strategically located over subcutaneous bony prominences. Specifically, these are the occiput, spinous processes, scapular spines, acromion processes, clavicles, sternum, and mandible. Interestingly, motion at the occipital-cervical junction is slightly increased by most cervical collars (54).

Soft cervical orthoses do little to decrease spinal motion (54). They are most commonly used after cervical strains or sprains, otherwise known whiplash-type injuries. Soft collars are worn for comfort.

Rigid cervical orthoses can provide varying degrees of immobilization depending on the construct material and the overall design. They are commonly used for emergent in-field cervical spine immobilization. An important design feature of all cervical orthoses is inclusion of a window or cutout over the anterior neck to allow access to a tracheostomy, if present. The most effective means of in-field cervical spine immobilization is strapping the chin and forehead to a rigid spine board (55). With this, addition of a cervical collar offers little additional stability.

There are many manufactured brands of rigid cervical collars. An x-ray study demonstrated that the NecLoc restricted more sagittal (flexion-extension), axial rotation, and lateral bending motion than the Miami J, Philadelphia, Aspen, or Stiffneck devices (3). The Miami J restricted extension better than the Philadelphia and Apsen collars (3). Despite these detected x-ray differences, the clinical superiority of one device over another has yet to be demonstrated.

There are a number of complications associated with use of cervical collars. Skin breakdown at bony prominences, in particular the occiput, mandible, and sternum, can occur. Up to 38% of patients with severe closed head injuries can develop skin complications with prolonged use (56). This places a priority on early injury exclusion (“clearing’’) of the cervical spine so that unnecessarily protracted collar wear can be avoided.

Cervicothoracic orthoses (CTO) include components that extend to the upper thorax. In this way, the braces provide more effective immobilization than simple cervical collars in all planes (54,57). Types include the sterno-occipito-mandibular immobilizer (SOMI), Minerva brace, and the Yale device. Mechanical studies have demonstrated that between 79% to 87% of sagittal motion, 75% to 77% of axial rotation, and 51% to 61% of lateral bending motion is restricted in these orthoses (58,59,60). The disadvantages of CTO braces are that they are more difficult to apply/remove and produce high resting pressures on the chin and occiput (61). Like simple cervical collars, CTOs have demonstrated increased motion at the occipitocervical junction. CTOs alone are not an effective means of immobilizing the spine below the C7–T1 junction.

Cervical traction can be applied to achieve a number of different goals. Traction can be used to help provide temporary immobilization of unstable cervical spine injuries (62). This may be of particular benefit when transferring patients to and from institutions or for patients awaiting operative stabilization.

Traction is most commonly used to realign or reduce cervical spine fractures or dislocations. This is achieved by application of gradually increasing amounts of weight in conjunction with frequent x-ray examination. Traction should not be used in cases in which occipitocervical instability or dissociation is present or suspected. Furthermore, traction should not be used to realign a subaxial spine injury if there is a concomitant type IIA traumatic spondylolisthesis of the axis. As such concomitant injuries may not always be initially recognized, light traction (5 to 10 lb) should be applied first, followed by a lateral x-ray to look specifically for signs of abnormal distraction between the cervical spinal segments or occipitocervical junction (63).

Two of the more common devices used to apply traction to the cervical spine are Gardner-Wells tongs and the halo-ring. In general, Gardner-Wells (G-W) tongs can be applied more quickly and easily than a halo-ring. However, G-W tongs are temporary devices that are usually removed after surgical stabilization. Traction is applied using a halo-ring if the planned definitive treatment includes halo immobilization.

Traction can be used to realign and reduce a variety of subaxial spinal injuries. Realignment of an injury can provide some, if not complete, spinal canal decompression. An example is reduction of bilateral facet dislocations. Fracture fragment retropulsion, such as that produced with burst-type injuries, is usually associated with vertebral body height loss and comminution. With application of cervical traction, ligamentotaxis can partially reduce these fragments and produce some degree of canal decompression. This maneuver, however, relies on sufficient continuity of the posterior longitudinal ligament. Segmental, traumatic kyphotic angulation can also be improved with application of traction.

Some authors advocate use of large amounts of weight (up to 140 lb) to reduce facet dislocations (64,65), however overdistraction should be avoided as it can lead to neurological decline (66,67). The biomechanical strength of the traction device is of particular significance if such practices are employed. Steel tongs are stronger than titanium-alloy or carbon-fiber tongs, though the latter two are MRI-compatible (68). Because pullout strength is diminished with repeat usage, it has been suggested that steel cranial tongs should be recalibrated after frequent usage (69). According to their manufacturers’ recommendations, carbon-fiber tongs should only be used once.

With improved techniques of rigid internal fixation of the cervical spine, halo immobilization has become a less popular form of treatment for subaxial cervical spine injuries. As discussed above, halo-based traction can be used, in particular when definitive treatment with a halo vest is planned. Halo fixation may also be a useful method of temporary stabilization of highly unstable cervical spine injuries in multiply injured patients who must undergo diagnostic and nonspine surgical procedures (78). The use of halo fixation for cervical facet dislocations has demonstrated a high rate of persistent instability and poor anatomic alignment (79,80,81). Varying rates of success have been documented for other types of subaxial cervical injuries (82,83,84,85). Despite this, halo vest immobilization remains a viable, minimally invasive method of stabilization of unstable cervical spine injuries for patients who might otherwise have contraindications to open surgical methods. Halo fixators are the best nonoperative method of resisting rotational and translational forces (86), while they are relatively ineffective at resisting axial compressive loads.

The optimal time to perform surgery, particularly in patients with neurologic deficits, still remains unclear. The two most commonly proposed benefits of earlier versus later surgery are improved rates of neurologic recovery and improved ability to mobilize the patient without concern of spinal displacement.

A number of animal studies have suggested a significant benefit to earlier decompression after acute spinal cord injury. In perhaps the most clinically representative spinal cord injury model reported, Dimar et al (90) found a nearly linear relationship between rate and extent of neurologic recovery and time to decompression in rats with experimentally induced incomplete spinal cord injuries with 35% canal compromise. Other animal studies using less clinically representative models, such as inflatable intracanal bladder compression, spinal cord cinching with cables, and cauda equina level injuries, have demonstrated similar time-dependent recovery rates.

To date, there is little human clinical evidence to support that early surgical decompression and stabilization improves
neurologic recovery rates. This may largely be the result of disagreement upon what is considered early versus late surgery. In the only randomized, prospective, controlled trial found in the literature, surgery performed for cervical spinal cord injuries less than 72 hours versus more than 5 days from the injury demonstrated no significant difference in motor scores at final follow-up (91). Notwithstanding the possibility of beta error, it is also possible that surgery within a critical time period within the first 72 hours to 5 days might have made a difference. Supportively, other nonrandomized prospective studies have demonstrated that surgery performed within 8 hours or within 24 hours from injury did not result in a better neurologic outcomes (92,93). In one interesting report, performing surgery within 8 hours from the time of injury was feasible in only 10% of cases (94).

One retrospective study found surgery performed within 72 hours had significantly better neurologic outcomes than surgery performed after 10 days from the injury. Many practitioners believe surgery should be delayed to allow for optimal medical stabilization of the patient and resolution of initial spinal cord swelling, hypothesizing that early surgery may be potentially detrimental. A number of clinical series have demonstrated that, in the least, surgery performed as soon as 8 hours does not appear to increase the rate of complications or lead to neurologic decline (94,95).

Patients with cervical compression fractures without posterior ligamentous injury can be treated nonoperatively. For injuries at C3 to C6, a rigid cervical collar usually suffices. For injuries of C7 or T1, a cervicothoracic brace may offer better control. Ultimately, if the injury is stable, the orthosis minimizes motion at the fracture site, which can decrease pain and facilitate resolution of muscular spasm. Attempts to correct minor kyphotic deformities are usually fruitless and are in reality unnecessary. A lateral x-ray should be obtained with patient sitting up or standing before discharge from the hospital, because this mechanical “litmus” test can sometimes demonstrate instability. The fracture is usually healed by 3 months, at which time flexion-extension views under the patient’s own control should be obtained to rule occult instability, either at the level of the injury or at a distant site, such as C1–C2, which may have been undetected at the time of initial presentation.

Surgical stabilization should be considered for patients with evidence of posterior ligamentous injury. This is suggested by segmental kyphosis greater than 11 degrees or substantial amounts of vertebral body wedging. Correlative numbers for the percentage of height loss and posterior ligamentous complex injury in the cervical have not yet been established. MRI evidence for ligamentous injury should also be considered. Before surgery, cervical traction with a slight extension vector can be used to realign the fracture. This is best performed in the awake patient. Though described in more detail for patients with facet
dislocations, a herniated disc may be rule out based on the MRI as well. If the spinal canal is clear, a single level posterior fusion with instrumentation can be performed.

Nonoperative treatment might be considered in neurologically intact patients with little vertebral body comminution and only
the mildest degree of canal compromise. There should be minimal kyphotic deformity (neutral or <5 degrees of relative kyphosis) and no indication of posterior ligamentous injury. Because of the potential for vertebral body collapse, it is the author’s preference to treat patients in a halo vest or rigid CTO. X-rays should be obtained with the patient standing or sitting, before discharge, and rigorous comparisons made to supine films. High-quality films are crucial. Patients should be followed weekly for the first month.

Patients with neurologic deficit, regardless of the integrity of the posterior ligamentous complex, should be surgically decompressed and stabilized. Retropulsed vertebral body fragments are most easily accessed through a direct anterior approach. The injured vertebral body should be removed, and the spinal canal fully decompressed. Intraoperative traction can help realign the spine if significant deformity exists (Table 39-4). The anterior column should be reconstructed with a bone graft or strut. It is the author’s preference to insert a rigid titanium mesh cage filled with salvaged bone in addition to cancellous iliac crest autograft. Bone should be tightly packed into the cage, which may also enhance the surface area contact with the end
plates. An anterior cervical plate is then applied to restore anterior stability. If the posterior ligamentous complex appears to be disrupted, it is the author’s preference to perform a posterior instrumented fusion, either during the same operation or delayed until the patient can better tolerate it.

There is a definite role for nonoperative treatment of cervical tear-drop fractures. Minimally displaced fractures with little kyphosis and no posterior ligamentous complex injury are stable. They can be treated in a rigid cervical collar or CTO, depending on the level of injury. Halo treatment can be used to treat unstable fractures. The apparatus can help realign fractures, which can effect canal clearance in patients with spinal cord injury. Importantly, halo treatment for unstable tear-drop fractures results in inferior x-ray results than anterior surgery, though neurologic and clinical outcome scores appear to be comparable (41). The halo should be kept in place for 3 months, provided acceptable alignment has been maintained. After the fixator has been removed, a flexion-extension x-ray should be obtained to confirm that stability has been achieved.

In patients with neurologic deficit, anterior corpectomy is usually performed to remove the offending retrolisthesed vertebral body (Table 39-5). This is followed by anterior strut grafting and rigid plating (Fig. 39-34). In some cases, an unstable tear-drop fracture can occur in a neurologically intact patient. In these cases, anterior surgery can include a nondecompressive corpectomy, which entails resection of the majority of the vertebral body back to, but not through, the posterior wall. This is best reserved for injuries without retrolisthesis. Some surgeons prefer to perform a single-level discectomy or partial corpectomy. However, extensive end-plate fracture appears to be a risk factor for failure of anterior alone constructs (112). If the initial injury demonstrated a large amount of translational deformity (>3 to 3.5 mm) and the facet joints appear to be gapped or widened, posterior surgery can be performed to provide additional stability.

In rare cases, posterior surgery alone can be elected (Fig. 39-35). This should be reserved for patients who are neurologically intact, have minimal vertebral body height loss, and have less than 30% inferior end plate involvement. A potential advantage

of this approach is that the fusion can potentially be limited to a single motion segment.

Most facet fractures are minimally displaced and can be considered mechanically stable. Such patients can be treated in a rigid
cervical collar for a period of 6 to 12 weeks, monitored by frequent x-ray examination. It is important to recognize, however, that they can be associated with occult unstable ligamentous injury (demonstrating little if any translational deformity at initial presentation). Many authors recommend routine MRI examination to rule out significant ligamentous damage. Disruption of the anterior longitudinal ligament and intervertebral disc are thought to play in important role in late displacement, although posterior ligamentous complex injury can also occur. Neurologic injury associated with the fracture is rare and, if present, is usually limited to a mild single root radiculopathy that often resolves without formal decompression. Flexionextension views should be obtained after the completion of collar immobilization to ensure stability. Although late displacement is considered to be an indicator of instability by most, it has been observed that this can autostabilize with time. It is unclear what the long-term consequences of fixed deformity on the overall clinical outcome.

Because of the inconsistent prevalence of ligamentous instability, some have aggressively recommended operative treatment for nearly all facet fractures. Others have focused the indications by employing judicious use of flexion-extension views or MRI. A validation study of the predictive value of MRI on displacement of facet fractures has yet to be published.

Either anterior or posterior surgery can be performed to treat facet fractures. Anterior surgery usually consists of a single-level interbody fusion with a plate (Fig. 39-38). The advantages of this approach are that the fusion rates are consistently high, a perceived lower infection rate than posterior approaches, and the ability to fuse only one motion segment. It is important to choose the correct level to fuse, which should be based on the facet joint involved, as opposed to the numerical level of the articular process fracture.

Posterior surgery entails stabilization and fusion and can be performed in a variety of manners. The most mechanically stable constructs are achieved with the use of bilateral lateral mass screws connected by rods or plates. However, bilateral screw placement can be precluded with large articular process fractures, which may necessitate instrumentation of the next adjacent nonfractured lateral mass. Although these issues are avoided by using interspinous process wiring, the stability of the construct is inferior and is precluded with lamina fractures. A combination construct, using an interspinous process wire and a unilateral lateral mass plate is another option (Fig. 39-39).

There is sparse literature concerning the operative outcomes of unilateral facet fractures. Because they are considered to be rotational injuries, they are often reported along with unilateral pedicle fractures and facet dislocations. In a combined retrospective review of posterior surgery and prospective analysis of anterior surgery, anterior surgery resulted in a superior x-ray outcomes (113).

The role of nonoperative treatment for facet dislocations is minimal. If it is elected, it should be reserved for unilateral facet dislocations in patients without any signs of neurologic injury or for those who are too sick to undergo surgery. Even in these cases, cervical orthoses are not usually considered adequate treatment. The inability of halo-fixation to treat facet dislocations has been demonstrated by Sears and Fazl (80), with more than 50% of patients exhibiting persistent instability after treatment.

Frequent x-ray examination is important. Evidence of autofusion

of the dislocated facet joint or, less frequently the disc space, is a sign of healing. After a course of about 3 months of halo immobilization, the device should be removed and the flexion-extension views should be obtained to confirm stability. Surgical fusion should be considered if instability is present.

The optimal treatment of unilateral or bilateral facet dislocations remains unclear. A number of different methods can be successful. It is important to recognize the limitations, advantages, and disadvantages of each. In a number of circumstances, one treatment method may be preferred over another.

The safety of closed reduction as well as the role of MRI have been discussed in previous sections. In summary, closed reduction appears to be a safe practice in the awake, cooperative patient who can be serially examined. If the patient does not fulfill these criteria, particularly in the incomplete spinal cord–injured patient, prereduction MRI is highly advised to detect a herniated disc (Table 39-7).

In the patient who has undergone successful closed reduction before operation, anterior or posterior surgery can be performed. There is no clear evidence of the superiority of one approach over another. Conceptually, posterior stabilization more directly addressed the instability caused by posterior ligamentous complex disruption. A postreduction MRI is usually performed before surgery (Fig. 39-40). Most surgeons feel uncomfortable
performing posterior surgery alone in the presence of a herniated disc after reduction in a neurologically intact patient. Anterior surgery can be performed to carefully remove the herniated disc.

There are pitfalls to both approaches. Anterior discectomy and fusion involves resection of one of the major remaining stabilizing structures—the anterior longitudinal ligament. In these approaches, one can easily overdistract the disc space while sizing and placing grafts. This can leave the facet joints gapped posteriorly, which can obviate posterior column load
sharing. Improper fit of the interbody device can place greater demands on the anterior plate and screws, which can lead to early hardware failure.

During posterior surgery, in particular lateral mass screw placement, injury to the adjacent intact facet joints should be avoided. Although posterior compression can aid in articular apposition, overcompression can increase intradiscal pressure. This can cause intraoperative disc herniations that have the potential to cause additional neurologic injury. Intraoperative spinal cord monitoring is useful in such situations.

Combined anterior and posterior surgery may also be performed. This is usually reserved for patients with more severe or missed injuries with fixed deformities. It is the author’s preference to perform anterior surgery followed by posterior stabilization for patients with highly unstable bilateral facet dislocations. If the facets are gapped or kyphosis remains after anterior surgery, posterior instrumentation is performed to avoid catastrophic construct failure. Anterior and posterior surgery are rarely necessary for unilateral dislocations.

Feldborg Nielson et al (114) found anterior fusion to result in better pain relief than posterior wiring without fusion for facet dislocations. The authors attributed this to persistent residual motion in the unfused cases. Razack et al (115) performed single-level anterior fusion with a titanium locking plate in 22 patients with bilateral facet fracture-dislocations. At an average follow-up of 32 months, only one case of instrumentation failure was reported, though all patients eventually had solid fusion and achieved stability. Vital et al (116) found that 91 bilateral facet dislocations could be successfully reduced using traction in 43%, manipulation under general anesthesia in 30%, and anterior surgical reduction in 27% of cases. Anterior discectomy and plate fixation was performed in all patients as definitive treatment. Shapiro (117) reported results of a series of 24 patients with unilateral facet dislocations. Although halo treatment was attempted in two patients who had undergone successful closed reduction, they had recurrent dislocations and eventually underwent surgery. Thus, all 24 patients underwent posterior interspinous process wiring. Fusion was successful in 23 (96%) of cases. In a later study by the same group (117), comparable clinical results were reported with interspinous wiring with lateral mass plate versus facet wiring with iliac crest for unilateral facet dislocations. Greater maintenance of kyphotic correction was observed in the plate group compared to the facet wire and iliac crest graft group. In a review of their cases, Beyer and Cabenela (72) and Beyer et al (73) found that unilateral facet dislocations or fracture-dislocations were better treated with operative
intervention, because nonoperative management led to an unacceptably high rate of late pain and instability. In a series of 36 patients, 36% of patients treated with a halo-fixator demonstrated anatomic alignment compared to 60% in the operative group.

These injuries are usually highly unstable. Thus, nonoperative treatment is rarely considered a definitive treatment. In cases in which the patient is too sick to undergo surgery, and is intact, a flexion moment can reapproximate the vertebral bodies. If reduction can be maintained, bony ankylosis of the disc space can occur. Flexion-extension views must be obtained to confirm adequate stability.

Anterior tension band disruption, consisting of discontinuity of the anterior longitudinal ligament, intervertebral disc, and possibly the posterior longitudinal ligament, is an ideal indication for anterior surgery. An anterior discectomy and plating with fusion can restore the tension band in its mechanical sense, just as posterior fixation addresses posterior ligamentous complex disruption with bilateral facet dislocations. To the author’s knowledge, there are no dedicated series of surgical fixation of anterior tension band injuries. Because the proposed mechanism is similar, extension-type cervical teardrop fractures can be successfully treated with anterior or posterior surgery.

The ankylosed spine should be considered as a long bone rather than the normal spinal column of articulating vertebrae. Bridging osteophytes, whether marginal as in ankylosing spondylitis (Fig. 39-42A) or nonmarginal and flowing as in diffuse idiopathic skeletal hyperostosis are the x-ray hallmarks of these diseases. They effectively fuse the spine into a solid, continuous piece of bone. As is true for an adult long bone (for example, femur) fracture, if there is a break in one region, it is likely to have propagated through both the anterior and posterior elements (Fig. 39-42B,C). Thus, fractures in patients with diffuse idiopathic skeletal hyperostosis or ankylosing spondylitis are almost universally unstable and should be treated as such.

The key to caring for patients with diffuse idiopathic skeletal hyperostosis or ankylosing spondylitis cervical fractures is early recognition in order to avoid catastrophic neurologic decline. Diagnosis is often missed, resulting in a very high rate of neurologic deficits in previously intact patients (119). Among all cervical fractures, ankylosing spondylitis has been demonstrated to be a significant risk factor for neurologic worsening (120,121). Patients should be immobilized as soon as the diagnosis is made, admitted to the hospital, and placed on strict log-roll precautions until definitive management has been decided. Patients with ankylosing spondylitis, with or without fracture, are predisposed to developing neurologic decline from epidural hematomata.

There are very little data available concerning the optimal treatment. There are advocates of nonoperative treatment. Halo fixation has been used to successfully maintain alignment and prevent displacement in patients with and without neurologic deficit (122); however, skin breakdown remains a concern, particular in patients who will remain recumbent for an extended period. Both anterior and posterior surgery have been advocated as well (123,124). Key principles in management should be kept in mind, regardless of the approach taken.

When placing patients in traction, the preexisting kyphotic deformity that is often present must be considered. Thus, inline traction, as is effectively used for most other cases, can lead to gross displacement and neural impingement. The awake patient should help guide the position of comfort when applying traction vectors. In cases of ankylosing spondylosis, kyphotic deformities usually necessitate a flexion vector.

This same concept is important when considering operative fixation. Both anterior and posterior implants should be contoured to fix the patients specific anatomy. Unless a concomitant osteotomy is being planned at the same time, internal fixation of the fracture should maintain or approximate the preinjury posture of the neck. This can often necessitate unusual implant contouring (Fig. 39-42D).

Patients often present with complete or incomplete spinal cord injury without x-ray signs of a frank injury, fracture, or ligamentous instability. Underlying cervical stenosis is often present, which can arise from degenerative changes or congenitally narrow canal (Fig. 39-43). This apparently increases the risk for neural injury with abrupt movements of the neck that otherwise are not severe enough to cause a mechanically destabilizing fracture or ligament injury.

There are few data concerning the best treatment. Nonoperative treatment can include a period of observation, in particular for patients with central cord lesions, in which a high percentage will have nearly complete resolution of their neural deficits. Long-term treatment is influenced by the presence of persistent signs or symptoms or myelopathy. Immobilization in a rigid cervical collar can be useful to avoid the extremes of motion, although patients usually will not voluntarily move their neck beyond these limits.

The role and timing of surgery is controversial. In a retrospective study, early surgery for patients with incomplete spinal cord injuries resulted in earlier neurologic recovery, with better motor scores at 1 month and 6 months (125). However, at 2 years, there was no statistically significant difference between the operative and nonoperative groups (125).

In the author’s practice, operative treatment is delayed. Following resolution of spinal shock, patients are observed for signs of neurologic recovery over a period of 2 to 3 days. If there are no signs of neurologic return, surgical decompression is performed with the hopes of improving the rate of recovery. Provided that neurologic examination demonstrates improvements in motor strength, surgery is postponed. Many patients will have complete motor and sensory recovery but demonstrate residual signs and symptoms of myelopathy (such as walking imbalance, diminished finger dexterity). In such cases, a decompressive procedure is performed electively weeks after the injury.